VOLUME 2
COMMENTS ON THE
ADVANCE NOTICE OF PROPOSED RULEMAKING
"OZONE - DEPLETING CHLOROFLUOROCARBONS
PROPOSED PRODUCTION RESTRICTION"
BY THE
ENVIRONMENTAL PROTECTION AGENCY
SUBMITTED BY
E. I. DU PONT DE NEMOURS & COMPANY (INC.)
WILMINGTON, DELAWARE
JANUARY 5, 1981
-------
VOLUME 2
Table of Contents
Page
X.. APPENDICES
A. Description of Major Pertinent Reports and
Submissions on the Chlorofluorocarbon/Ozone
Issue A-l-13
B. The Du Pont Development Program on the
Alternatives to Commercial Chlorofluoro-
carbons B-l-15
C. The Energy Consequences of Chlorofluoro-
carbon Regulation (Battelle Report) C-1-2S
D. A Comparison of Some of the Principal
Findings of the November 1979 National
Academy of Sciences' Report and the
October 1979 United Kingdom Department
of the Environment's Report D-l-5
E. Chlorofluorocarbons and Ozone - The
Science E-l-69
F. Effects of Ozone Depletion
1. Human Skin Cancer; Review by Professor
Frederick Urbach, M.D. F-l(1-199)
2. Measurement and Instrumentation; Review
by Dr. William H. Klein. F-2(l-19)
3. Agricultural Crops; Review by
Professor R. Hilton Biggs. F-3(l-13)
4. Aquatic Ecosystems; Review by
Dr. David M. Damkaer. F-4(l-36)
NOTE: Appendices G-L and Section XI-Bibliography-appear in
Volume 3.
-------
APPENDIX A
DESCRIPTION OF MAJOR PERTINENT
REPORTS AND SUBMISSIONS ON THE
CHLOROFLUOROCARBON/OZONE ISSUE
A-1
-------
Reports & Submissions
In this appendix we provide a listing and brief
abstract of important publications, reports, updates and
submissions on the Chlorofluorocarbon/Ozone Issue. While this
list by no means exhausts the available literature, it does
incorporate major documents which we believe should be reviewed
as part of any evaluation of this complex issue. Full copies of
the cited documents which are not already listed by EPA as being
in the public record are submitted as part of the Du Pont ANPR
comments.
A-2
-------
Reports & Submissions
1. Flurocarbon/Ozone Update, Du Pont, September, 1977.
This was the first in a continuing series of bulletins
published by Du Pont to provide customers with an overview of
the most current developments in the CFC/0., controversy.
This edition includes an assessment of the signi-
ficance of the ozone depletion theory, a discussion of the
September, 1976 NAS report, a review of technical developments
and future technical needs, and a brief look at the inter-
national situation.
2. Non-Aerosol Propellant Uses of Fully Halogenated
Halocarbons, Du Pont, March, 1978.
A comprehensive response to EPA's request for
information on non-aerosol propellant uses of fully halogenated
halocarbons, this Du Pont submission supplements information
provided to EPA in ten previous verbal presentations at EPA
public meetings.
The report describes the essentiality or benefits of
the major nonpropellant uses of chlorofluorocarbons. It also
provides a status report on the alternative fluorocarbon
development program and the possibility for control of
inadvertent CFC emissions.
3. The Ozone Controversy and Its Relationship to
Refrigeration and Air Conditioning, ARI, June, 1978.
This bulletin was published by the Air Conditioning
and Refrigeration Institute. It addresses principle issues in
the CFC/0- controversy, including: status of theory, essen-
tiality of refrigeration and air-conditioning, CFC emission
A-3
-------
Reports & Submissions
control, risks of regulating, economic scope, action of other
countries and regulatory options.
4. Fluorocarbon/Ozone Update, Du Pont, October, 1978.
The second in a series, this bulletin reviews the
CFC/0, issue and briefly explains EPA's Phase I regulations
banning aerosol propellant use of CFCs. It also offers a
simplified discussion of recent technical developments in the
science and comments on the program for future research.
5. Non-Propellant Uses of Fully Halogenated Halocarbon -
No 2, Du Pont, June, 1979.
This Du Pont report updates and supplements
information provided to EPA in the March 15, 1978 report of the
same title.
The specific areas of mention are the status of the
CFC/0., Depletion Theory, development programs on alternatives
to commercial CFCs, potential for CFC emission control and
refrigerant emission control through the use of Dytel® leak
detective.
6. The Fluorocarbon Industry Research Program and Current
Uncertainties in the Ozone Depletion Theory, CMA, November,
1979.
This report summarizes the research programs being
sponsored by the Chemical Manufacturers Association Fluorocarbon
Project Panel (CMA/FPP) to reduce uncertainties in the science
of the theory of depletion of ozone by CFCs. The summary
identifies major uncertainties and their significance, reviews
the pertinent research program and comments on the expected
results and timing.
A-4
-------
Reports & Submissions
7. Chlorofluorocarbons and Their Effect on Stratospheric
Ozone (Second Report), UK DOE, October, 1979.
This report is a comprehensive review of UK policy in
light of the state of knowledge of the science of ozone
depletion. Part 2 of the report is a detailed analysis of the
CFC/03 science prepared by "The Stratospheric Research
Advisory Committee" (STRAC). The committee consisted of experts
from the UK Meteorological office, universities, industry and
government laboratories.
8. Stratospheric Ozone Depletion by Halocarbons: Chemistry
and Transport, NAS, November, 1979.
This report was prepared by the Panel on Stratospheric
Chemistry and Transport (PSCT) of the National Academy of
Sciences (NAS). It is an update on the effects of halocarbons
on the stratospheric ozone layer based on available data at the
time of preparation. It also assessed the uncertainty limits of
predictions based on those available data.
9. The National Academy of Sciences (NAS) and U. K.
Department of the Environment (DOE) Reports, Du Pont,
November, 1979.
Du Pont prepared this comparison of the first eight
principle findings of the NAS report with quotations from the UK
DOE report addressing the same subject matter.
10. The Stratosphere; Present and Future, NASA Reference
Publication 1049, December, 1979.
The Clean Air Act Amendment of 1977 requires that the
National Aeronautics and Space Administration report its
A-5
-------
Reports & Submissions
assessment of the current state of knowledge of the stratosphere
to both the EPA and Congress biannually.
This report is the first in that series. Summary
documents of seven working groups organized by NASA to assess
the state of knowledge in different areas of the science were
assembled into this report.
11. Protection Against Depletion of Stratospheric Ozone by
Chlorofluorocarbons, NAS, December, 1979.
This report is a two part report of the work of two
committees of the National Academy of Sciences.
Part I - Causes and Effects of Stratospheric Ozone
Depletion.
Prepared by NAS Committee on Impacts of
Stratospheric Change (CISC), it is often referred
to as the CISC Report. It examines the effects,
including both biological and climatic effects,
which may occur as a result of the predicted ozone
depletion described in an earlier NAS report by
the Panel on Stratospheric Chemistry and Transport.
Part II - Alternatives for the Control of Chlorofluoro-
carbon Emissions and Options for
Implementation.
The NAS Committee on Alternatives for the
Reduction of Chlorof luorocarbon Emissions (CARCE)
examines the alternatives, costs, feasibility and
timing of possible methods for reducing chloro-
fluorocarbon emissions.
A-6
-------
Reports & Submissions
12. Comments on the National Academy of Sciences Report;
"Stratospheric Ozone Depletion by Halocarbons; Chemistry
and Transport", Du Pont , January, 1980.
A Du Pont submission to EPA, this report comments on
the uncertainty estimates and various technical points in the
science, and provides a comparison of the NAS report by the
Panel on Stratospheric Chemistry and Transport (PSCT) with the
UK DOE 1979 report.
13. Fluorocarbon/Ozone Update, Du Pont, January, 1980.
Du Font's third update in this series focuses pri-
marily on the NAS PSCT report. The report's treatment of
uncertainties and the impact of the report are discussed.
The update also compares comments in the NAS report
with comments from the UK DOE report on the same subjects.
14. An Overview of Industry Efforts to Investigate the
Potential for Chlorofluorocarbon (CFC) Emission Reduction,
Du Pont, February, 1980.
This update prepared for the EPA reviews the
activities in various industries to reduce emissions of CFCs.
Industries discussed include mobile air-conditioning, flexible
polyurethane and packaging foam, solvents, commercial
refrigeration and air-conditioning, industrial gas sterilization
and liquid food-freezing.
A-7
-------
Reports & Submissions
15. The "Importance of Chlorofluorocarbons and Polyurethane
Foams, SPI, March, 1980.
Published by the Urethane Division of the Society of
the Plastics Industry, this bulletin addresses the use of CFCs
in foams, the essentiality of the products, the lack of suitable
alternatives to CFCs and emission control of CFC processes.
16. A Critique of the Rand Corporation Draft Report, "Economic
Implications of Regulating Chlorofluorocarbon Emissions
from Non-Propellant Applications", Du Pont, March, 1980.
Du Font's critique discusses the Rand report's
strengths, data limitations, analytical assumptions, limitations
on the use of the findings, and areas needing clarification and
further work. Rand's treatment of individual uses of CFCs and
the theoretical, methodological, practical and legal aspects of
the economic incentive regulatory approaches are examined.
17. The Du Pont Development Program on Alternatives to
Commercial Chlorofluorocarbons, March, 1980.
This submission reports on the status of the Du Pont
alternatives program designed to identify and develop
fluorocarbon compounds that could be used in place of the CFCs,
but which would not contribute significantly to potential ozone
depletion.
A-8
-------
Reports & Submissions
18. Comments on the December, 1979 Report by the National
Academy of Science's Committee on Alternatives for the
Reduction of Chlorofluorocarbon Emissions (CARCE),
Du Pont; April, 1980.
Du Font's comments on the CARCE report are divided
into two categories: a review of general concepts underlying
the CARCE report and a review of the data and conclusions
developed by the CARCE report.
19. Fluorocarbon/Ozone Update, Du Pont, April, 1980.
Du Font's fourth edition of the series provides brief
discussions of the NAS CISC and NAS CARCE reports and includes
comments on international activity.
20. Summary; Research Program on Effects of Chlorofluoro-
carbons on the Atmosphere, CMA, May, 1980.
This report is Revision 13 of a summary of on-going
research programs, sponsored by the Chemical Manufacturers
Association, on effects of chlorofluorocarbons in the atmosphere.
These industry-sponsored programs are focused on
developing information needed to fill gaps in the existing
scientific knowledge of the potential effects of CFCs on ozone
and any subsequent environmental impact.
21. Energy Consequences of Chlorofluorocarbons Regulation,
Battelle, May, 1980.
Battelle prepared this comprehensive analysis of the
energy penalties of regulation of chlorofluorocarbons for
Du Pont.
A-9
-------
Reports & Submissions
The report analyzed four major uses of CFCs:
automotive air-conditioning, refrigeration, insulating foams and
liquid food-freezing.
22. Comments on the December, 1979 Report by the Committee on
the Impacts of Stratospheric Change (CISC) of the National
Academy of Sciences, Du Pont, May, 1980.
In this submission to EPA, Du Pont comments on the
assessment of the problem, the general conclusions and the key
findings of the report by the NAS Committee on Impacts of
Stratospheric Change (CISC).
23. Analysis of the Conclusions of the British and American
Reports on the Effect of Chlorofluorocarbons on the
Atmospheric Ozone, EEC, June, 1980.
This report to the EEC, prepared by G. Brasseur,
focused primarily on: (1) an analysis of various reports on the
CFC/OO controversy, including their conclusions, certainties
and uncertainties, and (2) developing a list of projects to
reduce the uncertainties in the next five years.
Side-by-side comparisons of the NAS Report and the UK
Report are included as well as an in-depth analysis of the
uncertainties in the science.
24. Chlorofluorocarbons and the Environment, EEC, June, 1980.
This report by the EEC Commission issued to the EEC
Council was a re-examination of the scientific data and a
re-examination of economic and technical data on CFC use in
member nations.
A-10
-------
Reports & Submissions
The re-examination of the science was based on Dr.
Brasseur's [Brasseur, 1980] report, highlighting only his major
findings.
25. Some Concerns with Recent EPA Communications on the
Chlorofluorocarbon/Ozone Issue, Du Pont/ June, 1980.
This document and a transmittal letter from C. N.
Hasten, Director of Du Font's "Freon" Products Division,
discusses concerns over inaccuracies and overstatements made by
EPA in a number of communications.
The document discusses various areas of the
controversy including: CFC use projections, scientific
uncertainties, detection of ozone depletion and depletion
potential of CFC-22.
26. An Assessment of the Chlorofluorocarbon/Ozone Problem,
CMA, June, 1980.
Prepared by the CMA Fluorocarbon Project Panel, after
a review of the data base used by groups commenting on the
subject of ozone depletion, this work examines the relative
validity of one-dimensional modeling and ozone trend analysis,
through an assessment of the uncertainties of each.
27. Fluorocarbon/Ozone Update, Du Pont, June, 1980.
In the fifth issue of this series, Du Pont reviews its
development program for alternatives to commercial chlorofluoro-
carbons. Included in the discussion are: 1) criteria for
alternatives, 2) a listing of chlorofluorocarbon alternatives,
3) the program status and plans, and 4) the timetables.
A-ll
-------
Reports & Submissions
28. Uncertainties in Chlorofluorocarbon Effects and
Stratospheric Ozone, SRI, July, 1980.
EPA sponsored a workshop, hosted by SRI, to evaluate
the critical issues and uncertainties that were hindering EPA's
ability to make a fully supportable decision on further CFC
regulation.
Eleven participants discussed the areas of CFC
releases, atmospheric transport and chemistry, climatic effects,
human health effects, biological/ecological effects and
economics.
SRI compiled the results of the workshop into this
report.
29. Research Program Directed by the Chemical Manufacturers
Association Fluorocarbon Project Panel, Progress Report,
CMA, July, 1980.
A progress report on the research programs directed by
the Chemical Manufacturers Association Fluorocarbon Project
Panel was submitted to UNEP in July, 1980.
Programs reviewed included Atmospheric Lifetime
Experiment, Ozone Trend Analysis, Atmospheric Measurements,
Modeling and Atmospheric Chemistry.
30. Fluorocarbon/Ozone Update, Du Pont, July, 1980.
In the sixth of this series, Du Pont provides an over-
view of the Battelle report, "Energy Consequences of Chloro-
fluorocarbon Regulation."
A-12
-------
Reports & Submissions
31. Uncertainties and Benefit-Cost Analysis of CFC Control,
August, 1980.
M. J. Bailey, of the University of Maryland, assembled
data on the significant effects of continued CFC emissions. His
report also reviews the theory, the actual technical
uncertainties and offers an appraisal of the reliability of data
and derived estimates. An assessment of the risk against the
likelihood of narrowing the uncertainties is also included.
32. Fluorocarbon/Ozone Update, Du Pont, October, 1980.
The seventh, and most recent, issue of the series
offers an overview of the theory and a review of the recent
developments in the science underlying the theory.
A-13
-------
APPENDIX B
THE DU PONT DEVELOPMENT PROGRAM
ON ALTERNATIVES
TO COMMERCIAL CHLOROFLUOROCARBONS
B-l
-------
Alternatives Program
Table of Contents
PAGE
1. INTRODUCTION 3
2. CURRENT COMMERCIAL CHLOROFLUOROCARBONS 4
3. PROPERTIES OF COMMERCIAL CHLOROFLUOROCARBONS 4
4. CRITERIA FOR ALTERNATIVE PRODUCTS 5
5. SCOPE OF DU PONT ALTERNATIVES PROGRAM 5
6. ALTERNATIVE CANDIDATES BY POTENTIAL APPLICATION 7
a. Refrigerants 7
b. Foam Blowing Agents 9
c. Cleaning Agents 10
d. Liquid Food Freezant 11
e. Sterilant Gas 11
7. PROGRAM STATUS AND PLANS 11
8. TIMETABLE FOR COMMERCIALIZATION 13
9. SUMMARY OF ALTERNATIVES STATUS 15
B-2
-------
Alternatives Program
1. INTRODUCTION
In response to the issuance of the Chlorofluoro-
carbon/Ozone Depletion Theory, the Du Pont Company and others
launched major research efforts to resolve the uncertainties in
the underlying science in order to quantitatively verify or
disprove the theory. However, as the outcome of this research
could not be predicted, Du Pont also initiated a program to
examine how the hypothetical stratospheric chlorine burden
resulting from the emissions of chlorofluorocarbons (CFCs) might
be reduced. This program had two thrusts: 1) To assess the
technological and economic feasibility of reducing CFC emissions
through conservation and/or recovery/recycle, and 2) To
identify and develop fluorocarbon compounds that could be used
in place of the CFCs but which would not contribute to the
hypothesized environmental risk. This paper reports the status
of the alternatives program.
B-3
-------
Alternatives Program
2. CURRENT COMMERCIAL CHLOROFLUOROCARBONS
Six major chlorofluorocarbons are manufactured in
commercial quantity:
CFC-11 (CC13F)
CFC-12 (CC12F2)
CFC-22 (CHC1F2)
CFC-113 (C2C13F3)
CFC-114 (C2C12F4)
CFC-115 (C2C1F5)
These compounds are key factors in maintaining important
segments of the national economy. Their applications include
the energy transfer fluid in refrigeration and air-conditioning,
the blowing agent and insulating gas for plastic foams, the
solvent cleaning and drying medium in the manufacture of
electronic and mechanical equipment and military hardware, the
inerting agent in sterilizing gases, and the freezing agent in
some speciality food freezing. Because the uses are dependent
upon the specific properties of the chlorofluorocarbon employed,
the compounds are not interchangeable among the applications.
3. PROPERTIES OF COMMERCIAL CHLOROFLUOROCARBONS
The commercial compounds contain at least one
carbon-fluorine bond in the molecule and have a unique
combination of properties including low heat of vaporization,
nonflammability, low chemical reactivity, and low toxicity.
These properties result in low energy consumption, very few
material compatibility problems and good safety-in-use. Also,
they give rise to a high degree of atmospheric stability.
B-4
-------
Alternatives Program
4. CRITERIA FOR ALTERNATIVE PRODUCTS
To justify consideration as an alternative product, a
compound must meet the technical performance of current
commercial chlorofluorocarbons with regard to product
performance, toxicity, and safety-in-use. It must meet a
criterion for environmental acceptability (as yet undefined)
opposite its potential for ozone depletion. In addition, a
commercial process for its manufacture must be available, or be
developed, an economic incentive for manufacture must exist and
the cost of the compound must be compatible with its
value-in-use.
A commercial process should (1) utilize raw materials
currently available or readily developed, (2) generate a minimum
of by-products for disposal or treatment to meet environmental
regulations, (3) meet health standards for the protection of
workers, (4) provide a product quality suitable for end-use
applications, (5) be based on technology within the capability
of current materials of construction and design knowledge, and
(6) be affordable as to total capital required for
construction. All of these parameters must be met with the
expectation of a suitable return on the investment required.
5. SCOPE OF DU PONT ALTERNATIVES PROGRAM
The most promising candidates have been identified as
either fluorocarbons containing no chlorine (since these
compounds would not contribute to the hypothesized ozone
destruction) or chlorofluorocarbons containing hydrogen (since
these compounds should degrade to a large extent in the
troposphere, through reaction with hydroxyl radicals). This led
to an examination of all practical fluorocarbon and
chlorofluorocarbon compounds meeting one or the other of these
criteria. Only fourteen compounds were found to comprise this
category:
B-5
-------
Alternatives Program
CFC-21 (CHC12F)
CFC-22 (CHC1F2)
CFC-31 (CH2C1F)
FC-32 (CH2F2)
CFC-123 (C2HC12F3)
CFC-124 (C2HC1F4)
FC-125 (C2HF5)
CFC-132b (C2H2C12F2)
CFC-133a (C2H2C1F3)
FC-134a (C2H2F4)
CFC-1415 (C2H3C12F)
CFC-142b (C2H3C1F2)
FC-1433 (C2H3F3)
FC-152a (C2H4F2)
These compounds have been or are being evaluated
extensively for product performance (as refrigerants, foam blow-
ing agents and solvents), safety (flammability and toxicology)
and manufacturing capability. Thus far, only CFC-22, FC-134a,
CFC-141b, CFC-142b and FC152a have survived all the tests
performed. However, the results of all long-term toxicology
studies, which would be necessary before more broad use of these
compounds would be permitted, are not yet available. Only
CFC-22 has a commerical process.*
Over the past five years, Du Pont has expended over
$15 million of technical effort by its chemists, engineers and
toxicologists to:
• Make sample quantities for testing.
• Formulate products and evaluate performance.
• Perform toxicological tests.
CFC-142b and FC-152a are manufactured in very limited quantities
B-6
-------
Alternatives Program
• Obtain data to calculate atmospheric stability.
• Carry out process research for manufacture.
6. ALTERNATIVE CANDIDATES BY POTENTIAL APPLICATION
A summary of the status of the search for alternatives
by application follows:
a. Refrigerants
i. Mobile Air-Conditioning -
• CFC-12 is the refrigerant used in mobile
air-conditioning.
• CFC-22 and CFC-114 mixtures or CFC-22 and
CFC-142b mixtures potentially could be used in
present equipment designed for CFC-12. Both
CFC-22 and CFC-142b, however, have been found
to produce weakly mutagenic effects in the
Ames Test. Long-term inhalation tests on
CFC-22 and CFC-142b are in progress in order
to ascertain the significance of the Ames Test
results. Additional problems with CFC-142b
are that, by itself, it is flammable under
some conditions. Consequently, long-term
field testing would be required prior to
commerciali zation.
• The use of CFC-22 alone would require major
redesign of equipment.
• FC-134a is a direct substitute for CFC-12
based on laboratory and wind tunnel tests.
However, extensive field testing has not been
B-7
-------
Alternatives Program
performed. More importantly, no commercial
process for its manufacture has been found
despite a large research effort. Toxico-
logical studies are incomplete and limited by
material availability.
ii. Home Appliances -
• CFC-12 is the refrigerant used in home appli-
ances .
• Mixtures of CFC-22 and CFC-142b could be used
with little redesign. Equivalent performance
to CFC-12 has been demonstrated.
• The use of CFC-22 alone would require redesign.
• FC-134a is a direct substitute for CFC-12
based on 20 months of performance testing.
• Toxicological testing and/or process limita-
tions for CFC-22, CFC-142b and FC-134a as
discussed above apply equally to this
application.
iii. Store Refrigeration (Non-Frozen Food) -
• CFC-12, CFC-22 and CFC-502 are all used.
• CFC-22 has caused some performance problems.
The industry prefers CFC-502.
• CFC-502 already is replacing the use of CFC-12
in new equipment.
B-8
-------
Alternatives Program
• Toxicological testing limitations on CFC-22
apply equally to this application.
iv. Chillers -
• CFC-11, CFC-12 and CFC-114 are all used.
• Specific design requirements limit the
potential for a single alternative replacement.
b. Foam Blowing Agents
i. Polyurethane Insulating Foam -
• CFC-11 is the product used in polyurethane
insulating foam.
• The use of any of the alternative candidates
examined will result in some decrease in the
insulating value from that of CFC-11.
• CFC-123 and CFC-141b have been used to produce
this foam and are thought to be technically
feasible as replacements — but a special
polyol will need to be developed.
• No commercial process now exists for the
manufacture of either compound. Toxicology of
both compounds is incomplete.
ii. Flexible Polyurethane Foam -
• CFC-11 is the major CFC product used in flex-
ible polyurethane foam.
B-9
-------
Alternatives Program
• CFC-123, CFC-133a and CFC-141b are technically
feasible as replacements. CFC-133a has poor
toxicity. lexicological testing and process
limitations of CFC-123 and CFC-141b apply
equally to this application.
iii. Polystyrene Insulation and Packaging Foam -
• CFC-12 is the major CFC product in these uses.
• CFC-124 and FC-134a should be technically
feasible. No commercial process exists for
either compound. Toxicological testing is
incomplete.
• CFC-22 and CFC-142b mixtures were evaluated
but performance was poor.
iv. Polyethylene Foam -
• CFC-12 and CFC-114 are the major products used
in polyethylene foam.
• We have not identified a promising alternative
candidate.
c. Cleaning Agents
• CFC-113 is the product used in precision
cleaning, e.g., electronics and computers.
• All possible alternatives either failed to
meet product requirements, are toxic or are
flammable. The closest alternative, CFC-132b,
is inadequate in performance since it is too
strong a solvent.
B-10
-------
Alternatives Program
d. Liquid Food Freezant
• CFC-12 is the only fluorocarbon or chloro-
fluorocarbon the Food & Drug Administration
has approved for this use.
• There is no market justification to develop an
alternative product.
e. Sterilant Gas
• CFC-12 is the product used as an inerting
agent in sterilizing gases.
• There is no market justification to develop an
alternative product specifically for this
use. However, it is possible that one or more
of the alternatives, upon passing all testing,
could be adopted for this application.
7. PROGRAM STATUS AND PLANS
The research and development effort on alternatives is
continuing:
• Acute and short-term toxicity tests have been
performed on:
CFC-21 CFC-133a
CFC-22 FC-134a
CFC-31 CFC-141b
CFC-123 CFC-142b
CFC-124 FC-152a
• Work has been terminated on CFC-21, CFC-31, and
CFC-133a due to adverse toxicological findings.
B-ll
-------
Alternatives Program
• A long-term toxicology study of CFC-22 for
carcinogenecity is under way in Europe. A similar
test on FC-152a is in progress in the United
States with results expected by mid 1981. The
same test has just been initiated on CFC-142b in
the United States with results expected by late
1982.
• Small scale field tests of CFC-22 and CFC-142b in
automotive air-conditioning are underway. Field
tests of FC-134a will require production of sample
material.
• Basic data has been collected over the last 20
months on the performance of FC-134a in a house-
hold refrigerator.
• Basic data on equipment modification needed to
replace CFC-12 with CFC-22 or CFC-22 and CFC-142b
mixtures in refrigeration and air-conditioning are
being developed.
• Work is continuing on short-term tests which would
enable the determination of long-term insulating
performance of alternatives in polyurethane foam.
• Work has been terminated on CFC-123, CFC-124 and
FC-125 due to lack of an acceptable manufacturing
process.
• Process research is continuing on commercial
routes to FC-134a and CFC-141b.
B-12
-------
Alternatives Program
8. TIMETABLE FOR COMMERCIALIZATION
Any alternative will require up to ten years for
commercial production (assuming all technical and toxicological
programs yield favorable results) as follows:
• Construction and operation of pilot plant to develop
process design data and to obtain material (40 to 50
tons) for toxicity testing: 2 to 3 years.
• Long-term chronic toxicity testing: 3 years.
• Development of basic design data for commercial plant,
including waste disposal: 2 years (can coincide with
toxicity testing).
• Purchase of plant equipment and the preparation of
plant site: 2 years.
• Obtain environmental permits: 1 year (can coincide
with site preparation).
• Construction of plant: 1 to 3 years.
• Start-up to full commercial operation: 1/2 to 1 year.
• Prudent business decision would hold purchase of
equipment and plant construction for favorable
toxicological results.
1 Although processes exist for CFC-22, CFC-142b and FC-152a,
significant expansions or new facilities would be required
for their increased production and for their raw material
needs.
B-13
-------
9. SUMMARY OF ALTERNATIVE STATUS
00
I
Fluorocarbon
#
and Formula
11 CC13F (a)
12 CC12F2 (a)
Potential Application
Blowing agent, refrigerant
Refrigerant, blowing agent,
Manufacturing
Process
Yes
Flammability Toxicology
No Low
113 CC12FCC1F2 (a)
114 CC1F2CC1F2 (a)
21 CHC12F
22 CHC1F2
food freezant, sterilant
Solvent, refrigerant
Blowing agent, refrigerant
Replacement for CFC-11;
blowing agent
Replacement for CFC-12;
Yes
Yes
Yes
Yes
No
No
No
No
Low
Low
Low
Toxic; Dropped
Weak mutagen, Life-
31 CH2C1F
32 CH2F2 (b)
123 CHC12CF3
124 CHC1FCF3
125 CHF2CF3 (b)
132b CH2C1CC1F2
133a CH2C1CF3
refrigerant
Aerosol propellant
No primary application; low
boiling compound; Dropped
Replacement for CFC-11;
blowing agent, refrigerant
Replacement for CFC-12;
refrigerant, sterilant
Refrigerant
Replacement for CFC-113;
too strong a solvent;
Dropped
Blowing agent, propellant
Yes
Not commercial
Not commercial
No; Dropped
No; Dropped
No; Dropped
No
Not commercial
in U.S.
No time test underway
Yes Toxic; Dropped
Yes Low
No Low
No Low
No Not known
No Very incomplete
Embryotoxic;
No Dropped
-------
Fluorocarbon
f
and Formula
134a CH2FCF3 (b)
141b CH3CC12F
142b CH3CC1F2
143a CH3CF3 (b)
152a CH3CHF2 (b)
Potential Application
Replacement for CFC-12;
refrigerant, others?
Replacement for CFC-11;
blowing agent
Manufacturing
Process
No
Yes, developmental
Blowing agent, refrigerant Yes
Refrigerant Not commercial
Propellant, refrigerant Yes
Flammability Toxicology
Very incomplete;
testing in
No progress
Yes Weak mutagen
Weak mutagen;
Lifetime
Yes test underway
Yes Incomplete
Yes Low; Lifetime test
underway
i
M
Ul
(a) existing products for comparison; (b) contains no chlorine.
-------
X. APPENDIX C
THE ENERGY CONSEQUENCES OF
CHLOROFLUOROCARBON REGULATION
(BATTELLE REPORT)
C-l
-------
Energy Consequences
of CFC Regulation
The following discussion on the energy consequences of
chlorofluorocarbon regulation summarizes work done by Battelle,
Columbus Division under contract to E. I. du Pont de Nemours &
Company.
The discussion is organized for various levels of
interest and varying need for detail.
• The first section is the Summary and Conclusions
which deal only with the net results and conclusions
to be drawn.
• Next is the Introduction describing the objectives
and scope, and methodology.
• Following next are brief discussions of each of the
four major CFC use areas investigated, giving some
of the rationale involved.
/
References and appendices cited in this section refer to
the full report from Battelle [Battelle, 1980]. The full report,
including appendices giving detailed rationale and calculations
as well as tabulation of the results, is included in our
submission.
C-2
-------
Energy Consequences
of CFC Regulatfon
TABLE OF CONTENTS
Page
Summary and Conclusions 4
Introduction 10
Objective and Scope 11
Methodology 11
Automotive Air-Conditioning 14
General Considerations s 14
Energy Penalties 16
Refrigeration 18
General Considerations 18
Energy Penalties 19
Insulating Foams 24
General Considerations 24
Energy Penalties 24
Liquid Food-Freezing 27
General Considerations 27
Energy Penalties 27
C-3
-------
Energy Consequences
of CFG Regulation
FINAL REPORT
on
ENERGY CONSEQUENCES OF
CHLOROFLUOROCARBON REGULATION
to
E. I. du Pont de Nemours
& Company
from
BATTELLE
Columbus Division
by
P. R. Beck, J. M. Corliss,
M.E.D. Hillman, J. L. Otis
and S. G. Talbert
SUMMARY AND CONCLUSIONS
The overall results of this study are presented in Table 1
in summary form and in somewhat greater detail 1n Table 2. The results
show that there is indeed a significant energy penalty associated with
a ban on chlorofluorocarbons in the various applications studied.
The results indicate there would be a net energy penalty of
5.5billion or 9.5 billion gallons of fuel equivalents in the tenth
year of the ban and a cumulative total of 27.9 billion or 49.8 billion
gallons for the first decade following the ban, depending on whether
or not R-22 is adopted as a substitute for the banned CFCs. Adoption
of R-22 leads to the smaller penalty.
Text continues on page c-7
c-4
-------
TABLE 1. SUMMARY OF ENERGY PENALTIES ASSOCIATED WITH A
BAN ON USE OF CHLOROFLUOROCARBONS
(millions of gallons fuel equivalent)
o
Automotive A1r Conditioning
Home & Store Refrigeration'8'
Insulating Foams
Liquid Food Freezing
Total
Using
1981
48
169
231
9
457
R-22 Where Possible
1990
326
2,003
3,169
12
5,510
Decade
2,090
10,371
15,360
106
27,927
Using Next Best
1981
161
446
231
9
847
1990
1,070
5,267
3,171
12
9,520
Alternative
Decade
6,870
27,461
15,365
106
49,802
(a)
Includes losses due to outdoor compressor when using ammonia and Incremental
effect of elimination of CFC's 1n both refrigerant and foams.
9
-------
Energy Consequences
of CFC Regulation
TABLE 2. ENERGY PENALTIES ASSOCIATED WITH A BAN
ON THE USE OF CHLOROFLUOROCARBONS
(millions of gallons of fuel equivalent)
Using R-22
Where possible
.
Automotive Air Conditioning
Domestic & Commercial Refrigeration
Refrigerators
Freezers
Centrifugal Chillers
Food Store Freezers
Beverage Coolers
Mobile Units
Unit Coolers
Ice Makers
Water Coolers
Insulation
Commercial Construction
Residential Construction
Industrial Construction
Refrigeration
Tanks and Pipes
Transportation
Incremental Combined Effect
Refrigerators
Freezers
Mobile Units
Losses Due to Outdoor Compressor
Liquid Food- Freezing
Totals
1981
48
162
77
19
9
3
39
0
3
11
1
231
117
11
4
68
30
1
7
3
4
0
—
9
457
1990
326
1930
907
200
115
31
540
0
28
128
6
3169
1707
174
52
833
393
10
73
36
35
2
—
12
5510
Decade
2090
9980
4750
1080
600
170
2670
0
.150
660
30
15,360
7740
840
260
4420
2040
60
391
190
191
10
—
106
27,927
Using Next
Best Alternative
1981
161
173
77
19
7
3
39
13
3
11
1
231
117
11
4
68
30
1
7
3
4
0
266
9
847
1990
1070
2084
907
200
90
31
540
154
28
128
6
3171
1707
174
52
833
393
12'
73
36
35
2
3110
12
9520
Decade
6870
10,770
4750
1080
470
170
2670
.790
150
660
30
15,365
7740
840
260
4420
2040
65
391
190
191
.10
16,300
106
49,802
C-6
-------
Consequences
or CFC Regulation
It should first be noted that these penalties describe only
the first decade of the ban and that the annual penalty will undoubtedly
continue to grow for many decades beyond the first. This Is because
complete displacement of the more efficient CFCs such as R-ll and R-12
will not be completed within 10 years. Most of the applications con-
sidered have a useful life beyond 10 years, and particularly 1n the
case of insulation, replacement of the more effective CFCs would not be
complete for several decades. Thus the energy penalty would continue to
grow well beyond the first decade. No attempt has been made to project
the magnitude of these future increases because of the obvious unrelia-
bility of such long-range forecasts. Nevertheless, 1t is certain that
the annual penalty will continue to grow well beyond the first decade.
Alternative projections are given for the first decade because
of the unique characteristics of refrigerant R-22. Whether or not R-22
represents a viable alternative depends upon EPA actions regarding its
future and the attitude of industry regarding its economic feasibility.
R-22 differs from other CFCs in that it contains a single hydrogen atom.
The presence of this hydrogen atom makes it susceptible to degradation
by reaction with hydroxyl radicals. Ironically, this "weakness" causes
it to disassociate in the lower levels of the atmosphere. Less, there-
fore, survives to diffuse to the upper atmosphere where it may cause
harm to the ozone layer. The impact is variously estimated to be
10 to 30 percent of the impact caused by fully halogenated CFCs. While
scientists may not agree on just how much the impact is reduced, 1t
is clear that R-22 is a preferred material from the standpoint of the
impact on the ozone layer.
It is also clear that R-22 1s a preferred material in the
refrigeration and automotive air conditioning applications from the
energy conservation standpoint. It would therefore appear to be a
desirable compromise in the event that other CFCs are banned in these
applications.
However, this study has focused only on the energy consequences
of a ban on CFCs, and the economic Impact of a conversion of current
refrigeration systems to R-22 has not been evaluated. At any given
C-7
-------
Energy Consequences
ot CFC Regulation
temperature R-22 has a substantially higher vapor pressure than R-ll
and R-12 and would require a complete redesign of the refrigeration and
automotive units for future use. This, of course, would be a costly
procedure which the industry may not choose to undertake. Further, since
higher pressures would be involved, the redesigned equipment would
undoubtedly be more expensive and heavier than that currently in use
and could be viewed as unsalable. Finally, although R-22 represents a
substantial improvement with respect to possible damage to the ozone
layer, it is not regarded as completely harmless. In weighing the sub-
stantial business risks involved, plus the possibility of future
restrictions on R-22, industry may choose to reject 1t as an alterna-
tive to the currently used CFCs. Rather than speculate further on how
industry will view R-22 as an alternative, Battelle has chosen to simply
present the consequences if R-22 is used or if industry selects the next
best alternative. The difference in the results is about 22 billion
gallons for the decade and 4 billion gallons in the tenth year.
Numbers like 50 billion gallons of fuel per decade have some
meaning to petroleum economists and others closely involved with the
magnitude of the numbers relating to the energy crisis. However, such
numbers need to be related to more familiar benchmarks for those outside
the energy industry. Following are a few such benchmarks for the energy
penalty associated with the 9 billion gallons of fuel calculated for the
tenth year of a ban on CFCs and R-22.
• 9.5 percent of our 1978 crude oil imports
• 38 percent of our 1978 oil imports from Iran
t 45 percent of the current annual oil production
from Alaska's North Slope
• Energy to supply 11 cities the size of Toledo
(population about 500,000) for 1 year
(excluding gasoline)
-• Energy output of 29 nuclear power plants
(nearly half the number in operation)
• 18 times the expected petroleum savings
envisioned for 1985 through the use of
"gasohol"
c-8
-------
Energy Consequences
of CFC Regulation
• 73 percent of the third quarter 1978 balance
of payments deficit (assuming oil imports are
$25 per barrel)
• The fuel required to drive 12 million average
cars (about 10 percent of all U.S. cars) for
1 year.
The primary conclusion to be drawn Is quite simple. If the
use of CFCs is banned in the applications examined, there will be
additional and serious aggravation of our current energy problems. The
banning of some CFCs but not R-22 might reduce the Impact of the regu-
lations, provided Industry 1s willing to take the business and economic
risks involved.
With more extensive computation, additional data, more detailed
market information, or simply different, but still realistic assumptions,
one might derive a different net total which may be somewhat higher or
lower than presented here. The calculations leading to these results
are, in fact, presented in some detail in the Appendices to this report
to facilitate an examination of new or different data. The conclusions,
however, will remain the same: A ban on the use of CFCs will have an
adverse and serious impact on an already serious energy problem.
C-9
-------
Energy Consequences
of CFG Regulation
INTRODUCTION
Chlorofluorocarbons (CFCs) constitute a series of specially
designed synthetic fluids originally developed to perform as refrigerants.
They are of low toxicity and chemically Inert. Over the years, 1n
addition to refrigeration, they have found a number of applications,
including aerosol propel1 ants and blowing agents for Insulating foams.
Most of these applications relate directly to their unique physical
properties and their safety features.
However, in recent years it has been claimed that the CFCs
diffusing into the upper atmosphere are having a depleting effect upon
the protective ozone layer. Because of this, the Environmental Protection
Agency has already banned the use of CFCs as aerosol propel!ants and
is considering further restrictions upon their use in other applications.
In most major applications, the CFCs are used because of their
advantageous thermodynamic properties. It is therefore logical that
exclusion of their use in such applications would carry an associated
energy penalty. In 1979 the Freon® Products Division of the E. I. du
Pont de Nemours & Company conducted a preliminary examination of the
energy impact of a ban on CFCs in several major applications. The results
of this preliminary study indicated that there would be, indeed, a large
penalty if CFCs were banned and industry was forced to use less optimal
materials. Based on these indications, Du Pont then decided to have a
more thorough examination conducted by an independent organization, and
Battelle was selected to carry out this work. This report contains the
results of that effort.
c-io
-------
Energy Consequences
of CFC Regulation
Objectives and Scope
The primary objective of this program was to develop a realistic
estimate of the energy costs associated with a ban on the use of chloro-
fluorocarbons in major applications. The applications covered are:
• Automotive air conditioning
0 Home and store refrigeration
• Insulating foams
• Liquid food freezants.
.
The study focused only on energy consumption. No analysis of
economics or safety was performed except that these factors were given
qualitative consideration in selection of alternatives to be used for
comparison.
Only systems and equipment currently in use were considered.
New inventions or developments would have been considered 1f it were likely
they would have a near-term impact. However, no such developments or
breakthroughs for these applications are apparent.
Finally, no consideration was given to developing technologies
involving the use of CFCs for other applications such as heat pipes, solar
heating, etc. These situations represent an additional penalty in terms
of lost opportunity but do not fit well with the "here and now" analysis
of existing areas.
Methodology
The general approach employed in all cases consisted of five
steps:
• Selection of one or more reasonable alternatives
to the use of CFCs
_• Development of appropriate per unit engineering
parameters
• Calculation of the per unit energy penalty
Involved
• Development of market data for the period
-1981 to 1990
t Extension of the total energy penalty for
e»r:!» year of the decade.
c-ll
-------
Energy Consequences
of CFG Regulation
Implicit 1n the selection of alternatives is the presumption
that there will be no near-term technical breakthrough providing easy
solutions to the problem. For example, it appears very unlikely that
any new type of refrigeration equipment will be developed over the next
decade. Therefore, the systems considered are conventional heat pump
or compressor-type refrigeration, absorption systems, and air-cycle
systems.
Likewise, it is presumed to be highly unlikely that alternative
fluids to the CFCs will be developed. The CFCs were specifically
designed to optimize their thermodynamic properties in a nonflammable,
nontoxic fluid. By definition these materials must be simple molecules,
consisting of a very few atoms. Were there any obvious alternatives
to the current CFCs, they very likely would have been developed long ago.
While scientists are always reluctant to say "It can't be done", they
generally agree development of alternatives to CFCs having equivalent
safety and thermodynamic properties is extremely unlikely.
Battelle therefore believes the alternatives given consideration
in this report are the most reasonable and appropriate for the time period
being considered. The development of basic engineering calculations to
determine per unit energy penalties required a certain number of assumptions
regarding typical systems. For example, in refrigeration, certain
assumptions are necessary regarding the ambient conditions in which the
unit is operating. Also, in the case of insulation, certain assumptions
were necessary regarding the overall structure of the insulated surface.
Such assumptions are detailed in the appendices along with the calcu-
lations such that if the reader wishes, he or she may test the impact of
different sets of assumptions. However, Battelle believes that the
conditions and structures selected for this set of calculations are
reasonably typical and representative of actual situations. The calcu-
lations themselves are relatively straightforward and follow standard
engineering practice.
*Hydrochlorofluorocarbons, for Instance CHC1F9 (R-22), constitute a special case.
See page c-7 c_12
-------
Energy Consequences
of CFC Regulation
In most cases market data was derived from generally accepted
authoritative sources, and assumptions were avoided wherever possible.
A 11st of references 1s provided at the end of each appendix. In a
few cases even the authoritative sources disagreed, and it was necessary
to select one source or develop some compromise. For example, there
1s considerable disparity 1n the opinions as to what percentage of the
time automobile air conditoners are actually 1n operation, and the
resolution of this disparity 1s detailed in Appendix C.
Considerable amounts of the market data were derived from the
Rand Corporation draft report on the economic implications of regulating
chlorofluorocarbon emissions. It should be noted that this Is a working
draft dated September 1979, with the finalized version of the report
to be Issued later this year. However, it is Battelle's understanding
that there will be no changes between this draft and the final report in
the market data for the applications with which this report 1s concerned
(Reference B-3, Appendix B). However, should there by any significant
changes, the detail of marketing calculations and forecasts 1s Included
in the Appendices so that they may be incorporated by the reader.
The time period 1981 to 1990 was selected for the market fore-
casts primarily because of the availability of data for that period,
especially data from the Rand Report mentioned above. This does not
necessarily presume that a ban on chlorofluorocarbons would have an
impact as early as 1981; rather 1t presumes that the Impact during any
10-year period in the relatively near future would be roughly the same
as those for which the forecasts have been made.
The results of multiplying the market forecasts by the penalties
per unit result 1n the total energy penalty for the various applications.
These results are in dissimilar units, I.e., gallons of gasoline, kilowatt
hours, gallons of fuel oil, etc. For purposes of this report, comparison
and summation were converted to the equivalent gallons of fuel (140,000
Btu energy content) as a common denominator. This unit was chosen because
of the more universal familiarity of the public with the meaning of a
gallon of fuel, e.g., gasoline, as opposed to such terms as Btu's, quads
and kilowatt hours. The actual penalty, of course, if a ban 1s enacted,
will be some mlxutre of gallons of gasoline or fuel oil, tons of coal,
units of nuclear power, hydroelectric, etc.
C-13
-------
Energy Consequences
of CFC Regulation
AUTOMOTIVE AIR CONDITIONING
Supporting data, references, and calculations for the discussion
of automotive air conditioning are contained in Appendices B, C, and K at
the end of this report.
General Considerations
The popularity of automotive air conditioning has grown steadily
since its inception. Currently, nearly three-fourths of all domestically
produced passenger cars, about half of the light trucks, and half of the
imported automobiles are equipped with air conditioners. In addition,
there are substantial after-market sales for installation for nonfactory
installed air conditioners. These air conditioners, at present, consume
an estimated 500 million gallons of gasoline per year and will consume
naarly 22 billion gallons over the course of the next decade. The popu-
larity of air conditioning in automotive vehicles is no longer considered
a matter of personal comfort. It has come to be regarded as a necessity
by many, especially those who are subjected to prolonged periods of
driving under heat stress conditions. The air conditioner, of course,
plays a key role in reducing the debilitating effect of heat stress and
thereby contributes to safety, health, and productivity. The benefits
derived from automotive air conditioning are quite parallel to those
involved in the air conditioning of homes and commercial buildings but
with significantly greater emphasis on safety.
R-12 is the universally used refrigerant in automotive air
conditioning systems. The systems have been designed specifically for R-12,
and no other refrigerant can be substituted directly in these systems.
Thus, if the use of R-12 is prohibited, a redesign of the air conditioning
equipment will be necessary. Prelminiary calculations eliminated absorption
and air-cycle systems as alternatives to the current compressor CFC units.
Thus, the alternatives considered in this report involve substitution of
other fluids for systems similar to the current R-12 units but appropriately
modified for the substitute fluids. The use of R-22, ammonia, and propane
were examined.
C-14
-------
Energy Consequence^
of CFC Regulation"^"
Two penalties are associated with a switch from R-12 to any
of the above three alternatives. One is associated with the increased
weight of the vehicle. This is because the alternatives operate under
conditions of higher pressure and will require stronger components
throughout the unit. Whether, in fact, dependable units can be designed
within the limits of economic feasibility is open to question. However,
only the energy considerations are within the scope of this study.
The second penalty reflects changes in the efficiency caused
by the change of refrigerant and by the design requirements necessi-
tated by the use of refrigerants which otherwise expose occupants to
unacceptable toxicological and flammability hazards.
It should be noted that the penalty associated with the effec-
tiveness of the unit applies, of course, only while the air conditioner
is on, probably one-third of the time on the average, but the penalty
associated with increased weight is, in effect, 100 percent of the miles
driven.
c-15
-------
Energy Consequences
of CFC Regulation
Energy Penalties
Energy penalties associated with a ban on CFCs are shown in
Table 3. The results show that the increased fuel consumption if R-12
is banned but R-22 continues to be used will be about 2 billion gallons
of fuel for the decade, or about 10 percent more than the current fuel
used for air conditioning if R-12 were continued in use.
If R-22 should also be banned, the next best alternative
would appear to be an ammonia system. The penalty here for the decade
is on the order of 7 billion gallons of fuel, which is roughly one-third
more than the current fuel used for air condition'irg if R-12 were to
continue in use.
The use of propane resulted in higher penalties due to both
cycle efficiency and weight increase. Therefore, only the ammonia and
the R-22 systems are considered in the summary and conclusions.
It should be noted that ammonia and propane are both flammable
and that ammonia is also toxic. Ammonia is a disabling lachrymator.
Conceivably either or both could be prohibited from use in motor vehicles
for safety reasons. Also, because of these characteristics, systems
employing these fluids are assumed to require a secondary loop (a secondary
heat exchanger) located outside the cab of the vehicle so that the
refrigerant does not enter the area occupied by passengers. This further
increases both the weight and cost of the system and lowers efficiency.
Thus while ammonia and propane may be viable alternatives from an energy-
penalty standpoint, they may eventually prove impractical on the basis
of economics and safety.
R-22, on the other hand, is nonflammable and relatively nontoxic
and, therefore, would neither require a secondary loop nor present an
increased hazard on the highway. Its use would, however, still require
redesign of equipment and undoubtedly additional costs.
C-16
-------
TABLE 3. ENERGY PENALTIES ASSOCIATED WITH USE OF ALTERNATIVES
TO R-12 VEHICULAR AIR CONDITIONING
(millions of gallons per year or decade)
o
i
Penalties Associated with Refrigeration
Cycle Efficiency (Excluding Weight Penalties)
If Only CFCs
Banned
Use of R-22:
1981 34
1990 226
Decade 1450
1981
1990
Decade
If CFCs and
R-22 Banned
Use of Ammonia:
147
969
6230
Use of Propane:
193
1270
8190
Penalties Associated with Refrigeration*
Cycle Efficiency Combined with Weight*3'
If Only CFCs If CFCs and
Banned R-22 Banned
Use of R-22 Use of Ammonia
48 161
326 1070
2090 6870
Use of Propane:
207
1370
10,100
(a) Height penalties for either ammonia or propane systems are likely to be greater than the 12
pounds estimated for R-22. However, a 12 pound weight penalty has been assumed for ammonia
and propane systems.
-------
Energy Consequences
of CFC Regulation
REFRIGERATION
Supporting data, references and calculations for the material
in the discussion of refrigeration are contained in Appendices D, G, J
and K.
General Comments
Refrigeration is the largest and probably the most important
of the current uses of CFCs. In almost all cases our fresh food supply
1s dependent on some form of cooling, from packing and shipment through
warehousing and retail and into the home. Even fruits and vegetables
not thought of as requiring refrigeration must be moved through the
distribution system at controlled temperatures.
In addition to the food industry, there are a host of industrial
applications where chillers or coolers are necessary.
Of these items, the familiar household refrigerators and freezers
are the largest market segment. However, this market is largely saturated,
and growth of new sales is proportional to new household formation:
sales are mostly for replacement purposes. Sales in other segments, notably
beverage coolers and mobile transportation ("reefers"), are more dynamic.
Refrigerant R-12 is by far the most broadly used refrigerant in
this category. Some R-22 and R-502 is used in food store freezers. R-ll,
R-12, R-500, R-114, and R-22 are used in centrifugal chillers, but R-ll
and R-12 are 90 percent of the total.
Alternatives to the use of CFCs in these applications include
ammonia, propane, air-cycle, and absorption cycle. In all cases, ammonia
was found to be the best non-CFC alternative. Ammonia, of course, is
both toxic and flammable and would not be permitted in homes, stores, etc.,
where it might present a hazard. Therefore, the compressor unit would
necessarily be remote from the cooling unit, requiring a secondary loop
using a safe fluid to cool the loop within the house or store. Such units
are not Inconceivable and would resemble current home air conditioning
Installations with an added external heat exchanger. However, there would
be a substantial economic penalty compared to an ordinary refrigerator or
freezer as well as an energy penalty.
C-18
-------
Energy Consequences
of CFC Regulation
In addition to the energy penalty associated with the reduced
effectiveness of ammonia In the case of refrigerators, there 1s also a
penalty for the lost utilization of waste heat. A household refrigerator
1s, In effect, a heat pump transferring energy from food, air, etc.,
within the refrigerator to the home. Thus the electrical energy required
to operate the refrigerator results In reduced heating fuel demand. This
energy, of course, is wasted if the compressor unit 1s located outside
the home. Of course, the reverse 1s true If the home 1s being air
conditioned. The net effect is calculated in Appendices G and J.
Further, if there were .a simultaneous ban on use of CFCs in a
refrigerant and 1n Insulating foams, the energy penalty associated with
the refrigeration cycle would be increased. That is, because the
insulating would be less effective, more units of cooling would be required
from a unit operating at reduced efficiency. This effect 1s also detailed
in Appendixes G and J.
In the event that R-22 were permitted in future refrigerant use,
a secondary outdoor loop would not be required for safety reasons. There
would be a reduced economic penalty; a higher pressure system would still
be required, but this would certainly cost less than an exterior loop.
Also, there would be no net heating penalty for waste heat because the
unit could safely be within the home.
Energy Penalties
The energy penalties associated with a ban on CFCs and the
use of ammonia as an alternative for the refrigeration cycle efficiency
alone are shown In Table 4. Also shown are the penalties 1f R-22 were
allowed. Only in the case of mobile refrigeration is the difference
significant. There are other differences 1n food store freezers and
centrifugal chillers, but the numbers are small and disappear upon
rounding. ~
Table 5 shows the additional Impact of a simultaneous ban
on CFCs for Insulation, as well as refrigeration for refrigerators and
freezers and mobile units. The Impact Is the same whether R-22 or
Text continues on page c-22
C-19
-------
Energy Consequences
of CFC Regulation
TABLE 4. ENERGY PENALTIES ASSOCIATED WITH A BAN
ON CFCs FOR REFRIGERATION
(millions of gallons of fuel equivalent)
Using Ammonia
Refrigerators*
Freezers*
Centrifugal Chillers
Food Store Freezers*
Beverage Coolers*
Mobile Units
Unit Coolers*
Ice Makers*
Water Coolers*
Total
1981^
77
19
7
3
39
13
3
11
1
173
1990
907
200
90
31
540
154
28
128
6
2084
Decade
4750
1080
470
170
2670
790
150
660
30
10,770
Using R-22
1981 ~
77
19
9
3
39
0
3
11
1
162
1990
907
200
115
31
540
0
28
128
6
1955
Decade
4750
1080
600
170
2670
0
150
660
30
10,110
^Secondary loop assumed when using ammonia.
C-20
-------
Energy Consequences
of CFC Regulation
TABLE 5. INCREMENTAL IMPACT OF A SIMULTANEOUS BAN ON CFC
USE IN BOTH REFRIGERATION AND INSULATION
(millions of gallons of fuel equivalent)
Using Either Anmonla or R-22
as Alternative Refrigerant
1981 1990 Decade
Refrigerators 3 36 190
Freezers 4 35 191
Mobile Units 0 2 10
Total 7 73 391
C-21
-------
Energy Consequences
of CFC Regulation
ammonia 1s used, and amounts to an additional 391 million gallons for the
decade. If data were available, similar calculations for other commercial
refrigeration would probably add another 100 million gallons to the total.
Finally, Table 6 shows the additional net energy penalty for
refrigerators only if the compressor 1s located outside of the residence,
taking into account the effect on both heating and air conditioning
requirements.
The total of all of these effects for the decade amounts to
10.5 billion gallons if R-22 were allowed and 27.5 billion gallons if
ammonia were used.
C-22
-------
Energy Consequences
of CFC Regulation
TABLE 6. IMPACT OF LOCATING REFRIGERATOR COMPRESSOR
UNIT OUTSIDE OF HOME
For a home not
air conditioned
For an air conditioned
home
Net
Average if 61% are
air conditioned
9.11 million Btu's per year lost heat
9.11 million Btu's per year lost heat
-4.70 million Btu's per year lost cooling
4.41 million Btu's per year lost heat
6.24 million Btu's per year
44.6 gallons of fuel equivalent per year
For: 1981
1990
266 millions of gallons of fuel equivalent
per year
3110 millions of gallons of fuel equivalent
per year
millions of gallons of fuel equivalent
16,300 f0r the decade
C-23
-------
Energy Consequences
of CFG Regulation
INSULATING FOAMS
Supporting data, references and calculations for this discussion
of insulating foams are contained in Appendices E, F and G.
General Considerations
CFCs are employed as the blowing agents in rigid polyurethane
and polystyrene (PS) foams, which are used extensively as insulation in
residential and commercial construction, freezers and refrigerators, and
transportation equipment. The blowing agents generally used are R-ll
and R-12. In addition to being nonflammable and nontoxic, the CFCs have
lower heat transfer coefficients than other usable gases.
In the case of residential, commercial and industrial con-
struction, the insulation may be applied as slabstocki laminated board,
filled spray or poured-in-place. In transportation, slabstock, spray or
poured are used. Tank insulation is virtually all sprayed, and refriger-
ation and pipe insulation is poured.
In the case of polystyrene foams, the use of CFCs is for reasons
of safety and performance. Pentane is an effective substitute from the
standpoint of producing suitable cell size and foam structure, but its
flammability constitutes a severe production hazard.
In polyurethane and polystyrene foams used for Insulation, the
CFC is retained in the cells over long periods and contributes materially
to its insulating properties. Foams blown with CFC have on the order of
twice the insulating value of foams blown with carbon dioxide.
Energy Penalties
Table 7 shows the energy penalties associated with elimination
of the use"of CFCs as blowing agents in the various applications listed.
Insulating materials are only one component contributing to
the Insulating value of a structure. Masonry, wood, sheathing, etc., all
contribute something. The structures used as a basis for the table
calculations are detailed in Appendix F, and the alternatives are shown
1n Table E-7.
C-24
-------
Energy Consequences
of CFC Regulation
TABLE 7. ENERGY PENALTIES ASSOCIATED WITH THE ELIMINATION
OF CFCs IN INSULATING FOAMS
(millions of gallons of fuel equivalent)
Commercial Construction
Residential Construction
Industrial Construction
Refrigeration
Tanks and Pipes
Transportation Equipment
Total
1981
106
8
4
68
30
1
217
1990
1565
126
52
833
393
10
2979
Decade
7010
610
260
4420
2040
60
14,403
C-25
-------
CFC Consequences
of CFC Regulation
In most cases, these alternatives presume some thickness con-
straint. For example, refrigerators cannot be made larger to accommodate
thicker Insulation because of constraints Imposed by 2x1 sting doorway
sizes. The Interior size could be reduced, but smaller capacity would
result 1n some combination of more units or more frequent trips to the
grocery and greater gasoline consumption.
Tank Insulation presents a special problem 1n that there is no
real constraint on the thickness of Insulation used. Fiberglass is not
currently used because 1t requires protection from the elements, especially
rain. Polyurethane foam is self-protecting. If one assumes effective
protection for fiberglass can be developed and Is economically practical,
the thickness can be made as great as needed to eliminate any energy
penalty. Therefore, the table shows a range of penalties for equivalent
thicknesses to no loss. To date the alternative has frequently been
foam or no insulation.
The uninsulated alternative is unlikely to remain viable so
was not considered, although it would lead to much larger energy penalties.
C-26
-------
Energy Consequences
of CFC Regulation
LIQUID FOOD-FREEZING
Datat references and calculations relating to this discussion
of liquid food-freezing are contained in Appendix H.
General Considerations
Liquid food-freezing refers to the use of a freezant R-12 (IFF)
to spray foods or directly immerse them in cold liquid. The process differs
from cryogenic freezing with liquid nitrogen or liquid carbon dioxide in
that the freezant is condensed and recovered rather than vented to the
atmosphere. The liquid must be therefore easily condensed. In addition,
it must be inert, of low toxicity, and, of course, high purity.
IFF is used only for certain rather delicate types of food. The
air blast technique used for most foods is cheaper but tends to cause
^dehydration or other damage to some products. Among the foods where IFF
'is used are clams and shrimp, french-sliced green beans, berries and
similar freeze-fragile items. For these products, air blast is not an
alternative if quality is a constraint. Therefore, only cryogenic nitrogen
and carbon dioxide are considered as alternatives.
Energy Penalties
The energy penalties associated with discontinuation of the use
of R-12 in IFF applications are shown in Table 8. These data show there
1s an expected penalty of about 12 million gallons of fuel equivalent
in the tenth year of a ban and over 100 million gallons for the first
decade.
These projections have assumed a rather modest growth for LFF.
The growth might well be greater except that the threat of a ban on the
use of R-12 has already had an impact on the installation of new systems.
The projections, therefore, reflect the current outlook rather than
speculation on what the situation might have been.
C-27
-------
Energy Consequences
of CFC Regulation
TABLE 8. ENERGY PENALTIES ASSOCIATED WITH USE OF
LIQUID NITROGEN AND LIQUID CARBON DIOXIDE
AS A SUBSTITUTE FOR R-12 IN FOOD FREEZING
(millions of gallons fuel oil equivalent)^'
•Year
Fruits
Seafood
Meat
and Vegetables
and Specialties
Total .
1
6
1
1
9
981
.45
.38
.39
.22
1
8
1
1
12
990
.42
.81
.81
.04
Decade
73
15
15
105
.94
.87
.94
.75
(a) Battelle calculation based on Tables H-2 and H-4,
028
-------
X. APPENDIX D
- A COMPARISON OF SOME OF THE PRINCIPAL FINDINGS -
THE NATIONAL ACADEMY OF SCIENCES REPORT
"STRATOSPHERIC OZONE DEPLETION BY HALOCARBONS:
CHEMISTRY AND TRANSPORT", NOVEMBER, 1979
AND
THE UNITED KINGDOM DEPARTMENT OF THE ENVIRONMENT
REPORT "CHLOROFLUOROCARBONS AND THEIR EFFECT ON
STRATOSPHERIC OZONE", OCTOBER, 1979
D-i
-------
NAS/UK DOE Comparison
Introduction
The National Academy of Sciences' (NAS) November, 1979
report [NAS, 1979a] on the effect of chlorofluorocarbons (CFCs)
on the ozone layer differs sharply from the conclusions reached
in a report released in October, 1979 by the Department of the
Environment of the United Kingdom [UK DOE, 1979].
The two reports, drawn up by qualified scientific
bodies, have arrived at significantly different conclusions on
matters ranging from uncertainties in the calculations to the
significance of missing chemistry in the computer model.
Since the subject and conclusions of these reports are
matters of international consequence, the Du Pont Company
believes there should be a resolution of these scientific
differences regarding the effect of CFCs on the ozone layer prior
to further unilateral regulatory action by the United States.
A comparison 'of some of the major differences between
these two reports follows:
D-2
-------
NAS/UK DOE Comparison
NAS REPORT
•There is agreement with
previous reports that
continued release of
halocarbons into the atmo-
sphere will result in a
decrease in stratospheric
ozone.
UK REPORT
•The validity of the hypothesis
is still in doubt.
•Basic scientific understand-
ing, although progressing
rapidly, is still inadequate
in many respects.
•The report concludes that
present understanding of ozone
depletion is limited and is
based on model assumptions
which have not been adequately
validated.
•New values for some of the
chemical rate coefficients
have increased the predicted
ozone reduction resulting
from continued release of
chlorofluoromethanes
(CFM's).
>The results concur with other
studies that the predicted re-
ductions in total ozone amount
are greater than those estima-
ted in 1975 at the time of the
preparation of Pollution Paper
No. 5. Whether the statements
realistically describe what is
likely to happen in the atmo-
sphere depends on the validity
of the calculations and also
their coverage including the
simultaneous effects of other
atmospheric pollutants.
•The uncertainties in the chemi-
cal rate coefficients, in
atmospheric transport, and in
the use of one-dimensional
models have been combined to
give an overall uncertainty
range of a factor of 6 within
a 95 percent confidence
level.
•The uncertainty range means
that for the case of continued
release of CFM's at the 1977
level there is 1 chance in 40
that the ozone depletion will
be less than 5 percent and 1
chance in 40 that it will be
greater than 28 percent.
•It is not therefore realistic
to assign overall uncertainty
limits to our calculated ozone
perturbations; deficiencies in
our basic knowledge of the
processes establishing the
composition of the stratos-
phere and in the modelling
technology cast doubts on
their validity.
D-3
-------
NAS/UK DOE Comparison
•There have been considerable
improvements in the computer
model and in the laboratory
and atmospheric measurements,
which have reduced the uncer-
tainty range.
•The STRAC report [The UK
Stratospheric Research Ad-
visory Committee] deals ex-
tensively with the uncertain-
ties in the model results.
Not all of them could be asse-
ssed quantitatively and it is
not possible to assign error
ranges to these estimates that
allow for all the uncertain-
ties. These have, however,
widened rather than narrowed
since Pollution Paper 5 was
published [1976].
•Even allowing for the best
professional judgment of the
possibility that some impor-
tant chemical reaction has
been overlooked or that
there remain large errors
in the measured chemical
rate coefficients, we
believe that there is a 3
out of 4 chance that con-
tinued release of CFM's
at the 1977 level will
result in an ozone deple-
tion that lies in the
range of 9 to 24 percent.
•Any uncertainty analysis would
be incomplete as it would
encompass only quantifiable
sources of uncertainty.
In contrast the uncertainty
from the different assumptions
made in simulating the trans-
port processes and in the
modelling procedures cannot
yet be quantified, nor can any
firm statement be made regard-
ing the uncertainties arising
from possible deficiencies in
our knowledge of the detailed
chemistry and photochemistry.
D-4
-------
NAS/UK DOE Comparison
•Although there are a few excep-
tions, the comparison between
the models and measurements
of substances in the present
stratosphere is considered
to be satisfactory within the
uncertainties of the measure-
ments. We, therefore, believe
that the projections for
depletion are valid
within the stated uncertainty
ranges.
•A certain amount of success in
the simulations has clearly
been achieved both in the
STRAC programme and elsewhere
but there are major discrepan-
cies. Moreover, unless one is
prepared to reject evidence of
actual measurements in favor
of theoretical calculations it
must be concluded that there
are still important gaps in
our understanding and know-
ledge of the process that
determine stratospheric com-
position. It follows that if
we are not satisfied with our
present ability to simulate
the particularly on the
grounds that there may be
major error or omissions in
current theory, we must be
very cautious in accepting
quantitatively any predictions
from accompanying perturbation
experiments.
•These findings, together with
other discrepancies between
model calculations and mea-
surements brings into question
the validity of the models
presently used to predict
ozone perturbations.
The UK report concludes "In the light of the many
uncertainties still prevailing such a reduction [the voluntary
steps underway in the EEC to reduce aerosol propellant uses of
CFC's by 30 percent from the 1976 level by 1982 ] appears to be
adequate pending further
warranted at present."
research. Strict regulation is not
The United States already
propellant uses of CFCs.
has banned essentially all aerosol
D-5
-------
X. APPENDIX E
CHLOROFLUOROCARBONS AND OZONE - THE SCIENCE
E-l
-------
Science
Table of Contents
Paqe
1. SUMMARY AND CONCLUSIONS 5
A. MODELS DO NOT PREDICT FUTURE REALITY 5
B. IT IS UNREASONABLE TO EXPECT ACCURATE 6
FUTURE PREDICTIONS FROM MODELS
C. CURRENT MODEL RESULTS INDICATE A SMALLER 6
POTENTIAL FOR FUTURE OZONE DEPLETION THAN
CITED BY EPA; ADDITIONAL REDUCTIONS ARE LIKELY
D. REGULATORY ACTION AGAINST CFCS IS NOT NEEDED AT 8
THIS TIME TO PROTECT STRATOSPHERIC OZONE
2. CRITICAL REVIEW OF CURRENT CHLOROFLUOROCARBON/OZONE 10
SCIENCE
A. OZONE OBSERVATIONS 10
1. Statistical Analyses 11
2. Umkehr Measurements 14
3. Tests of Other Model Calculations 15
B. CHLOROFLUOROCARBON EMISSIONS AND TROPOSPHERIC 16
PROCESSES
1. Atmospheric Lifetime Experiment 16
2. Tropospheric Sinks 18
3. Chlorine Removal by Rainout 20
E-2
-------
Science
C. TRANSPORT IN THE ATMOSPHERE 21
. 1. Local Averaging 21
2. Eddy Diffusion 22
3. The Assumptions 23
4. Testing the Approximations 23
5. Estimating Uncertainty Due to Transport 25
6. 2-D Transport 26
D. STRATOSPHERIC CHEMISTRY 27
1. Measurement Uncertainty 28
2. Temperature and Pressure Dependence 28
3. Product Identification 29
4. Third-Body Dependencies 30
5. Pressure Dependencies 31
6. H0x Chemistry 32
7. Excited States 35
8. Solar Flux 36
9. Unknown Chemistry 36
E. STRUCTURE AND SCOPE OF MODELS 38
1. Stratospheric Chlorine Inputs 39
2. Other Time-Varying Chemical Inputs 40
3. Coupled Nature of Atmospheric Processes 42
4. Atmospheric Variability 43
F. MODEL RESULTS 47
1. Comparison of Calculated and Measured 47
Concentrations
2. Comparison of 2-D and 1-D Model Results 51
3. Current Model Ozone Depletion Calculations 52
4. The Effects of Postponed Regulations 55
G. OZONE AND ULTRAVIOLET RADIATION 57
1. Latitudinal Variation in Calculated 57
Depletion
-------
Science
2. Natural Ultraviolet Radiation 58
3. Action Spectra 59
4. Dose and Dose-Rate 61
H. UNCERTAINTIES 62
1. Current Uncertainties 63
2. Reduction of Uncertainty 67
E-4
-------
Science
1. SUMMARY AND CONCLUSIONS
To appreciate both the general status of atmospheric
science and the impact of recent developments, some synthesis is
necessary. The following conclusions are supported by the de-
tailed discussion in section 2.
A. CONTRARY TO A POPULAR MISCONCEPTION, MODELS DO NOT PREDICT
FUTURE REALITY.
