AN  ASSESSMENT  OF  THE  RISKS OF
          STRATOSPHERIC MODIFICATION
             Volume Hi:   Chapters 7-18
                   Submission to the

                Science  Advisory  Board
        U.S.  Environmental  Protection  Agency
                           By

             Office of Air  and  Radiation
        U.S.  Environmental  Protection  Agency
                     October 1986
          Comments  should be  addressed  to:

                    John  S. Hoffman
   U.S.  Environmental Protection Agency,  PM  221
                  401  M Street, S.W.
                Washington, D.C.   20460
                          USA
The  following report is being submitted to the Science Advisory Board an^
to the Public for review and comment.  Until the Science Advisory  Board
review has been completed and the document is revised, this assessment
does not represent JEPA's official position on the risks associated with
Stratospheric-Modification.  This report has been.'written as part'"of-the.
activities of the EPA's congressionally-established Science Advisory Board,
a public group providing extramural advice on scientific issues.   The' Board
is structured to provide a balanced independent expert assessment  of scientific
issues it reviews, and hence, the contents, of this report do not necessarily
represent the views and policies of the EPA nor of other agencies  in the
Executive Branch of the Federal Government.  Until the final report is
available, EPA requests that none of the information contained in  this
draft be cited or quoted. Written comments should be sent'to:  John S. Hoffman
at the EPA by November 14, 1986.

-------
Chapter 7

-------
                                CHAPTER 7

                        NONMELANOMA  SKIN TUMORS
SUMMARY

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

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

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

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

        Ib.   Higher incidence rates  occur among occupational groups with
             outdoor exposures compared to those with indoor exposures.

        Ic.   A latitudinal and UV radiation gradient exists showing the
             highest incidence rates in geographic areas of relatively high UV
             radiation exposure.

        Id.   Skin pigmentation provides a protective effect.

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

        If.   There is a predisposition for nonmelanoma skin tumors to develop
             among light-skinned individuals who are susceptible to sunburn
             and who have red/blond hair, blue/green eyes, and a Celtic
             .heritage.

    2.   AVAILABLE EPIDEMIOLOGICAL EVIDENCE SUGGESTS THAT SCC AND BCC RESPOND
        DIFFERENTLY TO DIFFERENT LEVELS OF SOLAR EXPOSURE.  IT HAS BEEN
        HYPOTHESIZED THAT CUMULATIVE UV RADIATION HAS A GREATER EFFECT ON THE
        DEVELOPMENT OF SCC THAN ON BCC.

        2a.   The BCC/SCC incidence ratio decreases with decreasing latitude.

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

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

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

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

        3a.   UV radiation produces malignant nonmelanoma skin tumors in
             animals.  UVB wavelengths have been shown to be most effective in
             producing these tumors.
                            * *  DRAFT FINAL  * * *

-------
                               7-3
    3b.  UVB has been shown to cause a variety of DNA lesions, to induce
         neoplastic transformation in cells, and to be a mutagen in both
         animal and bacterial cells.

4.  SEVERAL RESEARCHERS HAVE INVESTIGATED THE CHANGES IN THE INCIDENCE OF
    NONMELANOMA SKIN TUMORS THAT MAY RESULT FROM INCREASES IN EXPOSURE TO
    SOLAR UV RADIATION.

    4a.  Several studies have provided estimates of a biological
         amplification factor, BAF (also referred to as an amplication
         factor, AF), which is defined as the percent change in tumor
         incidence that results from a one percent change in UVB
         radiation.  The results from six studies produced an overall AF
         range for all nonmelanoma skin tumors from 1.8-2.85, although AF
         estimates for one sex may be even larger (e.g., 1.4-3.2 for males
         from Fears and Scotto (1983)).

    4b.  AF estimates are generally higher for males than for females and
         increase with decreasing latitude.  In addition, the AF estimates
         for SCC (e.g., 1.6-7.1) are higher than the AF estimates for BCC
         (e.g., 1.1-2.6).  This finding is consistent with observations
         that the BCC/SCC ratio decreases with decreasing latitude and
         that BCC is more likely to develop on unexposed sites.

    4c.  There is, however, no general concensus on the best quantitative
         measure to use in estimating the effect of ozone depletion on UVB
         dosage.  Although AF estimates appear to be relatively
         insensitive to the method used to measure UVB dosage, the
         estimated effect of ozone depletion on nonmelanoma skin cancer
         incidence is sensitive to the choice of an action spectrum.  For
         example, using the DNA action spectrum, Scott (1981) estimated
         that a 5 percent ozone reduction would lead to an increase in BCC
         in Minneapolis-St. Paul of 7.7 percent for males and 5.8 percent
         for females.  Using an action spectrum based on the RB meter,
         with the same functional model and cancer incidence data, Scotto
         et al. (1981) estimated corresponding increases to be 5.6 percent
         for males and 4.4 percent for females.  The reason for this
         difference is that the optical amplification factor  (percent
         change in weighting UVB wavelengths) is very sensitive to ozone
         depletion.
                      * * *  DRAFT FINAL  * * *

-------
                                   7-4
INTRODUCTION

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

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

    To evaluate the role of solar radiation, and UVB in particular, in the
development of nonmelanoma skin tumors, this chapter reviews the biology of
the skin and nonmelanoma skin tumors focusing on their responses to solar
radiation.   Evidence from molecular and animal studies relating nonmelanoma
skin tumors to ultraviolet radiation (UVR) is discussed.  Epidemiological
evidence linking exposure to solar radiation and development of nonmelanoma
skin tumors among human populations is also reviewed.  Finally, the available
dose-response relationships that estimate potential changes in the incidence
of nonmelanoma skin tumors due to changes in exposures to  UVR are presented.

BIOLOGY OF  NONMELANOMA  SKIN  TUMORS:  LINKS TO UVB

Biology and Photobiology of the Skin

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

    Skin Biology

    There are three principal layers of skin: the epidermis, the dermis,  and
the panniculus adiposus (Exhibit 7-1).  These layers vary  in thickness
depending on their location.  The epidermis normally consists of about 6-10
cell layers and is typically 60 urn-100 urn thick.  The thinner epidermis is
found on the head, trunk, and upper limbs and the thicker  epidermis is found
on the lower limbs.  The epidermis on the palms and soles  is as much as ten
times thicker than that on the head and trunk.  The stratum corneum is the
outermost,  more dense layer of the epidermis and generally comprises between 8
um-15 urn of the epidermis1 thickness (Pearl 1984; Fitzpatrick and Soter
1985).  The dermis is typically between 1700 urn and 2000 urn thick and the
subcutaneous layer, the panniculus adiposus, is typically 4000 um-9000 urn
thick (Fitzpatrick and Soter 1985).
                          * * *  DRAFT FINAL  * *

-------
                                     7-5



                                 EXHIBIT 7-1

                        Organization of the  Adult Skin
          •      y   '
       S. . f- .rL-^C-^
-------
                                   7-6
    The epidermis is principally a mixture of three cell types of different
embryonic origin and function; these are (in order of percent composition) the
keratinocyte (80 percent), the melanocyte (5-10 percent) and the Langerhans
cell (5-10 percent).  The dermis is composed mainly of connective tissue
fibers.  These fibers, which are secreted by cells called fibroblasts, are
responsible for the skin's resilience and elasticity.  Many skin changes
associated with aging are caused by the impact of solar radiation on these
dermal fibers.   The subcutaneous layer is a specialized layer of connective
tissue that functions as a cushion between the bone and the epidermis and
dermis.  It consists primarily of fat cells.

    Melanin is a pigment that absorbs light in the broad range of 250 nm-1200
nm; however, its absorption increases steadily towards the shorter wavelengths
(Anderson 1983).  It is synthesized in melanocytes from tyrosine and deposited
in a protein matrix in organelles termed melanosomes (Romsdahl and Cox 1976).
Melanosomes, in turn, are eventually destributed to keratinocytes which are
vertically arranged in the skin.  Melanin is a protein containing aromatic
amino acids, urocanic acid, carotenoids (in the stratum corneum only), and
nucleic acids.   In humans and other mammals there are two predominant forms of
melanin:  eumelanin, which is brown or black, and phaeomelanin, which is
yellow or auburn (the pigment responsible for red hair).  The melanosomes
containing these two types of melanin differ structurally in that those with
phaeomelanin are round and have a protein matrix with a "tangled" appearance
whereas eumelanin-containing melanosomes are round and have a characteristic
lamellar structure  (Norlund 1981).

    It is the degree of melanosome aggregation and their number which
determine skin color.  Melanosomes in keratinocytes from light-skinned
individuals are characteristically found as aggregates, whereas melanosomes in
keratinocytes from dark-skinned individuals, because they are larger,
generally occur singly.  Extensive studies have indicated nonsignificant
differences in the number of melanocytes in various racial groups, although
there are variations by anatomic region.  The highest concentration of
melanocytes is in the cheek (2,310 + 150 melanocytes/mm2) and the lowest is
in the back and thigh (890 + 70/mm2 and 1,000 + 70/mm2, respectively)
(Briele and das Gupta 1979).

    Photobiology

    The interaction of sunlight with the skin is a complex process.  It
involves the transfer of energy from sunlight to various molecules in the skin
layers and the subsequent cellular responses to this energy transfer.  About
95 percent of incident radiation penetrates the skin while the remaining 5
percent is reflected by the stratum corneum.  Two processes, scatter and
absorption, determine the penetration of radiation into the skin.
Measurements of the transmission of UVR through isolated epidermis tissue from
medium complexion Caucasian skin  (Bruls et al. 1984; Kaidbey et al. 1979)
indicate that between 1 percent-20 percent of 295 nm UVB radiation would reach
the basal layer.  For 340 nm UVA radiation, as much as 20 percent-60 percent
may penetrate to the basal layer  (Kubitschek et al. 1986).  Above 400 nm  (in
the visible range)  transmission approaches 90 percent (Anderson 1983).  The
                          * * *  DRAFT FINAL  *

-------
                                   7-7
amount of radiation reaching the basal layer is a function of the thickness of
the epidermis and stratum corneum as well as their content of
radiation-absorbing molecules called chromophores.   There are a large number
of chromophore types in the skin only one of which is melanin.  Most of the
optical absorbance within the skin is, however, attributable to melanin as
graphically shown in Exhibit 7-2.

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

        Solar Radiation Dose Reduction

    Dose-reduction via tanning may be either an immediate or a delayed
response.  The immediate response, sometimes referred to as immediate pigment
darkening (IPD) (Pathak 1985), is due not to an increase in epidermal
melanocytes or melanin but rather to changes in the melanin already present in
the skin.  Melanin in both melanocytes and keratinocytes evidently is
oxidized..  In addition, the specific distribution of melanosomes within the
melanocyte is altered (Gange and Parrish 1983).  The most effective
wavelengths for producing IPD range from 320 nm-400 nm.  The immediate
response may begin within one minute following exposure, is maximal at about
one hour, and is gone in about four hours.

    The delayed response involves the de novo synthesis of new melanosomes
that are subsequently transported to keratinocytes where they produce a tan
within about 10 hours after exposure.  This response may continue for several
days with the maximal tan being achieved in about one week.  With time the tan
wears off as the keratinocytes containing the extra pigment are sloughed off.
In addition to the increased production of melanosomes, the number of active
melanocytes in the skin also increases probably as a result of increased
mitosis and recruitment of dormant melanocytes.  The delayed tanning response
is induced predominantly by wavelengths in the UVB range, although UVC, UVA
and visible light can also induce the response.

    There is a wide variation in the actual tanning response of individuals to
sunlight exposure.  Skin tanning responses have been generally classified into
six skin types categories as outlined in Exhibit 7-3 (Pathak 1985).

    The remaining dose-reduction response involves the increased production of
keratinocytes and a subsequent thickening of the epidermis (keratinocyte
hyperplasia).  Studies in humans have shown that after a single exposure to
UVB, a sustained increase in epidermal mitoses occurs leading to increases in
the thickness of the epidermis and stratum corneum of between 1.5- and 3-fold
over the course of one to three weeks.  The UV dose first induces a transient
depression in macromolecular synthesis in which DNA, RNA and protein synthesis
are markedly reduced and then elevated.  The response is maximal within 24 to
48 hours after irradiation but may continue for as long as one week.  UVB and
UVC are the most effective wavelengths for inducing this response.
Keratinocyte hyperplasia protects individuals with a poor tanning ability
                            * *  DRAFT FINAL  * * *

-------
                                    7-8
                                EXHIBIT 7-2
       Ultraviolet Absorption Spectra of Major Epidermal Chromophores
2 00
                       220
                               740     260     280
                                   WAVELENGTH, NM
                                                   300     320
Experimental Conditions:
        DOPA-melanin,  1.5 mg%  in  HO;
                         -4
        urocanic acid,  10   M  in  HO;
        calf thymus DNA, 10 mg% in  H?0 (pH 4.5);
        tryptophan  (TRP), 2 x  lo"4  M  (pH 7);
        tyrosine (TYR),  2 x 10"10 M (pH 7).
Source:  Anderson (1983).
                            * *   DRAFT FINAL  * *

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

            Skin Types and Skin Tanning Responses
    Skin Type                 Skin Tanning Response
        I           Very sensitive.   Tans  little or  not  at
                    all.  Usually has red/blond hair,  blue
                    eyes,  freckled skin.   May develop  severe
                    sunburn after one hour of sun exposure  in
                                                    o
                    summer.   MED:  15 mJ/cm2-30 mJ/cm .   No
                    IPD response.

       II           Sunburns frequently, tans poorly.  MED: 20
                    mJ/cm2-35 mJ/cm2.  Weak IPD response.

      Ill           Burns  first and tans  later.   Has IPD
                    response.

       IV           Lightly tanned skin. Dark eye and  hair
                    color.  In summer, skin color changes  from
                    light  brown to olive or medium brown.
                    Strong IPD response.

        V           Usually dark brown or  black eye  and  hair
                    color.  Acquires deep  tan.  Intense  IPD
                    response.

       VI           Darkly pigmented, e.g., African  and
                    American Blacks.
IPD = Immediate Pigment Darkening

MED = Minimal Erythemal Dose
                   * * *  DRAFT FINAL  *

-------
                                   7-10
because the disulfide rich keratin synthesized by the keratinocytes absorbs
photons in the UVC and UVB ranges.  Among lightly pigmented individuals, this
skin thickening is probably the most important dose-reducing response whereas
among dark-skinned individuals, tanning is the more important response (Gange
and Parrish 1983).

        DNA Repair Mechanisms

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

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

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

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

    Another efficient DNA repair mechanism is called excision repair.
Excision repair works more slowly than photoreactivation but may repair not
only dimers but also other kinds of damage, e.g., methylations.  In the case
of dimers, excision repair involves the activity of three enzymes, an
endonuclease specific for dimers, a polymerase, and a ligase.  The
                          * * *  DRAFT FINAL  * * *

-------
                                   7-11
endonuclease attacks the DNA at or adjacent to the dimer and introduces a nick
in the DNA.  The DNA polyraerase removes the damaged DNA segments while
utilizing the opposite strand as a template for new synthesis, and finally a
ligase joins the newly synthesized DNA to the preexisting strand.  Other kinds
of DNA damage such as methylations require different first enzymes but the
activities of the polymerase and the ligase remain the same (Spikes 1983).

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

Biology of Nonmelanoma Skin Tumors

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

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

    Squamous cell carcinomas are neoplasms of the epidermis that differentiate
towards keratin formation (Scotto and Fraumeni 1982).  SCCs are far less
fastidious in their growth requirements in vitro and will grow under a
variety of conditions.  They are generally more aggressive than BCCs and
account for about three-fourths to four-fifths of the deaths attributable to
nonmelanoma skin cancer (Dunn et al. 1965).

Cellular  Studies

    Solar radiation and its component UV and visible wavelengths have been
observed to cause cellular changes, such as the induction of cell division
and/or differentiation, the loss of specialized functions, mutation,
                          * * *  DRAFT FINAL  * *

-------
                                   7-12
transformation, and death.  Evidence from cellular and molecular studies
clearly indicate that UVR can cause transformation in mammalian cells in the
absence of any confounding immunological, hormonal, or physiological effects
encountered in vivo.  Shorter UVB wavelengths are generally more effective
in transforming mammalian cells.  In addition, the experiments conducted in
vitro may be used to extrapolate effects in vivo.

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

    Exposure to UVR has also been found to result in the impairment of antigen
presenting cell function in both the mouse and man (Greene et al. 1979).  In
animals with primary UVR induced tumors, this impairment results in the
generation'of UV-tumor-antigen specific suppressor lymphocytes (Kripke et al.
1977; Daynes et al. 1977).

    Ultraviolet light is mutagenic in both bacterial and mammalian cells.  The
importance of mutation to an assessment, of the role of UVB in nonmelanoma skin
cancer development lies in a theory of carcinogenesis which suggests that
somatic mutation in mammalian cells is the first step down a pathway which
includes malignant transformation and ends in neoplasia and metastasis (Trosko
and Chu 1975).  In UVR-treated bacteria, researchers have concluded that
pyrimidine dimers are responsible for much of the observed mutagenesis (Hall
and Mount 1981).  UVR treated cells, if subsequently treated with
photoreactivating light, show a substantially decreased frequency of mutants
per survivor compared to UVR-treated cells without subsequent photoreactivation
treatment.  Peak et al. (1984) derived action spectra for DNA dimer induction,
lethality, and mutagenesis in E. coli over wavelengths from 254 nm-405 nm
(Exhibit 7-4) and found that all three endpoints decreased in efficiency in a
similar fashion as the wavelengths of radiation increased.  From 300 nm-320
nm, all characteristics showed differences of about 2.5 orders of magnitude.
Furthermore between about 250 nm and 320 nm, the values for the three
endpoints either coincided with or closely paralleled Setlow's (1974) proposed
average DNA action spectrum.

    Experiments have also been performed using fibroblasts from xeroderma
pigmentosum (XP) patients.  These patients have a genetic inability to repair
UV-induced pyrimidine dimers.  Setlow et al.  (1969) compared the responses of
XP fibroblasts to normal human fibroblasts upon irradiation to UVR.  They
observed that normal human fibroblasts which were UVR irradiated, then
stimulated to undergo cell division by immediate replating, showed much higher
numbers of mutations than similarly irradiated normal cells which were
incubated for seven days as nondividing monolayer cultures.  These results
indicated that normal human cells can repair the damage induced by UVR
(presumably by excision repair).  XP cells, by comparison, produced a
                          * * *  DRAFT FINAL  * * *

-------
                                         7-13
                                    EXHIBIT 7-4

                Ultraviolet Action Spectra  for DNA  Dimer Induction
                                     n, • Relative dimer yield per quantum
                                      —A Relative lethality per quantum
                                          Relative mutagenicity per quantum
                                     	Average DNA spectrum
                                          (Setlow, 1974)
                                          Xenon lamp
                                     A, o Hg lines
                                                (Tyrrell, 1973)
                    250
Source:   Peak et  al.    (1984).
    350       400
WAVELENGTH  (nm)
                              * -> *
                                  *   DRAFT  FINAL   »  * *

-------
                                   7-14
comparable number of mutants whether replated immediately or given seven days
of incubation.  Since XP cells are unable to excise pyrimidine ditners, these
experiments suggest that unexcised dimers present at the time of cell division
may have been responsible for the production of mutations.

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

    Doniger et al. (1981) developed action spectra for transformation,
lethality, and thymine dimer formation using a monochromatic light source.   In
dose-response studies comparing pyrimidine dimer formation and transformation
of Syrian hamster embryo cells, the slopes of the dose-response curves were
not always parallel.  The discordance was greatest at 290 nm.  The lowest
exposure required for equivalent cell transformation, lethality, and
pyrimidine dimer formation was at 270 nm.  Comparing results at 290 nm,
297 nm, and 302 nm, the respective doses required were 2.3 J/m2, 8.7 J/m2,
and 25 J/m2 for dimer formation and 7.4 J/m2, 47 J/m2, and 97 J/m2 for
transformation.  A similar spectrum was found for induction of anchorage
independent growth in human fibroblasts (Sutherland et al. 1981).  These
authors also found a maximum effectiveness at 265 nm and that transformation
at 290 nm was six times more effective per photon than that at 297 nm.

Molecular Studies:  DNA and  UV Radiation

    Chemical changes and biological damage induced by ultraviolet light
require the absorption of light energy  (photons) by molecules within the
target.  Each type of molecule is capable of absorbing radiation only in
specific wavelength ranges.  Examination of the absorption spectra of
molecules in biological systems indicates that a number of biomolecules can
absorb radiation in the 220 nm to 400 nm region and thus may be critical
targets for detrimental UV effects (Spikes 1979).  This section will review
information only on the most important molecule, DNA, because of its putative
role in mutagenesis, transformation and carcinogenesis.

    As already mentioned, two possible types of mechanisms can induce DNA
damage --a direct mechanism resulting from absorption of energy by DNA and an
indirect mechanism involving reactive oxygen species.  A number of different
lesions may be induced in DNA by UV irradiation.  These include  (1) pyrimidine
dimers, (2) pyrimidine adducts, (3) single strand breaks, (4) double strand
breaks and (5) protein-DNA crosslinks.  Different wavelengths have different
efficiencies .for the production of these lesions.
                              *  DRAFT FINAL  * * *

-------
                                   7-15
    Pyrimidine Dimers and Adducts

    As described in an earlier section, pyrimidine dimers may be formed by
direct absorption of photons in the UVB wavelength range.  Adjacent pyrimidine
molecules on the same strand of DNA become linked together by a cyclobutane
ring between the 5 and 6 carbon atoms of each residue.  Thymine dimers are the
most likely pyrimidine dimers because the excited state of thymine is a much
lower energy state than that of cytosine .   Cytosine co-dimers are also
possible, as are heterodimers between thymine and cytosine or uracil (Spikes
1982).

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

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

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

    Brash and Haseltine (1982) have shown that there is a linear relationship
between base damage incidence and mutation incidence.  For shorter wavelengths
(UVB) it seems clear that pyrimidine dimers and 64 photoadducts are involved
in mutagenesis, but at 365 nm the correlation between dimers or adducts and
mutagenesis is not known (Peak et al. 1984).

    DNA Single-Strand Breaks

    DNA single-strand breaks can be induced directly by UVB radiation or
indirectly by UVA and visible light.  Analysis of the relative efficiencies
for the induction of single-strand breaks reveals an action spectrum that
corresponds with nucleic acid absorption at or below 313 nm (Peak and Peak
1986a).   Above 313 nm to wavelengths as long as 546 nm, single-strand breaks
have been detected in human fibroblasts with an action spectrum in the visible
range resembling that of riboflavin (Tyrell 1984).  Relative to thymine
                          * *. *  DRAFT FINAL  * *

-------
                                   7-16
dimers, single-strand breaks are induced only to a small extent by 254 nm
radiation, but as the wavelength is increased the proportion of single-strand
breaks to thymine dimers increases.  At 313 nm in E. coli, one single-strand
break is induced for every nine thymine dimers (Cerrutti and Netrawalli
1979).  Some single-strand breaks induced at 313 nm may be due to indirect
effects from photosensitizers and oxygen dependent mechanisms (Miguel and
Tyrell 1983), but the spectral analysis indicates that the majority of
single-strand breaks induced at this wavelength are due to direct effects
(Peak and Peak 1986a).

    DNA Double-Strand Breaks

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

    DNA-Protein Crosslinks

    DNA-protein crosslinks can be induced by borderline UVB radiation (290 nm)
via a direct photon-absorbing mechanism in human cellular DNA.  Below 320 nm,
there are approximately 40 DNA-protein crosslinks per lethal event.  As cells
can survive 2xl03 DNA-protein crosslinks induced at longer wavelengths (405
nm),  it appears that such a small number of DNA-protein crosslinks is not
important in UV-induced cell lethality.  This assumes, however, that there are
no interactions between dimers and DNA-protein crosslinks (Peak et al. 1985b).

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

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

-------
                                   7-17
Animal Studies

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

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

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

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

    There has been a great deal of experimentation to determine if there is
dose-rate reciprocity in UV carcinogenesis.  Blum et al.  (1942) studied the
effect of the intensity of the dose on the tumor latency period and reported
that for a 10-fold dose range of between 4.3 J/m2/sec and 0.42 J/m2/sec,
no significant differences were found.  In a second study using a wider range
                          * * *  DRAFT FINAL  * * *

-------
                                   7-18
of intensities and greater numbers of animals he reported that above 0.4
J/m2/sec the effectiveness of UVR in inducing tumors was almost independent
of dose-rate, but below 0.4 J/m2/sec the effectiveness fell off rapidly with
intensity.

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

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

Immunosuppression

    The risk of developing nonmelanoma skin tumors may be influenced by the
inmune status of a UVR-exposed animal or human.  UVR itself is
immunosuppressive (Kripke et al. 1977; Spellman et al. 1977),  inducing the
production of a class of lymphocytes, T suppressor cells, which depress the
host's ability to respond to UVR-induced tumor antigens (Fisher and Kripke
1977, 1978; Spellman and Daynes 1977, 1978; Daynes et al. 1979).

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

    Among systemically immunosuppressed individuals, such as kidney transplant
patients, the incidence and aggressiveness of nonmelanoma skin cancer may be
increased.  The available data on the increased risk among immunosuppressed
patients are generally more convincing for SCC than for BCC (Pollack et al.
1982).

EPIDEMIOLOGICAL EVIDENCE

Occurrence and Trends

    Nonmelanoma skin tumors are among the most common malignant neoplasms
occurring in white populations.  Based on a one-year survey (1977-1978)
conducted by the National Cancer Institute (NCI), it was determined that
nonmelanoma skin cancers developed in approximately 400,000 white Americans
each year (Scotto and Fraumeni  1982).  The annual age-adjusted incidence rate
                          * * *  DRAFT FINAL  * *

-------
                                   7-19
for this survey period was estimated to be 232.6/10s among whites.  For
comparison, the estimated incidence rate for all other cancers among whites in
the United States based on 1973-1976 data from the Surveillance, Epidemiology
and End Results (SEER) Program was 318.9/105 (NCI 1985).  Among blacks, the
annual age-adjusted incidence rates for nonmelanoma skin tumors and all other
cancers (NCI 1985) were 3.4/105 and 347.3/105, respectively.

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

    These relative differences in BCC and SCC (and CMM) persist when males and
females are examined separately even though the incidence in males generally
exceeds that in females.  Exhibit 7-5 presents age-adjusted BCC and SCC
incidence data for white American males and females (Scotto and Fraumeni
1982).  The incidence of BCC was approximately four to six times greater than
the incidence of SCC.

    The overall case fatality rate for nonmelanoma skin tumors is
approximately 1 percent (Epstein et al. 1984).  SCCs are, however, generally
more aggresive than BCCs, accounting for about three-fourths to four-fifths of
the deaths attributable to nonmelanoma skin cancer (Dunn et al. 1985; NAS
1982).  SCCs may metastasize shortly after their appearance or even at the
time of diagnosis (Epstein 1983).  The rate of metastasis among SCC patients
has been estimated by several researchers to range between 2 percent and 20
percent.  Epstein (1983) reported that 25 percent of SCC patients with
metastisis were alive five years after the diagnosis, 13 percent were alive 10
years after diagnosis, and only 8 percent were alive.after 15 years.
Approximatley two-thirds of metastasizing SCCs were observed to occur on the
face and dorsa of the hands.  BCCs, in contrast, rarely metastasize even when
present for many years (Pollack et al. 1982)  Their aggressiveness may be
correlated with the site of tumor origin.  According to Pollack et al. (1982),
only about 100 cases of metastatic BCC have been reported.

    Time Trends

    Several researchers have observed that the incidence of BCC and SCC have
been increasing over the past several decades (Lee 1982, NAS 1982).  Epstein
et al. (1984) has pointed out that the rate of increase of SCC is greater than
that of BCC.  Scotto and Fraumeni  (1982), however, noted when comparing the
                          * * *  DRAFT FINAL  * * *

-------
                            7-20
                       EXHIBIT 7-5

     Comparison of Age-Adjusted Incidence Rates  Per  100,000
for Squamous Cell Carcinoma (SCC)  and Basal  Cell Carcinoma (BCC)
       Among White Males and Females in the  United States
                                    Age-Adjusted
                                   Incidence Rate/105
           Skin Tumor Type          Males      Females
SCC
BCC
65.4
246.6
23.6
150.1
           Source:  Scotto and Fraumeni  (1982).
                   * * *  DRAFT FINAL  * * *

-------
                                   7-21
NCI 1971-1972 and 1977-1978 survey data that the observed incidence increases
applied mainly to BCC.  The authors noted that the incidence of BCC among
United States whites increased by approximately 15 percent-20 percent over the
six-year period between surveys.

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

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

    Anatomical Distribution

    The predominant anatomical sites for both SCC and BCC and for both males
and females are generally the face, head, and neck.  Exhibit 7-6 presents the
distribution of BCC and SCC by sex for tumors occurring in whites in the
United States (Scotto and Fraumeni 1982).  The data were collected as part of
the 1977-1978 NCI survey.  The face, head, and neck accounted for 60 percent
or more of the total nonmelanoma skin tumors.  Most of the remaining BCCs
occurred on the trunk whereas most of the remaining SCCs occurred on the upper
extremities.  Skin tumor data from 1974-1978 from Switzerland in Exhibit 7-7
shows similar tumor site distributions  (Lee 1982).  Scotto and Fraumeni (1982)
noted that the tendency for nonmelanoma skin tumors to develop in exposed
areas was consistent with the belief that solar radiation is a dominant risk
factor for nonmelanoma.

    There are, however, some general differences in the distributions of SCC
and BCC which suggest that these nonmelanoma skin tumors may respond
differently to different dosages of sun exposure.  The available
epidemiological data generally show that BCC is more likely to develop on
regularly unexposed sites compared with SCC.  Pollack et al. (1982) noted that
although it is well-established that sun-exposed areas are more prone to BCC,
in at least one study Urbach et al. (1972), approximately one-third of the
BCCs occurred in light-protected regions.  Laerum and Iversen (1981) observed
that among a group of SCC and BCC cases, 90 percent of the BCCs and
approximately 50 percent of the SCCs occurred on the head and neck.  They
                          * * *  DRAFT FINAL  * * *

-------
                      7-22
                  EXHIBIT 7-6

      Percentage of Tumors by Anatomic Site for
    Nonmelanoma Skin Cancer Among White Males and
           Females in the United  States
           (1977-1978 NCI Survey Data)
BCC
Anatomic Site
Face, Head and Neck
Trunk
Upper Extremities
Lower Extremities
Other Sites
Male
81.2
12.0
4.9
1.3
0.5
Female
84.1
8.9
3.4
2.9
0.7
sec
Male
74.8
4.5
18.1
1.3
1.4
Female
60.1
5.3
25.8
5.7
3.2
Source:   Scotto and Fraumeni  (1982).
             * * *  DRAFT FINAL * * *

-------
                       7-23
                   EXHIBIT 7-7

 Distribution by Sex and  Anatomic Site of Nonmelanoma
       Skin  Tumors:  Canton of Vaud, Switzerland
                    (1974-1978)
BCC
Anatomic Site
Head and Neck
Trunk
Upper Limbs
Lower Limbs
Other
Total
Male
540
(71.2)
130
(17.1)
14 '
(1.8)
15
(2.1)
58
(7.6)
758
(100.0)
Female
505
(78.7)
76
(11.9)
17
(2.6)
18
(2.8)
26
(4.0)
642
(100.0)
sec
Male
202
(80.5)
8
(3.2)
26
(10.3)
5
(2.0)
10
(4.0)
251
(100.0)
Female
103
(72.5)
11
(7.7)
19
(13.4)
7
(4.9)
2
(1.4)
142
(100.0)
Source:   Levi  and Chapallaz (1981)  as  cited  in Lee (1982)
                * *  DRAFT FINAL  * * *

-------
                                   7-24
estimated that of these tumors, two-thirds of the BCCs occurred on sites of
the head and neck receiving the highest UV radiation doses (e.g., the nose)
whereas all the SCCs occurred at these sites.  Hilstrom and Swanbeck (1970)
examined only SCC cases and observed that about 80 percent of the SCCs
occurred on the head.  Facial SCCs occurred relatively equally between males
and females, but of the SCCs on the external ear, 90 percent occurred among
males and only 10 percent among females.  Emmett (1982) examined only BCC
cases and observed that 75.5 percent occurred on the head and neck, 16 percent
on the limbs, and 8.4 percent on the trunk.  Lee (1982) noted that although
BCC and SCC tend to be concentrated on exposed sites, the distribution of BCC
did not precisely correspond with sun-exposed areas.

Exposure Factors

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

    Prolonged sun exposure is considered to be the dominant risk factor for
nonmelanoma skin tumors among light-skinned populations (NAS 1982; Scotto and
Fraumeni 1982; Greene and O'Rourke 1985; Lee 1982; Beral and Robinson 1981).
In its 1982 report, the National Academy of Sciences (NAS) wrote that experts
agree that exposure to sunlight causes 90 percent or more of the BCCs and SCCs
in the United States.  Several observations supporting this hypothesis were
cited in the NAS report (NAS 1982) and by other researchers • (Scotto and
Fraumeni 1982; Urbach 1983; Beral and Robinson 1981; Scotto et al. 1981;
Laerum and Iversen 1982; Emmett 1982).  These observations include:

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

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

        •   the latitudinal and UV radiation gradient showing the
            highest incidence rates in geographic areas of
            relatively high UV radiation exposure;

        •   the increase in incidence rates with increasing age;

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

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

        •   the predisposition for nonmelanoma skin tumors to
            develop among light-skinned individuals who are
            susceptible to sunburn and who have red/blond hair,
            blue/light eyes, and a Celtic heritage;
                          * * *  DRAFT FINAL  * * *

-------
                                   7-25
        •   the increased incidence of SCC in individuals treated
            with high intensity UVB sources and oral methoxsalen or
            topical coal tar; and

        •   the capacity of UV radiation to induce nonmelanoma
            skin tumors in experimental animals.

    The observed variations in nonmelanoma incidence by latitude in particular
support an association with sun exposure.  For example, Scotto and Fraumeni
(1982) observed, based on the 1977-1978 NCI survey data, that the incidence of
SCC and BCC in the United States showed a latitudinal gradient with higher
rates in the south.  These results are displayed in Exhibits 7-8 and 7-9 for
white males and white females, respectively.  (Also shown in these figures are
1973-1976 CMM data from the SEER program.  The latitudinal gradients for CMM
were least pronounced.)  It should be noted, however, that only a small number
of data points (as shown in Exhibits 7-8 and 7-9) were used to examine the
latitudinal gradients of BCC and SCC.

    Variations in the incidence ratio of BCC to SCC by latitude again suggest
that the two forms of nonmelanoma skin tumors respond differently to solar
exposure.  The results of MacDonald and Bubendorf (1964, as cited in Vitaliano
and Urbach 1980) showed that the BCC/SCC ratio decreased from approximately
10:1 in northern United States cities to approximately 2:1 or 3:1 in southern
rural areas.  Similarly, Scotto and Fraumeni (1982) noted that the ratio of
SCC incidence to BCC incidence increased with decreasing latitude and
increasing sunlight exposure (see Exhibits 7-8 and 7-9).  Pollack et al.
(1982) commented that dosimetry studies revealed a poor correlation between
BCC density in a site and UV radiation dose.  They suggested that etiological
factors for BCC other than UV radiation, such as the presence of areas of
scarring or epidermal nevi, may exist.  Among blacks, in whom nonmelanoma skin
tumors occur rarely compared to whites, areas of trauma or scarring may be
important sites for the development of SCC (Scotto and Fraumeni 1982).

    Vitaliano and Urbach (1980) examined several risk factors for SCC and BCC
in a case-control study.  The study included 366 BCC and 58 SCC cases seen at
the Tumor Clinic of the Skin and Cancer Hospital in Philadelphia (dates were
not specified).  A group of 294 white controls without carcinoma were selected
from the skin and cancer outpatient department.  Information compiled on each
case and control included cumulative solar exposure based on vocational and
military history, time spent sunbathing, and participation in outdoor sports
(as a spectator or participant).  Exposure was divided into four categories:
£1=0-9,999 hours, £2=10,000-19,999 hours, £3=20,000-29,999 hours, and
£4=30,000 or more hours.  The host factors that were examined included
complexion (pale or mild-dark), age (0-59 years or 60 and over), and ability
to tan (tans or burn-sensitive).

    Based on a statistical logistic regression of the case-control data,
Vitaliano and Urbach (1980) identified the most important risk factors for BCC
and SCC as follows:
                                 DRAFT FINAL  *

-------
                                     7-26
                                 EXHIBIT 7-8

                Annual Age Adjusted Incidence Rates for Basal and
              Squamous Cell Carcinomas (1977-1978 NCI  Survey Data)
             and Melanoma  (1973-1976 SEER Data)  Among White Males*
      500
         L
      TOO —
«••

a
a
w
,|
• •• —
"3
a
f£
2«5
1
A

o
a
I
5
6
^
3
a
1
0
a
u
c
c

-------
                                    7-27
                                EXHIBIT 7-9

               Annual  Age Adjusted  Incidence Rates for Basal and
             Squamous Cell  Carcinomas (1977-1978 NCI  Survey Data)
           and Melanoma (1973-1976 SEER  Data)  Among White Females*
   3
   a
                              XXjkK UinUkVtOlCT IUV-41 KAOIATIOM ("OCX
* According to annual UVB measurements  at  selected  areas of the United States,
with regression lines based on an exponential model.

Source: Scotto and Fraumeni (1982).
                                 DRAFT FINAL  * * *

-------
                                   7-28
        BCC:    cumulative solar exposure (p<0.001) » ability
                to tan (p<0.001) » age (p<0.005) »
                complexion (p<0.025)

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

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

    Exhibit 7-10 presents the estimated relative risks (RRs) of BCC and SCC
for the combinations of risk factors considered in the Vitaliano and Urbach
(1980) study.  Vitaliano and Urbach (1980) concluded that the most important
difference between SCC and BCC was their relationship with cumulative '
exposure.  As shown in Exhibit 7-10, a higher exposure level was required for
BCC than for SCC to reach similar RRs.  The authors noted that the maximum
response of BCC to solar exposure occurred in exposure category E4 (30,000 or
more hours) whereas for SCC it occurred in exposure category E3 (20,000-29,999
hours).  They observed that the results were consistent with the belief that
exposure to UV radiation has a greater effect on the development of SCC than
on BCC although an association between BCC and sunlight does exist.

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

Host  Factors

    As already mentioned, in addition to exposure factors, host factors play
an important role in the risk of nonmelanoma skin tumors.  Characteristics
associated with pigmentation, including eye, hair, and skin color, ability to


                          * * *  DRAFT FINAL  * * *

-------
                                   7-29
                                EXHIBIT 7-10

               Estimated Relative Risks of Basal and  Squamous
            Cell  Carcinomas for 32  Combinations of Risk Factors
   Exposure Grade
(Total Exposure, hr)
                                     Tans
  0-59 yrs      60+ yrs
                	Burns-Sensitive
                 0-59 yrs        60+ yrs
Dark   Pale   Dark    Pale   Dark    Pale    Dark     Pale
                                             Basal  Cell  Carcinoma
E4 (30,000 or more)
E3 (20,000-29,999)
E2 (10,000-19,999)
El (0-9,999)
E4
E3
E2
El
 3.19  4.94
 2.86  4.43
 1.77  2.75
 1.00  1.55
 7.09 22.79
 7.09 22.79
 4.42  5.72
 1. 92  1.00
 4.99   7.76
 4.49   6.95
 2.79   4.32
 1.57   2.43
                                           Squamous  Cell  Carcinoma
6.10
5.47
3.39
1.91
9.43
8.47
5.26
2.96
9.57
8.58
5.32
3.00
28.61  90.12
28.61  90.12
17.94  23.08
 7.76   4.03
                      14.80
                      13.29
                       8.25
                       4.65
26.61  84.66 107.70  347.08
26.61  84.66 107.17  347.08
16.60  21.41  66.99   86.52
 7.19   3.74  29.19   15.06
Source:  Vitaliano and Urbach (1980).
                              *  DRAFT FINAL  * * *

-------
                                   7-30
tan, tendency to sunburn, and ethnic origin are particularly important.
Evidence presented in the 1982 NAS report and by numerous other researchers
has identified those with fair skin, blue/green eyes,  red/blond hair,  a
tendency to burn and to rarely tan, and a Celtic heritage as being at  high
risk of developing nonmelanoma skin tumors (NAS 1982).

DOSE-RESPONSE  RELATIONSHIPS

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

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

Previous Analysis

    Dose-response relationships for exposure to UVR and nonmelanoma skin
tumors have been developed in both epidemiological and experimental studies.
Five major dose-response analyses are reviewed below;  three are based  on
epidemiological (ecological) studies, one is based on an animal experiment,
and the last is based on a molecular experiment.  In each of these studies,
UVB exposure is estimated by weighting individual wavelenths in the UVB
spectrum according to biological effectivenss.  A single weighting approach is
referred to as an action spectrum.  Several action spectra based on the  UVB
measurement approach used are available including Robinson-Berger (RB) meter,
human erythema and Setlow DNA.  The erethyma and DNA action spectra weight
those wavelengths thought to promote sunburn and damage to DNA more heavily
than other wavelengths.

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

    Since these early works, additional epidemiological data were made
available by NCI.  A new survey of skin cancer incidence was conducted by NCI
in 1977-1978 for eight locations across the U.S.  This  data consists of newly
                            * *  DRAFT FINAL

-------
                                   7-31
diagnosed cases of nonmelanoma skin cancer, and are age-, sex-, and cell
type-specific for whites only.  In addition, new readings from RB meters at
five locations in the U.S. were provided.  These readings combined with the
earlier RB readings provided UVB data for all eight locations included in the
1977-1978 NCI survey.

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

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

    Scotto efal. (1981) and Fears and Scotto (1983)

    Scotto et al. (1981) estimated values for the BAF by correlating the
incidence of nonmelanoma skin cancer with UVB dosage for eight locations in
the U.S.  Separate estimates are presented by location, sex, and cell type.
The functional form of the model is

                           In (R) = ln(a) + b U + e

where R is nonmelanoma skin cancer incidence, U is UVB radiation dosage, a and
b are constants, and e is an error term.  Incidence rates were obtained from
the 1977-1978 NCI survey.  UVB dosages were obtained from RB meters located at
major metropolitan airports at each location.  The coefficients were estimated
using ordinary least squares regression.

    Scotto et al. (1981) found that although incidence rates varied by age,
BAFs did not vary significantly by age.  Age-adjusted incidence rates were
used for all estimates.  Their estimates of the BAFs for BCC ranged from 1.32
to 2.59 for males and 1.06 to 2.07 for females.  The estimates of the BAFs for
SCC were slightly greater, from 2.08 to 7.09 for males and 2.18 to 4.30 for
females.  In both cases, higher estimates were obtained for locations at lower
latitudes.

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

                        In (R) = In (a) + b In  (U) + e
                            * *  DRAFT FINAL  * *

-------
                                   7-32
Separate estimates are presented by sex, location, and percent increase in UVB
dosage but not by cell type.  The data used were the same as those used in the
earlier study.

    Fears and Scotto (1983) concluded that the relationship between UVB
radiation and nonmelannoma skin cancer incidence could be modeled using either
the log-linear or the log-log form.  They noted that there was no strong
reason to prefer one form of the model over the other.  The BAF estimates
presented in this study ranged from 1.4 to 3.2 for males and  1.1 to 2.3 for
females.  Higher estimates were obtained for locations at lower latitudes and
higher percent changes in UVB dosage.

    Scott (1981)

    Scott used the log-linear equation as specified in the Fears and Scotto
(1983) study discussed above.  Incidence data were also obtained from the
1977-1978 NCI survey.  However, rather than using the RB meter data for UVB
dosage, Scott (1981) used the action spectrum for DNA damage.

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

    For example, using the DNA action spectrum, Scott (1981) estimated that a
5 percent ozone reduction would lead to an increase in BCC in Minneapolis-St.
Paul of 7.7 percent for males and 5.8 percent for females.  Using the RB
meter, with the same functional model and cancer incidence data, Scotto et al.
(1981) estimated corresponding increases to be 5.6 percent for males and 4.4
percent for females.  The reason for this difference lies in the differential
sensitivity of different weighting functions to ozone depletion -- the optical
amplification factor.  Wavelengths at 295-300 nm change much more than at
315-320 nm.  Thus the shape of the action spectrum significantly influences
the effects of a given ozone depletion.

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

    Rundell (1983)

    The Rundell study is the third epidemiological study to use the 1977-1978
NCI survey data.  Measurements from the RB meter were used to estimate UVB
exposure.  However, Rundell used a different dose-response model than that


                          * * *  DRAFT FINAL  * * *

-------
                                   7-33
used by Fears and Scotto (1983) or Scott (1981) to estimate the BAFs.   The
Rundell model used age as a proxy for cumulative sun exposure and estimated
skin cancer incidence at a single location as a function of this proxy
variable.  Separate BAF estimates were made by sex, cell-type, and location.

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

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

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

    de Gruijl et al. (1983)

    The dose-response analysis presented by de Gruijl et al. (1983) was based
on data from an animal experiment.  Mice were exposed to radiation from a
sunlamp for 3.25 hours per day at seven different levels ranging from
0 mJ/cm2 to 190 mJ/cm2.  The response was measured by observing the
initiation and growth of tumors on the mice.  The results of this study showed
that the initiation of tumors was dependent on the dose, but that the growth
and yield of tumors were independent of dose.  He also concluded that
reciprocity was not supported, except at high dosage levels.

    Kubitschek et al.. (1986)

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

            Sm = / E(8,X,w) T(X) E(\) dX
                          * * *  DRAFT FINAL

-------
                                   7-34
        where

            Sm       = Rate of induction of mutation
            9        = Solar zenith angle
            X        = Wavelength
            W        = Amount of stratospheric ozone
            E(0,X,W) = irradiation at earth's surface
            T(X)     = rate of transmission through skin
            s(X)     = action spectrum from E. coli

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

Uncertainties  in Estimates of Dose-Response

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

    A major issue affecting the uncertainty of the estimtes is the
availability of epidemiological data.  All of the epidemiological studies
described above use the 1977-1978 NCI survey for nonmelanoma-tumor incidence
data.  However, data are not currently available for individual UVB dosage.
Therefore, researchers have had to use a single measure of UVB dosage for all
individuals in a given location.  Actual dosages recieved will vary by
individual because of behavioral factors such as time spent outdoors.  The
incidence of nonmelanoma skin cancer will also vary by individual because of
genetic characteristics such as skin color.  If these factors are correlated
with location, the estimates could be biased.  In addition, the issue of peak
versus cumulative dose has not been addressed in these studies; although basal
and squamous cell skin cancers clearly have somewhat different etiologies,
suggesting somewhat different responses to UVB.

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

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


                          * * *  DRAFT FINAL  * * *

-------
                                   7-35
    There is also statistical uncertainty in the measurement of UVB flux.
This uncertainty is not captured in the regression analysis, which considers
the UVB values as certain.  As a result, the dose-response coefficients are
subject to a greater degree of statistical uncertainty than is indicated in
the regressions.  Incorporating statistical uncertainty in the estimates of
UVB would probably result in different dose-response coefficients, but the
direction and magnitude of this difference is not known (Zeger (no date)).