Model calculations account only for those aspects of the
future which are specifically included as input. Thus, all
fluxes of chemical species to the atmosphere are held constant
beyond some given year (usually the present) in all steady-state
calculations. Even for time-dependent scenarios, calculations
generally include only changes in CFC flux. Where specific
calculations involving more than one potential perturbation are
done, the combined effects are found not to be simply additive.
Even if projected variations could be included simultaneously,
the calculated results are limited by the accuracy of the input
projections - none of which are well defined.'
Model success in prognostic applications must neces-
sarily be preceded by diagnostic success. Yet, models fail to
reproduce the current best (if imprecise) picture of the present
day stratosphere based on measurements. A particular example is
lower stratospheric CIO concentrations - overestimated by models,
and directly responsible for much of the model-calculated ozone
depletion. Ozone measurements also do not confirm the changes
predicted by models for either short- or long-term perturbations.
E-5
-------
Science
B. ON THE BASIS OF PRESENT KNOWLEDGE, IT IS UNREASONABLE TO
EVEN EXPECT ACCURATE FUTURE PREDICTIONS FROM MODELS.
The uncertainties associated with models are numerous,
large, and often inherently unquantifiable. These uncertainties
arise from model inputs, model design, calculated simplifica-
tions, and incomplete testing of assumptions. The chlorine
contribution to the stratosphere from CFCs is not well-known.
Critical reaction rates are often poorly defined. The treatment
of transport is uncertain.
There are also many potentially important interactions
which are simply not understood. New chemistry is still possible
and indeed being discovered. The 1-D approximation is only now
being tested, and even at that it is being tested only by a 2-D
approximation. Local variability on short temporal and spatial
scales is ignored for lack of both information and a way to study
it. Feedbacks are omitted or guessed. It is simply unreasonable
to ask more of models than they are capable of delivering. The
failure of models to reproduce the present signals an immediate
caution on any extrapolations.
C. CURRENT MODEL RESULTS, TO WHATEVER EXTENT THEY CAN BE RELIED
UPON, NOW INDICATE A SMALLER POTENTIAL FOR FUTURE ADVERSE
EFFECTS THAN THAT GIVEN IN 1979; PRELIMINARY RESEARCH INDI-
CATES FURTHER REDUCTIONS TO COME.
According to a recent publication [Wine et al. , 1980],
revised chemical rate data has reduced the LLL model base case
calculated ozone depletion to 13.9%. The NAS [1979a] reported
the earlier LLL result, 18.6%, and revised it downward to 16.5%,
based on expectation of small tropospheric sinks and feedback
mechanisms. A similar treatment prorated to the more recent base
case of 13.9% would revise it downward to approximately 12.3%.
However, as discussed elsewhere, we believe the NAS treatment to
E-6
-------
Science
be inadequate (an attempt to quantify the inherently unquantifi-
able) . Therefore, we will use the 13.9% as a base case model
calculation for the ensuing discussion.
Other revisions, some to be recommended in the next NASA
revision, and others awaiting confirmation, would improve model
agreement with measurements of CIO concentrations by reducing
lower stratospheric OH levels. The concomitant reduction in
calculated steady-state ozone depletion reduces the figure to
below 6-9%.
Preliminary Atmospheric Lifetime Experiment (ALE)
results [CMA, 1980a] indicate a lifetime for CFC-11 of half that
derived from known stratospheric destruction, implying a
tropospheric sink equal to that in the stratosphere. If the
preliminary estimate is correct, ALE will confirm the lifetime of
CFC-11 by 1982. If CFC-12 behaves similarly, as one would
expect, the chlorine contribution of CFCs-11 and 12, and the
calculated ozone depletion would both be halved. Along with the
other likely revisions, this reduces the figure to less than 5%
depletion at steady-state.
The expected doubling in C02 is calculated to reduce
calculated CFC-induced depletion by about 3-5%. That is, on the
basis of recent developments and more complete analysis, the net
model-calculated potential future ozone change due to continued
release of CFCs at recent levels may well be as low as 2% when
current results are confirmed by additional measurements.
A 2-D model calculation of latitudinal depletion
distribution has shown that the physical amplification factor for
conversion to a DUV change is near 1 rather than 2 or more [Pyle
and Derwent, 1980]. Thus the risk of a 44% increase in DUV,
estimated a year ago, may now be reasonably assessed as a risk of
as little as a 2% change in DUV. (The most recent Du Pont 2-D
-------
Science
model results suggest that the latitudinal variation in calcula-
ted depletion may be considerably less than that of Pyle and
Derwent [1980], and the physical amplification factor may lie
between 1 and 2).
The reduction from 18.6% to as low as 2% or below for
steady-state depletion implies a comparable reduction in
calculated present day depletion to a few tenths of a percent -
reducing the disagreement with the positive trend observed with
ozone trend analysis. Thus, ozone measurements lend further
credence to the assessment that the theoretical risk is now far
less than that perceived in 1979 by the NAS panel.
D. THE GOAL OF MAINTAINING STRATOSPHERIC OZONE CONCENTRATION AT
NEAR CURRENT LEVELS CAN BE MET WITHOUT IMMEDIATE REGULATORY
ACTION AGAINST CHLOROFLUOROCARBONS.
Significant reductions in the perceived risk based on
model-calculated ozone depletion reduce the need for immediate
regulatory decisions. Increased awareness of uncertainties at
the same time decreases the risk of a wrong decision. Ozone
measurements provide reassuring evidence of the situation as it
really exists at present - a small long-term trend toward
increasing ozone. This result is in sharp contrast with
now-dated EPA-cited calculations, and is indirect confirmation of
preliminary research results which would further reduce the
theoretical risk.
The near future promises many significant tests of the
theory and improvements in modeling capability. The additional
depletion associated with even a five-year delay in U.S. regu-
lation was small even before recent model revisions. Thus, the
benefits of a delayed decision, scientific as well.as economic,
combined with the likelihood of a significantly better decision
-------
Science
with each passing year, far outweigh the very small current
estimates of the risk arising from delay.
And even overriding those issues is what appears to be
the most significant development to date. Refinements in ozone
trend analysis have given us a capable early warning system for
long-term changes in ozone concentration, no matter what their
cause. Those changes, if they occur, will be detected at a level
of less than 1.5 percent [Reinsel e_t al. , 1980; St. John, 1980a;
1980b; St John et a_l. , 1980]. If CFCs are deemed to be the
cause, the ozone change can be contained to very low levels -
levels below those contemplated recently by EPA as potentially
achievable goals [Jellinek, 1980a; EPA, 1980d]. No immediate
regulation is required to better this goal.
E-9
-------
Science
2. CRITICAL REVIEW OF CURRENT CHLOROFLUOROCARBON/OZONE SCIENCE
In each of the subsections which follow, an area of
research is reviewed. Particular attention is given to recent
developments, remaining uncertainties, and conclusions justified
on the basis of present knowledge. Specific detailed support for
scientific arguments made elsewhere in the submission is provided
here.
A. OZONE OBSERVATIONS
Measurements of ozone have been suggested as a valuable
tool for both scientific and regulatory analyses of potential
ozone depletion [CMA, 1980a; Du Pont, 1978; 1980a; Ward, 1979;
Masten, 1980]. The information available from the ozone record,
via various analytical and statistical analyses, provides insight
into the factors affecting ozone concentration and a test of
theoretical predictions. The former, of course, leads to greater
understanding of natural ozone variations, simplifying the task
of discerning unnatural changes. The latter, both constrains to
some extent the confidence placed in predictive models, and
potentially leads to insights in modeling as well. Ozone mea-
surements, in and of themselves, do not obviate the need for a
scientific resolution of the theory of ozone depletion. The
science must be resolved to explain what is or is not happening.
Nonetheless, the use of ozone measurements as an early warning of
unusual change in ozone should be of value in regulatory analy-
ses , and the existence of an early warning system permits more
time to be taken to resolve the underlying science with confi-
dence that a dangerous problem is not developing.
Two recent reports [NAS, 1979a; NASA, 1979] have con-
sidered the early warning possibility in a limited application,
i.e., the potential for detecting a CFC-induced decrease in
ozone. The implicit premise in those analyses is that detection
E-10
-------
Science
of a trend (or lack thereof) is a useful exercise only if the
trend is large enough to permit determination of its cause. This
reasoning neglects the value of information on the simple
existence or nonexistence of a trend. That information should be
of particular value to a regulator faced with inadequate and
uncertain information regarding potential (and calculated
current) changes in ozone concentration.
The approach to the regulatory question chosen by the
NAS panel was direct: Assuming the then-current model results
were accurate predictions, could a regulator afford to wait until
the theory has been verified by ozone measurements before taking
action? Through arguments, some of which will be discussed
below, that question was answered in the negative. On that
basis, any current value of the measurements to regulatory
decision-making was dismissed. Yet, the theory may not be
correct; ozone may not change, or it may increase. The
regulator need not decide now that regulation be put off for 20
years. He or she must decide only whether to regulate now or to
postpone regulation for some additional period of research and
evaluation. Moreover, the regulator's concern should be
protection of the ozone layer,not protection from some
calculated inherent danger of CFCs.
In the following discussion, this broader application of
ozone measurements is discussed, along with the recent develop-
ments in their analysis. Ozone measurements do function as an
early warning system for long-term change in ozone concentration.
The importance of this contribution cannot be overlooked.
1.' Statistical Analyses
The type of ozone observational data available, and
developments in analysis of this data through 1979, have been
discussed in some detail by the NAS panel [1979a]. Several
-------
Science
averaging techniques for deriving global ozone rely on graphical
and visual analysis for trend detection, e.g., [Angell and
Korshover, 1978; 1980; Miller e_t al. , 1979]. The more recent
applications of statistical trend analysis [Hill and Sheldon,
1975; Hill et. a_l. , 1977] and' spectral analysis [NAS, 1979a] to
Dobson ozone measurements allow more specific and apparently more
sensitive approaches to trend detection. The most recent work by
three groups has contributed considerably to that relatively
young research area.
St. John [1980a; 1980b] has introduced . several
refinements to the original methods of Hill [Hill and Sheldon,
1975; Hill £t al_. f 1977]. Rather than the zero trend followed by
linear trend model for CFM induced depletion, St. John regresses
directly against a 1-D atmospheric-model-calculated depletion
curve. The regression coefficient or CFM-parameter may be
directly translated into a net ozone change occurring over the
period 1958-79. In further testing of the data, St. John notes
that differentiation of the CFM curve from other similar slowly
changing long-term trends is difficult given the near absence of
any trend. Thus the estimated trend is better described as the
overall long-term net trend in ozone.
For the period 1958-1978 St. John finds a slight but
statistically insignificant increase in ozone of 0.3 + 1.2%.
Stated uncertainties are 95 percent confidence limits, and are
derived from variability among the estimates for individual
Dobson stations. This approach to uncertainty defines
statistically the uncertainties associated with any long-term
effects that vary from one station to the next, and furthermore
it removes those effects as contribution to the observed trend,
isolating them as uncertainties. It is not necessary, therefore,
to rely on estimates of those uncertainties. Indeed, the NAS
estimates for instrumental uncertainties, meteorological bias,
E-12
-------
Science
and geographical biases all appear to be conservative over-
estimates when compared to the true data-derived uncertainties.
Reinsel e_t al. [1980] at the University of Wisconsin
have applied similar methods using the linear trend represen-
tation for the CFM trend, and for the period 1970-1978 find an
increase of 0.3 + 1.4%. Included in that uncertainty is a
conservative estimate of correlations (i.e., lack of indepen-
dence) among Dobson stations clustered geographically. Other-
wise, Reinsel's et al. and St. John's results are comparable and
in notably good agreement.
The Wisconsin group has also examined the more complete
satellite data now available in order to test the global
representativeness of their chosen set of 36 Dobson stations.
Satellite data display a negative time trend with respect to
ground based measurements. The investigators in the satellite
measurement program concur that a degradation in satellite
performance took place, and that the satellite trend is not
an indication of an actual decrease in ozone [Cunnold, 1980;
Stolarski, 1980]. The degradation may, in fact, hide an
increasing trend observed in the Dobson measurements.
Nonetheless, the global satellite trend may provide useful
information. A comparison of Dobson station data with the
satellite data for the same isolated geographical regions shows a
difference identical to that found with the global satellite
data. Thus, the Dobson station locations are shown to be
representative of the global trend in ozone. While no inference
is possible regarding absolute calibration, the Dobson locations
would appear to reproduce global trends with no geographical
distribution bias.
Watson [Watson, G.S., 1980] has also continued the
spectral analysis of the Dobson ozone record discussed by NAS
[1979a] , and reports a change in the seventies of +0.6 +_ 1.9% in
E-13
-------
Science
ozone concentration. The uncertainty stated here is somewhat
broader than those discussed above, including some partial
measure of the possibility a trend is masked by natural
variations. Again, however,, the results are similar to those of
other statistical techniques.
In conclusion, ozone appears to be increasing, and none
of the estimated 2v uncertainty ranges includes the 1979 calcu-
lated "to date" ozone depletion of 2.1%. Recent changes to the
model (discussed in ensuing sections) reduce those calculated
figures enough to move them to within the confidence range for
measured ozone. In that sense, the ozone measurements support
the appropriateness of such model revisions, in which case the
theory as cited by EPA overestimates depletion.
&
There is, of course, the alternative possibility that
another long-term trend may be offsetting CFC induced depletion,
resulting in little net change to date of global ozone. Either
possibility has important regulatory implications, but must be
considered along with other information available from ozone
measurements.
2. Umkehr Measurements
As pointed out by the NAS panel [1979a] , Umkehr measure-
ments of ozone vertical concentration profiles provide access to
measurements of local ozone concentration in the region where the
percentage impact calculated to have occurred is largest -
approximately 5% near 40 km. Angell and Korshover [1980] have
examined data through 1979 for the 32 - 46 km layer and conclude
"there is still no evidence of an anthropogen i ca 1 ly-_i nduc_ed_
decrease in ozone in this sensitive layer." The Princeton group
has begun analysis of the Umkehr data, and initially have noted
an ozone increase during the 1970's. These are particularly
interesting results in light of the general confidence often
E-14
-------
Science
expressed that stratospheric chemistry dominates transport at
these altitudes, and is also well understood there. A conflict
with model calculations for the upper stratosphere is, therefore,
a more serious challenge to the theory. The significance of that
challenge will be better defined as the analysis is completed and
uncertainty limits are published within the next year or two.
3. Tests of Other Model Calculations
Based on model calculations numerous other stratospheric
perturbations are thought to influence ozone concentration.
Aside from the phenomenon known as sudden stratospheric warming,
the NAS panel notes that "strong evidence is lacking for other
natural or man-induced perturbations to the ozone distribution."
These include volcanic activity [Steed e_t al. , 1980; Angell and
Korshover, 1973; 1978], nuclear explosions [Angell and Korshover,
1973; 1978], and the 11-year solar cycle [Angell and Korshover,
1973; 1978]. The accumulated evidence suggests that the
discrepancies- between models and measurements for ozone
perturbations extend beyond the CFC-effect to include other
aspects of model chemistry or feedback through atmospheric
dynamics ignored completely in the models.
Taken together, the accumulated evidence available from
the ozone measurement record leads to considerable skepticism
concerning model calculations - skepticism that is justified by
the extensive and increased model uncertainty at the present
time.
With this background we turn now to an examination of
the various steps in the theory and the status of the model
inputs and results.
S-15
-------
Science
B. CHLOROFLUOROCARBON EMISSIONS AND TROPOSPHERIC PROCESSES
The initial determinant in quantifying potential
stratospheric ozone depletion is, of course, that amount of
chlorine contributed to the stratosphere by CFC emissions. The
first controlling factor is production and release. Much
attention has been paid to release scenarios in official studies
[NAS, 1979a; 1979b; Rand, 1980; UK DOE, 1979] and certain factual
errors have been noted [Du Pont, 1980a; Masten, 1980]. The fact
is that 1979 production of CFCs-11 and 12 is down by 17.5 percent
from peak levels of 1974 [CMA, 1980a; 1980b]. Unrestricted
growth is an unrealistic assumption given current regulatory
initiatives elsewhere and continued expected regulatory concern.
Concern over potential future regulation has a large dampening
impact on growth because of the perceived extremely high risk
associated with any' significant capital investment in a
threatened product. The only reasonable conclusion is that
growth in production will be limited (by current capacity and
regulatory concern) and will be relatively slow, unless and until
the science surrounding the Chlorofluorocarbon/Ozone Theory is
satisfactorily resolved.
1. Atmospheric Lifetime Experiment
A second factor influencing the stratospheric chlorine
contribution is transport through the troposphere. NAS
acknowledges the possibility of some CFC destruction in the
troposphere, and indeed revised its central value for
steady-state depletion downward to account, for its likelihood
[NAS, 1979a]. However, the magnitude of this "correction" was
determined arbitrarily after attempting to enumerate all the
possibilities and assign lov/er bounds to each. Given the paucity
of available research results, such a task is formidable.
E-16
-------
Science
A more direct route to the answer on the tropospheric
sink question is available from the Atmospheric Lifetime Experi-
ment, currently underway under CMA sponsorship [CMA, 1930d].
Although no results were available to the NAS Panel, a report of
interim results was made to EPA [CMA, 1980a] and the scope of the
program and its goals were clearly defined in the published
literature [Cunnold e_t al. , 1978], In short, a series of
frequent, carefully calibrated measurements is being made over a
period of at least three years to determine the rate at which
atmospheric CFC concentrations are increasing. Along with
accurate production figures and use distributions, the measured
trend allows determination of the atmospheric lifetimes for the
compounds. To the extent that this lifetime is smaller than that
calculated assuming only stratospheric destruction, the CFCs are
being destroyed by some other mechanism. In the trend method of
data analysis, uncertainty limits rapidly converge as the data
set lengthens. A total of three years of measurements will be
sufficient to identify a total lifetime significantly shorter
than the stratospheric lifetime if current data quality is
maintained [CAP Associates, 1980].
The advantages indicated by the trend method of analysis
are both the rapid convergence of uncertainty limits and the
independence of the result from absolute calibration of the
measurements. This becomes especially important given the
current data which indicate more CFC-11 and CFC-12 in the
atmosphere than calculated emissions would account for. Both the
stratospheric sink and a possible tropospheric sink would
increase this discrepancy. Nonetheless, the trend method, which
is independent of absolute calibration, indicates a lifetime
shorter than that based on stratospheric photolysis alone. The
atmospheric overburden indicated by the budget or global burden
analysis is currently under investigation to identify the source
of the discrepancy.
E-17
-------
Science
A best fit of the trend in the data set available to
date (18 months) indicates a lifetime approximately one half of
that due to known stratospheric sinks for CFC-11. CFC-12 is
expected to behave similarly to CFC-11 but to exhibit somewhat
more stability to any sink mechanism. The uncertainty range
remains very wide with the short data set available currently.
However, the significance of the method lies in the rapid
convergence of those limits. Within two more years a tropo-
spheric sink important enough to drastically reduce calculated
ozone depletion will be indicated if its exists—as preliminary
results indicate.
2. Tropospheric Sinks
It will, of course, be equally important to eventually
justify such a result by demonstrating a sink mechanism. Work by
Ausloos and Rebbert [1980] has identified destruction of CFCs-11
and 12 absorbed on sand and irradiated with ultraviolet light.
Interpretational difficulties arise in attempted extrapolation of
the laboratory results to the real atmosphere, however. The
experimental rate of destruction was found to be sensitive to
parameters such as relative humidity. The appropriate conditions
for atmospheric encounters of chlorofluorocarbon and sand are not
known.
An alternative approach to an understanding of the sand
mechanism and assessment of its significance has arisen from the
work of Crescentini and Bruner [1979; 1980]. Bruner has noted
large variability in CFC-21 measurements as a function of season,
location, and wind direction. CFC-21 has been suggested as a
possible decomposition product for CFC-11 [Glasgow et al. , 1977]
and its concentration has been measured by several groups [Singh
e± al. , 1977; Penkett e_t al. , 1980; Crescentini and Bruner,
1979] . The concentration is often found to exceed expected upper
limits. (According to current understanding, CFC-21 occurs only
E-1C
-------
Science
as a minor by-product in chemical manufacturing processes).
Crescentini and Bruner [1980] noted a correlation of high CFC-21
concentrations with air masses which have recently traversed
desert regions and of low concentrations with clean ocean air.
The work is currently being extended to further define the
correlation and search for corresponding decreases in CFC-11
concentration. The research also is information which has been
made available to the EPA, although not in advance of the NAS
report.
CFC-22, which is a commercial product, has also been
detected in the atmosphere at concentrations well above those
expected on the basis of available production estimates. In
analogy with the possible formation of CFC-21 from CFC-11, it is
also possible that CFC-12 decomposes by some mechanism to form
CFC-22.
The work is indicative of the progress being made to
determine experimentally whether tropospheric CFC sinks do indeed
exist. Although the NAS minimized the likelihood of significant
sinks by assuming only a small one was possible, the evidence
supporting the conclusion was not exhaustive by any means and
represents an element of acknowledged subjectivity in the report.
Other work in progress includes consideration of a sink not
thoroughly examined previously. A group at Du Pont is studying
possible destruction of CFCs at plant surfaces. The available
surface area represents over six times the total world land
surface and provides a reducing medium that may be capable of
inducing reaction of the compounds. Even a slow reaction on such
a scale could have profound effects.
The troposphere is a much more complicated chemical
system than the stratosphere and is correspondingly less well
understood. One would expect tropospheric influence on potential
ozone depletion to be small, but the two known areas of impor-
E-IS
-------
Science
tance contribute to a major uncertainty. Both the quantity of
CFCs which reaches the stratosphere undestroyed and (see below)
the quantity of chlorine leaving the lower stratosphere with
rainfall are open to question. Current information indicates
preliminarily that as little as half the currently assumed con-
tribution of CFCs is actually realized in the stratosphere.
Nonetheless, within as short a time as two years, research will
provide quantitative evidence for a tropospheric sink if it
exists. An understanding of the sink, if it is verified to
exist, may be somewhat slower in coming.
3. Chlorine Removal by Rainout
A final area of some uncertainty in stratospheric
chlorine contribution is the rate and maximum altitude for
removal of soluble atmospheric gases in rainfall. The concen-
tration of chlorine in the lower stratosphere, where most ozone
depletion is calculated to take place, is tightly coupled to the
maximum altitude of rainout in the computer model atmosphere
[Miller et al., 1980a] . The real world data available provides
little guidance to modelers, leaving significant uncertainty in
this model parameter.
E-20
-------
Science
C. TRANSPORT IN THE ATMOSPHERE
An understanding of how chemical species move around in
the atmosphere is critical to any evaluation of the net effects
of a given species on the atmospheric chemical system. The rate
at which any simple chemical reaction proceeds is governed by the
local concentrations of the reacting species. Those concentra-
tions are in turn determined by previous chemistry and by trans-
port into and out of the region of interest. Both the time
increments and the distance increments over which large concen-
tration fluctuations may occur are small. In addition, the
problem is a three-dimensional one. The appropriate tool for the
problem is a three-dimensional model, with time increments of the
order of seconds, and distance increments of centimeters to
meters -- a model which is not likely to be developed for
decades. Thus research on atmospheric processes has necessitated
the development of simplified models employing a variety of
approximations. This section examines the transport parameteri-
zation of one- and two-dimensional models, noting the critical
assumptions and associated uncertainties.
1. Local Averaging
It is not feasible to treat the extremely small-scale
variations in transport velocities and species concentrations in
a computational model of the atmosphere, although such variations
may be as large as the average quantities (e.g., direction
reversal for winds). With respect to transport, such variations
are primarily important insofar as they compromise other
assumptions the modeler must make to deal with the variations,
i.e., eddy diffusion. This subject is discussed below. The
effects of such averaging on chemical terms in the model is
addressed in the section on atmospheric variability (E-4).
E-21
-------
Science
2. Eddy Diffusion
In two-dimensional models, one averages not only over
local variations but also over longitude. The resulting mass
balance equations for each chemical species contain not only the
mean vertical and latitudinal motions, but also covariance terms
which are averages of the product of fluctuations in wind
velocities and fluctuations in species concentrations. If
velocities and concentrations are simultaneously large, net
transport is large, etc. However, the true atmosphere has not
been measured, as the average covariance is not well defined.
Reasoning that net transport occurs only if the initial
concentration is different in two adjacent regions, and further
that such transport is most likely to be determined by the growth
in mean concentration, Reed and German [1965] introduced the
so-called "K-theory" in which eddy fluxes (i.e., transport in
addition to mean flow) in the meridional plane are proportional
to mixing ratio gradients. The constants of proportionality,
Kw' Kw7» K-7w' an(3 %•-„ are assumed to be species-independent, but
yy yz zy zz
functions of latitude, altitude, and time of year. These
assumptions provide "closure" to the mass balance equation. That
is, they permit calculation without knowledge of the fluctuations
themselves or of their covariance. One needs only averaged
mixing ratios and some information on the magnitude of the pro-
portionality constants.
/
For one-dimensional models the averaging is extended to
include latitude -- leaving only altitude as a spatial variable.
The covariance terms now include latitudinal fluctuations, as
well as longitudinal, temporal, and small-scale local fluctua-
tions. Here the entire set of covariance terms is assumed to be
expressible as the product of a single diffusion coefficient, KZ,
and the vertical concentration gradient. With this final stage
of averaging, there is no mean flow term remaining, since there
E-22
-------
Science
is no overall transport of mass in the vertical direction either
away from or toward the earth's surface. In one-dimensional
models, eddy diffusion is the only transport to be considered.
3. The Assumptions
The National Academy of Sciences report outlines the
assumptions inherent in eddy diffusion:
"...to justify the use of the diffusion
approximation in a 1-D model, the mean correlation
(between velocity fluctuations and concentration
fluctuations) must not only be proportional to the
negative gradient of the mean mixing ratio for all
species, but the coefficient of proportionalities,
K, must also be independent of the species." [NAS,
1979a, p. 51].
Of course, a similar statement may be made regarding two-dimen-
sional eddies. No direct tests are available for these assump-
tions, but caution is in order.
The diffusion coefficients are not necessarily truly
independent of the species. In recent theoretical work on the
two-dimensional eddy-diffusion approximation, Pyle and Rogers
[1980a; 1980b] note that the assumption of independence is valid
only for very long-lived or inert tracers. To the extent that a
species reacts, the diffusion coefficient may be chemistry-
dependent. For very reactive species, chemistry could dominate
creating serious errors. The principles apply equally well to
the further simplified eddies of 1-D models.
4. Testing the Approximations
The danger of acccepting the eddy diffusion approxima-
\
tion blindly is equally well noted by NAS:
-------
Science
"It is not very probable that such stringent conditions
[see previous quotation] can be met in view of the fact
that photochemical sources and sinks, vertical fluxes,
and chemical reactions all vary with latitude, longti-
tude, and height for each gas, giving rise to many
nonlinear effects." [NAS, 1979a, p. 51].
The panel concludes, however, that the approximation would still
be useful (for 1-D) if K profiles derived by fitting measurements
of upward moving trace gases are similar in shape and comparable
in magnitude to those derived from downward fluxing trace gases.
In fact, for the several long-lived gases examined, the
comparison is reasonably good.
However, the rationale presented has a major flaw. As
indicated by Pyle and Rogers [1980a; 1980b] , it is just these
long-lived species, such as N-0, 0., and CH4, for which the
approximation should be most applicable. Yet it is the
short-lived, highly reactive species that may influence ozone
concentrations, and, therefore, which must be accurately modeled.
It is satisfying that upward and downward moving species behave
similarly, but not sufficiently so to justify application of the
same transport parameters to all atmospheric species.
The long-lived species are initially chosen for the test
by valid reasoning. Their concentrations are largely governed by
transport and boundary conditions, with fairly well understood
chemical sources and sinks. Shorter-lived species are themselves
far more dependent on chemistry, and any fitting attempt would be
as much a test of model chemistry as of model transport, and
therefore of little value as a test of transport parameteri-
zation. If eddy transport itself is a function of chemistry as
Pyle argues, eddy transport of short-lived species remains as an
untested assumption in current models.
£-24
-------
Science
5. Estimating Uncertainty Due to Transport
The uncertainty in calculated ozone depletion arising
from transport is taken by NAS to be dominated by the effect on
CFC lifetimes. Since the CFCs are long-lived trace species, this
aspect of the uncertainty has indeed been tested by their
analysis and may well be correct. That all other uncertainties
are smaller has not been demonstrated.
A change in the constant K profile for all species is
said "generally" to provide a negative feedback on the change in
ozone destruction, so that the uncertainty in A. Q-^ is less than
or equal to that in K. This ignores the uncertainty in K-theory
itself, i.e., the assumption that it is a reasonable way of
expressing atmospheric transport. Also discarded is the
possibility that the theory is basically sound, but the
coefficients are strongly species dependent. The NAS reasoning
is as follows:
"It is possible, however, if transport errors in
different species were to add randomly rather than in
the correlated fashion of the 1-D models, that the
overall uncertainty in ozone change could be somewhat
larger than the uncertainty in CFM lifetimes. We
believe it likely, however, that the uncertainties due
to transport are smaller than the uncertainty in CFM
lifetimes and that our error estimate [a factor of 1 +_
0.3] is a conservative one." [NAS, 1979a, p. 58].
Although the results of Pyle were not yet available to the NAS,
the possibility of species-dependent transport was known, but was
entirely untested for the most critical species, and was dis-
carded completely on the sole basis of belief.
E-25
-------
Science
6. 2-D Transport
The theory of 2-D eddies has been discussed already, and
the problems are similar to those of 1-D eddy diffusion except
for those associated specifically with latitudinal averaging.
Yet while the four-parameter (K , K , K , K ) expression of
yy y £ zy z z
eddy transport perhaps allows for more realism, it also requires
the derivation of four parameters rather than one. The limita-
tions of the data base alluded to by NAS for 1-D eddy diffusion
therefore become more critical, and the individual parameters
less well defined.
An additional requirement in 2-D modeling is a parame-
terization for mean flow in the meridional plane. It has been
argued [Pyle and Rogers, 1980a; 1980b] that, in the Eulerian
representation, the eddy diffusion tensor must be non-symmetric
(i.e., K ? Kzv^ * Tne implied advective transport could instead
be incorporated with the mean flow in a Langrangian representa-
tion, leaving the K-tensor symmetric and reducing by one the
number of necessary parameters [Miller e_t al. , 1980a, 1980b; Pyle
and Rogers, 1980a; 1980b] . Two-dimensional modeling remains a
very young research area, though, and it is likely that expansion
of the data base will permit (in the future if not already) 2-D
models to further improve transport representation and, at the
same time, provide a much more realistic simulation of the atmo-
sphere than that in 1-D models. Both 1-D and 2-D models, how-
ever, will still be subject to the untested assumptions accom-
panying the eddy diffusion theory.
E-25
-------
Science
D. STRATOSPHERIC CHEMISTRY
While tropospheric processes, transport, and model
construction all influence the magnitude of calculated ozone
reduction, the process if it occurs is a chemical one. The
active catalytic species are generated and lost in chemical and
photochemical reactions, the ozone destruction cycle is chemical,
and the chemical interactions among all the atmospheric trace
species govern the degree of effectiveness of the cycle.
Furthermore, despite the confidence expressed by the NAS Panel on
Stratospheric Chemistry and Transport in the body of chemical
knowledge assembled for their report [1979a], it was chemistry
which contributed the major uncertainty according to their
analysis.
Since the 1979 report, research in kinetics and photo-
chemistry has provided even more evidence of the weaknesses in
the current data base. Several new results and remeasurements of
"well-known" reactions underline the conclusion of Smith [1978]
that the uncertainties have been repeatedly underestimated. Pos-
sible reasons for the underestimation are not difficult to find.
Available techniques are continually pushed to the limit of their
capabilities. The atmospheric system is not amenable to simple
duplication in the laboratory. Finally, the questions being
asked of kineticists and photochemists are demanding. The
molecular interactions of radical species are not nearly so well
understood mechanistically as those of more stable molecules, and
conventional extrapolations over temperature and pressure ranges
may often be unjustified. Products are not always the expected
ones, and even minor product branching may have major con-
sequences.
E-27
-------
Science
1. Measurement Uncertainty
The first source of uncertainty which must be considered
in any evaluation is simply the degree of confidence in current
data -- soundness of techniques, carefulness of the experi-
menters, and accurate expression of confidence limits. The
number of reactions to be considered makes this alone a source of
major uncertainty in model-calculated depletion. To that must
be added the effect of sensitivity of calculated depletion, both
to individual errors and to combination of errors. As will be
explained below for the case of HC^NC^/ a significant error in
one reaction leads to underestimated sensitivities for other
reactions.
2. Temperature and Pressure Dependencies
Next, the data base itself suffers problems in its ap-
plication to the stratosphere. Reaction rates have rarely been
measured under the full range of temperature and pressure con-
ditions present in the stratosphere. The most direct, and
presumably most accurate, laboratory techniques require low
pressures, making the high pressure, low tempeprature conditions
of the lower stratosphere almost impossible to simulate. Yet it
is in the lower stratosphere where the most complicated chemistry
is important and where most ozone depletion has been calculated
to take place. Many reactions in the current data set have never
been examined for pressure dependence, yet similar reactions have
been (surprisingly) found or postulated to exhibit pressure
dependence (e.g., OH + CO, HO- + H02). Unusual temperature
dependencies abound — H02 + H02, H02 + CIO, OH + HNOj, for
example. In many cases, it is apparent that reaction mechanisms
are simply not well understood, and a great deal of research
remains needed to solidify the data base. Such questions are
critical when considering the extensive coupling among chemical
families and the fact that many of the reactions involved are
I? _-.Q
ij — £. O
-------
Science
among those to which calculated depletion is most sensitive. The
results of the past year for OH + H202 [Keyser, 1980; Sridharan
et al., 1980] and OH + HN03 [Wine et al., 1980] both lie outside
95 percent confidence range given in the NASA [1979a] evaluation.
The NAS estimation of uncertainty in A CK was, of course, based
upon such evaluations.
3. Product Identification
A third area of uncertainty minimized in earlier evalua-
tions is the identification of reaction products. Sze [1979] has
noted that if the reaction OH + CIO has a branch of as little as
10 percent to HCl + 02, the calculated ozone depletion is reduced
significantly. Yet model calculations assume the products to be
entirely H02 + Cl. Although product studies are in progress,
previous work limits that branch only to <" 35 percent [Leu and
Lin, 1979]. A similar situation exists for a possible HCl pro-
duct from the reaction H02 + CIO, although an upper limit of 3
percent has been established at room temperatures [Stimpfle et
al. , 1979; Leek ^t a_l. , 1980; Leu, 1980].
The significance of product identification was made more
apparent recently by Molina _et al. [1980] in their study of
chlorine nitrate formation. Previous work [See NASA, 1979] on
this reaction, CIO + N02 + M, had measured the rate by monitoring
reactant disappearance. Molina e_t al. monitored formation of
C10N02 product, and observed a rate one third that of the
previous studies. It has been postulated that the remaining
reactant disappearance leads to production of another isomeric
product whose chemistry in unknown, but which is likely to be a
less effective chlorine sink than chlorine nitrate. A less
effective sink could increase calculated ozone depletion.
Because of the major role played by ClON02 in current models of
E-29
-------
Science
the stratosphere and the potential for significant new chemistry
involving any isomer, the enlarged uncertainty here deserves
rapid attention.
Identification of photolysis products has also in many
cases not been resolved, with chlorine nitrate again as an
example [See NASA, 1979]. A more critical case is the possi-
bility of HCl production from HOCl photolysis, being studied
currently by Birks. The examples mentioned are by no means
exhaustive. As with temperature and pressure dependencies, a
careful review of the CODATA summary [1980] of evaluated data,
and comparison with NASA report RP-1049 [1979] reveals many
inadequacies in the data base. The hundreds of references
contained in those documents serve to identify the assumptions
incorporated in current models. As the new results continue to
point out, such assumptions are all too often not warranted.
4. Third-Body Dependencies
Hamilton and Lii [1977] and others [Lii e_t al. , 1979;
Cox and Burrows, 1979] , have found the reaction rate for H02 +
H02 to be dependent on the concentration of water vapor in the
reaction region and proposed that a complex, H02«H20, may
participate in the reaction, effectively catalyzing it. Con-
siderable study will be required to understand the mechanism of
such action, but a possible role for the complex in other H02
reactions is suggested by these results. A complex with water,
unless very strong, is of significance only in the troposphere
and lower stratosphere, given the relative aridity of the
stratosphere.
A potentially more important complex has been postulated
by Prasad [1980], In examining laboratory data on chlorine-
sensitized ozone decomposition, Prasad has found that apparent
E-30
-------
Science
inconsistencies can be explained in terms of a C10»02 complex.
The complex could involve significant amounts of CIO. The
sensitivity of such an association complex to temperature could
help to explain the large variability among the several in situ
measurements of stratospheric CIO by Anderson ejt al. , [1980a] .
Reactions of the complex might also serve to decrease the
efficiency of the chlorine catalytic cycle through the following
"do-nothing" cycle:
CIO + 02 + M * C10»02 + M
C10-02 + 0 >-OClO
OC10 + hv »> CIO + 0
A complex involving 02 would not be limited by 02 concentration,
but only by the temperature and pressure conditions required for
its formation. Recent work of Anderson et al., [1980b] suggests
that the role of this complex is very limited, but the notion of
oxygen complexes, like that of water complexes, remains generally
untested.
5. Pressure Dependencies
Negative temperature dependencies noted for several
radical-radical reactions also imply the possible role of
complexes as intermediates. Should that be the case, product
distributions would be strongly influenced by the structure of
the complexes, and product identification becomes even more
essential. Additional variation of rate with altitude again adds
to rate uncertainty.
Pressure dependence of 3-body reactions are also not
always well defined. An example is the reaction for chlorine
E-31
-------
Science
nitrate formation, where the fall-off in the observed rate does
not fit the theoretical fall-off curve [Watson, R. T., 1980].
6. HO^ Chemistry
X ~™ " ••" ' ' " " '
One might note from the preceding discussion that many
of the major types of uncertainty in stratospheric chemistry
involve HO species, OH and HO,. A review of those reactions to
X £f
which calculated ozone depletion is thought to be most sensitive
indicates that several of those reactions also involve either OH
or H02« Finally, many of the most surprising developments from
research, during the past year in particular, once again involve
the so-called odd hydrogen species. The chemistry of these
species, particularly HO,, [Watson, R. T., 1980], is not at all
well understood.
The principal interaction between odd hydrogen and odd
chlorine species is through the reaction:
HO + HC1 > H20 + Cl
This reaction controls the rate at which chlorine is removed from
the temporary sink HC1. As HCl, chlorine is isolated in a
relatively stable form, and does not contribute to depletion.
Chlorine atoms, on the other hand are a reactant in the catalytic
destruction cycle. A variety of atmospheric measurements
indicate indirectly that the model-calculated concentrations of
OH in the lower stratosphere are too high. Through the reaction
above, this leads to exaggerated amounts of calculated ozone
depletion. Consequently, special importance has been attached to
the study of reactions by which OH may be converted to H20 and
removed more rapidly from the stratosphere. Several such
reactions are currently being studied.
-------
Science
Perhaps the most important, and almost certainly the
most uncertain, of these reactions is:
OH + HOj >H2° + °2
A number of studies have been completed, and they generally fall
into two groups: studies by relatively direct methods at low
pressure giving a slow reaction rate [Hack et al. , 1978; Burrows
et al., 1978a; 1978b; Chang and Kaufman, 1978], and studies by
more indirect methods at higher (atmospheric) pressure giving
rates up to an order of magnitude faster [Demore, 1979;
Hochanadel et al., 1972]. Furthermore, the reaction is not
expected to be pressure-dependent, making resolution of the
differences difficult. At present, there is no direct method
available for the entire range of pressure necessary. The most
recent studies corroborate earlier indirect results [Kurylo,
1980; Demore, 1930]. A fast reaction rate, even if only at high
pressures and low temperatures (i.e., the lower stratosphere)
would significantly reduce calculated depletion. The value of
such reduction and those arising from other recent developments
will be discussed thoroughly in the section on model results.
A surprising result obtained recently by Wine et al. ,
[1980] is a significantly faster rate for the reaction OH + HNOj,
previously thought to be known quite well. In addition, a
negative temperature dependence was observed, making the reaction
especially important in the cool lower stratosphere. Kurylo
[1980] has recently given preliminary confirmation for the Wine
et al. , result, although Marinelli _et al. , [1980] measured a
somewhat slower rate. The unexpected Arrhenius parameters have
led the researchers to an investigation of the products. The
-------
Science
products currently assumed by modelers are H20 + NO-,/ but the
temperature dependence observed is uncommon for a simple hydrogen
abstraction reaction. Other products such as H202 and N02 are
possible, although Marinelli e_t al. [1980] have concluded that
H20 and NO-, are indeed the major products. Again here, the model
calculations are quite sensitive to this reaction and its
products.
In other HO chemistry, Molina and Molina [1980a; 1980b]
X
found that the association complex between H02 and N02, i.e.,
peroxynitric acid (H02N02), photolyzes more slowly than previous
results [Graham et. al. , 1978] had implied. As a consequence,
other reactions of the molecule assume greater significance in
atmospheric models. The LLL model used by NAS, along with
several other models, had omitted H02N02 from earlier calcula-
tions, assuming it photolyzed rapidly back to the initial
reactants. With the longer lifetime of H02N02, the reaction OH +
H02N02 may now provide another route for removal of odd hydrogen
from the lower stratosphere. Trevor et al. [1980] have used a
direct method to determine a very fast upper limit for the rate
of the OH + H02N02 reaction. Littlejohn and Johnston [1980] very
recently reported the first actual measurement of this rate, as 2
-12 3
+_ 1 x 10 cm /molecule - sec. — 5 times faster than the NASA
[1979] recommended rate. And Barker ejt al [1980] , in a continua-
tion of the Trevor study by an alternate method, have just
a
reported rate measurements of OH + H09N09 in the range of 4 to 5
-12
x 10 at four different temperatures. A fast rate for the OH +
HOpNO-, like other rate increases mentioned in this section,
again converts odd hydrogen to water more quickly and leads to
decreases in calculated depletion.
The significance of these developments goes beyond their
effects on calculated depletion. The changes came in unexpected
places. They belie the great confidence expressed in the current
recommended rate set used by modelers and raise questions about
E-34
-------
Science
the true range of uncertainty in calculated ozone depletion which
arises from "known chemistry".
7. Excited States
The role of excited states in stratospheric chemistry is
a question often discussed, but rarely with satisfactory
resolution. Sufficient excitation could produce dramatic
differences in reaction rates. Certainly, however, most excited
state species produced as a result of photolysis or greatly
exothermic reactions will be rapidly quenched. The 0( D) atom is
an exception already incorporated in models, although there
remains some doubt about the absolute quantum yields for its
production [NASA, 1979; Chang, 1980]. Electronically excited 02
and 0-, have been proposed, e.g., Prasad and Burton [1979], to
play a role, but no examples have yet been found. These are
particularly important in that they may provide additional
sources of odd oxygen (i.e., atomic oxygen and ozone).
Vibrationally excited OH radical is a likely product of several
reactions, but its survival relative to quenching is likely to be
small, and only a large source would make this species
significant.
There is also a more indirect way that excited states
may relate to our knowledge of stratospheric chemistry. In
laboratory kinetic studies the radicals of interest may also be
produced in excited states. Failure to equilibrate these species
sufficiently can lead to erroneous reaction rates. Ravishankara
and Wine [1980] note a likely example in their work on the Cl +
CH^ reaction. The two spin states of the ground electronic state
of chlorine atoms are separated in energy by such an amount that
the lower state reacts endothermically with methane, while the
upper state reaction is exothermic. Ravishankara postulates that
this may explain the difference apparent among previous studies
of the reaction since the measured rate will depend on the
E-35
-------
Science
relative populations of the two states. Furthermore, the rate
appropriate for use in stratospheric models will depend on the
stratospheric populations. Further consideration of these ideas
will certainly lead to refinements in stratospheric models or our
confidence in them, but at this stage, the ideas represent an
additional new source of uncertainty.
8. Solar Flux
Recent efforts to determine more accurately the solar
ultraviolet flux reaching the stratosphere are discussed by
Nicolet [1980] . Reductions from the flux levels currently used
by modelers have been considered. If incorporated, the rate of
ozone production is decreased, and several other photochemical
reactions are affected. The net result could be a slight change
in calculated ozone depletion. Once again the point to be
stressed is the significant uncertainty in model inputs.
9. Unknown Chemistry
Several other facets of stratospheric chemistry may have
potential impacts on ozone perturbations, but are currently not
well enough understood to permit conclusions. First among these
is the process of halocarbon oxidation. Current models 'assume
immediate release of all chlorine atoms upon initial photolysis
of CFCs 11 and 12, and most neglect any subsequent chemistry of
the fragments. Indeed, all the chlorine from stratospheric CFCs
11 and 12 is probably released eventually, and a time delay would
be relatively unimportant. However, the possibility remains that
a chlorine-containing intermediate would be sufficiently
long-lived to be transported to the lower stratosphere and
removed directly. This would reduce the effective chlorine
contribution of CFCs to the stratosphere, and thereby reduce any
ozone depletion which might be calculated to occur.
E-36
-------
Science
The methane oxidation cycle is expected to provide a
small source of odd oxygen through interactions with odd hydrogen
and odd nitrogen species. However, this cycle is not yet well
understood. Many of the methane oxidation products included in
current models have not been observed in the atmosphere, so veri-
fication of the current model chemistry is not possible. The rel-
evance to calculated ozone depletion lies in proper quantifica-
tion of the ozone source and in exploring possible interactions
with the chlorine cycle.
Likewise, sulfur chemistry is thought to be irrevelant
to the ozone depletion calculations, but again the knowledge is
minimal. Since much of the sulfur is rapidly converted to
aerosol, the question becomes one of heterogeneous chemistry.
Here again the neglect of this chemistry in current models is
based more on lack of information than on confidence in its
unimportance.
A final note on unknown chemistry concerns ozone
production. A variety of possible reaction schemes which lead to
production of ozone (or equivalently, oxygen atom) have been
proposed [Prasad and Burton, 1979; Miller ej: al. , 1980c] . None
have been verified, although experimental work is still in
progress. The schemes proposed are all dependent on chlorine,
and may provide explanations for stratospheric measurements such
as Anderson's observation of normal ozone in the presence of
abnormally large amounts of chlorine [Anderson et al., 1980a], or
Dobson and Umkehr measurements of increasing ozone levels over
the past decade despite increasing concentrations of
stratospheric chlorine (discussed in the section on ozone
measurements) . It is important either to verify or to rule out
such chemistry to reduce the very large uncertainty in "unknown"
chemistry. Of course the very nature of the word "unknown"
implies that the topics addressed here do not necessarily
encompass all such chemistry. Other surprises may exist as well.
E-37
-------
Science
E. STRUCTURE AND SCOPE OF MODELS
The previous sections on the troposphere, transport, and
chemistry describe the foundations of our knowledge of the atmo-
sphere in general and the chlorofluorocarbon effect in parti-
cular. To achieve some understanding of the CFC effect requires
synthesis of this knowledge. Computer models of the atmosphere
provide the tool. One must realize, however, that models are
only a tool — compilations of knowledge and not duplications of
the atmosphere itself.
Modelers attempt to verify their results by comparison
with measurements made at a point in time. However, the time
propagation of models is rarely tested (few comparable measure-
ments exist) , and the comparisons are not always successful
[Steed e_t al. , 1980; Angell and Korshover; 1973; 1978]. Indeed,
Prather ej: al. [1979] argue that the concept of the atmosphere
employed in current models may suffer from physical instabilities
such that only minor changes in input, e.g., total NO , may cause
X
gross changes in the chlorine chemistry calculated by the model.
This is especially true for high latitude conditions. Given the
very nonlinear system of equations to be solved, such a result is
not surprising.
•
Nonetheless, models are an exceptionally useful diagnos-
tic tool. They have helped researchers identify weaknesses and
gaps in knowledge and have guided subsequent research toward the
more crucial problems. The dynamical nature of models also
allows a consideration of the time evolution of the atmosphere
under a given set of conditions. To the extent that the "present
day" results of a model are correct and the dynamical equations
are correct, one can calculate the changes to be expected under a
totally specified set of future conditions.
E-38
-------
Science
However/ as discussed earlier, the limitations of the
current data base cause concern. Neither the transport nor the
chemistry in current models is likely to be complete or exact.
The time propagation of any errors is also likely to exaggerate
them, particularly if physical instabilities of the sort
mentioned by Prather e_t a_l. [1979] are present.
Even without the concern over data base and conceptual
uncertainties, true prognostication with models is difficult. As
mentioned, one must make assumptions about the "future" in order
to do a calculation. The simplest, most common, and probably
most misleading assumption is that all input fluxes of chemical
species and solar energy remain at their present (or other well
defined) levels. From the point of view of the research scien-
tist, this is a useful if not essential exercise. It allows the
models to achieve a steady-state, i.e., a condition where no
further change is noted from one model-year to the next. Given
numerical differences in modeling techniques, this steady-state
condition is most amenable to intercomparison of models. How-
ever, this is not by any means a simulation of the real future,
as an experienced modeler will readily admit. The model behavior
is that to be expected if the data base is complete and accurate,
jLE the conditions are accurately descriptive of the future, and
if the model structure does not impose unwarranted constraints.
The data base has been discussed, and the latter two conditions
are addressed in this section.
1. Stratospheric Chlorine Inputs
The sources of stratospheric chlorine are not all well
defined [NASA, 1979; Jesson, 1980] . Background chlorine levels
in models are based on estimates of known natural sources (e.g.,
CH-jCl) and estimates of industrial sources. In "prognostic"
calculations, estimated current sources of most compounds are
presumed to remain constant into the model "future". Sometimes
-------
Science
the model flux of CFCs-11 and 12 is allowed to change to
investigate the possible impacts of such changes. Realistically,
however, the future fluxes (both natural and man-made) are not
well known - but are unlikely to remain constant. An excellent
example is methyl chloroform, for which both industrial
production and atmospheric concentrations are known to be
increasing at the present time. Given nonlinear effects even in
chlorine chemistry itself [Cook e± al., 1980], the marginal
effects of continued CFC release are not accurately calculated
under such unrealistic constraints. Volcanoes may also provide a
time-varying source of atmospheric chlorine [Steed ej: a^. , 1980],
It may well be impossible to account for all such
variations in chlorine release, and that in itself does not
demean the results of model calculations. It does, however,
provide a perspective for viewing calculations extrapolated to
the "future".
2. Other Time-Varying Chemical Inputs
CFCs are not the only stratospheric perturbation to have
been considered by modelers. However, until recently, each such
perturbation has been considered individually with the assumption
that little coupling takes place among the various chemical
families. Molecules such as H02N02, ClON02, and HOC1 along with
several fast reactions such as H02 + NO, CIO + NO, and OH + HCl
have served to better define the coupling among HO , NO and Cl
a j\ x
species and their chemistry.
NO perturbations, possibly in the form of an increased
A
N20 source from fertilizer use, cannot reasonably be ignored in
consideration of a CFC perturbation. To the extent that an N20
source is increased, the calculated CFC effect on ozone is
decreased [Miller e_t al. , 1980c; NASA, 1979]. A recent set of
calculations [NASA, 1979] shows the CFC effect is roughly halved
E-40
-------
Science
if the ^0 source doubles. A "prediction" for the future is
improperly named if such changes are not taken into account. A
"what if" question which isolates a single perturbation is irre-
levant to the question of appropriate immediate action. One
reacts with justification to what one expects will happen taking
all variables into account. Focusing on a single variable and ig-
noring others is unreasonable.
Honest evaluation of the future has become increasingly
difficult as our knowledge has improved. A case in point is the
carbon dioxide perturbation. The primary effect of the well
documented increases in C02 [NASA, 1979; NAS , 1979a] is on the
radiative and thermal properties of the atmosphere. Chemical
effects, although potentially large, are secondary. Most
chemical models parameterize radiation and temperature fields
rather than calculating them explicitly, and the radiation models
generally use only simplified chemistry not capable of accounting
for a possible CFC effect. Recently, however, modelers have
begun to combine the two more effectively [Penner, 1980a; 1980b;
Haigh and Pyle, 1980; Callis and Natarajan, 1980]. The new
calculations show that a CC^-induced cooling of the stratosphere
may well cause increases in ozone concentration of up to 6 per-
cent. Taken in conjunction with a CFC perturbation, the net
effect is a reduction of CFC-induced depletion (by approximately
3-5 percent with current model chemistry) . [Penner, 1980b] .
A doubling of both N^O and COp/ the first of which is
not unexpected and the second of which is likely, reduces the
calculated CFC effect on ozone by well over half. Yet both are
customarily ignored in discussing the "dangers" of CFCs. Calcula-
tions of a postulated CFC scenario continue to be referred to as
"predictions of the future" despite the obvious fact that they
are neither predictions nor an accurate account of the future.
E-41
-------
Science
3. Coupled Nature of Atmospheric Processes
The atmosphere is a complex coupled system involving
inputs of chemicals and subsequent reactions, radiation inputs
and radiative processes, and transport. Radiation influences
chemistry both through photolysis and temperature. Dynamics is
influenced by radiation via heating and convective processes and
in turn serves to transport and distribute thermal energy.
Dynamics as well alters local chemical composition, which, in
turn, both drives chemical reactions and influences the flux of
radiation through the atmosphere. Finally the chemistry of the
atmosphere itself determines the chemical composition under the
influence of all the other processes.
In attempting to model the atmosphere with a 1-D model,
these processes are not all treated. Horizontal transport is
ignored and vertical transport is paramaterized and not
interactive with other changes. In particular, the coupling with
dynamics and radiation is omitted. The difficulty with such
approximations lies in the tacit neglect of feedback mechanisms.
The lack of calculational experiments has limited ability to
estimate the magnitude of these effects. They have been noted
and subjective guesses as to their magnitude have been made [NAS,
1979a; UK DOE, 1979]. For the moment, however, they remain not
well understood, and not at all rigorously quantified - even to
the extent of establishing confident upper limits.
Feedbacks are, of course, second order effects --
responses to a response. Thus, for very small perturbations,
they may be reasonably given proportionally less emphasis.
However, the sorts of chemical change postulated to arise from
CFCs are large enough to lead to more serious errors. The
severity is compounded by the basic lack of understanding of
feedbacks as much as by the computational difficulty of including
them.
E-42
-------
Science
4. Atmospheric Variability
As models are designed, calculated chemical species
concentrations are slowly varying functions of altitude and time
(and latitude in 2-D models). The realistic behavior, however,
is far removed from the model representations. Water has been
observed to exhibit significant stratification on relatively
small altitude scales [Kley £t al., 1979]. Total column
fluctuations occur for N0? (Noxon, 1979) and OH [Burnett and
Burnett, 1979]. I_n situ measurements of CIO also show marked
variability at all altitudes [Anderson e_t al. , 1980a; 1980b;
Parrish £t al, 1980]. Even source molecules such as ^0, CH^,
and CFCs are found to exhibit variability [Ehhalt, 1978; Goldan
ert al. , 1980]. These measurements are all examples of
variability on time and distance scales which models are unlikely
to be able to account for, since models necessarily calculate
averaged concentrations.
Beyond this sort of variability, there are the longi-
tudinal, latitudinal, and seasonal variations, all of which are
ignored by 1-D models. Such fluctuations and gradients in
species concentration can become significant in the reaction rate
calculations performed by models. To take the case of a simple
bimolecular reaction, the rate is conventionally expressed as:
k • A • B
where k is the temperature dependent rate constant, and A and B
are the local reactant concentrations. In a one-dimensional
model, what we want is the average rate of reaction, i.e., the
average of that product over time, latitude, longitude, and
altitude increment. We denote the average by:
V _ f '3
1J 1 -I
-------
Science
k • A • B
To get that result, one needs a value for the product at each
time and point in space -- unavailable information. The
information we can approximate, however, is the individual
averages of those quantities, i.e., k, A and B, so the models
employ the product:
k" • A • B
The two expressions of the product are not necessarily equal.
Suppose one considers the averages for two nearby locations. At
the first, k, = 1, A, = 2, B, = 0; at the second, k~ = 1, A2 =
0, and B = 2. Now:
k • A • B = 1/2 (kj_ • A! • Bj_ + k2 • A2 • B2)
= 1/2 (1 • 2 • 0 + 1 • 0 • 2)
= 1/2 (0 + 0)
= 0
This is true average reaction rate in that region. The model,
however, performs a slightly different calculation:
"k • "A • "i = 1/2 (kj_ +k2) • 1/2 (Aj_ + A2) • 1/2 (El + B2)
=1/2 (1+1) • 1/2 (2 + 0) • 1/2 (0 + 2)
= 1 • 1 • 1
= 1
This sort of averaging error can, and does, take place with each
of the kinds of averaging incorporated into model calculations.
That much is known. What is not known is the magnitude of such
errors.
Furthermore, that magnitude is very difficult to check
against reality. One is forced to compare model with model. For
example, a 2-D model can be averaged over latitude to give a
E-44
-------
Science
"1-D" result, which can then be compared to 1-D models. In fact,
Tuck [UK DOE, 1979], has compared 3-D model results with a 2-D
simulation for particular reactions and found that the errors may
indeed be very serious. It is, of course, possible that the
errors will tend to cancel each other, but for an extremely
nonlinear system that is an improbable outcome.
Spatial variability also presents difficulties for
interpreting model results and comparing them with field measure-
ment data. Any atmospheric field quantity, X, may be expressed
as the sum of an average over longitude, X, and the deviation
from that average, X1, i.e., X = X + X1. Now a measurement
provides the quantity X, whereas a model calculates X. For a
valid comparison, one must either know X1 or have sufficient
measurements to establish a reliable X from field measurements of
X.
Carrying the situation to 1-D models, the situation is
further complicated by latitude. The longitudinal average X may
be further broken down into a global average X and a deviation of
the longtudinal average from the global average X" so that
"x = X~ + X"
and X = * + X1 + X"
In analogy to the 2-D situation, a comparison of 1-D results for
1( requires either knowledge of both X1 and X" or sufficient data
to provide a reliable average field measurement of X. In
general, field data are simply not sufficient to define either
the fluctuation terms or the necessary averages, and the desired
comparison is not possible.
-------
Science
An appreciation of these problems is essential to any
comparison of model results with field measurements. Thus,
statements that current agreement is reasonably good [NAS, 1979a]
are implicitly qualified by the knowledge that only gross errors
in magnitude or errors in shape are likely to be detectable. The
calculated averages may well be quite different from either
individual measurements or the real averages occurring in the
stratosphere.
E-46
-------
Science
F. MODEL RESULTS
Despite the extensive uncertainty associated with models
and the data upon which they are based, models are indeed useful
and their results are important. The dangers lie in the inter-
pretation of the results without appropriate caution. The
principle applications lie first, in using comparisons of
calculated and measured results to give general indications of
model inadequacy or adequacy, and second, in providing qualita-
tive, if not quantitative, information on impacts of changes in
either model input (e.g., release scenarios) or data base (e.g.,
rate constants). In either case, appropriate attention to uncer-
tainty is a necessity.
1. Comparison of Calculated and Measured Concentrations
Numerous reviews comparing theoretical and experimental
profiles of chemical species mixing ratios (or concentrations)
have been published. NASA [1979], NAS [1979a], and the UK DOE
[1979] have all addressed this subject, and the citations
contained in those documents provide an even more complete body
of information. Miller e_t al. [1980c] reviewed the situation for
one-dimensional model results, and later [Miller jit: al. 1980b]
for 2-D results based on the Du Pont models. These models agree
well with others [see NASA, 1979] and, taken together, have the
further advantages of containing identical treatments of the
chemical scheme, and hence providing some information on the
effects of including the latitudinal dimension. The following
discussion summarizes the current situation.
a). OH, H02, H202. For the odd hydrogen species, the
available field measurement data are very limited. For both 1-D
and 2-D models, where information is present, the comparisons
between model and measurement are quite good. The major
difficulty is the lack of experimental profile measurements below
E-47
-------
Science
30 km, the lower limit of measurements by Anderson [1976], In
the section on chemistry, a reference was made to the implied
discrepancy in the lower stratosphere. OH concentration strongly
influences the ratios N02/HN03 [Harries, 1978; Evans jjt_ al.,
1976; Loewenstein ej: al., 1978] and HC1/C10 [see Miller ^t al. ,
1980a; 1980b; 1980c], both of which appear to be underestimated
in model results. Such a comparison is difficult, however, due
to the different techniques and concomitant different absolute
accuracies of measurements for each separate species. The ratios
are often derived from data taken at different times, and even
different locations. Decreased concentrations of OH at low
altitudes in model calculations would nevertheless improve the
comparison by increasing the model ratios. Recent changes lead
to some, but relatively insignificant improvement.
b) . Total NO . Calculated total odd nitrogen in the
A,
upper stratosphere (approximately 35-45 km) is near 25 ppb with
NASA RP 1049 [1979] model chemistry. Expected increases in the
recommended photolysis rate for NO and other recent changes in
chemistry [Chang, 1980], may improve this situation to some
extent, but do not remove the problem. (Before the recommended
NO photolysis rate was slowed in 1979, the model-calculated total
odd nitrogen in this region was 15-18 ppb [Miller e_t aj^. , 1980c] ,
whereas measurements indicated less than 15 ppb [e.g., Horvath
and Mason, 1978]. The difference occurs in both 1-D and 2-D
models.
c) . HNO-j. Nitric acid remains greatly overestimated by
models in the upper stratosphere [NASA, 1979]. The problem
persists in 2-D calculations [Miller et al. , 1980a] , and in
calculations with the most recent chemistry.
d). N02. The underestimation of N02 below 30 km noted
by NASA [1979] is still characteristic of model results, and like
other NOX discrepancies is unimproved in 2-D models. Calculated
E-48
-------
Science
profiles between 20 and 30 km lie below measured values at all
latitudes, but particularly the higher latitudes. This deviation
is consistent with calculated total column NCU which does not
show the variation with latitude characteristic of measured
values. Like other discrepancies, it is not removed with recent
chemistry.
e). HO^NOp. Employing reduced photolysis cross
sections of Molina [1980a; 1980b] in recent model calculations
has led to an additional discrepancy. Whereas Murcray e_t al.
[1973] have given an upper limit of 0.4 ppb at 25 + 4 km, 64
degrees N latitude, model results for that location range from
1.5 to 2.0 ppb. Although the Littlejohn and Johnston [1980] rate
for the reaction OH + H02N02 reduces the calculations to
approximately 0.5 ppb, Murcray [1980] claims recent measurements
reduce his upper limit to approximately 0.1 ppb. Thus, other
HC^NC^ reactions may well be important.
f) . 0-j. While 1-D model results indicated a possible
underestimate of ozone at high altitudes (50 km) , the represen-
tation appears to be improved in the current 2-D simulation. New
chemistry recommendations lead to small overestimates in ozone.
g). C10N02, Cl. Information on these critical chlorine
species is limited, but agreement appears to be reasonable where
comparison can be made.
h) . HC1. Although the calculated profiles for HC1 in
both 1-D and 2-D models lie near the measured profiles throughout
the stratosphere, the qualitative comparison reveals a signif-
icant discrepancy. (A detailed explanation is given by Miller
et al., [1980a]). Whereas the measured profiles generally
display a monotonic increase of mixing ratios with increasing
altitude, the calculated profile shows a rather steep increase in
the lower stratosphere with relatively constant mixing ratio
E-49
-------
Science
above approximately 25 km. The result is an underestimate by
models at higher altitudes, reasonable agreement in the middle
stratosphere, and an overestimate below 25 km. The improvement
anticipated with the development of 2-D modeling [MAS, 1979a] has
not materialized, and the discrepancies seem to be indicative of
problems with model chlorine chemistry, rather than a phenomenon
of meridional transport [Miller et al., 1980a]. The most recent
reaction rate revisions significantly increase the disagreement.
i. CIO. Qualitatively, the comparison of calcu-
lated and measured CIO profiles is similar to that for HCl. For
CIO, however, the best agreement seems to come at slightly higher
altitude. Near 40 km, the calculated profile lies near the lower
end of the so-called "normal" range of CIO profiles measured by
Anderson e_t al. , [1980a] . At lower altitudes, however, the
calculated profile exceeds measurement. More significantly, the
decrease in mixing ratio with decreasing latitude below 30 km is
not nearly so steep in the calculated profiles as that observed
in the measurements. Thus, at 25 km, the calculated profiles in
the NAS report [1979a] exceeded the measurements by over 100
percent. Furthermore, those models also calculated the majority
of ozone depletion to take place between 20 and 30 km, and that
depletion is directly related to these overestimated CIO concen-
trations. Just as with HCl, NAS [1979a] expressed expectation
that improvement would be forthcoming with inclusion of latitu-
dinal transport in 2-D models; and again, the 2-D treatment at 30
degrees N latitude using equivalent chemistry is very similar to
the 1-D result [Miller et al., 19SOa]. The most recent model
chemistry significantly decreases both this discrepancy and
calculated ozone depletion. Complete removal of the discrepancy
will lead to still lower calculated depletion.
A problem which has as yet found no model solution
is that of the very high CIO measured on_p,ne_ occasion by
Anderson et al., [1980a]. While this represents only a single
E-50
-------
Science
balloon flight, the unique results remain significant. The
possible explanations have been discussed often, e.g., Anderson
et al. , [1980a; 1980b] and Miller e_t al. , [1980c] . However, the
crucial point is the simultaneous measurement of ozone made on
the same flight. Ozone levels were found to be normal in spite
of the elevated chlorine monoxide. The measurements were made up
to 40 km where one ordinarily assumes the chemistry to be simpler
and relatively better understood than that in the lower
stratosphere. Yet, the facts don't fit the theory. Here, as in
other cases of real world- measurements which conflict with the
theory, it has seemed more appropriate to some to discard the
evidence rather than to question the theory.
2. Comparison of 2-D and 1-D Model Results
In the preceding section, several instances were noted
where calculated two-dimensional model vertical mixing ratio pro-
files for 30 degrees N latitude are in substantial agreement with
1-D results. As mentioned, a more complete comparison is made by
Miller e_t a^. , [1980a, 1980b] . Complete evaluation of the newest
calculations is still in progress. While this conclusion may
provide some support for the contention that 1-D models may be
most representative at that latitudinal zone, it also dispels the
notion that the major discrepancies in C1X species are minor
transport effects with little implication for calculated ozone
depletion. On the contrary, those discrepancies will merit
greater attention with regard to determining their true origin.
The revisions of several reactions, particularly those of OH
radical, are merely the beginnings of that work, and will not
solve the problem completely.
Two-dimensional models are also capable of providing
more complete information regarding calculated ozone depletion.
While it is not obvious for 1-D results whether the calculated
depletion represents a global average or the results for a parti-
E-51
-------
Science
cular latitude, the 2-D model provides a latitudinal distribution
of depletion. Pyle and Derwent [1980] £ind that the depletion
calculated by their model occurs primarily at high latitudes in
winter. This coincides with the measured maximum in the annual
ozone variation [NASA, 1979] . The minimum in solar ultraviolet
flux, which occurs simultaneously, implies that the link between
global average ozone depletion and global average increase in UV
is not likely to be one of amplification to the extent expected
at the time of the 1979 NAS reports. Pyle and Derwent find a 1
to 1 relationship between the percentage changes, half of that
estimated in 1979. Other models will soon be capable of pro-
viding confirmation for this result, and of assessing the effect
of new recommendations in chemistry, which may decrease the
latitudinal variation somewhat.
The additional information in the latitudinal and
temporal variations in the theoretical ozone depletion will also
allow for more sophisticated analyses of ozone measurements, with
the aim of identifying any trends in ozone concentrations which
may be related to model calculations. In summary, 2-D models, as
they reach greater levels of sophistication and validation, are
likely to provide considerably more complete information than 1-D
models. Such information allows better understanding of the
theorized depletion and its implications for biological systems.
3. Current Model Ozone Depletions Calculations
Perceptions of potential ozone depletion according to
theoretical calculations have changed significantly over the past
year among scientists who have followed the recent developments
in chemistry. The of ten-referred-to figures quoted by NAS
[1979a] were calculated with the Lawrence Livermore Laboratory
(LLL) model. A recent paper [Wine et al. , 1980] notes the
changes in those results that came with adoption of the chemistry
recommended in 1979 by NASA, along with other minor revisions
E-52
-------
Science
(such as the change in OH + H2°2^ * Rather than 18.6% calculated
steady state depletion as the base case, the model now calculates
13.9% ozone depletion for continued release of CFCs-11 and 12 at
1977 release rates. The NAS figure was adjusted from 18.6% to
16.5% based on some expectation of tropospheric sinks and nega-
tive feedback mechanisms, and a proportional adjustment might be
applied to the 13.9% figure with the same assumptions (yielding
approximately 12.3%).
Several newer reaction rate measurements may reduce this
figure much further. Wine et al. , [1980] have reported a rate
for OH + HNOo which, when included in the LLL model, reduces
calculated depletion to 9.5%, if the reaction products are those
currently assumed (H20 + N03). (NASA recommends use of this rate
in its latest evaluation [Chang, 1980]). Despite this large
revision due to the change in this reaction rate, it was not
identified in the NAS list of reactions to which 03 is most
sensitive. Should the reaction produce an alternative, and less
likely, set of products (H200 + N02^' calculated depletion would
be reduced from the 13.9% figure to near 12%.
Further, a change from the estimated rate for OH +
H02N02 to the measured value which is now available [Littlejohn
and Johnston, 1980] also reduces the 13.9% figure to below 12%.
Also, without including changes in the other OH reactions, a move
to the fast rate for OH + HO- would reduce calculated depletion
from 13.9% to below 11%. Several other reaction rates, including
H02N02, are revised in the latest NASA recommendation [Chang,
1980]. As a group they also tend to lower calculated ozone
depletion.