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

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

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

Sensitivity of Dose-Response Estimates to Choice of Action Spectrum

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

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

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

-------
                                   7-36
    However, because the optical amplification factor is sensitive to ozone
depletion, (that is, because UVB at 295-300 nm increases more than at 315 nm-
320 nm) estimates of the effect of ozone depletion on UVB dosage were found to
differ significantly by choice of action spectrum.  .Use of an action spectrum
based on the effects on a biological target, of course, makes sense.
Nevertheless, an issue still exists about whether DNA, erethyma, mouse edema,
or some other biological action spectrum should be used.  The Setlow DNA
action spectrum was found to give the highest estimates of the effect of ozone
depletion on UVB dosage of the three examined (see Exhibit A-4 in Appendix
A).  Obviously, the choice of an action spectrum will influence the estimates
of the effect of ozone depletion on skin cancer incidence.
                          * * *  DRAFT FINAL  * *

-------
                                   7-37
                                APPENDIX A

              Sensitivity Analysis of the Selection of Action
                     Spectra on Dose-Response  Estimates
             Relating UVR Exposures to Nonmelanoma Skin Tumors
    This appendix presents the results of an analysis of the relationship
between changes in ozone levels and incidence rates of nontnelanoma skin
tumors.  Estimating the impact of ozone change on tumor incidence requires two
pieces of information: (1) the response of UVB radiation to changes in ozone,
and (2) the dose-response relationship between UVB exposure and the incidence
of nonmelanoma skin tumors.  This analysis focuses on the sensitivity of the
selection of an action spectrum used to describe the relationship between UV
exposure and the incidence of nonmelanoma skin cancer.

    UVB Exposure and Nonmelanoma Skin Cancer

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

                In (R) = a + b In (UV) + e

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

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

    Incidence of Nonmelanoma Skin Cancer

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

    UVB Radiation

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

-------
                                   7-38
    Data on UVB radiation were obtained from the NASA UVB model (Serafino and
Frederick 1986).  This model uses mathematical relationships to estimate UV
solar flux along different wavelengths (UVA and UVB) reaching the Earth's
surface at any time of year and location on the globe.  To estimate UV flux,
the model uses information about global ozone abundance, terrain height, cloud
cover, cloud transmission, and surface characteristics.

    The NASA model provided estimates of UVB radiation for each of the eight
U.S. locations.  The data were provided for all three action spectra, and for
each action spectrum, data were presented for three different time periods --
annual, a day in the month of June, and a clear day.  In addition to UVB data
at current ozone levels, the NASA model provided estimates of radiation
reaching the Earth's surface at different ozone levels.

    To analyze the significance of measuring UVB radiation using different
action spectra and time periods, Exhibit A-l shows correlations among the NASA
UVB data for the three action spectra and for annual, month of June, and clear
day time periods.  The exhibit also shows correlations among actual UVB
measurements (Scotto, Fears, and Fraumeni 1981) and the modeled data.  The
strong correlation among these UVB data suggests that dose-response
coefficients are not highly sensitive to the different measures.  Nonetheless,
a regression analysis was performed using dose-response coefficients for the
annual time period for all three action spectra.  Before performing the
regression analysis, two weights were applied to the data.  First, the data
were weighted by the population of each respective location.  Exhibit A-2
shows the populations of these locations.  Second, as a correction for
heteroskedasticity, the data were weighted by the inverse of the estimated
variance of the logarithm of the age-adjusted incidence rates.

    Exhibit A-3 presents regression estimates of dose-response coefficients
for basal and squamous cell skin cancers by sex.  Separate estimates are given
for all three action spectra (annual only).  The exhibit shows a low, middle,
and high estimate.  The low value represents the estimated coefficient minus
one standard error.  The high coefficient adds one standard error.  The
dose-response coefficients for SCC are higher than those for BCC for each of
the action spectra.  In addition, the coefficients for SCC were found to be of
greater statistical significance.  These findings are consistent with the
belief that cumulative UVB radiation has a greater effect on the development
of SCC than on the development of BCC..  As already mentioned, epidemiological
observations indicate that the ratio of basal to squamous cell skin cancer
decreases with decreasing latitude and that BCC is more likely to develop on
regularly unexposed sites.  In addition, the sensitivity of basal cell cancer
incidence rates to UVB was higher for males than for females.  The relative
sensitivity by sex is reversed for SCC.

    It is important to recognize that the NASA models UVB flux values are
subject to statistical uncertainty.  This uncertainty is not captured in the
regression analysis which treats the UVB values as certain.  As a result, the
dose-response coefficients are subject to a greater degree of statistical
uncertainty than is indicated in the regressions.  Incorporating statistical
                          * * *  DRAFT FINAL  * * *

-------
                                        7-39
                                    EXHIBIT A-1

               Correlation of Alternative Measurements  of  UVB  Radiation
                      for Ten Locations in the United States
                        Annual
              RB  b
            Actual   RB    Erethyma   DNA
     Clear Day
RB
Erethyma   DNA
                     Month of June
RB
Erethyma   DNA
Annual b
RB Actual
RB
Erethyma
DNA

1.00 0.979
1.00



0
0
1


.988
.990
.00


0.
0.
0.
1.

987
984
999
00

0.940
0.964
0.941
0.932

0.977
0.975
0.985
0.984

0.977
0.970
0.986
0.987

0.882
0.935
0.883
0.866

0.955
0.975
0.959
0.949 .

0.967
0.989
0.971
0.963
Clear Day
RB
Erethyma
DNA
Bfnth of June
RB
Erethyma
DNA

1.00 0.986 0.955 0.938
1.00 0.999 0.869
1.00 0.850

1.00



0.967
0.947
0.936

0.978
1.00


0.968
0.960
0.951

0.967
0.999
1.00
 The ten locations include:   Seattle,  Washington;  Minneapolis,  Minnesota;  Detroit,
Michigan; Des Moines, Iowa;  Salt Lake  City,  Utah;  San Francisco-Oakland,  California;
Atlanta, Georgia; Dallas-Ft.  Worth,  Texas;  New Orleans,  Louisiana;  Albuquerque,  New
Mexico.  Data for Des Moines  and Dallas-Ft.  Worth  were included in  the correlations;
however, because nonmelanoma  skin cancer rates were unavailable,  these locations were
excluded from the dose-response regression analysis.
 Annual UVB data from Scotto, Fears and Fraumeni (1981).
measurements taken at National Weather Service stations.
             These data are actual UVB
 Model data from the NASA UV model.
                               * * *  DRAFT FINAL  * * *

-------
                          7-40
                      EXHIBIT A-2

Population  Weights for Ten Locations in  the  United States
                    Location            Weight


             Seattle                    1,607,469

             Minneapolis-St.  Paul        2,113,533

             Detroit                    4,353,413

             Salt Lake City3            1,461,000

             San Francisco-Oakland      3,250,630

             Atlanta                    2,029,710

             New Orleans                 1,076,204

             Albuquerque0               1,303,000
             a
              State of Utah.

             b
              San Francisco.

             c
              State of New Mexico.

             Source:   State and Metropolitan Area Data
                      Book, 1982,  U.S.  Department of
                      Commerce, Bureau  of the Census,
                      U.S. Government Printing Office
                      August 1982.
                 * * *  DRAFT FINAL  * * *

-------
                                   7-41
                               EXHIBIT A-3

           Estimated Dose-Response Coefficients (and t-Statistics)
                  for Basal  and Squamous Cell  Skin  Cancers
                      (UVB Dose-Skin Cancer Incidence)
              Robinson-Berger         Human  Erethyma       	Set low DNA	
              a                b      a               b      a                b
           Low      Mid    High    Low      Mid     High    Low      Mid    High


Basal
  Male     1.300   1.810   2.330   1.020    1.410    1.800   0.932   1.290   1.650
                   (3.53)             .      (3.60)                  (3.60)

  Female   0.443   1.050   1.670   0.346    0.809    1.270   0.316   0.739   1.160
                   (1.73)                   (1.75)                  (1.75)

Squamous
  Male     1.940   2.840   3.730   1.540    2.210    2.880   1.420   2.030   2.640
                   (3.20)                   (3.31)                  (3.33)

  Female   1.880   3.060   4.230   1.570    2.420    3.260   1.470   2.220   2.980
                   (2.60)                   (2.87)                  (2.94)
a
 Estimated coefficient minus one standard error.

b
 Estimated coefficient plus one standard error.
                          * * *  DRAFT FINAL  * * *

-------
                                   7-42
uncertainty into the estimates of UVB would probably result in different
dose-response coefficients, but the direction and magnitude of this difference
is not known.

    Ozone Levels and Changes in UVB Exposure

    The NASA UVB model was used to estimate changes in UVB radiation
associated with changes in ozone levels.  Exhibit A-4 shows representative
estimates for San Francisco.  In the exhibit, a two percent reduction in ozone
resulted in an estimated 1.6 percent increase in UVB radiation as measured on
a RB meter.  The estimated percentage increase is greater using the human
erethyma or Set low DNA action spectra.

    Relationship Between Ozone Changes and the Incidence of
    Nonmelanoma Skin Tumors

    The NASA estimates of the sensitivity of UVB radiation to changes in ozone
were combined with the UVB dose-skin cancer incidence coefficients.  Exhibits
A-5 and A-6 present the results of this analysis for an estimated 2 percent
and 10 percent depletion in the ozone layer for San Francisco.  For example,
Exhibit A-5 indicates that the steady-state incidence rate of basal cell skin
cancer could increase by 2.98 percent for males and 1.72 percent for females
in response to a 2 percent depletion in ozone.
                            * *  DRAFT FINAL  * *

-------
                              7-43
                          EXHIBIT A-4

Estimated Percentage Changes in UVB Radiation  in San  Francisco
        For a Two and  Ten Percent Depletion  in Ozone
                                Ozone Depletion  (percent)
                                     2           10
             Robinson-Berger         1'. 6           8.6

             Human Erethyma          3.5          19.0

             Setlow DNA              4.3   -       23.0
                     * * *  DRAFT FINAL  *  * *

-------
                                 7-44
                             EXHIBIT A-5

  Percentage Change in  Incidence of Basal and Squamous Cell Skin Cancers
          for a Two Percent Depletion  in  Ozone for San Francisco
            Robinson-Berger         Human Erethyma            Set low DNA	
           Low    Mid    High     Low    Mid    High      Low   Mid    High
Basal
  Male     2.13    2.98    3.86     3.55   4.95    6.36     3.99    5.56    7.17
  Female   0.72    1.72    2.75     1.19   2.81    4.44     1.33    3.15    4.99

Squamous
  Male     3.20    4.72    6.24     5.41   7.86   10.36     6.14    8.89    11.72
  Female   3.10    5.09    7.11     5.52   8.64   11.81     6.36    9.76    13.32
                            *  DRAFT FINAL  * * *

-------
                                  7-45
                              EXHIBIT A-6

   Percentage Change in Incidence  of Basal and  Squamous  Cell Skin Cancers
           for a  Ten Percent Depletion  in Ozone for San Francisco
              Robinson-Berger         Human Erethyma       	Setlow DNA	
            Low     Mid     High     Low     Mid    High     Low     Mid    High
Basal
  Male     11.34   16.14    21.24    19.10   27.34   36.14   21.15   30.41   40.45
  Female    3.73    9.07    14.80    6.11   14.87   24.32    6.72   16.43   26.97

Squamous
  Male     17.39   26.46    36.12    30.20   46.05   63.82   33.95   51.87   72.19
  Female   16.81   28.78    41.86    30.88   51.40   74.84   35.34   57.93   84.68
                          *  *  *  DRAFT FINAL  * * *

-------
                                   7-46
                             REFERENCES
Anderson, R.R.  "Tissue optics and photoimmunology," Chapter 4, In:
    Photoimmunology, J.A. Parrish, M.L. Kripke, and W.L. Morison, (eds).   New
    York: Plenum Medical Book Co. pp 61-76 (1983).

Beral, V. and N. Robinson.  "The relationship of malignant melanoma, basal and
    squamous skin cancers to indoor and outdoor work,"  Br J Cancer
    44:886-891  (1981).

Blum, H.F.  "Wavelength dependence of tumor induction by ultraviolet radiation.
    JNCI 3:533-537 (1943).

Blum, H.F., H.G. Grady, and J.S. Kirby-Smith.  "Relationships between dosage
    and rate of tumor induction by ultraviolet radiation," JNCI 3:91-97
    (1942).

Briele, H.A. and T.K. das Gupta.  "Natural history of cutaneous malignant
    melanoma," World J Surg 3:255-270  (1979).

Bruls, W.A.G., H. Slaper, J.C. van der Leun,  and L. Berrens.   "Transmission of
    human epidermis and stratum corneum as a function of thickness in the
    ultraviolet and visible wavelengths,"  Photochemistry and Photobiology,
    40(4):485-494 (1984).       ;   .

Burnham, D.K., L.C. Gahring, and R.A. Daynes.  "Clonal origin of tumors
    induced by ultraviolet radiation," JNCI 76:151-158 (1986).

Cole, C.A., D. Forbes, and R.E. Davies.  "An action spectrum for UV
    photocarcinogenesis," Photochem Photobiol 43:275-284 (1986).

Daynes, R.A.,  M.K. Schmitt, L.K. Roberts and C.W. Spellman.  Phenotypic and
    physical characteristics of the lymphoid cells involved in the immunity to
    syngeneic UV-induced tumors.  J. Immunol. 122:2458-2464 (1979).

De Gruijl, F.R., J.B. van der Meer, and J.C.  van der Leun.   "Dose-time
    dependency of tumor formation by chronic UV exposure," Photochem
    Photobiol 37:53-62 (1983).

Dunn, J.E. Jr., C.A. Levin, G. Linden et al.   "Skin cancer as a cause of
    death," Calif Med 102:361-363 (1965).

Eastcott, D.F.  "Epidemiology of Skin Cancer in New Zealand," NCI Monograph No.
    10:141-151.  (1963).

Emmett, A.J.  "Basal cell carcinoma," Chapter 4, Malignant Skin Tumors.
    Emmett, A.J. and M.G.E. O'Rourke, Eds.  Churchill Livingston, New York
    (1982).
                            * *  DRAFT FINAL  » *

-------
                                   7-47
Epstein, W.L.  "Malignant Sun Induced Squasmas Cell Carcinoma of the Skin,"
    J. Dermatol Surg Oncl, 9:505-506 (1983).

Epstein, W.L., J.C. Bystryn, R.  Edelson, P.M. Elias, D.R. Lowy, and S. Yuspa.
    "Nonmelanoma skin cancer melanomas, warts and viral oncogenesis," J  Amer
    Acad Dermatol, 5(2):960-970  (1984).

Epstein, J.H., and W.L. Epstein.  "Cocarcinogenic effects of ultraviolet light
    on DMBA tumor initiation in albino mice," J.  Invest Derm 39:455-460
    (1962).

Fears, T.R. and J. Scotto.  "Estimating Increases in skin cancer morbidity due
    to increased in ultraviolet ratiation exposure," Cancer Investigation,
    l(2):119-26 (1983).

Fears, T.R.,  J. Scotto, and M.A. Schneiderman.  "Mathematical models of age
    and ultraviolet effects on the incidence of skin cancer among whites in
    the United States," Am J. Epidemiol 105:420-27 (1977).

Findlay, G.M.  "Ultraviolet light and skin cancer," Lancet ii:1070-1073
    (1928).

Fisher, M.S., and M.L. Kripke.  Systemic alteration induced in mice by
    ultraviolet radiation and its relationship to ultraviolet carcinogenesis.
    Proc. Natl. Acad. Sci. USA.  74:1688-1692, (1977).

Fisher, M.S., and M.L. Kripke.  Further studies on the tumor-specific
    suppressor cells induced by ultraviolet radiation.  J. Immunol.
    121:1139-1144, (1978).

Fitzpatrick,  T.B.  and N.A. Soter.  "Pathophysiology of the skin," In
    Pathophysiology:  The Biological Principles of Disease, L.H. Smith, Jr.
    and S.O.  Thier (eds). pp. 1297-1330  (1985).

Flaxman, B.A.  "Growth in vitro and induction of differentiation in cells of
    basal cell cancer," Cancer Res 32:462-469 (1972).

Freeman, R.G.  "Data on the action spectrum for ultraviolet carcinogenesis,"
    JNCI 55:1119-1122 (1975).

Gange, R.W. and J.A. Parrish.  "Acute effects of ultraviolet radiation upon
    the skin," Chapt 5, In: Photoimmunology J.A.  Parrish, M.L. Kripke, and
    W.L. Morison (eds), New York: Plenum Medical Book Co., pp 77-94 (1983).

Goodman, G.J., R.  Marks, T.S. Selwood, M.W. Ponsford,  and W. Pakes.
    "Nonmelanotic skin cancer and solar keratoses in Victoria -- clinical
    studies II," Aust J Derm 25:103-106 (1984).

Green, A.C.,  and M.G.E. O'Rourke.  "Cutaneous malignant melanoma in
    association with other skin cancers," JNCI 74(5):977-980 (1985).
                              *  DRAFT FINAL  *

-------
                                   7-48
Green, A.C., AES.  "Ultraviolet exposure and skin cancer: response," Am J.
    Epidemiol 107:277-80 (1978).

de Gruijl, F.R., J.B. Van der Meer, and J.C. Van der Luen.   "Dose-time
    dependency of tumor formation by chronic UV exposure," Photochem
    Photobiol. 37(l):53-62 (1983).

Harris, T.J.  "Squamous cell carcinoma," Chapter 5, In:  Malignant Skin
    Tumors.   A.J. Emmett and M.G.E. O'Rourke, (eds.), Churchill Livingston,
    New York (1982).

Hillstrom, L. and G. Swanbeck.  "Analysis of etiological factors of squamous
    cell skin cancer of different locations," Acta Dermatovener, 50:129-133
    (1970).  .

Hsu, J., P.D. Forbes, L.C., Harber, and E. Lakow.  "induction of skin tumors
    in hairless mice by a single exposure to UV radiation," Photochem
    Photobiol 21:185-288 (1975).

Hu, F.  "Melanocyte cytology in normal skin, melanocytic nevi and malignant
    melanomas:  A review," In:  The Pathology of Malignant Melanoma, A.B.
    Ackerman, (ed.), New York:  Masson Publishing  USA., pp 1-22  (1981).

Kaidbey, K.H., P.P. Agin, M.M. Sayre, and A.H. Kligman.  "Photoprotection
    by melanin -- A comparison of Black and Caucasian skin," J Amer .Acad
    Dermatol 1:249-260 (1979).

Kent,  A.  "Pathology of basal cell carcinoma," In:  Cancer of the Skin,
    R. Andrade, S.L. Gumport, and G.L. Popkin (eds.), W.B.  Sanders Co.,
    Philadelphia, pp. 845-882 (1976).

Kligman, L.H. and A.M. Kligman.  "Histogenesis and progression of ultraviolet
    light-induced tumors in hairless mice," JNCI 67:1289-1297 (1981).

Kripke, M.L.  "Latency, histology, and antigenicity of tumors induced by
    ultraviolet in three inbred mouse strains," Cancer Res 37:1395-1400
    (1977).

Kubilus, J., H.P. Baden, and N. McGilvray.  "Filamentous protein of basal cell
    epithelioma:  Characteristics in vivo and in vitro," J Natl Cancer Inst,
    65:869-875 (1980).

Kubitschek,  H.E., K.S. Baker, and M.J. Peak.  "Enhancement of mutagenesis and
    human skin cnacer rtes resulting from increased fluences of solar
    ultraviolet radiation," forthcoming in Photochemistry and Photobiology
    (1986).

Laerum, O.D. and O.H..Iversen (eds.).  "Biology of Skin Cancer  (Excluding
    Melanomas)," A series of workshops on the biology of human cancer.  Report
    No. 15,  International Union Against Cancer, UICC Technical Report Series,
    Volume 63, Geneva (1981).
                          * * *  DRAFT FINAL  * * *

-------
                                   7-49
Lee, J.H.  "Melanoma and Exposure to Sunlight," Epi Rev,  4:110-136,  (1982).

Marks, R., M.W. Ponsford, T.S. Selwood, G.  Goodman, and G.  Mason.
    "Non-melanotic skin cancer and solar keratoses in Victoria," Med J Aust
    2:619-622  (1983).

National Academy of Sciences  (NAS), Causes  and Effects of Stratospheric Ozone
    Reduction:  An Update," National Academy Press, Washington, D.C. (1982).

National Cancer Institute (NCI), 1985 Annual Cancer Statistics Review,
    Presented to National Cancer Advisory Board, National Institutes of
    Health, Bethesda, Maryland (1985).

National Research Council.  "Environmental  Impact of Stratospheric Flight:
    Biological and Climatic Effects of Aircraft Emissions in the
    Stratosphere," Climatic Impact Committee, Washington, B.C.:  National
    Academy of Sciences (1975).

Nordlund, J.J.  "Genetic basis of pigmentation and the disorders of
    pigmentation," In:  Pathology of Malignant Melanoma,  A.B. Ackerman (ed.),
    New York:  Masson Publishing, USA pp 23-45 (1981).

Pathak, M.A.  "Activation of the melanocyte system by ultraviolet radiation
    and cell transformation," Ann N Y Acad  Sci, 453:328-339 (1985).

Peak, M.J., J.G. Peak, M.P. Moehring, and R.B. Webb.  "Ultraviolet action
    spectra for DNA dimer induction, lethality, and mutagenesis in
    Escherichia coli with emphasis on the UVB region," Photochem Photobiol,
    40(5):613-620 (1984)..

Pearl, D.H.  "A stocastic cancer model involving the environment of
    transforming cells with application to  skin cancer," Ph.D. Thesis,
    University of California at Berkeley  (1984).

Pinkus, H.  "Premalignant fibroepithelial tumors of skin," Arch Dermatol
    Syphelol, 67:98-615 (1953).

Pollack, S.V., J.B. Gaslen, E.F. Sherertz,  and B.V. Jegasothy.  "The biology
    of basal cell carcinoma:  A review," J Amer Acad Dermatol 7:569-577
    (1982).

Romsdahl, M.M. and I.S. Cox.  "Biological aspects of pigment cells and
    malignant melanoma," In:  Neoplasms of the Skin and Malignant Melanoma,
    Year Book of Medical Publishers, Inc.,  pp 251-277 (1976).

Rundell, R.D.  "Promotional effects of ultraviolet radiation on human basal
    and squamous cell carcinoma," Photochem Photobiol, 38(7);569-575 (1983).

Rusch, H.P., B.E. Kline, and C.A. Baumann.   "Carcinogenesis by ultraviolet rays
    with reference to wavelength and energy," Arch Path 31:135-146  (1943).
                          * * *  DRAFT FINAL  * * *

-------
                                   7-50
Scott, E.L.  "Summary of studies made since December 1979:  Estimating the
    increase in skin cancer that will result from the depletion of
    stratospheric ozone," University of California, Berkeley,  Unpublished
    (transmitted to J.R. Hibbs, U.S. Environmental Protection Agency,
    Washington, B.C.) (1981).

Scott, E.L., and M.L. Straf.  "Ultraviolet radiation as a cause of
    cancer," pages 529-546, Origins of Human Cancer, Book A, Incidence of
    Cancer in Humans, edited by H.H. Hiatt, J.D.  Watson, and J.A.  Winsten,
    Vol. 4 of The Cold Spring Harbor Conferences  on Cell Proliferation, New
    York, N.Y.:  The Cold Spring Harbor Laboratory, (1977).

Scotto, J. and J.F.. Fraumeni, Jr.  "Skin (other than melanoma)," In:   Cancer
    Epidemiology and Prevention, D. Schottenfeld  and J.F. Fraumeni (eds.),
    Philadelphia, W.B. Saunders Co., pp. 996-1011 (1982).

Scotto, J., T.R. Fears,  and J.F. Fraumeni, Jr., Incidence of Nonmelanoma Skin
    Cancer in the United States, National Cancer  Institute, U.S. Department
    of Health and Human Services, Publ. No. (NIH) 82-2433 (1981).

Scotto, J., A.W. Kopf, and F. Urbach.  "Nonmelanoma skin cancer among
    Caucasians in four areas of the United States," Cancer 34:1333-1338,
    (1974).

Spellman, C.W., and R.A. Daynes.  Modification of immunologic potential by
    ultraviolet radiation.  II.  Generation of suppressor cells in short term
    UV-irradiated mice.   Transplantation 24:120-126 (1977).

Spellman, C.W., and R.A. Daynes.  Properties of ultraviolet light-induced
    suppressor lymphocytes within a syngeneic tumor system.  Cell. Immunol.
    36:383-387 (1978).

Spikes, J.D.  "UV-damage in biological systems,"  In:  Experimental and
    Clinical Photoimmunology, R.A. Daynes, and J.D. Spikes (eds),  CRC Press
    Inc.

Spikes, J.D., M.L. Kripke, R.J. Conner, and E.J.  Eichwald.   "Time of
    appearance and histology of tumors induced in the corsal skin of C3Hf mice
    by ultraviolet radiation from a mercury arc lamp," JNCI 59:1637-1643
    (1977).

Stenbeck, F.  "Species-specific neoplastic progression by ultraviolet light
    on the skin of rats, guinea pigs, hamsters and mice," Oncology 31:309-225
    (1975).

Strickland, P.T., F.J. Burns, and R.E. Albert.  "induction of skin tumors in
    the rat by single exposure to ultraviolet radiation," Photochem Photobiol
    30:683-688 (1979).

Urbach, F., D.B. Rose and M. Bonnem.  "Genetic and environmental interactions
    in skin carcinogenesis," In:  Environment and Cancer, Williams and
    Wilkins Co.:  Baltimore, pp. 355-371 (1972),  as cited in Pollack et al.
    (1982).

                          * * *  DRAFT FINAL  * * *

-------
                                   7-51
Van der Luen, J.C.  "Yearly Review:  UV-Carcinogensis," Photochem and
    Photobiol 39(6):861-868 (1984).

Vitaliano, P.P. and F. Urbach.  "The relative importance of risk factors in
    nonmelanoma carcinoma," Arch Dermatol, 116:454-456 (1980).

Winkelman, R.K., P.E. Zollman, and E.J. Baldes, "Squamous cell carcinoma
    produced by ultraviolet light in hairless mice," J. Invest Derm
    40:217-224 (1963).

Zeger, S.L., L. Cameron, E. Gold, and E. Emmett.  "Nonmelanoma skin cancer and
    ultraviolet radiation," Johns Hopkins School of Hygiene and Public Health,
    unpublished (undated).
                          * * *  DRAFT FINAL  * *

-------
Chapter 8

-------
                                CHAPTER 8

                      CUTANEOUS MALIGNANT MELANOMA
SUMMARY

    In 1985, there were 25,000 cases of cutaneous malignant melanoma (CMM)  in
the United States and 5,000 deaths.   For more than a decade, there has  been
serious concern that CMM is at least partially caused by ultraviolet-B
radiation (UVB).  However, several aspects of the scientific information about
CMM have puzzled researchers and have contributed to uncertainty about  the
relationship of CMM to solar radiation, in particular UVB.   In the past
several years, some progress has been made in understanding CMM and its
possible relation to solar radiation, and there now exists  an array of
evidence which indicates that exposure to solar radiation,  and UVB in
particular,  is a likely cause of CMM.

    Supporting evidence includes the fact that people who lack the protection
of pigmentation which blocks penetration of ultraviolet radiation (UVR) in  the
skin have higher CMM incidence rates; a correlation, in well-designed
ecological studies, of higher CMM rates with decreasing latitude and
increasing UVB levels; the association of freckling and nevus formation (risk
factors for CMM) with solar exposure; the differences between natives and
immigrants to sunny climates in CMM rates; high rates of CMM in patients who
are genetically deficient at repairing DNA damage induced by UVB; and the
indication,  in controlled studies, that sun exposure at early ages and  of an
intermittent nature results in higher CMM risks.

    Information that has created uncertainty about the relationship between
solar radiation and CMM has included: a failure to find latitude gradients  in
some ecological studies; the fact that outdoor workers have lower CMM rates
than indoor workers (but higher non-melanoma skin cancer rates); and the fact
that anatomic sites with lower sun exposure have high CMM rates.

    The latter evidence has made it clear that the relationship between solar
radiation and CMM is not a simple one.  In fact,  this evidence can only be
reconciled with the hypothesis that solar radiation is an etiological factor
in CMM if one concludes that cumulative dose based on total number of hours in
the sun is not the appropriate dose parameter to correlate with CMM
development.  This has led to the emergence of hypotheses that postulate the
importance of intermittent exposure to high fluxes of UVB (resulting in large
UVB doses) in the causation of CMM.

    Unfortunately, at this time, there have been no studies which examine the
correlation between CMM and actual measured UVB doses received by individuals,
thus making it impossible to prove various hypotheses on CMM.  The balance  of
available evidence appears to support the conclusion that solar radiation is
at least a major cause of CMM.  Other evidence appears to indicate that
cumulative sun-exposed hours is not an appropriate measure of dose for  CMM,
and that CMM has a complicated etiology.
                          * * *  DRAFT FINAL  * * *

-------
                                   8-2
    A variety of different kinds of evidence supports the hypothesis that UVB
is a component of solar radiation that causes CMM.  Most important is the fact
that xeroderma pigmentosum patients, who are genetically deficient in
repairing UVB-induced DNA damage, have significantly higher rates of CMM than
the general population.  Another fact that supports the hypothesis that UVB is
a component of solar radiation that causes CMM is the fact that, in
experimental studies, UVB is the most mutagenic and carcinogenic waveband as
well as being the most active at causing immunosuppression.  All of this
information suggests that UVB is the primary component of solar radiation that
causes CMM.  Together, this evidence does not provide proof that UVB is the
waveband of solar radiation that causes CMM; however, the preponderance of
evidence makes UVB a likely cause of CMM.

    Finally, although much remains uncertain about the relationships between
solar radiation UVB and CMM, enough information exists to make reasonable
estimates of the dose-response relationships which exist between UVB and CMM.
These can then be used to estimate future cases and mortality of CMM if ozone
depletion occurs.  The possible relationships based on an improving but still
deficient database in the United States indicate that each 1 percent ozone
depletion will produce around a 1 to 2 percent increase in incidence and 0.8
to 1.5 percent increase in mortality.  At this time, it is impossible to
estimate how incidence and mortality of CMM will change in other countries
whose populations have unknown skin pigmentation and behavior.
                            * *  DRAFT FINAL  * *

-------
                                   8-3
FINDINGS

1.  CUTANEOUS MALIGNANT MELANOMA (CMM) IS A SERIOUS LIFE-THREATENING DISEASE
    THAT AFFECTS A LARGE NUMBER OF PEOPLE IN THE UNITED STATES.

    la.  CMM incidence and mortality is increasing among fair-skinned
         populations.  These increases appear to be real and are not just the
         result of improved diagnosis and reporting.

    Ib.  In 1985, there were 25,000 cases of CMM and 5,000 fatalities related
         to melanoma in the United States.  If current trends continue even
         without ozone depletion, the lifetime risk based on current data
         appears to be about 1 in 150 in the United States; compared to less
         than 1 in 66 currently observed in Australia.

2.  LIMITATIONS IN THE DATABASE PREVENT ABSOLUTE CERTAINTY ABOUT THE
    RELATIONSHIP OF SOLAR RADIATION, UVB. AND CUTANEOUS MALIGNANT MELANOMA.

    2a.  There is no animal model that uses UVB to induce melanomas
         experimentally.

    2b.  There is no experimental in vitro model for malignant
         transformation of melanocytes,

    2c.  .No epidemiologic studies of CMM have been.conducted in which
         individual human UVB exposures (and biologically effective doses of
         solar radiation) have been adequately assessed.

3.  EVALUATION OF THE EPIDEMIC-LOGICAL AND EXPERIMENTAL DATABASES REQUIRES
    CLOSE ATTENTION TO THE RELATIONSHIP OF WAVELENGTH AND DOSE AND TO THE
    VARIATIONS OF SOLAR RADIATION IN THE AMBIENT ENVIRONMENT.

    3a.  Ozone differentially removes wavelengths of UVB between 295-320 nm;
         UVA (320-400 nm) is not removed above 350 nm and visible light
         (400-900 nm) is not removed.  It totally removes all UV-C (i.e.,
         wavelengths less than 295 nm).

    3b.  Wavelengths between 295 nm and 300 nm are generally more biologically
         effective (i.e., damage target molecules in the skin including DNA)
         than other wavelengths in UVB and even more so than UVA radiation.

    3c.  Latitudinal variations exist in solar radiation; model predictions
         indicate that the greatest variability is seen in cumulative UVB
         (e.g., monthly doses) followed by peak UVB (highest one-day doses)
         and then cumulative UVA.  Peak UVA does not vary significantly across
         latitudes up to 60°N.  Greater ambient variation also exists in UVB
         than UVA by time of day.
                          * * *  DRAFT FINAL

-------
                                   8-4
    3d.  The biologically effective dose of radiation that actually reaches
         target molecules depends on the duration of exposure at particular
         locations, time of day, time of year, behavior (i.e., in terms of
         clothes and sunscreens), pigmentation, and other characteristics of
         the skin including temporal variations (e.g., in tanning and in skin
         thickening).

    3e.  Cloudiness and albedo, although causing large variations in the
         amount of UVB and UVA, do not greatly change the ratio of UVB to UVA.

    3f.  Ozone depletion is predicted to cause the largest increases in
         radiation in the 295-299 nm UVB range, less in 300-320 nm UVB range,
         and UVA is virtually unaffected.

    3g.  Cutaneous malignant melanoma has a number of different histologic
         types which vary in their relationship to sunlight, site, and racial
         preference and possibly in their precursor lesions.  Assessment of
         these types is not consistent among registries, thus confounding
         attempts to evaluate associations between CMM and solar radiation.

    3h.  Melanin is the principal pigment in skin that gives it color; melanin
         absorbs UV radiation very effectively, thus the darker the skin, the
         more the basal layer is protected from UV.

4.  A LARGE ARRAY OF EVIDENCE SUPPORTS THE CONCLUSION THAT SOLAR RADIATION IS
    ONE OF THE CAUSES OF CUTANEOUS MALIGNANT MELANOMA.

    4a.  Whites have higher incidence and mortality rates of CMM than blacks.

    4b.  Light-skinned whites, including those who are unable to tan or who
         tan poorly, get more CMM than darker-skinned whites.

    4c.  Sun exposure leading to sunburn apparently induces melanocytic nevi.

    4d.  Individuals who have more melanocytic nevi develop more CMM; the
         greatest risk is associated with a particular type of nevus -- the
         dysplastic nevus.

    4e.  Sunlight induces freckling, and freckling is an important risk factor
         for CMM.

    4f.  Incidence has been increasing in cohorts in a manner consistent with
         changes in patterns of sun exposure, particularly with respect to
         increasing intermittent exposure of certain anatomical sites.

    4g.  Immigrants who move to sunnier climates have higher rates of CMM than
         populations who remain in their country of origin.  Immigrants
         develop rates approaching those of the adopted country; this is
         particularly accentuated in individuals arriving before the age of
         puberty (10-14 years).
                            * *  DRAFT FINAL  * * *

-------
                                   8-5
    4h.  CMM risk is associated with childhood sunburn; this may reflect an
         individuals pigmentary characteristics or may be related to nevus
         development.

    4i.  Most studies that have used latitude as a surrogate for sunlight or
         UVB exposure have found an increase in the incidence or mortality of
         CMM as one approaches the equator.  A recent study on incidence that
         used measured UVB and survey data found a strong relationship between
         UVB and incidence.  Another study that used modeled UVB data and an
         expanded database on mortality found a strong UVB-mortality
         relationship.

    4j.  Patients with xeroderma pigmentosum who cannot repair UVB-induced
         lesions in skin DNA have a 2,000-fold increase in CMM by the age of
         20.

    4k.  One form of CMM, Hutchinson's melanotic freckle melanoma, appears
         almost invariably on the chronically sun damaged skin of older people.

5.  SOME EVIDENCE CREATES UNCERTAINTY ABOUT THE RELATIONSHIP BETWEEN SOLAR
    RADIATION AND CUTANEOUS MALIGNANT MELANOMA.

    5a.  Some ecologic epidemiology studies have failed to find a latitudinal
         gradient for CMM.

    5b.  Outdoor workers generally have lower incidence and mortality rates
         for CMM than indoor workers, which appears incompatible with a
         hypothesis that cumulative dose from solar exposure causes CMM.

    5c.  Unlike basal cell and squamous cell carcinomas, most CMM occurs on
         sites that are not habitually exposed to sunlight; this contrast
         suggests that cumulative exposure to solar radiation or UVB is not
         solely responsible for variations in CMM.

6.  ULTRAVIOLET-B RADIATION IS A LIKELY COMPONENT OF SOLAR RADIATION THAT
    CAUSES CUTANEOUS MALIGNANT MELANOMA (CMM),

    6a.  Xeroderma pigmentosum patients who fail to repair UVB induced
         pyrimidine dimers in their DNA have a 2,000-fold excess rate of CMM
         by the time they are 20.

    6b.  UVB is the most active part of the solar spectrum in the induction of
         mutagenesis and transformation in vitro.

    6c.  UVB is the most active part of the solar spectrum in the induction of
         carcinogenesis in experimental animals and is considered by most to
         be a causative agent of nonmelanoma skin cancer in humans.

    6d.  UVB is the most active portion of the solar spectrum in inducing
         immunosuppression which may have a role in melanoma development.
                              *  DRAFT FINAL

-------
                                   8-6
    6e.   The limitations in the epidemological and experimental database leave
         some doubt as to the effectiveness of UVB wavelengths in causing CMM.

7.   WHILE SOME UNCERTAINTY EXISTS,  INCREASES IN THE INCIDENCE AND MORTALITY
    OF CUTANEOUS MALIGNANT MELANOMA ARE LIKELY AS A RESULT OF OZONE
    MODIFICATION.   ALTHOUGH IMPERFECT,  INFORMATION EXISTS TO ESTIMATE CHANGES
    IN INCIDENCE AND MORTALITY IF THE OZONE LAYER IS MODIFIED.

    la.   Well-designed ecological studies using survey data and measured UVB
         provide a reasonable estimate of dose response relationship for UVB
         and CMM.   A one percent change in ozone is likely to increase
         incidence by between slightly less than 1% and 2%, depending on the
         choice of action spectrum.

    7b.   Using predictions of UVB from a NASA model and a database for CMM
         mortality developed by EPA/NIA, a dose-response relationship for CMM
         mortality has been developed.   It is estimated that a one percent
         change in ozone would result in a 0.8 to 1.5 percent change in CMM
         mortality.

    7c.   Uncertainty exists about the appropriate action spectrum to be used
         in estimating dose, the best functional form for dose-response, and
         the best way to characterize dose (peak value, cumulative summer
         exposure, etc.).  Histologically different CMM (or possibly CMM
         located at different anatomical sites.) are likely to have different
         dose-response relationships.  Current estimates of dose-response
         relationships fail to consider these histological or site differences.

    7d.   Additional problems for projecting future incidence and mortality of
         CMM in the U.S. include a lack of a good database with regard to
         variations in skin pigmentation and human sun-exposure behavior among
         different populations and estimates of how these relationships may
         change in the future.

    7e.   At this time, projections of CMM increases in other parts of the
         world are not possible due to insufficient access to necessary data.
                          * * *  DRAFT FINAL  *

-------
                                   8-7
INTRODUCTION

    This chapter is designed to examine the current state of knowledge with
regard to the relationship between human cutaneous malignant melanoma and
solar radiation,  in particular UVB.  Two previous reviews examining this
question came to the following somewhat contradictory conclusions:

        "The NCI [National Cancer Institute] is collecting skin
        melanoma incidence from ten locations in the United States.
        Earlier data analyses indicate that skin melanomas, which
        predominate in Caucasians, are related to UVB exposure.  A
        preliminary finding is that the incidence is increasing and
        that the rate of increase may be greater than expected from
        earlier surveys.  The dose-response relationship between
        skin melanoma and UVB appears to be more complicated than
        that observed for nonmelanomas of the skin." (NAS 1978)

        "...since 1976, the case for an association between UVB and
        melanoma has been weakened rather than strengthened by the
        results of additional clinical, pathological and
        epidemiological studies.  Furthermore (with the exception of
        a single animal), it has not been possible to use UVB alone
        to induce melanomas in experimental animals." (NAS 1982)

These contrasting conclusions require that a closer examination be  made of-the
relationship between melanoma and UVB.

    Since publication of the second NAS document in 1982, more information has
been gathered; this chapter reviews briefly both the old and the new
information in order to ascertain what can be said at the current time about
the relationship of UVB to melanoma.  A much more detailed review of the
relevant information is presented in Appendix A.

BACKGROUND INFORMATION

    A number of factors can modify the amount of solar radiation which should
be considered the "dose" received by a particular target.  This section
presents a brief overview of factors which can influence this "dose" -- a
detailed discussion is provided in Chapter 2 of Appendix A.

Background on  Solar  Radiation and the Concept of Dose

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

-------
                                   8-8
    A location's latitude determines the average angle between the location
and the sun.  This angle, in turn, determines the thickness of the atmosphere
that photons must pass through before reaching the location.  The lower the
angle, the longer the path through the ozone-containing stratosphere that
photons must travel and the greater the amount of absorption that can occur.
A longer path also allows longer exposure to aerosols and, thus, more scatter
-- thereby additionally reducing the amount of energy reaching a location.
Because ozone absorbs different wavebands differently, however, all photons
are not retarded equally, instead differential gradients for various
wavelengths occur.  Exhibit 8-1 shows estimates from the UV Model, developed
by Serafino and Frederick (1986), which predict that at 12 noon on March 21,
the flux of UVB at 295-299 nm can be expected to vary from the equator to 70°N
by a factor of over 100, while UVA at 335-339 nm can be expected to vary by a
factor of 5.

    Season also has a tremendous impact on the flux of solar radiation at a
given location.  Again, this occurs because the relationship of a location to
the sun changes with season.  Exhibit 8-2 shows model estimates of how the
various wavelengths vary by month for Washington, D.C.  Energy at 295 nm
increases by about a factor of 10 from winter (December) to Spring (March) and
by about another factor of 10 by midsummer (July).  UVB at 305 nm shows much  .
less variation and UVA very little, if any, variation.

    Exhibit 8-3 shows the impact that varying the time of day has on estimates
of UVB.  Clearly, the time of .day of exposure is a very important factor in
determining the amount of UV energy received at the various wavelengths.
Variations in cloud cover and surface albedo also cause variations in ambient
solar energy levels, however, there is not much differential variation by
wavelength (for additional details, see Chapter 2, Appendix A).

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

    The amount of energy which is actually effective is determined by the
action spectrum of the effect of concern.  If, for instance, damage to DNA is
the underlying effect linking solar radiation to cutaneous malignant melanoma,
then not all of the photons delivered to the target molecules will be
effective.  Instead, the absolute amounts of energy delivered in the various
wavelengths need to be weighted according to the action spectrum presented in
Exhibit 8-4.  There are a number of different effects which might underlie the
relationship between solar radiation and CMM (details of which are presented
in Chapters 2 and 3 of Appendix A); a key characteristic of them is that
energy in the UVB wavelengths (295-320 nm) tends to be much more biologically
active.

    How much target cells can potentially receive depends on the time a person
spends outside during certain periods of the day and days of the year.  The
total amount of energy delivered at a given location provides an upper bound
                                 DRAFT FINAL  *

-------
                                   8-9
                                EXHIBIT 8-1


              Variation  in UV Radiation  by Latitude  as  Percent

                of Levels  at the Equator on March 21 at Noon
     100.0
 0)
 CP
 c
 H3
£.
u

•U
 c
 a;
 u
 1-1
10. Q
                                     I      I   I   I   I      I   I   I   I
       1.0 _
                                                                       nm
                        20             40             60



                           Degrees North Latitude
This graph shows the variation estimated by the NASA model for five wavebands

of ultraviolet radiation for a single  day, high noon, in March.  The variation

would be different for other days.
                          * * *   DRAFT FINAL  * *

-------
                                   8-10





                                EXHIBIT 8-2


                 UV Radiation  by Month  in Washington,  D.C.
    0)
    en
    c
    (0
   .c
   u

    Qj
   JJ

   (0
   r— I

   0)
            I  I  I  I  I  I  I  I  I  I  I  I  I  I  I  I  IAI  I    I  I  I  I    I  I  I  I
This graph shows that the NASA model estimates that radiation at 295 nm has a

much larger proportional gain than at 305 or higher wavelengths.  Note that

the 335 nm line is almost coincident with the x axis,  indicating low monthly

variation in Washington, D.C.
                          « * *  DRAFT FINAL  * * *

-------
                           8-11
                        EXHIBIT 8-3

              Relative Change in UV Flux by Hour
               in Washington,  D.C., on June 21
(X 100)
      I   I  I  I   I  I  I  I   I  I  I   I  I  I     I  I  I   I  I  I   I  I  I
                                                  5 pm
8 pm
                               Hour
                   * * *  DRAFT FINAL  * * *

-------
             §
              do-*
;«0-2|
                                  8-12






                               EXHIBIT 8-4


                       Average DNA Action  Spectrum
                                                         10'
                                                         10'
                                                        to
                                                        •10'
                                                            a:
                                                            tf
                                                          10-
                                                           
-------
                                   8-13
of exposure, not the actual exposure.  Few individuals are out in the sun
during all daylight hours, therefore actual exposure is correspondingly
reduced.  As indicated above, the seasonal and hourly variations in incident
energy from solar radiation can be considerable particularly in those
wavelengths that are the most biologically effective (295-299 nm) so all hours
in the sun cannot be considered equal.  In addition, people living in areas
having the same number of sunlit hours may have additional behavioral
differences that modify the amount of radiation reaching the skin (the actual
exposure dose).  For example, some people wear lots of clothing, others do
not.  Some people wear sunscreens, while others use sun reflectors to gather
more solar radiation.

    In addition, the transition from the amount of energy to which an
individual is exposed to the amount of energy which is delivered to a target
molecule (the biologically effective dose), involves the interplay of factors
such as skin thickness and pigmentation.  Keratin and melanin both absorb
ultraviolet radiation.  The greater the skin thickness and degree of
pigmentation, the greater the amount of energy is prevented from penetrating
the two potential targets.  Thus earlier exposure to solar radiation and skin
response can influence doses of solar radiation, by wavelength, reaching
possible targets.

Background  on Melanoma and the Concept of Response

    Malignant melanoma is a rare tumor that arises as the result of the
neoplastic transformation of a melanocyte.  Melanocytes are pigment-producing
cells.  Exhibit 8-5 shows the relationship of the melanocyte to other cells in
the epidermis and other skin layers.

    The major function of the melanocyte is the production and distribution of
melanin to keratinocytes -- the major cell population of the skin.  Melanin is
a pigment which absorbs UV light over a broad range of wavelengths (25 nm to
1200 nm); it is the principal chromophore responsible for the difference
between black and white skin in the transmittance of UV light (shown in
Exhibit 8-6).

    The biology of cutaneous malignant melanoma is complex; there are several
morphologic types of tumor that may have different pathways of histogenesis;
they tend to behave differently in terms of age at appearance and
characteristic location, and yet have common elements in their progression.
Brief information is presented below; more detailed information is found in
Appendix A.

    Terminology for the four principal classes of melanomas was proposed by
McGovern et al. (1973) and has been subsequently modified (Smith 1976; Elder
et al. 1980).  The four principal types of melanoma are given below:

        (1)  Melanoma arising in Hutchinson's melanotic freckle
             (HMF); it occurs predominantly on sites receiving the
                          * * *  DRAFT FINAL  * * *

-------
                                      8-14
                                  EXHIBIT 8-5


                  Location  of Melanocyte in the Epidermis





                              Cornlfltd ctll*;


                           M«ltnln du»t


                         M«l«nosom«s.