In combination, these latest revisions shift the calcu-
lated result to approximately 7% or even less, a dramatic change
from the 18.6% reported only a year ago. (Using a similar set of
r.eaction rates, Penner [1980b] reports 5.5%, but notes that
E-53
-------
Science
necessary changes in transport rates will increase this figure
somewhat). (A change in OH + HO^ would further reduce calculated
depletion). In retrospect, the concerns generated by the NAS
report [1979a] and the increase in calculated depletion since the
previous report [NASf 1976] were somewhat excessive. Only one
year later the model results are near those of the first report
[NAS, 1976] and are likely to move to well below that level. At
the same time, the confidence in models themselves has been con-
siderably shaken, both by 2-D results/ and by the increased
awareness of critical and unjustified assumptions in the models.
The preliminary reports of the ALE program provide
initial evidence contrary to the assumption in all depletion
calculations that CFCs are not destroyed in the troposphere. If,
for instance, CFCs-11 and 12 are both found to have atmospheric
lifetimes, equal to half the stratospheric photolysis lifetime,
calculated depletion would likewise be reduced by half, i.e.,
from the most recent values of 6-9% down to 3-4.5%.
The observed trend in C02 concentrations, should it
continue until CC>2 has doubled as expected, would imply a growth
of ozone estimated at 4 to 6% [Penner, 1980b]. This additional
factor (perhaps slightly reduced to 3-5% by interaction with a
chlorine perturbation) may be applied to any of the calculated
figures mentioned previously. Any trend in ^0 is calculated to
further reduce the potential impacts of CFCs.
To claim the current evidence proves CFCs to have a
negligible effect on the ozone is an exaggeration, but the evi-
dence indicates that such an outcome is highly possible. And
very definitely, the evidence shows that the large effects
"predicted" so confidently and so recently are now much less
likely.
E-54
-------
Science
4. The Effects of Postponed Regulations
A proper regulatory assessment of a potential, but
ill-defined and unverified, threat to the environment requires
attention to the time dependence of the potential effects.
Specifically, the reliance on calculated steady-state ozone
depletion numbers can distort the perception of how fast the
potential problem is developing and how urgent the need for
action. It is illustrative to consider how that potential
predicted effect would be changed by a regulatory decision made
now as compared to one made at some future time. That point was
made in an earlier Du Pont submission to EPA [Du Pont, 1980b]
regarding the NAS CISC report [1979b]. Assuming an extreme
regulatory response of a total U.S. ban on CFCs-11 and 12,
combined with constant continued release by the rest of the
world, Du Pont 1-D model results showed that deferral of such
action from the end of 1980 to the end of 1985 would have a
calculated maximum incremental effect of 0.2 percent on the time
dependent ozone depletion curve. That is, the depletion at any
time in the future is calculated to increase by less than 0.2
percent as a result of deferring the regulation for five years.
That maximum difference would occur while depletion is calculated
to be still relatively small; ultimate depletion would be
Identical at steady-state. This would be true of any reduction
in emissions by a comparable amount (approximately 30%). It is
important to realize too that, those calculations were made with a
model chemistry which calculated 18.3% steady-state depletion for
continued release at 1975 rates, well above current LLL results.
These calculations represent an overestimate based on best
current knowledge.
The significance of those results lies not in arguments
for arbitrary deferral, but in support of a cautious regulatory
position in the face of severe uncertainties which can be
resolved by the ongoing academic, government, and industry
E-55
-------
Science
research programs. Regular evaluations can be made to reassess
the situation.
It may, therefore, be concluded that premature regula-
tion has virtually no tangible or calculated benefits to the
environment, yet it may entail serious costs addressed elsewhere
in this document. At the same time, the likelihood of a more
soundly based decision increase with each research result as more
aspects of the atmospheric system are better defined. Several of
the numerous results to be expected in the next few years have
been mentioned above and will be more completely considered in
section H on uncertainties.
E-56
-------
Science
G. OZONE AND ULTRAVIOLET RADIATION
A persistent problem with understanding theorized ozone
depletion lies in determining its significance. Qualitatively,
it is understood that less ozone will result in more ultraviolet
radiation reaching the surface of the earth. Increased ultra-
violet may be related to certain biological effects. However,
neither of the relationships is well-defined. Moreover, worst-
case scenarios differ tremendously from other possibilities.
Lack of data frustrates confident differentiation among the
possibilities. Below we highlight key variables; detailed
discussion appears in Appendix F.
1. Latitudinal Variation in Calculated Depletion
Potential ultraviolet changes have generally been
calculated in terms of a global average depletion calculated with
1-D models, and a relative amplification factor based on solar
spectrum and global average ozone concentration. Two-dimensional
models have now demonstrated a correlation between calculated
depletion and natural seasonal and latitudinal variations in
ozone concentration [Pyle and Derwent, 1980]. The correlation is
such that a calculated global average change in ultraviolet flux
is considerably less than that calculated in the one-dimensional
approximation, even when the global average ozone depletion from
the 2-D model is equal to that in the 1-D model. Accurate
quantitative evaluation of this correlation is a priority task
for the newly developed 2-D models.
To assess the importance of any change in ozone
concentration, one must understand just what the change implies,
i.e., how does ultraviolet vary in each location. Then one must
further understand the importance of such changes in ultraviolet.
Just how large are the changes compared to actual latitudinal and
temporal variations?
E-57
-------
Science
2. Natural Ultraviolet Radiation
Natural ultraviolet fluxes to the earth's surface have
been both measured and calculated. Field measurements are made
difficult both by instrument sensitivity and by meteorological
considerations [See Appendix F-2]. In order to determine either
average or representative daily fluxes, one must consider normal
patterns of cloud cover and associated meteorology as a function
of season and location.
Calculations based on solar UV flux and ozone concentra-
tions are most often based on clear sky conditions, neglecting
the attenuating effect of clouds. Occasionally, however, a cloud
cover correction is applied and comparison of measured and
calculated data is possible. Such comparisons reveal discre-
pancies of up to 25 percent. A more thorough discussion of these
issues is given by Klein [See Appendix F-2] and much of the
following discussion relies on that analysis.
Despite the difficulties and associated uncertainties, a
picture of the UV radiation flux to the earth is beginning to
emerge. An almost startling discovery is the range of variation
in UV flux with both latitude and season. Measurements discussed
by Klein have been made in Panama (9 degrees N latitude) and in
Rockville, Maryland (39 degrees N latitude). UV input was mea-
sured in 5 nanometer intervals. The measured radiation in Panama
was found to be more than twice that occurring in Rockville for
representative days of each season. Using the DNA action
spectrum of Setlow (1S74) , the UV flux was weighted and summed
according to the NAS [1979b] definition of "damaging" ultraviolet
(DUV) . The inferred DUV dose at Panama was found to be four to
five times as large as that received in Maryland, a dramatic
change over only 30 degrees of latitude. As Klein notes: "This
certainly would imply that one should be cautious about predicted
effects of 20-45 percent increases in UVB in temperature
E-58
-------
Science
latitudes." Even such seemingly significant changes are
literally small compared to natural latitudinal variations.
Furthermore, as noted elsewhere in this document, such changes
are based on calculated depletion levels which are beyond current
estimates and which represent a steady-state situation which is
neither realistic nor likely to occur.
The natural latitudinal variations are not cited to
imply that a geographic redistribution of population would be an
appropriate or an adequate response to ozone depletion. They do,
however, give some perspective on changes in UV levels. Also, a
better knowledge of such variations, both in UV and DUV (or other
appropriate weighting), allows realistic interpretation of
effects research in biological systems - both epidemiological and
experimental studies. An accurate picture of normal latitudinal
and seasonal variations must also be used with time and latitude
dependent calculations of ozone change in order to determine
realistic effects of diminished ozone levels. The several
correlations and nonlinear variations of calculated depletion,
ozone itself, and incident solar UV all introduce errors in
estimated overall effects when each step in the sequence is
treated using global annual averages. As a simple obvious
example, large depletion at the poles during polar night has
virturally no effect on incident UV at the pole, whereas a much
smaller ozone change at the equator in summer may cause a rather
large absolute change in local UV flux.
3. Action Spectra
One might very reasonably assume that different wave-
lengths of ultraviolet light will interact differently with
biological systems. Therefore, in studying the effects of
changes in ultraviolet it is appropriate to consider a UV flux
weighted by some action spectrum. It is also reasonable to
believe that different biological effects occur by different
E-59
-------
Science
mechanisms and, therefore, would exhibit different action
spectra. Therein lies the major problem in interpreting the
effects of even well-defined changes in incident UV. For most of
the postulated potential effects, no action spectrum has been
determined [NAS, 1979b]. For others, arguments have been made
that a given measured action spectrum is appropriate, but no
actual relationship has been demonstrated.
In the face of this lack of certainty, the term
"damaging ultraviolet" has been used by NAS [1979b] to describe
UV changes weighted by the DNA action spectrum given by Setlow
[1974]. At the same time, Appendix D of the NAS report [1979b]
makes clear the very large differences in percentage change of
weighted UV which may be derived using different action spectra
but identical ozone depletion. Thus, the magnitude of an effect
depends not only on the magnitude of ozone depletion and the
change from normal UV levels but also on the appropriate action
spectrum. A review of the NAS appendix makes clear that the
DNA-damage action spectrum is a worst-case among the various
action spectra considered thus far. The DNA-damage action
spectrum has not been directly related to any of the effects
discussed by NAS, but only chosen as a plausible relationship.
The arbitrariness of the "folklore" value of 2 for a biological
amplification factor (relative change in biologically effective
ultraviolet for a given change in ultraviolet itself) is made
quite clear by NAS.
As summarized by the authors of the NAS CISC report:
"The complex relationships between RAF [relative
amplification factor] and latitude, season, percen-
tage ozone-layer reduction, and weighting function
create considerable uncertainty in any simple statement
about the DUV dose change expected for any particular
ozone depletion. This uncertainty is compounded with
E-60
-------
Science
the uncertainty in the relationship of the biological
response to the DUV dose in assessing the ultimate
biospheric impact." (emphasis added). [NAS, 1979b,
p. 311].
Yet such a "simple statement" is included as Key Finding No. 7 of
the same report! Reiteration of the Key Finding No. 7 statement
is now cited as partial justification for regulatory action, and
the numerous strong caveats which should be applied are virtually
ignored.
4. Dose and Dose-Rate
A final potentially important factor in understanding
changes in ultraviolet flux is an appreciation of natural
temporal variations in ultraviolet on a daily basis. UV flux, of
course, follows the diurnal variations of sunlight in general.
As such, the maximum occurs in the mid-day hours with decreased
levels toward morning and evening. As a consequence, it may be
appropriate to'consider not total daily dose or annual dose of
UV, but changes in peak dose, peak dose rate, or average dose
rate.
It has been argued (Damkaer, See Appendix F-4) that dose
rate may be equally or more important than total dose in
determining biological effects. Again, this suggestion should be
acknowledged and accounted for in any attempt to quantify the
effects of calculated ozone depletion.
E-61
-------
Science
H. UNCERTAINTIES
An uncertainty is almost always more easily defined than
quantified. Despite their mention of most of the significant
uncertainties, the NAS Panel on Stratospheric Chemistry and
Transport has been severely criticized over its attempts to
quantify those uncertainties [Du Pont, 1980a; CMA, 1980]. The
criticism addressed both the propriety of certain quantifications
and the estimates themselves.
Faced with an uncertainty for which either theory or
experiment provides well defined limitations, it is reasonable to
report a range of uncertainty. For other cases, however, where
uncertainty centers around a plausible but otherwise untested
assumption, only subjective considerations will allow the
assignment of a numerical value to uncertainty. Such estimates
are often useful to an investigator determining the future course
of research, but only with the full realization that they do not
constitute scientific evidence. Use of subjective uncertainty
ranges easily can be misleading — a decision which seems to be
justified by the evidence may in fact not be so.
The fear of inadvertent misuse of the uncertainty esti-
mates prompted a portion of our earlier criticism and now appears
to be well-founded. Despite statements by the NAS panel that the
question of unknown chemistry broadened the 95 percent confidence
limits on calculated steady-state depletion to 3 to 30 percent,
the principal findings' statement of 5 to 23 percent is the one
which is cited. Not only is that estimate given undue credence
(it already includes some "unquantifiable" uncertainties), but an
acknowledged (if underestimated) source of uncertainty is
summarily ignored.
In short, a regulator must be aware of the distinction
between what has been scientifically demonstrated - via the
E-62 .
-------
Science
scientific method - and what are subjective or intuitive
conclusions of the scientists themselves. This applies both to
atmospheric science in general and to evaluations of uncertainty.
Although some caveats were included in the NAS report [1979b],
the distinctions were apparently not sufficiently clear, and the
uncertainty estimates were in general not well supported.
1. Current Uncertainties
A brief review of general areas of uncertainty will
demonstrate the significance of several major concerns which may
be hidden by the assignment of.subjective estimates.
a). Known Chemistry. Previous estimates of the
uncertainty in calculated ozone depletion due to uncertainties in
known chemistry have generally relied on model sensitivity
studies, coupled with confidence ranges for individual reactions.
Despite evidence by Smith [1973] that such methods have in the
past led to underestimation, the caution was unheeded. Again,
remeasurements of "known" reactions (OH + HgOp an<3 OH + HNCU)
have found rates outside previous uncertainty ranges, with
consequent larger effects on calculated ozone depletion than
anticipated. Not only uncertainty, but also the sensitivity of
calculated ozone depletion to given reactions have been under-
estimated. Although the calculated depletion still falls within
the range suggested for errors in known chemistry, one must
remember that only a few remeasurements have been made. The fact
remains that the premises (individual reaction rate uncertainties
and sensitivities) supporting the overall estimated range have
not held up under scrutiny.
A further example is found in the case of peroxynitric
acid (PNA). Based on the previously "known" ultraviolet
absorption cross sections, PNA and its "unknown" chemistry was
omitted from both model calculations and uncertainty analyses
E-63
-------
Science
used by MAS. The molecule was deemed to have little potential
influence on results. However, a remeasurement of its photolysis
cross sections [Molina and Molina, 1980a; 1980b] has implied a
significantly longer photolysis lifetime in the stratosphere, and
has made other reactions more important. In particular,
calculated depletion varies by over a factor of two depending on
the chosen rate for the reaction OH + H02N02, which has only
recently been measured to be 5 times or more faster than the
estimated value modelers have had to rely on.
b). Transport. The uncertainties due to transport have
been evaluated on the basis of changes in calculated A 0-, only
with respect to changes in the eddy diffusion parameterization.
The changes in parameters were restricted by assuming that
changes in calculated CFC lifetimes dominated other errors.
These assumptions have been challenged, and indeed it may be the
applicability of the eddy approximations themselves to
short-lived species which introduces the major uncertainty. In
any case, such factors have been ignored in the NAS estimates.
c). Tropospheric Sinks. Uncertainties in the chlorine
contribution to the stratosphere from CFC release have been eva-
luated by attempting to quantify, on a case-by-case basis, every
previously postulated mechanism for a tropospheric sink. An
estimate is then made of an arbitrary and unspecified "likely"
sink and possible limits to its effectiveness. This approach is
valid only if all possible sinks have been considered; complete-
ness has not been demonstrated, and without it the uncertainty is
unquantifiable without information directly from atmospheric CFC
concentration changes.
d). 3-D Chemistry and Feedbacks. The problem with this
potential source of error is, of course, that it is unquantifi-
able. While the possible sources of error have generally been
identified, quantitative tests of many assumptions regarding 1-D
E-64
-------
Science
model construction and omitted feedback mechanisms are simply not
yet available. Although the NAS expects that some reduction of
calculated depletion will take place (the "correction factor"
applied to calculated depletion), the chosen range of possible
effects is necessarily arbitrary. At the same time, the spatial
and temporal averaging assumptions in 1-D models are fully
capable of leading to severe errors in chemical treatment.
e). Systematic Errors and Omitted Chemistry. The
question of systematic errors and omissions in chemistry was
discussed briefly by NAS [1979a] , and an "educated guess" of the
associated uncertainty was given, but not included in a table
purporting to be a summary of contributions to predicted ozone
change and error estimates. Perhaps such "guesses" were omitted
because of concern over their arbitrariness, but other equally
subjective estimates were included in the table. This contrib-
ution to uncertainty is actually a critical one. As an example,
possibilities for chlorine-catalytzed ozone production have been
suggested in the literature, but not disproven scientifically.
No matter how doubtful they may be in someone's educated opinion,
the possibilities still exist. Unknown chemistry involving
H02N02 (now demonstrated to be important by an unexpected
revision in "known" chemistry) accounts for the entire factor of
2 range given by NAS for this error contribution. Other
contributions, appropriately included, would broaden this range
in a realistic assessment. It may be argued that the overall
range for possible calculated steady-state ozone change must
include the potential for positive changes in ozone levels. In
any case, the unknown is a major source of potential error in
ozone depletion calculations and should not be considered as
subordinate to, or less real than, other sources of error.
f). Ozone Trend Analysis. In its chapter on ozone
change projections (a far better choice of words than
predictions), the NAS Panel notes that all change estimates are
E-65
-------
Science
made independently of ozone trend estimates. Also mentioned is
the possibility "to improve" our estimates by making use of the
information given by the observed ozone change." (emphasis
added). [NAS, 1979a, p. 194].
It is possible to repeat the analysis using current
results for both ozone trend and model calculations. Further-
more, the improved ability to analyze ozone measurements makes
the exercise much more meaningful now than in its application by
NAS. The exercise makes use of all information, rather than
relying solely on the theoretical computer model results and
neglecting the measured real-world data.
In this exercise the variance and central value of trend
estimates are combined with those of model calculations covering
the same time period to give a "best estimate" of the 1973 ozone
change based on both trend analysis and the model prediction" of
-1.3% [NAS, 1979a, p. 195] (emphasis added). Since several
2
contributions to the ozone trend variance estimate, [(1.75%) ]
have now been determined statistically, (measured variance
2
approximately (0.7) ), it is appropriate to redo that analysis.
The best trend estimate now is +0.3% (through 1978). Model
calculations for depletion through 1973 recently hovered near
-1.5% (corresponding to approximately -13% depletion at steady-
state) , and the variance assumed can be taken from the NAS error
analysis. (The latest calculation would give a still smaller
present day figure). One standard deviation = 0.35 x calculated
depletion = 0.35 x 1.5% = 0.53%. The variance is, therefore,
2
(0.53%) . The derived combined "best estimate" for ozone change
through 1978 is, therefore, now about -0.84% or lower according
to this "best estimate".
In an extension of that analysis, according to NAS
[1979a], all other estimates can be multiplied by the propor-
tionality factor, in this case 0.84/1.5 = 0.56, to account for
E-66
-------
Science
the current information given by trend analysis. Applied to the
steady-state calculated depletion of 13% used here (revised LLL
model plus minimal correction for OH + HNCU), the "best estimate"
for steady-state depletion by this method is 7.3%. This analysis
still ignores many other recent developments which further reduce
the model estimate, and, therefore, the "best estimate," but is
given to demonstrate the information contained in trend analysis
estimates - information that should not be overlooked.
The NAS Panel chose to call such a correction
speculative because of the uncertainty in their own estimates of
variance. The comparable variance has now, however, been
determined from the ozone data, and indeed, the present day
variance in ozone data is found to be similar to that in model
calculations. This is especially true when one realizes that NAS
has likely underestimated the model variance by a significant
amount.
2. Reduction of Uncertainty
In pondering the necessity of a regulatory decision, one
must pay attention not only to current uncertainties but to the
time scale on which they may be diminished. Attendees at a
Stanford Research Institute Workshop considered such questions on
a very broad scale [SRI, 1980]. It is also, of course, relevant
to consider research in a more detailed sense, i.e., to examine
likely near-term research results. For major developments
affecting major uncertainties, such analysis is a necessity.
a). An experiment intended to determine experimentally
the quantitative relationships between catalytic species (CIO,
OH, H02 and N02) and odd oxygen (0 and O-.) in the stratosphere is
currently being developed by J. G. Anderson. All species will be
measured simultaneously as a function of altitude in the upper
E-67
-------
Science
stratosphere (approximately 28 to 38 km). The upper stratosphere
has been chosen because relative species concentrations are
determined primarily by chemistry, due to the long transport
lifetimes in that region. Measurements will be made with an
instrument package that is raised and lowered repeatedly
(approximately 10 times) through the relevant altitude range.
Variability of concentrations over the time and spatial scale of
the measurements will be sufficient to establish the sought-after
relationships, independent of any model results. A comparison
with the relationships derived from models for the same altitude
range will help to define the accuracy of the models in that
region. By accounting in some way for the role of transport, it
is possible to extend the measurements into the middle and lower
stratosphere. In theory, this experiment is capable of providing
definitive information regarding the stratospheric chemical
aspects of ozone depletion. The reel-up-reel-down engineering
will be tested in a flight planned for the spring of 1981, and
a flight of the full instrument package flown is planned later
that year. Analyses and more flights should follow shortly
thereafter.
b) . The Atmospheric Lifetime Experiment enters its
third year in 1981. Analysis of its results will address the
question of tropospheric sinks for CFCs, a second major aspect of
the theory. A sink important enough to have a major impact on
calculated depletion would be evident in the complete three-year
data. Thus, an answer should also be available in 1982.
c) . The rapid further development of two-dimensional
models in the near future promises to be a third major event of
the next few years. The ability to test directly certain
currently unquantifiable uncertainties of modeling answers a
critical need. One-dimensional models are being increased in
E-68
-------
Science
sophistication, with radiative-convective models being coupled to
photochemical models for investigation of feedbacks and combined
perturbations.
d). Research in kinetics and photochemistry is expected
to continue at its customary rapid pace. However, even in this
area, the near future is an especially important time given the
several high-sensitivity, critical reactions currently under
study, e.g., OH + H02N02, OH + CIO, OH + H02 and others.
e). In the area of field measurements, concentration
profiles for a number of species will no doubt be better defined
by ongoing programs of experiments. Of special interest is the
increased development of ground based techniques. [Parrish et
al., 1980] have utilized such a technique to measure CIO radical
column densities (finding results in near agreement with
Anderson's "normal" range for CIO), and are also capable of
measuring 0-,. Vertical profiles can be derived from the data.
The ability to take measurements on almost a daily basis is a
great advantage of this method over infrequent balloon flight
measurements, and leads to much more rapid building of a data
base suitable for model validation or invalidation.
f) . The improvements to be expected in ozone trend
analysis have already been quite fully described. The principal
efforts of future research will be to devise new ways of removing
explainable variations from the ozone record to better expose the
existence or nonexistence of a trend. Already the method pro-
vides the desired margin of safety to permit a postponement in
regulations until a better decision can be made.
E-69
-------
Effects—Skin Cancer
Potential Effects on Humans of Alteration of the
Stratospheric Ozone Layer.
A Critical Review of the Report of the Committee on
Impacts of Stratospheric Change of the National Academy of Sciences
Entitled: Protection Against Depletion of
Stratospheric Ozone by Chlorofluorocarbons,
Washington, B.C., 1979
by
Frederick Urbach, M.D,,, F.A.C.P.
Director, The Center for Photobiology,
Skin and Cancer Hospital of Philadelphia
Professor and Chairman, Department of Dermatology
Temple University School of Medicine
The Skin and Cancer Hospital
3322 North Broad Street
Philadelphia, PA 19140
(215)221-3924
(APPENDIX F-l)
F-l-1
-------
Transraittal letter for
Appendix F-l
TEMPLE UNIVERSITY
SCHOOL OF MEDICINE
PHILADELPHIA, PENNSYLVANIA 19140
THE CENTER FOR PHOTOBIOLOGY
FREDERICK URBACH, M.D., F.A.C.P.
PROFESSOR AND DIRECTOR
215-221-3924
P. DONALD FORBES, Ph.D.
ASSOCIATE DIRECTOR
215-221-3920
DANIEL BERGER, M.S.E.E.
215-221-3937
HAROLD F. BLUM. Ph.D.
215-221-3924
RONALD E. DA VIES, Ph.D.
215-221-3960
MAHENDRA K. LOGANI, Ph.D.
215-221-3961
STANLEY S. MANN, Ph.D.
215-221-3916
THOMAS R. C. SISSON, M.D.
215-221-3071
WILLIAM H. COLE
ADMINISTRATOR
215-221-3926
SKIN AND CANCER HOSPITAL
3322 N. BROAD STREET
PHILADELPHIA, PA. 19140
December 12, 1980
Dr. Richard B. Ward
E. I. DuPont de Nemours & Company, Inc.
Petrochemicals Department
Wilmington, Delaware 19898
Dear Dr. Ward:
Enclosed please find my report entitled, "Potential Effects on Humans of Al-
teration of the Stratospheric Ozone Layer: A Critical Review," which was prepared
at your request for use by E. I. DuPont de Nemours & Company, Inc.
Please feel free to use this report in any way that you deem appropriate.
I propose to submit all or part of this report to scientific journals for pub-
lication, and will keep you informed of the contents and the journals to which I
will submit the material for publication.
If you have any questions or comments, please do not hesitate to let me know.
Sincerely,
Frederick Urbach , M.D.
FU/jbh
enclosures
Z-967
-------
Effects—Skin Cance.
TABLE OF CONTENTS
Page
A.I. Summary F-l-5
A.2. Recommendations F-l-11
B. Historical Review of the Problem F-l-14
C. Potential Changes in "Biologically Effective" UVR Reaching
Earth Secondary to Anthropomorphic Alteration of the
Stratospheric Ozone Concentration F-l-28
C.I. Conclusions and Recommendations F-l-28
C.2. Action Spectra for Some Effects of UVR on the Skin F-l-28
C.2.1. The Human Erythema Action Spectrum F-l-28
C.2.1.2. Physiologic Mechanisms Involved in
Skin Erythema Development F-l-33
C.2.2. The Skin Carcinogenesis Action Spectrum F-l-37
C.3. The Effect of Weighting of Solar (Earth) Level UVR with
Various Action Spectra F-l-40
C.4. Comparison of Effect of Various Skin Action Spectrum
Weighting of Solar UVR with "Real," i_n vivo Experiments F-l-44
D. The Risk of Increase in Nonmelanoma Skin Cancer Due to
Increased Earth-Level Solar UVR F-l-48
D.I. Conclusions F-l-48
D.2. the Natural History of Nonmelanoma Skin Cancer F-l-51
D.2.1. Anatomic Distribution of Basal Cell and Squamous
Cell Skin Cancer F-l-54
D.2.2. Age and Duration of Exposure and Nonmelanoma
Skin Cancer F-l-57
F-l-2
-------
Effects—Skin Cancer
D.2.3. Race and Ethnic Extraction ' F-l-63
D.2.4. Geographic Distribution F-l-65
D.2.5. Relationship to Environmental Exposure and
Changing Incidence with Time F-l-77
D.3. Experimental Sk-in Carcinogenesis F-l-82
D.3.1. Mechanisms of UVR Carcinogenesis F-l-83
D.3.2. Tumor Types F-l-83
D.3.3. Ultraviolet Radiation as an Initiating Agent F-l-88
D.3.4. Interactions between Ultraviolet Radiation and
Chemicals F-l-89
D.3.5. Physical and Quantitative Aspects of UV
Irradiation in Animal Studies F-l-91
D.3.6. Dose Response Relationships F-l-93
D.3.7. The Immune Response to Tumor Induction -p-I-95
E. The Risk of Increase of Malignant Melanomas of the Skin Due to
Increases Earth Level Solar UV Radiation F-l-98
E.I. Conclusions F-l-98
E.2. The Natural History of Malignant Melanoma Skin Cancer F-l-102
E.2.1. Anatomic Distribution and Histopathology F-1-104
E.2.1.1. Lentigo Maligna Melanoma F-l-104
E.2.1.2. Superficial Spreading Melanoma F-l-105
E.2.1.3. Nodular Melanoma F-l-106
E.2.1.4. Classification of Malignant Melanoma F-l-108
E.2.1.5. Acral Lentiginous Melanoma F-l-110
E.2.2. Age, Sex and Site Distribution of Malignant
M i F-l-111
Melanoma
F-l-3
-------
Effects—Skin Cancer
E.2.3. Race, Ethnic Extraction and Heredity F-l-117
E.2.4. Malignant Melanoma and Exposure to Chemical
Agents F-l-121
E.2.5. Geographic Distribution of Malignant Melanoma F-l-122
E.2.6. Sunlight and Malignant Melanoma F-l-128
E.3. Experimental Models for Malignant Melanoma F-l-135
F.I. Forecasting the Effects of a 5-Year Delay in Regulation
of CFC Emissions in the U.S. F-l-137
F.I.I. DuPont's Comments on the NAS (1979) Report F-l-140
G. Critique of the NRC Report: Protection Against Depletion of
Stratospheric Ozone by Chlorofluorocarbons F-l-155
G.I. Chapter 3 - Human Health Effects F-l-155
G.2. Appendix C - Solar UV Irradiation at the Earth's
Surface F-l-164
G.3. Appendix D - The Biologically Effective UVR F-l-164
G.4. Appendix F - Factors in UV Dose-Response of
Npnmelanoma Skin Cancer and Malignant Melanoma F-l-164
G.5. Appendix G - Further Detail on Malignant Melanoma F-l-166
G.6. Appendix H - Preventive Measures Related to Malignant
Melanoma F-l-167
H. Recommendations for Further Research F-l-169
I. References F-l-175
F-l-4
-------
Effects—Skin Cancer
A.I. Summary
The high energy, short wavelength portion of the solar electro-
magnetic spectrum (wavelengths shorter than 320 nm) is potentially very
detrimental to living cells and tissues. A low concentration of ozone
formed in the stratosphere absorbs most of the photons of ultraviolet
radiation (UVR) and thus prevents most of them from reaching earth.
However, even in the presence of this ozone layer, which varies in
thickness in various latitudes and at various seasons, a biologically
significant amount of UVR reaches the surface of the earth.
It is a working assumption that byproducts of human activity in
recent years and in the foreseeable future may penetrate to the level
of ozone formation in the stratosphere and could result in a depletion
of this important "ozone shield."
Model calculations performed in the past decade have resulted in
the suggestion of a parametric range for the potential decrease in
stratospheric ozone due to various causes of from 1% to 50% (median
15%). A recent report of the National Academy of Sciences (NAS) en-
titled, "Protection against Depletion of Stratospheri.c Ozone by Chloro-
fluorocarbons" (NAS, 1979) estimates a median decrease in stratospheric
ozone at equilibrium of 16.5%, assuming continuing discharge of chloro-
fluorocarbons (CFC) into the atmosphere.
The major effects on humans of UVR in the UVB (320 to 280 nm) range
are on the skin and the eyes. Acute effects consist of "sunburn," an
inflammatory response of the tissues which may be no more than mild
redness, or slight stinging of the eyes, or may develop into the equiva-
lent of second degree (blistering) "burns." The acute effects of single
F-l-5
-------
Effects—Skin Cancer
overdoses of UVB are transient, heal without scarring, and in the skin
lead to adaptive changes of skin thickening and pigmentation, which
afford some degree of protection. The only established positive (bene-
ficial) effect of UVB in humans is the production of vitamin D precur-
sors in the skin, which are absorbed into the bloodstream and prevent
rickets, a serious vitamin deficiency disease. It should be recognized
that most work has been on the harmful effects of UVB, and relatively
little attention has been given to possible beneficial effects, however.
Repeated UVB exposure, prolonged over many years, can result in
chronic degenerative changes in skin, characterized by skin "aging"
and the development of premalignant and malignant ski.n lesions. The
skin cancers of man can be broadly divided into two types: nonmelanoma
skin cancer (NMSC) and malignant melanoma (MM).
There is excellent, although circumstantial, evidence that NMSC
is primarily due to repeated exposure of the skin to UVB. The major
arguments in favor of such a causal role of UVR in NMSC are:
• The most frequent location of NMSC is on the most exposed skin
sites (head, neck, arms, hands).
• Pigmented races, who sunburn much less readily than people with
white skin, have much less NMSC and when it does occur, it does
not affect the sun exposed sites.
• Among white skinned people, those who spend more time outdoors
and live in areas of greater UVR exposure (near the equator and
in tropical and semitropical areas) have much greater risk for
NMSC.
• Genetic diseases resulting in greater sensitivity to solar UVR
are associated with premature NMSC development (albinism).
F-l-6
-------
Effects—Skin Cancer
• Skin cancers of the NMSC type can readily be induced in the skin
of experimental animals, and the upper wavelength limit for this
is 320 nm, similar to the spectral, range producing sunburn in man.
Recent epidemiologic research has resulted in the development of
a preliminary dose-response relationship of NMSC to UVB. Reasonable
assumptions, based on tissue culture and animal experiments, for the
relationship of various wavelengths of UVR to NMSC induction have been
proposed. While there is still discussion about the exact relationships
of various parts of the UVR to carcinogenesis, existing data appear
reasonably sufficient for the development of model systems that allow
for some predictions as to the effect of reductions of stratospheric
ozone to the probable increases in the incidence of NMSC secondary to
increase in UVB.
Because of uncertainties in the amount of UVB actually reaching
various populations, and because parts of any population are signifi-
cantly more sensitive (for inherited reasons) to the chronic effects
of UVR, estimates of possible NMSC increases have ranged from a few
thousand to many hundred thousand of additional cases i.n the U.S.A.
alone. Projected estimates are. based on available data (which is im-
proving) and best estimates of their meaning (where considerable, dif-
ferences of opinion still exist).
In summary, there is reason to believe that most NMSC (but not
all, for about one-third of basal cell cancer occurs in areas not re-
ceiving much UVB and not showing other sunlight damage) is causally
related to chronic, repeated UVB exposure, and that a dose-response
function exists that can, with considerable future refinement, be used
F-l-7
-------
Effects—Skin Cancer
to predict changes in NMSC incidence given alterations of the UVB cli-
mate.
The situation is somewhat different where MM is concerned. The
NAS report (NAS, 1979) uncritically suggests that the relationship of
MM to solar UVR exposure is basically similar to that of NMSC, and that
the same model systems apply equally to NMSC and MM for predictive pur-
poses.
There is considerable evidence that such assumptions are not ten-
able. The matter is of great importance, since MM is a serious, often
(40%) fatal cancer. Although much less frequent in incidence, its po-
tential ability to cause death makes it of greatest importance to at-
tempt to develop a model for its causation, which may allow predictions
as to future incidence to be made.
Because of its seriousness, recording of cases of MM is usually
very good, so that epidemiologic. studies that are quite reliable have
been carried out in many countries in the past three decades. Very
briefly, the following are the. salient findings:
• There has been a consistent, worldwide increase in MM, incidence
rates increasing 3 to 7% per year, leading to a doubling in 10
to 15 years. This trend for increasing incidence began with people
born at or before the. turn of the century, and has been progres-
sively accelerating, so that younger persons have a greater risk
for development of MM than older ones. This is very unusual be-
havior for a malignant tumor.
• There are striking differences in anatomic distribution by sex
- MM is much more frequent on the. backs of men and the lower legs
F-l-8
-------
Effects—Skin Cancer
of women. The incidence in those areas has been increasing rapid-
ly, in contrast to MM of the head and neck area, and on the feet,
which rose only slightly. This distribution is entirely different
from that of NMSC.
Most MM is a disease of young adults. In contrast to NMSC, inci-
dence of MM rises sharply during adolescence, plateaus in middle
age, and rises again in old age. NMSC begins to appear in late
middle age and rises sharply in old age.
In contrast to NMSC, only about 10% of MM (a subtype that closely
resembles squamous cell cancer in its biologic behavior) show sig-
nificant evidence, of chronic solar skin damage.
While there are latitude gradients for MM within some countries,
the normal gradient of increasing incidence from higher to lower
latitudes is reversed in central Europe, and, particularly in the
Scandinavian countries, show incidence rates greatly in excess
of those expected based on relative UVB intensity found at their
latitude.
From available data, it appears that the latent period for MM
development is short, perhaps even 3 to 5 years. The latent period
for NMSC is certainly in excess of 20 years.
In contrast to NMSC, which affects mainly those chronically ex-
posed to sunlight, i.e., outdoor workers, MM is much more frequent
in white collar, educated, more affluent city dwellers.
As yet, no animal model for experimental production of MM exists
using UVR. Only recently has it been possible to produce MM ex-
perimentally in guinea pigs, using a chemical carcinogen.
F-l-9
-------
Effects—Skin Cancer
From the foregoing it is obvious that there are major differences
in most features of NMSC and MM.
It is obvious that, if UVR is causally related to the development
of MM, the mechanisms are very different. Certainly, except for the
10% Lentigo Maligna Melanoma that appears on the head and neck of old
persons, most MM can not reasonably be caused by the cumulative effect
of repeated exposures to UVR, as is the case with NMSC. The various
theories that have been proposed for its etiology are:
Related to UVR:
• Exposure to intermittent, high dose exposure, presumably due to
social changes, i.e., sunbathing, weekend gardening, golf, tennis,
etc. and vacation trips to sunny places (Spain, North Africa, the
Caribbean, etc.).
• Production by UVR of a "circulating factor" which causes distant
effects on precursor lesions.
• "Initiation" of melanocytes, either by chemicals or UVR, with ei-
ther UVR or chemicals acting as a "promoter" (i.e., two stage car-
cinogenesis).
• "Promotion" of preexisting, abnormal precursor cells or lesions
by UVR (precursor lesions perhaps genetically determined).
In any case, it is obvious that at this time, there is no possi-
bility of assuming a reasonable dose-response relationship of UVR to
MM development. In the absence of such information, attempts to predict
what changs may take place in MM incidence if the intensity or spectrum
of UVR changes are not appropriate.
There is certainly no rationale for using the models proposed for
NMSC development secondary to UVR exposure for evaluation of MM changes.
F-l-10
-------
Effects—Skin Cancer
It is in making such assumptions that the NAS report errs serious-
ly. There are a number of technical errors in the report. The most
glaring are: e.g., the wrong dose for production of minimal erythema,
the uncritical use of the deoxyribonucleic acid (DNA) action spectrum
used to derive the term "DUV" (damaging ultraviolet), the neglect to
heed warnings from agencies and investigators that the relationship
of MM to UVR differs significantly from that of NMSC, and that a com-
plicated mixture of environmental and host factors relate to MM, etc.
There are other technical problems - fabrics do not allow signifi-
cant amounts of UVB to pass, thus the assumption that such transmission
allows exposure of lightly clad backs is faulty.
The authors frequently confuse intensity and dose. There are fre-
quent speculations for which no real data exist, e.g.:
"MM does not correlate with cumulative lifetime exposure, which
suggests that intermittency of exposures may be important." (NAS,
1979, p. 325)
The first clause of that sentence is correct; the conclusion is a specu-
lation for which no real evidence exists.
Finally, calculations, using worst case assumptions based on NAS
data, strongly suggest that a 5 year delay in regulation of CFG will
not have a discernible effect on increases in incidence of NMSC or MM.
A.2. Recommendations
1. Measurements of Actual Skin UVB Dose
In order to establish more accurate dose-response relationships
between incidence of skin cancer and UVR exposure, it is necessary to
develop, test and use personal UVR dosimeters that allow estimation
of actual received dose of various populations.
F-l-11
-------
Effects—Skin Cancer
Two prototypes, based on the Robertson-Berger system for measuring
erythema effective UVR, are being developed at present. Such instru-
ments are urgently needed and should be used by persons of various types
of outdoor exposure, and at various latitudes.
2. Continuation and Extension of Actual UVB Climatology
While hardly perfect, data obtained from the present network of
Robertson-Berger meters has been most useful for modeling of UVR dose
by latitude. Additional stations need to be established, particularly
in the tropics and at high latitudes. The data obtained needs to be
acquired and disseminated on a more rapid and regular basis (at least
yearly) and be accessible to all investigators interested in these prob-
lems .
3. Animal Experiments investigating effects of dose rate, frac—
tionation of dose effects and interaction between wavelengths need to
be performed. The use of appropriate solar simulators and instruments
allowing tests of interaction of various wavelength bands is of con-
siderable importance.
Investigation is needed of the influence of change in flux of UVB
on skin carcinogenesis. Preliminary experiments show that protracting
the delivery of a dose of UV (i.e., altering flux) has a significant
effect on skin carcinogenesis. The direction of the effect is unex-
pected - lower flux can be more effective.
It is recommended that the effect of varying interval periods
during carcinogenesis experiments be studied in detail. The dose-
response model of Blum and the model for effect of "effective exposure"
of Robertson demonstrate that intervals of exposure are of importance
F-l-12
-------
Effects—Skin Cancer
in skin carcinogenesis. This may be one of the most important experi-
ments, since in nature, change in ozone will affect flux and spectral
distribution of UV but not the time relationship during which people
will be exposed.
At present, the only workable animal model for MM is the guinea
pig system reported by Pawlovski et al. (1980). However, this utilizes
chronic skin painting with a chemical carcinogen only. It is recom-
mended that experiments using the same guinea pigs but using UVB ex-
posure be performed. In addition, the interaction of chemical carcino-
gens and UVR should be investigated.
4. A Central Laboratory for measurement, calibration and reference
standards for UVR light sources and measuring devices needs to be estab-
lished, so that data obtained by various investigators with different
equipment can be appropriately compared.
F-l-13
-------
Effects—Skin Cancer
B. Historical Review of the Problem
Solar radiation is a very important element in our environment,
and yet, because of its ubiquity, the wide scope of its chemical and
biological effects is often not fully appreciated. The fact that solar
energy fixation makes life possible is widely known. It is not so
generally appreciated that many of the effects of solar radiation are
detrimental. Most people are aware that a painful sunburn can be caused
by excessive exposure to the sun. There are also more subtle effects
of sunlight on living cells, including the production of mutations.
The development of skin cancer may follow sufficient chronic exposure
to sunlight.
The shortwave portion of the solar spectrum is potentially very
detrimental to plant and animal cells. A low concentration of ozone
in the upper atmosphere filters out these harmful wavelengths of UV
radiation and thus prevents most of it from reaching the surface of
the earth. The formation of this ozone shield in geologic time was
most likely a prerequisite for the evolution of terrestrial life as
we know it. However, even in the presence of this ozone layer, a bio-
logicaly significant amount of UV radiation does reach the surface of
the earth.
It is a working assumption that human activity has not yet appreci-
ably modified the earth level solar spectrum. Initially, concern was
raised that exhaust products from a large fleet of supersonic transports
(SSTs) might seriously reduce the stratospheric ozone layer that shields
the earth's surface from solar UV radiation (CIAP, 1975). More recent-
ly, calculations of rate constants for an important chemical, reaction
F-l-14
-------
Effects—Skin Cancer
(H09 + NO), using improved techniques, suggest that NO injection will
£- X
not cause a serious perturbation of ozone. However, data presented
to a very recent conference on stratospheric 0~ have again changed the
predicted effect of NO on ozone.
More recently, it has been reported that continuing discharge of
chlorofluorocarbons (CFG) into the atmosphere or a major series of
atomic explosions above the earth's surface could have an even more
significant effect on the ozone layer (Hampson, 1974; NAS, 1979).
Recent computer calculations suggest that if chlorofluorocarbon
releases were to continue at the 1977 rate, a stratospheric ozone deple-
tion of about 16.5% could eventually occur, with about half this deple-
tion (or 8%) attained in 30 years (e.g., Stief et al. , 1978; Dickinson
et al., 1978; Crutzen et al., 1978; NAS, 1979).
The shortest wavelength of solar radiation observed on earth at
sea level, at the time of maximum solar elevation, is about 288 nm,
although the photosphere of the sun radiates significant amounts of
very much shorter UV radiation, since it corresponds to a brightness
r*
temperature of about 5,000 K.
The available UV solar radiation is depleted as it penetrates the
earth's upper atmosphere. This depletion is, in part, due to scattering
by air molecules, but it is mostly due to absorption by molecular oxygen
and ozone which is further enhanced by scattering. Molecular oxygen
absorption becomes strong at wavelengths below 200 nm and is important
to the thermal balance of the upper atmosphere above 50 km. The sig-
nificant spectral regions for ozone absorption in the UV are the con-
tiguous Muggins (320-360 nm) and Hartley (220-320 nm) bands. These
F-l-15
-------
Effects—Skin Cancer
bands absorb quite weakly at 360 nm, but rapidly increase in strength
with decreasing wavelength, reaching an absorption maximum at 255 nm.
At sea level, the effect of atmospheric ozone becomes significant below
325 nm, and there is a very sharp drop in transmission, from approxi-
mately 9070 of extraterrestrial UV radiation at 325 nm to 170 or less
at 295 nm. Thus, the absorption of UV radiation by ozone occurs pri-
marily in a wavelength region shown to be effective in the production
of skin erythema, chronic skin and eye damage, and skin cancer. Thus,
if a reduction in atmospheric ozone would occur secondarily to altera-
tions of conditions in the stratosphere, concern arises as to potential.
increases in the incidence of sunburn, degenerative skin and eye dis-
orders, and skin cancer in man (CIAP, 1975; NAS, 1973, 1975, 1979).
A parametric range for the potential decrease in stratospheric
ozone due to various causes of from 1 to 50 (median 15) percent has
been assumed by various authors (CIAP, vol. 5, 1975; NAS, 1975, 1979).
Based on the admittedly inadequate epidemiologic data, this has given
rise to estimates of increases in skin cancer ranging from a few thou-
sand to many hundreds of thousands of additional cases in the United
States alone. The huge range in estimates results in part from the
present uncertainty about the shape of the "biologic action spectrum"
for skin cancer induct ion and the lack of precise knowledge of the amount
of "carcinogenic" UV radiation reaching populations at risk. It is
important to recognize that projected estimates are based on available
data and best estimates of their meaning. Additional surveys of natural
UV insolation, analyses of additional epidemiologic data, and further
experiments with animals would greatly reduce some of the uncertainty.
F-l-16
-------
Effects—Skin Cancer
Although potential effects of increased solar irradiance on man,
particularly in the form of skin cancer, is a prominent and sensitive
issue, impacts of this increased irradiance on organisms of terrestrial
(both agricultural and nonagricultural) and aquatic ecosystems may ulti-
mately be of much greater concern. Certainly our understanding of the
sensitivity of plants, invertebrates and vertebrates to increases of
UV irradiance is not nearly as well developed as is the case for homo
sapiens. Without a reasonable understanding of what a unit decrease
in ozone means for the biosphere, there is no perspective upon which
to base decisions concerning acceptability of "allowable" change.
Depletion of stratospheric ozone and consequent increase of UV
irradiance attributable to SSTs of the size and configuration of Con-
corde are now contemplated to be rather small compared to present-day
ambient fluctuations or to the diminutions of ozone postulated to result
from other activities of man (NAS, 1979). On the other hand, present
estimates of ozone reduction due to CFC release are estimated to reach
as high as 16.5% (NAS, 1979). Thus there exists a critical need to
investigate the effects of enhanced UV irradiance on organisms. It
would not be unreasonable to postulate that the added stress imposed
by UV radiation increase due to any one agent, though small, could be
a critical amount when added to that increase of UV resulting from ef-
fluents of super- and subsonic aircraft, diffusion of chlorofluoro-
carbons into the stratosphere, release of nitrogen oxides into the
stratosphere by nuclear explosion, etc. Since chemical injectants from
several sources may be producing additive effects on ozone depletion,
that increment of diminution resulting from any one activity must be
F-l-17
-------
Effects—Skin Cancer
evaluated in a biological context in light of other potential changes.
For example, the effects of a 1% decrease of ozone from current levels
can not be compared with the effects of a 1% decrease for an atmosphere
already partially depleted by other causes.
State of Current Knowledge Involving Effects of
Enhanced UV Radiation on Organisms
Photobiological research involving that region of the UV spectrum
under consideration (280-320 nm, referred to as UVB) has received little
attention. Most UV photobiology has employed 254 nm radiation because
it is easily generated by low pressure mercury vapor lamps and is ef-
ficiently absorbed by nucleic acids. Since wavelengths shorter than
285 nm are effectively absorbed even by an atmosphere severely depleted
of ozone, this "254 nm photobiology" is of little value in quantita-
tively assessing the impact of increased solar UVB irradiance (NAS,
1973). Only recently have experiments been initiated to specifically
address the question of the biological effects of increased UVB irradi-
ance as would occur with stratospheric ozone depletion. This work was
sponsored initially by the Department of Transportation Climatic Impact
Assessment Program (DOT/CIAP) and continued by the Biological and Cli-
matic Effects Research (BACER) program, now Stratospheric Impact Re-
search and Assessment Program (SIRA) of the U.S. Environmental Protec-
tion Agency.
These preliminary experiments and the earlier photobiological
literature indicate that: (1) Most observed biological effects of UVB
radiation are decidedly detrimental; (2) Most organisms have developed
a capacity to avoid excessive normal solar UVB radiation by absorption
F-l-18
-------
Effects—Skin Cancer
of Che radiation before it reaches sensitive physiological targets (e.g.
by cuticular waxes and pigment on plants; by feathers, fur, pigments,
etc. of animals), to avoid exposure by behavioral patterns, or to toler-
ate a certain UVB radiation stress such as by molecular, repair mechan-
isms; (3) CIAP (1975) indicated that the capacity to avoid or tolerate
this UVB radiation may be limited for most organisms, and that some
organisms may be already existing near a threshold where an increased
UVB irradiance regime would be detrimental. However, the relevance
of such experiments to field conditions on land and in oceans is uncer-
tain, because of difficulty in simulating appropriate conditions.
Biomedical Implications of UV Radiation Changes
for Humans
Acute Effects on the Skin. Erythema solare, or more commonly,
sunburn, consists in its mildest form of a reddening of the skin which
appears one to six hours after exposure to erythemogenic UVR and grad-
ually fades in one to three days. In its more severe forms, erythema
solare causes inflammation, blistering and peeling of the skin; it is
followed by tanning of the skin, which becomes noticeable within two
or three days following irradiation (Giese, 1965).
Although sunburn will occur with shorter exposures if stratospheric
ozone is reduced, protective measures now available (clothing, chemical
sunscreens, etc.) will prevent this from becoming a seriously limiting
factor for human activities.
Acute Effects on the Eye. That the biologic effect of UVA (320-
400 nm) and UVB on living tissue was known in very early times is in-
dicated by Xenophon's mention of snow blindness in his treatise Anabasis
F-l-19
-------
Effects—Skin Cancer
(379-371 B.C.)- Coordinated studies on the harmful effects of UV were
begun in the nineteenth century, but were directed at the effects of
the electric light on the eye.
Although more energetic than the visible portion of the electro-
magnetic spectrum, UV radiation is not detected by the visual receptors
in mammals, including man. Thus, exposure to UV may result in ocular
damage before the recipient is aware of the potential danger. Many
cases of keratitis of the cornea and cataracts of the lens due to UV
exposure have been reported. The UV radiation involved was produced
by welding arcs, high-pressure pulsed lamps, and the reflection of solar
radiation from snow and sand.
In the UVB region reaching earth from the sun, which has most bio-
logical effectiveness (290-320 nm), the action spectrum for photokera-
titis is almost identical to that for skin erythema. Thus all the con-
clusions about increased hazard from a change in this UV band applicable
to acute effects on the skin apply equally well to the eye (Pitts,
1970).
Vitamin D. Sunbathing is popular and there is a widespread feeling
that "sunlight is good for you," but the physiological benefits that
presumably underlie the feeling of wellbeing have not been adequately
explained or studied. Since recent concerns have centered on ozone
depletion and UVB increases, most recent research on the effects of
UVB has examined potential detrimental effects of increased UVB.
The only thoroughly established beneficial effect of UV radiation
on the skin is the conversion of 7-dehydrocholesterol to vitamin D3.
Renewed interest and recent progress in understanding of the mechanisms
F-l-20
-------
Effects—Skin Cancer
of vitamin D production and its metabolism and functioning have been
made by several investigators, including DeLuca's (1971) group at the
University of Wisconsin, where Steenbock did the classic work on conver-
sion of ergosterol in plant foods to dietarily effective vitamin D by
UV radiation. Steenbock's work was done in a period of rapid progress
in biochemistry and nutrition. However, as DeLuca has said, the clin-
ical existence of rickets and designation of vitamin D as a vitamin,
rather than a prohormone, is a "complication of civilization." Through-
out most of human existence 7-dehydrocholesterol was converted in the
skin by UV radiation to vitamin D.j. Only in northerly climates, and
particularly with the advent of cities, was the sun-produced vitamin 0
insufficient to prevent rickets in growing children and osteomalacia
in adults. Both of these diseases are produced by defective bone cal-
cification. Loomis (1967) describes the particular cruelty to poor
children inflicted by a tax on windows in urban housing. There is also
a claim that Neanderthal man was rachitic at a time of rapid climatic
change.
Anthropologists have long looked upon white skin as an adaptation
to low levels of UV radiation, with selection for lighter skin perhaps
\
occurring very rapidly during cloudiness associated with glacial periods
because of the impossibility of a woman with a rachitically deformed
pelvis to give birth. (This assumption is open to many questions.)
Loomis (1967) endeavored to extend the implications of the vitamin
D theory of skin color to the idea that white man in the tropics should
be expected to suffer from vitamin D intoxication. This idea was chal-
lenged in a series of letters to the editor of Science (Blois, 1968;
F-l-21
-------
Effects—Skin Cancer
Blum, 1968) and elsewhere (Daniels et al., 1972) on the grounds that
such intoxication had not been observed in white men in the tropics
and that Loomis was assuming higher exposures and efficiencies than
were likely.
There is evidence that increased dietary intake of vitamin D (al-
most always vitamin D,,) can produce vitamin D intoxication. Whether
this represents a difference in route of administration or a difference
between vitamins D? and D^ does not appear in the literature and is
an important area for urgent investigation. We know that vitamin D
intoxication can be produced by oral administration, and we know that
this effect appears to summate with UV effects on the skin, but we can
not at present answer the question, "Would an increase in environmental
UV radiation lead to vitamin D intoxication or poisoning in normal man
or other animals?"
Chronic Detrimental Effects of UV Radiation
UV Radiation and Skin Aging. Solar or actinic changes in human
skin (farmers' skin, sailors' skin) are shown by atrophy, freckling
with hyper- and hypopigmentation, dilated blood vessels and a yellow
discoloration due to increases of abnormal elastic tissue. Wrinkling
is also a prominent feature. Most attention related to hazards of in-
creased UV radiation in man's environment has been centered on skin
cancer, but for every patient who develops actual cancer, there are
probably hundreds who develop readily visible actinic changes. This
may be no problem in a "weather-beaten sailor" or a "red-neck farmer,"
but it can be a devastating problem in a woman who likes golfing or
boating and may look in some ways as if she were 70 years old when she
F-l-22
-------
Effects—Skin Cancer
is 50. Significant increases in UV radiation will make this a much
more common problem that can not be treated adequately or cured the
way skin cancer can be, and may lead to more psychological effects than
skin cancer itself.
Nonmelanoma Skin Cancer. In section C of this report, the existing
data for etiology, epidemiology and experimental production of nonmela-
noma skin cancer are described in detail.
Human Skin Cancer Production by UV. Examination of the sun's role
in the production of human skin cancer does not lend itself to direct
experimentation. However, extensive astute observations have strongly
suggested the etiologic significance of light energy in the induction
of these tumors. Skin cancers in.Caucasians in general are most pre-
valent in geographical areas of the greatest insolation and among people
who receive the most exposure, i.e. men who work outdoors. They are
rare in Negroes and other deeply pigmented individuals who have the
greatest protection against UV light injury. Further, the lightest
complexioned individuals, such as those of Scottish and Irish descent,
appear to be most susceptible to skin cancer formation when they live
in geographic areas of high UV exposure. When skin cancers do occur
in the darkly pigmented races, they are not distributed primarily in
the sun-exposed areas as they are in light skinned people. The tumors
in these pigmented individuals are more commonly stimulated by other
forms of trauma, such as chronic leg ulcers, irritation due to the lack
of wearing shoes, the use of a Kangri (an earthenware pot that is filled
with burning charcoal and strapped to the abdomen for warmth), the wear-
ing of a Dhoti (loin cloth), and so on. In contrast, the distribution
F-l-23
-------
Effects—Skin Cancer
of skin cancer in the Bantu albino and in patients with xeroderma pig-
mentosutn follows sun exposure patterns.
Blum (1959), Urbach et al. (1969, 1972) and most recently Emmett
(1974) have reviewed the evidence supporting the role of sunlight in
human skin cancer development. Briefly, the main arguments are:
• It is clearly established that superficial skin cancers occur most
frequently on the head, neck, arms and hands, parts of the body
habitually exposed to sunlight.
• Pigmented races, who sunburn much less readily than people with
white skin, have very much less skin cancer and when it does occur,
it affects areas not exposed to sunlight most frequently.
• Among Caucasians there appears to be much greater incidence of
skin cancer in those who spend more time outdoors than those who
work predominantly indoors.
• Skin cancer is more common in white skinned people living in areas
where insolation is greater.
• Genetic diseases resulting in greater sensitivity of skin to the
effect of solar UV radiation are associated with marked increases
and premature skin cancer development (albinism, xeroderma pig-
mentosum).
• Superficial skin cancers, particularly squamous cell carcinoma
of the skin, occur predominantly on the areas receiving the maximum
amounts of solar UV radiation and where histologic changes of
chronic UV damage are most severe.
• Skin cancer can be produced readily on the skin of mice and rats
with repeated doses of UV radiation and the upper wavelength limit
F-l-24
-------
Effects—Skin Cancer
of the most effective cancer producing radiation is about 320 nm,
that is the same spectral range that produces erythema solare in
human skin.
Though these arguments do not constitute absolute proof, there
is excellent epidemiologic evidence supporting the role of sunlight
in nonmelanoma skin cancers.
Malignant Melanoma. In section D of this report, the existing
data for etiology, epidemiology and experimental production of malignant
melanoma are described in detail.
Suffice it to state here that, while malignant melanoma as an en-
tity has been known and described for over 150 years, the possible etio-
logic relationship to solar ultraviolet radiation has only been sug-
gested in recent years. Until 1948, MM was not separated from other
skin tumors in international statistical classifications, and thus know-
ledge of its variation from population to population was effectively
prevented. An equally important factor is the lack of concentration
of MM on light-exposed skin surfaces, in obvious contrast to the distri-
bution of nonmelanoma skin cancer.
The original and strongest evidence for the possible importance
of exposure to sunlight as a causal factor for development of MM in
white people is the apparent influence of'latitude of residence on the
incidence and mortality of MM. As will be seen in section D, this phe-
nomenon is not simple, or even realistically similar to the latitude
gradients found regularly for nonmelanoma skin cancer.
However, since 35 to 40% of all patients affected by MM die of
their disease, and since the incidence rate of MM is rising at an alarm-
ing rate, the need to determine etiologic factors in this malignant
tumor is urgent.
F-l-25
-------
Effects—Skin Cancer
Summary
Solar ultraviolet radiation shorter than wavelength 320 nm is bio-
logically highly effective, and the majority of the effects on living
organisms are detrimental. Most of the UVR emitted by the photosphere
of the sun is effectively absorbed by a thin layer of ozone produced
photochemically from oxygen in the stratosphere. There is considerable
concern that anthropomorphic effluents, particularly NO , Cl and Br,
can alter stratospheric ozone concentrations. Concern in recent years
has been primarily directed at NOx emitted from aircraft capable of
stratospheric flight, and Cl derived from destruction of chlorofluoro-
carbons in the stratosphere. Estimates of a calculated 16.5% reduction
in stratospheric ozone concentration (at equilibrium) have been made,
based on incomplete knowledge of stratospheric chemistry and various
model systems.
It should be noted, however, that preliminary calculations using
two dimensional modeling (Pyle and Derwent, 1980) suggest calculated
ozone depletion and calculated DUV increase is not uniform with latitude
and season. If this is correct, estimates of effects would have to
be revised downwards.
It is certain that a reduction in stratospheric ozone concentration
will result in an increase of biologically effective UVR reaching earth.
Among the known effects of chronic exposure of skin to UVB are degenera-
tive changes in the connective tissue of the skin leading to "skin
aging" and the induction of nonmelanoma skin cancer. The degree to
which such changes take place is markedly affected by dose and dose-
rate of UVR, as well as pigmentation and genetic predisposition.
F-l-26
-------
Effects—Skin Cancer
Much less clear is the relationship of UVB exposure of skin to
the incidence of malignant melanoma, a potentially fatal skin malig-
nancy.
F-l-27
-------
Effects—Skin Cancer
C. Potential Changes in "Biologically Effective" UVR
Reaching Earth Secondary to Anthropomorphic Alteration
of the Stratospheric Ozone Concentration
C.I. Conclusions and Recommendations
Of the various action spectra proposed for skin carcinogenesis,
the portion of the DNA action spectrum up to about 280 nm published
•
by Setlow (1974) appears to be closest to experimental observations
in mice.
The exact shape of an action spectrum used to weight calculated
or measured spectral data of solar UVR is of critical importance for
the estimation of any effects attributable to changes in quality and
quantity of solar UVR reaching biologic targets.
Recommendations
1. In any calculations concerning biologic effects of UVR it is neces-
sary to check with known observed biologic reactions by utilizing actual
values substituted in any model system, to determine whether the effect
matches the prediction.
2. Considerable further experimentation is needed to determine the
exact shape of a skin carcinogenesis action spectrum, particularly with
wavelengths between 300 and 280 nm. Animal experiments, using filtered
light sources simulating field conditions, are feasible.
C.2. Action Spectra for Some Effects of UVR on the Skin
C.2.1. The Human Skin Erythema Action Spectrum
During the 1920's and early 1930's, the relationship between the
degree of erythema produced and a given wavelength of UV was studied
by several teams of scientists. Hausser and Vahle (1927) reported the
F-l-28
-------
Effects—Skin Cancer
first precise determination of the action spectrum for the erythema
of human skin; a double peak was shown with maxima at about 250 and
297 nm and a minimum at about 280 nm. In these and related studies,
the skin of several individuals was exposed to UV radiation from a mer-
cury lamp passed through a double-quartz-prism monochromator, and the
influence of wavelength, exposure time and exposure rate upon the na-
ture, degree and course of erythema was examined. Similar action spec-
tra were published by Lukiesh et al. (1939), using a mercury arc lamp
and filters, and by Coblentz and Stair (1934), using a quartz-prism
monochromator.
Coblentz and Stair (1934) formulated a standard skin erythema curve
(Fig. 1). This curve represented the relative effectiveness of equal
amounts of energy at different UV wavelengths for producing erythema.
The most effective wavelength in the then accepted standard curve was
297 nm;. 250 nm being 60% and 280 nm about 5% as effective as 297 nm.
Recent reports have shown that, although 280 nm is somewhat less ef-
fective than adjacent wavelengths in producing erythema, this reduction
in efficiency is by no means as low as the old standard curve suggests.
In addition, doubt has been cast on the peak of efficiency being at
297 nm. A comparison of various action spectra for UV erythema up to
1966 has been carried out by Johnson et al. (Fig. 2). One important
factor to consider when comparing results from different laboratories
is the region of the skin that was used as the test site. Olson et
al. have reported that the minimum erythemal dose may vary considerably
with different anatomical locations.
Recently, Magnus, Everett et al., Freeman et al., and Berger et
al. have reported UV action spectra that have several features in common
F-l-29
-------
Effects—Skin Cancer
240 250 260 270 280 290 300 310 320
Wovelengths (nyim)
Figure 1. Action spectrum for erythema production. L, H and T, 1930;
Luckiesh, Holladay and Taylor (1930). H & V, 1927; Hausser (1928).
C, S & H, 1931; Coblentz, Stair and Hogue (1931). H & V, 1922; Hausser
and Vahle (1922). [From W. W. Coblenz, R. Stair and J. M. Hogue, Bureau
of Standards Journal of Research, 8, p. 541 (1932).] (From Coblentz
and Stair, 1934.)
F-l-30
-------
Effects—Skin Cancer
250 270 290 310
•Standard curve'I.C.I. Berlin (1935)
Luckiesh, Hoiladay S Taylor (1930)
Luckiesh 8 Taylor (1939)
CoWentz, Stair a Hogue (1932)
250 270 290 310
Freemen et al. (1966)
250 270 290
Everett eto/.( 1965)
310
Wavelength (nm)
Figure 2. Action spectra for UV erythema. Hausser (1928): Forearm
skin. Spectral lines from medium pressure mercury arc separated with
quartz prism. Luckiesh et al. (1930): Back skin. Spectral lines from
medium pressure mercury arc separated with filters. Mathematical cor-
rection for spectral contamination. Luckiesh and Taylor (1939): Upper
arm, inner aspect. Source and corrections as for Luckiesh et al. (1930)
Coblentz et al. (1932): Forearm skin. Spectral lines from medium pres-
sure mercury arc separated with quartz prism. Freeman et al. (1966):
Abdomen skin. Continua from three high pressure xenon arcs, dispersed
through Bausch and Lomb grating monochromators to single focus. Slit
widths: 2.7 mm 1.5 mm. 100% = 8.0 x 10^ ergs cm" . Everett et al.
(1965b): Back skin Continuum from high pressure xenon arc dispersed
through Bausch and Lomb grating monochromator. Slit widths: 5.4 mm
3.0 mm. 100% = 6.6 x 10 ergs cm~
(From Johnson et al., 1968.)
F-l-31
-------
Effects—Skin Cancer
but also show some major differences. Cripps et al. were in agreement
with these authors that the most effective wavelengths are in the 250-
260 nm range and that effectiveness falls off rapidly above 300 nm.
In the region of 270-280 nm, Magnus found 270 nm to be the least ef-
fective wavelength. Freeman et al., who used abdominal skin as the
test site, showed a marked depression at 280 nm, whereas Everett et
al. reported a plateau between 280-290 nm. However, these latter au-
thors, although using the skin of the back of the trunk, recorded their
observations at 8 hours after the exposure rather than at 24 hours.
The results of Cripps et al. show a definite depression in the 270-280
nm region, with the least effective wavelength appearing at 275 nm at
24 hours and 280 nm at 8 hours. Berger et al. showed that the time
of determining erythema after irradiation had a great influence on the
shape of the action spectrum, particularly on the relative effectiveness
of the 250 and 297 peaks, with erythema due to 250 nm peaking 8 hours
after irradiation and that due to 297 nm at 24 hours.
The shortest wavelength that appears at present in terrestrial
sunlight is approximately 288 nm. Thus, in a normal person, the solar
erythemogenic region may be considered to extend from about 290 to 320
nm, but the lower wavelength limit is of course asymptotic. Cripps
et al. reported that, in the wavelength interval between 292.5 and 298.5
nm, the most effective wavelengths for producing erythema were 292.5
and 294 nm. Freeman et al., using rabbit skin and human abdominal skin,
found 292 nm to bethe most effective, and the preliminary results of
Berger et al. suggest that the most efficient wavelength in this region
is between 290 and 294 nm.
F-l-32
-------
Effects—Sk.in Cancer
A reexamination of the erythema response of human skin to UV radia-
tion was carried out by Berger et al. (Fig. 3). Using high-intensity
monochromatic radiation with minimal stray-light characteristics, they
found that stray light significantly affected the measurement of the
minimal erythema dose at wavelengths between 303 and 313 nm. Utilizing
the skin action spectrum cited above and the solar UV output spectrum,
the most effective radiation for producing erythema from solar radiation
was calculated to be between 305 and 310 nm, depending on solar alti-
tude.
The type of erythema produced depends somewhat on the wavelength.
Erythema from shorter wavelengths appears earlier and fades quicker
than that produced by longer wavelengths. Further, short wavelength
erythema is pink in contrast to the deeper red from longer wavelengths.
Blum and Terus (1945) found that large doses of deeper penetrating
radiation around 300 nm inhibited the action of wavelengths at 254 nm,
probably by injuring superficial blood vessels.
C.2.1.2 Physiologic Mechanisms Involved in Skin Erythema Development
Erythema caused by ultraviolet radiation is confined to the exposed
areas and reflects blood vessel dilation and increased blood flow in
the dermis. It is often assumed that the initial photochemical reaction
F-l-33
-------
Effects—Skin Cancer
per
cent
220
140
100
70
30
8 hrs MED
74 hrs MED
24 hrs 30 R
\
N
nm 254
280
297 303 313
Figure 3. "Action spectrum" of human skin. Averages of values for
five subjects, abdominal skin, second exit slit. Note great similarity
for wavelength from 297 to 313 nm, and marked differences for 8 h MED,
24 h MED and a curve constructed by using values for moderate erythema
(Kodak Color Balancing filter 30 R) (From Berger et al., 1962.)
F-l-34
-------
Effects—Skin Cancer
is in the epidermis, where photon absorption by keratinocytes may lead
to liberation of intracellular substances which diffuse into the papil-
lary dermis to cause vasodilation. This diffusion theory is supported
by the existence of a latent period between exposure and erythema, and
by the fact that much of the radiant energy of the erythemogenic wave-
length region is absorbed by the epidermis. However, there may also
be direct injury to the vascular endothelium or to other sites in the
dermis.
Most studies of the mechanism of "sunburn" have used artificial
sources of UVB or sources in which the spectral distribution is such
that UVB is assumed to provide the major erythemogenic influence. In
experimental animals, the vascular response to ultraviolet radiation
is biphasic. A transient immediate vasopermeability is followed, after
a latent period of 2 to 8 hours, by a delayed, prolonged, increased
vasopermeability and vasodilation. In some animal models the initial
vasopermeability is accompanied by a faint erythema which may begin
during exposure. This immediate effect has been attributed to histamine
release possibly due to a direct effect of photons on dermal mast cells.
There is evidence that serotonin may also play some role. In rats and
guinea pigs, serotonin and histamine antagonists suppress the immediate
phase of UV vascular responses. In vivo studies of human skin have
shown transient appearance of kinins within minutes after ultraviolet
radiation. Kinin was not found after the onset of delayed erythema.
The mediators of the delayed phase of ultraviolet induced vascular
response have been difficult to define. Antihistamines do not suppress
the delayed erythema phase of the vascular response to ultraviolet
F-l-35
-------
Effects—Skin Cancer
radiation in guinea pig, rats or man. Kinins have been stated to be
absent or not elevated. Delayed erythema was not suppressed by using
various inhibitors of proteases, plasminogen activators, or kallikrein.
Serotonin has been found in urine following ultraviolet exposure but
the significance of this finding is not clear.
More recently prostaglandins, a group of long-chain fatty acids
with yasoactive properties, have been implicated as possible mediators
of the delayed phase of erythema. Prostaglandins are produced in human
and animal skin (PGE and PGF groups of prostaglandins), intradermal
injection of prostaglandins produces erythema (PGE mainly, PGF group
are much less active) and furthermore the production of prostaglandins
increases following UV exposure. Indomethacin is a potent inhibitor
of the conversion of arachnidonic acid to active prostaglandin, a reac-
tion catalyzed by the enzyme prostaglandin synthetase. Topical indo-
methacin in both humans and guinea pigs produces a profound and pro-
longed blanching of UVB induced delayed erythema, and in humans intra-
dermal indomethacin has been shown to consistently decrease erythema
from UVB but not from all sites irradiated with UVC (wavelength 280nnm);
these effects appear to be due to inhibition of prostaglandin synthe-
tase.
It has been suggested that the complex reaction known as "sunburn"
may result from hydrolytic enzymes and possibly other substances re-
leased by lysosomes within keratinocytes. Johnson found that irradia-
tion with wavelengths shorter than 320 nm reduced the extractable acid
phosphatase in mouse ears, but did not deplete histochemically demon-
strable succinic dehydrogenase. He felt this was evidence for lysosomal
labilization in the absence of nonspecific leakage due to cell death.
F-l-36
-------
Effects—Skin Cancer
Histochemical studies of human skin exposed to ultraviolet light
were consistent with the theory that specific damage to lysosomal mem-
branes caused partial to complete rupture and release of enzymes.
The release of lysosomal enzymes may not only lead to damage of
the keratinocytes but may also release enzymes, vasodilator substances,
or subsequently formed cell-breakdown products into the dermis, where
they may directly or indirectly lead to erythema. It has been suggested
that the immediate erythema which results from ultraviolet radiation
results from disruption of lysosomes of endothelial cells with release
of chemical mediators. The delayed erythema could then result from
secondary diffusion of proteinases from the epidermis secondary to lyso-
somal rupture. It is also possible that direct photon damage to the
lysosomes of mast cells of the dermis or endothelial cells of dermal
blood vessels may play a role in the delayed erythema of sunburn.
Multiple chromophores may exist in skin, and radiation probably
leads to activation of a complicated cascade of mediators whose final
endpoint is erythema. Multiple pathways may exist. It is also possible
that photons have direct effects on blood vessels or nerves. Ultra-
violet light causes dilatation of isolated exposed dermal blood vessels.
Dermal proteins may be directly changed by radiation.
C.2.2 The Skin Carcinogenesis Action Spectrum
Human skin cancer, especially basal cell and squamous cell car-
cinoma, are closely associated with chronic repeated exposure of skin
to solar ultraviolet radiation (cf. D.2 below). Three types of evidence
indicate that the most effective wavelengths are shorter than wavelength
320 nm: (1) extensive experiments in mice show that wavelengths longer
F-l-37
-------
Effects—Skin Cancer
than 320 nm are almost ineffective for induction of skin cancer (Blum,
1959; Forbes et al., 1978; Freeman, 1978); (2) Wavelengths shorter than
320 nm are highly effective in inducing photochemical changes in DNA
and killing of cells in tissue culture (Setlow, 1974; Rothman and Setlow
1979). Furthermore, damage to DNA is considered to be one of the events
leading to carcinogenesis (Smith, 1976), and a number of carcinogenic
chemicals mimic UVR damage to DNA (Regan and Setlow, 1973); (3) The
effective wavelengths for human skin erythema production are below 320
nm. Also, individuals who sunburn easily and have high exposure to
solar UVR have a much higher incidence of nonmelanoma skin cancer than
those who sunburn rarely and have little exposure to the sun (Vitaliano
and Urbach, 1980).
This last observation has been used as a basis for assuming that
the human skin erythema action spectrum and the skin carcinogenesis
action spectrum are closely related (CIAP, 1975, vol. 5).
In 1974, Setlow suggested that a slightly altered DNA action spec-
trum would be a better representative than the skin erythema action
spectrum for skin carcinogenesis. He compared data obtained in bac-
teria, and later in cultured mammalian cells (Rothman and Setlow, 1979)
to the skin erythema action spectrum, and found that below 297 nm these
two action spectra differ by less than 20% provided the spectra were
appropriately normalized (Fig. 4). Given present day conditions of
stratospheric ozone, the erythema action spectrum would only slightly
underestimate the carcinogenic effect if DNA were the target, since
almost no radiation shorter than 297 nm reaches earth. The major dif-
ference of importance is that the erythema action spectrum turns down-
ward in effectiveness below 292 nm, while the DNA action spectrum
F-l-38
-------
Effects—Skin Cancer
I |i I I I I I I I I
i 1-° -
o 10-'.
UJ
QL
i 10-2-
^—
CO
z
% 10-3-
5.
o
o
g io-4-
^j
o
CO
UJ
g 10-=-
<
I
__1
UJ
" 10-'-
^V O ^^
\ /
°\ / SUNLIGHT
V
o /
\l
fc
k.
j \ .
/ \'
\ '•
1 \ \ ERYTHEMA
| V8'-
' 9 \
1 ' •
1 \ '
1 DNA A
\
^0 *
^•v
^^^s
-10"
-1011 I
CC
111
o.
co
{N '
e
£
-109 <
z
^
— _
-108 —
t
I
I
-106
i r iiiiir
260 280 300 320 340 360
WAVELENGTH (nm)
Figure 4
(from Setlow, 1974)
F-l-39
-------
Effects—Skin Cancer
continues to rise steeply in effectiveness to 265 nm. This difference
in shape is of considerable importance if stratospheric ozone were to
decrease - associated with an absolute increase in solar UVR shorter
than 320 nm is a shift to lower wavelengths (Fig. 5).
That there is indeed justification for using the DNA action spec-
trum in some form is strongly suggested by the experiments of Forbes
(1978). Utilizing a xenon arc filtered so as to represent the solar
spectrum of thinner and thinner layers of atmospheric ozone, Forbes
showed in mouse carcinogenesis experiments that the addition of small
amounts of radiation of wavelengths shorter than 300 nm very signifi-
cantly increased experimental skin cancer induction. These experiments
strongly suggest that the DNA action spectrum indeed is applicable to
this problem rather than the erythema action spectrum (Fig. 6). Never-
theless, there is a caveat - experiments using UVR of 254 nm (Forbes
et al., 1975) show very little carcinogenic effectiveness. Thus, ab-
sorption by overlying dead layers of skin must have an effect. At this
time it is not known just where the peak effectiveness of UVR even for
mouse skin is located, except that it is shorter than 297 nm and longer
than 254 nm.
C.3. The Effect of Weighting of Solar (Earth) Level
UVR with Various Action Spectra
It has been pointed out above that some reasonable representation
of the DNA action spectrum, obtained in vitro, appears to be on theo-
retical and experimental grounds the most likely weighting function
for estimation of the carcinogenic effect of solar UVR.
F-l-40
-------
Effects—Skin Cancer
100.0
p
35
in
(A
111
10.0-
1.0-
0.1-
0.01-
0.001
0.0001
0.00001
_j I i I I I
ACTION SPECTRA
DNA t ERYTHEMA
- Normalized to297nm
ERYTHEMA -
Normalized to
297nm
/
DNA - Normalized
265 nm
265 275 285 295 305 315 325 335
nm
Figure 5 .
F-l-41
-------
Effects—Skin Cancer
10
15 20
Weeks
Figure 6. Tumor yield vs. time (expt. E). (From Forbes et al., 1978.)
F-l-42
-------
Effects—Skin Cancer
280 290
300 310
x (nm)
320
330 340
Figure 7. Comparison of some measurements of the spectral efficiency
for erythema and analytic representations of the data by least-squares
fitting. CS, BUD and CR denote data by Coblentz and Stair (1934); Ber-
ger et al. (1968), and Cripps and Ramsay (1970), respectively. (From
Green et al., 1974.)
F-l-43
-------
Effects—Skin Cancer
It has been shown that for wavelengths longer than 297 nm (which
includes virtually all present solar UVR reaching human populations)
there is a great similarity in slope and only 20% difference in quantity
between these two action spectra. As a matter of fact, if one goes
from 30 N northward, the DNA weighted irradiance decreases faster than
the erythema weighted irradiance (see Table 2 below, C.4 ).
However, numerical results of weighting of calculated or measured
solar spectra indicate a very considerable sensitivity to the specific
choice of action spectrum. This has been elegantly shown by Green et
al. (1974), These authors compared calculated dose rates vs. solar
zenith angle for four different erythema action spectra. In Fig. 7
are compared the action spectra obtained by Coblentz and Stair (1934),
Berger et al. (1968) and Cripps and Ramsay (1970) respectively. Table
1 shows the results of such a study. Dose rates obtained with the
Berger et al. (1968) action spectrum were approximately 80% of those
calculated from the Coblentz and Stair (1934) data. This indicates
that the marked differences between these response functions in the
320 to 297 region have very little effect on the erythema dose calcula-
tions. However, for the "long tailed" Cripps-Ramsay spectrum and for
the "sharp" Green and Mo (1974) spectrum, significant relative changes
with latitude and declination of the sun will arise (see also below,
C.4).
C-. 4 . Comparison of Effect of Various Skin Action Spectrum Weighting
of Solar UVR with "Real" In-Vivo Experiments
In almost all calculations made for modeling purposes, either only
relative (or ratio) increases of biologically effective UVR have been
F-l-44
-------
Effects—Skin Cancer
Table 1
Erythema Radiation Doses
Dose rate vs. solar zenith angle for different erythema action spectra
Angle
(deg.)
0
10
20
30
40
50
60
70
80
Coblentz & Stair
(1934)*
i
2.56 x 10
2.45 x 10"1
1
2.12 x 10
1.65 x 10"1
1.11 x 10"1
6.11 x 10~2
2.47 x 10~2
6.17 x 10~3
8.90 x 10~4
Berger et al.
(1968)**
0.799
0.799
0.800
0.801
0.802
0.804
0.804
0.802
0.796
Cripps & Ramsay
(1970)**
0.869
0.877
0.902
0.953
1.049
1.233
1.625
2.560
4.787
Green & Mo
(1974)**
0.540
0.538
0.529
0.514
0.491
0.454
0.390
0.258
0.057
*Dose rate in erythemal-effective W/m .
**Dose rate relative to Coblentz-Stair value.
(From Green et al., 1974)
F-l-45
-------
Effects—Skin Cancer
used or dose-rates have been calculated and annual doses based on such
calculations have been compared, without any one having put "real" num-
bers to such models.
*t i
To give examples:
A. Using a reasonable calculated irradiance for noon summer solar
UVR at 30 N latitude and the Cripps-Ramsay action spectrum, and
the data for minimal erythema doses (MED) at their peak value,
a minimal erythema dose for untanned human skin should be obtain-
able in 111 seconds, or less than 2 minutes. All known actual
observations show that at least 10-12 minutes are required (Nacht-
wey, personal communication).
B. Using a typical skin erythema spectrum in which relative ef-
fectiveness is related to wavelength by an analytical representa-
tion determined by a nonlinear least square analysis of five pub-
lished erythema action spectra, and an MED aft 300 nm of 206 J/m2,
the time necessary to produce an MED at 30°N is 12 minutes, or
a time consistent with actual observation (Nachtwey and Rundel,
in press).
Using the same MED, but the doses calculated for various ac-
tion spectra by Green et al. (1974) one finds (assuming a sun angle
of 20°):
Coblentz & Stair 16.19 minutes
Berger et al. 12.96 minutes
Cripps and Ramsay 14.61 minutes
Green and Mo 8.57 minutes
Again, the Berger et al. spectrum, most similar to the one
used above, gives reasonable results.
F-l-46
-------
Effects—Skin Cancer
Finally, Table 2 below is instructive in terms of the actual ir-
radiances under simulated field conditions. (Parenthetic irradiances
are normalized to 30°N.)
W/M2
Source
Sun 0
Sun 30° N
Sun 40° N
Sun 50° N
Unweighted
Irradiance
270-360 nm
34.5
33.6
31.7
28.6
Erythema
Weighted
Irradiance
386 x 10"3 (1.31)
294 x 10~3 (1.0)
238 x 10~3 (.81)
_3
178 x 10 (.61)
DNA
Weighted
Irradiance
11.14 x 10~3 (1.51)
7.39 x 10"3 (1.0)
5.53 x 10~3 (.75)
3
3.77 x 10 (.51)
DNA + Skin
Transmission
Irradiance
1.84 x 10~3 (1..41)
1.3 x 10~3 (1.0)
1.01 x 10~3 (.78)
0.72 x 10~3 (.55)
Table 2.
(Modified from Nachtwey and Rundel, in press)
F-l-47
-------
Effects—Skin Cancer
D. The Risk of Increase in Nonmelanoma Skin Cancer
Due to Increased Earth-Level Solar UVR
D.I. Conclusions
From a few detailed epidemiologic investigations and the data
available from a world compilation presented in Cancer Incidence in Five
Continents some figures for incidence of basal cell and squamous cell
carcinoma can be obtained which show a geographic variation from annual
skin cancer incidences as low as 5/100,000 to over 200/100,000. From
these data it is also possible to estimate that the incidence of skin
cancer approximately doubles for every 10 degrees of latitude, provided
that the population is of reasonably similar genetic stock.
Various models have been proposed, relating a decrease in strato-
spheric ozone to potential increases in NMSC. Most of these are based
on two steps: a calculation of the expected increase of "biologically
effective" ultraviolet radiation, and then a number relating UVR to
increase in NMSC.
Existing models suffer from several major uncertainties which can
significantly affect the resultant estimates: The choice of the appro-
priate action spectrum to determine the magnitude of the physical ampli-
fication factor (0, to UVR); the choice of the configuration and data
source for estimation of dose-response relationships leading to the
biologic amplification factor (UVR to cancer); the potential confounding
effects of age, genetic background, occupation, etc., etc. affecting
actual dose of UVR received by the skin.
At this time, an overall amplification factor (0^ to cancer) of
approximately 4 times is predicted by the most recent models. However,
F-l-48
-------
Effects—Skin Cancer
no separate analyses have as yet been performed for basal cell carcinoma
and squamous cell carcinoma, which have clearly different relationships
to UVR. Additionally, preliminary two dimensional modeling of ozone
depletion (models in which the north-south or latitude dimension is
expressed) suggest this factor may be substantially reduced due to the
distribution of ozone depletion towards the poles and to the winter
season (Pyle and Derwent, 1980).
It has been suggested many times that prolonged exposure to sun-
light can result in the development of skin cancer in man. Furthermore,
the person most likely to develop skin cancer has been characterized
(mostly by inference) as having a light complexion, blond or red hair,
and blue eyes. He or she supposedly sunburns and freckles easily and
is often of Celtic ancestry.
As a result of the early studies of Unna and Dubreuilh and many
others since the turn of the century, a number of arguments support the
belief that sunlight is a causal factor in human skin cancer. Among
the main arguments in favor of this assumption are:
1. It is clearly established that skin cancer occurs most fre-
quently on head, neck, arms and hands.
2. Pigmented races, who sunburn much less readily than do people
with white skin, have much less skin cancer, and when it does
occur, it is not predominantly found on light-exposed areas.
3. Among Caucasians there appears to be a greater prevalence of
skin cancer in those who spend more time outdoors than in those
who work predominantly indoors.
4. Skin cancer is more common in white-skinned people living in
areas where insolation is greater.
F-l-49
-------
Effects—Skin Cancer
5. Skin cancer can be produced readily on the skin of mice and
rats with repeated doses of ultraviolet radiation; the upper
wavelength limit of cancer-producing radiation is about 320
nm, i.e., the same spectral range that produces sunburn in
human skin.
Although statements asserting one or more of the above assumptions
have been frequently published, adequate epidemiologic information to
substantiate them has been found difficult to obtain. The main reasons
for this difficulty are due to the fact that cancer of the skin is ex-
tremely common in whites (amounting to 1/3 to 1/2 of all cancers of
all sites in the United States) and that it is rarely fatal.
Unlike cancer of most other sites, most skin cancers are diagnosed
and treated in physicians' offices and the apparent diagnostic advantage
of accessibility of the lesions can be nullified by the need for compre-
hensive coverage of many sources of reports outside hospitals that main-
tain different practices for the recording of data. Records in physi-
cians' offices often may indicate little more than the patient's name
and omit such items as age and residence that are helpful for identi-
fication of individuals in order to eliminate duplicate case reports
and are necessary for the calculation of rates specific for age and
other demographic factors. Also, many specimens escape microscopic
examination and, what is worse, the percentage not so examined will
vary among communities and with such host characteristics as age and
site of the lesion. Skin cancers are usually not lethal, and mortality
data can not be relied on to yield information on differentials in risk
by age, sex, race and other characteristics with the same precision
as for other sites with less favorable prognoses.
F-l-50
-------
Effects—Skin Cancer
Despite these difficulties, certain facts about skin cancer seem
to be established beyond dispute. These include the low risk in heavily
pigmented peoples, the continued rise in incidence rates throughout
the life span, and the higher rates among people living in subtropical
and tropical latitudes.
D.2. The Natural History of Nonmelanoma Skin Cancer
There has been a recent reawakening of interest in the field of
photobiology. This has been largely due to rapid advances in photo-
chemistry, particularly due to a much better understanding of photobio-
logic phenomena occurring at the molecular level, and a recent concern
about potential alteration of the ozone layer. Furthermore, the great
improvement in design, versatility and intensity of modern light sources
and the development of accurate measuring devices capable of determining
intensity of very narrow bands of light even in the ultraviolet region
of the electromagnetic spectrum have made more elegant studies possible.
That sunlight can cause demonstrable acute and chronic changes
in apparently normal skin certainly has been known since antiquity.
"...I am dark, because the sun has scorched me." (Song of Solomon).
Charcot (1859) determined that ultraviolet radiation caused acute ery-
thema and Unna proved that pigmentation can be induced by ultraviolet
radiation. It soon became apparent that the capability of the skin
to react to light by pigmenting was most variable, and that this varia-
bility pertained not only to different races but also to individuals
of apparently similar ancestry.
Unna (1894) noted severe degenerative skin changes on exposed areas
of the skin of sailors and associated these with the development of
skin cancer which was seen with great regularity in his clinic in Ham-
burg, an old Hanseatic seaport town. Dubreuilh (1896, 1912), studying
F-l-51
-------
Effects—Skin Cancer
skin diseases in the Bordeaux region of France, noted the frequent oc-
currence of keratoses and skin cancer in vineyard workers, while the
nearby city dwellers showed few such lesions. These observations were
soon confirmed by Shield (1899), Hyde (1906), and others, who noted
a high incidence of skin cancer in rural areas of the U.S. and Austra-
lia, where light exposure was much more intense than in central Europe.
It is interesting (and almost prophetic) that Bruusgaard in 1922 con-
sidered the sailors' skin cancer as being due to a combination of sun-
light and coal tar (to which sailors were heavily exposed in those
days).
Following the clinical observation of a relationship of chronic
sunlight exposure to skin cancer, there was much discussion among der-
matologists as to whether this association applied to all white-skinned
people or, as Haxthausen (1929) had proposed, really occurred only in
those carrying a fo rme fruste trait of xeroderma pigmentosum. This
view began to change when Findlay in 1928 showed that daily irradiation
of mice with ultraviolet radiation from a mercury arc caused the induc-
tion of skin cancers. Incidentally, Findlay also noted that when mice
were tarred prior to ultraviolet radiation exposure, the period neces-
sary for the induction of skin cancer was reduced. In 1933 he was able
to produce skin cancer in rats with ultraviolet light, and his findings
were soon reproduced by Putschar and Holtz. In a series of studies
between 1930 and 1936, Roffo showed that skin cancer could be induced
in rats with natural sunlight as well as with mercury arc radiation,
and carried out the first real epidemiologic study of human skin cancer.
As had Dubreuilh, he pointed out that the same skin areas most likely
to develop skin cancer also showed a great tendency to develop hyper-
keratoses, and considered these keratoses as premalignant lesions.
F-l-52
-------
Effects—Skin Cancer
Finally, Roffo carried out the first action spectrum studies of skin
photocarcinogenesis: he showed that clear window glass was sufficient
to stop skin cancer production by both natural sunlight and mercury
arc radiation, thus setting an approximate limit for the effective UV
radiation of shorter than 320 nm.
Classical evidence in support of the role of sunlight, particularly
of ultraviolet radiation, as a causal factor in human skin cancer com-
prises a number of factors. • Most of these factors may be summarized
by six associations of skin cancer:
1. Association with exposed areas of the skin. Among Caucasians,
skin cancers occur most frequently on parts of the body most ex-
posed to sunlight - the head, neck, arms and hands, and the up-
per back.
2. Association with protection against ultraviolet radiation.
Among races with dark skin, in which pigment filters UV radiation,
there is very little skin cancer and the disease does not pre-
dominantly occur in areas of the skin exposed to the sun. Sunburn
and skin cancer arise in the same tissue, and UV radiation is known
to cause sunburn. It appears that those who are more susceptible
to skin cancer sunburn more easily. Caucasians of Celtic origin
are more susceptible to both skin cancer and sunburn; those of
Latin origin, less.
3. Association with the amount of exposure to the sun. Among
Caucasians there appears to be a greater prevalence of skin cancer
among those who spend more time outdoors.
4. Association with intensity of solar insolation. As one ap-
proaches the equator, measures of incidence of skin cancer among
F-l-53
-------
Effects—Skin Cancer
white-skinned people increase, as do the amount of solar radiation
and intensity of UV radiation.
5. Association of ultraviolet radiation in laboratory studies.
Skin cancer can be produced in mice with repeated doses of UV
radiation in the same spectral range that produces sunburn in human
skin.
6. Association with insufficient ability to repair DNA damaged
by ultraviolet radiation. Those with the recessively inherited
disease xeroderma pigmentosum, who consequently have a defect in
the repair of DNA damaged by UV radiation, have a predisposition
to the development of skin cancer. Such victims are often photo-
sensitive, and their tumors appear to be induced by sun exposure.
They usually die as a result of skin malignancies before reaching
adult life.
D.2.1 Anatomic Distribution of Basal Cell and Squamous Cell Cancer
Numerous studies have shown that skin cancers arise (in those sus-
ceptible, i.e., white skinned people) primarily on sunlight exposed
sites. From these studies it is demonstrable that about 90% of all
basal cell carcinomas (BCC) and more than half of all squamous cell
carcinomas (SCC) occur on the head and neck. The majority of those
squamous cell cancers not occurring on the head and neck are found on
hands and forearms, and the ears of females are markedly protected
(Table 3).
Comparing the sites of. non-melanoma cancers with studies made of
the geometry of insolation of the head and neck areas, it becomes clear
that 2/3 of all basal cell carcinomas occur on the skin sites receiving
F-l-54
-------
Effects—Skin Cancer
Table 3
Anatomical Distribution of Keratoses, Basal Cell
Carcinomas and Squamous Cell Carcinomas
Anatomical Site
Forearm and Wrist
Hands and Fingers
Side of Face
Ears
Scalp, Temple,
Forehead
Nose
Neck
Lower Limb
Upper Arm and Shoulder
Trunk & Elsewhere
Circumocular Region
Lower Lip, Chin
Upper Lip
Keratoses
Males
7,
47.7
19.0
7.3
6.4
5.0
4.9
2.8
2.6
1.6
1.1
0.9
0.5
0.2
Females
7o
39.3
22.7
11.5
0.5
8.4
8.7
1.1
2.6
0.8
1.0
1.8
0.9
0.7
Squamous Cell
• Carcinomas
Males
7o
22.9
15.3
13.7
8.8
5.5
7.2
8.0
5.0
1.5
3.1
3.3
4.1
1.6
Females
7.
15.8
19.5
17.4
1.6
12.5
12.5
4.9
3.3
0.5
2.2
4.3
3.3
2.2
Basal Cell
Carcinomas
Males
%
5.2
1.0
28.9
4.7
12.5
16.7
9.1
0.7
1.9
4.0
11.5
2.2
1.5
Females
7o
2.7
1.8
24.0
1.7
16.2
27.4
4.4
0.5
1.6
2.1
9.8
4.6
3.2
Ratio of Basal
Cell Carcinomas
to Squamous Cell
Carcinomas
Males
0.9
0.3
7.9
2.0
8.5
8.8
•4.3
0.5
5.0
4.9
13.3
2.0
3.5
Females
1.0
0.5
8.0
6.0
7.6
12.7
5.2
0.8
1 7 . 0*
5.5
13.9
8.1
8.5
1
Sources: Keratosis data from three regional surveys, cancers from Queensland Radium Institute
records
Only one squamous cell carcinoma was found at this site.
(From Silverstone & Searle, 1970)
F-l-55
-------
Effects—Skin Cancer
the highest UVR doses, and that virtually all squamous cell carcinomas
occur at these sites (Tables 3 and 4).
Table 4
Area
BCC
Head & Neck
(unshaded)
Head & Neck
(shaded)
sec
Head & Neck
(unshaded)
Head & Neck
( shaded)
Skin & Cancer Hospital
(Urbach 1972)
Male %
63.3
36.7
100.0
0
Female %
63.0
37.0
88.0
12.0
Total
63.2
36.8
94.0
6.0
Radiumhemmett
(Magnusson, 1935)
Male %
57.1
42.9
84.2
15.8
Female %
67.3
32.7
90.0
10.0
Total
1
62.2
1
37.8
86.5
13.5
Comparison of distribution of BCC and SCC over protected and unpro-
tected areas of head and neck. (The 6 and 13% incidence in shaded
areas at Skin and Cancer and Radiumhemmett represents 1 and 3
patients respectively.)
F-l-56
-------
Effects—Skin
D.2.2 Age and Duration of Exposure and Non-Melanoma Skin Cancer
Age specific incidence races have been obtained in several recent
skin cancer surveys. In Figs. 8&9 are given such data for the results
of the Third National Cancer Survey, carried out in four locations of
the U.S., and in Table 5 the data collected by Gordon et al. from the
literature for widely geographically separated areas. All figures are
for white skinned Caucasians only. In Figure 10 A & B , the graphs
are recorded on a log-log scale (log incidence against log age). If
the graphs for the various countries were parallel, this would mean
that their relative incidence rates were independent of age. The graph
for Queensland has a distinctly more shallow slope, indicating a rela-
tively greater spread of skin cancer into the younger age groups.
The most recent age specific incidence rates have recently been
published for 1977-78 and eight areas in the U.S. by Scotto et al.
(1980). These are given in Table 5 following.
From the factors listed in the introduction to this section and
studies utilizing Robertson type UVR meters in Australia, Philadelphia
and Galway, it is possible to arrive at some very crude estimates of
the amount of UVR required for susceptible individuals to develop skin
cancer. Clearly such an estimate has to be, at this time, considered
to be extremely crude and very preliminary. Robertson estimates as
follows:
Individual
Child (Goroka)
Outdoor
(North Queensland)
Outdoor (Victoria)
Outdoor (Galway)
| Fraction
I in Sun
I 3/5
1
1 2/5
1
1 1/2
I 3/5
Duration
(years)
6
20
30
50
S.U.
(per year)
4000
2600
1600
1000
Total
S.U.
24,000
52,000
48,000
50,000
S.U. - "Sunburn" unit is equivalent to 12-15 minutes of noon
equatorial clear day solar UVR.
F-l-57
-------
Effects—Skin Cancer
0-
o
a.
o
o
o
•
o
o
oc.
UJ
a.
VS
3200
2600
2400
2000
1600
1200
800
400
0
OflLLRS
IS- 25- 35- 45- 55-
24 34 44 54 64
flG€ CROUP
65- 75-
74 84
85 *
Figure 8. Age-specific incidence rates of nonmelanoma skin cancer among
Caucasian males. (From Scotto et al., 1974.)
F-l-58
-------
Effects—Skin Cancer
ae.
CL
O
0.
O
O
O
•
O
O
oc
UJ
d.
(O
*-
UJ
UJ
h-
a
QC
3200
2800
2400
2000
1600
1200
800
400
0
OflLLflS
niNN
-SflN FRflN
, - JOWfl
15-
24
25- 35- 45- 55- 65- 75-
34 44 54 64 74 84
nee GROUP
85*
Figure 9. Age-specific incidence rates of nonmelanoma skin cancer among
Caucasian females. (From Scotto et al., 1974.)
F-l-59
-------
'inofe j
Nonmelanoma Skin Cancer in the United States — NCI Incidence Surveys
TABLE 7, ANNUM. ACS-SPECIFIC, CRUDE, AGE-ADJUSTED (1970 0 S STANDARD) CKIK CANCER INCIDENCE KATES
PEB 100,000 POPULATION BY GEOGRAPHIC AREA, 1977 - 1970, ALL ANATOMICAL SITES - HHITE, DOTH SZXE3
SAH
OY-3 RADIATION INDEX" .
DEGREES N LATITUDE
<15
15-21
25-31
' 35-11
1
V 55-61 •
o 65-71
75-01
05 +
CRUDE RATE
ACE-ADJUSTED RATE
STANDARD ERROR X 1.96
SEATTLE
{MUG CO) .
101
'47.5
0.1
1.6
36.0
121.7
303,5
516.1
712.6
1125.7
1100.0
200.6
100.7
0.0
KINN-
ST PAUL
106
11.9
0.7
5.9
21.1
95.7
2 96.. 5
190.1
806.3 .
1200.9
1559.7
105.2
193.3
6.6
DETROIT
(SMSA)
110
12.2
0.3
3.5
1fi.7
ns.7
206.9
361.1
503.5
005.7
036.0
138.0 '
135.6
. 1.0
UTAH mucisco-
(STATE) OAKLAND
(SHSA)
•10. 7 t
0.8
8.5
56.2
213.2
517.9
881.9
1U12.2
1962.1
2323.1
271.5
337.6
11.3
151
37.8
1.3
5.7
30.1
133.1
311.0
61U.6 • -
871.6
1009.7
993.6
212.1
213.0
5.3
• ATLANTA
(SHSA)
33.7
***
8.1
70.0
219.9
639.2
1120.5
1676.0
2062.2
2027.7
330. U
393.0
12.3
NEV
ORLEANS
(HETRC)*
17C
30.0
1.3
5.5
. 60.5
212.0
629.9
1089.8
1505.3
2165.2
2350.2
399.1
301.2
11.3
NEV
MEXICO
(STATE)
197 t
35. It
1.1
1.S
38,5
107.9
515.7
917.3
1150.0
2071.9
2079.3
295.6
336.7
11.7
ALL
SURVIJ
AREAS
(101-197)
(17.5-30.0)
0.7
5.6
37.1
113.1
361.9
638.8 .
977.2
1305.2
• 1357.0
229.1
231.3
2. .6
•INCLUDES THBEE PARISHES: JEPFEItSON, ORLEANS, AND ST BERNARD
••INDEX « ESTIMATED OV-D COUHTS PER 10,000 PER ANHDH
tUYD COUNTS AHD LATITUDES ARE FOR ALBUQUERQUE, NEW HEXICO AND SALT LAKE CITI, UTAH RESPECTIVELT
W
Hi
Hi
fl>
O
rt
en
I
I
tn
(From Scolito cl. ;i I . ,
n
m
5
D
-------
Effects—Skin Cancer
INCIDENCE
PER
MILLION
10,000
5,000
0-31%
(1) Queensland Coastal Regions
(2) El Paso/Texas (Non-Latins)
(3) Cape Province, S.A. (Whites)
(4) South-west England.
30
40 50
AGE IN YEARS
60
70 80
Figure 10A
(From Gordon and Silverstone, 1976)
F-l-61
-------
Effects—Skin Cancer
INCIDENCE
PER
MILLION
10.000
5,000
1000
500
100
50
10
5
1 *
20
(1) Queensland Coastal Regions
(2) El Poso,Texas (Non-Latins)
(3) Cape Province, S.A. (Whites)
(A) South-west England
30 40 50
AGE IN YEARS
60
70 80
Figure 10B
(From Gordon and Si 1verstone, 1976)
F-l-62
-------
Effects—Skin Cancer
D.2.3 Race and Ethnic Extraction
It has long since been known that whites have a much higher inci-
dence of skin cancer than pigmented races (Urbach, 1963). The higher
incidence observed in sun exposed areas of skin of whites and albinos
is attributable to the chronic effects of UVR. Basal cell carcinoma
in particular is very rare in pigmented races (Higginson and Oettle,
1960; Urbach, 1963).
In contrast to the anatomic distribution of squamous cell carcinoma
in whites (70% on head and neck), Isaacson (1978) found only 1/3 of
squamous cell carcinomas in urban blacks in South Africa to be on the
head and neck. Furthermore, the male-female ratio was 60/40, as com-
pared to 85/15 in whites. Basal cell carcinoma was almost absent.
Furthermore, the age of onset was much earlier in blacks than whites,
50% of squamous cell carcinomas occurring before age 50!
In several carefully controlled studies, comparing white patients
with non-melanoma skin cancer to age-sex matched controls from the same
populations, it has been shown that patients with skin cancer have much
greater frequency of certain genetically transmitted traits than those
without skin cancer (Urbach et al., 1972; Gellin et al., 1966).
Skin cancer patients, as a group, had greater frequencies of light
eyes, complexion and hair color than did the controls. They also sun-
burned more frequently and tanned less easily than did the controls,
though the degree of tan achieved was only marginally less. Male cancer
patients had experienced greater cumulative outdoor exposure than had
male controls. The proportion of Irish and English (including Scots
and Welsh) ancestry was greater in cancer patients, while the Slavic
and "other" categories of national origin were smaller.
F-l-63
-------
Effects—Skin Cancer
Patients with basal cell and squamous cell carcinoma differed with
respect to age and sex distribution. In the case of basal cell car-
cinoma, the two sexes were almost equally represented; almost 15% of
the patients were less than 50 years of age, and nearly 65% were less
than 70 years old. In the squamous cell carcinoma group, on the other
hand, males outnumbered females by more than 6 to 1, and all patients
were at least 50 years of age, with more than 70% of them being over
70 years of age. Light (blue or green) eye color was also somewhat
more frequent in the squamous cell carcinoma group, but in the other
coloring characteristics, complexion and hair, the two cancer groups
were not significantly different. The two groups were also indistin-
guishable with respect to tanning and sunburning characteristics and
national origins. They differed significantly with respect to total
outdoor exposure, with a much larger fraction of the squamous cell group
appearing in the maximum exposure category (Vitaliano and Urbach, 1980).
These studies and the clinical impressions of many astute physi-
cians show a distinct association of skin cancer with light color eyes,
fair complexion, light hair color, poor ability to tan, ease of sun-
burning and a history of repeated severe sunburn. Furthermore, whenever
looked for, there is an increase in Celtic stock among skin cancer
patients.
But this clearly is not the whole story. There are many people
o
with origins other than Celtic who match these characteristics, and
it is apparent that the Celts inherit, but are by no means the sole
possessors of, a sun-sensitive, poorly adapting skin. Yet it is clear
that genetic factors are of great importance. In the three careful
F-l-64
-------
Effects—Skin Cancer
prevalence studies of Silverstone (1960), Macdonald and Bubendorf (1964)
and O'Beirn et al. (1968), the prevalence of skin cancer rises with
advancing age at about the same rate in Queensland and in Galway, dif-
fering primarily by age of first lesion. In El Paso the lesions begin
early, but the slope of the prevalence curve is much less. This is
probably the result of genetic differences since the population studied
in Queensland and Ireland is mainly Celtic, while the El Paso population
has few Celts and probably much Mexican admixture (Urbach et al., 1972).
D. 2.4 Geographic distribution
Incidence data for skin cancer, other than melanoma, must be
treated with considerable reserve. Many cancer registries do not
register non-melanoma skin cancer at all, and those that do are uni-
formly incomplete, since most of these tumors are treated in doctors'
offices and either not at all reported or reported without histologic
verification.
The best overview of geographic distribution has been given by
Gordon and Silverstone (1976) and is shown in Tables 6 and 7. The re-
sults are shown in the form of a cross classification of skin cancer
rate against latitudinal "zone." The tabulation is based on data for-
mulas, but in each cell of the table a figure is given for the average
ratio of female to male rates for the countries listed in that cell.
Some indication of the correlation between latitude and skin cancer
4
incidence is obtained by observing the "diagonal" pattern made by the
group of ethnically similar peoples starting with Sweden at "top left"
and running down to Queensland at "bottom right." On the other hand,
the more pigmented groups tend to be concentrated at the "top right"
of the tables.
F-l-65
-------
Effects—Skin Cancer
Average Incidence
of Skin Cancer per
100.000
(both sexes)
0.0-4. '.1
5.0- m.ii
20.0- 24. !l
30.0-4'.). VI
A B
Sweden** Poland
Rumania
Denmark
F/M = 0.48 F/M = 0.84*
Finland United Kingdom
German Dem. Rep.
F M = 0.88 F, M = 0.725
Canada
F. M =.0.r>7
Zone6
C
Japan
F/M = 0.80
Yugoslavia
F'M = O.Sli
New York Stale**
F. M = o.iili
Ne\ ad.i
F'M = (l.(>4
D
S.A. Cape Bantus
S. A. Cape Colored
Natal Africans
Natal Indians
Bulawavo Africans
F/M= i.27
Texas (Latin)
F/M= 1.01
E
Bomha\
Nigeria
F/M = 1.4>l
Jamaica
F/M = 0.94
Colombia
Puerto Rito
F/M = O.H5
50.0-W.9
100 and mcr
Victoria (Australia)
Tasmania (Australia)
F M = 0.4!l
S. A. Cape (whites)
Texas lnon-l.alin)<
Queensland (whiies)J
F;M = O..V.I
"Zonal latitudes are: A. ah
'''Zoning based on male ra
^Poland = (1.84: Rumania
ttExcludiriK New York Cit
§L'.K. = 0.38; G.D.R. = <>.-
UMales. 133.0: females. 72
T.Males. 168.2: lemales. 10
JMales. 265.1; females. 15
ve till": B. 4ric-till=: C. H;>°-4.ri°: 1). u'll'-:!.')°; E. II*- 20°.
es onh.
I.Oil: Denmaik = O.S'.I.
1; HungaiA = 0. 88.
1.
8.
Table 6. Global distribution of skin cancer incidence other than mela-
noma; male data with corresponding female to male ratios (F/M) (From
Gordon and Silverstone, 1976.)
F-l-66
-------
Effects—Skin Cancer
Average Incidence
of Melanoma per
100,000
(both sexes)
0.0-O.tl
1.0-1.9
2.0-2.9
3.0-3. SI
4.0-4.9
5.0 and over
A B
Rumania (Banal)
F/M = 0.5(1
United Kingdom
Poland
Hungan
F/M= 1.51
Finland Canada
Denmark
German Dem. Rep.
F/M = 0.83 F/M = 1.3'.'
Norwav
Sweden
F/M= IJ2
Zone*
C
Japan
F/M = 0.83
Yugoslavia
F/M = 1.47
New Zealand Maoris
New York Staiet
F/Mtt
Nevada
Connecticut
Victoria (Australia)
F/M = 1.28
New Ze.iland§
(Europeans)
F/M = 1.47
D
S.A. Cape Colored
Natal Indians
Bulawayo Africans
California Negroes
Hawaiian
F/M= 1.92
S. A. Cape Bantu
Natal Africans
Texas Latins**
F/M = 2.84
Israel (Native
News)
F/M = 1.55
Texas (non-Latin)
California (whites)
S.A. Cape (whites)
F/M= 1.30
Hawaii (Caucasian)
F/M = 0.78
Queensland (whites)||
F/M- 1.22
E
Bombay
Puerto Rico
F/M" 1.81
Jamaica
Nigeria
F/M= 1.24
Colombia
F/M = 0.45
•Zonal latitudes are: A. above 60°; B. 45°-60°; C. 35°-45°; D. 20°-35°; E. ()°-20°.
"S.A. Cape Bantus = 4.71; Natal Africans = 3.00; Texas Latins = 0.80.
tExtluding New York City.
ttN.Z. Maoris = 4.38; N.Y. State = 1.13.
SMales. 6.2; females. 9.1.
[[Queensland is half in zone D and half in zone E: male;.. 14.3; females, 17.4.
Table 7. Global distribution of melanoma incidence; male data with
corresponding female to male ratios (F/M) (From Gordon and Silverstone,
1976.)
F-l-67
-------
Effects—Skin Cancer
U.S.A.
The main sources for incidence data of skin cancer in the U.S.
came from the National Cancer Survey performed in eight cities in 1947,
a survey in Iowa performed in 1950 and in New York state (excluding
New York city) in i960. The most extensive data is that of the Texas
Tumor Registry, obtained by McDonald, and covering the years 1944 to
1966, and the most recent survey is that of the Third National Cancer
Survey, performed in 1974, and a special survey of the U.S. National
Cancer Institute performed in 1977-78.
All these data show a marked increase in the incidence of skin
cancer in the past three decades and a pronounced north-south gradient.
Table 8 shows data obtained by -Hocnszel in the years 1947-60. Tables
9 and 10 show the data of McDonald for six regions in Texas (1964), and
Table H shows the data of the Third National Cancer Survey (Scotto et
al., 1974). Table 12 shows that in the U.S., skin cancers make up from
40 to 116% of all other cancers of all sites in Caucasians (Scotto
et al., 1974).
Thus it appears that in Iowa, for instance, the incidence of skin
cancers in males rose from 61.4/100,000 in 1950 to 174/100,000 in 1972.
This roughly threefold increase is also noticeable in all areas in
Texas, with the exception of El Paso, where the incidence rate actually
decreased, and Houston, where the increase apparently represented only
a doubling. The Texas data show that incidence rates increase with
descending latitude, but not in a stepwise fashion. In the most recent
five-year period (1962-1966) the rate for El Paso was 183/100,000, that
for San Antonio (1 further south) 147.35/100,000. In San Antonio,
F-l-68
-------
Table 8
Age-adjusted"-'- skin cancer incidence per 100,000 population by sex for selected characteristics: 8 U.S. cities"""'' (1947),
Iowa (1930), and New York state, exclusive of New York city (1958-60)
Characteristics
All skin cancers***
Region
North
South
Residence
Urban
Rural
Site
Face, head and neck
Upper extremities
Trunk
Lower extremities
Histologic type
Melanoma
Basal and basosquamous
eel 1 carcinoma
Squamous cell
carcinoma
Male
8 Cities**
48.2
33.4
143.3
48.2
39.4
3.7
2.8
1.4
1.7
21.8
12.1
Iowa
61.4
72.1
52.1
48.6
5.3
}3.4
•J
Not
com-
puted
New York
41.3
45.6
42. 3 +
30.3
3.1
3.0
1.0
2.3
26.8
7.2
Female
8 Cities"""'f Iowa
34.9
25.4
88.0
34.9
28.2
1.8
2.9
1.5
2.2
16.7
6.7
35.3
45.4
24.9
27.4
2.1
13 7
_}
Not
com-
puted
New York
27.6
31.1
25. 8+
20.5
1.4
2.3
1.4
2.6
19.4
2.9
Ratio: Male/female
8 Cities-'--"
1.38
1.31
1.63
1.38
1.40
2.1
0.97
0.93
0.77
1.31
1.81
Iowa
1.74
1.59
2.1
1.77
2.5
?0.92
(w • ^
"'
New York
1.50
1.47
1.64
1.48
2.2
1.30
0.71
0.80
1.38
2.5
(Ti
*Age-adjusted to total population of continental United States, 1950.
—Whites only
***Includes newly diagnosed cases of all types and sites, with or without microscopic confirmation.
+Residents of rural areas of nonmetropolitan counties in New York State
Sources: 8 cities - unpublished tabulations from 10-city cancer morbidity surveys. Iowa - supplemented by unpub-
lished tabulations. New York - unpublished tabulations furnished by Bureau of Cancer Control, New York State
Department of Health.
(from Haenszel, 1963)
M
H,
Ml
fl>
O
rt
(A
I
en
O
O
n>
H
-------
Effects—Skin Cancer
Table 9
AVEKACE ANNUAL ACE-AUJUSTF.D IXClDKXCt KATES
PEU 100.000'
SIX REGIONS IN TEXAS
1944-1955
MAl.CS
\V>UTE NONVVIIITt M'AMMl >UKNAM_t.l>
KUMDEH HATE NOV.SKK K-\T£ NUMHKIt RAT*.
Replon A}
(El Paso)
1944-48
19-19-53
19S-1-5S
1959-63
1944-60
• 1949-65
1952-65
Region D
(San Antonio)
1944-45
1949-53
1954-58
1959-63
1944-65
1949-66
1962-65
Region El
(Laredo)
1944 -4S
! 1949-53
I. ' 1954-58
. 1959 -G3
1944-6G
1949-G5
1952-66
Region F}
(Harlingen)
, 1944-48
1 1949-53
. 1954 -5S
1959-63
,. 1944-6G .
1949-65
1952-65
Region G
•(Corpus Cliristi)
1944-48
1949-53
19.54-58
1959-53
•1944-65
1949-55
1952-G6
Region It
(Houston)
1944-48
1949-53
1954-5S
1959-63
'. 1944-66
1949-65
1952-66
326 215.G
493 ' 254.3
494 19S.S
566 201.3
2207 205.8
15S1 203.5
579 183.0
SIS 5S.-2
765 71.2
872 70.9
1807 131.3
539S 97.8
48SO • 106.3
2137 147.4
•
16 ' 34.4 '
39 63.5
69 93.5
1GO 214.2
412 - 135.4
396 154.8
209 25G.G
•
31 16.1
224 93.5
729 274.8
796 277.9
2251 193.8
2230 '229.6
&41 284.6 .
•
276 105.5
615 180.0
11G3 • 273.4
1577 323.9
4900 264.5
4624 293.1
1916 371.0
. *
ftf~A 4rt 1
**6* 4U-i
503 52.2
896 73.3
1354 . 84.6
4032 71.5
3S1S 7G.9
1G5G 92.7
• 29
30
31
41
159
130
45
1 £ 26
5 4.7 35
-•> 20 51
5 3.8 67
14 2.7 247
13 3.1 221
3 2.4 91
3'
12
15
50
117
H'
^
R7
0 1
5
19
M&
86
' '95
2S2
277
' • . - 125
. *
0 0 18
2 5.6 ' 36
5 17.2 46
6 17.6 59
17 11.5 195
17 13.7 ISO
7 13.9 63
B 38 1
° ~
. 12 4.3 . 3
14 4.4 4
15 ^ is
59 4.0 29
51 4.0 25
14 2.7 »
30.7
22~6
26.0
25.6
2-1.7
24.5
6.9
10.4
12.S
13.1
13.1
14.4
' 1S.3
3.6
9.4
12.3
32.9
19.3
22.8
430
"»fc^.\^
2J3
98
36.9
34.4
26.2
31.2
42.8
18.1
29.7
30.3
36.2
30.2
32.7
34.2
1.S
4.8
6.7
14.7
9.9
11.1
20.5
1970 Standard population.
figured for Spanish surname* and non
oxists" ofKces not iiicluUed.
sotiwme
>uriu
-------
Effects—Skin Cancer
Table 10
GANGER OF THE SKIN EXCLUDING MELANOMA
AVERAGE ANNUAL AGE-ADJUSTED INCIDENCE KAILS
.
PEK 100.000*
SIX REGIONS IN TEXAS
1944- 19 5S
FFV.\:.F:S -
WHITE . NON-.-'illTE
SI'AMSM >CI
NUMD'EK RATE NUMBES RATE NUMi-.iiK
Region A}
(El Paso)
1944-48
. 1949 -53
1954-55
1959-63 '
1944-65
1949-66
' 1952-65
Region D
(San Antonio)
1944-45
1949-53
1954-55
1939-63
1944-65
1949-65
1952-65
Region Et
(Laredo)
1944-4S
1949-53
1954-5S
1959-63
1944-65
1949-65
1952-65
Region Ft
(Harlingen)'
1944 —IS
1949-53
1954-5S
•1959-63
. 19-14-66
1949-65
1952- 6G
Kegion C .
(Corpus Christ!)
1944-48
1949-53
1954-5S
1959-63
• 1914-65
1949-65
1962-66
• lli-pun II
(Houston)
1044-43
. 19-19-53
1954 -5S .
1939 -Go
1044-CG
1949 -.GO
19G2-Gr,
• 1970 St:ti»d»nt pi
IK»(cs figured for
173 105.6
240 119.8
326 120.2 '
380 122.3
1349 117.3
1176 119.2
36S 105.9
359 .35.6 2 1.7
516 41.4 4 3.6
595 41.1 10 7.6
1310 78.5 " 6 3.8
3SS2 59.8 25 4.3
3523 64.3 24 4.8
1621 90.2 8 4.9
8 23.1
• •* ft I C
lo 24.6 .
33 49.4
' 94' 124.2
213 71.8
205 81.1
112 131.2
13 7.9
124 55.5
399 146.1
411 132.9
1237 104.3
1224 120.8
453 139.3
125 ' 47.1 0 ' 0
370 105.6 2 . 3.-.
6G5 15-5-0 1 2-0
1007 .193.7 < 9-S
2035 1S3.4 S 4.3
28 11 H0.3 S 5.-
1176 20G.9 4 5.3
•
- 153 19.3 J2 '5-3
259 23.5 10 4.3
' 5G9 39.5 9 2.4
HUO 43.1 1* 3.1
S3S2 3G.O =*> 3.6
2229 3S.7 4o J^
932 -M— 'G
SS^rnanv.a ««d ,:o,WSp.nW, >«
25
31
35
44
174
149
f\
50
25
35
52
77 •
257
241
105
4
16
20
40
104
100 •
41
6
IS
70
S4
260
" 254
120
22
: 3S
.J9.
44
192
170
60
3
5
S
S
35
32
13
niamccl.
CNAMKP
KATE
'
20.7 - .
2G.4
f\ "1 O
22.5
m f\
2^.0 •
24.0 .
24.7
rj t "T
^.-t.l
7.4
10.0
11.4
15.5
13.2
14.4
15.7
4.1
10.S
145
22.7
15.3
17.5
21.7
3.1
8.5
29.9
29.2
23'i
27.5
37.7
24.5
33.1
3-1.9
' 24.6
30.1
31.2
3i3
12.4
7.3
S.5
9.9
11.4
1U
13.1
J|)cm.;iU.I<>psts- Olliccs •«« incmuw-
(From MacDonald anH Rnbendorf,
F-l-71
1964)
-------
Effects—Skin Cancer
Table 11
Annual Age-Adjusted Skin Cancer Incidence Rates* for Caucasians by Sex,
Cell Type, and Area, 1971-72
Cell type
Dallas-Fort Worth
M
A.
T
Iowa
Patients
M I F
Minneapolis-
St. Paul
T I M I F
| San Francisco-
| Oakland
! T I M I F
All patients | 379 j 539 | 259 | 124 174 | 83 | 151 1 201 | 115 | 184 ] 250 | 133
Basal cell | 286 | 394 | 205 | 93 j 123 j 69 j 129 j 165 | 103 j 153 j 198 | 117
Squamous cell | 83 | 124 j 51 j 28 | 47 | 13 | 21 j 35 j 12 j 28 j 45 | 15
Both basal and j I j I I I I I I I I !
squamous | 10 j 21 j 3 | 3 j 4 j 1 | 1 j 1 j 0 j 3 | 7 j 1
(From Scotto et al., 1974)
Table 12
Age-Adjusted Incidence Rates'" Among Caucasians for Nonmelanoma Skin
Cancers and Cancers of All Other Organs Combined by Area
Dallas-Ft. Worth
Iowa
Minneapolis-
St. Paul
San Francisco-
Oakland
All cancers excluding
nonmelanoma skin
cancers, 1969-70**
326
297
320
355
Nonmelanoma
skin cancers
379
124
151
184
Nonmelanoma skin
as percent of al
cancers
116
42
47
52
cancers
1 other
*Per 100,000 population standardized to the population of the U.S. for 1970
**Preliminary, unpublished data from the Third National Cancer Survey.
(From Scotto et al., 1974)
F-l-72
-------
Effects—Skin Cancer
•a large pare of Che permanent population consists of retired military
personnel and their families. Most of these people spent much less
time in the sunny south than the permanent population in El Paso. The
highest incidence rate was found at Corpus Christi at 28 N latitude
(371/100,000). This is significantly greater than the incidence at
Harlingen (284.5/100,000) at 26°N latitude. At Corpus Christi an ad-
ditional factor to exposure is operating: a great preponderance of
Celtic inhabitants which migrated to that area about a century ago.
Thus the disproportionately high incidence of skin cancer in Corpus
Christi is due to a combination of intense insolation, and a very sus-
ceptible population, a situation similar to that existing in Queensland,
Australia.
Europe
The data available for incidence and prevalence of skin cancer
in Europe are fragmentary and extremely difficult to interpret, since
surveys, such as they are, were performed by variously rigorous methods,
and over at least three decades, during which the incidence of skin
cancer has greatly increased. However, it is clear that just as in
the U.S. and Australia, a marked north-south gradient exists.
In 1967, Ball et al. reported skin cancer incidences as follows:
Sweden - males 7.4, females 4.5/100,000; Denmark - 26.7/18.0; German/ -
12.9/9.5; Yugoslavia - 15.8/23.4; England - 34.9/15.9. More recent
reports suggest incidences in the German Democratic Republic of 43.9
for males and 40.3 for females, in Bulgaria of 42.2 for rural and 25.0
for urban people, and a staggering incidence for precancerous skin
lesions of 2450/100,000 in the Kirghiz S.S.R.
F-l-73
-------
Effects—Skin Cancer
Africa, India, Japan
Information relating to skin cancer in Africa, gathered from var-
ious sources, mainly the reports of Oettle and J.N.P. Davies, show that
the native Africans have extremely low rates of non-melanoma as well
as melanoma skin cancer.
The incidence rates in the Johannesburg Bantus and in Uganda are
of the order of 1-2/100,000 for non-melanoma skin cancer (almost all
squamous cell carcinomas of the lower extremities). Isaacson (1979)
reported an incidence of 1/100,000 for Soweto (South Africa) blacks.
The exceptions are albino natives, who are extremely prone to the
development of skin cancer. In South Africa, albinism is quite common
among the Bantus. Oettle estimated a crude annual incidence rate of
579 for male and 408 for female albino Bantus for squamous cell car-
cinomas of the skin. Interestingly, the rates for basal cell carcinoma
were only 36/100,000 for both albino sexes combined (actually only one
case was found!).
India
Incidence data for skin cancer in India is not available. However,
it is clear that the majority of skin cancers seen in hospitals occur
in whites. Among native Indians, special types of squamous cell car-
cinomas of the skin are found, which are presumably due to extreme heat,
combustion and chronic friction such as Kangri, Dhoti and Chutta cancer
(Urbach, 1963).
Japan and Taiwan
While again incidence figures are not available, it is clear that
skin cancer is uncommon in Japan and Taiwan. This is also borne out
F-l-74
-------
Effects—Skin Cancer
by a reversal in Japan and Taiwan of the usual basal cell to squamous
cell carcinoma ratio seen in whites. Squamous cell carcinomas are two
to three times more common than basal cell cancers, and may arise at
the site of premalignant skin changes such as burns, chronic trauma,
or secondary to arsenic ingestion (Urbach, 1963).
Australia and New Zealand
Probably the best skin cancer surveys of recent years have been
carried out in Australia, particularly in Queensland, by Silverstone
and Searle (1970). In Table 13 are reported their values for incidence
of skin cancer in various parts of Australia.
Table 13
Skin Cancer in Australia
Annual Incidence per 100,000 population of sex stated
State
Victoria
Queensland
Brisbane
Townsville
Crude
Crude
68.5
265.1
242
466
Male
Age
Standardized
66.6
265.1
Crude
50.5
174
170
300
Female
Age
Standardized
38.5
155.8
How do these incidences in Queensland compare with those elsewhere?
One has to look for comparable demographic data in those areas of the
world which are warm and sunny and to which people from the British
Isles and northern Europe migrated.
F-l-75
-------
Effects—Skin Cancer
Examples of High Incidences of Skin Cancer
Skin Cancer - Annual Incidence per 100,000 by Sex
| Country
JS.W. England
| South Africa - Cape Whites
[Texas (Non-Latin)
1
[Queensland -''(Whites)
Male
28
133
168
265
Female
15
72
106
156
"'This means that in Queensland about 3,500 to
4-,000 people develop a skin cancer for the
first time each year.
When age specific incidences for these regions are calculated it
can be shown that skin cancer tends to appear at an earlier age in the
Queensland population than in those dwelling in other areas.
The percentages of the aged male population who develop skin cancer
per year are: southwest England 3.1; non-Latins, Texas 7.5; Cape Whites
10.2; Queensland 17.0. The corresponding prevalence rates for females
are: 1.5%, 5.6%, 5.2%, 10.5%. For long it was said that the Celtic
and Anglo-Saxon population of north Queensland was the only successful
northern European settlement of a permanent kind in the tropics, hence
one must expect to find a high incidence of skin cancers.
What is the effect of latitude within Australia? An approximation
can be made. If the male incidence in Hobart is unity, in Brisbane
it is four and in Townsville it is eight. These places correspond to
south latitudes of about 43 , 28 and 19 respectively. The proportion
of males in Queensland between 20 and 80 years who would be found with
F-l-76
-------
Effects—Skin Cancer
at least one skin cancer or history of same, varied between 5% or 6%
in the subtropical areas around Brisbane to 12% to 13% in the tropical
north of the state. As a generalization the proportion of females with
skin cancer in the various parts of Queensland was about 60% of the
proportion found in males.
If one plots the global distribution of annual UV solar radiation
as put forward by Schulze and Grafe (1969) against incidence of skin
cancer on a global basis a case can be made that a doubling of incidence
occurs for every 10 decrease in latitude. The Australian data would
approximate to this, particularly in Queensland where the difference
in incidence between Brisbane and Townsville is about two to one and
the difference in latitude about 9 . (A doubling effect with every
8 to 9 decrease in latitude would fit Australian estimates of inci-
dences more accurately.)
In New Zealand, not as much work has been done on skin cancers
as in Australia, however, good estimates of skin cancer incidence
have been reported by Eastcott. These are:
For basal cell carcinoma 113/100,000, for squamous cell carcinoma
38/100,000, and for malignant melanoma 5.5/100,000. These rates are
considerably lower than Australia, but New Zealand is much further from
the equator and thus receives less UV radiation.
D. 2.5 Relationship to Environmental Exposure and Changing Inci-
dence with Time
All available evidence suggests that virtually all human skin squa-
mous cell cancer but only 2/3 of basal cell cancer are related to total
accumulated lifetime dose of biologically effective solar ultraviolet
F-l-77
-------
Effects—Skin Cancer
radiation (Urbach, 1969; O'Beirn et al., 1968; Gellin et al., 1966;
Scotto et al., 1974; Vitaliano and Urbach, 1980). It is also clear
that pigmentation of the skin confers a marked protection and that there
is considerable variation in the sensitivity to UVR among "white" peo-
ple, and that much of this variation is genetic (Urbach, 1969). It
is highly associated with "ease of sunburning" and ability to tan
(Vitaliano and Urbach, 1980). However, given sufficient UVR exposure
for sufficient time, most white people can develop NMSC. The great
difference in susceptibility, which is by no means regularly distributed
geographically, is one of the major confounding factors for any dose-
response calculation in man.
All investigators who have attempted to model dose-response effects
for NMSC agree that there are also a number of latitude-related vari-
ables which modify the relation of actually received UVR dose to inci-
dence of NMSC. This conclusion is based in part on the indication of
non-linearity in incidence-latitude figures, and in part on the intui-
tive concept that latitude-related climatic and behavioral effects must
modify the UVR dose actually reaching the skin of a target population.
There is obviously a relationship between latitude and solar energy
intensity, and this gradient is exaggerated in the UV portion of the
spectrum. Although the shape of this relationship is relatively complex
and dependent on a number of variables (altitude, albedo, cloud cover,
aerosols, etc.) for a range of northern mid-latitude values (30 to
50 . N) , the form closely approximates a straight line for calculated
(Green, 1978) and integrated skin erythema action spectrum weighted
values (Scotto et al., 1980).
F-l-73
-------
Effects—Skin Cancer
The relative effects of the "lifestyle" factor are largely specula-
tive. MacDonald (1971) suggested that it contributes about 507» of the
gradient, and van der Leun and Daniels (1975) considered this not an
unreasonable estimate. To accept such an estimate in conjunction with
an exponential dose-incidence relationship, however, implies that these
"other factors" are also exponentially related to latitude. If the
assumed magnitude of these factors is speculative, an assumed exponen-
tial relationship is considerably more so.
If the input of potentially effective UVR is linearly related to
latitude, and if the factors relating this input to NMSC are nonlinearly
latitude dependent, it follows that the relationship of skin cancer
incidence to latitude must be nonlinear. Indeed, the epidemiologic
evidence available (Gordon, 1976; Scotto et al., 1980) shows this to
be the case.
In order to obviate the unknown "other factors," various modelers
have used a two step procedure: in the first step, the amount of
average stratospheric ozone is related to the biologically effective
UVR dose (physical amplification factor); in the second step the annual
UVR dose is related to estimates of skin cancer incidence (biological
amplification factor) (Green and Mo, 1975; Urbach et al., 1975; Scott
and Straf, 1977).
The advantage of this approach is that one can explicitly account
for dose differences which are not related to the amount of strato-
spheric ozone. No attempt was made to account for possible genetic
or behavioral differences between the various populations.
A fundamentally different approach is to evaluate the effect of
an ozone reduction indirectly by using a dose-response theory to
F-l-79
-------
Effects—Skin Cancer
describe the effect of an increment in the annual UVR dose (Rundel and
Nachtwey, 1978). This approach resembles the aforementioned "two step
approach." It also uses the same two steps. The difference is that
the step in which the annual UVR dose is related to the incidence is
not based on epidemiologic data, but on a dose-response model inferred
from animal experiments. The advantage of this approach is that one
deals with one population (hairless mice) and does not have the problem
of genetic, environmental and behavioral differences at different loca-
tions (Rundel and Nachtwey, 1978).
Finally and most recently, De Gruijl and van der Leun (1980) have
developed a dose-response model based on animal experiments. Using
this model they derived a mathematical formula which gives the relation-
ship between a fractional increment in the animal UVR dose and the cor-
responding fractional increment in skin cancer incidence.
For such model calculations, estimates of the "physical amplifica-
tion" factor are necessary. Extensive calculations have been made by
Green et al. (1974), Mo and Green (1974), Johnson et al. (1976), Scott
and Straf (1977). These calculations are based mostly on extensive
UVR measurements performed by Bener (1972), estimates of local average
0., concentrations, estimates for effect of cloud cover, etc. With minor
variations, amplification factor (optical) has been found to be approx-
imately 1.7 (Green and Mo, 1975) based on an action spectrum similar
to DNA.
There is still extensive discussion among modelers whether the
biological dose response relationship follows an exponential or a power
law relationship. Nevertheless, irrespective of the model used, the
F-l-80
-------
Effects—Skin
"biological amplification factor" ranges in the vicinity of 2x (Urbach
et al., 1975; Green et al., 1976; Green, 1978; Scotto et al., 1980;
de Gruijl and van der Leun, 1980). It should, however, be pointed out
that, particularly if the power law applies, the "biological amplifica-
tion factor" (AFg) will, be less away from the equator and more near
the equator (e.g. , Seattle, 47.5 N, AF_ 1.47; New Orleans, 30 N, AFn
B o
2.57) (Scotto et al., 1980).
Thus, the general estimate for the overall amplification factor
obtained from the most recent models remains at 4x, i.e. a 170 reduction
in ozone may result in a 4% increase in nonmelanoma skin cancer. Pre-
liminary results from two dimensional modeling suggest the nonuniform
°3
distribution ofAdepletion could result in a lower factor (Pyle and Der-
went, 1980).
One other point needs to be made before the above described dose-
response relationships are put to general use. That is that there is
a significant difference in risk for developing basal cell or squamous
cell cancer of the skin. It has been pointed out before that of basal
cell carcinomas, about 1/3 seem to have no relationship to UV exposure
(Urbach et al., 1972; Gellin et al., 1966). Also, unpublished data
of the most recent epidemiologic survey performed by NCI (Scotto, per-
sonal communication) show a much shallower north-south incidence slope
for basal cell carcinoma than squamous cell carcinoma. Finally, Vital-
iano and Urbach (1980) have found that the increased risk for the most
susceptible population is very much greater for development of squamous
cell carcinoma than basal cell carcinoma, and that, given the same level
of cumulative lifetime solar exposure, subjects over 60 years of age
had a higher risk for development of NMSC than those at younger ages.
F-l-81
-------
Effects—Skin Cancer
Clearly, the existing models need considerable refining before
realistic risk estimates relating reduction in ozone to increase in
NMSC can be made, and present models certainly overestimate the risk.
Finally, as far as changing incidence of NMSC with time is con-
cerned, there is a small amount of data that the incidence has been
increasing by 2-3% per year in the past decade (Scotto et al., 1980).
This has been happening although there has been no measurable decrease
in ozone in that time period (Berger, personal communication; London,
1980). It may be the reflection of a continuing increase in exposure
for sociologic reasons - i.e., "a tan is beautiful."
D.3 Experimental Skin Carcinogenesis
Effects of Long-Term Exposure of Skin to UVR
UV irradiation induces an inflammatory response and ulceration
in both the epidermis and the dermis, the latter being infiltrated with
leukocytes in the region of the lesions, and to a much lesser extent
between them. These lesions ulcerate and the epidermis may disappear
for a time in the center. However, peripherally there is particularly
active hyperplasia. The basal membrane (between the epidermis and the
dermis) may disappear for a time in the regions of these "open" lesions.
Between the lesions, the infiltration of leukocytes is relatively
slight.
Injury to the epidermis and dermis, brought about by long-term
i
exposure to UVR, leads to dermal alteration, fibrosis and elastosis,
as well as to epidermal atrophy. However, experimental production of
cutaneous elastotic changes in animals by artificial UV irradiation
has only been reported rarely. Using histochemical methods, Sams et
al. (1964) demonstrated focal dermal elastosis in mice after prolonged
exposure to artificial UVR. UVR-induced changes in connective tissue
were also seen in rat skin by Nakamura and Johnson (1968).
F-l-82
-------
Effects—Skin Cancer
D.3.1 Mechanism of UVR Carcinogenesis
In order to understand the possible mechanisms of UVR carcino-
genesis, it has been necessary to study the formation and excision of
pyrimidine dimers, unscheduled DNA synthesis, and all the data valid
for the range from bacteria to mammalian cells.
It is now generally believed that UVR carcinogenesis results from
a succession of events originating in a photolesion of the genetic
material.
From the numerous studies that have led to an elucidation of some
of these mechanisms, it emerges that faulty DNA repair can increase
the frequency of carcinogenesis in the following ways: by causing al-
terations in DNA, which also find expression in an increased frequency
of chromosome aberrations and a rise in mutation rate; by increasing
the rate of transformation of normal cells into cancer cells; and by
facilitating the expression of latent oncogenic viruses able to trigger
cancerous growth (Setlow, 1973; Trosko and Chu, 1973).
Errors induced by DNA repair during the initiation phase of car-
cinogenesis seem to be the most likely mechanism leading to UVR cancers.
D.3.2 Tumor Types
Epidermal tumors. The first visible step in UVR-induced epidermal
tumor formation in animal skin consists of cell proliferation, i.e.,
an increase in the number of squamous cells and cell layers, which
gradually become papillomatous in character (Stenback, 1978). This
is accompanied by an increase in cellular atypia, 'nuclear enlargement,
hyperchromatism, indentation, and prominence of nucleoli. This basical-
ly proliferative response is frequently replaced by a dysplastic
F-l-83
-------
Effects—Skin Cancer
pleomorphism, occasionally with pseudoepitheliomatous hyperplasia-like
features, which ultimately invade the dermis. The tumors first seen
are acanthomatous papillomas (trichoepitheliomas) , with a predominantly
epithelial component or fibropapillomas in which the tumor is composed
of a fibrous stroma covered by squamous epithelium.
The malignant tumors that ultimately develop are squamous cell
carcinomas of different types including: solid keratin containing
tumors; moderately differentiated, individually keratinizing tumors
with distinct intercellular bridges; and less-differentiated, non-
keratinizing spindle cell tumors, in which ultrastructural analysis
reveals squamous cell patterns.
Keratoacanthomas, i.e., proliferating epithelium on a cup-shaped
base, are relatively more frequent in animals of different species re-
ceiving large doses of UVR than in animals treated with chemical car-
cinogens.
Epidermal tumors are easily induced by different agents in mouse
skin (Stenback, 1969). Winkelman et al. (1960) reported the production
of squamous cell carcinomas on the backs of hairless mice exposed to
UVR. Further studies established that carcinomas could be induced in
this animal almost to the exclusion of sarcoma formation (Epstein and
Epstein, 1963). In early studies, epithelial tumors were reported in
both rats and mice (Findlay, 1930; Herlitz et al., 1930; Putschar and
Holz, 1930), but in later studies the deeper lying dermal tumor response
to UVR predominated. No mention of skin sarcomas was made, however,
in the studies of Beard et al. (1936) on albino rats in which 12 animals
exhibited 9 carcinomas of the external ear, 6 sarcomas of the eye and
1 carcinoma of the skin of the nose.