                       Kcratlnocyu
     Stratum comeum<

     Qranular layer


 Keratlnocyte layer
                            Dendrltea
                            Dellverlna  ^^>«g
                                         ~
EPIDERMIS
     Baaal call lay*
                                               ollaganoua
                                              fiber
                             alum eomwMi   •

                               ar call layer
                                             EPIDERMIS
                          •quamou* eeH layer


                            •aaal cell layer
     Melaneeytei



Basement membra
                                             PAPILLARY DCRMIS
                                            -RETICULAR DERMIS
                                            _ SUBCUTIS
                        EPI-

                      \DERMIS
                             * * *  DRAFT FINAL   * * *

-------
    100  -
49
U
c.
o
c
a
      230
                                    8-15





                                EXHIBIT 8-6


                  Comparative Transmittance of UV Radiation
4OO
eoo
                                                 800
                            Wavelength  ( nm)
 Comparison of measured  epidermal transmittance A,  Caucasian; B, Dark Black.


 Source:  Wan et al.  1981.
                          * * *  DRAFT FINAL  * *

-------
                                   8-16
             greatest cumulative exposure and is thought to have the
             strongest relationship to cumulative solar exposure.

        (2)  Superficial spreading melanoma (SSM); it shows the
             strongest preference for sites that are intermittently
             exposed (trunks in males, lower extremities in
             females). It also shows a stronger association with
             nevi-possible precursor lesion.

        (3)  Nodular melanoma (NM); it has growth characteristics
             which differ from that of SSM and HMFM, although these
             types may progress to NM.

        (4)  Unclassifiable melanoma (UCM).

A fifth type of melanoma -- acral lentiginous melanoma (ALM) -- has been
distinguished with a site preference for soles, palms, and subungual surfaces.
It is virtually the only form observed in Blacks and in people who have never
lived south of the Artie Circle (Clarke et al. 1986).  Thus in characterizing
response to dose, aggregation of all types of melanoma will tend to obfuscate
real relationships.

EPIDEMIOLOGIC  EVIDENCE

    With the above background, this chapter will now review the epidemiologic
and experimental evidence relevant to assessing the role of solar radiation
and UVB in the development of CMM.  Because of the extent of the information
to be reviewed, the epidemiologic information has been described in five
segments:  time trends, anatomical site distribution, exposure factors, host
factors, and miscellaneous factors.  This is followed by a brief review of the
experimental evidence and a discussion of possible dose-response relationships.

Time Trends

    As indicated in Exhibit 8-7, there have been sharp increases in incidence
and mortality due to CMM reported in Caucasian populations worldwide (Magnus
1982; Rousch et al. 1985b; Osterlind and Jensen 1986).  Based on observational
and analytical evidence, most experts agree that these trends in CMM incidence
and mortality are genuine, and not due to increases in the registration and
diagnosis of the disease (Magnus 1975; Pakkenen 1977; Malec et al. 1977).

    Birth cohort analyses of CMM incidence rates across six nations show
consistent increases in CMM for cohorts born between 1910-1930, whereas later
birth cohorts (1940-1950) that show these increases do so to a lesser extent
(Muir and Nectoux 1982).  Most authors that have conducted cohort analyses of
CMM incidence and mortality rates conclude that virtually all secular
increases in CMM are due to cohort effects  (Magnus 1981, 1982; Houghton et al.
1980; Muir and Nectoux 1982; Holman et al.  1980; Cooke et al. 1983).
                              *  DRAFT FINAL  * * *

-------
                                                 EXHIBIT  8-7

                               Increases  in  Incidence  and Mortality  Rates  From
                                  Malignant  Melanoma  in Different  Countries
Fi rst Period
of Observation


1 nc idence
Inc idence
Morta 1
Morta 1
Morta 1
Morta 1
Morta 1
Morta 1
ity
«ty
ity
ity
ity
ity
1 nc idence
Inc idence
Morta 1
ity
Country
New York State
Norway
Norway
Canada
United Kingdom
Austra 1 ia
Denmark
Sweden
Connect (cut
U.S.A.
U.S.A.
Sex
M
F
M
F
M
F
M
F
Both
M
F
M
F
M
F
M
F
WM
WF
WM
WF
T ime
1941-1943
1941-1943
1955
1955
1956-1960
1956-1960
1951-1955
1951-1955
1950
1931-1940
1931-1940
1956-1960
1956-1960
1956-1960
1956-1960
1935-1939
1935-1939
1974
1974
1950
1950
Rate
6
per 10
1.2
1.8
1.8
2.6
1.6
1.3
0.7
0.6
0.5
1.0
0.8
1.6
1.6
1.7
1.1
1 . 1
0.9
6.7
6.0
1.0
0.8
Second Period
of Observation
Time
1967
1967
1970
1970
1966-1970
1966-1970
1966-1970
1966-1970
1967
1961-1970
1961-1970
1966-1969
1966-1969
1966-1968
1966-1968
1975-1979
1975-1979
1983
1983
1977
1977
Rate
6
per 10
3.4
2.9
6.3
6.8
2.7
1.8
1.4
1.2
1.0
3.6
2.5
2.4
2.1
2.1
1.5
8.2
6.8
9.6
8.3
2.6
1.6
Tota 1
Percent
Increase
176
65
264
195
69
36
93
107
100
267
227
49
32
30
40
645
656
43
38
160
200
Number of
Yea rs
25
25
15
15
10
10
15
15
16
30
30
10
10
9
9
40
40
10
10
27
27
Average
Annua I
Percent
Increase*
7.
2.
17.
13.
6.
3.
6.
7.
6.
8.
7.
4.
3.
3.
4.
16.
16.
4.
3.
5.
7.
0
6
6
0
9
6
2
1 I
^ ^J
9
6
9
2
3
4
1
4
3
8
9
4
*  Computed by dividing total  percent  increase  in  incidence  or  mortality  rate  by  the  number  of  years  between
   the first and second periods  of  observations.   The  figures  in  this  column do not  represent compounded
   annual growth rates.

Adapted from:  Elwood and Lee  (1976);  NCI  (198  );  NCI  (1985a);  NCI  (1985b).

-------
                                   8-18
    In most countries, the first signs of increasing rates are seen in cohorts
born around 1900, though increases in cohorts born as early as 1865 are
observed in Australia and New Zealand (Holman et al. 1980; Cooke et al.
1983).  In Norway and several other countries, there is a slight tendency for
a slowing of increase in incidence in cohorts born around and after 1930
(Magnus 1981).  Stabilization of mortality rates is also occurring in cohorts
born 1925-1939 and later in countries such as Australia, New Zealand, England,
Wales, and Finland (Lee and Carter 1970; Teppo 1978; Holman et al. 1980; Cooke
et al. 1983).

    Age-adjusted incidence rates of CMM of the head, neck, and face for birth
cohorts born after approximately 1900 have not differed markedly; however,
age-adjusted incidence rates of CMM of the lower extremities among females and
the trunk among males have increased in successively younger birth cohorts
born during the first half of this century (Magnus 1981; Houghton et al. 1980;
Cooke 1984; Stevens and Mookjaukar 1984).

    On the basis of the Connecticut Tumor Registry data, modeling of the CMM
incidence rates observed in subsequent birth cohorts indicated that the
incidence rate of CMM in the 1955 birth cohort will rival those for colon
cancer, currently the third most common cancer in Connecticut (Rousch et al.
1985a).

    Those anatomical sites that show the greatest amount of increase are not
the sites that receive the greatest amount of cumulative solar exposure.  This
finding tends to lead to a conclusion that cumulative solar exposure is not
causally associated with melanoma; as discussed in more detail below,
cutaneous melanoma may be related to some other measure of exposure to solar
radiation.

Anatomical Site Distribution

    Cutaneous malignant melanoma has a unique anatomical distribution that has
been the subject of numerous epidemiological investigations.  Research efforts
have focused on the site-specific trends in CMM incidence related to sex, age,
race, histogenic type, birth cohort, and season.  The most pronounced
differences have been associated with gender, race, and birth cohort.

    The overall site distribution is presented in Exhibit 8-8.  There is wide
variability among the studies presented in the exhibit, making it difficult to
generalize about these data.  However, as a rule, among white populations, the
upper extremities have the lowest overall proportion of melanomas, while the
lower extremities have the highest proportion.  The variability is most
probably due to differences among the study populations, such as their sex and
racial distribution.

    Gender is one of the factors associated with the most pronounced
differences in CMM site distribution.  Exhibit 8-9 lists data from 14 studies
that reported CMM site distribution by sex.  Data in the table indicate that
most of the studies observed higher incidences of CMM on the lower extremities
among females and on the trunk among males than on other parts of the body.
                                 DRAFT FINAL  * * *

-------
                                              LXIIIBIT 8-8

                       An;iLomic S i le Distribution of Cutaneous Malignant Melanoma
                                      (Percentage or Total  Tumors)
Local i on
a
United States
(caucas ian )
Texas
Texas
Al abama
Now Mexico
New South Wa les
Queens land
New Zea 1 and
(Maori/Polynesian)
I srae 1
Norway
F inl and
Japan
Hong Kong
Uganda
Yea rs
1978-1981
W4-1966
19-3 1,- 1970
1955-1980
1966-1977
1955-1980
1977
1963-1981
1960-1972
1955-1970
1953-19/3
1961-1982
196'l-1982
1963-1966
Samp le S i ze
<4.86H
911
510
537
1)03
1.110
690
2U
966
2,5'41
2.501
5<»6
'13
152
Head/Neck
20
22
25
27
27
in
21
13
16
b
22
19
15
7
8
T runk
35
25
16
28
29
37
3U
13
25
c
»43
37
20
7
f
1 1
Extremi
Upper
23
19
21
19
19
I'l
20
0
21
8
10
13
d
21
5
ties
Lower
22
13
37
23
25
33
2'l
5'»
38
18
26
i|6
e
63
9
72
Source
Scotto 1986
MacDonald 1976
Smith 1976
Ba Ich et a 1 . 1982
Pa thak et a 1 . 1982
Ba Ich et a 1 . 1982
Little et a 1 . 1980
Moss 198»»
Ana i se et a 1 . 1978
Magnus 1973
Teppo et a 1 . 1978
Takuhashi 1983
Col 1 ins 198'l
Kiryabwire et al. 1968
                                                                                                             DO
                                                                                                             I
b
 Based  ofi data  from Seattle,  Detroit, Iowa, Utah, San Francisco/Oakland, Atlanta,  and New Mexico.
>
 (ace.

 Neck/truck.
J
 17  percent  of"  the total  on the hands.
•»
 56  percent  or  the total  on the feet.
r
 "Skift" and  geni ta I s.
J
 Feet and legs.

-------
              EXHIBIT 8-9

Anatomic Site Distribution of Cutaneous
      Malignant Melanoma  by Gender
 (Percentage of total  tumors  by gender
Location
(Ethnic Group) Years

United States 1980

b
United States 1978-1981


North America
(Caucasian )

Europe
(Caucas ian)

Texas (White) 19'4U-1966

Hawaii (Caucasian) 1960-1977

New Mexico (Anglos) 1966-1977

Israel 1960-1972

Scotland 1961-1976

Finland 1953-1973

Denmark 19'l3-1957

Queensland 1977

Texas 195*4-1970

Ext rem i t ies
Sample Size
Predomi nant
't,5'l5



-------
                                           EXHIBIT 8-9 (Continued)
Loca t ion
(Ethnic Group) Years
Extremi t i es
Sample Size Sex
Head/Neck Trunk
Upper
Lower Source
Mixed Study Populations
Israel (Mixed Race) 1961-1967


Japan 1961-1982


Texas (Spanish • 1911-1966
surname )
Texas (Non-White) 1911-1916

New Mexico 1969-1977
( H i span i cs )
Hawa i i 1960-1970
( Non-Caucas ian )

Hawaii 1960-1977
( Non-Caucas ian)


Uganda 1963-1966


368 F
M

516 F

M
206 F
M
30 F
M
35 f
M
66 . F
M

61 F

M

152 F

M
15
17

20

13 .
22
18
6
31
23
23
26
16

26

16

6

10
21
35

19

23
16
22
6
--
36
16
17
19

17

20
h
11
h
11
15
5

18

1 1
18
1 1
18
8
32
23
13
19

13

20

1

5
12 Movshovitz and Modan
25 1973
d
12 Takahashi 1983
e
51
30 MacOonald 1976
30
53 MacDonald 1976
51
9 Pathak et al . 1982
8
11 Hinds and Ko 1 one 1 1980
17
r
11 Hinds 1979
9
11
i
78 Lewis 1967
j
71
                                                                                                                      OO
                                                                                                                      I
 Percentages may not total  100 percent because of rounding errors and exclusion of "other sites" or MM.
b
 Based on data  from Seattle,  Detroit,  Iowa,  Utah,  San Francisco/OakI arid,  Atlanta,  and New Mexico.

 Truck includes scrotum and "unspecified"  melanomas.
d
 29 percent of  the total  were on the  feet.
B
 39 percent of  the total  were on the  feet.
f
 22 percent of  the total  were on the  feet.
3
 12 percent of  the total  were on the  feet.
h
 "Skin" and geni ta I s.

 63 percent of  the total  were on the  feet.
j
 6*1 percent of  the total  were on the  feet.

-------
                                   8-22
Most authors have concluded that the observed differences in site distribution
by gender are not incompatible with the role of sunlight as a major etiologic
factor (Pathak et al. 1982, Crombie 1981; Hinds and Kolonel 1980).

    Race and gender also are associated with pronounced effects in site
distribution of CMM.  The predominant difference is a higher percentage of CMM
occurring on the feet and, in some cases, on the hands of darker skinned
ethnic groups.  These groups have a much lower incidence than the light-
skinned ethnic groups, so that the higher percentages on feet may not
represent a higher incidence.  Trauma has been suggested as a possible reason
for the increased percentage of melanomas occurring on feet (Lewis 1967; Hinds
1979).  Although in a recent review of the subject, Briggs (1984) concluded
that there is no unequivocal evidence for a role of trauma in the vast
majority of melanomas.  One group has concluded that sunlight is not an
important risk factor for melanoma (Hinds 1979; Hinds and Kolonel 1980) for
any site among non-Caucasians.

    There appear to be site-dependent differences in melanoma incidence by
age, with melanomas of the face showing one pattern and melanoma of trunks and
lower extremities, another.  In several instances the differences have been
identified as being between continuously exposed sites (e.g., head) and
intermittently exposed sites (e.g., trunk and lower extremities).  As a
general rule the pattern observed for the exposed sites is characteristic of
that observed for many other cancers --a slow increase with age, up to the
age of 40 or 50, followed by a rapid increase with age thereafter (Magnus  .
1981; Teppo 1982; Holman et al. 1980).  For the intermittently-exposed sites,
the pattern is quite different with a relatively steep increase which peaks
between 40 and 60 (depending on the data set) and levels off or decreases
thereafter (Magnus 1981a,b; Teppo 1982; Holman et al. 1980).

    There are also differences in the anatomic distribution of melanomas
associated with different histologic (histogenetic) types of CMM.  HMFM and
its precursor lesion, HMF, are observed predominantly on the face and other
exposed sites, whereas SSM was most commonly observed on less exposed sites
(trunk and lower extremities) (Holman et al. 1980, Smith 1976, Adler and Gaeta
1979, Pondes et al. 1981).

    Very few studies have examined the relationship of geographic area to
differences in a anatomic site distribution.  In Norway, Magnus  (1973) noted
that the only site lacking a definite north/south gradient was the foot, and a
comparison of Alabama patients to Australian patients noted that whereas
melanomas of the trunk were the most common in both sets of patients,
melanomas of the lower extremities were more common in Australians than in the
Alabama patients, and head and neck tumors showed the reverse trend (Balch et
al. 1982).

    A final important parameter that has been associated with differences in
site distribution of CMM is the temporal pattern of exposure.  This is a
complex parameter that is, as indicated earlier, difficult to separate from
the observed age effects, but it is fairly clear that there are differences in
the anatomic distribution of melanomas, which are associated with the year of
                          * * *  DRAFT FINAL  * *

-------
                                   8-23
diagnosis and/or birth cohort.  It is difficult to generalize, however, the
conclusion drawn by Boyle et al. (1983) that there were cohort effects of
melanoma incidence that differed by site and that the site effects differed by
sex.  The overall impression from the studies that looked at these cohort
effects is, as well, that birth cohort-site differences are greater for sites
that are intermittently exposed (e.g., trunk and lower extremities) than for
those that are always exposed (e.g., face).

Exposure Factors

    A number of studies have examined how various indicators or measures of
exposure are related to melanoma incidence or mortality.  This section reviews
information from ecological studies that have looked at the correlation of
melanoma incidence or mortality on the one hand, and geographical location and
migration on the other.  Studies that have examined the relationship between
type of exposure (e.g., early, intermittent) and melanoma incidence are also
discussed.

    Geographic Location

    Within predominantly Caucasian nations, most of the ecological studies of
melanoma and latitude show increasing melanoma incidence and/or melanoma
mortality with decreasing latitude, leading to the hypothesis that melanoma is
associated with ultraviolet radiation because of the strong correlation
between UV and latitude.  As a general rule, those studies that failed to find
this association did not adequately account for pigmentation differences
(Crombie 1979) or had other serious methodological flaws (poor measurement of
UVB) (Baker-Blocker 1980).

    In general, melanoma rates (incidence and/or mortality) were found to be
higher in areas closer to the equator, in coastal rather than inland areas,
and in urban rather than rural areas within various nations (Holman and
Armstrong 1984; Green and Siskind 1983; Magnus 1981a, 1981b).

    Migrant Behavior

    In an Australian study (Holman and Armstrong 1984), age at arrival in
Australia was more important than duration of residence with respect to the
risk of SSM, with arrival before age 10 incurring a risk approaching or
exceeding the risk of SSM for those born in Australia.  The odds ratio
decreased when the age at arrival was 10-14 years and stabilized at a
significantly lower level for those arriving at or after the age of 15.

    Immigrants moving to sunnier climates in which the native CMM incidence
rates exceed those of the immigrant's country of origin initially tend to have
lower CMM risks than the native population.  These risks increase, however,
with increasing duration of residence or earlier age of arrival in the adopted
homeland (Holman and Armstrong 1984).
                          * * *  DRAFT FINAL  * *

-------
                                   8-24
    Type of Sun Exposure

    Early Exposures.  CMM risk is associated with childhood sunburn (Green et
al. 1985, Elwood et al. 1985b) but this association appears only to reflect an
individual's pigmentary characteristics as related to poor sun tolerance.  One
study that evaluated the risk associated with outdoor work before college
found a significantly elevated odds ratio for those who had worked outdoors
compared with those who had not (Paffenbarger et al. 1978).

    Intermittent ("Summer Sun") Exposure.  One commonly examined surrogate
for intermittent exposure has been a parameter related to a history of sunny
vacations (Eklund and Malec 1978; Klepp and Magnus 1979; Adam et al. 1981).
Different studies have used slightly different measures of this type of
exposure behavior (e.g., one study evaluated those who "tanned themselves
while on holiday abroad" (Adam et al. 1981), another evaluated those with a
history of long (30 days or more), sunny vacations as a child (Lew et al.
1983).  In general, however, there was a positive association between CMM risk
and increasing number of "sunny vactions" taken.

    The strongest evidence for an association between a measure of sunny
vacation exposure comes from a case-control study conducted in Western Canada
(Eldwood et al. 1985a, 1985b).  Analysis of data from this study of 595
melanoma case-control pairs indicated that there was a significant (p^O.OOl)
trend of increasing risk of CMM which increasing numbers of sunny vacations
per decade, with a relative risk of 1.8 associated with four or more sunny
vacations per decade.  In addition, this relative risk remained significant
(RR = 1.7, 95% C.I.:1.2-2.3) even after controlling for significant host
factors (hair color, skin color, history of freckles) and ethnic origin.

    At least one study that evaluated a different measure of intermittent
exposure found no significant association between intermittent exposure and an
increased incidence of any histologic type of CMM (Holman and Armstrong
1984).  This study of 507 pairs (matched according to sex, age, and residence)
used as its surrogate for intermittent exposure a variable based on the ratio
of recreational outdoor sun exposure in summer to total outdoor sun exposure
in summer.

    In an additional report of the same study (Holman et al. 1986), an
increased SSM risk for some summer sun activities at early ages (from history
or estimated) was observed for boating, fishing, and female sunbathing  (on
trunk).  This study also found some increased risks for summer sun exposure by
clothing habits.  The primary CMM site was more frequently "sometimes exposed
to sun" than usually exposed or usually covered, except for SSM, which showed
increasing risk with increasing exposure.  The risk of developing CMM of the
trunk (especially SSM) was found to be greater for women whose bathing suits
covered less of their bodies between 15 and 24 years of age.

    A seasonal pattern with a summertime peak was found for CMM incidence
rates (female in U.S. (Scotto and Nam 1980), both sexes in Hawaii (Hinds et
al. 1981)).  This seasonal pattern  may be related to greater awareness of
skin changes in summer months or could indicate promotional effects of UVB
exposures.
                            * *  DRAFT FINAL  *

-------
                                   8-25
    Sunburn.   A number of case-control studies that have examined the
relationship of CMM to history of sunburn have found statistically significant
associations.  Most of these studies did not, however, adjust for pigmentary
factors.  One study that controlled for pigmentary factors found a significant
association (Lew et al. 1983); however, in this study, biased selection of
controls was likely, leaving this association in doubt.  The conclusion drawn
from studies that appeared to be subject to the fewest biases was that the
presence or absence of factors associated with a tendency to sunburn, rather
than the sunburn itself, determine the CMM risk.

    A study from Western Australia (Holman et al. 1980) found significantly
increased total CMM risk, SSM risk, and HMF risk with increased annual hours
of bright sunlight at residence, with increasing actinic skin damage, and with
previous nonmelanoma skin cancer.

    Assessment of individual exposure through questionnaires.  In Western
Canada, an analysis of total sun exposure showed some increased risks in
higher exposure groups compared with the lowest exposure group but none were
statistically significant nor was there a significant trend of increasing risk
(Holman et al. 1986).

    In Queensland, Australia, elevated CMM risks were associated with
increasing estimated total hours of sun exposure at ages 10 years and older
after adjustment for exact age, presence of nevi on arms, hair color, and
sunburn propensity.  The confidence intervals for intermediate and higher
levels of exposure include unity.

    Indirect measures of cumulative sun exposure (i.e., assessment of numbers
of hours exposed) show correlations with melanoma risk that were significant,
borderline significant, and not significant.  Several studies have evaluated a
direct measure of received dose (i.e., actinic skin damage), which is also a
measure of the skin's responsiveness to insult.  In these studies, melanoma
risk increased significantly with the degree of actinic skin damage.

    Exposure During Sunspot Activity.  Most studies that examined the
relationship between CMM incidence rates and sunspot cycles found high
correlations (cite studies).  Different studies have observed differences in
the lag time between the period of sunspot activity and changes in CMM
incidence rates.

    Most of the information presented above tends to support a relationship
between melanoma and some function of increased exposure to solar radiation.
What the exact function of sunlight exposure may be is unclear.  There is some
indication that early  (childhood) exposure may be important, and other
information implicates intensive exposure such as that which occurs during
summertime and times of sunspot activity; however, no single kind of exposure
clearly stands out.  None of these studies considered UVB, which varies
significantly with time of day.  Thus, by using sun exposed hours as a dose
variable, such studies may be inaccurately assessing dose, especially when
they compare across occupations whose sun-exposure behavior may vary
significantly.
                          * * *  DRAFT FINAL

-------
                                   8-26
Host Factors

    There is clear evidence that certain host factors are associated with the
risk of developing CMM.  It is possible that the factors described below
reflect the capability to modify the dose of solar radiation potentially
received by a target cell or modify the ability of the cell to respond to
insult.

    Epidemiologic studies have shown that Caucasian populations have much
higher rates of CMM incidence and mortality than other racial groups.  In
South Africa, CMM was approximately six times more frequent in whites than in
blacks.  Broken down by sex, the ratio of white to black CMM incidence was
greater for males (13:1) than for females (4:1).  Black/white differences in
the United States population were even greater, although this may be due to
differences in the quality of cancer registry data between South Africa and
the United States.  Based on 1983 SEER data, white:black ratios were 19:1 and
8:1 in males and females, respectively.  In addition, whites in New Zealand
experience much higher incidence rates than New Zealand Maoris and
Polynesians.  Likewise, American Indians experience much lower, rates of CMM
than American whites.