F-l-84
-------
Effects—Skin Cancer
The difference in the distribution of tumor types, with sarcomatous
growth predominating in haired mice and carcinomas in man (Urbach, 1969)
may, in part, be explained by the difference in penetration of UVR,
a greater amount reaching the dermis in mice (Kirby-Smith et al., 1942;
Everett et al., 1966).
Dermal tumors. Another type of neoplastic progression seen in
mice, particularly after intensive treatment with large doses of UVR
over a short time period, consists of ulceration, scarring and the sub-
sequent formation of dermal tumors. These tumors begin as aggregates
of regularly built, elongated cells with small monomorphic nuclei.
Epithelial proliferation is occasionally observed as a secondary phe-
nomenon. The tumors rarely extend grossly through the surface. In
the early stages, they appear to be papillomas, although they consist
entirely of fibroblastic cells. Some tumors are remarkably acellular,
with a prominent fibrillary pattern. The tumors are composed of large,
polymorphic cells with prominent nuclei. Ultrastructural analysis shows
the predominant cellular components - a dark cell type, with hyperchro-
matic nuclei and scanty cytoplasm, and a light cell type, with large
nuclei and abundant cytoplasmic ribosomes (Stenback, I975a). The same
cell types are also seen in malignant tumors, in which the cellular
polymorphism frequently is considerable, with nuclear atypism and en-
largement, numerous nucleoli and a generally disorganized arrangement.
Sarcoma induction is partly species-specific, as these tumors were not
seen in UV-irradiated Syrian golden hamsters (Stenback, 1975b) nor were
they seen in hairless mice (Epstein and Epstein, 1963) or guinea pigs,
susceptible to chemical sarcoma induction (Stenback, 1969, I975b).
F-l-85
-------
Effects—Skin Cancer
An infrequent neoplastic alteration in several animal species is
the vascular tumor (Stenback, I975b). This begins as a proliferation
of dilated vascular spaces with regular endothelial lining. Rarely,
the endothelium proliferates to the point of forming angiosarcomas,
or invasive tumors composed of large, atypical cells arranged in a
nodular pattern.
The role of the dermis in epidermal tumor formation has been empha-
sized by numerous investigators (Orr, 1938; Mackie and McGovern, 1958).
A proliferation of elastic tissue was induced experimentally in mice,
by Sams et al. (1964), through repeated exposure to UVR. Similarly,
Magnus and Johnson (1965) stimulated formation of elastotic tissue,
following early destruction of elastic fibers, with radiation of 300 nm
from a monochromatic source. Nakamura and Johnson (1968) reported that
dermal elastic tissue proliferation occurred in albino rats after
chronic irradiation with UVR, only after discontinuation of the expo-
sure. It was postulated by Johnson et al. (1968) that this change was
the result of photochemically-induced alterations in fibroblast function,
rather than the degradation of normal elastic fibers. In support of
this concept, Epstein et al. (1969) noted that unscheduled DNA synthesis
occurred in connective tissue cells of the upper dermis within minutes
of exposure to UVR shorter than 320 nm, demonstrating a direct effect
of UVR on dermal fibroblasts.
Because of its frequent association with skin cancer formation,
actinic elastosis has been considered to play an important role in tumor
development. However, Sams and his co-workers (1964) and Graham and
Helwig (1965) demonstrated that actinic elastosis was not essential
F-l-86
-------
Effects—Skin Cancer
for Che development of epidermal malignancies. Furthermore, the experi-
mental production of elastosis in animals has not been associated with
cancer formation, nor has UVR-induced experimental cancer depended on
the presence of this change (Epstein and Epstein, 1963).
Adnexal tumor formation is not as common in UVR-treated animals
(Stenback, 1975) as in, for example, carcinogen-treated rats (Zackheim,
1964; Stenback, 1969). Hyperplasia and cystic disorganization of hair
follicle walls is very common, but rarely progresses to grossly visible
neoplasia. Trichoepitheliomas with barely visible follicular arrange-
ments are rarely seen. Even more uncommon are hamartomatous tumors,
hair follicle-derived trichofolliculomas and sebaceous gland tumors.
Sebaceous gland epitheliomas and carcinomas are even rarer. In a study
in 1930, Putschar and Holtz reported only a very small number of basal
cell carcinomas in rats.
Species-Specificity
Three specific factors - pigment, hair and thickness of the stratum
corneum - have been found to alter susceptibility to tumor induction.
It was found that pigmented mice required significantly more radiation
to induce tumors than albino animals. Hair offered even greater protec-
tion (Blum et al., 1959), and thus the hairless rat appeared to be a
likely subject for tumor induction studies. However, the results of
Hueper's (1941) extensive studies indicated that this animal was, in
fact, quite resistant to UV penetration because of its thick stratum
corneum. Since pigment, hair and the stratum corneum were limiting
factors, the ears of albino mice and rats became the traditional test
sites for experimental production of cancer by UV irradiation. A re-
markable amount of quantitative data has been accumulated using this
F-l-87
-------
Effects—Skin Cancer
system. The usual tumor produced in this tissue was a sarcoma (Roffo,
1934; Grady et al., 1943b). Thus, the albino ear model could not be
used for evaluating qualitative changes associated with epidermal car-
cinogenesis, which is the primary process induced by UVR in human skin.
Winkelman and his co-workers (1960) reported the production of
squamous cell carcinomas on the backs of hairless mice exposed to UVR.
Further studies established that carcinomas could be induced in this
animal almost to the exclusion of sarcomas (Epstein and Epstein, 1963;
Epstein, 1965). In addition, UVR-induced pigmented tumors were reported
in pigmented hairless mice (Epstein et al., 1967). Thus, the hairless
mouse has provided a model for both the qualitative and quantitative
examination of the carcinogenicity of UVR.
Though penetration of UVR appears to be of obvious importance,
other factors also influence the type of growth induced by UVR. Grady
et al. (1943a) found that the size of individual doses did not have
any effect on the carcinoma/sarcoma ratio in the albino, hairy mouse
but that reduced intervals between exposures increased the number of
epidermal carcinomas. These findings suggest that various tissues re-
spond differently to the carcinogenic effects of UVR (Stenback, I975a).
In part, this may be associated with differences in penetration of var-
ious wavelengths of UVR. Furthermore, there are great species differ-
ences in the repair capability of cells.
D.3.3 Ultraviolet Radiation as an Initiating Agent
The two-stage concept for skin tumor formation proposed by Beren-
blum and Shubik (1947) supposed formation of dormant tumor cells by
a single application of a carcinogen. These latent tumor cells were
F-l-88
-------
Effects—Skin Cancer
provoked by the subsequent application of a promoter to form visible
tumors. In his studies on the induction of skin cancer by exposure
to UVR, Blum (1969) indicated that the process was continuous beginning
with the first exposure and progressing to ultimate tumor formation.
Blum's conclusions were based on experiments in which he (Blum et al.,
1943) and Rusch et al. (1941) could not produce tumors, unless exposures
were carried out over a minimum of 1\ months, regardless of the amount
of energy used. Blum's experiments suggested that, with shorter expo-
sure periods, tumor formation was not accelerated enough to become
visible within the lifetime of the experimental animal. Epstein and
Roth (1968), using a single exposure to UVR as an initiator and treat-
ment with croton oil as a promoter, concluded that croton oil stimulated
tumor formation, the characteristics of which were established by
initial exposure to UVR. The results of these studies were signifi-
cantly different from those encountered when a chemical carcinogen was
used as the initiator (Stenback, 1969).
D.3.4 Interactions between Ultraviolet Radiation and Chemicals
Chemically-enhanced photocarcinogenesis
An equally significant problem concerns photo-induced carcino-
genesis following the application to the skin of agents which are photo-
toxic, but not in themselves carcinogenic.
A portion of the sunlight spectrum is carcinogenic, even in the
absence of an exogenous photosensitizer. At the current rate of intro-
duction of new compounds into the environment, it has become increas-
ingly important to determine whether a readily demonstrable property,
such as phototoxicity, can be used to predict compounds or treatment
regimes that could enhance photocarcinogenesis.
F-l-89
-------
Effects—Skin Cancer
Concepts of chemical interaction with UVR-photocarcinogenesis are
of recent origin. Blum (1969) and Emmett (1973) reviewed a number of
reports dealing with the influence of phototoxic substances on photo-
carcinogenesis. The results frequently appeared to be in disagreement,
a situation possibly reflecting differences in technique, including
solvent, routes of administration, light sources, criteria for tumor
recognition, and in statistical evaluation (Blum, 1969). In addition,
characteristics of some compounds (toxicity, carcinogenicity, insta-
bility) rendered their interactions with light complex and their analy-
sis difficult.
The relative enhancing effects on photocarcinogenesis of two widely
recognized photoactive compounds, 8-methoxypsoralen (8-MOP) and anthra-
cene, were studied by Forbes et al. (1970). Both compounds were photo-
toxic, but only the 8-MOP solutions markedly enhanced photocarcino-
genesis. Thus, the ability of a chemical to induce phototoxicity is
not always sufficient to augment photocarcinogenesis.
Interactions Between Light and Chemical Carcinogens
The fact that UVR can alter several phenanthrene carcinogens photo-
chemically has been known for some time. The studies of Davies et al.
(I972a,b) showed that the carcinogenicity of 7,12-dimethylbenz(a)anthra-
cene (DMBA) was reduced by light according to the demonstrable photo-
chemical lability of the compound. There was also evidence that an
additional time-dependent factor could influence this effect. Thus,
it appears that, at least in the case of DMBA-treated animals, light
may contribute in two opposing ways: (a) by degradation of the carcino-
gen to noncarcinogenic products, and (b) by stimulating a phototoxic
response that appears to coincide with a relatively increased tumor
yield.
F-l-90
-------
Effects—Skin Cancer
Depending on the wavelengths of the UVR used, carcinogens can be
photodegraded to a less carcinogenic compound, or can induce photo-
toxicity which may augment carcinogenesis or cause such a severe local
phototoxic reaction that the epithelial skin cells are nearly all de-
stroyed. Thus, either enhancement or inhibition of skin carcinogenesis
may occur, depending upon the carcinogen and the wavelength of the light
source used.
UVR-induced carcinogen formation
The photochemical conversion of sterols to carcinogenic substances
has been proposed as a potential explanation for the cancer-causing
effects of light upon skin (Black and Douglas, 1973). It has recently
been demonstrated, in vitro, that one such compound, cholesterol-5a-
oxide, which possesses carcinogenic properties (Bishoff, 1969), is
formed in human skin exposed to UVR (Black and Lo, 1971).
D.3.5 Physical and Quantitative Aspects of Ultraviolet Irradiation
in Animal Studies
Carcinogenic action spectrum
Determination of the effective wavelengths or "action spectrum"
is one of the primary objectives in the study of photobiological re-
sponses. However, data are not available for the action spectrum of
UVR-induced cancer formation. The paucity of this information for one
of the most extensively studied photobiological reactions is due to
a number of factors, including the large number of potential wave-
lengths, the considerable number of animals necessary and the length
of time (a matter of many months or years) required for exposure to
each wavelength, the difficulties in immobilizing experimental animals,
F-l-91
-------
Effects—Skin Cancer
and the need for an especially good monochromator with practically no
stray light contamination. Though the complete curve of the carcino-
genic spectrum is not known, certain aspects have been determined by
less sophisticated methods. Roffo (1934) reported that window glass
filtration eliminated the carcinogenic effects of sunlight on white
rats. Thus the offending rays of the sun would be found approximately
between 290-320 nm. A number of investigators using mercury arc and
fluorescent sun lamps with filters have confirmed that, under their
experimental conditions, 320 nm represented the longer wavelength limit
for cancer formation (Griffin et al., 1955; Blum, 1969). Furthermore,
carcinogenic responses have been produced by radiation as short as 230.2
nm (Roffo, 1934), and skin cancer has long been known to be induced
by UVC and UVB. Thus the action spectrum appears to include wavelengths
between 230 and 320 nm, but wavelengths between 290 and 320 nm have
been shown to have significantly greater carcinogenic effects than UVR
shorter than 260 nm (Rusch et al., 1941; Blum, 1943; Blum and Lippincott
1943; Kelner and Taft, 1956).
Freeman (1978) performed a series of experiments to provide more
specific comparative data by testing the hypothesis that the action
spectrum for carcinogenesis parallelled that for erythema. In these
studies, squamous cell carcinomas developed at approximately the same
rate and frequency, when UVR exposure was proportional to that for ery-
thema, with a decreasing potency from 300 to 320 nm. No tumors occurred
in mice exposed to 290 nm. These cancer-producing wavelengths are also
responsible for the normal phototoxic sunburn reaction. Longer UV and
visible light are neither erythema-producing nor carcinogenic under
ordinary conditions.
F-l-92
-------
Effects—Skin Cancer
It can not be assumed that the action spectra for human skin ery-
thema and mouse skin photocarcinogenesis are similar, unless a common
chromophore or action mechanism is involved. Setlow (1974) proposed
that the common denominator was the action spectrum for affecting DNA.
Making some allowance for the skin transmission of UVR, he showed that
the shapes of action spectra for DNA, erythema and possibly skin cancer
production were similar and could be made to coincide.
D.3.6 Dose-Response Relationships
The second law of photochemistry (the law of reciprocity of Bunsen
and Roscoe) states that photochemical action depends only on the product
of the light intensity and the duration of exposure. This law, however,
holds only for primary photochemical action, and can not be applied
to secondary reactions. Since the biological endpoints that can be
observed, such as erythema, pigmentation, skin cancer production, etc.
are certainly indirect effects, and since we still know little about
the primary photochemical reactions that underlie them, it is not sur-
prising that "reciprocity" holds only for some of the effects studied.
In the first quantitative photocarcinogenesis experiments ever
performed, Blum (1969) found that, within relatively narrow limits (ap-
proximate factor of 5), differences in dose, intensity or interval be-
tween doses did not alter the shape or slope of tumor incidence curves,
but only their positions on the log-time axis. Blum, however, was care-
ful to point out that this was only true as long as the experimental
conditions remained the same until the time the tumors appeared.
With the accumulated data, he surmised that UVR-induced cancer
formation was a continuous process that began with the initial exposure
F-l-93
-------
Effects—Skin Cancer
and that the appearance of tumors within the lifetime of the animal
depended on sufficient acceleration of the growth process.
In the majority of studies on photocarcinogenesis, fixed doses
of UVR have been given at a fixed dose rate, and the interval between
doses altered, but in increments of at least 24 hours. Such experi-
ments, while very valuable, are far removed from the conditions found
in nature under which human skin is exposed. Man is exposed to a rela-
tively low UVR flux that varies with time of day, season and environ-
mental conditions, such as cloud cover, and also during the exposure
period.
Two recent animal experiments have shown that both varying the
UVR dose increment and varying the dose-rate while the daily dose remains
constant affect UVR-induced skin carcinogenesis.
In the first experiment, groups of hairless mice were exposed to
doses of UVR from a bank of "fluorescent sun" (FS) lamps known to pro-
duce skin cancer in these animals. Equal doses of UVR were delivered
in periods of 5 minutes, 50 minutes or 500 minutes. Thus, while the
doses (given 5 times weekly) were the same, the flux varied by a factor
of 10 or 100. Striking differences in both tumor development time and
tumor yield were noted. The animals given the total UVR dose in 5
minutes developed tumors later and in smaller numbers than those given
the same total dose in 50 or 500 minutes (Forbes, personal communica-
tion). Thus, protracting the UVR dose over longer time periods resulted
in a striking increase in the carcinogenic effects of the radiation.
In another experiment, mice were exposed to UVR doses per day dif-
fering by a factor of two. As Blum had found previously, the lower
F-l-94
-------
sttects—Skin Cancer |
daily dose resulted in the delayed onset of first tumors without sig-
nificantly changing the shape of the response curve (Forbes, 1978).
Physical factors influencing UVR carcinogenesis
Although the tumor-promoting properties of such physical factors
as freezing, scalding and wounding have been described for chemical
carcinogenesis systems, little information is available about the ef-
fects of these factors on UVR-induced cancer formation. Bain and Rush
o
(1943) reported that increasing the temperature to 35-38 C accelerated
the tumor growth rate. The stimulating effects of heat on UVR carcino-
genesis were confirmed by Freeman and Knox (I964b). Heat also enhanced
the acute injury response to UVR.
Temperature does not affect the photochemical reactions that follow
UV irradiation, but it does affect many of the biochemical reactions
that follow the initial photochemical change (Blum, 1941, 1969). Al-
though it is known that heat adversely affects photosensitivity (Lipson
and Baldes, 1960) and other phenomena of light injury (Bovie and Klein,
1919; Hill and Eidenow, 1923), and that heat alters the effects of x-
ray (Carlson and Jackson, 1959), the influence of heat on burns produced
by sunlight or UVR has rarely been considered (Freeman and Knox, I964b).
Other studies have shown that high winds and high humidity signifi-
cantly increase tumor incidence (Zilov, 1971; Owens et al., 1977).
D. 3.7 The Immune Response to Tumor Induction
A number of studies have shown that the immune status of the host
and tumor induction are potentially interactive processes. Chemical
carcinogens cause alterations of the host immune-response, the type
and extent of which depend on the tumor-inducing agent (Curtis, 1975).
F-l-95
-------
Effects—Skin Cancer
UVR also profoundly affects immunological reactivity, particularly the
immune response to skin tumors induced by UVR. Studies leading to this
conclusion were prompted by an observation by Kripke (1976) that tumors
induced by UVR in CoHf mice were highly antigenic and are usually im-
munologically rejected when transplanted to normal, nonirradiated syn-
genic recipients. This raised the question as to why these tumors were
able to grow progressively in their primary host without succumbing
to immunological rejection. In an extensive series of experiments,
Kripke and Fisher (1976) found that pretreatment of mice with UVR for
periods of time too short to induce skin tumors made them unable to
reject transplants of UVR-induced tumors, even though such transplants
were immunologically rejected by unexposed animals. This indicates
that UVR—exposed mice are systemically altered in a way that prevents
immunological rejection of highly antigenic UVR-induced tumors.
Similarly, inability of unexposed secondary hosts to reject UVR-
induced tumors after transfer of. lymphoid cells from UVR-treated mice
has been established and demonstrates the immunological nature of the
systemic alteration in the UVR-treated mice (Fisher and Kripke, 1977).
Furthermore, the failure of lymphoid cells from UVR-exposed mice to
react against UVR-induced tumors is due to the presence of suppressor
T lymphocytes in the lymphoid organs of UVR-treated animals. In spite
of their inability to reject highly antigenic UVR-induced tumors, UVR-
exposed mice respond normally to most other antigens (Kripke, et al.,
1977; Norbury et al., 1977). The one exception is that UVR-treated
mice have a transient defect in antigen processing in the skin, which
is reflected in their inability to develop contact hypersensitivity
reactions (Jessup et al., 1978).
F-l-96
-------
Effects—Skin Cancer I
The finding that a selective immunological defect precedes the
appearance of UVR-induced primary tumors suggests that the immune system
might control early UVR-induced skin cancers and that tumors ultimately
appear because of this interference by UVR with host defense mechanisms.
The carcinogenic action of polycyclic hydrocarbons has been as-
sociated with their immunosuppressive action (Stenback, 1969). Immuno-
deficiency states and immunosuppression therapy are both associated
with an increased tumor incidence. Immunosuppressive agents, such as
antilymphocyte serum, enhance both chemically- and UVR-induced tumor
formation (Nathanson et al., 1973, 1976).
F-l-97
-------
Effects—Skin Cancer
E. The Risk of Increase of Malignant Melanoma of the Skin
Due to Increased Earth-Level Solar UV Radiation
E.I. Conclusions
There appear to be significant differences in the age/sex ratios,
incidence, geographic distribution and anatomic location between NMSC
and MM.
NMSC is a disease increasing sharply with advancing age, beginning
in the 50+ year olds. It is significantly more frequent in males (par-
ticularly squamous cell cancer), affects primarily the head and neck,
and to a much lesser extent the shoulders, upper chest and forearms
and hands. It is very uncommon below the midchest. There is a dis-
tinct, striking latitude gradient both within and between countries,
following the intensity pattern of solar UVR. The disease is uncommon
in heavily pigmented persons, even in areas of intense insolation', and
usually occurs in these at sites of chronic trauma. There is almost
always evidence of severe solar elastosis (sunlight damage to the con-
nective tissue) in the skin in which the tumors arise. Lesions clini-
cally, histologically and biologically almost identical with squamous
cell carcinoma occur spontaneously in animals on exposed skin in heavily
insolated areas, and can be reproduced with excellent dose-response
characteristics in experimental animals exposed to appropriate wave-
lengths of UVR.
In contrast, MM is a relatively uncommon tumor, primarily occurring
in man. The sex ratio varies from 1:1 to 1:1.2 in favor of females.
In contrast to NMSC, the anatomic distribution of MM does not follow
the most UVR exposed sites. About 10% of MM are on the head and neck.
This type of lesion most often represents LMM, and has basically the
same characteristics as squamous cell carcinoma: location on most ex-
posed sites of head and neck, low incidence before age 50, with a rapid
F-l-98
-------
Effects—Skin Cancer
and progressive rise in frequency with advancing age, almost uniform
presence of solar elastosis in the adjacent skin and low aggressiveness.
While NMSC is extremely uncommon in pigmented races, MM occurs
about one-fifth as frequently in such people as in white-skinned pa-
tients, and there is mostly found (75-85%) on the foot and lower leg.
Another 6 to 11% of MM appear to be of genetic origin, as evidence
by familial and multiple lesions, and peculiar precursors found early
in life (B-K mole).
The remaining approximately 75 to 80% of MM in white persons have
interesting attributes: The incidence rises sharply from adolescence
to early adult life (particularly on the legs of women), plateaus
through middle age, and rises again in old age (because of appearance
of LMM). The male/female incidence shows a preponderance of young fe-
males (less than 40 years old) and the sites of greatest incidence dif-
fer, being trunk in males and lower leg in females. Of interest is
the observation that, although the various populations studied live
in such disparate areas as Finland (north of 60 latitude) and Queens-
land, Australia (25 -15 south latitude) and thus are exposed to hugely
different amounts of solar UVR, the relative proportion of MM affecting
various body sites has remained quite stable until recently. In at
least two areas (Norway and Hawaii) the differences in incidence of
MM between males and females on the most affected sites (back in men
and legs in women) seem to be disappearing in the past decade only.
The latitude gradients for incidence (and mortality) of MM exist
within certain countries, but not in others. Thus there are real lati-
tude gradients for MM in Norway, Sweden, Great Britain and the U.S.A.,
F-l-99
-------
Effects—Skin Cancer
but less striking, or even reversed gradients in Western Australia,
across central Europe and a partial latitude gradient in eastern Aus-
tralia, where in Queensland the incidence of MM is less in the tropics
than in the subtropical areas.
In contrast to NMSC, the populations most affected are not the
outdoor workers, but rather the white-collar, more educated, more af-
fluent people, and the demonstrable concentration of MM in large cities
cancels out a latitude gradient in such places as Finland and Western
Australia, where this has been investigated.
The worldwide rapid increase in the incidence of MM (and, inter-
estingly, much slower increase in mortality rates, as if the MM were
becoming less aggressive) has been attributed to changes in lifestyle,
with greater exposure to solar UVR during leisure activities and vaca-
tions. Since the more affluent can be considered to be more likely
to partake in such activities, the reasoning goes that men removing
their shirts outdoors, and women wearing shorter skirts, account for
the peculiar anatomic distribution of MM.
The lack of evidence for chronic solar damage of skin in which
MM appear, the young age of a majority of patients, the variation in
latitude gradients, the peculiar anatomic distribution not matching
most exposed skin areas and the preponderance in city dwellers suggest
very strongly a significant difference in pathogenesis of NMSC and MM,
at least as far as the significance of solar UVR is concerned.
Except for LMM, other MM are certainly not related to chronic,
repeated solar UVR damage resulting from accumulated dose.
Whether the suggestions that acute, intermittent exposure to solar
UVR, or some dependency on intensity of irradiation, or some interaction
F-l-100
-------
Effects—Skin Cancer
of UVR with chemical or precursor lesions are the basis of MM etiology
needs yet to be determined.
The absence of a good animal model for MM, at least for the rela-
tion of UVR to MM, makes such studies difficult.
F-l-101
-------
Effects—Skin Cancer
E.2. The Natural History of Malignant Melanoma Skin Cancer
The first case of malignant melanoma ever recorded seems to have
been treated by John Hunter in 1787, although he never described the
disease as such. Laennec first reported what we now know as MM in 1812,
and coined the term "melanoses" (melas - Greek = black).
The classic and prophetic reports in the English language are those
of Norris (1820, 1857). His description of eight cases is fascinating
(Norris, 1857): "The disease often occurs in those persons who have
moles on various parts of the body, most resided in very smoky coal
and iron districts, there is a strong tendency to hereditary predispo-
sition and it is a disease allied to cancer."
He noted that his patients had light colored hair and a fair com-
plexion and that trauma may accelerate the growth of a melanoma, that
while often black, the lesions could be almost lacking in pigment and
have blue, pink or grey areas, and may disseminate widely throughout
the body. In other words, Norris, 150 years ago, had made the most
salient observations regarding this uncommon, but serious skin cancer
(Davis, 1980).
There are significant differences between many aspects of natural
history of nonmelanoma and melanoma skin cancer. For NMSC, three kinds
of evidence - latitude dependence, body location and relation to outdoor
sunlight dependence- all combine to point closely to exposure to solar
UVR as the prime cause.
F-l-102
-------
Effects—Skin Cance;
The situation is different for malignant melanoma, as will be seen
in the following. The latitude dependence, only recently pointed to
as the prime indicator of relationship to solar UVR, is becoming less
striking, indeed in some areas invalid, as better epidemiologic studies
become available. The body location does not fit areas of maximal solar
exposure, and differs greatly from NMSC. In contrast to NMSC, malignant
melanoma appears to have a predilection for middle class, primarily
indoor workers and their wives.
There is general consensus of those having studied the natural
history of malignant melanoma that, if there is a relationship to solar
UVR exposure, the mechanisms involved must be very different from those
operating in NMSC, and that the cumulative effect of chronic, repeated
UVR exposure can not be held responsible for malignant melanoma induc-
tion.
The present lack of an animal model for malignant melanoma and
the uncertainties regarding dose-response characteristics prevent the
conversion of projections of change in UVR intensity to estimates of
extra deaths.
F-l-103
-------
Effects—Skin Cancer
E.2.1 Anatomic Distribution and Histopathology
E.2.1.1 Lentigo Maligna Melanoma
(Hutchinson's melanotic freckle, circumscribed precancerous
melanosis of Dubreuilh)
This kind of malignant melanoma begins as a small, irregular
freckle-like lesion, almost always located on the head and neck, and
then evolves over many years in a distinctive way to become an invasive
malignant melanoma distinguishable from other forms of malignant mela-
noma. The early, noninvasive stages are referred to as "Lentigo Ma-
ligna" and the late, invasive stages as L-M melanoma. This has a better
prognosis than other forms of melanoma (Clark et al., 1969).
L-M melanoma occurs primarily on the exposed surfaces of the body
of elderly patients with white skin,most of whom have fair complexions.
The early, noninvasive lesion is an irregular, tan, freckle-like pro-
cess, from 0.3 to as much as 4 cm in diameter, most commonly found on
the temples or cheeks, particularly of females. Histologically there
is a distinct increase in the number of epidermal melanocytes, which
vary from normal to distinctly atypical, to frankly bizarre in struc-
ture. When invasive L-M melanoma develops, usually after 10 to 30 years
of a slowly growing Lentigo, the appearance changes. The outline of
the lesion becomes irregular and develops a kaleidoscope of colors,
including tan, brown, black, elevated blue-black nodules, and sometimes
translucent whitish-blue areas. These latter represent areas of spon-
taneous regression in parts of the lesion, which however is not a sign
of cure; the lesion continues to develop in adjacent areas. Histo-
logically, the L-M melanoma is identical to lentigo maligna except
for invasion of the dermis by malignant melanocytes.
F-l-104
-------
Effects—Skin Cancer I
There is also almost uniformly histologic evidence of severe "solar
elastosis," dermal connective tissue changes typical of chronic, severe
sunlight exposure. Patients with lentigo maligna and L-M melanoma often
have a history of solar keratoses, basal cell and squamous cell cancers
of exposed areas (McGovern, Little, Freeman, Robertson, personal com-
munications; Larsen and Crude 1979).
The distribution (78% on head, neck and upper trunk - Larsen and
Crude, 1979), the almost uniform presence of severe solar elastosis
of 'the adjacent skin, the slow progression and good prognosis (Wayte
and Helwig, 1968) clearly separate this form of malignant melanoma from
all .others.
The anatomic location, incidence in the older (60+) age group,
associated solar dermal damage and slow development strongly suggest
an etiologic relationship to chronic solar UV exposure. In terms of
frequency, approximately 10 to 1570 of all malignant melanomas are of
the L-M melanoma type.
E.2.1.2 Superficial Spreading Melanoma
Superficial spreading melanoma (SSM) is the most common of the
four clinical types of malignant melanoma. It constitues from 35 to
45% of all such tumors (Larsen and Crude, 1979).
SSM is a complex lesion that presents two quite distinct patterns
of growth, which have been described as the radial and the vertical
growth phase. The radial growth phase is a period of gradual peripheral
enlargement ("spreading") of a relatively flat pigmented lesion.
Sooner or later, usually after a period of one to five years, a
nodule may develop in SSM and this constitutes the onset of the vertical
F-l-105
-------
Effects—Skin Cancer
growth phase (VGP). This appears to be caused by growth of a morpho-
logically distinct new type of cell, with a capability for deeper inva-
sion and metastasis.
Examination of the cut surface of the lesion with the dissecting
microscope usually shows pigment filling the papillary dermis (much
of the pigment will be in melanophages deep to the tumor). The nodule
(or VGP), if present, consists of a pigmented or, more frequently,
amelanotic, soft fleshy tumor that may extend into the reticular dermis.
Wherever one judges invasion to be the deepest, levels of invasion and
thickness are determined.
The hallmark of SSM is the large epithelioid cell. Proliferation
occurs initially in the basal keratinocytic layers, and although dis-
positions of individual cells are frequent at the lesional margins,
nests of melanocytes are invariably present elsewhere within the lesion.
These nests differ from those of nevi in the following ways:
1. Variation in size and shape.
2. Tendency to fuse with adjoining nests and to spread up the
sides of rete ridges (McGovern, 1976).
3. Indistinct margins with an ill-defined separation between
melanoma cells and keratinocytes (Price et al., 1976).
A. Separation of nests or single cells from the basal lamina with
upward migration to produce a pattern of large, relatively
pale cells in the epidermis which resembles that of Paget's
disease (McGovern, 1970).
E.2.1.3 Nodular Malignant Melanoma
Nodular malignant melanoma (NMM) is, by definition, pure vertical
growth phase disease. More specifically, nodular melanoma is invasive
melanoma without the preceding and adjacent intraepidermal component.
F-l-106
-------
Effects—Skin Cancer
It is a tumor that arises by direct tumor progression without a discern-
ible evolution through demonstrable lesions analogous with radial growth
phase disease.
A nodular melanoma presents as a mass that expands only slightly
in radial axes as it becomes progressively elevated. Growth is usually
rapid and the history commonly short (under 12 months). Although a
history of preexisting nevus may be obtained, the spreading variegated
lesion of the radial growth phase distinctive for most melanomas is
not present, and the lesion is relatively symmetrical with discrete
borders. Even in the so-called amelanotic melanoma, a few flecks of
pigment are usual, and the total absence of melanin (to careful hand-
lens inspection) suggests another diagnosis, especially Spitz tumor
or pyogenic granuloma. Ulceration is a frequent but by no means invari-
able finding. The epidermal fine surface marking pattern is usually
obliterated as a result of stretching of the epithelium by growth of
the tumor in the papillary dermis or because of epidermal invasion.
Ulceration may occur with spontaneous hemorrhage in the form of a bloody
ooze or with crusting.
As in SSM, epithelioid cells are usually predominant. They are
usually pigmented but sometimes the majority are amelanotic. A few
pigmented cells are almost invariably present, and they contain dusty
fine melanin granules that represent abnormal melanosomes. The tumor
cells commonly invade the overlying epidermis in a pagetoid fashion,
and if these cells extend a short distance lateral to the deep dermal
component, SSM may be simulated. The invasive cells may extend through-
out the full epidermal thickness to produce frank ulceration. Neither
F-l-107
-------
Effects—Skin Cancer
epidermal invasion nor ulceration is an invariable finding, however;
a "grenz" zone of uninvolved dermis may separate epithelium and dermal
component, and in some such cases, it may not be possible to demonstrate
an origin of the tumor in the overlying epidermis.
The dermal component of a nodular melanoma is almost invariably
level III or IV. The cells are arranged in large nests within which
there is no stromal response. Between the nests, fibrosis and new ves-
sel formation are usual, and there is a scattered lymphohistiocytic
infiltrate with melanophages and, quite frequently, plasma cells.
E.2.1.4 Classification of Malignant Melanoma
The better prognosis of superficially invasive melanomas has been
appreciated by a number of writers (Mehnert and Heard, 1965; Lund and
Ihnen, 1955; Clark et al., 1969; Nicolle et al., 1960). Clark in 1967
proposed a classification of melanomas in terms of the level of invasion
in relation to the interface between the papillary and reticular dermis
(Clark, 1967; Clark et al., 1969). This classification was accepted
by an international group of pathologists in 1973 (McGovern et al.,
1973) and remains in wide use today. The basis of classification into
five levels of invasion is generally understood, although problems of
interpretation arise from time to time. The levels are as follows:
Level I: Confined to the epidermis (in situ)
Level II: Invasive into the papillary zone of the dermis
Level III: Tumor fills and expands the papillary dermis
Level IV: Invasion of the reticular dermis
Level V: Invasion of the subcutaneous fat
The classification above is useful, in that the prognosis for sur-
vival is excellent in level I and II melanoma, but becomes materially
F-l-108
-------
Effects—Skin Cancer
worse when the tumor reaches the reticular dermis (level III), probably
because of emergence of a new type of tumor cell with altered proper-
ties constituting the vertical growth phase of the disease (Wanebo et
al., 1975; Foulds, 1979).
The major histologic features of malignant melanoma are shown in
the table below.
Table 14
Histologic features of Malignant Melanoma
Pagetoid spread
Dermal invasion
Lack of maturation in dermal component
Dusty pigment
Pigment in deep dermal lesional cells
Inflammatory response and regression
Mitotic activity
Cytologic atypia
Variable cytologic patterns within the lesion
Ulceration
Epitheloid or spindle cells; large nucleoli
Nuclear hyperchromatism and pleomorphis
Irregular nests: Size variation
Poor definition
Fusion
Dyshesion
P-l-109
-------
Effects—Skin Cancer
E.2.1.5 Acral Lentiginous Melanoma
Melanomas of the hands and feet are an uncommon variety, accounting
for about 5% of malignant melanomas in white skinned people, but 50
to 75% of all malignant melanomas in genetically deeply pigmented per-
sons (see below, Section E.2.3).
In contrast to the more common varieties, acral lentiginous mela-
noma shows distinctive clinical and histological changes.
The typical clinical appearance is of a black, rounded nodule.
Many of the tumors are exophytic , and ulceration is common. Most les-
ions are accompanied by a peripheral macular hyperpigmented halo, which
may extend for several centimeters about the nodule or ulcer.
Histologically, the macular portion shows the pattern of radial
growth phase, with a diffuse lentiginous hyperplasia of atypical melano-
cytes in the basal portion of the epidermis, which itself is hyper-
plastic, with elongated rete ridges. The dysplastic melanocytes are
separated from neighboring cells by clear areas, the papillary dermis
shows a dense, lymphohistiocytic infiltrate.
In the nodules, vertical growth of pigmented, spindle celled mela-
nocytes infiltrates the dermis. When found, most lesions are deeply
invasive (Clark level 3).
Acral lentiginous melanoma is more common in males (M/F ratio 2:1)
and the duration of the lesions prior to diagnosis varies from months
to a few years. Most common age group is 50-60 years.
In a recent series (Coleman et al., 1980), 50% of patients, had
melanoma metastatic to the inguinal nodes, and a poor life expectancy.
P-l-110
-------
Effects—Skin Cancer
E.2.2 Age, Sex and Site Distribution
In recent reports of several large series of cases of MM, certain
observations occur uniformly. These are: a consistent, world-wide
increase in MM, primarily in lesions involving the back of men and the
lower legs of women (Magnus, 1977; Holman et al., 1980); a progressive
decrease of age of onset (Malec and Eklund, 1978; Lee et al.', 1979),
again world-wide, in the same anatomic locations; and great difference
in incidence and localization of MM between white and nonwhite popula-
t ions.
As will be shown, it is important to examine not only age and sex
incidence ratios, but to compare such factors in relation to location
of the MM lesions. There is uniform agreement that, with the exception
of LMM, MM is a type of tumor which attacks younger age groups. There
is, among white populations, a greater incidence in females than in
males (particularly England and Wales). The male to female incidence
ratios vary from 1:1 to 1:1.2 in several studies (McDonald, 1948; Lee
et al., 1971; Beardmore, 1972; Shaw et al., 1977; Sober et al., 1979;
Jensen and Bolander, 1980).
The differences between males and females are markedly affected
by age group; for instance, E. McDonald (1948) found a preponderance
of females in the under 40 year old age group.
| Age | Male | Female
I I I
| 20 years | 2.9% | 6.270
I I I
20-40 | 12.2% | 22.6%
I I
40-60 | 44.8% | 32.2%
I I
60 + | 40.1% | 39.0%
(From MacDonald, 1948;
Connecticut, 1935-1946)
F-l-111
-------
Effects—Skin Cancer
Similarly, Beardmore (1972) finds in Queensland, Australia between
1963 and 1969 45% males and 5.5% females with MM. The age breakdown
is given below for the incidence of MM in tropical and subtropical areas
of Queensland:
Tropical | Subtropical
Age | Male | Female
Male | Female
10 -
20 -
30 -
40 -
50 -
60 +
0.9 | 2.3
8.4
14.7
28.5
19.2
11.7
22.0
1.2 | 2.3
7.8
17.4
21.9 | 23.5
21.8
25.6
28.5 | 24.2 | 29.5
14.3
25.7
30.0
25.2
29.2
Incidence/100,000 adjusted to Australia
population.
Notable again is the preponderance of young females (40 years old or
less), and the absence in increase in MM as solar intensity increases
(i.e., tropical vs. subtropical areas).
The age difference becomes even more striking when one examines
the various sites of appearance of the different types of MM. For in-
stance, Bakes and McMillan (1973) report: LMM, 11%, age group affected
70 to 79; SSM, 50%, age group affected, 40 to 49; NMM 39%, age group
affected, 30 to 70. Sober et al., in 1130 cases of MM, found 69.6%
SSM, 15.7% NMM and 4.7% LMM, with mean ages of 47.3, 50.4 and 69.4 re-
spectively. The most extensive analysis of site, sex and age relation-
ships of MM has been carried out by Beardmore (1972). It is instructive
to compare the sex and site analysis in two widely separated areas of
o o
the world - Queensland, Australia (latitude 28 to 17 south) (Beardmore
1972) and Finland (latitude 60° to 70° north) (Teppo et al., 1978).
The data are given below:
F-l-112
-------
Effects—Skin Cancer
Some of the above observations can be explained on the basis of
the different biologic behavior of the different types of MM. LMM is
a very slowly progressing neoplasm, developing over a period of decades.
It is the most common melanoma found on the face and neck area, and
clearly related to chronic solar UVR exposure, as evidence by the uni-
form presence of dermal solar elastosis, and frequent association with
NMSC (Teppo et al., 1978; Larsen and Crude, 1978). This explains the
low incidence in younger patients and the inexorable continuing rise
in incidence in older age. The frequent location on the ears of males
(Beardmore, 1972) is also due to the exposure of this area in males,
given the hairstyles existing until recently.
The most recent and fascinating study has been prepared by Magnus
(Cancer, in press), reviewing the very extensive MM material of the
Cancer Registry of Norway. Between 1955 and 1977, over 5000 new cases
of MM were registered, allowing significant analyses according to co-
hort, sex, age and primary site. Magnus finds: a continuous, exponen-
tial rise in the incidence of MM over three decades, at the rate of
about 77o per year; a change in incidence over time by anatomic sites
in the most recent decade and a striking cohort effect.
Significant changes took place between the 1955 to 1970 and 1970
to 1977 periods. There was a reduction in the annual percentage in-
crease for MM of the trunk in males coupled with an increase in this
body site in females, and the reverse was found for MM of the lower
limb. In other words, the incidence trends for these two sites have
now become almost identical, in contrast to the situation a decade ago.
Magnus1 cohort analysis strikingly shows a marked cohort variation
on the trunk and lower limb, with the incidence being higher in the
younger generation.
F-l-114
-------
Effects—Skin Canceq
Also, the incidence in the face/neck area is markedly higher in
persons born 1890-1909, and lower in those born in 1930-1939 than the
incidence on trunk in males and lower leg in females.
Although Magnus concludes that the most likely cause for these
observations is a change in lifestyle, leading to greater exposure to
solar radiation in the younger cohorts, he warns that there can hardly
be a simple relationships between the cumulative dose of sunlight and
MM. He points out that the predominant incidence of NMSC on the face
of old people suggests that total dose rather than intensity of solar
radiation is essential. However, since it can reasonably be assumed
that sun exposure to the face/neck area is reasonably stable to an ad-
vanced age, and since the shape of the incidence curves for trunk/lower
limb deviates from that of the face/neck, this must imply that the car-
cinogenic exposure on these sites can not be stable, but increases
during adolescence and early life. Thus in MM, assuming UVR is causal,
it is not solely a question of accumulated dose, particularly in view
of the almost total absence of evidence of solar elastosis in the ad-
jacent dermis, and the absence of association with NMSC (Teppo, 1978).
Thus, the pathogenesis of NMSC and MM skin cancer may be quite different
as far as the role of solar radiation is concerned, and the latent per-
iod for MM must be significantly shorter than that of NMSC (Lee, 1972;
Elwood and Lee, 1975; Fears et al., 1976).
The striking predilection of SSM to affet the legs of women and
backs of men is not so easily explained. Since both, and particularly
the former, appear quite early in life, it must be assumed that the
precancerous period is short, of the order of a few years. Clearly,
F-l-115
-------
Effects—Skin Cancer
this type of lesion is not reasonably due to chronic repeated solar
UVR exposure, particularly in view of the almost total absence of evi-
dence of solar elastosis in the adjacent dermis, and the absence of
association with NMSC (Teppo et al., 1978). The nature of the carcino-
genic process causing this most common type of MM is not obvious.
F-l-116
-------
Effects—Skin Cancer
E.2.3 Race, Ethnic Extraction and Heredity
The susceptibility of Caucasians with fair, poorly tanning, easily
sunburned skin to develop nonmelanoma skin cancer has been documented
frequently (Molesworth, 1927; Urbach et al., 1976). That a similar
effect may exist in malignant melanoma has been reported in various
ways. Gellin et al. (1969) in a case-control study showed that, while
there was no significant association with complexion or ability to tan,
there were significant correlations of malignant melanoma with light
eye color, light hair and estimated outdoor exposure. However, Gellin
et al. found that patients with MM were much less likely to have another
skin cancer than patients with BCC, where development of another skin
cancer was very frequent. Bart and Scholl (1973) also found light eye
color in malignant melanoma patients.
Lane-Brown et al. (1971, 1973) found a disproportionate representa-
tion of persons with a Celtic genetic heritage among malignant melanoma
patients in Australia, with 50% of malignant melanoma patients being
at least one-half Celtic extraction (as compared to 23 to 28% of non-
cancer controls). This was as high as the Celticity found in NMSC in
the same study.
Among non-Caucasians, malignant melanoma is much less frequent
than among white-skinned people, Scotto et al. (1974) reported an inci-
dence in the U.S. of 4.47/100,000 in whites, and 0.8/100,000 in blacks.
Thus, the incidence in U.S. blacks was about 20% of that in whites,
a relatively much higher proportion than that found for NMSC.
In sharp contrast to the distribution in whites, non-Caucasians,
and blacks in particular, have a very high incidence of malignant mela-
noma on the foot (whites, 10-15%; blacks, 60-75%; Asians, 30-35%).
F-l-117
-------
Effects—Skin Cancer
Crombie (1979) in an extensive analysis of 59 population based regis-
tries, showed again that malignant melanoma occurs in Africans primarily
on the foot. He found marked heterogeneity among the nonwhite groups.
The tumors in Africans were mostly on the lower limb, particularly the
foot (60-75% of total MM) (Oettle, 1966; Lewis, 1967; Davies, et al.,
1968; Camain et al., 1972).
In Asian populations 40% of MM are of the acral lentiginous type,
and are rising at a rate of 57G per year in reasonable accord with wes-
tern trends. Haverkamp and Radman (1979) reviewed the experience of
American blacks - in a series of reports dating back to 1939, 50 to
76.5% of all malignant melanoma occurred on the foot. It has been sug-
gested that this disproportionate incidence on the foot is due to a
relative absence of nevus-cell nevi on the skin of non-Caucasians, ex-
cepting palms and soles (Coleman et al., 1980; Van Scott et al., 1957;
Kenny, personal communication). Thus, if nevus-cell nevi sometimes
are precursors or can be stimulated to develop MM, this maldistribution
could in part account for this phenomenon.
Instructive is the recent report of Hinds (1979) concerning inci-
dence in the non-Caucasian population of Hawaii. In contrast to rela-
tive incidence of 2.6% on feet in white males and 2.7%, in white females,
the relative incidence of malignant melanoma in non-Caucasians on the
feet was 41.5%, for males and 21.7% for females.
Finally, a report by El Bolkainy and Ebeid (1973) from Egypt is
instructive. Egyptians are racially Caucasians, but are of much darker
skin color than Europeans. In spite of the high solar exposure of most
inhabitants, malignant melanoma appears to be rare in Egypt. As in
F-l-118
-------
Effects—Skin Cancer
non-Caucasians, 43% of malignant melanomas occurred on the feet. In
contrast to blacks, where malignant melanoma is rare on the head and
neck, 26%, occurred in this area, about the same as in whites. However,
in contrast to whites, lentigo maligna melanoma is rarely, if ever,
encountered in Egypt. Since this latter lesion is the one commonly
associated with chronic UVR exposure in whites, it appears that the
heavy pigmentation of the Egyptians protects their skin (also evidenced
by the rarity of NMSC in native Egyptians). If that is the case, it
is difficult to conclude that the other malignant melanoma of the ex-
posed head and neck skin are due to solar UVR.
There is now clear evidence that a real genetic element is among
the factors responsible for the causation of malignant melanoma. Esti-
mates for 6 to 11% of malignant melanoma lesions to be of genetic origin
have been made by various authors (Andersen et al., 1970, 1971; Wallace
and Exton, 1972; Wallace et al., 1973). Familial patients manifest
a younger age distribution, a significantly earlier average age at first
diagnosis, increased frequency of multiple primary melanomas and a sig-
nificantly higher survival rate. The genetic mechanisms underlying
the familial type of melanoma is apparently complex and may involve
several autosomal gene loci, in addition to a cytoplasmic component
transmitted by an affected or carrier female. Neither the site nor
the type of lesion is specific, except that patients with familial
malignant melanoma and their relations seem to have distinctive melano-
cytic nevi, designated by Clark et al. (1978) the "B-K mole syndrome."
Such peculiar nevi begin in childhood, but the progressive cellular
atypia described occurs later, possibly because such genetically altered
F-l-119
-------
Effects—Skin Cancer
nevi are more susceptible to neoplastic transformation than the common
acquired melanocytic nevi. There are doubtless other phenotypes associ-
ated with familial melanoma, and additional studies need to be done
before this syndrome can be understood.
However, the various hereditary and ethnic factors affecting inci-
dence and distribution of malignant melanoma can not be only or even
primarily related to the effect of UVR on the skin. Although it is
tempting to read into the lesser frequency of malignant melanoma in
non-Caucasians the simplistic idea that such is due to protection by
pigment, clearly the matter is more complicated. The very high fre-
quency of malignant melanoma on the feet of non-Caucasians, the normal
distribution of malignant melanoma on the head and neck of pigment pro-
tected Egyptians and the frequency of familial incidence are not easily
explainable in this fashion.
Furthermore, American blacks have lower rates than American whites
for MM of the ureal tract and the mucous membranes (Scotto et al.,
1976).
F-l-120
-------
Effects—Skin Cancer
E.2.4 Malignant Melanoma and Exposure to Chemical Agents
The evidence for a causal association between chemical agents and
MM is meager. Isolated cases or small clusters of MM apparently related
to exposure to a variety of chemical agents have been reported. Levo-
dopa used for therapy of Parkinson's disease has been incriminated,
primarily because it is a common intermediate in the biosynthesis of
catecholamines and melanins. Skibba et al. (1972) reported a patient
who developed a recurrence of previously excised MM, and multiple pig-
mented skin lesions, some of which were histologically primary MM.
On the other hand, Sober and Wick (1978) found only one patient of 1,099
MM cases of the melanoma cooperative group who ingested Levodopa. They
concluded that if this drug played a part in the induction of MM, it
must be inconsequential.
Beral et al. (1977) noted higher incidence rates of MM in women
who had used contraceptives, particularly longterm users. They sug-
gested that the known effect of estrogen, augmented by the simultaneous
administration of progesterones, as stimulant of melanogenesis was the
effective mechanism.
More recently, Bahn et al. (1976) found a very small MM cluster
f
(2/31 exposed) in men heavily exposed to PCBs. However, further study
of this possible carcinogenic association is warranted, and the epi-
demiological significance of this apparent relationship between MM and
occupational exposures to chemicals is not known.
Pell et al. (1978) in a careful and detailed study of cancer inci-
dence and mortality in the DuPont Company found a significant increase
in the standardized incidence ratio for MM in male employees (SIR =
F-l-121
-------
Effects—Skin Cancer
123). This held for both wage and salaried employees. While the SIR
for females was also increased, it did not reach levels of significance,
lagging considerably behind cancer of the cervix and leukemia.
The excess of incidence in male employees was slightly greater
in the salaried group. One possible explanation offered was the com-
paratively high concentration of employees in the southeastern U.S.,
where mortality from MM is especially high.
Of interest also is that the standardized mortality ratio was sig-
nificantly less .than expected in males, and less, but not significantly
so, in females.
It may be suggested that this phenomenon is due to early recogni-
tion and treatment of lesions.
No discernible relationship to any specific chemical agent was
found in this study.
E.2.5 Geographic Distribution of Malignant Melanoma
The two major findings that support the contention that NMSC is
primarily caused by chronic exposure to solar UVR are the anatomic dis-
tribution of the cancers (on sun exposed areas) and the striking lati-
tude gradient of their incidence.
It has been pointed out in a previous section that the anatomic
distribution of MM does not clearly parallel the presumed sites of maxi-
mum solar exposure. In the case of one type of MM, LMM, the character-
istics of age distribution, anatomic distribution and histologic evi-
dence of solar connective tissue damage all seem, for practical purposes,
to be identical to the features of squamous cell carcinoma of the skin.
It must thus be assumed that solar UVR exposure is the primary cause
of LMM. However, this then applies to only about 10% of MM (Larsen and
Crude, 1979).
F-l-122
-------
Effects—Skin Cancer
Cutchis (1978) was one of the first to point out that the geo-
graphic distribution of MM, reported in Vol. Ill of Cancer Incidence
in Five Continents (Doll et al., 1976) shows significant anomalies if
the assumption is made that exposure to solar UVR is the most signifi-
cant factor in the causation of MM.
The major discrepancies found at that time were the relatively
large incidence of MM in the Scandinavian countries, particularly Norway
(Magnus, 1975, 1977) and Sweden (Malec and Eklund, 1978), the lack of
any evidence for a latitude gradient (or even a reversal from the usual
N/S increase) for central European countries (Crombie, 1979); and the
peculiar male-female ratio for MM, which usually remains below unity,
in sharp contrast to NMSC, where the M/F ratio exceeds 4:1 or more
(Urbach, 1972).
The data described above have been accumulated by the International
Agency for Research in Cancer (Doll et al., 1976) from worldwide cancer
registries. They certainly represent information based on various time
periods and acquired by different means. Since the incidence of MM
has been rising sharply world wide, it may not be too surprising if
these data did not represent strictly comparable enumerations.
If solar UVR is the primary cause of MM, then, as in NMSC, there
should exist recognizable latitude gradients for the incidence and mor-
tality of MM, with rates increasing toward the tropics. The existence
of such latitude gradients has been reported from Norway (Magnus, 1975),
Sweden (Magnus, 1977; Eklund and Malec, 1978), Finland (Teppo et al.,
1978), and the U.S.A. (Cutler and Young, 1975; Fears et al., 1976).
There are, however, inconsistencies in a number of these studies
which have recently been noticed. Of major importance have been the
F-l-123
-------
Effects—Skin Cancer
observations that major cities show a .disproportionately high incidence
of MM, which could not be explained on latitude alone, and that most
Nordic countries show incidence rates for MM greater than can be ex-
pected based on their latitude (Lee and Issenberg, 1972; Eklund and
Malec, 1978; Crombie, 1979; Viola and Houghton, 1978; Houghton et al.,
1980; Jensen and Bolander, 1980).
For instance, Teppo et al. (1978) found a distinct north-south
gradient for MM incidence along Finland, with the incidence rates 30%
or more greater in the south of that country. However, when the inci-
dence rates were adjusted to urban/rural population ratios, the latitude
gradient markedly diminished for the 1953-59 time period, and disap-
peared for 1961-73. They concluded that there was a marked effect of
urbanization and that factors other than latitude operate. Eklund and
Malec (1978) performed an epidemiologic study similar to that of Teppo
(1978) and Magnus (1975, 1977). Their major findings were: Sweden
has a much higher incidence rate for MM than central Europe. There
was a distinct north/south gradient along Sweden, but the major cities
had a disproportionately high incidence of MM. The difference in the
cities was primarily due to increases of MM of the trunk and arms of
men and the arms and legs of women. There was no disparity in incidence
of MM of the head and neck between rural and city areas. The differ-
ences could not be explained on latitude alone.
In contrast to the above, Beardmore (1972) and Little et al. (1980)
noted that in Queensland, Australia, MM was more frequent in subtropical
than in tropical areas, and Holman noted a reverse gradient (i.e., away
from the equator) in Western Australia. This finding was ascribed to
the above noted increased incidence of MM in large cities.
F-l-124
-------
Effects—Skin Cancel
Most recently, Crombie examined data from a large series of tumor
registries for Europe (Doll et al., 1976) and concluded that the inci-
dence of MM increased with increasing latitude in central Europe from
Italy to Germany, the reverse of that found in England and Wales, the
Nordic countries and the U.S.A. He investigated possible confounding
effects and concluded that the result was highly significant and not
due to confounding factors. He also noted the apparently excessively
high incidence of MM in Scandinavia and the excess found in large
cities. Crombie concluded that this latitude reversal may be due to
darker pigmentation of southern Europeans' skin, but pointed out that
the effect of change in susceptibility would have to be very large to
overcome a latitude effect.
Jensen and Bolander (1980) reviewed again the IARC and newer inci-
dence data. They comment on great variation in incidence of MM by coun-
try, the low incidence in countries with predominantly pigmented inhabi-
tants and point out that there are latitude gradients within countries
(England and Wale's, Sweden, Canada, U.S., Norway) which increase as
latitude decreases and exceptions to this phenomenon (Western Australia,
Texas, Finland) (Table 18).
Lee (1977) was the first to point out a peculiar association of
incidence of MM with certain, unexpected, occupational groups. He ob-
served, first in England and Wales, that the occupational groups that
suffered the highest mortality from melanomas of the skin were clerical
and professional workers and their wives. This initial finding has
been confirmed by Viola and Houghton (1978) for Connecticut, by Holman
et al. (1980) for Western Australia and summarized by Lee and Strickland
(1980).
F-l-125
-------
Effects—Skin Cancer
Am
Zone
M*le*
Hommn
Femil**
Femmn
AFRICA,—AFRIQUE
Nigeria, Ibadan — Nigeria, Ibadan
South Africa, Cape
Afrique du Sud, Le Cap
White —Blancs
Coloured—Metis
Bantu—Bantous ,
Natal
African—Africains ,
Indian — Indiens
Southern Rhodesia, Bulawayo (African) .
Rhodesia du Sud. Bulawayo (Africains)
AMERICA — AMERIOUE
Brazil — Bresil
Recife
S3o Paulo
0.9
3.1
0.2
0.7
0.8
0.9
3.0
Canada
Alberta
British Columbia
Colombia britannique
Manitoba
Maritime Provinces
Provinces maritimes
Newfoundland—Terre-Neuve
Quebec—Quebec
Saskatchewan
Colombia, Cali —Colombie, Cat!
Cuba
Jamaica,Kingston ..<
JamaTque, Kingston
United States of America
Etats-Unis d'Amerique
Alameda, White—Blancs
Black—Noirs
Connecticut .'
Detroit — Detroit '
White—Blancs
Black—Noirs
El Paso
Spanish—Espagnols
Other White—Autres Blancs
Hawaii
Hawaiian—Hawaiiens
Caucasian—Caucasians
Chinese—Chinois
Filipino—Philippins
Japanese—Japonais
Jpwa
New Mexico — Nouveau-Mexique
Spanish — Espagnols
Other White—Autres Blancs
Arher. Indian — Indiens d'Amerique
New York State — Etat de New York .
Puerto Rico—Porto Rico
San.Francisco Bay Area
Bate de San Francisco
White—Blancs
Black—Noirs
Chinese—Chinois
Utah
ASIA — ASIE
India.Bombay—Inde.Bombay .....
Israel — Israel
All Jews—Population juivetotale
Bom in Israel—Natifs d'lsrael ....
1.6
2.2
2.2
3.6
2.6
2.0
1.6
1.4
23
2.1
0.5
1.4
5.4
0.7
45
2.7
0.5
0.0
3.8
0.9
6.8
0.0
0.3
0.3
3.2
0.8
4.8
0.7
3.4
0.7
6.3
0.9
0.7
55
0.2
3.4
4.7
2.2
5.2
1.0
3.3
2.4
0.0
2.6
1.2
13
2.7
4.8
3.4
2.7
1.9
13
3.4
2.0
0.3
1.1
5.9
0.5
4.3
3.1
0.6
1.0
4.8
1.0
5.7
0.0
0.0
0.3
2.7
0.9
5.3
1.2
3.0
0.8
6.6
0.6
0.0
5.1
0.2
4.5
7.0
Ana
Zone
ASIA(concl.) — ASIE(fin)
Israel IconcU — Israel (fin)
Born in Europe or in America
Natifs d'Europe ou d'Amerique
Born in Asia or in Africa
Natifs d'Asie ou d'Afrique
Non-Jews — Non-Juifs
Japan — Japon
Miyagi
Okayama
Osaka
Singapore — Singapour
Chinese — Chinois
Malay — Malais .'.'.....
Indian — Indiens
EUROPE
Denmark — Danemark
Finland — Finlande ..'
German Democratic Republic
Republique dsmocratique allamande
Germany, Federal Republic of
Allemagne. Republique federate d'
Saarland — Sarre
Hungary — Hongrie
Szabotcs
Vas •-
Iceland — Islands
Malta — Malte
Urban — Population urbaine
Rural — Population rurale
Poland — Pologne
Cieszyn, Nowy Sacz
Cracow — Cracovie
Katowice
Warsaw, City — Varsovie, ville
Warsaw, Rural — Varsovie. campagne
Romania — Roumanie
Timis .'
Spain — Espagne
Zaragoza — Saragosse
Sweden — Suede
Switzerland — Suisse
Geneva — Geneve
United Kingdom — Royaume-Uni
Oxford
Sheffield
South Metropolitan Cancer Registry . .
Registre sud-metropolilain du cancer
SouthWest— Sud-Ouest
Liverpool
Ayrshire
Yugoslavia — Yougoslavie
Slovenia — Slovenie
MJI«»
Homm<»
3.4
0.9
0.7
0.3
0.1
0.3
0.6
0.4
0.4
2.9
23
2.1
2.4
2.3
1.6
1.8
1.6
05
5.4
6.7
4.5
0.6
1.7
1.2
2.3
03
IS
0.3
4.1
33
1.3 '
2.2
1.1
1.6
1.6
0.9
2.0
13
Fem^j
FBmr^
4.8
05
0.1
0.1
0.0
0.2
0.3
0.0
0.9
43
2.8
2.3
1.3
2.3
1.4
2.1
3.6
0.9
5.7
65
5.1
2.5
2.0
15
2.1
0.3
1.4
0.3
- 43
13
2.2
3.1
15
23
4.0
2.0
23
23
OCEANIA — OCEANIE
New Zealand — NouveHe-Zelande
Maori—Maoris 1-9
Non-Maori—Non-Maoris 9.4
• World Standard ?opulJt:on (DOLL. R. ET Al_ 19701. and
Sowct: WATERNOUSE. J. ET At, 1976).
15
11.7
Table 18. Age standardized incidence rates of malignant melanoma of
skin. (From Jensen and Bolander, 1980.)
F-l-126
-------
Effects—Skin Cancer
The size and consistency of the relationship of MM risks to some
i
factors associated with better education, high social status, or more
money, and the presence of the relationship in both employed men and
among their wives suggest that the effect is real. It suggests that
the relationship is a biological one between MM incidence and some fea-
ture of life associated with education or economic status. It may also
explain the above noted preponderance of MM incidence to large cities,
since the more affluent are more likely to live in cities and their
suburbs.
Teppo, Pukkala et al. (1980) in a remarkable study of the relation-
ship of "Way of Life and Cancer Incidence in Finland" confirm the pre-
dominance of MM in the towns and low rates in rural areas. They show
a consistently positive association with parameters describing the
socioeconomic level of the municipality, particularly for MM in rural
areas. A negative association was seen between farming and forestry
and MM.
In contrast, for cancer of the lip (the best reported form of NMSC)
a strong inverse association was seen between the incidence of lip can-
cer and that describing the standard of living of the municipality.
No other primary cancer site displayed suchstrong negative associations
than that found for lip cancer. The risk for lip cancer was highest
in the farming, forestry and fishing occupations, and in rural areas.
Thus a striking difference between MM and lip cancer was found as far
as income, city dwelling and outdoor occupations are concerned.
It has been previously pointed out that the incidence of MM has
been rising rapidly in white populations for many years. For example,
F-l-127
-------
Effects—Skin Cancer
in Norway a continuous rise in incidence of about 7°/c per year has been
observed since 1955 (Magnus, 1980). A major component of the causation
of this rising incidence is a systematic increase in risk of succes-
sively later born cohorts (Lee and Carter, 1970; Elwood and Lee, 1974;
Magnus, 1975; Eklund and Malec, 1978; Lee, et al., 1979; Holman et al.,
1980). This cohort increase began as early as the last quarter of the
nineteenth century in Australia (Holman et al., 1980) and about 1900
in Norway (Magnus, 1975). It is of interest that cohorts born since
1925-30 show a stabilization of this increased risk, but at a higher
rate (Elwood and Lee, 1975; Holman et al. , 1980). The increase and
cohort variations seem to be much greater for MM of trunk and lower
limbs than for MM of the face-neck area (Magnus, 1980). This indicates
that the trend in carcinogenic exposure through life may be different
for face-neck and trunk-leg areas. Why people born 5 years later should
go through life with a substantial increase in risk of dying from MM
compared with their elder peers in the same population is not known
(Lee et al. , 1979).
E.2.6 Sun Exposure and Risk of MM
In addition to the above described geographic variations of MM
incidence and mortality, several other observations have been proposed
so as to relate the risk of MM development to sunlight exposure.
Anaise et al. (1977) and Movshovitz and Modan (1973) reported that
MM was found more frequently in European Jews than in African or Asian
Jews in Israel; that the incidence was higher in Israeli born, European
descended Jews than recent immigrants and attributed this to greater
UV exposure in a sunny country. However, Hinds and Kolonel (1980) and
F-l-128
-------
Effects—Skin Cancer
Holman et al. (1980) noted reverse conditions, i.e. more MM in recent
immigrants to Hawaii or Western Australia than in the native born,
European descendants. Houghton et al. (1978) noted a cyclic increase
in MM superimposed on the previously reported continuing rise in MM
incidence two to three years after sunspot maxima in Connecticut, New
York and Finland, but not in Norway. Wigle (1978) confirmed this obser-
vation in Canada. Houghton et al. ascribed this to a decrease in
stratospheric ozone associated with increases in galactic cosmic rays.
They suggested ozone decreases of the order of 1 to 3% secondary to
sunspot influence. However, no such evidence of ozone decrease has
been observed in the long term Dobson measurements reported by London
(1980), nor have such changes in ozone been noted in association with
sunspot activity (London and Reber, 1979).
Scotto and Nam (1980) observed a strong seasonal pattern with sum-
mertime peaks for MM of females (particularly legs), and a similar peak
for MM of upper extremities of men. This was also noted by Malec and
Eklund (1978).
On the other hand, Leach et al. (1978) observed: A consistent
increase in stratospheric ozone over England (in keeping with the re-
sults of London, 1980), an accompanying fall in NMSC in Bristol, but
a steadily rising incidence of MM opposite to the ozone concentration
in the stratosphere.
It has been suggested by a number of observers (Lee, 1972; Fears
et al., 1976; Magnus, 1975, 1980) that the increase in the incidence
of MM, particularly among the affluent population in the cities, may
be due to unaccustomed, intermittent overexposure to sunlight on vaca-
tions, weekends, etc. Furthermore, the changes in clothing and outdoor
F-l-129
-------
Effects—Skin Cancer
recreation habits in the decades since World War II would allow more
exposure (men going shirtless, women with shorter skirts and nylon
stockings, more sun exposure for cosmetic tanning purposes, etc.).
To test this hypothesis, Eklund and Malec (1978) investigated the
issuance of passports in Sweden, reasoning that to get to a sunny area
one had to leave Sweden and go south. Indeed, it was found that more
passports were issued in cities than in rural areas, but the difference
in passport issue could at most explain 25% of the increase of MM in
Swedish cities. Furthermore, Brodthagen (personal communication) in-
vestigated the frequency of Danes participating in charter tours since
1950 (Fig. 11), the number of persons in organized tours of more than
three days' duration traveling to areas 35 to 40 latitude by year
since 1970 (Fig. 12), and the number of persons traveling to various
areas on these tours (Fig. 13). As can be seen, the frequency of Danes
leaving the country has increased dramatically (Fig. 11), but there
has been little change in the past decade in those going to sunny areas,
the greatest increase had been in travelers to England (i.e., the same
latitude), probably because of favorable foreign money exchange. It
thus appears that greater exposure to sunlight in the south, at least
for Danes, is not likely to be the major cause of the increase in MM
incidence.
In summary, it appears that the pathogenesis of NMSC and MM may
be quite different as far as the role of UVR is concerned. The pre-
dominance of NMSC on the head and neck of older persons strongly sug-
gests that cumulative, total UVR does is involved. The relatively early
onset and different distribution of MM, as well as the secular changes
F-l-130
-------
Effects—Skin Cancer
Number of Persons on Charter Tours from Denmark
(Company: Sterling Airways Ltd.)
Log Scale
10x10 -
614,507
50 54 58 62 66 70 74 78 YEAR
Figure 11
F-l-131
-------
Effects—Skin Cancer
Number of Persons in Organized Tours
(duration more than 3 days)
number x 10
i
400-
350-
300-
250-
200-
150-
100-
50
Latitude 35-40°
'Spain
Yugoslavia
70 71 72 73 74 75 76 77 YEAR
Figure 12
F-l-132
-------
Effects—Skin Cancer
number xlO
12-
10-
8-
6-
4
2
sr=r--: Latitude 25-35u
-.-•-:- Latitude 45°
Latitude 45-50°
,England
.-Austria
"**--Africa
France
^-Middle East
1 - 1
[— — i i i i
70 71 72 73 74 75 76 77 YEAR
Figure 13
F-l-133
-------
Effects—Skin Cancer
described above (cohort effects), strongly suggest that the latent per-
iod for MM is very much shorter than that for NMSC (Lee, 1972; Magnus,
1980).
Lee (1972) in an elegant analysis of the potential mechanisms of
MM induction suggested:
(1) that the causative mechanism for MM is complex and specific,
and certainly not "chronic irritation" (i.e., cumulative UVR
damage);
(2) that exposure to the eliciting agent operates with a short
latent period, so that MM develops in the susceptible persons
reasonably soon after they have a possibility of developing
the tumor
(3) that either a "systemic agent" which operates at a distance
is involved, or that the skin of susceptible persons contains
"dominant precursors," i.e. something akin to premalignant
clones of cells that are stimulated into fully developed can-
cers by an environmental agent, of which UVR may be one.
In summary, the geographic variation of the incidence of MM is
not nearly as clear-cut as is that of NMSC. While there are respectable
latitude gradients within some countries, these are affected by the
very real increase in MM incidence in big cities, which, at least in
Scandinavia, are located in the southern portions of the countries.
In contrast to this is a reverse latitude gradient along central Europe,
which appears to be real. Further studies will be needed to sort out
the degree to which exposure to solar radiation is responsible for the
development of MM.
F-l-134
-------
Effects—skin
E.3. Experimental Models for Malignant Melanoma
Although malignant melanoma occurs spontaneously in most vertebrate
animals, only a few experimental models exist, of which only in one
case the induction of MM has been associated with UVR.
The genetic regulation of MM has been studied in the platy fish,
in which a tumor producing gene and associated regulatory genes have
been identified (Vielkind, 1976), in the Sinclair swine system, in
which the development of MM is also genetically determined (Millikan
et al., 1974). In neither case is the development of MM associated
with exposure to UVR.
What appears to be a realistic model for MM has been reported by
Pawlovski et al. (1980). 40% of 65 albino guinea pigs, painted chroni-
cally with 7,12 dimethylbenzanthracene (a known, potent skin carcinogen)
developed metastasizing MM which apparently arose by a malignant trans-
formation of carcinogen induced junctional nevi. The clinical and his-
tological events leading to the development of MM were found to be quite
similar to human MM in most aspects. No UVR exposure was used in this
system.
Melanomas have been induced in several different rodent systems.
Isolated lesions of MM were produced by carcinogen painting of
skin in guinea pigs by Berenblum (1949). In the studies of Edgcomb
and Mitchelich (1963) and Clark et al. (1976), again, no UVR was used.
The neonatal administration of urethane caused a high incidence
of melanoma in non-inbred Syrian hamsters (presumably malignant trans-
formation of the "pigment spot" in these animals) (Vessilinovitch et
al., 1970).
F-l-135
-------
Effects—Skin Cancer
In mice, two reports of what appears to be UVR related production
of MM have appeared. Epstein et al. (1967) induced "blue nevi" in pig-
mented hairless mice by painting with a single dose of DMBA, and pro-
moted with chronic administration of UVR. They produced many squamous
cell cancers, and late in the life of the treated animals very few lo-
cally invasive MM lesions appeared. Kripke (1979) reported that 1 of
40 animals used in a modified initiation-promotion experiment developed
a pigmented lesion, apparently a MM. In this experiment, UVR was used
as initiator and croton oil as promoter (basically the reverse situation
from Epstein et al.'s experiment, where carcinogenesis was induced with
a chemical [DMBA] and UVR was used as "promoter" ).
Finally, Kripke and Lill (1979) have observed potentiation of
growth rate to a B16 melanoma line in UVR treated mice. This effect
could be due to immunologic alteration of the host animals, as has been
reported by Kripke (1977) for UVR induced skin sarcomas.
F-l-136
-------
Effects—Skin Cancer
F.I. Forecasting the Effects of a 5-year Delay
in Regulation of CFC Emission in the U.S.
Even assuming that the data given in the NAS report are correct
(and it appears that due to recent changes the overall depletion of
ozone and increases in DUV are overestimated), the effect of a 5-year
delay in regulation of CFC emissions in the United States on skin cancer
incidence is not likely to be noticeable.
In DuPont's comments (DuPont, 1980) on the NAS report (NAS, 1979),
an appendix A calculated the differential ozone depletion which would
follow a ban on all production of CFC-11 and CFC-12 in the United States
in 1980 or 1985. The calculation assumes continued constant emissions
abroad, and calculates that U.S. emissions would reduce gradually and
terminate 20 years after the ban. The appendix follows (Section F.I.I,
pp. F-l-140 to F-l-154) and contains details of the calculation method-
ology and results.
The calculation shows that the difference due to a 5-year delay
is always small. During the first 30 years (to 2010), the difference
in ozone depletion increases to 0.2%, and thereafter slowly declines,
by 2100 the difference due to the 5-year delay is 0.06%.
DuPont used the Setlow DNA action spectrum (by analogy to the NAS
procedure) to calculate the incremental DUV from the 5-year delay.
The incremental DUV from a 5-year delay changes with the incremental
ozone depletion, reaching a maximum of 0.6% by 2010 and declining to
0.2%, by 2100. For the reasons discussed in Section C, this overesti-
mates the effective DUV.
F-l-137
-------
Effects—Skin Cancer
In Section D.2.5 we noted that existing data indicates that NMSC
incidence is increasing by 2-3% per year.
Using the 2-fold biological amplification factor - again a worst
case "model" which overestimates any calculated changes in NMSC inci-
dence due to ozone depletion and DUV increase - the incremental effect
due to the 5-year delay may be crudely translated into an increase in
NMSC which gradually increases to 1.2% over the period 1980-2010 (while
the estimated present trend would presumably increase incidence by 60-
90%, at 2-3% per year), and then declines through the rest of the
twenty-first century (while the present trend would presumably increase
incidence by a further 180-270%). <•
The natural variations are such that analyses of NMSC incidence
figures over many years are necessary to detect and quantify the exist-
ing incidence trend of 2-3% per year. A superimposed increase of less
than 1% in the incidence of NMSC over a human lifetime would be lost
in the natural variations and would not be discernible.
For the reasons given in Section E of the text, it is not possible
to relate any change in MM incidence to the small change in DUV which
would be associated with the 5-year delay. MM incidence is also rising
worldwide at a rate of 3-7%, and as for NMSC, these increases in inci-
dence are in the absence of any measurable decrease in stratospheric
ozone.
*It should be immediately evident that this use of statistical data
is not appropriate from any rigorous statistical or epidemiological
viewpoint. It does, however, indicate the relative magnitude of the
effect of current trends compared with changes in incidence due to a
small increase in DUV.
F-l-138
-------
Effects—Skin Cancer
Thus, the calculated effect of a 5-year delay in regulation of
U.S. emissions of CFCs, even if it were to actually occur, would not
produce noticeable increases in MM or NMSC incidence.
F-l-139
-------
Effects—Skin Cancer
F.I.I DuPont's Comments on the Report by the Committee
on Impacts of Stratospheric Change (CISC) of the
National Academy of Sciences
(E. I. DuPont de Nemours & Co., Inc., 1980)
Appendix A
Effect on Calculated Ozone Depletion of a Five-Year
Postponement in U.S. Regulation of Chlorofluorocarbons
F-l-140
-------
Effects—Skin Cancer
VI. APPENDIX
Effect on Calculated Ozone Depletion of a Five-Year
Postponement in U.S. Regulation of Chlorofluorocarbons (CFC)
The Du Pont one-dimensional atmospheric model has been
used to calculate the changes in ozone concentration and flux of
damaging ultraviolet radiation (DUV) predicted to result from
emissions of CFC-11 and -12 for the following three scenarios:
Case A. Continued release at 1978 levels.
Case B. U.S. production halted at the end of 1980.
Case C. U.S. production halted at the end of 1985.
The calculations permit the determination, as a function of time, of
the differences in calculated ozone depletion and in DUV flux which
would result from a five-year postponement in a ban on U.S. fluoro-
carbon production. Cumulative dosage information is also available
from the calculations.
1. Results and Discussion
A. Ozone Depletion
Calculated cumulative ozone depletion for the three cases
is presented graphically in Appendix Figure 1 [p. 25] through the
year 2080. A U.S. cutoff is seen to reduce the calculated depletion
by an amount which increases slowly until the steady state values
(asymptotes) are reached. The difference between Cases B and C
reaches a maximum at about 2010 and decreases thereafter. The lat-
ter two cases are identical at steady state. This is seen more
clearly in Appendix Figure 2 [p. 26], where the calculated additional
ozone depletion due to a five-year postponement is plotted by year.
The maximum difference in calculated ozone depletion is 0.2 percent,
from 5.7 percent to 5.9 percent, in the year 2010. By 2080, a U.S.
cutoff in 1985 is calculated to lead to 0.08 percent more ozone
depletion than the 9.2 percent calculated for a cutoff after 1980.
B. Increase in DUV
As explained in more detail later, the calculated changes
in stratospheric ozone content have been used to calculate changes
in the ground-level flux of DUV. Appendix Figure 3 [p. 27] presents
that change as a function of time analogously to Appendix Figure 1
IP- 25]. The same general comments regarding the effects apply here,
since the DUV flux is implicitly related to ozone content. The
average value of a given curve during some time interval represents
the percentage increase in calculated cumulative dose relative to
exposure to the pre-CFC DUV flux for the same number of years.
Appendix Figure 4 [p. 28] is a plot of the calculated
additional DUV flux attributable to a five-year postponement in
F-l-141
-------
Effects—Skin Cancer
the assumed U.S. cutoff (analogous to Appendix Figure 2 [p. 26] for
ozone depletion). The maximum difference is 0.6 percent in about
2010 and the difference declines slowly to no difference at steady
state. By 2090, the difference is 0.25 percent. Again, the area
under the curve may be used to determine the calculated difference
in cumulative dose due to a five-year postponement. For example,
the increased DUV dosage for a 30-year period is the average value
of the curve during that 30-year period.
C. Steady State Changes
Assuming both U.S. and world production remain constant
up to the point of a U.S. cutoff and world production remains
constant thereafter, the steady state changes in ozone concentration
and DUV flux are independent of the date of the U.S. cutoff. The
steady state results are summarized in the following table.
Calculated Steady State Results
AOp (percent) ADUV (percent)
No Cutoff (1978 Release) -16.5 55.7
U.S. Cutoff -11.8 36.0
2. Details of the Calculations
The following sections describe the model used, the
calculation of DUV flux, and the release scenarios employed in
the calculations.
A. Model
The model is essentially the same as that employed in the
1979 NASA stratospheric workshop12, with a few minor modifications.
The photolytic behavior of NO is calculated on the basis of the
recent results of Frederick and Hudson1-^, rather than the* earlier
evaluation of Cieslik and Nicolet-1- . This leads to a greater amount
of NOy in the stratosphere. In addition, the mixing ratio of water
vapor at the tropopause has been increased from 3ppm to 4ppm. The
NO modification tends to decrease the sensitivity of stratospheric
ozone to a chlorine perturbation, while the water vapor change tends
to increase the sensitivity. The net effect of both is an increase
in sensitivity by less than a factor of 1.1. The calculations for
the NASA workshop employed so-called "standardized" 1975 chloro-
fluoromethane (CFM) release rates totalling 750 million Kg/yr world-
wide. The worldwide release rates in 1978 were a factor of *1.2
lower than the standardized 1975 values, and form the basis for the
present release figures.
F-l-142
-------
Effects—Skin Cancer
B. DUV Calculation
The amount of destructive ultraviolet radiation at ground
level, or DUV, is defined in terms of its biologically harmful
effects on DNA molecules. DNA molecules exhibit varying degrees of
sensitivity to monochromatic light at different wavelengths. This
sensitivity is characterized by the so-called "action spectrum" of
the molecule, which provides a weighting factor as a function of
wavelength for the imposed radiation flux.
In the present calculations, the DNA action spectrum
developed by Setlow15 serves as the basis for the calculation of
the DUV flux. The calculation includes the effects of Rayleigh
scattering of solar radiation and diffuse reflection at the ground
with an albedo of 0.25. The fluxes calculated in this manner are
the integrals over all solid angles of the so-called radiation
intensity I. Procedures for performing the radiation transfer
calculation with Rayleigh scattering and albedo are described in
Miller et_ al_.16'17 In effect, the DUV flux is treated exactly as
if it were a photolytic reaction rate in the model, with the photo-
dissociation cross section replaced by the product of hJ and the
action spectrum. In the present evaluation we calculate globally
averaged values of the DUV fluxes, although, because the model is
one-dimensional, the effects of ozone column variations with .latitude
are not accounted for.
C. Release Rates
Release rates for CFC-11 and CFC-12 follow the standard
data for release through 1978 as compiled by Alexander Grant for the
Chemical Manufacturers Association. It is assumed that, if U.S.
production were to continue unaltered, the worldwide release of CFC-
11 and CFC-12 would proceed at their 1978 levels (283.6 million
Kg/yr CFC-11 and 340.9 million Kg/yr CFC-12) indefinitely. The U.S.
production rates of CFC-11 and CFC-12 just prior to the halt in
production are taken as 60.0 million Kg/yr and 125.0 million Kg/yr
respectively. These estimates are based on the non-aerosol
production figures for 197812'18, together with estimates of the
residual aerosol production following the 1979 ban. The distribution
by uses of CFC-11 and CFC-12 prior to the assumed halt are taken to
be as follows:
CFC-11 Production CFC-12 Production
Use (million Kg/yr) (million Kg/yr)
Hermetic Refrigeration 3.5 33.1
Non-Hermetic Refrigeration 5.5 59.9
Closed Cell Foams 25.9
" 25.9
Open Cell Foams & Other 20.8
Aerosol 4.3 6.1
Total 60.0 125.0
F-l-143
-------
Effects—Skin Cancer
above distribution is based on that which existed for non-aerosol
during the years 1977 and 1978, and on the assumed residual
production.
The assumed U.S. production rate of CFC-11 represents 21
of the worldwide CFC-11 release rate, while that of CFC-12
,-ii-esponds to 37 percent of the worldwide CFC-12 release. The
.
-------
Effects—Skin Cancer
the production halt. These are listed in Appendix Table 1 [p. 24]
and plotted graphically in Appendix Figures 5 and 6 [pp. 29, 30].
The results in the table can also be used to determine the implied
total accumulation of U.S. production in existing equipment prior
to the cutoff. The value obtained is equal to ^5 years of assumed
U.S. production for both CFC-11 and CFC-12. This is rather large
compared to the ^2 years production in existing equipment worldwide
calculated for the year 1978. The larger value of the accumulation
was anticipated, and is attributable to the assumed constant U.S.
production rate and distribution by uses. Worldwide, the
accumulation has been increasing rapidly over the past decade owing
to the influence of the distributional changes that have occurred
in the U.S. The release of *5 years of U.S. CFM production during
the 20 years following the cutoff is realistically an upper bound
to the actual discharge that would occur. The release figures in
Appendix Table I [p. 24] can, in this sense, be considered to be
conservative, since they will result in a maximum amount of esti-
mated ozone depletion. A lower bound to the release is of course
provided by assuming no accumulation, and a release rate which
responds instantaneously to the production rate. The actual release
rate is expected to reside between these two limiting situations
(assuming no further growth in U.S. production).
F-l-145
-------
Effects—Skin Cancer
APPENDIX TABLE I
CFC-11 and CFC-12 Releases Following U.S. Cutoff
Year
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
CFC-11 Release
(million Kg/yr)
283.6
266.7
252.6
250.5
248.4
245.9
244.0
242.1
240.6
239.0
237.7
236.1
234.6
233.0
231.6
230.3
228.8
227.6
226.2
225.0
223.6
CFC-12 Release
(million Kg/yr)
,7
,3
340.9
320.0
298
289
277.6
265.6
255.9
249.2
244.6
241,7
239.0
236.0
232.1
228.8
225.8
223.2
221.2
219.2
218.2
217.2
215.9
F-l-146
-------
o
c
nl
CJ
T 1 1 1 T
STEADY STATE - CASE A
15
APPENDIX FIGURE 1
CALCULATED OZONE DEPLETION
10
STEADY STATE - B AND
CAS
o
CASE B
I960
1980
2000
2020
2040
2060
2080
-------
0)
o
c
0
c
-H
CO
I
tn
•P
o
0)
M-I
W
O
Q_
I—
GO
Q_
CD
APPENDIX FIGURE 2
CALCULATED OZONE DEPLETION
FOR FIVE-YEAR POSTPONEMENT
0.2
en
o
CO
Q_
UJ
(=3
LU
Q_
0.1
1980
2000
2020
2040
2060
2080
YEAR
-------
-------
U
c
•H
I
Cfi
t^O
~ (1)
by M-I
«— m
^ w
CD 0.6
Q_
I—
oo
CD
Q_
CD
~ 0.4
CJD
«a:
•zz.
CD
C=>
C=i
-------
0)
o
c
(0
u
c
-H
U)
-p
o
OJ
00
«=c
290
280
270 -
260 '
250
240
230
22Q
0
APPENDIX FIGURE 5
CFC-11 RELEASE FOLLOWING U.S. CUTOFF
10
15
20
YEARS AFTER CUTOFF
-------
J-l
0)
u
£
td
u
c
-H
I
I
CO
-p
u
W
oo
«=c
350 -
300*
250
APPENDIX FIGURE 6
CFC-12 RELEASE FOLLOWING U.S. CUTOFF
200
m
H
I
H
I
Pn
10
15
20
YF4RS,
-------
Effects—Skin Cance
VII. REFERENCES
1. National Academy of Sciences: Protection Against Depletion
of Stratospheric Ozone by Chlorofluorocarbons, Part I,
Report of the Committee on Impacts of Stratospheric
Change, 1979.
2. National Academy of Sciences: Stratospheric Ozone
Depletion by Halocarbons: Chemistry and Transport,
Report of the Panel on Stratospheric Chemistry and
Transport, 1979.
3. Du Pont: "Comments on the National Academy of Sciences
Report, Stratospheric Ozone Depletion by Halocarbons:
Chemistry and Transport." Submitted to EPA, January,
1980.
4. Reference 1, Part II, Report of the Committee on Alter-
natives for the Reduction of Chlorofluorocarbon Emissions,
1979.
5. Du Pont: "Comments on the December 1979 Report by the
National Academy of Sciences Committee on Alternatives
for the Reduction of Chlorofluorocarbon Emissions (CARCE).
Submitted to EPA, April, 1980.
6. National Academy of Sciences, Halocarbons: Environmental
Effects of Chlorofluoromethane Release, Report of the Com-
mittee on Impacts of Stratospheric Change, 1976.
7. Cutchis, P. On the Linkage of Ultraviolet Radiation to
Skin Cancer. Institute for Defense Analysis. Report to
the Federal Aviation Administration, Department of
Transportation, Report No. FAA-EQ-78-19, September 1978.
8. D'Arge, R.C., Daly, G., and Patten, C.W. Economic and
Social Measures of Biologic and Climatic Change. Climatic
Impact Assessment Program (CIAP) Monograph 6. Department
of Transportation DOT-TST-75-56 (1975).
Technology Assessment of the Fluorocarbon/Ozone Depletion
Problem, Systems Control, Incorporated. Report under
Grant No. ERS 77-09248 to the National Science Foundation
(1979).
Hoch, I., Climate, Energy Use and Wages. Chapter 4 in
Some Economic Aspects of Controlling Ozone Depletion,
A Report to the Environmental Protection Agency from
the University of Maryland. EPA Grant R 805411-01
(SR-EDAF), October'1979.
F-l-153
-------
Effects—Skin Cancer
9. Reference 1, p. 118, Table 4.2.
10. Ward, R.B., (Du Pont) to Klauder, D. (EPA). "Chemical
Manufacturers Association Fluorocarbon Project Panel
Research Program and Scientific Uncertainties".
November 19, 1979.
11. Summary: Research Program on Effect of Chlorofluoro-
carbons on the Atmosphere - Revision 12. Chemical
Manufacturers Association. November 30, 1979.
12. National Aeronautics and Space Administration. "The
Stratosphere: Present and Future," NASA RP1049, 1979.
13. Frederick, J.E. and R.D. Hudson, J. Atmos. Sci. 36, 737-
745, 1979.
14. Cieslik, S. and M. Nicolet, Planet. Space Sci. 21, 925-
938, 1973.
15. Setlow, R.B., Proc. Nat. Acad. Sci. U.S. 71, 3363-3366,
1974.
16. Miller, C., P. Meakin, R.G.E. Franks, and J.P. Jesson,
Atmos. Environ. 12, 2481-2500, 1978.
17. Miller, C., D.L. Filkin, and J.P. Jesson, Atmos. Environ.
131, 381-394, 1979.
18. McCarthy, R.L., F.A. Bower, and J.P. Jesson, Atmos. Environ.
11, 491-497, 1977.
F-l-154
-------
Effects—Skin Cancer
G. Critique of the NRC Report: Protection Against
Depletion of Stratospheric Ozone by Chlorofluorocarbons
G.I. Chapter 3 - Human Health Effects
To: Introduction (NAS, 1979, pp 74-75)
It is correct that "the epidemiologic evidence for the causative
role of damaging UVR (DUV) in the development of nonmelanoma skin cancer
of the exposed areas (the face and neck) is indisputable." (NAS, 1979,
p 74). However, the exposed areas of skin in people are not only the
face and neck, but also the upper back and the dorsae of the hands
and forearms, where SCC is commonly found but BCC is quite uncommon.
It is correct that "the susceptible population includes (one should
really say: consists of...) white, fair skinned persons "ho sunburn
easily and who receive prolonged occupational or recreational exposure."
There is only the most preliminary data that "a subset of this white
population [has]... a deficient DNA repair mechanism" (NAS, 1979, p
74) - this is based on two small studies on the repair capacity of
lymphocytes circulating in the blood of patients with actinic keratoses.
The meaning of these preliminary observations is by no means clear.
It is of interest that the same authors have not found such an abnor-
mality of DNA repair capacity in patients with malignant melanoma.
None of the above detracts from the clear and indisputable associa-
tion of most (but not all — see below) nonmelanoma skin cancer and
the cumulative effect of repeated, chronic exposure to solar UVB.
The situation is much less clear as far as malignant melanoma
(MM) is concerned. In contrast to the statement "recent evidence sup-
ports the previous conclusions of CISC that the development of malignant
F-l-155
-------
Effects—Skin Cancer
melanoma of the skin of man is related to DUV exposure," (NAS, 1979,
recent evidence casts doubt on a direct relationship of MM to solar
UVR exposure.
The latitude gradients vary from country to country, and in Europe
are even reversed. While the incidence of MM has been increasing world-
wide, the one body site where there has been no or little increase
has been on the most exposed area, i.e. the head and neck. The assump-
tion that females' legs are more exposed than men's because of the
wearing of skirts is not really borne out, when one considers that
the amount of UVR (excepting when on snow or ice, where UVR reflection
is high) that reaches the legs in the upright position is only 15-20%
of that reaching the head and neck.
The cohort effect described is real, but, on closer inspection.
began at least about the year 1900, and perhaps earlier in Australia.
It thus appears dubious that this can be primarily ascribed to increased
recreational exposure.
The only certain things about malignant melanoma are that there
has been a real worldwide increase in incidence (incidentally in the
absence of any change in stratospheric ozone), and that it is primarily
the middle class males and their wives who show this increase, i.e.
not those habitually and occupationally exposed to UVR.
To: General Aspects of Sunlight Exposure (NAS, 1979, pp 75-77)
It is true that "a single 15 to 20 minute exposure of white skin
to the sun at noon in middle latitudes during summer can produce a
slight sunburn" (modifications underscored, NAS, 1979, p 75). The
figure given for the dose for such an exposure is peculiar. 1.2 x
10 J/m is about the total solar exposure, assuming a solar constant
F-l-156
-------
Effects—Skin Cancer
0
of 1333 w/m and 15 minutes of exposure. This figure is misleading
and meaningless - unless the UVB dose is given (200-250 J/ro of energy
equivalent to the effectiveness of wavelength 297 nm), the number is
/• o
not useful. One could be exposed to the 1.2 x 10 J/m dose through
3 mm of window glass and have n_o biologic effect.
The statement that 1/3 of the white population will not develop
a protective tan has to be taken with a great deal of caution. In
the U.S., it is more like 15%, in central Europe perhaps 5%, in Ireland
and Scotland perhaps 50%.
To: Skin Cancer Caused by Exposure to Sunlight (NAS, 1979, pp 77-
79)
The statement that "...the evidence associating UVB with malignant
melanoma is not so strong as for nonmelanoma, we believe that the only
action spectra that we can prudently use for any quantitative estimate
of the potential hazard arising from an increase in UVB are those given
in (figure location)" (NAS, 1979, p 79) is certainly true, at least
in .part, if the assumption is made that UVB is the predominant cause
of MM. Whether this assumed conclusion is tenable is discussed below.
Comment should be made on the uncritical use of the DNA action
spectrum that is used to derive "DUV" in this monograph. It is clear
that the known DNA action spectra in the UVR have a slope similar to,
but somewhat steeper than, the accepted skin erythema action spectrum.
The major difference occurs at wavelengths shorter than 295 nm, where
the erythema effectiveness becomes less, while the DNA effectiveness
continues to rise.
Since the only cells in human skin normally capable of cell divi-
sion (and thus capable of being transformed to cancer cells) are the
F-l-157
-------
Effects—Skin Cancer
basal cells and possibly the melanocytes, the radiation must reach
the nuclei of such cells to be effective. These cells are located at
the bottom of the epidermis, and effective radiation must pass several
cell layers (containing DNA in their nuclei, but being unable to divide)
before it can reach its target. Since shorter UV radiation is more
absorbed than longer UVR, the "real" effective action spectrum for
susceptible cell damage will have a shape somewhere between the DNA
absorption spectrum and the erythema action spectrum. Our own experi-
ments in hairless mice, showing a strikingly increased photocarcinogenic
effectiveness of wavelengths shorter than 295 nm,must be taken with
caution, since the mouse skin averages 2-3 cell layers above the basal
cells, and human skin has 5-10 cell layers able to absorb radiation.
Studies to determine actual transmission of UVR to the basal cell layer
of human skin are of critical importance, since, as pointed out before,
the shape of the action spectrum can have a striking effect on calcula-
tions of "DUV" (which is an action spectrum-weighted number), particu-
larly when solar UVR increase is being considered.
To: Evidence for Sun-Induced Skin Cancer (NAS, 1979, pp 80-85)
The logic of the introductory statements as far as NMSC is reason-
able and in keeping with best present knowledge. It is unfortunate
that basal cell carcinoma, as described in NAS, 1979, p 84, Figure
3.4, should have been used as the example for documentation that "the
skin cancers should be heavily concentrated on the parts of the body
most exposed to the sun," (NAS, 1979, p. 81) since 1/3 of the basal
cell carcinomas on the head and neck occur in areas receiving, for
anatomic reasons, very little UVR and where the adjacent skin shows
little or no solar damage (inner and outer canthus, upper lid, post-
auricular area, etc.).
F-l-158
-------
Effects—Skin Cancer
Whether solar radiation alone causes skin cancer ("pattern 1")
or whether the dependence on exposure is more complex ("pattern 2"),
the logical reasoning is consistent, but the arguments are circular:
Since all the evidence strongly points to major differences in causal
mechanisms between NMSC and MM, and since considerable data as to what
is happening is available, it is not difficult to determine criteria
and then show that the observations match the proposed criteria. So,
the criteria become self-fulfilling prophecies, since they were drawn
from existing data.
Nevertheless, if the assumed criteria are correct (and of course
they fit present day data), it must be clear that predictions of future
effects are at best very difficult (e.g., the most recent statement
in the 2nd Biennial Report to Congress of NCI - 1979: "A study of
skin melanoma incidence data from 10 locations in the U.S. has also
produced preliminary results. Although incidence is increasing, the
dose-response relationship between UVB radiation and melanoma suggests
a complicated involvement of environmental and host factors requiring
further investigation.").
The NAS report, nevertheless, states further, "(for MM) the lati-
tude dependence appears well established" (NAS, 1979, p 82). Data
quoted above (section E) shows that this statement is at best a half-
truth. Latitude gradients are reversed in Europe, Western Australia
and not present in Finland, and incidence maxima do not correspond
to latitude - e.g. higher in Stockholm than in England or Connecticut.
The anatomic distribution of MM only vaguely corresponds to insola-
tion - the greater frequency of MM on head and neck in males is con-
siderably weighted by the presence of lentigo maligna melanoma, which,
F-l-159
-------
Effects—Skin Cancer
while comprising only 10% of MM, really seems to be due to chronic
insolation. MM on the scalp has nothing to do with UVR, since hair
is an almost perfect neutral density filter, and MM does not appear
excessively on bald scalps. The unlikeliness of the really greater
incidence of MM on the legs of women to be related to sunlight has
been commented on - unless all these ladies sunbathe in skirts.
The putative relationship between clothing styles and incidence
of MM on legs is at best tenuous, since north Australian men, who
habitually wear shorts all year long, have low incidence of leg MM,
while both in Hawaii and recently in Norway, the difference in the
incidence of MM on legs between males and females has almost disap-
peared.
It is certainly true that there has been a worldwide increase
in the incidence of MM. To ascribe this mostly to changes in exposure
patterns appears premature. It is also true that the dose-response
relation to DUV is complicated and unlikely to involve many other vari-
ables.
To: Body Location (NAS, 1979, pp 86-89)
There are several statements where knowledge is implied but no
good data to back the statements exists:
E.g.: "Most (MM) occur on lightly covered or occasionally un-
uncovered regions of the body...Very few have been seen
(either in males or females) on the regions ordinarily
covered by bathing suits" (NAS, 1979, p 87). If the in-
duction period is short (5 years or less), where are the
MM on the abdomen and lower back of women who, at least
in the U.S., have been wearing two—piece bathing suits
at least since 1950?
F-l-160
-------
Effects—Skin Cancer
.Is there data on the frequency of men exposing their
trunks to the sun during recreation?
.The data quoted on transmission of UVR by fabrics of
Infante and Daniels (NAS, 1979, p 87) is wrong. Unfor-
tunately, they neglected to remember the almost universal
use of optical brighteners which resulted in spurious
measurements not due to UVR.
Realistic studies show maximal transmission of even
thin woven cloth on the order of less than 10% below 320
nm (Herskowitz and Dobes, in press).
.Why does lack of change of incidence on face or feet
in Norway (and other areas) during a period of rapid rise
in incidence on other areas support the hypothesis that
"change in behavioral pattern is an important factor in
the striking increase in incidence of MM since World War
II" (NAS, 1979, p 89)?
Even if one assumes that some factor in the skin (pre-
cursors, nevi, etc.) are rate-limiting for the effect
of solar UVR, it is unreasonable to expect that indoor
workers and their wives have already had maximal effective
UVR doses to the face, and certainly they receive more
UVR to the face than in the past if their backs and legs
now have more "recreational" exposure.
To: Forecasting the Effects of DUV Increase (NAS, 1979, pp 96-99)
The best present estimates suggest that the biologic amplification
factor (to convert from DUV increase to an increase in NMSC incidence)
is 2. In other words, for a 1% increase in DUV, it may be expected
F-l-161
-------
Effects—Skin Cancer
that there will be a 2% increase, at equilibrium, in NMSC. It should
be noted, however, that the overall effect of ozone depletion on NMSC
incidence may be reduced when two dimensional modeling is used (Pyle
and Derwent, 1980).
The same relationship does not hold for MM. At this time, no
reasonable calculations for the biological amplification factor for
MM can be calculated. Even for NMSC, a simple number for this factor
is an oversimplification.
It is true that, given a decrease in 0.,, the intensity of DUV
will build up at higher latitudes, towards that previously found at
lower latitudes. However, the dose received by an individual at risk
will be smaller at the higher latitudes even then.
This is due to the fact that the dose (i.e. intensity x time)
varies so strikingly with sun angle; the sun angle of course will not
change with ozone concentration. Also, since the time each day that
the sun angle exceeds 55 from the vertical varies with latitude by
season, and since ozone concentrations vary with latitude by season,
the effects of a per cent decrease at one latitude can not simply be
equated with a shift to a lower latitude.
The final paragraph of this section contains a peculiar set of
arguments:
."An increase in MM deaths..., if it occurred, would be delayed,
well beyond the onset of a DUV increase, while the accumulated
dose builds up in individuals" (NAS, 1979, p 99). If the assump-
tion is that MM has a short incubation period, and is not related
to chronic, repeated UVR exposure, but is due to "intermittent
overdoses," then that statement is peculiar. Perhaps that is
F-l-162
-------
Effects—Skin Cancer
quibbling, because technically there will of course in either
case be a delay. However, the deaths due to MM should rise then
much faster than the incidence in NMSC.
In any case, the proposed calculation is not correct. If
a reduction of ozone by 16% causes a 44% increase in DUV, and
the biologic amplification factor (DUV to skin cancer) is x2 ,
the final increase would be 88% (i.e., 2 x 44%), not 64%. Table
3.5 (NAS, 1979, p 102), attributed to Scott and Straf, shows that
for MM mortality, an estimated 20% reduction of ozone would, by
their calculations, result in an increase of 1.5 to 2 times (HANES
TCNS or TCNS only) in Minneapolis-St. Paul (northern U.S.) and
a 1.5 to 4 times increase for the same data in Dallas-Fort Worth
(southern U.S.). All these figures, from the best estimates avail-
able at present, are significantly below the overall estimate
quoted above in the chapter.
All data available suggest that the north-south increase
in skin cancer has a nonlinear relationship - a considerably dif-
ferent assumption from the one used in all predictions, including
those in this monograph.
In summary, the descriptions of available data fit best present
knowledge as far as NMSC is concerned. There are serious problems
in trying to relate the incidence of MM in any straightforward way
to DUV exposure. The obvious evidence for the interaction of factors
other than DUV exposure makes risk calculation at this time extremely
hazardous. Certainly it is not reasonable to use the same transfer
functions which are appropriate for NMSC to estimates of changes of
MM.
F-l-163
-------
Effects—Skin Cancer
It is unfortunate that, at this time, a more reasonable approach
to this important problem does not suggest itself. However, as Lewis
Thomas once wrote:
"The skeptics in medicine have a hard time of it. It is much
more difficult to be convincing about ignorance concerning disease
mechanisms than it is to make claims for full comprehension, es-
pecially when the comprehension leads, logically or not, to some
sort of action. When it comes to serious illness, the public
tends, understandably, to be more skeptical about the skeptics,
more willing to believe the true believers." (Nature 284:298-299,
1980)
G. 2 Appendix C, Solar UV Irradiance at the Earth's Surface
(NAS, 1979, pp 300-303)
Excellent, simple description of a very complicated problem.
Does point out the difficulties inherent in making calculations of
spectrum and amount of UVR reaching earth, and the uncertainties in-
herent in all calculations.
G.3 Appendix D, The Biologically Effective UVR (NAS, 1979, pp 304-318)
Excellent, competent description of the problems inherent in
weighting solar UVR to obtain a "biologically effective" unit. The
magnitude of the differences due to different action spectrum assump-
tions is clearly shown, as is the problem inherent in normalization
of various action spectra.
G.4 Appendix F, Factors in UV Dose-Response of NMSC and MM
(NAS, 1979, pp 325-334)
It is correct that "NMSC appears to correlate best with cumulative
lifetime UVB exposure, therefore, older persons will have more cumula-
tive effects" (NAS, 1979, p 325).
F-l-164
-------
Effects—Skin Cancer
Again, a generalization. Obviously, one has to live long enough
to accumulate enough DUV to have an effect, but age alone is not the
criterion - it is exposure - in one of our studies (Urbach et al.,
1976) people of the same age can have as much as 25-fold differences
in outdoor exposure.
"Melanoma does not so correlate, which suggests that intermittency
of exposures may be important" (NAS, 1979, p 325). The first clause
of that sentence is correct; the conclusion is a speculation. One
might equally well conclude that therefore:
• DUV is not causally related to MM
• Predisposed skin requires much less accumulated dose to cause
MM (e.g. if there is an "initiated" skin area that is "promoted"
by UVR).
• Interaction of DUV with some other agent (chemical? viral?) is
effective.
• Etc.
Effect of Skin Color (NAS, 1979, p 328)
While people with genetically dark skin have indeed less MM, the
difference in incidence is not as striking as in NMSC (1:5 compared
to less than 1:50, with almost no basal cell carcinoma). When MM occurs
in pigmented people, 50+% occurs on the foot - indeed a lighter skin
area, but one that does not receive any solar UVR, and both palms and
soles are so heavily keratinized that it is almost impossible to sunburn
these areas because of the thick keratin layer that is dead and absorbs
virtually all UVR.
Precursor Lesions (NAS, 1979, p 330)
Familial incidence, of genetic origin, occurs in about 10% of
F-l-165
-------
Effects—Skin Cancer
all MM. These are mainly in younger groups and multiple lesions occur.
UVB can not be incriminated.
UVA-Induced Carcinogenesis (NAS, 1979, p 331)
The first experiment quoted was indeed reported in 1973. However,
it has since been found by Forbes (and reported at meetings where at
least two of the committee members were present) that this effect of
UVA potentiating UVB effects was due to the phenomenon of protraction
(Forbes, 1978).
It was found that, when the same dose (and spectrum) of UVR, from
the same light sources, was given over a 5 minute, a 50 minute or a
500 minute time period each irradiation day, the dose delivered in
500 minutes was very much more effective than that given in 5 minutes.
Repetition of the original experiment, with the original two light
sources, adjusted to deliver the daily UVB dose at the same rate showed
that additional UVA had n_o potentiating effect (Forbes et al., 1978).
While it is possible to produce skin cancer in hairless mice with
UVA (filtered to remove all UVB) it takes about 1600 times the daily
UVB dose to produce such effects.
Sunlight ('1 July 1976 - cloudless day, actual measurements, 40°N
latitude) delivers from sunrise to sunset 31 times as much UVA as UVB.
In terms of skin minimal erythema doses, that amounts to 17 MED
UVB and 2.25 MED UVA to a horizontal observer. Clearly, the contribu-
tion of UVA to NMSC induction in nature is negligible.
G.5 Appendix G, Further Detail on Malignant Melanoma
(NAS, 1979, pp 335-341)
Cohort Experience (NAS, 1979, p 336): Recent reports (see section
E) show that the increase in incidence in MM began with cohorts born
before 1900, possibly as early as 1885 in Australia.
F-l-166
-------
Effects—Skin Cancer
Geographic Variation (NAS, 1979, p 337)
More recent reports show that latitude gradients for MM vary great-
ly in various parts of the world. In part this is due to the peculiar
predilection of MM to occur in white collar, affluent people, who are
mostly city dwellers. In Finland, Norway and Sweden, the major cities
are in the south. In Western Australia, the gradient is reversed,
apparently because of the major city, Perth, being much further from
the equator than the rest of the state.
Even in Queensland, the MM incidence is higher in the subtropical
than in the tropical regions.
That these "deviations" are not just due to such factors as suscep-
tible population, climate, etc. is shown by the fact that the gradients
for NMSC show very little "deviation."
Types of Melanoma (NAS, 1979, p 337)
The four major types of MM are described. No mention is made
of the fact that of these, lentigo maligna melanoma behaves exactly
as a UVB induced skin cancer, i.e., squamous cell carcinoma. It is
most frequently found in the elderly, on the most light exposed areas,
develops slowly, and metastasizes late. Histologically, solar UVR
induced changes are almost always present, and, in the opinion of most
expert pathologists, a sine-qua-non for the diagnosis of this lesion.
Again, much is made of the localization in areas that are light
exposed, including legs of women. The problems that this attribution
causes have been discussed before.
G.6 Appendix H, Preventive Measures Related to Melanoma
(NAS, 1979, pp 342-344)
This brief chapter should be entitled: Preventive Measures Related
to Skin Cancer (modification underscored).
F-l-167
-------
Effects—Skin Cancer
The program outlined is reasonable and objective. It certainly
will lower the incidence 'of NMSC. Whether it will affect MM (except
for the self-examination, which can successfully lead to early diagnosis
and treatment) depends on the degree to which MM incidence is related
to UVB exposure.
F-l-168
-------
Effects—Skin Cancer
H. Recommendations for Further Research
Recommendation 1; Present Ozone Levels
That more accurate data on present ozone levels be obtained, par-
ticularly near sites where epidemiologic data have been or will be col-
lected.
Present Dobson ozone stations are scattered and make a limited
number of observations. Satellite ozone estimation offers greater in-
formation over wider areas and longer periods. Comparison of relative
pertinent usefulness of Dobson and satellite ozone measurement is re-
quired.
It must be anticipated that there will be processes other than
the SSTs which may perturb the steady ozone state. There is thus an
urgent need for ozone baseline data which are reliable, accurate, and
continuous. It is important to attribute any changes reliably to the
correct cause.
Recommendation 2: UV Radiation Reaching Earth
That accurate measurement be continued of the UV radiation reaching
earth from sun and sky:
(a) To establish baseline levels.
(b) To establish the range of existing natural variation.
(c) To monitor for possible steady change resulting from various
causes.
(d) To establish more reliably the relationship between ozone
and effective UV for various results.
Two useful classes of measurement have already been instituted:
A. Spectroradiometer
Bener at Davos has completed extensive studies on sun and sky show-
ing absolute intensities at all UV wavelengths and indicating apparent
F-l-169
-------
Effects—Skin Cancer
influences of altitude, ozone and elevation of the sun above the hori-
zon. It is recommended that further work should proceed at this exist-
ing observatory. (Other sites may also be desired.)
B. UV-B Meters
A network of UV-B meters (based on the design of Robertson as im-
proved by Robertson and Berger) has been established under the auspices
of the CIAP and NOAA. These instruments automatically integrate the
intensities of all UV-B wavelengths weighted to closely match the action
spectrum for production of erythema in untanned white human skin. They
are rugged, operate continuously with minimum attention, are relatively
inexpensive and stable and give an immediate approach to the overall
effectiveness of the sun's radiation as regards long-term damage to
human skin.
Present UV-B meters are located (usually at weather bureau sta-
tions) at the following locations: Philadelphia (Base), PA; Des Moines,
IA; Minneapolis, MN; Bismarck, ND; Tallahassee, FL; Mauna Loa, HI;
Brisbane, Australia; Aspendale, Australia; Davos, Switzerland. Further
stations will be established within the next three months at: Honey-
brook, PA; Tucson, AZ; Rochester, MN; Hamburg, Germany; Poland (2);
Italy, It is recommended that this network be continued and that con-
tinued assessment of results be supported.
Recommendation 3; UV Measurement Stations
That UV measurement stations be established at additional sites
with significance to the predictions desired for the ozone program.
Recommendation 4: Appropriate Improvement of the UV-B Meter System
The present UV-B meters (provided to meet the urgent demands of
the CIAP program) include minor instrumental errors which can and should
F-l-170
-------
Effects—Skin Cancer
be reduced to meet the more exacting requirements of a program ext.ending
forward for a longer period.
(a) Cosine error resulting from flat filters can be reduced by
hemispherical filters and changes in phosphor geometry.
(b) Temperature errors can be reduced by thermostatic control
to nearly constant temperature.
(c) Safer and quicker handling of data can be provided by improved
computer access through a process such as tape recording.
(d) Ensuring greater long-term reliability of results by improved
calibration procedures.
Recommendation 5: Personnel Monitors
Studies with personnel monitors (such as the KPR type already ap-
plied in Switzerland) are needed to relate variations in personal be-
havior and orientation with the indications of fixed outdoor detectors.
This factor does vary from one community to another and present
lack of knowledge on it acts to confuse application of existing data
on chronic skin damage.
Recommendation 6; Incidence of Skin Cancer
A better baseline for present incidence of skin cancer in man is
required:
(a) To make more reliable predictions of possible change.
(b) To attempt to recognize change in years hence data on inci-
dence (or prevalence) of skin cancer. This should be obtained
in at least 6-8 areas separated by at least 300 miles north-
south over a latitude span reaching beyond the most populated
areas. Similar data are needed nearer the equator. It is
F-l-171
-------
Effects—Skin Cancer
of paramount importance that all these studies be peformed
according to the same protocol so that valid comparisons may
be made. Promising areas for such studies, in addition to
the U.S., are Australia (particularly Queensland), Scandinavia
and South Africa.
Recommendations: Animal Experiments Associated with Development
of Skin Cancer
Recommendation 7: UV-B/Solar Spectrum Interaction
Animal experiments associated with development of skin cancer.
Some of these are already planned and to some extent operating as pre-
liminary trials. They should lead to a separation of the relative quan-
titative significance of differences in:
(a) Total dose
(b) Dose rate (flux)
(c) Diurnal and overall fractionation
(d) Action spectrum
(e) Combination of longwave UV and visible radiation with the
active "carcinogenic" wavelengths.
Separation of these different factors is necessary to permit proper
evaluation of even existing dose-response information, and to make it
possible to enhance predictability of effects due to changes in the
radiation climate.
A study is required of the effect of interaction between UV-B and
the rest of the solar spectrum in relation to skin tumor development.
Experiment has already shown that this effect exists.
F-l-172
-------
Effects—Skin Cancer |
It is particularly pertinent in this field where ozone reduc-
tion will affect UV-B only.
Recommendation 8: UV-B Flux
Investigation is needed of the influence of change in flux of UV-
B on skin carcinogenesis. Preliminary experiments show that protracting
the delivery of a dose of UV (i.e., altering flux) has a significant
effect on skin carcinogenesis. The direction of the effect is unex-
pected - lower flux can be more effective.
Recommendation 9; Intervals of UV Exposure
It is recommended that the effect of varying interval periods
during carcinogenesis experiments be studied in detail. The dose-
response model of Blum, and the model for effect of "effective exposure"
of Robertson, demonstrate that intervals of exposure are of importance
in skin carcinogenesis. This may be one of the most important experi-
ments, since in nature, change in ozone will affect flux and spectral
distribution of UV but not the time relationship during which people
will be exposed.
\
Recommendation 10: Animal Models for Malignant Melanoma
At present, the only workable animal model for MM is the guinea
pig system reported by Pawlovski et al. (1980). However, this utilizes
chronic skin painting with a chemical carcinogen only. It is recom-
mended that experiments using the same guinea pigs but using UVB ex-
posure be performed. In addition, the interaction of chemical carcino-
gens and UVR should be investigated.
Training and Reference Standards
Recommendation 11: There is a dearth of trained photobiologists in
the world. Primarily this is due to the fact that photobiology is a
F-l-173
-------
Effects—Skin Cancer
very young branch of science, and that at this time no real Departments
offering graduate training exist in this field. This is particularly
true of the animal experimentation area.
It is recommended that efforts be made to develop and support one
or two centers capable of offering such graduate training. Nuclei for
such centers exist at Harvard and Temple Universities.
Recommendation 12: A serious problem exists due to' the fact that var-
ious investigators use a wide variety- of UVR sources and measuring de-
vices which are not "standardized" in the usual sense in which this
word is used.
It is recommended that a central radiometric laboratory be estab-
lished and supported, capable of making accurate spectral and intensity
measurements of varius light sources, of developing and supplying ref-
erence sources, and of calibrating and testing measurement devices in
the UVR. Such activities would make it possible to compare results
obtained in many laboratories critically to each other, something which
is not possible at present.
F-l-174
-------
Effects—Skin Cancer
I. References
Anaise, D. , Steinitz, R. and Ben Hur, N. Solar radiation: A possible
etiological factor in malignant melanoma in Israel.
42:299-304, 1978.
Andersen, D. E., Smith, L., Jr. and McBride, C. M. Hereditary aspects
of malignant melanoma. J. Am. Med. Assoc. 200:741-746, 1967.
Bahn, A. K., Rosenwaike, I. et al. Melanoma after exposure to PCBs.
N. Engl. J. Med. 259:450, 1976.
Bain, J. and Rush, H. P. Carcinogenesis with UV radiation of wave-
lengths 2800-3400 A. Cancer Res. 3:425-430, 1943.
Bakes, L. and MacMillan, A. L. Malignant melanoma in East Anglia,
England. Brit. J. Dermatol. 88:551, 1973.
Bart, R. S. and Scholl , S. Eye color in darkly pigmented basal cell
carcinomas and malignant melanomas. Arch. Dermatol. 107:206-207,
1973.
Beard, H. H. , Boggess, T. S. and von Haam, E. Experimental production
of malignant tumors in the albino rat by means of ultraviolet rays.
Am. J. Cancer 27:257-266, 1936.
Beardmore, G. L. The epidemiology of malignant melanoma in Australia.
IN: Melanoma and Skin Cancer. New South Wales Government Printer,
Sydney, 1972, pp 39-64.
Bener, P. Approximate Values of Intensity of Natural Ultraviolet Radia-
tion for Different Amounts of Atmospheric Ozone. European Res.
Office, U.S. Army (Contract DA&A 36-68-C-1017), 1972.
Beral, V. Ramcharan, S. and Favis, R. Malignant melanoma and oral con-
traceptive use among women in California. Brit. J. Cancer 36:804-
809, 1977.
F-l-175
-------
Effects—Skin Cancer
Berenblum, I. The carcinogenic action of DMBA on the skin of the mouse,
rabbit, rat and guinea pig- J. Natl. Cancer Inst. 10:167-174,
1949.
Berenbluro, I. and Shubik, P. The role of croton oil applications asso-
ciated with a single painting of a carcinogen, in tumor induction
of the mouse's skin. Brit. J. Cancer 1:379-382, 1947.
Berger, D. , Urbach, F. and Davies, R. E. The action spectrum of ery-
thema induced by UV radiation. IN: Proc. 13th International
Congress of Dermatology (W. Jadassohn and C. Schirrer, eds.),
Springer Verlag, Berlin, 1968, pp 1112-1117.
Bischoff, F. Carcinogenic effect of steroids. IN: Advances in Lipid
Research (Paoletti and Kritschevsky, eds.), vol. 7, New York,
Academic Press, 1969, pp 165-244.
Black, H. S. and Douglas, D. R. Formation of a carcinogen of natural
origin in the etiology of UV carcinogenesis. Cancer Res. 33:2094-
2096, 1973.
Black, H. S. and Lo, W. B. Formation of a carcinogen in human skin
irradiated in ultraviolet light. Nature (Lond.) 234:306-308,
1971.
Blois, S. Vitamin D, sunlight and natural selection. Science 159:652,
1968.
Blum, H. F. Wavelength dependence of tumor induction by ultraviolet
radiation. J. Natl. Cancer Inst. 3:533-537, 1943.
Blum, H. F. Carcinogenesis by Ultraviolet Light. Princeton, Princeton
University Press, 1959.
Blum, H. F. Quantitative aspects of cancer induction by UV light.
IN: The Biological Effects of Ultraviolet Radiation (F. Urbach,
ed.), Oxford, Pergamon Press, 1969, pp 543-549.
F-l-176
-------
Effects—Skin Cancer
Blum, H. F., Butler, E. G. et al. Irradiation of mouse skin with single
doses of ultraviolet light. J. Natl. Cancer Inst. 22:979-993,
1959.
Blum, H. F. and Lippincott, S. W. Carcinogenic effectiveness of UV
P
radiation of wavelength 2537 A. J. Natl. Cancer Inst. 3:211-216,
1943.
Blum, H. F. and Terus. The erythemal threshold for sunburn. Am. J.
Physiol. 146:107-117, 1946.
Bovie, W. T. and Klein, A. Sensitization to heat due to exposure to
light of short wavelengths. J. Gen. Physiol. 1:331-336, 1919.
Bruusgaard, cited in Rasch, -C. Some historical and clinical remarks
on the effects of light on the skin and skin diseases. Proc.
Royal Soc. Med. 20:11-20, 1926.
Camain, R. A., Tuyn, A. J. et al. Cutaneous cancer in Dakar. J. Natl.
Cancer Inst. 48:33-49, 1972.
Carlson, L. D. and Jackson, B. H. Combined effects of ionizing radia-
tion and high temperature on longevity of Sprague-Dawley rats.
Radiat. Res. 11:509-519,- 1959.
Charcot, P. Erytheme produit par 1'action de la Lumiere electrique.
Comptes Rendus Soc. Biol. (Paris) 5:63-65, 1858.
CIAP Monograph #5. Impacts of Climatic Change on the Biosphere. Cli-
matic Impact Assessment Program, Department of Transportation,
September 1975.
Clark, W. H., Jr. A classification of malignant melanoma in man cor-
related with histogenesis and biologic behavior. IN: Advances
in Biology of Skin, vol. 8, The Pigmentary System (W. H. Montagna,
ed.), Oxford, Pergamon Press, 1967, pp 621-647.
F-l-177
-------
Effects—Skin Cancer
Clark, W. H., Jr., From, L., Bernardino, E. A. et al. The histogenesis
and biologic behavior of primary human malignant melanoma of the
skin. Cancer Res. 29:705-726, 1969.
Clark, W. H. Jr., Mastrangelo, M. J. et al. Current concepts of the
biology of human cutaneous melanoma. Advances in Cancer Res. 24:
267-338, 1977.
Clark, W. K. Jr. and Mihm, M. C. Lentigo maligna and lentigo maligna
melanoma. Am. J. Pathol. 55:39-67, 1969.
Clark, W. H. Jr., Mihm, B. H. and Kligman, L. H. The developmental
biology of induced malignant melanoma in guinea pigs and a com-
parison with other neoplastic systems. Cancer Res. 36:4079-4091,
1976.
Clark, W. H. Jr., Reiner, R. R., Greene, M. et al. Origin of familial
malignant melanoma from heritable melanocytic lesions. Arch.
Dermatol. 114:732-738, 1978.
Coblentz, W. W. and Stair, R. U.S. Bureau of Standards J. Res. 12:13-
14, 1934.
Coleman, W. P., Davis, R. S., Reed, R. J. and Krementz, E. J. Treatment
of lentigo maligna and lentigo maligna melanoma. J. Derm. Surg.
and Oncol. 6:476-479, 1980.
Coleman, W. P., Lorig, P. R. , Reed, R. J. and Krementz, E. J. Acral
lentiginous melanoma. Arch. Dermatol. 116:773-776, 1980.
Cosman, B., Heddle, S. B. and Crikelsir, G. F. The increasing incidence
of melanoma. Plastic and Reconstructive Surg. 57:50-56, 1976.
Cripps, D. J. and Ramsay, C. A. Ultraviolet action spectrum with a
prism-grating monochromator. Brit. J. Dermatol. 82:584-592, 1970.
F-l-178
-------
Effects—Skin Cancer
Crombie. 1. K. Variation of melanoma with latitude in North America
and Europe. Brit. J. Cancer 40:774-781, 1979.
Crutzen, P. J., Isaksen, I. S. A. and McAfee, J. R. The impact of
the chlorocarbon industry on the ozone layer. J. Geophys. Res.
83:345-363, 1978.
.Curtis, G. L. Initiation-promotion, skin carcinogenesis and immuno-
logical competence. Proc. Soc. Exp. Biol. Med. 150:61-64, 1975.
Cutchis, P. On the Linkage of- Solar Ultraviolet Radiation to Skin
Cancer. Institute for Defense Analysis. Washington, D.C., U.S.
Department of Transportation, 1978, pp 105-106.
Cutler, S. J. and Young, J. L. Third National Cancer Survey: Incidence
Data. NCI Monograph 41, DREW Publication No. (NIH)75-787. Bethesda
National Cancer Institute, 1975.
Daniels, F. Jr., Post, P. W. and Johnson, B. E. Theories of the role
of pigment in the evolution of human races. IN: Pigmentation:
Its Genesis and Biologic Control (V. Riley, ed.), New York, Apple-
ton-Century-Crofts, 1972.
Davies, J. N. P., Tauk, R. , Meyer, R. and Thursten, P. Cancer of the
integumentary tissues in Ugandan Africans. J. Natl. Cancer Inst.
41:31-51, 1968.
Davies, R. E. , Dodge, H. A. and Austin, W. A. (a) Carcinogenicity of
DMBA under various light sources. Proc. 9th Intl. Cong, of Photo-
biol., p. 247, 1972 (abstract)
Davies, R. E. , Dodge, H. A. and DeShields, L. H. (b) Alteration of
the carcinogenicity of DMBA by light. Proc. Am. Assoc. Cancer
Res. 13:14, 1972 (abstract)
F-l-179
-------
Effects—Skin Cancer
Davis, N. C. Malignant melanoma in Queensland. Austral. J. Dermatol.
19:13-18, 1978.
Davies, N. C. William Morris, M.D. , a pioneer in the study of melanoma.
Med. J. Austral. 1:52-54, 1980.
DeGruijl, F. R. and van der Leun, J. C. A dose-response model for
skin cancer induction by chronic UV exposure of a human population.
J. Theor. Biol. 83:487-504, 1980.
DeLuca, H. F. Vitamin D: A new look at an old vitamin. Nutr. Rev.
29:179-181, 1971.
Department of Health, Education and Welfare. Second Biennial Report
to the Congress on Research Activities of Relevance to the Clean
Air Act. Bethesda, National Cancer Institute, 1980.
Dickinson, R. E., Liu, S. C. and Donohue, J. M. Effect of chlorofluoro-
methane infrared radiation on zonal atmospheric temperature.
J. Atmos. Sci. 35:2142-2151, 1978.
Doll, R., Muir, C. and Waterhouse, J. Cancer Incidence in Five Con-
tinents, Vol. III. International Union Against Cancer, IARC #15,
1976.
Dubreuilh, W. Des Hyperkeratoses circonscrites. Ann. Derm. Syph.
(Ser. 3)7:1158-1204, 1896.
Dubreuilh, M. W. De la melanose circonscrite precancereuse. Ann.
Derm. Syph. (Paris) 3:129-151, 205-230, 1912.
DuPont Company, Submission to the Environmental Protection Agency,
"Comments on the December 1979 Report by the Committee on the
Impacts of Stratospheric Change (CISC) of the National Academy
of Sciences." May 1980.
Eastcott, D. F. Epidemiology of skin cancer in New Zealand. IN:
Monograph #10, The Biology of Cutaneous Cancer. Bethesda, National
Cancer Institute, 1963.
F-l-180
-------
Effects—Skin Cancer
Edgcomb, J. H. and Mitchelich, H. Melanomas of the skin of guinea
pigs following the application of a solution of DMBA in benzene.
Acta Union International Centre Cancrum 19:706-707, 1963.
Eklund, G. and Malec, B. Sunlight and incidence of cutaneous malignant
melanoma. Scand. J. Plast. Reconstr. Surg. 12:231-241, 1978.
Elwood, J. M. and Lee, J. A. H. Recent data on the epidemiology of
malignant melanoma. Semin. Oncol. 2:149, 1975.
Emmett, E. A. Ultraviolet radiation as a cause of skin tumors. CRC
Crit. Rev. Toxicol. 2:211, 1974.
Epstein, J. H. Comparison of the carcinogenic and cocarcinogenic ef-
fects of UV light on hairless mice. J. Natl. Cancer Inst. 34:741-
745, 1965.
Epstein, J. H. and Epstein, W. L. A study of tumor types produced
by UV light in hairless and hairy mice. J. Invest. Dermatol.
41:463-473, 1963.
Epstein, J. H., Epstein, W. L. and Nakai, T. Production of melanomas
from DMBA-induced "blue nevi" in hairless mice with ultraviolet
light. J. Natl. Cancer Inst. 30:19-30, 1967.
Epstein, J. H. , Fukuyama, K. and Dobson, R. Ultraviolet light carcino-
genesis. IN: The Biologic Effects of UV Radiation (F. Urbach,
ed.), Oxford, Pergamon Press, 1969, pp 551-568.
Epstein, J. H. and Roth, H. L. Experimental ultraviolet light carcino-
genesis: A study of croton oil promoting effects. J. Invest.
Dermatol. 50:387-389, 1968.
F-l-181
-------
Effects—Skin Cancer
Everett, M. A., Bell, R. and Hagans, J. A. Melanogenesis and skin
cancer. Arch. Dermatol. 82:908, 1960.
Everett, M. A., Yeargens, E., Sayre, R. M. and Olson, R. L. Penetration
of epidermis by ultraviolet rays. Photochem. Photobiol. 5:533-
542, 1966.
Fears, T. R. , Scotto, J. and Schneiderman, M. A. Skin cancer, melanoma
and sunlight. Am. J. Publ. Health 66:461-464, 1976.
Findlay, G. M. Ultraviolet light and skin cancer. Lancet 2:1070-1073,
1928.
Fisher, M. S. and Kripke, M. L. Systemic alteration induced in mice
by UV light irradiation and its relationship to UV carcinogenesis.
Proc. Natl. Acad. Sci. 74:1688-1692, 1977.
Forbes, P. D. , Davies, R. E. and Urbach, F. Phototoxicity and photo-
carcinogenesis: Comparative effects of anthracene and 8-methoxy-
psoralen in the skin of mice. Food Cosmet. Toxicol. 14:243, 1970.
Forbes, P. D. , Davies, R. E. and Urbach, F. Experimental ultraviolet
photocarcinogenesis: Wavelength interactions and time-dose rela-
tionships. IN: NCI Monograph #50, 1978.
Forbes, P. D., and Urbach, F. Experimental modification of photocar-
cinogenesis. I. Fluorescent whitening agents and shortwave UVR.
Food and Cosmet. Toxicol. 13:335-337, 1975.
Foulds, L. Neoplastic Development. New York, Academic Press, 1969,
vol. 1, p 75.
Freeman, R. G. Data on the action spectrum for ultraviolet carcino-
genesis. IN: NCI Monograph #50, 1978.
F-l-182
-------
Effects—Skin Cancer
Freeman, R. G. and Knox, J. M. Ultraviolet-induced corneal tumors
in different species and strains of animals. J. Invest. Dermatol.
43:431-436, 1964.
Gellin, G. A., Kopf, A. W. and Andrade, R. Basal cell epitheliomas:
A controlled study of associated factors. Arch. Dermatol. 91:38-
45, 1965.
Gellin, G. A., Kopf, A. W. and Garfinkel, L. Malignant melanoma.
A controlled study of possibly associated factors. Arch. Dermatol.
99:143-148, 1969.
Giese, A. C. Photophysiology, volume 4. New York, Academic Press,
1965, pp 139-202.
Gordon, D. and Silverstone, H. Worldwide epidemiology of premalignant
and malignant cutaneous lesions. IN: Cancer of the Skin (R.
Andrade, ed.), Philadelphia, W. B. Saunders, 1976, pp 405-455.
Grady, H. G., Blum, H. F. and Kirby-Smith, J. S. Types of tumors in-
duced by ultraviolet radiation and factors influencing their rela-
tive incidence. J. Natl. Cancer Inst. 3:371-378, 1943.
Graham, J. H. and Helwig, E. B. IN: Advances in Biology of the Skin,
vol. 2 (W. Montagna, ed.), Oxford, Pergamon Press, 1965, pp 277-
327.
Green, A. E. S. Ultraviolet exposure and skin cancer response. Am.
J. Epidemiol. 107:277-280, 1978.
Green, A. E. S., Findley, G. B. , Klenk, K. F., Wilson, W. M. and Mo, T.
The ultraviolet dose dependence of nonmelanoma skin cancer inci-
dence. Photochem. Photobiol. 24:353-362, 1976.
F-l-183
-------
Effects—Skin Cancer
Green, A. E. S. and Mo, T. Proc. Third Conf. on CIAP. DOT-TSC-OST-
74-15, pp 518-522, 1974.
Green, A. E. S., Mo, T. and Miller, J. H. A study of solar erythema
radiation doses. Photochem. Photobiol. 20:473-482, 1974.
Griffin, A. C. , Dolman, V. S. et al. The effect of visible light on
the carcinogenicity of ultraviolet light. Cancer Res. 15:5.23,
1955.
Haenszel, W. Variations in skin cancer incidence within the U.S.
NCI Monograph 10:225-243, 1963.
Hampson, J. Photochemical war on the atmosphere. Nature 250:189-191,
1974.
Hausser, K. W. and Vahle, W. Sonnenbrand und Sonnenbrauniing. Wiss.
Veroff. Siemenskonzern 6:101, 1927.
Haverkamp, J. and Radman, 0. G. Malignant melanoma of the foot in
black patients. J. Natl. Med. Assoc. 71:353-356, 1979.
Haxthausen, H. and Hausman, W. Die Lichterkrankungen der Haut. Urban
und Schwarzenberg, Vienna, 1929.
Herlitz, C. W. , Jundell, I. and Wahlgren, F. Malignant tumors in white
mice caused by ultraviolet radiation. Acta Paediat. 10:333-347,
1930.
Herskowitz, L. J. and Dobes, W. Photoprotective qualities of clothing.
In press
Higginson, J. and Oettle, A. G. Cancer incidence in the Bantu and
"Cape colored" races of South Africa. J. Natl. Cancer Inst. 24:
5389, 1960.
F-l-184
-------
Effects—Skin Cancer
Hill, L. and Eidenow, A. Biological action of light. I. Influence
of temperature. Proc. Royal Soc. (Biol.) 95:163-180, 1923.
Hinds, M. W. and Kolonel, L. N. Malignant melanoma of the skin in
Hawaii, 1960-1977. Cancer 45:811-817, 1980.
Holman, C. D. D. , James, I. R. , Galley, P. H. and Armstrong, B. K.
An analysis of trends in mortality of skin from malignant melanoma
of the skin in Australia. In press, 1980.
Holman. C. D. J., Mulroney, C. D. and Armstrong, P. K. Epidemiology
of pre-invasive and invasive malignant melanoma in Australia.
Int. J. Cancer 25:317-323, 1980.
Houghton, A., Flannery, J. and Viola, M. V. Malignant melanoma in
Connecticut and Denmark. Int. 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
1:759-760, 1978.
Hueper, W. C. Cutaneous neoplastic responses elicited by ultraviolet
rays in hairless rats and in their haired litter mates. Cancer
Res. 1:402-406, 1941.
Hyde, J. On the influence of light in the production of cancer of
the skin. Am. J. Med. Sci. 131:1-22, 1906.
Isaacson, C., Selzer, G., Kaye, V. et al. Cancer in urban blacks of
South Africa. S. African Cancer Bull. 22:49, 1978.
Jensen, 0. M. and Bolander, A. M. Trends in malignant melanoma of
the skin. World Health Statistics Quart. 33:2-26, 1980.
F-l-185
-------
Effects—Skin Cancer
Jessup, J. M. , Hanna, N. et al. Mechanisms of depressed reactivity
to DNCB and UVR induced tumors during UV carcinogenesis in BALB/c.
Cell. Immune 1. 38:105-115, 1978.
Johnson, B. E. Ultraviolet radiation and lysosomes in skin. Nature
(Lond.) 219:1258-1259, 1968.
Johnson, B. E., Daniels, F. and Magnus, I. A. Response of human skin
to ultraviolet light. IN: Photophysiology, vol. IV (A. C. Giese,
ed.), New York, Academic Press, 1968.
Kelner, A. and Taft, E. B. The influence of photoreactivating light
on the type and frequency of tumors induced by UV radiation. Cancer
Res. 16:860-866, 1956.
Kirby-Smith, J. S. , Blum, H. F. and Grady, H. G. Penetration of ultra-
violet radiation into skin, as a factor in carcinogenesis. J.
Natl. Cancer Inst. 2:403-412, 1942.
Klepp, 0. and Magnus, K. Some environmental and bodily characteristics
of melanoma patients. Int. J. Cancer 23:482, 1979.
Kripke, M. L. Speculations on the role of ultraviolet radiation in
the development of malignant melanoma. J. Natl. Cancer Inst. 63:
541-547, 1979.
Kripke, M. L. and Fisher, M. S. Effect of UV light on the host response
to UV induced tumors. Proc. Am. Soc. Photobiol., 1976 (abstract)
Kripke, M. L. and Lill, P. H. Unpublished data. Cited by Kripke,
M. L., J. Natl. Cancer Inst. 63:544, 1979.
Kripke, M. L. , Lofgren, J. S. et al. In vivo immune responses of mice
during carcinogenesis by UV radiation. J. Natl. Cancer Inst.
59:1227-1230, 1977.
F-l-186
-------
Effects—Skin Cancer
Lancaster, H. 0. Some geographical aspects of the mortality from mela-
noma in Europeans. Med. J. Austral. 1:1082, 1956.
Lane-Brown, M. M. and Melia, D. F. A genetic diathesis to skin cancer.
J. Invest. Dermatol. 61:39-41, 1973.
Lane-Brown, M. M. , Sharpe, C. A. B. , MacMillan, D. S. and 'McGovern,
V. J. Genetic predisposition to melanoma and other skin cancers
in Australians. Med. J. Austral. 1:852, 1971.
Larsen, T. and Crude, T. H. A retrospective histological study of
669 cases of primary malignant melanoma in clinical stage I.
I. Histologic classification, sex and age of patients, localiza-
tion of tumors and prognosis. Acta Path. Microbiol. Scand. Sect.
A. 86:473-480, 1978.
Larsen, T. and Crude, T. H. A retrospective histological study of
669 cases of primary malignant melanoma in clinical stage I.
VI. The relation of dermal solar elastosis to sex, age and
survival of the patient, and to localization, histological type
and level of invasion of the tumor. Acta Path. Microbiol. Scand.
Sect. A. 87:361-366, 1979.
Leach, J. F., Beadle, P. C. and Pingstone, A. R. Effect of ozone varia-
tion on disease in Great Britain. I. Skin cancer. Aviation,
Space and Environ. Med. 49:512-516, 1978.
Lee, J. A. H. Sunlight and the etiology of melanoma. IN: Melanoma
and Skin Cancer, Sydney, New South Wales Government Printer, 1972,
pp 83-94.
Lee, J. A. H. Current evidence about the causes of malignant melanoma.
Prog, in Clin. Cancer VI:151-161, 1975.
F-l-187
-------
Effects—Skin Cancer
Lee, J. A. H. Current evidence about the causes of malignant melanoma.
IN: Clinical Cancer (Ariel, ed.), New York, Grune and Stratton,
1977, p 151
Lee, J. A. H. and Carter, A. P. Secular trends in mortality from malig-
nant melanoma. J. Natl. Cancer Inst. 45:91-97, 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. Brit. 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 malig-
nant melanoma in the populations of England and Wales, Canada
and the white population of the U.S. Am. J. Epidemiol. 110:734-
739, 1979.
Lee, J. A. H. and Merrill, J. M. Sunlight and melanoma. Lancet 1:550
1971.
Lee, J. A. H. and Storer, B. E. Excess of malignant melanomas in women
in the British Isles. In press, 1980.
Lee, J. A. H. and Strickland, D. Malignant melanoma: Social status
and outdoor work. Brit. J. Cancer 41:757-763, 1980.
Lewis, M. G. Malignant melanoma in Uganda. Brit. J. Cancer 21:483-
495, 1967.
Lipson, R. L. and Baldes, E. J. Photosensitivity and heat. Arch.
Dermatol. 82:517-520, i960.
Little, J. H. Histology and prognosis in cutaneous melanoma. IN:
Melanoma and Skin Cancer, Sydney, New South Wales Government
Printer, 1972.