    Within white populations, rates of CMM differ according to country of
origin.  CMM incidence rates for Hispanic whites in New Mexico, for example,
are much lower than those for non-Hispanic whites; individuals from the
Mediterranean countries in southern Europe tend to have lower rates than
Caucasians from northern Europe; individuals of Celtic origin in Australia
tend to have higher rates than non-Celtic individuals.  Variation in the
incidence of CMM within the Caucasian race is commonly thought to be a
function of variation in genetically determined pigmentary traits across
ethnic groups.

    Numerous epidemiologic studies have focused on identification of important
pigmentary characteristics in the etiology of CMM.  Exhibit 8-10 shows a
summary of findings from a number of these studies.  The following
associations between pigmentary traits and CMM risk have been found:

        (a)  Skin color -- in all studies reviewed fair complexions
             were associated with higher risks of CMM than were dark
             complexions.

        (b)  Hair and eye color --in most studies red and blonde
             hair in childhood were associated with increased risk
             of CMM relative to dark hair.  Blue eyes were an
             independent risk factor in only one of four
             well-controlled epidemiologic studies; however, this
             could be due to the homogeneous nature of most of the
             study populations.

        (c)  Freckling -- those who freckled readily were at
             consistently elevated CMM risk relative to other
             individuals.
                          * * *  DRAFT FINAL  * *

-------
                                   8-27
                         EXHIBIT 8-10

          Malignant Melanoma Risk  Factors  by Measures of
        Skin Pigmentation Within the Caucasian Population
     Study Reference
                                       Measures  of Skin Pigmentation
                                        Within Caucasian Population
                                                            Reaction
                                                             to Sun
                            Skin   Hair    Eye              (Tanning/
                            Color  Color  Color  Freckling  Sunburn)   Ethnicity
Lancaster and Nelson 1957

Gellin et al. 1969

Lane Brown et al. 1971

IARC 1976

MacDonald 1976

Klepp and Magnus 1979

Mackie and Atchinson 1982

Beral et al. 1983

Hinds and Kolonel 1982

Lew et al. 1983

Elwood et al. 1984

Holman and Armstrong 1984b

Graham et al. 1985
                              +

                              +

                              +
                                     +

                                     -f
+

+

+
+

+
+

+
NOTE:      + = Significant risk factor.
           - = Not significant risk factor.
       Blank = Not included in study.
                          * * *  DRAFT FINAL  * * *

-------
                                   8-28
         (d)  Reaction to sun exposure -- in most studies reviewed,
             individuals who usually burned and were unable to tan
             were at significantly higher risk of CMM than those who
             tanned easily.

Exposure to sunlight appears to encourage the appearance (and possibly the
disappearance) of nevi  (including dysplastic nevi) from the skin  (Kopf et al.
1978,  1985; Armstrong et al. 1986).  Dysplastic nevi are clearly a risk factor
for melanoma independent of freckling or other pigmentary characteristics; it
also appears that congenital nevi as well as acquired melanocytic nevi may
also be  risk factors (Elder et al. 1981).

    Information  from xeroderma pigmentosum (XP) patients indicates that
individuals who  have an inability to repair solar radiation-induced DNA damage
also have  a high incidence of CMM relative to the normal population (Kraemer
et al. 1984).  The best characterized defect in XP patients is an inability to
excise pyrimidine dimers (Cleaver 1983) which suggests that the repair of such
lesions  can be important to prevention of CMM development.

Miscellaneous Factors

    Although a number of possible etiologic agents for melanoma have been
evaluated  (e.g., exposure to chemicals or ionizing radiation), no strong
candidate  has emerged.  It has been suggested that the risk of developing CMM
may be elevated  among individuals exposed to fluorescent lighting at work;
however, several recent well-controlled studies have failed to find such an
association.  Indeed, one of these studies initially found a positive
association based on personal interviews, but a self-administered postal
survey of  the same study subjects did not confirm the initial findings.

    Total  CMM incidence has been found to be higher among professional and
administrative indoor office workers, but not among other indoor workers
relative to outdoor workers; however, the implications of this finding may be
confounded by differences in socio-economic status, melanoma site, and
histologic type.  There is a higher risk of indoor office workers developing
CMM on an  anatomical site that is usually covered  (e.g., the trunk) than on a
site that  is usually exposed (e.g., the face).  For usually covered parts of
the body,  the incidence of CMM among indoor office workers is higher than for
outdoor  workers.  And the reverse is also true; for usually uncovered parts of
the body,  the incidence of CMM is higher among outdoor workers than among
indoor office workers.

    It is  not clear why certain indoor workers are at greater risk for
developing CMM than outdoor workers, nor why in these individuals the affected
sites  are  those  not normally exposed.  One speculation is that indoor workers
have more  leisure time  and that they spend that time under conditions of solar
exposure that maximize.intensive UVB exposures of areas that are  not protected
by a tan;  however, there is no information currently available from a
•published  study  that would allow this hypothesis to be examined.  From current
data it  is clear that for certain indoor workers, cumulative sun  hours of
                           *  *  *   DRAFT  FINAL  *  *  *

-------
                                   8-29
exposure that is not associated with melanoma risk.  This information does not
rule out that some other measure of exposure (e.g., number of high UVB
exposures received at a time of low tan) may be associated with melanoma risk.

Weighing and Balancing the Epidemiologic  Evidence

    In summary, while there is no single piece of evidence which proves that
solar radiation causes CMM, the weight of the available epidemiologic evidence
clearly supports the conclusion that solar radiation plays a major role in the
development of CMM for susceptible populations.  Cumulative dose (i.e., total
sun exposed hours) clearly does not explain important variations in incidence
rates.  Although, from existing evidence it is not possible to determine on
the most accurate method of calculating dose, the collective evidence does
suggest that exposure to solar radiation in some form produces cutaneous
malignant melanomas in Caucasian populations.

EXPERIMENTAL EVIDENCE

    Much of the experimental evidence relevant to the relationship of CMM and
exposure to UVB has been reviewed on non-melanoma skin cancer.  To briefly
recapitulate, there are five important points:

        (1)  UVB is the most active portion of solar radiation
             responsible for the induction of adverse effects on
             mammalian systems; it has been shown in vivo and in
             vitro to induce transformation of mammalian epidermal
             cells and to be mutagenic to them as well.  All of
             these effects are thought to occur by a mechanism that
             involves damage to the DNA (Doniger et al. .1981)

        (2)  UVB at 295-300 nm is the most active waveband of UVR
             for these effects, as well as for the induction of
             pyrimidine dimers.

        (3)  Patients with xeroderma pigmentosum, who cannot repair
             pyrimidine dimers induced in DNA by UVB (which are
             thought to be important to skin cancer development)
             have significantly higher rates of CMM than normal
             individuals.

        (4)  UVR is carcinogenic in animals and it is the UVB
             wavelengths that  are the most effective at inducing
             cancer -- the shorter UVB wavelenths being more
             carcinogenic than the longer ones (Forbes et al. 1981);
             the tumors induced are principally fibrosarcomas and
             squamous cell carcinomas.

        (5)  UVR, and specifically UVB, can induce the generation of
             T suppressor cells, which specifically suppress the
             immune response to UV-induced tumors.  The suppressor
                          * * *  DRAFT FINAL  * *

-------
                                   8-30
             cells have been shown to be capable of shortening the
             latest period of tumors induced by UVB, and UVB-treated
             animals also appear to have increased susceptibility to
             transplanted melanomas.

Weighing and Balancing the Experimental Evidence

    The experimental data indicate that UVB irradiation damages DNA and
induces carcinogenesis and immunosuppression in animals.  CMM rates are
elevated in patients who cannot repair UVR induced lesions in DNA.

DOSE-RESPONSE  RELATIONSHIPS

    Dose-response relationships between exposure to UVB and the incidence and
mortality of CMM have been developed in several epidemiologic studies.

    Animal studies are inadequate for estimating CMM dose-response because
there is no animal model for melanoma induced by UVB.   Because the limitations
in animal studies appear insurmountable at this time,  epidemiologic studies
provide the best candidate dose-response relationships for risk estimation.
This section reviews four such studies.  Three of the studies -- Fears,
Scotto, and Schneiderman (1976); Fears, Scotto, and Schneiderman (1977);  and
Scotto and Fears (1986) -- develop dose-response relationships associating UVB
and melanoma incidence.  Two studies -- Fears, Scotto, and Schneiderman (1976)
and Pitcher (1986) -- estimate dose-response relationships for melanoma
mortality.

    Fears, Scotto, and Schneiderman (1976) present one epidemiologic analysis
that is potentially useful for estimating dose-response.  These authors use
two types of data to represent UVB:  (1) latitude -- latitude and UVB
radiation weighted for erethyma effectiveness correlate at 0.97, and (2)
monthly totals of erethyma-producing UV radiation expressed as Biologically
Effective Units (BEU) and derived from Schultze (1974).  Both types of data
were correlated with CMM mortality and incidence data.  Incidence information
for four cities was obtained from the Third National Cancer Survey (TNCS)
(1975), and mortality data were from the U.S. Cancer Mortality by County
database (Mason and McKay (1973).

    Exhibit 8-11 shows the results of a simple correlation model based on
latitude:

                              log R = a + PL

where R is the age-adjusted rate of incidence or mortality, a and & are
constants, and L is latitude.  The estimated coefficients were statistically
significant at p>0.01.  Exhibit 8-12 shows the results estimated using
BEU's.  These estimates were based on an exponential model and represent the
percentage changes in incidence and mortality estimated to occur with
increases in UV radiation dose of between 10 and 30 percent.  Note that the
use of an exponential model results in higher dose-response relationships for
higher base exposures.
                            * *  DRAFT FINAL  *

-------
                                         8-31
                                    EXHIBIT 8-11

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

Correlation
Coefficient*
MALES
Regression
Slope+S.D.

Doubling
Latitude

Correlation
Coef ficient*
FEMALES
Regression
Slope+S.D.

Doubling
Latitude
Melanoma       -.86
  Incidence

Melanoma       -.81
  Mortality
-.031 + .007   - 9.8e
-.017 + .002   -19.9'
-.83       -.038 + .007   -10.7'
-.71       -.014 4- .002   -22.2C
•''Simple correlation coefficient between log of incidence/mortality rate and latitude.

Source:  Fears, Scotto, and Schneiderman (1976).
                                * * •'•  DRAFT FINAL  * * *

-------
                                                 EXHIBIT 8-12

                      Estimated  Relative  Increases in Melanoma  Skin Cancer Incidence  and
                           Mortality Associated  with Changes  in Ereythema  Dose ay
                                    Fears,  Scotto,  and Schneiderman (1976)
                         (Figures  in Parentheses are 95 Percent Confidence Intervals)



Me lanoma
1 nc idence



Me 1 anoma
Morta 1 ity




BASE
BEU


650
850
1050


650
850
1050


10%


15 (7-24)
20
25


8 (6-10)
10
13
MALES
Increase in Total Dose
20%


32
44 (18-75)
57


16
22 (16-28)
28
FEMALES

30%


52
72
96 b/ (37-180)


26
35
45 (33-58)
Increase
10%


13
18
22


6
9
11
in Tota I
20%


29
39
50


13
18
22
Dose
30%


46
64
84 b/
•

21
28
36









00
1
UJ
KJ
ay A sample computation is as  follows:
     A 10% increase in total dose  at  48.25'N,  where  exposure  is  650  BEU,  equals  65  BEU.
     A change of 65 BEU is equivalent to  a  reduction in  latitude of  1.93'.
     A change of 1.93'  at  48.25'N  is  associated with an  18% increase in  nonmelanoma incidence.
b/ These estimates require extrapolation well  beyond  the  range  of  the  data.


Source:  Fears,  Scotto,  and Schneiderman (1976).

-------
                                   8-33
    There are several important limitations in this work.  In particular, only
four cities were used for melanoma incidence.  With a sample this small,
variations in confounding factors such as behavior, skin pigmentation, and
cloud cover may bias dose-response estimates.  As a result, caution must be
used in evaluating the error of estimate.

    In another study, the same authors estimated a dose-response relationship
for melanoma incidence using a power model (Fears, Scotto, and Schneiderman
1977):
                    P.. = b (U,) (A.)
where:
    P.. = probability of developing melanoma for jth age group at

          location i;

    U.  = annual UVB count at ith location;

    A.  = midrange of jth age group; and

    b, c, and k are constants.

This equation was estimated by fitting a log form of the model (In R.. = a +

c In U. + kLnA.  + E..) to age-specific melanoma incidence rates for both

males and females after weighting by the observed number of cases.  Annual
counts from Robertson-Berger  (RB) meters were used for UVB; the RB meter
attaches greater weight to longer wavelengths compared to the erythema action
spectrum.  Incidence data also were from the TNCS (1975).

    Exhibit 8-13 presents summary regression coefficients and statistics for
this age-specific analysis.  Applying the results in this exhibit, a one
                                                                    £
percent increase in UV radiation results in an estimated 100"'((1.01) -1)
percentage increase in melanoma incidence, or 2.47 for females and 2.24 for
males.  The authors stress, however, that these estimates may be biased by the
omission of location-specific demographic and environmental variables.

    Scotto and Fears (1986) estimate a dose-response relationship for melanoma
incidence using updated information about populations in a greater number of
cities.  The authors also introduce a new term (VAR) to adjust for the
presence of some host or environmental characteristics:

            Ln R..  = a + b In (Age.) + c In UVB + dVAR + e
                            * *  DRAFT FINAL  * •- *

-------
                                        8-34
                                   EXHIBIT 8-13

                 Summary of Fears, Scotto,  and Schneiderman  (1977)
              Regression Analyses  of Melanoma  Incidence Dose-Response-'
                       t-Statistic       a + SD          K + SD            c + SD
 SEX    R-Squared    (UV Coefficient)   (constant)   (age coefficient)  (UV coefficient)



Males         0.74            5-4      -35.6 +-6.5       .80 + .31         2.45 + .45

Females       0.62            4.6      -30.5 + 6.9       .29 + .34         2.23 + .48



•'•- In Ry = a + c In U.  + < In A.  + sy

Source:  Fears, Scotto, and Schneiderman (1977).
                                       DRAFT  FINAL

-------
                                   8-35
Annual counts obtained from Robertson-Berger meters were used for UV-B.  Seven
areas were included:  Detroit, Seattle, Iowa, Utah, San Francisco, Atlanta,
and New Mexico.  Separate estimates were provided for:  (1) anatomical sites
of melanoma -- FHN (face, hands, neck), UE (upper extremity), LE (truck and
lower extremity); and (2) age groups -- 20-39, 40-54, 55-64, 65-74.

    Scotto and Fears (1986) found that after controlling for confounding
variables each 1 percent increase in UVB causes melanoma incidence to increase
by less than 1 percent.   Exhibits 8-14 and 8-15 show results for different
anatomical sites.  The results are statistically significant (p<0.1).
Exhibit 8-12 describes the effects of adding different constitutional and
environmental variables  derived using step-wise multiple regression.  If the
DNA action spectrum is used, Scotto's results indicate slightly less than a 2
percent increase in morbidity for each 1 percent ozone depletion.

    Pitcher (1986) related data on melanoma deaths from the EPA/NCI data base
(U.S. EPA, 1983) to exposure data obtained from a model developed by the
National Atmospheric and Space Administration (NASA).  The NASA model provided
estimates of UV-B radiation for alternative locations measured at different
times and using various  biological action spectra.

    Pitcher fit two different models for each sex.  Defining DRM..  as the

death rate for the ith cohort, the jth SMA (location), and the kth time
period, he estimated:
            DRM.
. ..  = exp(b.  + b.AGE..  + b.EXP.)  +
ijk     r  0     i   ik    j    j
where AGE.,  is the age of ith cohort in the kth time period and EXP. is
         IK                                                        J
the exposure in the jth SMA.  In this model, the percentage change in the
death rate that results from a given percentage change in exposure depends on
the baseline exposure.  The percentage change in death rates is higher the
greater the baseline exposure.

    The second model differed from the first only by using the log of the
exposure variable.  In this model, a one percent increase in exposure
generates approximately the same percentage increase in the mortality rate
regardless of the baseline mortality.

    Using UVB radiation data from the NASA model for a clear day in June and
measured with the DNA action spectrum, Pitcher estimated constants on the
exposure term in the second formulation to be 0.85 for males (standard error
0.067) and 0.58 for females (standard error 0.078).  The clear'day values of
UV radiation can be considered peak estimates.  With these constants, a 1
percent increase in UV radiation results in an estimated increase in melanoma
mortality of about 0.85 percent for males and about 0.58 percent for females.
Exhibit 8-16 extends the estimates to larger relative changes  in UVB.  Since
the optical amplification factor for DNA is 2 percent increase in UVB for a 1
percent ozone depletion, the Pitcher data indicates that there would be a 1.70
increase for males and a 1.16 for females in melanoma mortality for a 1
percent depletion.


                          * * *  DRAFT FINAL  * * »

-------
                                   8-36
                                EXHIBIT 8-14

             Biological Amplification Factors* for  Skin Melanoma
              by Sex  and Anatomical Site Groups, Adjusting for
                     age and Selected Constitutional and
                             Exposure Variables
    VARIABLE
                                           MALE
                             FEMALE
Trunk/LE    FHN/UE    Trunk/LE
                      FHN/UE
Age
  6%
8%
5%
10%
Sunburn
Freckles
Scandinavian ancestry
Lt . hair color
Scot/Irish ancestry
Moles
Lt . eye color
Fair skin
Sunscreen use
Suntan lotion use
Radiation protection
Protective clothes
Hours outdoors during weekdays
Hours outdoors during weekends
4
5
5
5
6
6
6
6
4
5
5
8
7
5
8
7
9
8
6
7
8
8
7
8
8
10
9
7
5
5
5
5
4
5
4
4
4
4
6
5
5
8
10
10
10
10
8
11
9
9
10
10
10
10
10
13
"A biological amplification factor indicates the relative change  in  melanoma
incidence associated with a 10 percent relative increase in UVB.

Source:  Scotto and Fears (1986).
                          * * *  DRAFT FINAL  *

-------
                                   8-37
                                EXHIBIT 8-15

          Biological Amplification Factors* for Melanoma Incidence
              by Sex and Anatomical Site  Groups, Adjusting for
               Age and Combinations of Selected Constitutional
                           and Exposure Variables
ANATOMICAL
   SITE
            MALE
            FEMALE
Trunk/LE

  Variables
  included
  in model
FHN/UE

  Variables
  included
  in model
          3 Percent

Suntan lotion use
Scandinavian
Lt hair color
UVB radiation index
Hours outdoors during weekdays

          5 Percent

Scot/Irish
Suntan lotion use
Fair skin
UVB radiation index
Hours outdoors during weekends
           4 Percent

Suntan lotion use
Scandinavian
Lt hair color
UVB radiation index
Hours outdoors during weekends

        6 Percent

Scot/Irish
suntan lotion use
fair skin
UVB radiation index
*A biological amplification factor is a number that indicates the percentage
increase in melanoma incidence for a 10 percent relative increase in UVB.

Source:  Scotto and Fears (1986).
                          * * *  DRAFT FINAL

-------
                    8-38
               EXHIBIT 8-16

Estimated  Percentage Change in  Melanoma Mortality
     for Different  Percentage Changes in UVB
 (DNA Action  Spectrum for  a Clear Day  in June)
PERCENTAGE
INCREASE
IN UVB
1
2
5
10
20


MALES
0
1
4
8
17
.85
.7 .
.2
.4
.0


FEMALES
0
1
2
5
11
.58
.2 .
.9
.7
.0
     Source:  Derived from results of Pitcher
             (1986).
                  DRAFT FINAL  * *

-------
                                   8-39
    In the analyses done by Pitcher, the question was raised regarding which
measure of UVB dose would be most appropriate for estimating mortality risks
in the event of ozone depletion.  The regression analysis showed a more
significant statistical relationship using the DNA action spectrum for peak
values (June 15th) as the exposure variable, as compared to using other action
spectra (e.g., erythema and RB) and/or other time periods (e.g., annual
cumulative values or cumulative values for June, which would be another way to
measure peak values).  It must be emphasized, however, that the statistical
differences due to the different exposure measures were extremely small.
Therefore, the decision regarding which dose measure to use is primarily
judgmental, and should be based inter alia on hypotheses about disease
etiology.   While peak values were selected as the central case for analyses in
Chapter 18, the question exists as to whether this approach overestimates the
possible risks of ozone depletion.  Using peak UVB values as the exposure
measure may overstate these risks because the variability of peak values
across locations is less than the variability of UVB measured over different
time periods.  The lower variability of the peak values results in larger
estimates of the responsiveness of melanoma mortality to changes in UVB for
this exposure measure.  Consequently, the appropriate action spectrum for
estimating dose-response for melanoma mortality is an area requiring future
research and sensitivity analyses should be included in assessments of
melanoma mortality.

Limitations of Epidemiologic Studies

    A variety of problems exist in applying the results of the epidemiological
studies to an analysis of melanoma risks associated with ozone depletion; many
of these problems have been described in earlier sections.  In particular,
applying the dose-response relationships requires an assumption that certain
key factors will not change.  However, the proportion of individuals in the
population who are most susceptible to UV radiation, such as certain ethnic
groups, may change over time.  Migration was not considered in the studies; if
southern states experience large increases in population then the sensitivity
of future populations to UVB will be underestimated.  Individuals might also
change their use of sunscreens or their propensity to work outside work habits.

    The epidemiologic studies do not measure actual ambient UVB dose.
Therefore, UVB data used to estimate dose-response relationships will differ
from the actual potential ambient exposure of individuals at any location.
Other factors will also affect actual ambient exposures, such as skin color,
patterns of dress, work exposure, recreation, eating, and medical care and
intervention.  These factors can be expected, in varying degrees, to introduce
unexplained variation (i.e., noise) into any analysis of incidence or
mortality.  Finally, few data are available to assess the significance of
genetic factors that effect melanoma incidence but are not captured in
epidemiologic studies.
                          * * *  DRAFT FINAL  * * *

-------
                                   8-40
REFERENCES

Acquavella, J.F., Wilkinson, G.S., Tietjen, G.L., Key,  C.R.,  Stebbings,  J.H.,
and Voelz, G.L.  A melanoma case-control study at the Los Alamos National
Laboratory.  Health Physics 45:587-592 (1983).

Adam, S.A., Sheaves, J.K., Wright, N.H., Mosser, G., Harris,  R.W.,  and Vessey,
M.P.  A case-control study of the possible association between oral
contraceptives and malignant melanoma.  Brit J Cancer 44:45-50 (1981).

Anaise, D., Steinitz, R., and Ben Hur, N.  Solar Radiation:   A possible
etiological factor in malignant melanoma in Israel; A retrospective study
(1960-1972).  Cancer 42:299-304 (1978).

Armstrong, B.K., Holman, C.D.J., Ford, J.M., and Wooding, T.L.  Trends in
melanoma incidence and mortality in Australia.  In:  Trends  in Cancer
Incidence, Causes and Practical Implications, Magnus, K., ed., Hemisphere
Publications (1981).

Armstrong B.K.  Melanoma of the skin.  Brit Med Bull 40:346-350 (1984).

Austin, D.F., Reynolds, P.J., Snyder, M.A., Biggs, M.W.,  and Stubbs, H.A.
Malignant melanoma among employees of Lawrence Livermore  National Laboratory.
Lancet 2:712-716 (1981).

Bahr, A.K., Rosewaike, I., Hermann, N., Grover, P., Stellman,  J., O'Leary,  K.
Letter to the editor:  Melanoma after exposures to PCBs.   New Eng J Med,
August 9, p. 450 (1976).

Baker-Blocker, A.  Ultraviolet radiation and melanoma mortality in the United
States.  Env Res 23:24-28 (1980).