-------
Effects—Skin Cancer
Little, J. H., Holt, I. and Davis, N. Changing epidemiology of malig-
nant melanoma in Queensland. Med. J. Austral. 1:66-69, i960.
London, J. The observed distribution and variation of total ozone.
IN: Proc. NATO Advanced Study Institute on Atmospheric Ozone.
U.S. Department of Transportation, FAA-HAPP, Washington, D.C.,
pp 31-44.
London, J. and Reber, C. A. Solar activity and total atmospheric ozone.
Geophys. Res. Lett. 6:869-872, 1979.
Loomis, W. F. Skin pigmentation regulation of vitamin D synthesis
in man. Science 157:501-506, 1967.
Lund, R. H. and Ihnen, M. Malignant melanoma. Surgery 38:652-659,
1955.
Luckiesh, M. , Holladay, L. L. and Taylor, A. M. Erythemal and tanning
effectiveness of UV energy. Gen. Electric Rev. 42:274-278, 1939.
MacDonald, E. J. Malignant melanoma in Connecticut. IN: The Biology
of Melanoma, New York Academy of Sciences, 1948, pp 71-81.
MacDonald, E. J. Assessment of possible SST effects on the incidence
of skin cancer. Tech. Report, Department of Atmospheric Physics,
University of Arizona, 1971.
MacDonald, E. J. and Bubendorf, E. Some epidemiologic aspects of skin
caner. IN: Tumors of the Skin, Yearbook Medical Publisher, 1964,
pp 23 ff.
Mackie, B. S. and McGovern, V. J. The mechanism of solar carcino-
genesis: A study of the role of collagen degeneration of the
dermis in the production of skin cancer. Arch. Dermatol. Syph.
78:218-244, 1958.
F-l-189
-------
Effects—Skin Cancer
Malec, E. and Eklund, G. The changing incidence of malignant melanoma
of the skin in Sweden 1959-1968. Scand. J. Plastic Reconstr.
Surg. 12:19-27, 1978.
Magnus, I. A. and Johnson, B. E. Cited by Johnson, B. E. et al. IN:
Photophysiology, vol. 4 (A. C. Giese, ed.), New York, Academic
Press, 1965, pp 139-202.
Magnus, K. Incidence of malignant melanoma of the skin in Norway 1955-
1970. Cancer 32:1273-1275, 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. Excerpta Medica
Congress Series #375, 1975.
Magnus, K. Incidence of malignant melanoma of the skin in the five
Nordic countries. Int. J. Cancer 20:477-485, 1977.
Magnus, K. Habits of sun exposure and risk of malignant melanoma.
Cancer, 1980, in press.
i
McCarthy, W. H. , Black, A. L. and Milton, G. W. Melanoma in New South
Wales. Cancer 46:427-432, 1980.
McGovern, V. J. The classification of melanoma and its relationship
with prognosis. Pathology 2:85-98, 1970.
McGovern, V. J. Malignant Melanoma. Clinical and Histologic Diagnosis.
New York, John Wiley, 1976.
McGovern, V. J. Epidemiologic aspects of melanoma: A review. Path-
ology 9:233-241, 1977.
McGovern, V. J. and Lane-Brown, M. M. The Nature of Melanoma. Spring-
field, C. C. Thomas, 1969.
F-l-190
-------
Effects—Skin Cancer
McGovern, V. J. and Mackie, B. S. The relationship of solar radiation
to melanoblastoma. Aust. N.Z. Surg. 28:357, 1959.
McGovern, V. J., Mihm, M. C. , Bailly, C. et al. The classification
of malignant melanoma. I. Its histologic reporting. Cancer 32:
1446-1457, 1973.
McKnight, C. K. and Magnusson, B. Tumors in Iceland. I. Malignant
tumors of the skin. Acta Path. Microbiol. Scand. Sect. A 87:37-
44, 1979.
Mehnert, J. H. and Heard, J. Staging of malignant melanoma by depth
of invasion. A proposed index to prognosis. Am. J. Surg. 110:168-
176, 1965.
Millikan, L. E. , Boylan, J. L., Hook, R. R. et al. Melanoma in Sinclair
swine: A new animal model. J. Invest. Dermatol. 62:20-30, 1974.
Mo, T. and Green, A. E. S. A climatology of solar erythema dose.
Photochem. Photobiol. 20:483-496, 1974.
Molesworth, E. H. Rodent ulcer. Med. J. Australia (Suppl), 1927, p
878.
Movshovitz, M. and Modan, B. Role of sun exposure in the etiology of
malignant melanoma. Epidemiological Conference. J. Natl. Cancer
Inst. 51:77, 1973.
Nakamura, K. and Johnson, W. C. Ultraviolet light induced connective
tissue changes in rat skin. J. Invest. Dermatol. 51:253-258, 1968.
Nathanson, R. B. , Forbes, P. D. and Urbach, F. UV photocarcinogenesis:
Modification of anti'lymphocytic serum or 6-mercaptopurine. Proc.
Am. Assoc. Cancer Res. 14:46, 1973 (abstract)
Nathanson, R. B. , Forbes, P. D. and Urbach, F. Modification of photo-
carcinogenesis by two immunosuppressive agents. Cancer Letters
1:243-247, 1976.
F-l-191
-------
Effects—Skin Cancer
NAS (National Academy of Sciences). Biological Impact of Increased
Intensities of Solar Ultraviolet Radiation. Washington, D.C.,
1973.
NAS (National Academy of Sciences). Environmental Impact of Strato-
spheric Flight: Biological and Climatic Effects of Aircraft Emis-
sions in the Stratosphere. Washington, D.C., 1975.
NAS (National Academy of Sciences). Report of the Committee on Im-
pacts of Stratospheric Change (CISC): Protection Against Depletion
of Stratospheric Ozone by Chiorofluorocarbons. Washington, D.C.
1979.
Nicolle, F. V., Mathews, W. H. and Palmer, J. D. Malignant melanoma
of the skin. Canad. J. Surg. 3:233-237, i960.
Nobil, El-Bokainy and Ebeid, A. M. Malignant melanoma in Egypt. Tumori
59:429-436, 1973.
Norbury, K. C. , Kripke, M. L. and Budman, M. B. In vitro reactivity
of macrophages and lymphocytes from UV irradiated mice. J. Natl.
Cancer Inst. 59:1231-1235, 1977.
Norris, W. Eight Cases of Melanosis with Pathological and Therapeutic
Remarks on That Disease. Langman, Brown, Green, Longman and
Roberts, London, 1857.
O'Beirn, S. F. , Judge, P., Urbach, F. et al. Skin cancer in county
Galway, Ireland. Proc. 10th Intl. Cancer Conference, Am. Cancer
Soc., 1968.
Oettle, C. H. Skin cancer in Africa. IN: Monograph 10, National
Cancer Institute, The Biology of Cutaneous Cancer. 1963, pp 197-
214.
O'Halloran, M. J. Skin cancer in Ireland. J. Irish Med. Assoc. 60:209-
213, 1967.
F-l-192
-------
Effects—Skin Cancer
Olson, R. L. , Sayre, R. M. and Everett, M. A. Effect of anatomic loca-
tion and time on ultraviolet erythema. Arch. Dermtol. 93:211-215,
1966.
Orr, J. W. The changes antecedent to tumor formation during the treat-
ment of mouse skin with carcinogenic hydrocarbons. J. Path. Bact.
46:495-515, 1938.
Owens, D. W. , Knox, J. H. et al. The influence of wind on chronic
ultraviolet light-induced carcinogenesis. Brit. J. Dermatol. 97:
285, 1977.
Pawlowski, A., Habermann, H. F. and Menon, I. A. Skin melanoma induced
by DMBA in albino guinea pigs and its similarity to skin melanoma
in humans. Cancer Res. 40:3652-3660, i960.
Pell, S., O'Berg, M. T. and Karrh, B. W. Cancer epidemiologic surveil-
lance in the DuPont Company. J. Occup. Med. 20:725-740, 1978.
Pitts, D. G. Amer. J. Optom. 47:535-546, 1970.
Price, N. M., Rywlin, A. M. and Ackerman, A. B. Histologic criteria
for the diagnosis of superficial spreading melanoma. Cancer 38:
2434-2441, 1970.
Putschar, W. and Holtz, F. Erzeugung von Hautkrebsen bei Ratten durch
Langedaurende Ultraviolett Bestrahlung. Ztschr. f. Krebsforsch.
33:219-260, 1930.
Pyle, J. A. and Derwent, R. G. Possible ozone reductions and UV changes
at the earth's surface. Nature 286:373-375, 1980.
Regan, J. and Set low, R. B. Two forms of repair in the DNA of human
cells damaged by chemical carcinogens and mutagens. Cancer Res.
34:3318-3325, 1974.
F-l-193
-------
Effects—Skin Cancer
Reiner, R. R. , Clark, W. H. Jr., Green, M. H. , Ainsworth, A. M. and
Fraumeni, J. F. Precursor lesions in familial melanoma. J. Am.
Med. Assoc. 239:744-746, 1978.
Ringentz, N. Cancer incidence in Finland, Iceland, Norway and Sweden.
Acta Path. Microbiol. Scand. Sect. A. Suppl. 224, 1971.
Robertson, D. F. (Thesis) Solar ultraviolet radiation in relation to
sunburn and skin cancer. University of Queensland, 1972.
Roffo, A. H. Cancer y Sol. Boll. Inst. de Med. Exper. para el estuda
y trata del Cancer. 10:417-439, 1933.
RosDahl, I. The epidermal melanocyte population and its reaction to
ultraviolet light. Acta Dermato-Venereologica Suppl. 88, 1979.
Rothman, R. H. and Setlow, R. B. An action spectrum for cell killing
and pyrimidine dimer formation in Chinese hamster V-79 cells.
Photochem. Photobiol. 29:57-61, 1979.
Rundel, R. D. and Nachtwey, D. S. Skin cancer and UV radiation. Photo-
chem. Photobiol. 28:345-356, 1978.
Rusch, H. P., Kline, B. Z. and Bauman, C. A. Carcinogenesis by UV rays
with reference to wavelength and energy. Arch. Pathol. 371:135-
146, 1941.
Sams, W. M. Jr., Smith, J. G. and Burk, P. G. The experimental pro-
duction of elastosis with ultraviolet light. J. Invest. Dermatol.
43:467, 1964.
Schulze, R. and Gra'fe, K. Consideration of sky ultraviolet radiation
in the measurement of solar UVR. IN: The Biologic Effects of
Ultraviolet Radiation (F. Urbach, ed.), Oxford, Pergamon Press,
1969, pp 369-373.
F-l-194
-------
Effects—Skin Cancer- I
Scott, E. L. and Straf, M. L. Ultraviolet radiation as a cause of can-
cer. IN: Origins of Human Cancer. Cold Springs Harbor Lab.,
1977, pp 529-546.
Scotto, J. , Fears, T. R. and Gori, G. B. Measurements of ultraviolet
radiation in the U.S. and comparison to skin cancer data. U.S.
DREW No. (NIH)76-1092, Washington, D.C., 1975.
Scotto, J., Fears, T. R. , et al. Incidence of nonmelanoma skin cancer
in the United States 1977-78. DHEW Publication No. (NIH)80-2154.
U.S. DHEW, National Cancer Institute, April 1980.
Scotto, J., Kopf, A. W. and Urbach, F. Nonmelanoma skin cancer in four
areas of the U.S. Cancer 34:1333-1338, 1974.
Scotto, J. and Nam, J. M. Skin melanoma and seasonal patterns. Am.
J. Epidemiol. 111:309-314, 1980.
Setlow, R. B. The relevance of photobiological repair. Ann. Acad.
Bros. Ciencas 45:215-220, 1973.
Setlow, R. B. The wavelengths in sunlight effective in producing skin
cancer: A theoretical analysis. Proc. Natl. Acad. Sci. 71:3363-
3366, 1974.
Shaw, H. M. , McCarthy, W. H. and Milton, G. W. Changing trends in mor-
tality from malignant melanoma. Med. J. Austral. 64:77-78, 1977.
Shield, A. M. A remarkable case of multiple growths of the skin caused
by exposure to the sun. Lancet 1:22-23, 1899.
Silverstone, H. Skin cancer in Queensland, Australia. IN: Report
of the Airlie House Conference, Monograph #50, National Cancer
Institute. DHEW Publication No. (NIH)78-1532, pp 230-232, 1978.
Silverstone, H. and Searle, J. H. A. The epidemiology of skin cancer
in Queensland. Brit. J. Cancer 24:235-253, 1970.
F-l-195
-------
Effects—Skin Cancer
Skibba, J. L. , Pinckley, J. et al. Multiple primary melanoma following
administration of levodopa. Arch. Pathol. 93:556-561, 1972.
Smith, K. C. The cellular repair of radiation damage. IN: Sunlight
and Man (T. B. Fitzpatrick et al., eds.), Tokyo, University of
Tokyo Press, 1974.
Sober, A. J., Blois, M. S. , Clark, W. H. Jr., Fitzpatrick, T. B., Kopf,
A. W. and Mihm, M. C. Primary malignant melanoma of the skin.
IN: Dermatology, Proc. 15th Intl. Cong., Mexico 1977. Excerpta
Medica, Intl. Cong, series #451, 1979.
Sober, A. J. , Lew, R. A., Fitzpatrick, T. B. and Marvell, R. Solar
exposure patterns in patients with cutaneous melanoma. Clin. Res.
28:561A,1980.
Sober, A. J. and Wick, M. M. Levodopa therapy and malignant melanoma.
J. Am. Med. Assoc. 240:554-555, 1978.
Stein, D. , Thind, I. S. and Louria, D. B. Melanoma of the skin in New
Jersey. J. of the Med. Soc. of New Jersey 75:391-393, 1978.
Stenback, F. Promotion in the morphogenesis of chemically inducible
skin tumors. Acta Path. Microbiol. Scand. Suppl. 208:1-116, 1969.
Stenback, F. (a) Cellular injury and cell proliferation in skin car-
cinogenesis by UV light. Oncology 31:61-65, 1975.
Stenback, F. (b) Species-specific neoplastic progression by ultraviolet
light on the skin of rats, guinea pigs, hamsters and mice. On-
cology 31:209-225, 1975.
Stenback, F. Life history and histopathology of ultraviolet light in-
duced skin tumors. IN: Intl. Conf. on UV Carcinogenesis. Mono-
graph #50, National Cancer Institute, 1978, pp 37-70.
F-l-196
-------
Effects—Skin Cancer
Stief, L. J., Michael, J. V. et al. Geophys. Res. Lett. 5:829, 1978.
Teppo, L. , Hakema, M., Hakulinen, T., Lehtonen, M. and Saxen, E. Cancer
in Finland. Acta Path. Microbiol. Scand. Sect. A, Suppl. 252,
1975.
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. , Hakulinen, T., Herva, A. and Saxen,
E. Way' of life and cancer incidence in Finland: A municipality
based ecological analysis. Scand. J. Social Med. Suppl. 19:50-
54, 1980.
Thomas, L. cited in Peto, R. Distorting the epidemiology of cancer:
The need for a more balanced overview. Nature 284:297-300, 1980
Toth, B. Susceptibility of guinea pigs to chemical carcinogens DMBA
and urethane. Cancer Res. 30:2583-2589, 1970.
Trosko, J. E. and Chu,H. E. Inhibition of repair of UV damaged DNA
by caffeine and mutation induction in Chinese hamster cells. Chem.
Biol. Interactions 6:317-332, 1973.
Unna, P. Histopathologie der Hautkrankheiten. August Hirschwald,
Berlin, 1894.
Urbach, F. (ed.) Monograph #10, National Cancer Institute, The Biology
of Cutaneous Cancer, 1963.
Urbach, F. Geographic pathology of skin cancer. IN: The Biologic
Effects of Ultraviolet Radiation (F. Urbach, ed.), Oxford, Pergamon
Press, 1969.
Urbach, F. and Davies, R. E. Estimate of ozone reduction in the strato-
sphere on the incidence of skin cancer in man. IN: CIAP Monograph
5, DOT-TST-75-55, National Technical Information Service, Spring-
field, VA, 1975.
F-l-197
-------
Effects—Skin Cancer
Urbach, F., Rose, D. B. and Bonnem, M. Genetic and environmental inter-
actions in skin carcinogenesis. IN: Environment and Cancer,
Williams and Wilkins, Baltimore, 1972, pp 355-371.
van der Leun, J. C. and Daniels, F. Jr. Biologic effects of strato-
spheric ozone decrease. A critical review of assessment. IN:
CIAP Monograph 5, Appendix B, p 7-105. DOT-TST-75-55, National
Technical Information Service, Springfield, VA, 1975.
Van Scott, E. J. , Reinertson, R. P. and McCall, C. B. The growing pre-
valence, histologic type and significance of palmar and plantar
nevi. Cancer 1:363-367, 1957.
Vesselinovitch, S. D., Mikhailovich, N. and Richter, W. R. The induc-
tion of malignant melanomas in Syrian white hamster by neonatal
exposure to urethane. Cancer Res. 30:2543-2547, 1970.
Vielkind, N. Genetic control of cell differentiation in platy fish
sword tail melanoma. J. Exp. Zool. 196:197-204, 1976.
Viola, M. V. and Houghton, A. Melanoma in Connecticut. Conn. Med.
42:268-269, 1978.
Vitaliano, P. P. and Urbach, F. The relative importance of risk factors
in nonmelanoma carcinoma. Arch. Dermatol. 116:454-456, 1980.
Wallace, D. C., Beardmore, G. H. and Exton, L. A. Familial malignant
melanoma. Ann. Surg. 177:15-20, 1973.
Wallace, D. C. and Exton, L. A. Genetic predisposition to development
of malignant melanoma. 1972.
Wanebo, H. J., Woodruff, J. and Fortner, J. G. Malignant melanoma of
the extremities. Cancer 35:666, 1975.
Wayte, D. M. and Helwig, E. G. Melanotic freckle of Hutchinson. Cancer
21:893-911, 1968.
F-l-198
-------
Effects—Skin Cancer
Wigle, D. T. Malignant melanoma of skin and sunspot activity. Lancet
2:38, 1978.
Word, W. H. Melanoma, carcinoma of the skin and sunlight. Australasian
J. Dermatol. 9:70-75, 1967.
Winkelman, R. K. , Baldes, E. J. and Zollman, P. E. Squ'amous cell tumors
induced in hairless mice with UV light. J. Invest. Dermatol. 34:
131-138, 1960.
Xenophon (translated by J. S. Watson). The Anabasis. London, Henry
G. Bohn, 1885, pp 126-127.
Zackheim, H. S. Origin of the human basal cell epithelioma. J. Invest.
Dermatol. 40:283-297, 1963.
Zilov, J. N. D. IN: Ultraviolet Radiation, Moscow, Medicina, 1971,
pp 237-241.
F-l-199
-------
COMMENTS ON THE NATIONAL ACADEMY
OF SCIENCES REPORT: "PROTECTION AGAINST
DEPLETION OF STRATOSPHERIC OZONE BY
CHLOROFLUOROCARBONS."
PREPARED FOR E. I. DU PONT DE NEMOURS & COMPANY, INC,
BY
DR. W. H. KLEIN
DIRECTOR
SMITHSONIAN INSTITUTION RADIATION BIOLOGY LABORATORY
ROCKVILLE, MARYLAND
DECEMBER, 1980
(APPENDIX F-2)
F-2-1
-------
Transmittal letter for
Appendix F-2
+ ' 12291543-45605-0849R
**TLX MSG RCVD**
WASHINGTON DC 12-29-80
SMITHSONIAN INSTITUTION
FROM WILLIAM H KLEIN DIRECTOR
SMITHSONIAN RADIATION BIOLOGY LAB.
TO RICHARD^VAPD _ PETRO CHiMlCAL DEPT.
THIS WILL ACKNOWLEDGE MY COMMENTS ON THE NAS REPORT.
-------
Effects - UV-B Measurement
COMMENTS ON THE NATIONAL ACADEMY
OF SCIENCES REPORT: "PROTECTION AGAINST
DEPLETION OF STRATOSPHERIC OZONE BY
CHLOROFLUOROCARBONS."
TABLE OF CONTENTS
Page
1. INTRODUCTION , F-2-3
2. DIFFICULTIES IN THE MEASUREMENT OF UV-B
A. UV-B Measurement as an Experimental
Complication F-2-3
B. Laboratory Measurements . F-2-A
C. Field Measurements F-2-A
D. Weather Corrections F-2-A
E. Significance of UV-B Measurement Difficulties F-2-6
3. LATITUDINAL VARIATION OF UV-B AND ACTION SPECTRUM SELECTION
A. Projected UV-B Changes versus UV-B Changes
with Latitude F-2-6
B. Comparison of Measurements at Various Latitudes F-2-6
C. Calculating DNA Dose from Five Nanometer
Energy Integrals F-2-8
D. Action Spectrum Choice F-2-11
E. Action Spectra other than DNA F-2-11
F. Si-gnificance F-2-13
A. SUPPLEMENTATION OF UV-B WITH LAMPS
A. Experimental Limitations and Sources of Error F-2-15
B. Physical Characteristics of Lamps F-2-16
C. Experimental Factors in the Use of Lamps F-2-16
D. Significance of Difficulties in Supplementing
UV-B with Lamps F-2-16
5. REFERENCES F-2-19
F-2-2
-------
Effects - UV-B Measurement
COMMENTS BY DR. W. H. KLEIN ON THE NATIONAL ACADEMY OF
SCIENCES REPORT, "PROTECTION AGAINST DEPLETION OF STRATOSPHERIC
OZONE BY CHLOROFLUOROCARBONS."
1. INTRODUCTION
I have examined the National Academy of Sciences (NAS)
Report entitled, "Protection against Depletion of Stratospheric
Ozone by Chlorofluorocarbons" (NAS, 1979) with particular emphasis
concerning measurements of UV-B (285-320 nm) and techniques of UV-B
supplementation. The comments concerning this aspect of the report
are my opinion and can only be attributed to me and no other person
or organization.
The report, in its entirety, is a rather comprehensive and
very good review of the ozone and UV-B situation. However, the
effort to summarize the complex problems does oversimplify and
possibly results in statements that may be misleading.
The following comments (Sections 2-4) are intended to be
constructive and informative, so that future experiments can be
improved.
2. DIFFICULTIES IN THE MEASUREMENT OF UV-B
A. UV-B Measurement as an Experimental Complication
NAS acknowledged the difficulties in the measurement of UV
radiation:
"Measurement of the UV radiation itself introduces
additional complications, which were especially troublesome
in earlier experiments. While spectroradiometric
measurements (that is, absolute measurements of the
radiation per unit area and unit spectral bandpass at a
sufficient number of representative wavelengths) have been
made recently by most investigators, this was not done in
most of the earlier studies. Instead, simpler dosimeters
were utilized, such as the Robertson-Berger meter, which
characterizes a broad waveband with respect to a weighting
function that may not be biologically correct. The
uncertainty introduced by meters of this kind is discussed
in Appendix E." (NAS, 1979, p. 66).
However, definitive information about the magnitude of the
difficulties was not provided.
F-2-3
-------
Effects - UV-B Measurement
B. Laboratory Measurements
The best spectral irradiance measurements that can be made
today between 250 and 350 nm are accurate to about 3 to 5 percent in
a rigidly controlled laboratory measurement. Depending on what kind
of standard source is used, calibration accuracy of UV-B instruments
can vary from less than 5 percent to about 10 percent (National
Bureau of Standards, 1977). Argon mini arcs have an uncertainty in
their irradiance of an estimated 6 percent for wavelengths greater
than~140 nm. The National Bureau of Standards (NBS) synchrontron UV
radiation facility can calibrate'spectrometer and photometer units
to better than 5 percent.
C. Field Measurements
The measurement of UV-B radiation under unfavorable field
conditions can .vary by as much as 25 percent (National Bureau of
Standards, 1977). Temperature, depending on the detector, has an
important effect on the accuracy of a measurement and normally under
field conditions this is not controlled. Therefore, some kind of
compensation is required for temperature-sensitive detectors. The
cosine correction and the transmission factors of a diffuser can
introduce sizeable errors if not determined and evaluated. Stray
light problems associated with unwanted visible light not being
sufficiently blocked out in daylight field measurements can
introduce significant errors.
D. Weather Corrections
If atmospheric conditions were always clear and
pollution-free, there would not be the need for accurate
measurements of UV-B, since the amount of energy would be primarily
dependent upon ozone concentration. However, there are not too many
clear days; therefore, clouds, aerosols, dust and other pollutants
such as sulfur dioxide should be considered. Table 1 shows a
comparison of the measured* Radiation Biology Laboratory (RBL) daily
averages of UV-B centered at 305 nm at 40°N latitude and calculated
values. In addition, the cloud correction figures used by Mo and
Green (1974) were applied to the calculated values. It is obvious
that deviations of calculated values from measured values under the
best of conditions range up to 25 percent. There is little doubt
that meterological conditions have a strong influence on the
quantity of UV-B reaching the surface of the earth, indicating that
monitoring of UV-B should be performed in order to evaluate properly
UV-B energy doses received by biological organisms (Klein and
Goldberg, 1978).
Instrument accurate to +_ 5 percent based on a NBS referenced
standard lamp.
F-2-4
-------
Effects - UV-B Measurement
Table 1
The calculated integrated daily global UV radiation for 5 nm
centered at 305 nm at sea level for a clear sky in units of
joules m-2. Average amounts of ozone for latitude and season
have been used in the calculations (Mo and Green, 197A). The
measured values are daily averages for the month.
Latitude AO°N
Month
Jan.
Feb.
Mar.
Apr.
May
-Dun.
July
Aug.
Sept.
Oct.
Nov.
Dec.
Calc.
62.78
15A.5
350.6
665.8
957. A
1163.
1230.
1087.
731.2
351.6
121. A
55.88
Meas.
A2.6A
137.6
239.2
532.9
5A3.0
687.2
NO DATA
NO DATA
NO DATA
271.8
10A.7
AA.78
Calc
Meas.
1.A7
1.12
1.A7
1.25
1.76
1.69
-
-
-
1.29
1.16
1.25
Cloud
Correction
.66A
.66A
.630
.759
.614
.6A7
-
-
-
.619
.670
.608
[Calc]
Meas.
Corrected
.98
.74
.93
.95
1.08
1.09
'
-
-
.80
.78
.76
Table 2
Comparison of relative DMA dose, daily total UV from 280 nm
to 322.5 nm (J m-2), and daily total insolation from 280 nm to
2800 nm U m-2) for representative days in Rockville, MD.
29 Mar 76 29 Jun 76 20 Sept 76 13 Jan 76
Daily DNA dose 24.12 45.34 30.10 4.60
Daily total UV 3.95 x 10* 5.6A x 10A A.60 x 10A l.AA x 104
Daily total
insolation
1.16 x 107 1.98 x
107 6.95 x 106
F-2-5
-------
Effects - UV-B Measurement
E. Significance of UV-B Measurement Difficulties
It is essential that instruments and standards be developed
to enable researchers to measure changes in UV-B with an accuracy
that will provide the data needed to follow changes to which
biological organisms will be exposed at the surface of the earth.
These measurements must be sensitive and accurate enough to indicate
relatively small changes in ozone, or to indicate small changes in
ultraviolet irradiation to which biological material will be
exposed. Calculations are only valid for essentially clear days.
They should not be used to determine UV-B exposure since the error
can be large.
3. LATITUDINAL VARIATION OF UV-B AND ACTION SPECTRUM SELECTION
A. Projected UV-B Changes Versus UV-B Changes With Latitude
The MAS Report commented:
"Recognizing all the above uncertainties, Figure 2.1
indicates the increases in the annual DNA-damaging UV at
40°N latitude expected from possible ozone-layer reductions
over the next century (assuming continued release of
chlorofluorocarbons at the 1977 rates). A 7.5 percent
ozone-layer reduction would, for example, lead to about a
19 percent increase in DNA-damaging UV, and a 16 percent
reduction to about a 44 percent increase. (These values
will be somewhat less at latitudes toward the equator and
greater at higher latitudes). The increase in DUV for a
given decrease in ozone concentration is larger than the
figure estimated in earlier reports, but this is hardly a
significant change in the light of remaining uncertainties
about the proper weighting function. The change arises
from more recent knowledge of the solar intensity
distribution at the shorter wavelengths, coupled with a
presumably better action spectrum (Appendix C)." (NAS,
1979, p. 62)
B. Comparison of Measurements at Various Latitudes
It is estimated, considering all the uncertainties during
continued 1977 releases of chlorofluorocarbons, that an increase in
UV-B reaching the earth's surface would be of the order of 19 to 44
percent. This is small compared to existing differences at latitude
39°N (Rockville, MD), 30.4°N (Tallahasee, FL) and 9°N (Panama), all
from actual UV-B totals and from DNA dose equivalents that follow
(see Table 4).
F-2-6
-------
Effects - UV-B Measurement
Table 3
Comparison of relative DNA dose and daily total UV from 280 nm
to 322.5 nm (3 m-2) for representative days in Tallahassee, FL.
20 Mar 76 21 Jun 76 28 Sept 76 06 Jan 76
Daily DNA dose 45.68 55.54 36.75, 16.00
Daily total UV 4.60 x 10* 4.57 x 10* 3.01 x 10* 2.42 x
No totals
available - -
Table 4
Comparison of relative DNA dose, daily total UV from 280 nm to
322.5 nm (J m-2), and daily total insolation from 280 nm to
2800 nm (J m-2) for representative days in Panama, and a
comparison with corresponding quantities for Rockville, MD.
23 Mar 76 30 Jun 76 27 Sept 76 18 Dec 75
Daily DNA dose 130.96 112.25 124.85 90.24
Daily total UV 1.02 x 105 9.50 x 104 1.01 x 105 8.32 x 10A
Daily total
insolation 2.12 x 10? 1.66 x 107 1.54 x 107 1.50 x 107
Ratios, Panama; Rockville
Daily DNA dose 5.4 2.5 4.1 19.6
Daily total UV 2.6 1.7 2.2 5.8
Daily total
insolation 1.8 0.8 1.0 2.2
F-2-7
-------
Effects - UV-B Measurement
C. Calculating DNA Dose from Five Nanometer Energy Integrals
A necessary part of doing UV mutagenesis studies is to
determine the dose of radiation received by the organisms of
interest. There are two ways to measure this dose with radiometric
instruments; either measure with an instrument that has the same
spectral sensitivity of DNA, or measure with an instrument of
different sensitivity and use the DNA-sensitivity relationships to
convert these measurements to some relative dose measure. In this
case, we have chosen the latter method, using an eight channel UV
radiometer that measures the integrated energy in a nominal
bandwidth of five nm, centered at 285, 290, 295, 300, 305, 310, 315,
and 320 nm (Goldberg and Klein, 1974). For the DNA sensitivity
curve (action spectrum) we have used the one from Setlow (1974), as
used by NAS. The choice of action spectrum represents a major
uncertainty itself, as discussed in 3.D and 3.E below.
Given a relative sensitivity function, the dose (D) is
defined:
° • J
E(x) I(X) dX
where E(X) is the relative sensitivity curve
I(X) is the spectral density of the light input
and X is the wavelength
Since we do not measure I(X) directly, but rather measure the energy
integrals over the eight channels (each 5 nm wide), we approximate:
8 Xi + 2.5
D = I (E (X^ ji(x) dx; Xi = 285, 290,..., 320
i = 1 Xi - 2.5
A further assumption is that dose and radiant flux are linearly
related with respect to time; doubling the exposure time while
cutting the flux in half should give the same dose. 'Dose units used
here are arbitrary; one dose unit is the equivalent of 1000 Jm-2
of 310 nm light. In terms of flux, one dose unit is equivalent to
10 W m-2 of 310 nm light for 100 seconds (or 5 W nr2 of 310 nm
light for 200 seconds).
The tables presented here show total dose, total UV (280 nm
- 322.5 nm), and total insolation (280 nm -2800 nm) for three sites
at four times of the year for representative days. Hourly totals
are given for dose and UV. Days are deemed representative on the
basis of having a daily total insolation that is near average for
the location and season. Tables 5-7 show hourly totals, and are
summed to provide the daily totals in Tables 2-4. Table 4 also
compares the ratios between Panama and Rockville, MD. Using the
Setlow DNA action spectrum the average DNA daily dose is 340 percent
larger in Panama, while the average daily insolation is only 28
percent larger in Panama.
F-2-8
-------
Effects - UV-B Measurement
Table 5
Hourly totals for relative DNA dose and tc
322.5 nm (J m-2) for representative daj
(1976).
29 Mar 76 29 Jun 76 20
Hour
4-5
5-6
6-7
7-8
8-9
10-11
11-12
12-13
13-14
14-15
15-16
16-17
17-18
18-19
19-20
Dose
0.00
0.02
0.25
0.86
3.23
4.40
5.06
4.58
2.53
0.92
0.30
0.04
0.00
Total
UV
1
192
1257
2559
4988
5956
6602
6432
4217
2078
1083
297
15
Dose
0.0
0.02
0.21
0.91
2.66
4.60
8.59
10.41
8.83
4.41
1.27
1.09
0.37
0.07
0.00
Total
UV
4
191
1068
2571
3880
5478
8632
9917
9029
5416
2328
2494
1416
442
31
Dose
0.00
0.02
0.28
3.04
5.20
6.71
6.27
4.31
2.45
0.61
0.05
0.00
Dtal UV from 280 nm to
fs in Rockville, MD
Sept 76
Total
UV
1
198
1201
5408
7397
8489
8033
6232
4381
1497
223
27
13
Dose
0.02
0.31
0.74
1.06
1.22
0.77
0.34
0.04
0.01
0.00
Jan 76
Total
UV
26
1428
2548
3001
3335
2232
1185
243
56
2
F-2-9
-------
Effects - UV-B Measurement
Table 6
Hour
5-6
6-7
7-8
8-9
9-10
10-11
11-12
12-13
13-14
-14-15
15-16
16-17
17-18
18-19
19-20
Note:
totals for relative DNA dose and total UV from 280
nm (J m-2) for representative days in Tallahassee,
20 Mar 76 21 Jun 76 28 Sept 76
Dose
0.00
0.10
0.77
2.63
4.33
5.68
10.02
8.61
7.56
3.93
1.65
0.39
0.02
Total
UV
9
403
1782
3683
4474
4979
7956
7014
7024
4560
2857
1138
113
Dose
0.00
0.06
0.54
2.13
5.09
7.47
9.45
8.55
9.44
6.04
3.23
2.47
0.89
0.16
0.01
Total
UV
6
261
1214
2924
4849
5678
6290
5557
6261
4368
3050
3003
1668
508
40
Dose
0.00
0.10
0.76
3.64
2.81
7.41
7.45
6.42
3.72
2.90
1.16
0.08
0.00
Total
UV
16
488
1364
3587
2200
5008
4899
4272
2887
3014
1934
423
9
nm to
FL.
06 Jan 76
Dose
0.00
0.04
0.27
1.27
2.95
3.93
3.63
2.45
1.13
0.31
0.03
Total
UV
8 '
198
745
2299
4183
5047
4766
3718
2260
903
116
The instrument used at Tallahassee differs from the others
in that instead of measuring a 5 nm wide integral centered
at 300 nm, it measured a ten nm wide integral at 297 nm.
This measurement is used with the 295 nm integral to
estimate the 300 nm integral.
F-2-10
-------
Effects - UV-B Measurement
Table 7
Hourly totals for relative DNA dose and total UV from 280 nm to
322.5 nm (J m-2) for representative days in Panama.
23 Mar 76 30 Jun 76 27
Hour
6-7
7-8
8-9
9-10
10-11
11-12
12-13
13-14
14-15
15-16
16-17
17-18
18-19
Dose
0.01
0.47
3.15
6.57
14.45
25.54
25.14
24.92
18.29
8.98
3.02
0.42
0.01
Total
UV
72
1361
497.1
6834
10892
16533
15552
16144
13898
9143
4788
1345
49
Dose
0.03
0.71
3.14
8.14
18.77
28.47
22.43
14.81
6.48
5.80
3.05
0.41
0.02
Total
uv •
205
1952
4732
8401
15056
19586
15039
10758
5928
6783
5176
1269
100
Dose
0.02
0.60
1.26
3.57
11.95
33.99
32.36
25.61
9.97
4.11
1.33
0.09
0.00
Sept 76
Total
UV
98
1645
2155
3899
9918
24206
22381
19123
9335
5531
2661
390
2
18 Dec
Dose
0.00
0.34
2.53
8.41
14.84
23.51
18.53
14.16
5.74
1.55
0.57
0.05
0.01
75
Total
UV
50
1319
4889
10121
13516
18229
13948
11513
5643
2217
1519
239
7
D. Action Spectrum Choice
The relative response of a system to various-wavelengths of the
visible and near-visible light spectrum is termed an action
spectrum. A particular action spectrum is a very specialized thing
and may apply only under the conditions used to obtain it.
Therefore, it is probably more appropriate to indicate that this is
an action spectrum and not the action spectrum.
E. Action Spectra Other than DNA
Figure 1 shows the weighting functions in current use for
biological effects. For our comparison of measured UV-B values and
calculated DNA dose for each latitude, we have used, and NAS used,
the most effective action spectrum for biological material.
Therefore, the maximum effect is shown and all the other weighting
functions will generally be considerably less.
F-2-11
-------
Effects - UV-B Measurement
z
LU
UJ
u.
-1
-2
-3
UJ
cr
O
O -4
-5
280 290 300 310 320
WAVELENGTH (nm)
Photosynthesis
Inhibition
330
340
Weighting functions in current use for biological UV effects.
FIGURE 1
(from MAS, 1979, p. 307)
F-2-12
-------
Effects - UV-B Measurement
F. Significance
/
There is already a natural variation with latitude that
greatly exceeds the anticipated change in UV-B due to ozone
changes. My opinion is that comparisons can be made between
latitudes using modern controlled environment facilities in
ultraviolet transmitting glasshouses. In fact, it is recommended in
the Report under Major Research Issues (MAS, 1979, p. 71) that such
a monitoring facility be established at various locations. The
action spectra for most biological effects of ultraviolet radiation
are not well known and caution should be used when using generalized
action spectra for estimating effects. Use of the DNA or Setlow
action spectrum represents a "worst case" choice.
"Another type of natural ozone variation offers an
opportunity to set limits on the magnitude of
ozone-depletion effects. These are cyclic variations,
amounting to approximately 5 percent total amplitude at
temperate latidudes, over a period of about a decade
(Angell and Korshover, 1973). Weighted for DNA-damaging
effectivenss, such ozone variations would produce roughly a
13 percent change in DUV. The common experience of
temperate areas of the world (which have repeatedly been
through such cycles) shows that changes of this magnitude
do not produce any spectacular effects on plants or animals
over the relatively few years of their duration. (Small
effects would, of course, tend to be blurred by the
ordinary variations in weather and other factors.) While
there seems to have been no concerted effort at detecting
effects due to these changes, it would seem safe to say
that most organisms can reasonably accommodate decade-long
oscillations in ozone concentration, of the order of one
third of the 16 percent change expected from continued CFC
release at current rates." (NAS, 1979, p. 64)
I agree with the above statement about cyclical variations
and according to Figure 2 (World Meterological Organization, 1977)
this kind of variation in ozone has been occurring for twenty or
more years. It would seem to me that most organisms could safely
handle a two or three percent change in ozone with little or no
detectable effect. It would probably not be a measureable
biological response with current techniques and methods.
F-2-13
-------
Effects - UV-B Measurement
30r-
20
10
-10
-20
6
4
2
0
-2
-4
•4
2
0
-2
-4
A
2
0
-2
-4
4
2
0
-2
-4
-6
-8
UJ
O
z
UJ
o
cr
UJ
a
(Note different scale
compared with other
regions below)
-10
TJ.S.S.R.
(7 stations)
EUROPE
(14 stations)
WORTH -
AKEEICA
(11 stations)
JAPAN
(3 stations)
INDIA
(5 stations)
1950
1955
iyoo
I9G5
1970
I97D
I960
Tine variation in total ozone in north temperate latitudes expressed as a percentage
deviation from the mean for the total length of record (the annual oscillation has
been removed). A 1-2-1 smoothing (divided by four) has been applied twice to the
successive seasonal values.
Vertical bars represent two standard errors of estimate based on individual station
values within the regions. Single-shafted arrows indicate occurrence of quasi-
biennial west wind maximum at 50 mb in the tropics;
A = Exuption of Mt. Agung (Indonesia) F = Eruption of Kt. Fuego (Guatemala)
N •» Large nuclear explosions S = Large solar proton event.
FIGURE 2
(from World Meterological Organization, 1977)
F-2-14
-------
Effects - UV-B Measurement
A. SUPPLEMENTATION OF UV-B WITH LAMPS
A. Experimental Limitations and Sources of Error
The NAS report acknowledges difficulties associated with
experimental supplementation of UV-B with lamps:
"The measured spectral composition of radiation from
the lamp sources is not the same as that of sunlight over
the biologically effective wavelengths (Figure E.I,
Appendix E). Consequently, in comparing its effects on
plants and animals with those of natural sunlight, a
weighting function based on the proper action spectrum must
be applied to the spectral distributions of both sources.
If the biological action spectra for UV radiation damage to
plants and animals were known precisely, one might make an
accurate comparison by this means and be able to predict
the consequences of ozone depletion from the lamp
experiments. However (as explained above and in Appendix
D), the action spectra for most biological effects of UV-B
radiation are not well known, and one must usually surmise
their form approximately. This can introduce an
uncertainty of as .much as twofold in predicting the
increased biological damage accompanying a 16 percent
reduction in the ozone layer." (NAS, 1979, p. 65)
and:
"Finally, the environmental conditions under which
many of the experiments have been performed may in some
cases have altered the sensitivity of plants and animals to
UV-B radiation. For example, the UV sensitivity of plants
appears to be as much as fourfold greater in the artifical
illumination of plant environment growth chambers or in
greenhouses, than in the open-field environment, possible
because of the different level of photosynthetic
illumination. Unfortunately, this means that the
experiments providing the most completely controlled
conditions (which should therefore allow more refined
testing) are not by themselves able to evaluate the
consequences of increased solar UV on plants. Also, in
most experiments, plants and animals have been subjected to
environmental conditions free from other stresses besides
the UV radiation. Thus, the interaction of such other
stresses with the UV-B has not yet been evaluated." (NAS,
1979, p. 66)
Further justification for these concerns follows below
F-2-15
-------
Effects - UV-B Measurement
B. Physical Characteristics
Experiments using fluorescent lamps to provide supplemental
UV-B must be carefully examined before accepting the results, even
on a qualitative basis. Lamps age and filters change and the
shortwave cutoff (Figure 3) often is not appropriate, as well as the
spectral quality and intensity of the rest of the spectrum. Figure
4 presents the aging effect on the emission of a UV-B lamp. Forty
percent of the energy is lost at the end of 2000 hours and the
decrease is relatively uniform after the first 10 hours of use.
This loss associated with the slow darkening of the filters
introduces significant errors in doses unless measured and adjusted
practically every day.
C. Experimental Factors in the Use of Lamps
One of the serious defects in UV-B supplemental lamps is
the inability to irradiate in the natural min-max-min mode as occurs
in natural daylight. This produces an improper dose rate and could
cause a completely different response to occur. Tables 5-7 show the
hourly doses of UV-B and also equivalent DNA dose at three latitudes
for a representative day, indicating change with time of day. It is
evident that a fixed intensity for four to six hours would not
simulate the natural exposure conditions. The lamp position is
fixed and usually produces a non-uniform source of UV-B energy, with
also a fixed shade over certain portions of the plant canopy. This,
then, results in less uniform visible light for photoreactivation
effects to occur. Most growth chambers cannot simulate the natural
daylight, neither from the spectral distribution nor intensity
standpoint and, therefore, results cannot be used to predict field
responses unless an appropriate transfer function is first
determined. It is plausible that this light quality and intensity
difference is the reason for a 4-fold difference in UV-B sensitivity
of plants grown in growth chambers versus natural conditions.
D. Significance of Difficulties in Supplementing UV-B with
Lamps
Extreme caution should be used in attempting to extrapolate
UV-B effects from growth chamber results to natural conditions
because of the inability to reproduce natural conditions. Even
experiments unde'r natural conditions with supplemental UV-B lamps
should be evaluated with the full knowledge that lamps and filters
change on a daily basis. Experiments should be conducted in which
monitoring of these factors are performed on a daily basis.
F-2-16
-------
Effects - UV-B Measurement
10
10'
10'
10'
Solar Irradiancv
fl'lO 0.20 / / 'OSO •<"> • cm Oione
I Q.30'0.401 |
280 290 300 310 320
WAVELENGTH (nm)
330
Spectral irradiance at a distance of 16.5 cm
from six FS40 sunlamps, filtered with 5-mil cellulose
acetate, superimposed on the solar spectral irradiance
(sun 30° from zenith) with different ozone thicknesses.
FIGURE 3
(from NAS, 1979, p. 320)
F-2-17
-------
c
0)
CD
(H
CD
0)
z:
CD
I
FIGURE 4. DECAY CURVE FOR SYLVANIA UV-B LAMP #2021
GTE SYLVANIA UV PHOSPHOR F48T12/2021/VHO
Test duration 2000.h (2 May 80 - 24 July 80)
en
4->
o
-------
Effects - UV-B Measurement
5. REFERENCES
Copies of the cited references appear in the Reference
Volumes.
Goldberg, B. and Klein, W. H. (197A). Radiometer to monitor low
levels of ultraviolet irradiances. Appl. Opt., jj (3), 493-496.
Klein, W. H. and Goldberg, B. (1978). Monitoring UV-B spectral
irradiances at three latitudes. Proc. Internatl. Solar Energy
Soc. Congress (Pergamon Press) !_, 400-413 (New Delhi, India,
January).
Mo, T. and Green, A.E.S. (1974). A climatology of solar erythema
dose. Photochem. Photobiol., 20, 483-496.
NAS - National Academy of Sciences - (1979). Report of the Committee
on Impacts of Stratospheric Change, ir\_ "Protection Against
Depletion of Stratospheric Ozone by Chlorofluorocarbons."
December, Washington, D.C.
National Bureau of Standards (1977). Symposium on Ultraviolet
Radiation Measurements for Environmental Protection and Public
Safety, June 8-9, Gaithersburg, MD.
Setlow, R. B. (1974). The wavelengths in sunlight effective in
producing skin cancer: a theoretical analysis. Proc. Nat.
Acad. Sci. USA. 73. (9), 3363-3366.
World Meteorological Organization (1977). UNEP Meeting of Experts:
Atmospheric Ozone. A survey of the current state of knowledge
of the ozone layer, March 1-9. Washington, D.C.
F-2-19
-------
COMMENTS ON THE NATIONAL ACADEMY
OF SCIENCES REPORT: "PROTECTION AGAINST
DEPLETION OF STRATOSPHERIC OZONE BY
CHLOROFLUOROCARBONS. "
PREPARED FOR E. I. DU PONT DE NEMOURS & COMPANY, INC,
BY
DR. R. H. BIGGS
PROFESSOR
INSTITUTE OF FOOD AND AGRICULTURAL SCIENCES
UNIVERSITY OF FLORIDA
GAINESVILLE, FLORIDA
DECEMBER, 1980
(APPENDIX F-3)
F-3-1
-------
Transmittal letter for
UNIVERSITY OF FLORIDA Appendix F-3
INSTITUTE OF FOOD AND AGRICULTURAL SCIENCES
•
GAINESVILLE. FLORIDA 32611
FRUIT CROPS DEPARTMENT
1172 MCCARTY
TELEPHONE: 904-392-1996
December 15, 1980
Dr. Richard B, Ward
Research Associate,Environmental
E.I. duPont de Nemours & Company
Wilmington, Delaware 19898
Dear Dr. Ward:
Enclosed please find my comments on the National Academy of
Science Report: "Protection against depletion of stratospheric
ozone by chlorofluorocarbons." I understand it will be appended
to a report you will be submitting to EPA.
I hope the comments are useful.
Sincerely,
Dr. R, H. Biggs
Professor (Biochemist)
-------
Effects - Crops
COMMENTS ON THE NATIONAL ACADEMY
OF SCIENCES REPORT: "PROTECTION AGAINST
DEPLETION OF STRATOSPHERIC OZONE .BY
CHLOROFLUOROCARBONS."
TABLE OF CONTENTS
Page
1. INTRODUCTION F-3-3
2. SUMMARY F-3-3
3. SPECIFIC COMMENTS ON THE REPORT F-3-A
A. General observations on an inadequate data
base for prediction. F-3-A
B. Biological Responses to UV-B radiation. F-3-A
(1) Action spectra and biological weighting
functions F-3-A
(2) Shielding and orientation F-3-5
(3) Field and environmental chamber problems F-3-6
(A) Significance of UV-B change F-3-10
(5) Major research issues F-3-10
A. REFERENCES F-3-13
F-3-2
-------
Effects - Crops
COMMENTS BY DR. R. H. BIGGS ON THE NATIONAL ACADEMY OF
SCIENCES REPORT, "PROTECTION AGAINST DEPLETION OF STRATOSPHERIC
OZONE BY CHLOROFLUOROCARBONS."
1. INTRODUCTION
In studying the new National Academy of Sciences (NAS)
report, "Protection against depletion of stratospheric ozone by
chlorofluorocarbons" (NAS, 1979) with particular emphasis on higher
plant systems, I had the reactions noted on subsequent pages. I
appreciate the opportunity to study this report and am hopeful that
my reactions to an extremely complex problem are useful. The
comments reflect my efforts, and so opinions stated can only be
attributed to me and no other person or organization.
Periodical review of this complex problem by the National
Academy of Sciences Committee on Impacts of Stratospheric Change is
extremely useful, and indeed, the present document is a valuable
general review of the UV-B radiation problem as it is envisioned
today. Careful reading of each section of the document does
indicate that conclusions based on facts are fair in light of my
knowledge of the problems. It is the "summaries of summaries," the
attempt to boil down the complex problem to a socially manageable
one, that I would criticize. Specific suggestions will be related
to the sections of the report dealing with higher plants and will
contain some suggestions as to future research endeavors that may
lend insight into plant responses to UV-B radiation that may have
some predictive value as applied to both cultivated and
non-cultivated ecosystems.
2. SUMMARY
There is no data base to predict that a catastrophe
will occur with crop plants as a result of decreasing
ozone predicted to occur at present chlorofluorocarbon
release rates or those after 5 years.
Growth chamber and greenhouse studies of cultivated
plants cannot be used at the present time to predict
yields under field conditions. These types of studies
should be continued along with field studies.
We need a much larger data base to predict effects of
increasing UV-B radiation on crop yields. Top priority
should be placed on field studies, action spectra, and
real-world documentation as related to changes in UV-B
irradiance in cropping systems and in non-cultivated
ecosystems.
F-3-3
-------
Effects - Crops
3. SPECIFIC COMMENTS ON THE REPORT
(NAS, 1979 - Chapter 2: Nonhuman Biological Effects, pp.
58-73; Appendix D: The Biologically Effective Ultraviolet
Radiation, pp. 304-318.)
A. General observations on an inadequate data base for
prediction
From my assessment of the Terrestrial effects, non-human
research, there is not a data base at this point to do much
predicting of the biospheric consequences of a stratospheric ozone
reduction. Although it is limited, the better data base is for
cultivated plants, and in this case, there is no reason to
anticipate an immediate catastrophe. In fact, what is evident is
that there are alternative ways to cope with this environmental
stress factor just as there are for high solar irradiances, drought,
winds, unfavorable temperatures, etc. Plants have mechanisms to
cope with UV-B radiation (285-320 nm) to varying degrees. The real
question that must be addressed is: when does the UV-B radiation
exceed a certain threshold that will affect yield, or some other
desirable plant trait of a crop or ecosystem? we cannot predict the
answer to this question with any precision as to yields from any
higher plant at the present time, nor can we predict what will
happen on specific ecosystems. Therefore, the position at this
point should be that there is the possibility of an additional
stress load on terrestrial plants, both cultivated and
non-cultivated, from an increase in UV-B radiation in the biosphere
should a 16% reduction in ozone ultimately occur. However, the
degree of uncertainty that is associated with the possibilities
would seem to indicate that the best course of action would be to
proceed for a limited period of time to mount a good research effort
to reduce (a) the uncertainties associated with knowing the degree
of stratospheric changes expected in relation to time, say five
years when some verification of whether stratospheric ozone changes
predicted by atmospheric scientists is actually occurring, and (b)
those uncertainties associated with biological effects of UV-B
radiation on plants. The latter will be dependent on rates of
changes, i.e., seasonal and long-term patterns of change in
stratospheric ozone.
B. Biological Responses to UV-B Radiation
(NAS, 1979, pp. 58-63)
(1) Action spectra and biological weighting functions
While there is no question that if DMA and proteins
absorb UV-B radiation there could be damage, there is a question
about how much UV-B radiation is actually absorbed by nucleic acids
and proteins that control plant growth and development, and the
relation of this absorbance to actual damage.
F-3-4
-------
Effects - Crops
We do not have a way at the present time of estimating
this damage directly and indirect methods are encumbered by many
sources of error, i.e., reflective and absorbance properties as
related to overall geometry of the plant; shape, structure and
orientation of individual organs; and internal structure of organs,
particularly as related to protective absorbers.
Biological weight functions are extremly important.
Our knowledge of how to make equivalences across the UV-B radiation
wave-band are based on only a few action spectra for biological
responses in this wave-band and these differ greatly (Caldwell,
1977). Since all quantitative predictions rely heavily on the
choice of the weighting function for evaluating potential damage,
there is reason for caution and a vital need for a better data base
in this area, particularly between the correct weighting function to
use as related to limiting factors in crop production or in
competititon between species in ecosystems.
Appendix D (NAS, 1979, pp. 304-318) is a good treatment
of the necessity for using biological weighting functions when
relating biological responses to the UV-B radiation wave-band. As
more data becomes available, i.e., photosynthesis, phytohormone
changes, partitioning of substrates between plant organs, etc., it
will be essential to supply the appropriate weighting function to
each response. The use of the weighting function that is based on
DMA has utility only because we lack the data based on action
spectra as related to specific physiological processes that may be
limiting factors under natural conditions. This demonstrates an
area of research that is vitally needed, that is, action spectra as
related to plant physiological responses to UV-B radiation.
(2) Shielding and orientation
Meristematic and other regenerative cells are generally
well shielded in most cases from UV-B radiation. In view of this,
it would seem important to stress avoidance mechanisms in more
detail than dealt with in the report, particularly as related to
photoprotection and future experimentation. For example, it could
be that the phototropic responses of leaf orientation as related to
incident radiation is an important factor in avoidance mechanisms.
The same would be true for orientation of meristematic organs. If
the phototropic response is heliotropic in nature, and this is the
case with certain varieties of soybeans, then comparisons between
controlled environmental growth chambers and field conditions will
have another confounding parameter. This may seem like a small
point but in models to predict yields of soybeans, leaf angle is a
major component (Duncan, 1971). Knowledge of these types of plant
responses lends further credence to implications in the report that
controlled environmental chambers data cannot be used to predict
plant responses, particularly yields in the field. Related to this
problem is the statement in the NAS report (NAS, 1979, p. 66) as
follows:
F-3-5
-------
Effects - Crops
"Finally, the environmental conditions under which
many of the experiments have been performed may in some
cases have altered the sensitivity of plants and
animals to UV-B radiation. For example, the UV
sensitivity of plants appears to be as much as fourfold
greater in the artificial illumination of plant
environment growth chambers, or in greenhouses, than in
the open-field environment, possibly because of the
different level of photosynthetic illumination.
Unfortunately, this means that the experiments
providing the most completely controlled conditions
(which should therefore allow more refined testing) are
not by themselves able to evaluate the consequences of
increased solar UV on plants. Also, in most
experiments, plants and animals have been subjected to
environmental conditions free from other stresses
besides the UV radiation. Thus, the interaction of
such other stresses with the UV-B has not yet been
evaluated."
It should be noted that even the "fourfold greater" estimate of
sensitivity is at this point only a rough guess, and this is for dry
matter production, not seed yields.
A listing of other avoidance mechanisms identified as
important to the UV-B radiation assessment as to damage, namely,
pigment changes (Caldwell, 1977) phytohormones (Lindoo e_t a.1. , 1979)
cuticles (Robberecht and Caldwell, 1978; Kossuth and Biggs, 1978)
and changes in plant forms (Biggs and Kossuth, 1978) also reinforces
the idea that we have very little knowledge at the present time of
transfer functions between growth and development of plants in
controlled environmental chambers and under field conditions when
exposed to enhanced UV-B irradiance levels. Thus a diligent and
consistent effort should be made by biologists and theoreticians to
establish the nature of these transfer functions, if they exist.
(3) Field and environmental chamber problems
I would agree with the last sentence of the
introductory paragraph (MAS, 1979, p. 63) "only beginnings have been
made in this direction," meaning in relation to experimental studies
on the general nature of biological responses both under field and
controlled chamber conditions. (Using the word damage at this point
could prejudice the research before it is conducted.)
While I would agree in principle with the idea that
"the epidemiological approach" would seem unworkable for estimating
effects of ozone depletion on plants because of many natural
uncontrollable factors, stratified experiments could take advantage
F-3-6
-------
Effects - Crops
of the natural variation in UV-B radiation at ground level to assess
plant responses. With many crop plants, we are able to optimize
water, mineral elements, temperature, gases (quantities and quality)
and in fact most factors except radiation. Optimization and
characterization of radiation for control environmental climate
chambers and greenhouses has been the single most troublesome
component. The use of natural solar fluences with UV-B radiation as
high as, or higher than, what would accompany a 16% ozone reduction
at 40°N latitude (for example, high altitude in Hawaii or other low
latitude and high altitude locations) to irradiate plants while
keeping other factors optimized would seem to be a good experimental
approach, especially if this is properly interfaced to conventional
controlled environmental phytotron and field experimentation. For
this integrative physiological approach, systems analyses will be
very critical to its success as a predictive tool.
As has been pointed out by Dr. w. H. Klein in his
comments*, experiments using fluorescent sun lamps to supplement
UV-B radiation, spectral quality and intensity, and the ON-or-OFF
mode of actuating the lamps leads to an unnatural exposure of plants
to UV-B irradiances. In addition to these factors inherent in lamp
enhancement studies, there is the problem of an unnatural albedo in
controlled climate chambers with highly reflective walls. This
enhances the albedo of UV-B irradiance over the direct component to
a much greater level than would generally occur in nature. This
could seriously hamper photo-control avoidance mechanisms and could
lead to an over-estimate of potential damage. The Duke Phytotron
study in the "C" type chambers by Biggs and Kossuth (1978) provided
much of the data for the NAS summary statement No. 1 (NAS, 1979, p.
67) as follows:
11 1. Tests of more than 100 species or varieties of
species in controlled environment growth chambers indicate
that approximately 20 percent are sensitive to daily UV-B
doses of the order of those delivered by Florida sunshine
at present ozone levels, while 20 percent were resistant to
doses four times greater than this, and the remaining 60
percent showed some intermediate sensitivity. These tests
would indicate that a significant fraction of the present
agricultural varieties are at present under UV stress and
would suffer decreased production with a 16 percent ozone
reduction."
This research was done in chambers with highly
reflective walls of polished aluminum. While enough caveats were
placed in the detailed accounts on these experiments, this summary
statement implies a level of damage that could lead to an
over-estimate of the problem. This means that summarizers should
make every effort to guard against misleading interpretation that
might arise in summary statements. For example, in the NAS Appendix
B (NAS, 1979, pp. 283-284):
*Note added by Du Pont. See Appendix F-2,
F-3-7
-------
Effects - Crops
"Over 100 species and varieties of agricultural plants,
as well as some nonagricultural higher plant species, have
now been tested for their comparative sensitivity to UV-B
radiation in controlled environmental growth chambers. A
short-term, but extensive screening program of 82 species
and varieties was recently undertaken at the University of
Florida in Gainesville (R. H. Biggs and S. V. Kossuth,
unpublished BACER report). Approximately 20 percent of the
species were classed as highly UV-B-sensitive species.
Many exhibited significant reductions in yield when exposed
daily to the lowest UV-B dose employed, which is
approximately equivalent to the amount of UV-B radiation
received in a half day of normal sunshine in the summer in
northern Florida. On the other hand, 20 percent of the
species suffered no apparent reduction in yield at even the
highest daily doses employed, which would correspond to
roughly twice the amount of solar UV-B radiation currently
received during a summer day in northern Florida. The
remaining 60 percent exhibited intermediate degrees of
sensitivity to UV-B radiation. Apart from reductions in
plant dry weight or yield, UV-B radiation also affected the
proportion of plant material represented in various plant
organs such as roots, shoots, and leaves. Although these
alterations might not necessarily be detrimental
themselves, they could result in undesirable consequences
for agricultural utilization of the plants." (Emphasis
added).
The word yield in this sense means reductions in the
dry weights of plants grown under these rather stratified conditions
in the phytotron. In subsequent summaries, there is the real
possibility that soybean will become the modifier and yields the
subject with a very active verb. A case in point is the fourfold
estimate statement in the second summary of findings (MAS, 1979, p.
677!
11 2. However, 15 species and varieties tested in the
open field appeared more UV resistant than plants grown in
laboratory environmental growth chambers. Where it is
possible to make comparisons of the same plants, the
differences in resistance are on the order of fourfold. If
this result turns out to be general, the higher sensitivity
indicated in the growth chambers does not represent
open-field behavior, and the expected consequences of ozone
depletion become considerably less than those tests would
indicate. Nevertheless, some species (sugar beets,
tomatoes, mustard, corn) still appear to be
affected—although the field experiments were necessarily
less well controlled, and the results less clear-cut, than
were the chamber studies." (Emphasis added).
F-3-8
-------
Effects - Crops
This was estimated from plant growth parameters so it
cannot be applied to yields. If anything this should be viewed as
an over-estimate in terms of effects on crop yield, particularly as
applied to seed crops.
Controlled environmental chamber studies cannot be used
to extrapolate to field conditions. They do provide a data base for
an identification of symptoms that might be encountered by higher
plants under stress from UV-B radiation. It may be that in
retrospect, after a sufficient comparative data base between plants
grown in controlled environmental growth chambers and under field
conditions have become available, that these and other controlled
environmental studies could be the basis for establishing transfer
functions for extrapolation between highly stratified studies and
field conditions, particularly biomass and form. At the present
time there is a lack of field related studies to establish any
transfer functions especially for yields or organs associated with
reproduction.
From our experience with field studies and analyzing
the data of others there are many variables encountered. One fact
noted from the CIAP studies (Biggs e_t al., 1975) was that plants
become more sensitive to UV-B radiation as irradiances in the UV-A
and visible portion of the spectrum are decreased.
Knowing this, our objective in the phytotron screening
test was to demonstrate the UV-B radiation at moderate fluence
levels, but under stratified conditions to demonstrate sensitivity,
does affect plants. These tests were designed to study the
responses of the various species to UV-B radiation stress to provide
a diagnostic symptoms base. It was assumed that plants that did not
demonstrate obvious symptoms of stress might be very tolerant under
field conditions. This latter is still a hypothesis. Even though
it would seem logical, it must be substantiated by fact in the
integrative physiological reactive mode under field conditions. A
further hypothesis was that it may be possible to establish the
degree of sensitization to UV-B radiation that arises as a result of
the stratified conditions. Verification of such a transfer function
has not been established but is seemingly related to low vs. high
irradiance in the visible portion of the spectrum.
No studies of plant responses under phytotron or
controlled environmental conditions per sje have ever been used
successfully to quantitatively predict crop yield under field
conditions (Loomis ejt al. , 1979). They have been used to
demonstrate that a certain parameter is critical to growth and
development and should be considered at some point in the hierarchic
levels of integration of a number of factors to predict the output
of the yield of a crop from a plant community.
F-3-9
-------
Effects - Crops
A good realization of the limitation of controlled
environmental growth chambers was outlined in a recent report on how
to make comparisons between laboratories as to reliability of growth
chamber generated data (Ormond et_ al., 1980).
It was once hoped that phytotrons would fill the role
of being a powerful tool for the integrative physiology of plants as
adaptive control systems. This has not been the case and it now
seems that crop modelers will be the principal consumers of plant
responses under phytotron, or control environmental conditions and
that models or system analyses will be the integrative tool (Loomis
£t al., 1979). If this is so and certain models have been
successfully applied to predict yields (Evans, 1977), experiments in
controlled environmental conditions, stratified greenhouse and field
tests and natural tests, should be interfaced to maximize
establishment of interfacing components for system analyses. Thus
an integrative approach would seem to have great merit to study the
problem.
(4) Significance of UV-B change
My best estimate at this time from my own studies and
reviewing others is that a 5 percent change in ozone, i.e., 13
percent UV-B irradiance increase, can be tolerated by most
cultivated plants adapted to grow in full sunlight. There may be
minimal effects on yields of some crops, particularly those related
to biomass production, but the data base does not allow for
quantitation in parameterizing the degree of effects on yield. One
would expect that the first observable changes would be at the low
latitudes where a predicted 5 percent ozone would be super-imposed
on a cyclic variation around a nodal point of less ozone in the
stratosphere.
(5) Major research issues: (MAS, 1979, pp. 70-72).
I would concur with the NAS statement number 1 (NAS, 1979,
p. 70) on recent research advances:
" 1. Larger numbers of agricultural plant varieties
have been experimentally tested by radiation supplement for
their sensitivity to UV-B. These tests have confirmed the
wide variation in sensitivity among different plant
varieties that was suspected earlier and have also
indicated that the sensitivity depends on conditions of
testing, but the tests leave the true sensitivity profile
of the most important agricultural plants in doubt."
A concerted effort should be made as I indicated in the
final paragraph of III.2.C, (page 10 of this Appendix).
F-3-10
-------
Effects - Crops
I would wholeheartedly support the suggestion that
action spectra are very much needed (NAS, 1979, p. 70) and would
make the task of constructing integrative physiological systems as
related to changes in cropping and ecosystems much easier than would
an empirical approach. To give proper biological weighting between
UV-B radiation increases and functional disorders, action spectra of
limiting processes are a must.
Both high-intensity, solar-simulators and
UV-transmitting greenhouses at high altitudes in low latitudes are
excellent suggestions (NAS, 1979, p. 71). My expansion on these
ideas would be to validate these tests further with field tests in a
system analysis approach.
A study to evaluate the feasibility of crop breeding to
develop more UV-resistant cultivars would not only provide more UV-B
resistant selections but would give some indication of the genetic
basis for adaptability (NAS, 1979, pp. 71-72.) and physiological
systems involved. Plants have evolved along with an evolution of
the atmosphere. There were earlier periods when there were much
higher fluences of UV-B radiation. Therefore, there are biological
mechanisms available for coping. The question is the rate of change
that can be handled by natural ecosystems. As viewed for cultivated
plants, what is the feasibility of crop breeding to aid in coping
with increased UV-B irradiances and genetically can these systems be
identified? Allied to this would be to address the question of how
many crop introductions in the tropics in high UV-B areas have
failed, or do poorly, because of this additional stress factor.
This could be a strategy to encourage research in this vital area
today which would lead to a better data base on the UV-B radiation
problem in general.
NAS1 last paragraph on major research issues (NAS,
1979, p. 72) is well stated. There must be a long-term commitment
to critical mass experimentation in this area. Facilities and
financial support must form part of a consistent and sustained
research program. The report of past efforts in this area, at best,
can only be viewed as the first stage of research. It was
constructed to ask the questions: Do we have a problem and what are
the possible effects? There is a several "quantum jump" required to
answer the question. Given a certain percentage depletion of ozone
what will be the real effects on plants in the biosphere? Thus, I
would agree with the general context of the concluding paragraph
(NAS, 1979, p. 72). Time is required for the development of a sound
data base. A "gearing up" and "gearing down" on a short turnaround
time has produced experiments that were hastily conceived and
conducted to assist in parameterizing the range of problems that
could be encountered by increased UV-B in the biosphere.
In addition to the suggestions in the report, I would
like to add the following from my perspective.
F-3-11
-------
Effects - Crops
Biologists need three critical items to address many
questions relevant to effects of the UV-B radiation on
plants (or animals):
a) Good irradiance sources of high intensity with proper
attentuating filters to simulate expected solar
irradiance in the UV-B wave-band.
b) Reliable and rugged spectroradiometers to validate
conditions in tests. This has to be coupled to
dependable calibration standards.
c) Systems capable of real-time analysis of variable
integrative physiological components that provide a
basis for predictive purposes.
Examination of UV-B radiation reflectors, e.g., dichroic
filters, for enhancement studies in the field should be
made. They could possibly simulate ozone depletion better
than currently available lamps.
When a space laboratory becomes a reality, priority should
be given to the testing of higher plants to
extra-terrestrial solar radiation using attenuating filters
to approximate ozone depletion.
F-3-12
-------
Effects - Crops
4. REFERENCES
Copies of the cited references appear in the Reference Volumes.
Biggs, R. H. and Kossuth, S. V. (1978). Impact of solar UV-B
radiation on crop productivity, Final report of UV-B biological
and climate effects research. Terrestrial. FY 77. Univ.
Florida, Gainesville, FL.
Biggs, R. H., Sisson, W. T., and Caldwell, M. M. (1975). Response
of higher terrestrial plants to elevated UV-B irradiance. In:
Nachtwey, D. S., Caldwell, M. M., and Biggs, R. H. (eds).
Impacts of climatic change on the biosphere, CIAP Monograph 5,
Part 1: Ultraviolet radiation effects. U.S. Dept. Trans.,
Springfield, VA, pp. 4-34 to 4-50.
Caldwell, M. M. (1977). The effects of solar UV-B radiation
(280-315 nm) on higher plants: Implications of stratospheric
ozone reduction. In: Castellani, A. (ed). Research in
photobiology. Plenum Publishing Corp., New York, pp. 597-607.