Balch, C.M., Soong, S., Milton, G.W., et al.  Changing trends in cutaneous
melanoma over a quarter century in Alabama, USA, and New  South Wales,
Australia.  Cancer 52: 1748-1753 (1983).

Balch, C.M., Soong, S., Milton, G.W., Shaw, H.M., McGovern,  V.J., McCarthy,
W.H., Murad, T.M.,and Maddox, W.A.  A comparison of prognostic factors and
surgical results in 1,786 patients with localized  (Stage  I)  melanoma trated.in
Alabama, USA, and New South Wales, Australia.  Ann Surg 6:677-684 (1982).

Balch, C.M., Karakousis, C., Mettlin, C., Natarajan, N.,  Donegan, W.L.,  Smart,
C.R., and Murphy, G.P.  Management of cutaneous melanoma  in the United States.
Surg Gyn Obstetrics 158:311-318 (1984).

Beral, V., Ramcharan, S., and Faris, R.  Malignant melanoma and oral
contraceptive use among women in California.  Brit. J.  Cancer 36:804-809
(1977).
                          -v * *  DRAFT FINAL  * * *

-------
                                   8-41
Beral, V., and Robinson, N.  The relationship of malignant melanoma,  basal and
squamous skin cancers to indoor and outdoor work.  Br.  J.  Cancer 44:886-891
(1981).

Beral, V., Evans, S., Shaw, H., and Milton, G.  Cutaneous  factors related to
the risk of malignant melanoma.  Brit J Dermatol 109:165-172 (1983).

Boyle, P., Day, N.E., and Magnus, K.  Mathematical modeling of malignant
melanoma trends in Norway, 1953-1978.  Am J Epidemiology 118:887-896  (1983).

Briele, H.A., and Das Gupta, T.K.  Natural history of cutaneous malignant
melanoma.  Wld J Surg 3:255-270 (1979).

Briggs, J.C.  The role of trauma in the aetiology of malignant melanoma:  A
review article.  Brit J Plastic Surg 37:517-516 (1984).

Cleaver, J.E.  Xeroderma pigmentosum.  Chapter 57 in Stanbury, J.B.,
Wyngarten, J.B., Frederickson, D.S., Goldsteen, J.L. and Brown, M.S.  (eds.)
The Metabolic Basis of Disease, McGraw-Hill Book Company,  New York, pp.
1227-1248 (1983).

Collins, R.J. Melanoma in the Chinese of Hong Kong: Emphasis on volar and
subungual sites.  Cancer 54:1482-1488  (1984).

Cook, P.J., Doll, R,, Fellingham, S.A.  A mathematical model for the  age
distribution of cancer in man. Int J Cancer 4:93-112 (1969).

Cooke, K.R., Skegg, D.C.G., and Fraser, J.  Trends in malignant melanoma of
skin in New Zealand.  Int J Cancer 31:715-718 (1983).

Cooke, K.R., Skegg, D.C.G., and Fraser, J.  Socio-economic status, indoor and
outdoor work and malignant melanoma.   Int J Cancer 34:57-62 (1984).

Cooke, K.R., and Fraser, J.  Migration and death from malignant melanoma.  Int
J Cancer 36:175-178  (1985).

Crombie, I.K.  Variation of melanoma incidence with latitude in North America
and Europe.  Brit J Cancer 40:774-781  (1979).

Crombie, I.K.  Distribution of malignant melanoma on the body surface.  Brit J
Cancer 43:842-849 (1981).

Dobson, A.J., and Leeder, S.R.  Mortality from malignant melanoma in
Australia:  Effects due to country of  birth.  Intl J of Epidemiology
11(3):207-211 (1982).

Doniger, J., Jacobson, E.D., Krell, K., Di Paolo, J.A.  Ultraviolet light
action spectra for neoplastic transformation and lethality of Syrian hamster
embryo cells correlate with spectrum for pyrimidine dimer formation in
cellular DNA.  Proc Nat Acad Sci (USA) 78(4):2378-2382 (1981).
                              *  DRAFT FINAL  * * *

-------
                                   8-42
Eklund, G., and Malec, E.  Sunlight and incidence of cutaneous malignant
melanoma.  Scand J Plas Reconstr Surg 12:231-241 (1978).

Elder, D.E., Jucovy, P.M., Tuthill, R.J., and Clark, W.H.  The classification
of malignant melanoma.  Am J Dermatopath 2:315-320 (1980).

Elwood, J.M., and Lee, J.A.H.  Trends in mortality from primary tumors of skin
in Canada.  CMA Journal 110:913-915 (1974).

Elwood, J.H., and Lee, J.A.H.  Recent data on the epidemiology of malignant
melanoma.  (1978)?  Journal?

Elwood, J.M., and Hislop.  Solar radiation in the etiology of cutaneous
malignant melanoma in Caucasians.  Natl Cancer Inst Monograph 62:167-171
(1982).

Elwood, J.M., Gallagher, R.P., Hill, G.B., Spinelli, J.J., Pearson, J.C.G.,
and Threlfall, W.  Pigmentation and skin reaction to sun as risk factors for
cutaneous melanoma:  Western Canada melanoma study.  Brit Med J 288:99-102
(1984).

Elwood, J.M., Gallagher, R.P., Hill, G.B., Pearson, J.C.G.  Cutaneous melanoma
in relation to intermittent and constant sun exposure:  The Western Canada
melanoma study.  Int J Cancer 35:427-433 (1985a).

Elwood, J.M., Gallagher, R.P., Davison, J. , and Hill, G.B.  Sunburn, suntan
and the risk of cutaneous malignant melanoma:  The Western Canada melanoma
study.  Br J Cancer 51:543-549 (1985b).

Fears, T.R., Scotto, J., and Schneiderman, M.A.  Skin cancer, melanoma and
sunlight.  Amer J Pub Health 66:461-464 (1976).

Fitzpatrick, T.B. and Soter, N.A.  Pathophysiology of the skin in
Pathophysiology:  The Biological Principles of Disease.  Smith, L.H. Jr. and
Thier, S.O. (eds.), pp. 1297-1330, 1985.

Forbes, P.D., Blum, H.F., Davies, R.E.  Photocarcinogenesis in hairless mice:
dose-response and the influence of dose-delivery.  Photochemistry and
Photobiology 34:361-365 (1981).

Gellin, G.A., Kopf, A.W., and Garfinkel, L.  Malignant melanoma:  a controlled
study of possibly associated factors.  Arch Derm 99:43-48 (1969).

Gordon, T., Crittenden, M., Haenszel, W.  End results and mortality trends in
cancer: II.  Cancer mortality trends in the United States.  Natl Cancer Inst
Monograph 6:133-350 (1961).

Green, A., and Siskind, V.  Geographical distribution of cutaneous melanoma in
Queensland.  The Med J Aust 1:407-410 (1983).
                              *  DRAFT FINAL

-------
                                   8-43
Green, A., Siskind, V., Bain, C., and Alexander, J.   Sunburn and malignant
melanoma.  Br J Cancer 51:393-397 (1985).

Haenszel, W.  Variations in skin cancer merdena within the United States.   NCI
Monograph 10:225-243 (1963).

Herron, J.  The geographical distribution of malignant melanoma in
Queensland.  The Med J Aust 1:892-894 (1969).

Hinds, M.W.  Anatomic distribution of malignant melanoma of the skin among
non-Caucasians in Hawaii.  Brit J Cancer 40:497-499  (1979).

Hinds, M.W., and Kolonel, L.N.  Malignant melanoma of the skin in Hawaii,
1960-1977.  Cancer 45:811-817 (1980).

Hinds, M.W., Lee, J., and Kolonel, L.N.   Seasonal patterns of skin melanoma
incidence in Hawaii.  Am J of Pub Health 71:496-499  (1981).

Hoar, S.K., and Pell, S.  A Retrospective Cohort Study of Mortality and Cancer
Indicence Among Chemists.  J Occup Med 23:483-494 (1981).

Holman, C.D.J., James, I.R., Gattey, P.H., and Armstrong, B.K.  An analysis of
trends in mortality from malignant melanoma of the skin in Australia.   Intl J
Cancer 26:703-709 (1980).

Holman, C.D.J., Mulroney, C.D., and Armstrong, B.K.   Epidemiology of
pre-invasive and invasive malignant melanoma in Western Australia.  Brit J
Cancer 25:317-323 (1980).

Holman, C.D.J., and Armstrong, B.  Skin melanoma and seasonal patterns.  Am J
Epidemiology 113:202 (1981).

Holman, C.D., Heenan, P.J., Caruso, V.,  Clancy, R.J., and Armstrong, B.K.
Seasonal variation in the junctional component of pigmented nevi.  Intl J
Cancer 31:213-215 (1983).

Holman, C.D., Armstrong, B.K., Heenan, P.H.  Guest editorial:  A theory of the
etiology and pathogenesis of human cutaneous malignant melanoma.  J Natl
Cancer Inst 71:651-656 (1983).

Holman, C.D.J., and Armstrong, B.K.  Cutaneous malignant melanoma and
indicators of total accumulated exposure to the sun:  An analysis separating
histogenetic types.  J Natl Cancer Inst 73:75-82 (1984a).

Holman, C.D.J., and Armstrong, B.K.  Pigmentary traits, ethnic origin, benign
nevi, and family history as risk factors for cutaneous malignant melanoma.  J
Natl Cancer Inst 72:257-266 (1984b).
                                 DRAFT FINAL  * * *

-------
                                   8-44
Holman, C.D.J., Armstrong, B.J., Evans, P.R., et al.  Relationship of Solar
Keratosis and History of Skin Cancer to Objective Measures of Actinic Skin
Damage.  Br J Dermatol 110(2):129-38 (1984c).

Holman, C.D.J., Armstrong, B.K., Heenan, P.J.  Relationship of cutaneous
malignant melanoma to individual sunlight-exposure habits.  JNCI 76:403-414
(1986).

Holmbert, B., Westerholm, P., Massing,  R.,  Kestrup, L.,  Gumaelius, H.,
Holmlund, L., Englund, A.  Retrospective cohort study of two plants in the
Swedish rubber industry.  Scand J Work Environ Health 9(2):59-68 (1983).

Houghton, A., Munster, E.W., and Viola, M.V.   Increased incidence of malignant
melanoma after peaks of sunspot activity.  The Lancet (April 8):759-760
(1978).

Houghton, A., Flannery, J., and Viola,  M.V.  Malignant Melanoma in Connecticut
and Denmark.  Am J Cancer 25:95-104 (1980).

Houghton, A.N., and Viola, M.V.  Solar radiation and malignant melanoma of the
skin.  J Am Acad Dermatol 5:477-483 (1981).

IARC monographs.  Biological data relevant to the evaluation of carcinogenic
risk to humans.  Appendix to Vol 40, in press, (1986).

Katz, L., Ben-Turia, S., and Steinitz,  R.  Malignant melanoma of the skin in
Israel:  Effect of.migration.  In: Trends in cancer incidence:  Causes and
practical implications. Magnus, K.  (ed). (1982).

Kiryabwire, J.W., Lewis, M.G.,  Ziegler, J.L., and Loefler, I.  Malignant
melanoma in Uganda.   East Afr Med J 45:498-507 (1968).

Kleinbaum, D.G., Kupper, L.L.,  and Morgenstern, H.  Epidemiologic Research:
Principles and Quantitative Methods.  New York:  Van Nostrand Reinhold  (1982).

Klepp, 0., and Magnus, K.' Some environmental and bodily characteristics of
melanoma patients:  A case-control study.  Int J Cancer 23:482-486 (1979).

Kopf, A.W., Lazar, M., Bart, R.S., Dubin, N., and Bromberg, J.  prevalence of
nevocytic nevi on lateral and medial aspects of arms.  J. Dermatol. Surg.
Oncol. 4124 (1978).

Kopf, A.W., Lindsay, A.C., Rogers, G.S., Friedman, R.J., Rigel, D.S., and
Levenstein, M.  Relationship of Nevocytic Nevi to Sun Exposure in Dyplastic
Nevus Syndrome.  J.  Am. Acad. Dermatol. 12:656-62 (1985).

Lancaster, H.O., and Nelson, J.  Sunlight as a cause of melanoma:  A clinical
survey.  The Med J Aust 6:452-456 (1957).
                                 DRAFT FINAL  * * *

-------
                                   8-45
Larsen, I.E., Little, J.H., Orell, S.R., and Prade, M.  International
pathologists' congruence survey on quantitation of malignant melanoma.
Pathology 12:245-253 (1980).

Lee, J.A.H., and Carter A.P.  Secular trends in mortality from malignant
melanoma.  J Natl Cancer Inst 45:19-97 (1970).

Lee, J.A.H., and Issenberg, H.J.  A comparison between England and Wales and
Sweden in the incidence and mortality of malignant skin tumors.  Brit J Cancer
26:59-66 (1972).

Lee, J.A.H., Petersen,  G.R., Stevens, R.G., et al.  The influence of age, year
of birth, and date on mortality from malignant melanoma in the populations of
England and Wales, Canada, and the white population of the United States.  Am
J Epidemiology 110:734-739 (1979).

Lee, J.A.H., and Strickland, D.  Malignant melanoma:  Social status and
outdoor work.  Brit J Cancer 41:757-763 (1980).

Lee J.A.H.   Melanoma.  In: Cancer Epidemiology and Prevention (Shottenfeld, D.
and Fraumeni, J.F., Jr., Eds.).  Philadelphia:   W.B. Saunders Company (1982).

Lee, J.A.H.   Melanoma and exposure to sunlight.  Epidemiologic Reviews
4:110-136 (1982).

Lee, J.A.H., and Storer, B.E.  Further studies on skin melanomas apparently
dependent on female sex hormones.  Intl J Epidemiology 11:127-131 (1982).

Lee, J.A.H.   Epidemiology of malignant melanoma:  10 Years' Progress.  Pigment
Cell 6:1-21  (1983).

Lew, R.A.,  Sober, A.J., Cook, N., Marvell, R., and Fitzpatrick, T.B.  Sun
exposure habits in patients with cutaneous melanoma:  A case control study.  J
Derm Surg Oncol 9:981-986 (1983).

Lewis, M.G.   Malignant melanoma in Uganda:  The relationship between
pigmentation and malignant melanoma on the soles of the feet.  Brit J Cancer
21:483-495 (1967).

Lilienfeld,  A.M., and Lilienfeld, D.E.  Foundations of Epidemiology.  New
York:  Oxford University Press  (1980).

Little, J.H., Holt, J., and Davis, N.  Changing epidemiology of malignant
melanoma in Queensland. Med J Aust 1:66-69 (1980)

MacDonald,  E.J.  Incidence and epidemiology of melanoma in Texas.  In:
Neoplasms and Malignant Melanoma.  Chicago: Year Book Medical Publishers,
Inc., pp 279 (1976).
                          * * *  DRAFT FINAL  * * *

-------
                                   8-46
MacKie, R.M., and Aitchinson, T.  Severe sunburn and subsequent risk of
primary cutaneous malignant melanoma in Scotland.   Brit J Cancer 46:955-950
(1982).

Magnus, K.  Incidence of malignant melanoma of the skin in Norway,  1955-1970:
Variations in time and space and solar radiation.   Cancer 32(5):1275-1286
(1973).

Magnus, K.  Epidemiology of malignant melanoma of the skin in Norway with
special reference to the effect of solar radiation.   In:   Biological
Characterication of Human Tumors, pp. 249-259, Exerpta Medica International
Congress Series No. 375, Amsterdam (1975).

Magnus, K.  Incidence of malignant melanoma of the skin in the five Nordic
countries:  Significance of solar radiation.   Int J Cancer 20:477-485 (1977).

Magnus, K.  Incidence trends in the nordic countries:  Effects of sun
exposure.  In:  Trends in Cancer Incidence, Causes and Practical Implications
(K. Magnus, ed.), Hemisphere Publication (1981a).

Magnus, K.  Habits of sun exposure and risk of malignant melanoma:  An analysis
of incidence rates in Norway 1955-1977 by cohort,  sex, age, and primary tumor
site.  Cancer 48:2329-2335 (1981b).

Malec, E., Eklund, G., and Langerlof, B.  Reappraisal of malignant melanoma
diagnosis in the Swedish Cancer Registry.  Acta Path Microbiol Scand 85:
707-712 (1977).

Malec, E., and Eklund, G.  The Changing Incidence of Malignant Melanoma of the
Skin in Sweden,  1959-1968.  Scand J Plast Reconstr Surg 12:19-27 (1978).

Mason, T., and McKay, F., U.S. Cancer Mortality by County, DHEW Publ. No.
(NIH) 74-615, U.S. Govt. Printgin Office, Washington, B.C. (1973).

McCarthy, W.H.,  Blach, A.L., and Milton, G.W.  Melanoma in New South Wales:
An Epidemiologic Survey  1970-76.  Cancer 46:427-432 (1980).

McGovern, V.J.,  Mihm, M.C., Bailly, C., Booth, J.C., Clark, W.H., Jr. Cochran,
A.J., Hardy, E.G., Hicks, J.D., Levene, A., Lewis, M.G., Little, J.H., and
Milton, G.W.  The classification of malignant melanoma and its histologic
reporting.  Cancer 32:1446-1457 (1973).

Moss, A.L.H.  Malignant melanoma in Maoris and Polynesians in New Zealand.
Brit J Plastic Surg 37:73-75 (1984).

Movshovitz, M.,  and Modan, B.  Role of sun exposure in the etiology of
malignant melanoma:  Epidemiologic enference.  J Natl Cancer Inst
51(3):777-779 (1973).
                          * * *  DRAFT FINAL  * *

-------
                                   8-47
Muir, C.S., and Nectoux, J.  time Trends:  Malignant Melanoma of the Skin.
In:    (1981?).

National Cancer Institute (NCI), Sondek, E., Young, J.L., Horn, J.W. and Ries
L.A.G. (eds), 1985 Annual Cancer Statistics Review (1985).

National Institute for Occupational Safety and Health (NIOSH).   Unpublished
data:  melanoma after exposure to PCBS (1976).

Osterlind, A., and Jensen O.M.  Trends in Incidence of Malignant Melanoma of
the Skin in Denmark 1943-1982.  Recent Results in Cancer Research 102:8-17
(1986).

Paffenbarger, R.S., Wing, A.L., and Hyde, R.T.  Characteristics in youth
predictive of adult-onset malignant lymphomas, melanomas, and leukemias:
Brief communication.  J Natl Cancer Inst 60:89-92 (1978).

Pakkanen, M.  Clinical appearance and treatment of malignant melanoma of the
skin.  Ann. Chir. Gynaecol. 66:21-30 (1977).

Pathak, D.R., Samet, J.M., Howard, C.A., and Key, C.R.  Malignant melanoma of
the skin in New Mexico 1969-1977.  Cancer 50:1440-46 (1982).

Pitcher H., Melanoma Death Rates and Ultraviolet Radiation in the United
States 1950-1979.  Unpublished manuscript (in press).

Pondes, S., Hunter, J.A.A., White, H., et al.  Cutaneous malignant melanoma in
south-east Scotland.  Quar J Med, New Series L. 197:103-121 (1981).

Roush, G.C., Schymura, M.J., and Holford, T.R.  Risk for cutaneous melanoma in
recent Connecticut birth cohorts.  Amer J Pub Health 75:679-682 (1985a).

Roush, G.C., Schymura, M.J., Holford, T.R., et al.  Time period compared to
birth cohort in Connecticut incidence rates for twenty-five malignant
neoplasms.  J Natl Cancer Inst 74:779-788 (1985b).

Rushton, L., Alderson, M.R.  An Epidemiological Survey of Eight Oil Refineries
in Britain.  Brit J Indust Med 38:225-234 (1981).

Scotto, J., and Fears, T.R.  The association of slar ultraviolet radiation and
skin melanoma among Caucasians in the United States.  Cancer Investigations
(in press).

Scotto, J., and Nam, J.  Skin melanoma and seasonal patterns.  Am J
Epidemiology 111(3):309-314 (1980).

Schulze, R.  Increase of Carcinogenic Ultraviolet Radiation Due to Reduction
in Ozone Concentration in the Atmosphere.  Proceedings of the International
Conference on Structure.  Composition and General Circulation of the Upper and
Lower Atmospheres and Possible Anthropogenic Perturbations.  Atmospheric
Environment Service, Ontario, Canda, Volume'1:479-493 (1974).
                            * *  DRAFT FINAL  * - *

-------
                                   8-48
Serafino, G. and J. Frederick (1986), "Global modeling of the ultraviolet
solar flux incident on the biosphere," prepared for the U.S. Environmental
Protection Agency, Washington, D.C.

Smith, J.L.  Histopathology and biologic behavior of malignant melanoma, pp.
293-330.  In:  Neoplasms of the Skin and Malignant Melanoma.  Chicago:  Year
Book Medical Publishers, Inc. 1976.

Stevens, R.G., and Moolgavkar S.H.  Malignant melanoma:  Dependence of
site-specific risk on age.  Am J Epidemiology 119:890-895 (1984).

Swerdlow, A.J.  Incidence of malignant melanoma of the skin in England and
Wales and its relationship to sunshine.  Brit Med J 2:1324-1327 (1979).

Teppo, L., Pakkanen, M., and  Hakulinen, T.  Sunlight as a risk factor of
malignant melanoma of the skin.   Cancer 41:2018-2027 (1978).

Takahashi, M. and Seiji, M. Acral melanoma in Japan.  In: Pigment cell.  Vol.
6. Malignant melanoma.  MacKie,  R.M. (ed).  Karger, Basel pp 150-165  (1983).

Teppo, L.  Incidence trends of Melanoma of the Skin in Finland by Histological
Type.  In:  Trends in cancer incidence; causes and practical implications.  K.
Magnus (ed.)  Hemisphere Publishing Corp., pp. 693-697 (1982).

Teppo, L., Pukkala, E., Kahama,  M., et al..  Way of life and cancer incidence
in Finland.  Scand J Soc Med (Suppl.) 19:50-54 (1980).  As cited in Lee (1982).

Vagero, D., Ringback, G., Kiviranta, H.  Melanoma and other tumours of the
skin among office, other indoor and outdoor workers in Sweden 1961-1979.  Brit
J Cancer 53:507-512 (1986).

Venzon, D.J., and Moolgavkar, S.H.  Cohort Analysis of Malignant Melanoma in
Five Countries.  Am J Epidemiology 119:62-70 (1984).

Viola, M.V., and Houghton A.N.  Solar Radiation and Cutaneous Malignant
Melanoma Hospital Practice.  June:97-106  (1982).

Wigle, D.T.  Malignant melanoma of skin and sunspot activity.  Lancet 2:38
(1978).
                                 DRAFT FINAL  *

-------
Chapter 9

-------
                                 CHAPTER 9

                  EFFECTS  OF SOLAR RADIATION ON THE IMMUNE
                     SYSTEM AND RESISTANCE TO INFECTIONS
SUMMARY

    Solar radiation has been found to have a detrimental  effect  on  the  immune
system of both humans and experimental animals.   In particular,  UV  radiation
(UVR) reduces the ability of the cell-mediated arm of  the immune response to
respond adequately to antigens.   UV irradiation of the skin  results in  the
generation of cells which suppress the immune response to an antigen
encountered via the skin.    In  a tumor bearing animal, this immunosuppression
may result in the outgrowth of tumor cells which in normal animals  would have
been destroyed by the immune system.  The cells in the skin  that seem to be
the most severely damaged are the Langerhans cells.  These cells are necessary
for the generation of an adequate immune response to those antigens which are
introduced through the skin.  Although it is not certain  that the damage to
the Langerhans cells is entirely responsible for the effect, it  is  clear that
UV irradiation of skin reduces the immune response in  that skin.  Even  more
important,  it is also clear that it is the UVB portion of the UV spectrum that
is responsible for the depression of the immune response.