Duncan, W. G. (1971). Leaf angles, leaf area, and canopy
photosynthesis. Crop Science 11, 482-485.
Evans, L. T. (ed) (1975). Crop Physiology. Cambridge Univ. Press,
pp. 374.
Kossuth, S. V. and Biggs, R. H. (1978). Sunburned blueberries.
Fla. State Hort. Soc. Proc. 9J., 173-175.
Lindoo, S. J., Seeley, S. B., and Caldwell, M. M. (1979). Effects
of ultraviolet-B radiation stress on the abscisic acid status of
Rumex patientia leaves. Physiol. Plant. 45, 67-72.
Loomis, R. S., Rabbinge, R., and Ng, E. (1979). Explanatory models
in crop physiology. Ann. Rev. PI. Physiol. 30, 339-367.
NAS - National Academy of Sciences - (1979). Report of the Committee
on Impacts of Stratospheric Change, ijn "Protection Against
Depletion of Stratospheric Ozone by Chlorofluorocarbons."
December, Washington, D.C.
Ormond, P., Hammer, A., Krizek, D. T., Tibbitts, T. W., McFarlane,
0. C., and Langhans, R. W. (1980). Base-line growth studies of
"First Lady" marigolds in controlled environments. J. Am. Soc.
Hort. Sci. 105, 632-638.
Robberecht, R. and Caldwell, M. M. (1978). Leaf epidermal
transmittance of ultraviolet radiation and its implication for
plant sensitivity to ultraviolet-radiation induced injury.
Oecologia 32, 277-287.
F-3-13
-------
COMMENTS ON THE NATIONAL ACADEMY
OF SCIENCES REPORT: "PROTECTION AGAINST
DEPLETION OF STRATOSPHERIC OZONE BY
CHLOROFLUOROCARBONS."
PREPARED FOR E. I. DU PONT DE NEMOURS & COMPANY, INC,
BY
DR. DAVID M. DAMKAER
AFFILIATE ASSISTANT PROFESSOR
DEPARTMENT OF OCEANOGRAPHY
UNIVERSITY OF WASHINGTON
SEATTLE, WASHINGTON
DECEMBER, 1980
(APPENDIX F-4)
F-4-1
-------
Transmittal letter for
Appendix F-4
University of Washington
WB-10
Seattle, WA 98195
December 16, 1980
Dr. Richard B. Ward
Research Associate
E. I. DuPont DeNemours & Company
Wilmington, Delaware 19898
Dear Dr. Ward:
As I mentioned in our conversation today, I have read and
approved the most recent draft of my comments on the National
Academy of Sciences Report (1979) "Protection against Depletion
of Stratospheric Ozone by Chlorofluorocarbons." You may con-
sider this the final transmittal of these comments. I trust
that they will shed light on at least my opinions of the NAS
review.
Sincerely yours,
David M. Damkaer
-------
Effects - Marine
COMMENTS ON THE NATIONAL ACADEMY
OF SCIENCES REPORT: "PROTECTION AGAINST
DEPLETION OF STRATOSPHERIC OZONE BY
CHLOROFLUOROCARBONS."
TABLE OF CONTENTS
PAGE
1. INTRODUCTION F-4-3
2. SUMMARY F-4-3
3. SPECIFIC COMMENTS F-4-4
A. Marine Phytoplankton - Natural communities
in situ F-4-5
B. Marine Phytoplankton - Single species
in laboratory culture F-4-6
I
C. "Aquatic Microorganisms, protozoa, algae,
and small invertebrates" F-4-7
D. Marine Zooplankton - (invertebrates) F-4-8
E. Marine Zooplankton (fish eggs and larvae) F-4-9
4. CONCLUSIONS F-4-11
5. REFERENCES F-4-13
6. CITATIONS FROM THE NAS REPORT F-4-14
F-4-2
-------
Effects - Marine
COMMENTS BY DR. DAVID M. DAMKAER ON THE NATIONAL ACADEMY OF
SCIENCES REPORT, "PROTECTION AGAINST DEPLETION OF STRATOSPHERIC
OZONE BY CHLOROFLUOROCARBONS."
1. INTRODUCTION
I have examined the title report, particularly the
sections on aquatic ecosystems. The comments that follow
represent my opinion and do not necessarily reflect the opinions
of the National Oceanic and Atmospheric Administration (my
employer), the Environmental Protection Agency (the supporters of
our research on UV effects) , or of the University of Washington
(with which I have a faculty affiliation). That said, I believe
that the National Academy of Sciences report (NAS) was an even-
handed, excellent review of the ozone/ultraviolet radiation
concern, and most of the real conclusions were fairly based on
what is known. It is unfortunate that some of these conclusions
were summarized and oversimplified in sections removed from the
more-detailed reviews. And I would criticize those areas
particularly since it is the summary sections that will be most
widely read by the bureaucracy and legislators.
2. SUMMARY
The compilers of the NAS report frequently admit to an
unfortunate lack of knowledge in the area of effects on aquatic
ecosystems. If the present observations were overwhelming,
consistent, and theoretically sound, perhaps the potential
consequences of increased biologically damaging ultraviolet
radiation (DUV) could be predicted. In the aquatic area, at
least, I believe that the studies as a whole do not yet permit a
conclusion one way or the other, and at this moment I tend to
believe that the effects will not be as severe as some have
concluded. Because there is much emotion connected with this
issue, I feel obliged also to state that I consider myself an
environmentalist, and not a tool of the establishment (whatever
that is). So I am especially concerned that certain unqualified
statements, like those implying that half or more of a population
will die with predicted DUV increases, might jeopardize
evaluations of any evidence. If subsequent research demonstrates
that there may be a serious problem, I hope I am among the first
to say so. In the meantime, I trust that the scientific
community will retain its credibility.
What information there is on UV effects on aquatic
ecosystems is the result of a "crash program," however laudable
certain facets have been. Probably the most valuable outcome of
the crash program has been the development of instruments.
Beyond that, in the application of instruments to the aquatic
area, there have arisen inevitable comparison gaps, and the total
effort cannot yet come up to the sum of its parts.
F-4-3
-------
Effects - Marine
Periodic NAS reviews will be indispensable in unifying
the experimental approach, techniques, and interpretations. In
some respects the inevitable diversity of initial investigations
has been good, but perhaps the time has come to determine which
areas and approaches will yield the more timely and the more
useful results. Comparison gaps will narrow as publications
become available, and as researchers meet and compare results.
Behind whatever factors that have contributed to this, I suppose
lack of money is the most mundane yet most important.
In short, one is frustrated when comparing the results
of different investigations. We are not yet near to
extrapolating the effects on one species to those of another, nor
can we say with much assurance how the same species might react
to UV in another locality. This is also true in studies of most
other parameters. But there are "standard methods" available in
traditional research areas, and UV effects research has not yet
reached that level. Consequently, there is variability in many
basic elements like types, combinations, geometry, filters, and
time-schedule-s of radiation sources; weighting factors; and
experimental design and interpretation. It is no wonder that
comparisons are difficult. An impression might be given by the
NAS report that there is a mass of corroborative data. But it
needs to be stated that there is perhaps not yet a body of
knowledge regarding UV effects on aquatic ecosystems.
3. SPECIFIC COMMENTS
In the NAS report, Key Finding 13 states:
13. Larval forms of several important seafood species,
as well as microorganisms at the base of the marine food
chain, would suffer appreciable killing as a result of a
16 to 30 percent ozone depletion. Present ignorance of
ultraviolet penetration into the waters that they
inhabit and of the depth distribution of the organisms
precludes an estimate of actual losses. (NAS, 1979,
p. 7).
Presumably this statement is based on the investigations reviewed
in the main body of the report. The Summary of Findings with
regard to aquatic ecosystems (NAS, 1979, pp. 68-70)* seems also
to be based on the same investigations. These investigations are
outlined in the NAS report's Appendix B (NAS, 1979, pp.
288-297)*, and include studies on:
A. Marine phytoplankton - natural communities in situ,
B. Marine phytoplankton - single species in laboratory
culture,
Cited pages from the NAS report are attached to these comments
as Section 6.
F-4-4
-------
Effects - Marine
C. "Aquatic microorganisms, protozoa, algae, and small
invertebrates,"
D. Marine zooplankton (invertebrates), and
E. Marine zooplankton (fish eggs and larvae).
A. Marine Phytoplankton - natural communities in situ.
The first group of studies is represented by the
observations of Jitts e_t al. (1976) and Lorenzen (1979). These
investigators observed that (a) the photosynthetic activity of
phytoplankton increased when natural solar incident UV was
removed by proper filters. The NAS conclusion is that (b)
natural solar UV decreases photosynthesis. Statements (a) and
(b) are quasi-equivalent, but the conclusions drawn from each
need not be the same. First of all, the experimental conditions
were not natural (over and above the fact that experimental
conditions never can be), in that UV, which is a natural
environmental component, was removed. This created some level of
optimal conditions and productivity increased. Implicit in
statement (b) , as it stands unqualified, is that further
increases in DUV will further decrease photosynthetic activity.
This is not necessarily true. Though plant systems are well
tuned to radiant energy, and usually respond rapidly and
predictably, there have not yet been sufficient experiments with
phytoplankton at elevated UV levels to eliminate the possibility
of threshold UV doses and dose-rates. That the activity of
phytoplankton increases as their environment goes from "natural"
to improved, does not necessarily mean that this activity will
decrease at the same rate while culture conditions go from
"natural" to somewhat worse. Depending on species and the many
plant pigment possibilities, there may be threshold levels beyond
current ambient levels, toward which UV and activity will not
change in the same degree. Thomson ^t al. (1980) show a
relationship between laboratory enhanced-UV and growth-rate of a
diatom (Melosira nummuloides) . They have estimated this to be a
linear relationship. But the regression is not particularly good
(r = 0.68) , and the individual observations might better fit a
parabolic or other regression, suggesting the existence of a
threshold (more about this under 2, below).
The NAS report points out in another section, 'that if
DUV weighting factors other than erythema (Lorenzen, 1979) or DNA
(Thomson et al., 1980) are used, predictions of increased DUV
(through ozone depletion) might be much less, as, for example,
with the application of a photosynthetic inhibition action
spectrum (NAS, 1979, p. 307-309)*. So that predictions of
decreased photosynthesis (even assuming a constant dose/effect
relationship) would also be less.
*See footnote on p. 4.
F-4-5
-------
Effects - Marine
In addition, there are a number of other uncertainties,
including the lethal limits of UV-B to algal cells, genetic
resilience, repair capability, the proportion of photosynthesis
at various depths and times, and the magnitude of vertical mixing
and exchange of UV-exposed populations.
B. Marine Phytoplankton - single species in laboratory
culture.
The NAS report refers to two studies on individual
phytoplankton species: Thomson, e_t aj^. (1980), and R. C. Worrest
(unpublished). The first study used diatoms that had been
pre-stressed by UV in an experiment on responses of an
attached-algae community. Therefore, there may have been a
danger in underestimating responses if_ there is some individual
variability in UV resistence (as is likely). The investigators
derived a relationship between DUV (DNA-weighted) and
growth-rate. As mentioned above, the observations are not
inconsistent with some threshold in the dose/effect relationship.
This would alter predictions based on the author's linear
regression. Also, as in the _i_n situ studies, they have included
data on enhanced growth-ratesBy" excluding UV. This, as
mentioned above, alters the perception of the dose/effect
relationship; if those (unnatural) levels are omitted, the linear
relationship would have a distinctly different slope, with a much
smaller predicted decrease in growth-rate, for increases in UV
through the levels of concern.
The experiments of Thomson e_t al. (1980) bracketed UV
irradiance levels and daily doses that wo~uld occur off the Oregon
coast (where the experiments were conducted). However, their
conclusions seemed to be based on comparisons w.ith a calculated
UV-| surface daily dose for June at 30 degrees NU) (Mexico), 141
Jm DNA* T^e autnors indicate that M. nummuloides has a wide
distribution. But the possibilities olgenetic, behavioral
(depth), and physiological adaptation with latitude may be very
important (as it is with temperature, for example), so that the
Oregon-derived dose/effect curve may not be valid for subtropical
populations. Beyond possible latitudinal biological effects,
calculations based on higher daily doses may be misleading, since
fixed percentage increases at high doses involve larger absolute
increments, and, therefore, larger growth-rate increments because
of the reported linear relationship. Using the "average clear
spring-summer" daily surface dose at 45 degrees N (Oregon) given
by the _aame authors in another paper (Karanas, e_t al., 1979),
55.9 Jm DNA» as t*ie base^*ne ^or calculating the predicted
(l)The NAS report, based on the Thomson et al., manuscript, says
40 degrees N=-but the published paper says 30 degrees N. I feel
that 141 Jm~ DNA is too high for Oregon, and have assumed that
30 degrees N is meant.
F-4-6
-------
Effects - Marine
growth-rate declines (using the relationship of Thompson et al.,
1980) a 15 percent DUV increase would result in a 3.2 percent
decrease in growth-rate (not 8 percent), while a 32 percent DUV
increase would result in a 7 percent growth-rate reduction (not
16 percent) .
The authors do not show their calculations for the
estimated reduced productivity throughout the upper two meters
off the Oregon coast ("integration of the percent reduction in
productivity for all depths throughout the upper two meters...").
But if they based these calculations on the percentage decreases
from a 30 degree N base DUV, certainly these integrated estimates
would also be high by about 2X.
I have not seen Worrest's (unpublished) report on
photosynthesis in the diatom Chaetoceros didymujs, which is
referred to in the NAS report. Reference is again made to
"reductions" of photosynthesis by natural UV, when compared to
controls receiving n£ UV. It may be more than semantics to
prefer that it be said that the unnatural lack of UV allows
increased photosynthesis. The choice of DUV weighting factors,
and the choice of curve fitting the data, would likely have a
great impact on the conclusions.
C. "Aquatic Microorganisms, Protozoa, Algae, and Small
Invertebrates."
The NAS report (NAS, 1979, p. 289)* refers to a series
of observations on dose/dose-rate responses in a large number of
aquatic organisms (e.g., Nachtwey, 1976). It is stated that UV
doses "comparable" to present solar doses can kill appreciable
percentages of these populations. In subsequent text (NAS, 1979,
p. 323)* it is stated unequivocally that the dosimetry in these
early experiments was misleading, and that the solar doses were
significantly overestimated. These experiments should be
repeated with spectroradiometric measurements. It is unfortunate
that the dose/dose-rate relationships suggested in these
experiments have not been pursued. It is becoming more and more
certain, I believe, that dose-rates or doses by themselves cannot
lead to reasonable predictions, but that it is necessary to
consider dose/dose-rate responses. That is, since there are
repair mechanisms, effects are determined not only by dose, but,
beyond a certain threshold, also by dose-rate.
*See footnote on p. 4.
F-4-7
-------
Effects - Marine
D. Marine Zooplankton (invertebrates).
The report by Damkaer e_t al. (1980) deals with a number
of common marine zooplankton species found near the surface at
least during some time in their life-cycle. At that time they
are presumably vulnerable to DUV. This report describes a
threshold level up to which np_ deleterious effects were noted.
In this respect, qualitatively, these observations agree with the
low dose-rate observations of Nachtwey (1976). These experiments
bracketed natural DUV doses and dose-rates throughout the season
of surface-occurrence. If only the higher dose-rates had been
utilized, no "thresholds" would have been observed. Damkaer
chose an erythemal weighting factor, which, in the ranges
considered, has nearly a linear relationship to the DNA
weighting, so that those results can be qualitatively and roughly
quantitatively compared to other observations using DNA-weight.
Probably the most serious short-comings of those investigations
are the lack of precise knowledge of penetration of DUV into
local seawater, and of the lack of information on the vertical
distributions of the organisms. The conclusion of Damkaer
et £l.(1980) , and of the NAS report, is that the tolerance levels
oT some of these zooplankton would be exceeded (at the surface)
toward the end of the near-surface cycle. This does not imply
catastrophe; some mechanism already limits this surface
occurrence, and perhaps it is UV even now. If so, or if
increased UV becomes a limiting factor, much more research needs
to be done before the extent and implications of those limits can
be determined. (More about such things below). Recent
evaluation of these data, in light of dose/dose-rate responses
and the estimated attenuation of DUV in seawater, suggests that
there would be a significant margin of safety from DUV, unless it
is only the late-occurring populations that are typically
successful, a determination of which is beyond present knowledge.
Another publication concerning marine plankton
invertebrates is that of Karanas et al., (1979). This was not
referred to in the NAS report, but Karanas1 paper will receive
wide distribution and doubtless will be considered by those who
must make decisions on the fate of CFCs.
Karanas studied the responses of the developmental
stages of the copepod Acartia clausii, the adult of which
approaches the morphological grade and size of the shrimp larvae
observed by Damkaer e_t_ al. Therefore, the two studies are
potentially complementary. Unfortunately, the DUV dose-rates
used for stressing the Acartia were unrealistically high (by
about 5X), even though the daily doses were comparable to natural
(Oregon) daily doses. If there were complete reciprocity this
would not matter, but it is doubtful if reciprocity is the rule,
particularly at lower (the natural) daily doses. Karanas et al.,
do not discuss reciprocity or its assumptions. I believe tTTat
they have misapplied Beer's Law, when they calculate the depths
of certain daily doses (Karanas et al., 1979, p. 1113). They
assume Beer's Law holds for doses as well as intensities, which
F-4-8
-------
Effects - Marine
may be true if the doses are given for the same time period. But
I do not believe that they can equate their experimental doses
over less than an hour with a natural daily dose obtained over
(say) 6 hours. There are no depths at which the laboratory
intensities are reached, or at which the necessary doses are
reached in the given experimental time. The conclusions of
Karanas e± al., should be re-examined. Namely, that "enough
radiation penetrates to a depth of about 1 m during every clear
summer day to eliminate about 50 percent of the early and late
naupliar developmental phases." What is the fate of the
remaining 50 percent, which may be somewhat weakened, on the next
day? Why are there still so many copepods?
E. Marine Invertebrates (fish eggs and larvae).
Hunter e_t al. , (1978) studied the sensitivity of eggs
and larvae of anchovy and mackerel. The anchovy were the more
sensitive, and most attention was given to these. I have
difficulty evaluating the reported LD values, especially when
the "survivors" showed retarded growtn and development and when
it seemed "unlikely that any of the damaged survivors, regardless
of dosage, would be able to feed successfully." This suggests
that the experimental doses were much too high. In an
unpublished study cited in the NAS report, Hunter £t_ al., did
lower dose-rate, and found a lower LDcnf indicating that the
effect of total dose depends on dose-rate and, therefore,
reciprocity would not apply. This unpublished study indicated,
but the NAS report did not mention, that, as before, even the
"surviving" larvae at "low" doses would probably not survive. I
believe that more research needs to be done with fish larvae
which would relate total lethal doses to dose-rates.
The NAS report states that the responses at these doses
probably represent a "worst case," say at June surface DUV
levels. That the anchovy eggs and larvae are at the surface from
January to June would indicate that earlier hatchings might avoid
UV stress. Therefore, the situation with anchovy larvae appears
analogous to that of the shrimp larvae (Damkaer et al. , 1980).
Recalculating the erythemal threshold levels Frombamkaer's
report (line A), to compare, in the following table, with the
lowest doses of Hunter et al., (line B) (at which significant
morphological damage occurs), I find:
F-4-9
-------
Effects - Marine
DNA-weight-Dose-rate DNA-weight Dose
"^) (Jm ^)
A .002 190
B .002 530
There is remarkable agreement.
But it should be stressed that Hunter's June DUV levels
are "worst case," and so would have been that of Damkaer et al. ,
if they had only considered end-of-the-season UV levels. There
is no natural law that says all those shrimp and fish larvae must
survive. Organisms place their bets on a lot of squares, and as
the season progresses not all bets will pay off. Populations do
not need to be winners on all squares; some squares never pay
off, and some only pay off once in awhile. All that is required
is a reasonable game. It appears that UV might be an important
regulator at some times. Perhaps increases will shorten the
seasons and decrease the odds then for population success, but I
do not believe that a case can be made one way or the other at
present.
This work of Hunter et al . , shows one very commendable
aspect, and that is the advantage taken in cooperation with the
adjacent Scripps Visibility Laboratory, so that these studies
amount to a joint effort of excellent biologists and excellent
physicists and instrument designers. One can expect good things
from such a combination.
Another hazard and uncertainty in laboratory work should
be mentioned here, and that is the difficulty of maintaining
certain organisms in good condition. For some common animals, it
is not yet possible to keep them alive, 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 mortality in the
controls for anchovy larvae, which were kept in small closed
containers, a poor but necessary approximation of the open ocean.
Karanas et al. , mentioned high temperature as a possible flaw in
their experiments, wherein the Acartia were kept at temperatures
near or exceeding reported optimum limits. No doubt the animals
would do better if kept in the field, but the technical
difficulties of experimenting with most of them there are
overwhelming^ As in some agricultural field tests * ' (NAS,
1979, p. 67) i actual field tolerance of UV is likely more than
that determined in the laboratory.
See footnote on p. 4.
added by Du Pont: This point is discussed by Dr. Biggs
in Appendix F-*3.
F-4-10
-------
Effects - Marine
The NAS report also points to possible UV dangers to
aquaculture, and the lack of information in this area is
mentioned. Since most aquaculture is in shallow bodies of water,
increases in UV may well become a serious problem. In the
manipulation of nature, normal developmental cycles are often
circumvented, and untimely exposures to UV might become the rule
in this area. However, especially in the smaller-scale projects,
there may even now be the possibility to increase production if
ponds are screened from UV. Removal of UV might add a new
dimension toward optimizing growth conditions.
Finally, the NAS report raises two issues of possible
relief for UV-affected organisms. The first is simply behavioral
change, and in the aquatic ecosystem this means avoiding the
near-surface layer. That this is not so simple is acknowledged.
In the first place, the UV would have to be sensed, and this is
not likely. Some recent experiments of our own have indicated
that some shrimp and crab larvae, euphausids, and copepods cannot
detect harmful doses of UV, and do not increase their depth in
large (2 m deep) cylinders. Even if they did alter their
preferred depth, they would likely lose some advantages which the
"normal" depth has offered them.
The second "strategy" to live with UV stress would be to
evolve protective coverings or to evolve new behavioral patterns
(by natural selection rather than by active sensing as above).
This could lead to new depths of occurrence or to new seasons at
the surface. For the larger organisms (above algae and
bacteria) , projected UV increases would probably occur in too
short a time to allow evolutionary changes. There is, however,
the famous case of the darkened moth in England, responding (by
natural selection) to increased air pollution. The time scales
here are not vastly different. However, not every organism would
find a way (or time) to cope.
4. CONCLUSIONS
In summary, the data available to the compilers of the
NAS report are not conclusive one way or the other. This is
admitted in the report, but some statements, (e.g., Key Finding
13) out of context might suggest a stronger foundation in fact.
UV does affect aquatic organisms adversely. To what extent it
will do so in nature, to individual species, populations, and
communities, cannot yet be predicted. There is enough evidence
to warrant (even demand) additional research, and pointing this
out may be the most important contribution of the NAS report.
Periodic reviews can serve to reevaluate new information and to
focus the investigative approach. I believe this is an
extraordinarily exciting research area, and one that is
potentially too important to ignore.
F-4-11
-------
Effects - Marine
The present information on the effects of enhanced solar
UV on aquatic ecosystems is clearly subject to differing
interpretation. With phytoplankton, the experiments under
decreased DUV probably should not be taken to mean that
productivity under enhanced UV will change with the same
magnitude. More serious is the lack of knowledge of
phytoplankton action spectra, and applying action spectra
applicable to other organisms can suggest a very wide range of
effects, from negligible to large. For zooplankton, some
experiments have not used DUV doses or dose-rates comparable to
solar UV doses, and have greatly underestimated experimental DUV
doses and, therefore, have overestimated the effects. Recent
experiments have yet to deal thoroughly with the likely
differences between reactions in the laboratory and nature, the
real decreases in UV with depth, the vertical distribution of
organisms, and the probabilities of UV levels with season and
seasonal occurrences of organisms.
If there can be any conclusion regarding UV effects on
natural populations, I believe it would be to say that the season
of near-surface occurrence might be shortened. This could become
a common factor in fluctuations of local populations in some
years, and may even lead to geographic restrictions for some
species. Against the other natural variability in aquatic
ecosystems, it would be very difficult to measure UV-caused
fluctuations. In the Earth's past, the rule has been expansion
and contraction of populations, so that in itself does not spell
catastrophe. As in other environmental questions, one must
decide what to trade. Might the UV question boil down to
refrigerators vs. shortened seasons and restricted distributions?
The same question arises with corn, coal and condominiums.
F-4-12
-------
Effects - Marine
5. REFERENCES
Copies of the cited pages of the NAS report appear in
Section 6 of this appendix. Copies of other references below
appear in the Reference Volumes.
Damkaer, D. M., Dey, D. B., Heron, G. A., and Prentice, E. F.
(1980). Effects of UV-B radiation on near-surface
zooplankton of Puget Sound. Oecologia, 44, 149-158.
Hunter, J. R., Taylor, J. H., and Moser, H. G. (1979). Effect of
ultraviolet irradiation on eggs and larvae of the northern
anchovy, Eugraulis mordax, and the Pacific mackerel, Scomber
japonicus^duringtheembryonic stage. Photochemistry and
PhotobioTogy, 29, 325-338.
Jitts, H. R., Morel, A., and Saijo, Y. (1976). The relation of
oceanic primary production to available photosynthetic
irradiance. Aust. J. Mar. Freshwater Res., 27, 441-454.
Karanas, J. J., Van Dyke, H., and Worrest, R. C. (1979).
Midultraviolet (UV-B) sensitivity of Acartia clausii
Giesbrecht (Copepoda). Limnol. Oceanogr., 24 (6) , 1104-1116.
Lorenzen, C. J. (1979). Ultraviolet radiation and phytoplankton
photosynthesis. Limnol. Oceanogr., ^4_(6) , 1117-1120.
Nachtwey, D. S. (1976). Potential effects on aquatic ecosystems
of increased UV-B radiation. Proceedings of the Fourth
Conference on the Climatic Impact Assessment Program, Hard,
T. M. and Broderick, A. J., eds., pp. 79-86.
DOT-TST-OTS-75-38, U.S. Department of Transportation,
Washington, D.C.
NAS-National Academy of Sciences-(1979). Report of the Committee
on Impacts of Stratospheric Change jri "Protection against
depletion of stratospheric ozone by chlorofluorocarbons.
Washington, D.C.
Thomson, B. E., Worrest, R. C., and Van Dyke, H. (1980). The
growth repsonse of an estuarine diatom (Melosira nummuloides
[Dillw.] Ag.) to UV-B (290-320 nm) radiation.Estuaries,
3(1), 69-72.
F-4-13
-------
Effects - Marine
6. CITATIONS FROM THE REPORT OF THE NAS COMMITTEE on Impacts of
StratosphericChange,"ProtectionAgainstDepletion of
Stratospheric Ozone by Chlorofluorocarbons." (Pages 67-70,
288-299, 307-309, 322-324).
F-4-14
-------
Effects - Marine
67
SUMMARY OF FINDINGS
The major information available on UV sensitivity of
organisms, which is helpful for evaluating the possible
effects of ozone depletion, is reviewed in Appendix B.
This information covers agricultural crops and land vege-
tation, domestic and wild animals, nonagricultural land
ecosystems, and aquatic organisms and ecosystems. The
information is summarized briefly here for reference.
h'ith Regard to Agricultural Plants
I. Tests of more than 100 species or varieties of
species in controlled environment growth chambers indi-
cate that approximately 20 percent are sensitive to daily
UV-B doses of the order of those delivered by Florida sun-
shine at present ozone levels, while 20 percent were re-
sistant to doses four times greater than this, and the
remaining 60 percent showed some intermediate sensitivity.
These tests would indicate that a significant fraction of
the present agricultural varieties are at present under
UV stress and would suffer decreased production with a
16 percent ozone reduction.
2. However, 15 species and varieties tested in the
open field appeared more UV resistant than plants grown
in laboratory environmental growth chambers. Where it is
possible to make comparisons of the same plants, the
differences in resistance are on the order of fourfold.
If this result turns out to be general, the higher sensi-
tivity indicated in the growth chambers does not repre-
sent open-field behavior, and the expected consequences
of ozone depletion become considerably less than those
tests would indicate. Nevertheless, some species (sugar
beets, tomatoes, mustard, corn) still appear to be af-
fected—although the field experiments were necessarily
less well controlled, and the results less clear-cut, than
were the chamber studies.
3. The existence of relatively more resistant varie-
ties of the same species offers the possibility of select-
ing plants better adapted to UV-B stress. Efforts could
be made to combine the UV resistance trait with the other
characteristics desired in these plants (high crop yields
and resistance to certain diseases, for example) through
systematic breeding. However, in much of the agriculture
of undeveloped countries, introduction of new varieties
would be difficult, and any emergence of more UV-resistant
F-4-I5
-------
Effects - Marine
68
varieties might have to depend on the slower process of
natural selection after the ozone decrease had occurred.
Kith Regard to Domestic and Wild Animals
4. The very incomplete information suggests that the
impact of ozone depletion on domestic and wild animals
would not be large. An increase in conjunctival carci-
noma (cancer eye) affecting some cattle breeds could be
expected, but the fraction of cattle affected would be
small.
h'i th Regard to Nonagricultural Land Ecosystems
5. Perturbations of nonagricultural ecosystems could
ultimately be as important as more direct effects on
human health or human food supplies, but .there is almost
no information with which to assess the consequences of
increased solar-UV radiation on such ecosystems. Because
nonagricultural plants show about the same range of sen-
sitivity as crop plants, reduced productivity of some
forest and grazing-land plants might occur. However,
change in the species composition of these systems seems
a more likely response. Such a change would alter the
character of the vegetation and might qualitatively change
the products provided. However, UV sensitivities and
their range of variation in wild species are known in so
few cases, and the theoretical modeling of ecosystems is
at present so primitive, that such changes cannot be pre-
dicted confidently.
The fact that ozone depletion would be a worldwide
perturbation makes its consequences more worrisome than
effects on simply a local environment. The presumed
effectiveness of past evolutionary adaptation, which de-
veloped over millennia, and the human need for a stable
framework in which to operate, argue that widespread
changes, occurring rapidly compared with the usual evo-
lutionary time scale, would be more likely to be harmful
than helpful. Consequently, where a prospective man.-made
change cannot be proven unequivocally beneficial, it
should be viewed with some suspicion.
With Regard to Aquatic Ecosystems
6. Experimental studies on over 60 aquatic micro-
organisms, protozoa, algae, and small invertebrates
F-4-16
-------
Effects - Marino
69
indicate that roost of them are sensitive to current water-
surface levels of DUV radiation. Because they currently
thrive in nature, attenuation of DUV by natural waters,
especially in productive areas, may play an important
role in their survival. (The degree of penetration of
DUV into such waters is not well characterized.)
7. UV-B irradiation of young anchovies by means of
lamps indicates that in surface waters this commercially
important species lives normally near its UV tolerance
limit. UV increases similar to those expected with a 16
percent ozone-layer reduction would, as a "worst case,"
kill over 50 percent of the anchovies in the top 10 m of
the clearest ocean water or else would require them to
substantially readjust their usual water depth to dimin-
ish UV exposure. Such avoidance, if it were to occur,
would result in a substantial shift in ecological rela-
tionships .
Similar studies show that mackerel are more UV resis-
tant than anchovies. .The "worst case," corresponding to
organisms exposed near the water surface, indicates essen-
tially no losses.
Analogous studies of crab and shrimp larvae indicate
that they too are near their DUV tolerance limit, con-
sidering their location in the water column (near surface)
and their reproduction/development season (late winter,
early spring). A 16 percent ozone-layer reduction might
substantially shorten their reproductive/development
season with unknown, but probably detrimental, conse-
quences to productivity.
This very limited sampling gives little perspective on
the UV sensitivity range of commercially important aquatic
organisms. If we presume that such a small sample is
roughly representative, an appreciable fraction of fishery
species young would be vulnerable to the DUV levels ex-
pected with a' 16 percent ozone reduction. They would
either be injured or would have to seek deeper water,
with unknown consequences on their subsequent prosperity.
8. The final conclusion depends on where the burden
of proof is presumed to lie: one cannot convincingly
demonstrate that a 16 percent ozone reduction would cause
significant decrease in yields from agricultural crops or
from nonagricultural land or aquatic ecosystems. Neither
is it clear that a 16 percent ozone reduction would not
entail important losses in these yields. The uncertainty
stems partly from ignorance of the proper weighting func-
tion for different wavelengths (i.e., action spectrum) to
be used in comparing solar and artificial source
F-4-17
-------
Effects - Marine
70
irradiations but mostly from the limited scope, and in
some cases incomplete design, of experiments and the lack
of corroboration by independent investigators. These
lacks, in turn, stem from the haste with which the re-
search efforts have been put together and the absence of
a systematic longer-range program to answer well-defined
Questions.
MAJOR RESEARCH ISSUES
Progress has been made on several questions since the last
full CISC report was published in 1976, although less than
could have been made, considering the need for information
to evaluate the biological effects of DUV.
The recent advances are:
1. Larger numbers of agricultural plant varieties
have been experimentally tested by radiation supplement
for their sensitivity to UV-B. These tests have confirmed
the wide variation in sensitivity among different plant
varieties that was suspected earlier and have also indi-
cated that the sensitivity depends on conditions of test-
ing, but the tests leave the true sensitivity profile of
the most important agricultural plants in doubt.
2. Tests with larval forms of several fishery species
also indicate that wide variation in sensitivity occurs
from one to the other and that some species have signifi-
cant sensitivity. The very small numbers tested, and
uncertainty about the likelihood of UV adaption or UV
avoidance by the animals under natural conditions, as
discussed in Appendix B, leaves the significance of these
results still unclear.
The points stated above need further investigation by
a sufficient number of independent investigators and need
to be pursued over a sufficient time to overcome the
spotty sampling, the possible statistical fluctuations in
small experiments, and, in some cases, the deficiencies
of experimental design that now limit interpretation of
the results. The effect of environmental conditions on
the UV sensitivity of plants needs to be explored.
Because the action spectra for different UV effects on
most higher plant and animal species are at present not
known, emphasis should be directed to their determination.
The results are needed to establish the weighting func-
tions for predicting consequences of ozone depletion.
F-4-18
-------
Effects - Marine
286
of those that are more sensitive (Caldwell, 1977; Fox and
Caldwell, 1978). Such changes in competitive balance
might amplify rather small sensitivity differences, af-
fecting subtle processes (pollination, relationships be-
tween predators and their prey, and biological fixation
of nitrogen, for example).
Changes in the competitive balance of plant species
have actually been demonstrated, using UV-B-producing
lamps to supplement normal sunlight (Fox and Caldwell,
1978). Although these supplements were greater than the
UV-B increase that would accompany currently expected de-
creases in stratospheric ozone, the results indicate some
of the processes that may be involved in species replace-
ments. Recently, A. Bogenrieder, B. Bruzek, and S. Kiliani
(University of Freiburg, unpublished data) showed that the
balance between plant species can also be altered when both
UV-B and the longer-wavelength, less biologically active
UV-A are filtered out of normal sunlight. This occurred
even at 50° N latitude, where the DUV dose rate is con-
siderably smaller than at lower latitudes. Such experi-
ments suggest that changes in nonagricultural ecosystems
may accompany small UV-B increases; but because so few
nonagricultural plant and animal species have been evalu-
ated, it is impossible at this time to predict which spe-
cies might benefit at the expense of others.
It is usually not easy to decide whether a specified
change in the species balance of an ecosystem should be
considered detrimental or beneficial to human welfare.
However, the presumed effectiveness of past evolutionary
adaptation, and the human need for a stable framework with-
in which to operate, make it prudent to regard any wide-
spread changes as likely to be harmful.
AQUATIC ORGANISMS AND ECOSYSTEMS
Marine and freshwater ecosystems are a vital element in
the complex environment that makes life as we know it pos-
sible. The finfish and shellfish used for human food, and
the organisms on which they feed, are obviously important.
In addition, aquatic organisms decompose organic wastes
(thus depolluting the waters), and marine algae serve as
a source of oxygen and as a sink for carbon dioxide.
Amounts of the latter gas in the atmosphere are measurably
increasing at this time, because of the very heavy, world-
wide combustion of fossil fuels.
Current levels of solar-UV radiation at the water sur-
face can be lethal to marine bacteria, and current surface
F-4-19
-------
Effects - Marine
289
UV levels, at wavelengths shorter than the cutoff point
for glass or Mylar (below 340 or 315 nm, respectively),
significantly reduce photosynthesis in marine algae when
these organisms are collected at various depths and brought
to the surface (Jitts et al., 1976; Lorenzen, 1975). Labo-
ratory studies on a variety of aquatic organisms support
these sunlight studies. Simulated doses of DUV radiation
at the water surface, produced by fluorescent sunlamps with
filters (see Appendix E), indicate that only a few hours of
radiation comparable with that in the near-noon summer sun-
light can kill 50 percent of the population in most of the
60 or so species tested (bacteria, .algae, protozoa, small
invertebrates) (Calkins, 1975; Calkins et al., 1978; Nacht-
wey, 1976). These studies thus suggest that aquatic eco-
systems may be extremely vulnerable to ozone depletion/
increased UV-B. However, the dosimetry for these experi-
ments was done with a Robertson-Berger Sunburning Ultra-
violet Meter (R-B Meter). As discussed in Appendix E, the
use of this meter for comparing DUV from filtered fluores-
cent sunlamps with that of sunlight may significantly over-
estimate sunlight sensitivity.
Most recent studies have employed better dosimetry, in-
volving spectroradiometric measurements weighted by either
an erythema or DNA-effective weighting function (see Appen-
dix D). Several of these studies provide quantitatively
useful data for assessment.
Photosynthesis in a marine diatom (Chaetoceros djdymus)
decreases exponentially with increasing DNA-weighted dose
(R. C. Worrest, Oregon State University, unpublished).
Recalculations of these data indicates that current summer-
time daily surface DUV doses at 40° N latitude at the water
surface can reduce photosynthesis (carbon dioxide uptake)
by 40 percent relative to controls receiving no UV-B. A
44 percent increase in DUV (such as would accompany a 16
percent ozone layer reduction at 40° N latitude) would de-
crease photosynthesis by 28 percent from that expected in
nature with current summertime daily surface doses. These
results assumed that the DNA-effective dose is the appro-
priate one to use, but this may not be the correct weight-
ing function for photosynthetic inhibition. Most important,
however, is the protection these organisms receive from
being located at some depth in the water.
Similar studies by Thompson et al. (1979) examined the
effects of simulated solar UV-B on the growth rate (cell
divisions/day) in an estuarine chain-forming diatom (Melo-
sira nurnnuloides), which attaches to surfaces and floats
in the water column. DUV doses comparable with current
F-4-20
-------
Effects - Marine
290
summer daily surface doses at 40° N latitude reduced the
growth rate by 35 percent relative to controls receiving
no UV-B. Any considerable increase in this DUV dose would
then be expected to reduce the growth rate for organisms
at the water surface.
These two studies indicate th?.t, with current surface
doses, these organisms are under UV radiation stress. Yet
they thrive in nature. This may mean that they must con-
stantly contend against the UV radiation in order to exist
in environments having other advantages and are consequently
not producing at the maximum possible rate. The situation
could be analogous to that of some terrestrial plants,
which show spectacular increases in yield if grown hydro-
ponically under optimum conditions.
In another quantitatively useful study, D. M. Damkaer
(NOAA Pacific Marine Environmental Laboratory, Seattle,
unpublished BACER report) examined the DUV sensitivity of
larval forms of shrimp and crab (two species). These or-
ganisms appear to be less sensitive than the bacteria,
protozoa, and algae mentioned above, in that several days
rather than several hours exposure are required to produce
an effect. Moreover, the organisms show a threshold of
tolerance to the killing effects of DUV, with no reduction
in rates of development at sublethal doses, compared with
control organisms. The various tolerance doses (weighted
with an erythemal weighting function) are plotted in Figure
B.I. In this plot, a short, horizontal line is positioned
vertically to show the tolerance dose of the organisms
(measured at the water surface) and horizontally to show
the period of the year during which the larval forms de-
velop. These tolerance doses are superimposed on curves
showing the annual DUV cycle of average daily doses at
the location of the experimental studies, assuming various
ozone-layer reductions.
During most of the reproductive/developmental period,
the tolerance threshold daily doses determined for these
organisms are greater than the average daily doses experi-
enced at current ozone levels, but there is not much margin,
the organisms being near their limits of tolerance. Al-
though daily DUV doses resulting from ozone-layer reduc-
tions as large as 40 percent would not exceed the daily
tolerance limits for part of reproductive/development
period (and would thus leave a window of safety at the
beginning of each group's surface dwelling season), this
surface-tolerating season could be significantly shortened.
There would be some shortening of it even at a 16 percent
ozone-layer reduction. Whether or not the populations
F-4-21
-------
Effects - Marine
291
4.0 f-
O
o
<
o
1.0-
FEB
MAR
JUNE JULY
AUG
SEPT
OCT
FIGURE B.I Daily solar DUV dose at the water surface
(erythermal action spectrum) in arbitrary units for vari-
ous ozone thicknesses at Manchester, Washington. Horizon-
tal bars, showing the time of year during which larvae of
various species develop, are positioned vertically to in-
dicate the tolerance daily dose.
could endure substantially reduced periods of reproduction/
development is not known. Success of any year's group of
larvae (the "year-class") depends on the timing of a great
number of factors besides UV-b levels (e.g., weather, food
supplies, and predators). Early larvae may do well one
year, but only late larvae may survive in a subsequent
year. An additional stress like increased UV-B is not
likely to be beneficial.
Hunter et al. (1979) examined the sensitivity of the
eggs and larvae of the northern anchovy and the Pacific
mackerel during their embryonic stages when they dwell
near the water surface. Using different types of filters
they established that the killing effect of UV-B was con-
sistent with a DNA-damage action spectrum. The doses
lethal to half the population (LDsg's) for fish embryos
irradiated at varying dose rates 7 hours per day for 4 to
5 days were 1150 J m~2 (joules per square meter) for an-
chovy and 1576 J m~2 for mackerel weighted for DNA damage
effectiveness, as described in Appendix D. Anchovy sur-
viving the lowest dose used (760 J m~2) showed brain and
eye lesions and retarded growth and.development. These
F-4-22
-------
Effects - Marine
292
results are compared in Figure B.2 with the DNA-effective
radiation penetrating into various natural waters in mid-
June at San Diego, California (33° N latitude) (Smith and
Baker, 1979). The comparisons indicate that brain and eye
damage and retardation of growth should occur in anchovy
larvae at the surface in about 4 days in June if the ozone
layer were reduced by 25 percent. Since spawning of these
creatures occurs mostly over a period (January-June) in
which the solar DUV dose rate increases by a large factor,
this represents a worst case.
In a subsequent study J. Hunter, 0. Taylor, and H. .G.
Moser (National Marine Fisheries Service, La Jolla, Cali-
fornia, unpublished BACER report) have extended their
studies with anchovy past the embryonic stage and into
the feeding stage. They employed a similar experimental
protocol but with lower daily dose rates, expecting that
the same total dose given over a longer period of time
might allow some repair and thus yield a higher 1,059.
Surprisingly, they found that exposures over 12 days at
6 h/day at various low dose rates yielded a lower 1,050,
675 J m (with a DNA damage weighting function). Refer-
ence to Figure B.2 indicates that the LD^o level is reached
at 4 m in clear ocean waters with current irradiances. If
a 25 percent ozone-layer reduction were to occur, the LDso
level is reached at about 7.5m. Growth inhibition, which
occurs at doses greater than 530 J m~^ (DNA-effective)
given over 12 days could be expected to occur under present-
day irradiances down to 5.5 m, and down to 8.5 m with a 25
percent ozone-layer reduction. Eggs and larvae of anchovy
commonly occur at these depths during this period of their
development. These calculations, made for a somewhat larger
ozone depletion than now envisaged, need to be repeated for
the 16 percent depletion figure but give some indication
of the general magnitudes of effects that might occur.
At first glance the large effects observed with doses
simulating current surface levels seem incompatible with
the productivity observed in nature. However, other fac-
tors, such as predators or food supply, may so limit the
population that UV-B becomes less important. Also, there
are reasons why these data may overestimate the- UV sensi-
tivity of the organisms tested. An obvious one is that
the organisms probably spend part of their time well below
the surface, where DUV is attenuated (Figure B.2). Another
is that in nature photoreactivating light is available at
an intensity more than an order of magnitude above usual
laboratory levels, which might increase photorepair of
DNA damage.
F-4-23
-------
Effects - Marine
293
(aim • cm! • 0.24
?88 I X 4 days- 1150)
o
o
o
Q
i
g
UJ
5
<
Q
DEPTH (m)
FIGURE B.2 Daily DUV dose (in mid-June at 33° N latitude
DNA-damage action spectrum) as a function of water depth in
three different types of ocean water. Curves show the dose-
depth relation with "standard ozone" thickness (0.32 atm cm)
and a 25 percent depletion of this value (0.24 atm cm) in
clear open ocean waters of low productivity (i.e., chloro-
phyll-a content 0.025 mg m~^ ( ); in moderately produc-
tive waters (chlorophyll-a content 0.5 mg itT^) (---); and
in the latter waters when a high concentration of dissolved
organic material_is also present ( ). The horizontal
line at 288 J m~^ day"1 shows the daily dose killing half
of an anchovy larval population in 4 days (LDso = 1150 J
m~2). The horizontal line at 190 J m~2 day"1 shows a daily
dose producing retardation and damage of these larvae in
4 days (760 J m~2). The horizontal line at 56 J m~2 day"1
shows the daily dose killing half of an anchovy larval
population irradiated for 12 days (LDso = 675 J m~2), a
time extending beyond the embryonic into the feeding stage.
The horizontal line at 44 J m~2 day"1 shows the daily dose
producing significant growth inhibition in 12 days (530 J
m"2).
F-4-24
-------
Effects - Marine
294
The actual location of organisms in the water column
and the amount of penetration of UV-B into natural waters
are not well known. In the case of anchovy irradiated for
12 days at current DUV irradiances, 50 percent would die
if they remained at a depth of 4 m or less in clear ocean
water or if they existed for the same period at 0.8 m or
less in the most UV-attenuating waters studied (Figure B.2)
'Even though the larvae of these animals occur through the
upper mixed layer, from the surface to the therrnocline,
the proportion of the population at various depths is not
known, and the effect on the population cannot now be pre-
dicted. Moreover, although the calculated irradiances of
Smith and Baker (1979) provide some indication of penetra-
tion of DUV radiation, their values are based on extrapo-
lations of measurements in only four water types at 310 nm
and above, and there have been no measurements in the more
effective spectral region for DNA-damage. The biological!}
important estuarine and near-coast environments have also
not been adequately studied. The near-coast measurements,
made in moderately productive waters, with a relatively
high concentration of dissolved organic material, indicate
a very strong attenuation of DUV (dotted lines, Figure B.2).
The extent to which this measurement is typical of the pro-
ductive regions of the globe is not known. Information on
the penetration of DUV into the waters where organisms
actually exist is critically important to the assessment
of the impact of increased DUV on seemingly very sensitive
aquatic organisms.
The impacts of ozone depletion on fishery species raise
a concern for supplies of human food derived from aquacul-
ture. Half the total protein intake for over 200 million
people in Southeast Asia, and additional millions in other
regions, is derived from pond-grown fish and shellfish
species. These organisms are reared in shallow ponds and
are dependent on microscopic plankton for food supply.
However, there has not been any experimental work to ex-
plore this potential problem.
The effect of more intense photoreactivating light will
depend on the degree to which DNA repair is light-limited
in the species in question. This is not known at present.
Directly determining whether photoreactivating light as
it occurs in nature strongly affects the sensitivity to
DUV has been difficult. The most practical source of pho-
toreactivating light of solar intensities is the sun it-
self (full-scale solar simulators being quite expensive
both to construct and maintain). Yet experiments in green-
houses or in the field suffer from considerable variability
F-4-25
-------
Effects - Marine
295
in environmental parameters, particularly temperature.
R. C. Worrest, Oregon State University (unpublished) was
unable to verify, in a greenhouse study, results obtained
in the laboratory, because of the substantial lack of re-
producibility from one greenhouse experiment to another,
and the question remains open.
In any event, we can say that any ozone-layer reduction
will lead to an increase in DUV at all depths in the water
column. Moreover, because photosynthetically active radia-
tion (PAR) will not change appreciably with ozone reduction,
the ratio of DUV to PAR will also increase at all depths in
the water column, with potential consequences to phytoplank-
ton. This would be true even if these organisms could mi-
grate deeper to avoid the increased DUV. Thus, on the basis
of current knowledge, we must accept the possibility that
ozone depletion could seriously affect the aquatic biosphere.
AVOIDANCE AND ADAPTATION
Most of our quantitative discussion has tacitly assumed
that the effects of a long-term depletion of stratospheric
ozone on plants and animals would be determined entirely
by the average increase in solar UV-B at the earth's sur-
face, i.e., that neither compensating behavioral changes
to reduce UV exposure nor adaptations to decrease sensi-
tivity would take place. Clearly, many organisms have
evolved strategies that reduce the UV exposure of their
sensitive structures. One example is protective outer
coverings, such as fur, feathers, or special pigments.
Some desert reptiles avoid daylight altogether and are
active only in the evening and nighttime hours, but those
species that are active by day have melanin accumulations
in their outer tissues that can prevent UV-B penetration
to physiologically sensitive points (Porter, 1967; 1975).
However, it is not clear how rapidly additional protective
strategies might emerge.
Aquatic organisms in water of sufficient organic con-
tent could readily find a much more protected environment
only a few meters deeper (cf. Figure B.2), but increased
UV-B radiation may not be directly sensed by these organ-
isms. Instead, they may (like human beings) respond to
concomitant visible light or radiant heat and rely on this
cue for avoidance of excessive UV exposure in normal sun-
light. Since changes in the amount of ozone in the atmo-
sphere would affect essentially only the UV-B radiation,
without an accompanying change in the directly perceived
cue, behavioral avoidance might be slow to evolve.
F-4-26
-------
Effects - Marine
296
Even if such avoidance did occur, changing the time of
day that an organism was active (in the case of animals),
or changing its location within an ecosystem, would alter
its relations with other species and might change the com-
petitive balance of species, the susceptibility of the
organism to predators or parasites, or, in the case of
plants, the availability of pollinators. Since most spe-
cies have .evolved over millenia to their niche in an eco-
system, such changes in relationships with other species
would probably not be advantageous.
Adaptation might occur as either a short-term physio-
logical response or as a longer-term evolutionary change.
An example of the former is the tanning of light-skinned
humans, in which the aggregation of melanin pigments in
outer layers of the skin, in response to a UV stimulus,
provides an improved barrier to subsequent UV-B penetra-
tion. Such protection is limited, compared with the
screening regularly available to dark-skinned races but
is nevertheless significant, as increased sunburn resis-
tance after tanning shows. There is limited evidence that
some plant species similarly respond to UV radiation by
increasing the amounts of certain pigments in their leaves
to provide an analogous protection (Robberecht and Caldwell,
1978). The increased physical screening in plants seems
to vary considerably from species to species and is also
very likely limited (Van et al., 1977; Robberecht and Cald-
well, 1978) .
Evolutionary Response to Increased UV-B Radiation
Considering that plants and animals have evolved effective
and sometimes very intricate mechanisms to cope with pres-
ent levels of solar UV-B radiation, the possibility exists
that some species could evolve an increased resistance.
Most natural species of plants and animals carry a certain
genetic variability, which provides the background on which
natural selection can act in the face of a new stress to
evolve a better adapted species. If all circumstances are
conducive, a rather rapid evolution of resistance can take
place, as in the case of the 250 species of insects and
crop pests that have evolved increased tolerance to widely
used insecticides—sometimes in periods of a few years
(Brown, 1960). Some plants have evolved tolerances toward
heavy metals, such as lead in mine tailings (Antonovics,
1975). On the other hand, many species have failed to
evolve resistance to new environmental perturbations—
F-4-27
-------
Effects - Marine
297
especially where these occurred rapidly—as exemplified
by continued sensitivity of falcons and ospreys to DDT.
For some organisms, the 35 years in which half the ozone
reduction may take place would only span a few generations
and is too short for such a mechanism to respond.
The time available for change is a crucial factor
(Antonovics, 1975), and in general, a higher probability
of evolving resistance is expected for species that are
very short-lived, have very large population numbers, and
naturally engage in crossbreeding. These characteristics
account for the success of insects in increasing their
resistance to insecticides, compared with the failure of
comparatively long-lived predatory bird species to do this.
However, no assured statements can be made beyond these
general expressions of probability.
REFERENCES
Antonovics, J. 1975. Predicting evolutionary response of
natural populations to increased UV radiation. CIAP
Monograph 5, Impacts of Climatic Change on the Biosphere,
Part 1, Ultraviolet Radiation Effects, D. S. Nachtwey,
M. M. Caldwell, and R. H. Biggs, eds. pp. 8-3 to 8-26.
DOT-TST-75-55, U.S. Department of Transportation, Wash-
ington, D.C.
Bogenrieder, A., and R. Klein. 1977. Die Rolle des UV-
Lichtes beim sog. Auspflanzungsschock von Gewachshaus-
setzlingen, Angew. Hot. 51:99-107.
Brandle, J. R. , W. F. Campbell, W. B. Sisson, and M. M.
Caldwell. 1977. Net photosynthesis, electron transport
capacity, and ultrastructure of Pisum sativum L. exposed
to ultraviolet-B radiation, Plant Physiol. 60:165-169.
Brown, A. W. A. 1960. Mechanisms of resistance against
insecticites, Ann. Rev. Entomol. 5:301-326.
Caldwell, M. M. 1977. The effects of solar UV-B radiation
(250-315 nm) on higher plants: implications of strato-
spheric ozone reduction, in Research in Photobiology,
Plenum Press, N.Y. pp. 597-607.
Calkins, J. 1975. Effects of real and simulated solar
UV-B in a variety of aquatic microorganisms—possible
implications of elevated UV irradiance. CIAP Monograph
5, Impacts of Climatic Change on the Biosphere, Part 1,
Ultraviolet Radiation Effects, D. S. Nachtwey, M. M.
Caldwell, and R. H. Biggs, eds. pp. 5-33 to 5-71. DOT-
TST-75-55, U.S. Department of Transportation, Washing-
ton, D.C.
F-4-28
-------
Effects - Marine
298
Calkins, J., J. A. Barcelo, P. Grigsby, and S. Martin.
1978. Studies of the role of solar ultraviolet radiation
in "natural" water purification by aquatic ecosystems.
Research Report No. 108, University of Kentucky Water
Resources Research Institute, Lexington, Kentucky.
Dickson, J. G., and M. M. Caldwell. 1978. Leaf develop-
ment of Rumex patientia L. (Polygonaceae) exposed to UV
irradiation (280-320 nm), Am. J. Bot. 65:857-863.
Fox, F. M., and M. M. Caldwell. 1978. Competitive inter-
action in plant populations exposed to supplementary
ultraviolet-B radiation, Oecologia 36:173-190.
Hunter, J. R., J. H. Taylor, and H. G. Moser. 1979. Ef-
fect of ultraviolet radiation on eggs and larvae of the
northern anchovy, Engraulis mordax, and the Pacific
mackerel, Scomber japonicus, during the embryonic stage,
Photochem. Photobiol. 29:325-338.
Jitts, H. R., A. Morel, and Y. Saijo. 1976. The relation
of oceanic primary production to available photosynthetic
irradiance, Aus. J. Marine Freshwater Pes. 27:441-454.
Lorenzen, C. 1975. Phytoplankton responses to UV radia-
tion and ecological implications of elevated UV irradi-
ance. CIAP Monograph 5, Impacts of Climatic Change on
the Biosphere, Part 1, Ultraviolet Radiation Effects,
D. S. Nachtwey, M. M. Caldwell, and R. H. Biggs, eds.
pp. 5-83 to 5-91. DOT-TST-75-55, U.S. Department of
Transportation, Washington, D.C.
Macdonald, E. J. 1975.. Association between cancer eye
and solar radiation. CIAP Monograph 5, Impacts of Cli-
matic Change on the Biosphere, Part 1, Ultraviolet Radia-
tion Effects, D. S. Nachtwey, M. M. Caldwell, and R. H.
Biggs, eds. pp. 6-20 to 6-22. DOT-TST-75-55, U.S. De-
partment of Transportation, Washington, D.C.
Nachtwey, D. S. 1976. Potential effects on aquatic eco-
systems of increased UV-B radiation, in Proceedings of
the Fourth Conference on the Climatic Impact Assessment
Program, T. M. Hard and A. J. Broderick, eds. pp. 79-86.
DOT-TST-OST-75-38, U.S. Department of Transportation,
Washington, D.C.
Pitts, D. G., and A. P. Cullen. 1976. Ocular ultraviolet
effects from 300 nm to 400 nm in the rabbit eye. NIOSH
Report for Contract CDC-99-74-12, October.
Porter, W. P. 1967. Solar radiation through the living
body walls of vertebrates with emphasis on desert rep-
tiles, Ecol. Monogr. 37:273-296.
Porter, W. P. 1975. Ultraviolet transmission properties
of vertebrate tissues, CIAP Monograph 5, Impacts of Cli-
matic Change on the Biosphere, D. S. Nachtwey, M. M.
F-4-29
-------
Effects - Marine
299
Caldwell, and R. G. Biggs, eds. pp. 6-3 to 6-15. DOT-
TST-75-55, U.S. Department of Transportation, Washington,
D.C.
Robberecht, R., and M. M. Caldwell. 1978. Leaf epidermal
transmittance of ultraviolet radiation and its implica-
tions for plant sensitivity to untraviolet-radiation
induced injury, Oecologia 32:277-287.
Sisson, K. B., and M. M. Caldwell. 1976. Photosynthesis,
dark respiration, and growth of Rumex patientia L. ex-
posed to ultraviolet irradiance (288 to 315 nanometers)
simulating a reduced atmospheric ozone column, Plant
Physiol. 58:563-568.
Sisson, W. B., and M. M. Caldwell. 1977. Atmospheric
ozone depletion: reduction of photosynthesis and growth
of a sensitive higher plant exposed to enhanced UV-B
radiation, J. Exp. Bot. 28:691-705.
Smith, R. C., and K. S. Baker. 1979. Penetration of UV-B
and biologically effective dose-rates in natural waters,
Photochem. Photobiol. 29:311-324.
Thompson, B. E., R. C. Worrest, and H. van Dyke. 1979.
The growth response of an estuarine diatom (Melosira
nummuliades [Dill.] Ag.) to UV-B (290-320 nm) radiation,
Estuaries (in press).
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.
Van, T. K., L. A. Garrard, and S. H. West. 1977. Effects
of 298 run-radiation on-photosynthetic reactions of leaf
discs and chloroplast preparations of some crop species,
Environ. Exp. Bot. 17:197-112.
F-4-30
-------
Effects - Marine
307
Photosynthesis
Inhibition
LLJ
z
UJ
o
01
cc
o
O
-4 -
-5 -
280
290
300 310 320
WAVELENGTH (nm)
340
FIGURE D.3 Weighting functions in current use for biologi-
cal UV effects.
latitude, season, and other factors. These tables became
the starting point for various investigations into biologi-
cal effects other than human skin response. It was believed
that, although human sunburn is not itself relevant to ef-
fects in other organisms, use of an erythemal weighting
function yielded a reasonable, relative DUV radiation mea-
surement for a typical biological effect. This use was
reinforced by the availability of the fairly inexpensive
F-4-31
-------
Effects - Marine
308
Robertson-Berger Sunburning Ultraviolet Meter, which was
reputed at that time to have a response spectrum (Figure
D.3, Curve B) sufficiently similar to the erythemal action
spectrum to serve as a practical dosimeter. However, as
is clear from the figure, the response spectrum of this
meter is really somewhat different from the erythemal
action spectrum.
Caldwell (1971) developed a generalized plant action
spectrum from a number of action spectrum studies with
plant material (Figure D.3, Curve C). Setlow (1974) com-
bined the results of action spectrum studies with bacteria
and viruses (where injury to DNA is known to be responsible
for the effects) to give a generalized DMA-damage action
spectrum (Figure D.3, Curve D). In unpublished studies
conducted for the EPA-BACER (Biological and Climatic Ef-
fects Research) Program, investigators at the USDA Agri-
cultural Research Center, Beltsville, Maryland, employed
a simplified weighting function, termed AI9 (Curve E), which
approximates the Setlow DNA-damage action spectrum. Inves-
tigators at the University of Florida (R. H. Biggs and
S. V. Kossuth, unpublished BACER report) employed an aver-
age weighting function, termed /5I21 (Curve F), that falls
between the DNA-damage action spectrum and the Caldwell
action spectrum in the short-wavelength portion of the
UV-B region. Figure D.3 also shows a weighting function
(Curve G) for inhibition of part of the photosynthetic
process occurring in isolated chloroplasts (Jones and Kok,
1966). This study is the only one known that has critically
examined the spectral influence of UV (and visible) radia-
tion on the photosynthetic reaction, independent of influ-
ences on cell growth and metabolism. Recent studies by
R. C. Smith, K. S. Baker, 0. Holm-Hansen, and R. Olson
(Scripps Institution of Oceanography, unpublished) have
shown that photosynthesis by phytoplankton exposed to solar
radiation through various filters is inhibited in a manner
consistent with the Jones and Kok action spectrum.
To understand the effect of the weighting-function
choice on the calculated DUV dose one may note in Figure
D.I that a given ozone-layer depletion increases the UV-B
irradiance by different amounts at different wavelengths.
The increase at a short wavelength (like 290 nm), where
J(X) is low may be over an order of magnitude more than
that at a longer wavelength (like 320 nm), where J(A) is
larger. Because of this fact the effect of decreased ozone
thickness on the DUV dose rate depends very much on the
relative values of E(X) at these longer and shorter wave-
lengths. For a given ozone change, the calculated change
F-4-32
-------
Effects - Marine
309
in DUV dose rate will be generally greater with weighting
functions that increase more toward the short wavelengths
than those that increase less.
Different weighting functions shown in Figure D.3 give
very different predictions of the effects of ozone deple-
tion. For example, about 50 percent of the inhibitory
effect of UV on the photosynthesis reaction results from
radiation greater than 390 rim, and 25 percent from radia-
tion between 340 and 390 nm (Curve G). Thus, 75 percent
of the photosynthetic inhibition by solar radiation results
from wavelengths that would not change with ozone-layer
reduction. Consequently, if this weighting function is
applicable to intact plants, a moderate ozone depletion
would not be likely to have a serious impact on photosyn-
thesis over that already being experienced. This contrasts
greatly with the predictions using an erythemal or a DNA-
damage weighting function (Curve A or D), which would pre-
dict a much larger effect.
The Robertson-Berger (R-B) meter reading also changes
less with a given change in ozone amount than does the DUV
dose rate calculated with an erythemal or DNA-damage weight-
ing function. One consequence of this fact is that an R-B
meter reading is less sensitive to the latitudinal change
of UV-B than is the DUV dose rate calculated with a DNA
damage action spectrum. Such a difference in the estimated
latitudinal gradient is important because some epidemiologi-
cal studies correlating human nonmelanoma skin-cancer inci-
dence with DUV have used R-B meter readings at different
cities to indicate the local UV environment. From such
correlations, projections of increased skin cancer for ~
particular ozone-layer depletions have been made. However,
if the DNA-damage weighting function is more appropriate
for skin cancer than the R-B meter response, and there is
preliminary evidence that this is so for skin-cancer induc-
tion in hairless mice (P. D. Forbes, F. Urbach, and R. E.
Davies, Temple University Skin and Cancer Hospital; R. D.
Rundle and D. S. Nachtwey, NASA Johnson Space Center, un-
published) , the latitudinal DUV dose gradient is actually
much steeper than the value measured. This would change
the estimated increase in skin-cancer incidence accompany-
ing a given ozone-layer depletion.
As shown in Figure D.4, a rather different latitude
dependence of solar-DUV dose rates is calculated with dif-
ferent action spectra, using the average values of ozone-
layer thickness appropriate for the season (the summer
solstice) and latitude. (To facilitate comparison, all
these curves are normalized to 30° N latitude, corresponding
F-4-33
-------
Effects - Marine
322
at the approximate latitude (30° N) of Gainesville,
Florida, where the experiments were performed.* Column 4
shows the percentage increase in this dose rate, with a
15 percent ozone reduction. Column 5 shows the corres-
ponding increase in weighted daily DUV dose rate in
J • m~2 • day~^ (the product of the values in Columns 3
and 4), and Column 6, the percentage reduction in yield
for a 15 percent ozone-layer reduction (the product of
Columns 5 and 2) .
It may be noted (Column 4) that the percentage increase
in DUV is lower with the Caldwell weighting function than
with the DNA-damage function, and still lower with the'
AI21 function. However, the yield reductions expected
with 15 percent ozone depletion (Column 6) are higher with
the AI21 and Caldwell weighting functions than with the
DNA-camage function. This seemingly paradoxical situation
arises because weighting functions like ^121 and Caldwell,
which give more emphasis to the longer UV-B wavelengths,
make the DUV dose rates for sunlamp radiation considerably
lower, relative to sunlight, than those calculated with a
DNA-camage weighting function. Consequently, in order to
simulate a certain increase in solar-DUV dose rate on the
basis of these weighting functions, more actual sunlamp
exposure is required than if the DNA-damage function were
employed. Hence the larger biological effect. Use of the
photosynthesis inhibition weighting function (Curve G of
Figure D.3, Appendix D), instead of the Caldwell curve,
would make even a larger difference. The question is, of
course, which of these Column 6 numbers is correct (i.e.,
what weighting function should be used)?
An analogous problem arises when a dosimeter, such as
the Robertson-Berger meter, is used to compare sunlight
with UV from sunlamps in an experimental setup. If the
DUV dose rate from sunlight (say, at noontime on the sum-
mer solstice, at 30° N latitude, with 0.29 atm • cm of
ozone overhead) were the same as that from a particular
set of filtered sunlamps, according to the R-B meter read-
ing, it would then be about sixfold less than the lamp
dose rate if recalculated on the basis of a DNA-damage
weighting function. That is, the sunlamp radiation would
actually contain six times the DNA-damaging UV of sunlight
*Note that, for the reasons explained at the beginning of
Appendix D, the raw DUV doses for different action spectra
in Column 3 cannot be directly compared (although the final
numbers in Column 6 can be).
F-4-34
-------
Effects - Marine
323
when adjusted to make the two give the same R-B meter read-
ing. This creates a large uncertainty in how experiments
should be set up and interpreted, which cannot be elimi-
nated until the correct weighting functions are known.
The instantaneous DUV dose rate may be an important
parameter in determining the biological effect, since high
rates of UV damage may overwhelm repair processes. Nacht-
wey (1975) has demonstrated a strong dependence on dose
rate of the effectiveness of filtered sunlamp radiation
for killing the alga Chlamydomonas reinhardi (Figure E.2).
These experiments employed a range of R-B meter readings
(expressed in "Sunburn Units," or SU) corresponding to sun-
light with present to moderately depleted ozone levels.
However, if the DNA-damage weighting function (rather than
the R-B meter response spectrum) is actually the appro-
priate one for killing of this organism, almost all of the
dose rates studied were greatly in excess of those given
by the sun, even with extreme ozone-layer reductions. Only
the points around 1 SU • h"1 are close to current noontime
2 3 4 5
DOSE RATE (SU • h"1)
FIGURE E.2 Lethal doses for 50 percent killing (LD50) and
90 percent killing (LDgg) of Chlamydomonas reinhardi ex-
posed to filtered fluorescent sunlamp radiation at differ-
ent dose rates. Doses and dose rates are given in terms
of R-B meter readings ("Sunburn Units," or SU) .
F-4-35
-------
Effects - Marine
324
irradiances on a DNA-damaging dose-rate basis, and at this
rate the time required for 90 percent kill amounts to a
full daylight period at noontime dose rates--i.e., it ex-
ceeds the daily dose naturally available. Since irradia-
tion protracted over a period of days might well give a
different dose response because of the better opportunity
afforded for repair, reports of biological effects result-
ing from a few SU delivered by fluorescent sunlamps must
be reevaluated with better information about the weighting
function. The conclusion that the organisms can be killed
by a few hours of summer sunlight under current irradiances
may or may not be valid.
An additional complication in experimental studies is
that, in most situations, the sunlamps are turned on for
from 4-6 h to obtain a given daily dose comparable with
that from the sun. This is obviously unnatural and sub-
jects the organisms to much higher intensities than the
sun delivers at some times of day (e.g., in midmorning
and midafternoon) but to lower intensities than the sun
delivers at noon. Only a few investigators have used the
more natural upward and downward stepping of irradiance,
which is a better practice.
Another problem is that the UV-A and visible components,
even with supplemental white fluorescent lamps, are gener-
ally more than an order of magnitude less intense than
that in sunlight. This unnatural situation may eliminate
some significant synergisms, positive or negative, between
radiations of different wavelength that are present in
sunlight.
REFERENCE
Nachtwey, D. S. 1975. Dose rate effects in the UV-B in-
activation of Chlamydomonas and implications for survi-
val in nature, in CIAP Monograph 5, Impacts of Climatic
Change on the Biosphere, Part 1, Ultraviolet Radiation
Effects. D. S. Nachtwey, M. M. Caldwell, and R. H. Biggs,
eds. pp. 3-105 to 3-119. DOT-TST-75-55, U.S. Dept. of
Transportation, Washington, D.C.
F-4-36
------- |