    One important facet of this  UVR-induced immunosuppression is the role it
plays in UVB-induced carcinogenesis; it affects the host's ability  to
recognize tumor-specific antigens present on the UVR-induced tumor  cells
thereby  increasing the risk for tumor development in  animal models, probably
by allowing the tumor to escape  the normal immune surveillance mechanisms.
Thus the tumor cells are allowed to divide and establish  a large and growing
tumor in the host.

    The immunosuppressive effects of UVR also are very likely to have a
deleterious effect on the immune response to those infectious diseases  that
enter through the skin, especially if the initial immune  response to the agent
takes place in the skin.  As yet, little research has  been done  in  this area.
Preliminary evidence indicates,  however, that UV irradiation during a first
cutaneous infection with two very different organisms, the parasite Leishmania
sp and Herpes simplex virus, may result in an impairment  of  the  immune
response of the host to subsequent infections.  In the case  of leishmaniasis,
this could lead to the development of the more lethal  form of the disease,
visceral leishmaniasis.

    Although there are no experimental data which specifically address  this
issue, all populations may be at risk from the immunosuppressive effects of
UVR.  In addition, individuals who are already immunosuppressed, e.g.,
transplant patients or individuals with immunodeficiency  diseases,  may  be at
greater risk.  In developing countries, particularly along the equator,
immunosuppressive effects occurring as the result of high solar  insolation may
exacerbate parasitic infections  of the skin such as leishmaniasis.
                              *  DRAFT FINAL  * * *

-------
                                   9-2
    The effects of UVB and solar radiation on the human immune system have not
been studied in sufficient detail to allow estimation of dose-response
relationships for these effects.  Qualitatively, it is known from animal
studies that the doses of UVB needed to induce immunosuppression are much
lower than those required for carcinogenesis.   This may mean that exposure to
low doses of UVR, even doses that do not cause a sunburn, may decrease the
ability of the human immune system to provide an effective defense against
neoplastic skin cells or skin infections.
                          * * *  DRAFT FINAL  * * *

-------
                                   9-3
FINDINGS

1.  UV-B SUPPRESSES THE IMMUNE SYSTEM IN ANIMAL EXPERIMENTS.

    la. Acute high doses of UV radiation cause a panimmunosuppression which is
        short lived.

    Ib. Low dose UV radiation also causes a depression in cell-mediated
        immunity which results in an inability to respond to an antigen which
        is presented to the animal through the UV irradiated skin, a
        depression of contact hypersensitivity.

    Ic. Chronic UV radiation or high doses of UV radiation cause a .depression
       . in cell-mediated immunity, delayed-type hypersensitivity, that results
        in an inability of the animal to respond to an antigen which is
        presented ot the animal through unradiated skin.

    Id. Both the effects of the local depression of contact hypersensitivity
        and the systemic suppression of contact hypersensitivity are mediated
        by a T suppressor cell which prevents the development of active
        immunity to the antigen.

    le. The immunosuppressive effects of UVR have been found to reside almost
        entirely in the UVB wavelengths.

2.  SUPPRESSION OF THE IMMUNE SYSTEM MAY PLAY AN IMPORTANT ROLE IN
    CARCINOGENESIS.

    2a. Animals which are UV-irradiated also develop T suppressor cells which
        interfere with the immune response to UV-induced tumors in such a way
        that the animals are more susceptible to the growth of autochthonous
        UV-induced tumors.

3.  LIMITED EXPERIMENTAL DATA INDICATES UV-B SUPPRESSES THE HUMAN IMMUNE
    SYSTEM.

    3a. Although there is not as much information on the effects of UVR on
        humans, there are several studies indicating that the immune response
        of humans is depressed by UVR and is depressed in UV-irradiated skin.

4.  UVB-INDUCED SUPPRESSION OF THE HUMAN IMMUNE SYSTEM IS LIKELY TO HAVE A
    DELETERIOUS EFFECT IN MANY HUMAN DISEASES.

    4a. Preliminary studies indicate that UVR may interfere with the earliest
        immune response to microorganisms that infect via the skin and prevent
        an effective immune response, thus predisposing to reinfection or
        chronic infection.

    4b. Two human diseases that may be influenced by UVB-induced immune
        suppression are Herpes virus infections and leishmaniasis.
                          * * *  DRAFT FINAL  *

-------
                               9-4
4c. Almost no research has been conducted on the influence of UVB on
    infectious diseases; additional investigation is clearly warranted.

4d. For at least one theory of the mechanisms of UVB suppression of the
    immune system (that involving uralonic acid), a possibility exists
    that non-whites, as well as whites, would be vulnerable toincreased
    human suppression caused by ozone depletion.
                      * * *  DRAFT FINAL  * * *

-------
                                   9-5
INTRODUCTION

    As indicated in Chapter 7, it has been known for years that ultraviolet
radiation is carcinogenic.  In the past decade or so, studies designed to
explore the mechanism of UVR-induced carcinogenesis revealed that UVR was a
potent but specific immunosuppressive agent which could abrogate the ability
of an animal to reject its own or a transplanted syngeneic UV-induced tumor.
Subsequent work investigating the mechanism of UVR-induced immunosupression
has indicated that UVR can have a number of effects on the immune system and
that these effects may modify a variety of disease processes.  It is the goal
of this chapter to review information on the effects of UVR on the immune
system, focusing first on what is known about the characteristics of UVR
induced immunosuppression followed by a brief review of what is known about
UVR's effects on the immune system as they relate to other cutaneous diseases
and then finally by a more cursory review of the impact of UVR on a variety of
other diseases, many having an allergic component.

EFFECTS OF ULTRAVIOLET  RADIATION ON THE IMMUNE SYSTEM

    Immediately after acute doses of UV irradiation, there is a suppression of
some immune parameters but after chronic UV irradiation most immune functions
return to normal.  However, research has provided evidence that the effect of
UVR on the immune response early in the course of UV treatment may lead to
events that allow initiated cells to escape immune surveillance and grow into
tumors in a sequence of events unique to UV carcinogenesis.

    Several parameters of immune function have received the greatest
attention, and they will be emphasized here; these are the effects of UVR on
tumor growth, the suppression of contact hypersensitivity and delayed type
hypersensitivity in vivo, and the effects on antigen presentation in vitro.
However, other effects have been noted and will be briefly discussed.  This
chapter will discuss briefly the effects of UVR that can be seen immediately
after treatment, i.e. the acute effects of UVR on the immune system and then
the effects of chronic UV irradiation on the immune system.   The latter two
immune functions are not identical, but are very similar.  In fact, all the
immune functions that are depressed by chronic UV irradiation fall under the
category of cell-mediated immunity, meaning that the effector mechanism
involves a cell rather than an antibody.  Contact hypersensitivity is the
response that is generated when a.person is sensitive to poison ivy.  The
cells responsible for the sensitivity are T cells, lymphocytes which require a
period of residence in the thymus during development to gain functionality.
The cells that actually produce the effect, such as the rash in poison ivy
sensitivity, are called effector cells and, in this case, are T cells.  Other
T cells also take part in the immune response and can either act to suppress
the immune response, suppressor T cells, or allow its expression, helper T
cells.

Effects of UVR on Tumor Growth

    The first indication that UVR might have an impact on the immune system
came from studies by Kripke (1974) who reported that most murine tumors
induced by UVR were rejected when transplanted into normal syngeneic


                          * * *  DRAFT FINAL  * * *

-------
                                   9-6
recipients whereas the same tumors grew progressively in immunosuppressed
mice.  This phenomenon was originally identified in a C3H mouse system in
which tumors were induced using a high intensity mercury arc lamp (Kripke
1974), subsequent studies indicated that it held true for other inbred strains
of mice irradiated using FS40 fluorescent sunlamps (Kripke 1977).

    Experiments designed to determine if the mechanism of this effect was that
UV irradiation of mice led to a generalized immunosuppression, evaluated a
variety of immune parameters using a number of different exposure regimens
(Kripke et al. 1977; Spellman et al. 1977; Norbury et al. 1977; Lynch and
Daynes 1983).  With a few exceptions, all of the functions examined (e.g.,
response to mitogens, allograft rejection, antibody production) remained
normal.  The exceptions included: a)when UV-irradiated mice were used as
recipients in a graft-versus-host assay, there was a small but significant
decrease in reactivity after 2 and 3 months of irradiation which returned to
normal at 4 months (Kripke et al. 1977); b) Reactivity to dinitrochlorobenzene
(DNCB) was depressed after 1 and 2 months of UVR but then also returned to
normal (Kripke et al. 1977) and c) naturally occurring cell-mediated
cytotoxicity in UV-irradiated mice was found to be transiently suppressed 6
days after UV radiation but was found to be normal after 5 to 10 weeks of UVR
(Lynch and Daynes 1983).

    It is clear from the above information that mice which have been
chronically irradiated with UV light are not generally immunosuppressed but
have normal immunity when tested over a very wide number of immune functions.
However, they have a specific immune deficit and are unable to reject
UV-induced tumors.

    Further experiments revealed that the mechanism for the lack of immune
reactivity to UV-induced tumors in UV-irradiated mice was a radiosensitive, la
positive, Lyt-1+2- (Ullrich and Kripke 1984) T suppressor cell (Fisher and
Kripke 1977, 1978, Spellman and Daynes, 1977, 1978,  Daynes et al, 1979).
These cells were specific for an antigen common to UV-induced tumors and did
not affect the ability of immune mice to reject tumors as a function of
immunity to tumor specific transplantation antigens (Roberts et al 1980) or to
histocompatibility antigens (Kripke and Fisher 1976).  Preliminary evidence
indicates that UV-irradiated mice immunologically recognize UV-induced tumors
(Romerdahl and Kripke, 1986), but that the suppressor cell acts to prevent the
development of cytotoxic effector cells.

    The suppressor T cells were found to be specific for UV-induced tumors,
with one exception.  The presence of UV-induced tumor specific suppressor
cells was not found to interfere with rejection of transplanted syngeneic
tumors (Kripke et al 1979) induced by methylcholanthrene and Moloney sarcoma
virus as well as spontaneous tumors.  Interestingly, growth of only one
spontaneous tumor, B16 melanoma, was enhanced in UV-irradiated mice.  The
enhanced growth of B16 melanoma in UV-irradiated mice was confirmed by Bowen
and Brody (1983) who reported an increased percentage of challenged mice which
bore tumors and decreased tumor latency.  It was also found that one can
achieve the same effects on the transplantation immunity to  UV-induced tumors
with sunlight as with those artificial sources of UV radiation,used in
experimentation (Morrison and Kelley 1985).
                            * *  DRAFT FINAL  * * *

-------
                                   9-7
    Experiments in which plastic filters were used to remove wavelengths from
the emission of Westinghouse FS40 sunlamps indicated that the ability of UVR
to induce susceptibility to UV-induced tumors lay in wavelengths below 315 nm,
for as the lower wavelengths were removed from the UVR, the effectiveness of
the UVR in inducing susceptibility to UV-induced tumors decreased.
Wavelengths above 315 nm were ineffective in inducing the susceptibility.
Reducing the dose rate through neutral density filters while keeping the total
dose constant did not change the proportion of mice which were tumor
susceptible; therefore there was reciprocity with regards to dose-rate and
time of UV exposure (De Fabo and Kripke, 1979, 1980).

      It has also been shown that UV irradiation can affect the latent period
and antigenic properties of tumors induced by other carcinogenic agents
including benzo(a.)pyrene and methylcholanthrene (Roberts and Daynes 1980).
Ebbesen (1981) reported an enhanced incidence of lymphomas in BALB/c mice
which had been UV-irradiated.

    The role that UVR induced immunosuppression plays in tumorigenesis by UVR
has been investigated by a number of laboratories (Fisher and Kripke 1982; De
Gruijl and Van Der Leun 1982, 1983).  In a series of experiments, it was found
that UV-induced tumorigenesis was enhanced in those mice reconstituted with
lymphoid cells from UV-irradiated mice as compared to mice reconstituted with
cells from normal mice (Fisher and Kripke 1982).  It was also found that
preirradiation of mice enhanced subsequent UV tumorigenesis in previously
unexposed skin (De Gruijl and Van Der Leun 1982, 1983), and that ventral UVB
irradiation greatly enhanced the susceptibility of mice to dorsal two-stage
carcinogenesis (Strickland et al 1985).  Thus it is clear that UVR induced
immunosuppression (and thus the presence of the T-suppressor cell) is of great
importance in determining if a tumor will develop in response to the
carcinogenic action of UVR.  This implies that immune surveillance plays a
role in retarding the growth of transformed cells and thus may at least
partially account for the long latent period of those tumors.

Effects of  UVR on Other  Immune Parameters

Suppression of Contact Hypersensitivity

    In subsequent investigations of the impact of UVR on immune responses, it
was discovered that UVR treatment resulted in a decreased capacity to elicit
contact hypersensitivity  (CHS) responses to antigenic chemicals applied
directly to the skin.  Low doses of UVR result in a local suppression of
contact hypersensitivity and animals may not be sensitized directly through
UV-irradiated skin although they are normal in their responses when sensitized
through skin not exposed to UVR.  Larger doses of UVR result in a systemic
suppression of contact hypersensitivity and animals are hyporesponsive
regardless of the site of antigen application.  In both cases antigen-specific
T suppressor cells have been reported to develop.

    The reason for the hyporesponsiveness of UV-irradiated animals is thought
to be the fact that under certain circumstances UVR can suppress what is known
as antigen presentation.  It had been discovered earlier (Shevach and
Rosenthal, 1973) in vitro studies that certain adherent cells in the spleen
and lymph nodes were required to present antigen to T cells in order to  elicit

                          * * *  DRAFT FINAL  * * *

-------
                                   9-8
a proliferative T cell response.  T cells without the aid of antigen
presentation by these adherent cells respond to antigen poorly or not at all.
Antigen presenting cells were found to be adherent to plastic and were found
in the spleen and among peritoneal exudate cells.  In addition, it was
discovered that certain cells in the epidermis, the Langerhans cells could
also act to present antigen to lymphocytes.  To examine the function of
antigen presentation, one may chemically link (derivatization) a molecule to
the antigen presenting cell.  Frequently, haptens are used.  A hapten is a
molecule that, when present by itself, does not elicit an immune response.
However, when linked to another, frequently larger molecule, a hapten can
induce a specific immune response.

Suppression of Systemic Hypersensitivity

    When adherent cells from the spleen or peritoneal exudate cells from
UV-treated donor mice were derivatized with trinitrophenyl (TNP), they could
not induce hapten-specific delayed-type hypersensitivity in UV irradiated
mice; whereas adherent TNP-derivatized from normal mice were able to induce
delayed-type hypersensitivity in UV-treated mice (Greene et al, 1979).  The
hyposensitivity could be transferred into a second recipient and a T
suppressor cell was shown to be present in UV-irradiated mice immunized with
splenic adherent cells from UV-irradiated mice.  Thus, the potential for
antigen presentation in UV-irradiated mice was changed in such a way that
there was apparently preferential induction of suppressor cells rather than
effector cells.  This was a systemic effect in that it was unnecessary
immunize through the UV-irradiated skin to show the suppressive effect.  Other
researchers also reported similar systemic suppression of antigen
presentation  (Noonan et al (1981b); Fox, 1981).

    Mice given UVR for an extended time period (5 wks) no longer exhibited
depressed antigen presentation (Gurish, 1982).  However, after 6 days of UVR,
although splenic adherent cells have depressed antigen presenting cells
function, lymph nodes adherent cell antigen presenting cell function was
elevated.  This elevation was abrogated by splenectomy prior to UVR.
Therefore decrease in splenic antigen presenting cell activity caused by acute
doses of UVR may be due to migration of splenic antigen presenting cells to
peripheral lymphoid tissues which drain the site of epidermal inflammation.

Suppression of Local Delayed Hypersensitivity

    In addition, UVR also locally suppresses the response to antigen which is
applied directly to the UV-irradiated skin.  Researchers have studied the
kinetics of the suppression.  Following a dose of UVB from FS40 lamps of
32,480 J/m2 in a single exposure, depression of the ability to be sensitized
to contact hypersensitivity in mice is observed immediately in UV-irradiated
skin whereas immunization in nonexposed skin elicits a normal response (Lynch
et al, 1983).  However, 3 days later, mice are not able to be sensitized
through non-UV-exposed skin  (Noonan et al 1981a).  At that time point, normal
densities of ATPase-positive Langerhans cells with normal histology were found
in the irradiated skin; yet normal contact hypersensitivity responses were not
found when dinitrofluorobenzene (DNFB) was applied to the UV-irradiated skin.
                          * * *  DRAFT FINAL  * * *

-------
                                   9-9
In addition, epidermal cells harvested from skin which was UV irradiated
immediately prior to the experiment were unable to function in inducing T cell
proliferation in vitro.  Epidermal cells from mice which had been exposed to
UVR three days previously had an enhanced ability to induce T-cell
proliferation whereas epidermal cells from non-irradiated skin were always
able to so function.  This was confirmed by the finding (Cooper et al 1985)
that the initial depression of antigen presentation in UV-irradiated skin is

quickly followed by the appearance of T6- Dr  cells that enhance epidermal
alloantigen presentation.  The local suppressive effect of UVB radiation was
also found to occur in guinea pigs (Haniszko and Suskind, 1963) (Morison and
Kripke (1984) (Morison et al. 1981).

    Application of DNCB to UVB-irradiated body wall skin of mice produces a
state of unresponsiveness that is associated with a Lyt-l+ T  suppressor cell
that acts on the induction phase of immunity (Elmets 1983) .  It was suggested
that UV impairs the antigen-presenting potential of Langerhans cells, in whose
absence other hapten-derivatized cells are able to deliver a tolerogenic
signal.

Effect of UVR on Skin Graft Rejection

    The ability of UVR to depress antigen presentation in vivo was tested to
determine if UV irradiation of skin grafts would be less immunogenic and thus
if UV treatment could .increase skin graft rejection time.  In the first
report,  full thickness skin grafts were UV-irradiated and transplanted onto
normal mice.  However, there was no effect on first set or second set graft
rejection times (Streilein et al 1980).  However, when isolated mouse tail
skin was UV-irradiated in vitro on both sides, transplantation across I/region
differences was significantly prolonged and 50 percent of the grafts were not
rejected at all (Claas et al 1985).  The prolongation of the skin graft
rejection time occurred over I-region differences but across H-2 differences,
usually considered a stronger difference, there was no prolongation of the
skin graft rejection time.

Effect of UVR on Lymphocyte Traffic

    Milton (1981) also showed that UVR of lymphocytes in vitro had no effect
on localization of lymphoid cells when those cells were injected into naive
recipients 1 hr after radiation, but significantly depressed localization in
lymph nodes of syngeneic and allogeneic recipients when lymphoid cells were
cultured 24 hr after UVR and then injected.   UVR of lymphocytes in vitro
followed by a 4 hr incubation led to significant alterations in the
localization of the lymphocytes to lymph nodes, gut, liver,  and kidneys
(Spangrude et al, 1983).  UVC, but not UVB, depressed the localization in the
spleen.  UVR of mice resulted in a dramatic and long-lasting increase in the
tropism of peripheral lymph nodes for circulating lymphoid cells.  Termination
of UVR did not reverse the tropism.  The homing patterns of lymphocytes from
long-term UV-irradiated mice did not vary from normal lymphocytes in UV
irradiated mice.  The alterations in tropism of lymphocytes from UV-irradiated
mice may play a role in the induction of a long lasting suppression.  In
addition, dermal vasculature of the skin has been found to expand in terms of
size and number of vessels (Spangrude et al 1983).  However, very few
inflammatory cells were detected in irradiated skin.

                          * * *  DRAFT FINAL  * * *

-------
                                 .  9-10
Effects of UVR on Lymphocyte Function In Vitro

    In another measure of the effect UVR on immune functions, lymphocytes
irradiated with UVC in vitro had nearly the same alloimmunogenicity in vivo
when measured by primary antibody cytotoxic response (Milton, 1981).   However,
low doses of UVR in vitro from an FS40 sunlamp abrogated the differentiation
events which led to the generation of functional cytotoxic T cells (Lynch et
al 1981).  T cells were found to be more sensitive to the effects than
accessory cells.  The effect seemed to be due to wavelengths in the UVB range
since much higher amounts of UVA were not inhibitory.

Photobiology and Wavelength Dependence of Immune Suppression

    The great majority of experimentation has been done with lamps which do
not accurately simulate sunlight.  (Details of the lamps used in the reserach
presented in thie chapter are present in Appendix 9-A.)  However, it was also
shown that sunlight could suppress contact hypersensitivity in both mice and
guinea pigs.   Exposures occurred in an unshaded area on sunny days in
September and early October at latitude 30, Frederick, MD.  Exposure of mice
on 3 consecutive days for 5 hours each resulted in a total dose of UVB of
+37,000-48,000 J/m2 and a 77% suppression of contact hypersensitivity.
Suppression was dose-dependent and could be transferred into naive recipients
with lymphocytes.  Mylar-filtered sunlight suppressed contact
hypersensitivity, but less than that produced by unfiltered sunlight (Morison
et al 1985).

    In a study of the photobiology of the suppression of contact
hypersensitivity by UVR, BALB/c mice were irradiated on the shaved dorsal skin
with FS40 sunlamps at a total dose of about 32,000 J/m2 and sensitized 5 days
later on the abdomen with dinitrofluorobenzene (DNFB) in acetone (Noonan et
al, 1981a).  Changes in dose fractionation over 5 days and changes in
dose-rate over 10-fold range did not change the total dose required to induce
systemic suppression.  Since a Mylar filter, which removes wavelengths less
than 315 nm,  abrogated the effect, it was concluded that depression of contact
hypersensitivity was due to UVB.  However, the UVC emitted by the FS40 lamps
may also play a role in depression of contact hypersensitivity.  It has been
shown to induce susceptibility to challenge with UV-induced tumors (Lill,
1984). Passive transfer of lymphocytes from mice which had been UV-irradiated
and then treated with DNFB as above showed that such treatment resulted in the
formation of antigen specific T suppressor cells.

    Elmets et al (1983) studied the wavelength dependence of the contact
hypersensitivity/suppressive effect of UVR and determined that the greatest
efficiency of suppression was seen at 297 nm.  Exposure to 400 J/m2 daily for
4 days resulted in 60 percent suppression whereas at wavelengths of 290 nm 700
J/m2 were required for equivalent suppression.  Above 297 nm, a dose of 1000
J/m2 did not induce unresponsiveness.  De Fabo and Noonan (1983) also
constructed an action spectrum but examined 10 different narrow band
wavelengths from 250 nm to 320 nm and found peak suppression of contact
hypersensitivity at 270 nm.  However, they also report that UVR at 295 nm is
slightly more effective at inducing hyposensitivity than UVR at 290 nm which
                            * *  DRAFT FINAL  * *

-------
                                   9-11
agrees with the data of Elmets et al (1983).  These wavelengths of UVR are
those which will be increased if ozone in the upper atmosphere is depleted.

Mechanism of  the  Induction  of Suppression

    Daynes et al (1985) proposed the hypothesis that recirculating Ia+ cells
have antigen presenting cell function.  Exogenous inflammatory stimulation of
the skin by UVR may result in a time-dependent redistribution of these cells
to peripheral compartments resulting in reduced antigen presenting cell
activity in the spleen and the enhanced antigen presenting cell activity in
the skin and draining lymph nodes.  They conclude that initially there may not
be a systemic defect, but a redistribution of antigen presenting cells.
However, the induction of suppressor cells in UV irradiated mice indicates
that the potential for antigen presentation has been altered.

    De Fabo and Noonan (1983) reported that their action spectrum of contact
hypersensitivity suppression is congruent with the absorption spectrum of
urocanic acid, a substance found in the epidermis.  The majority of urocanic
acid in the epidermis is found in the stratum corneum.  When the skin of mice
was stripped with adhesive tape four times before UVR, the stratum corneum is
removed.  If the mice are then treated with UVR, contact hypersensitivity
suppression does not occur.   They hypothesize that urocanic acid is a
photoreceptor which is located primarily in the stratum corneum and which
induces hyporesponsiveness upon UV irradiation.

    It has also been possible to transfer suppression of contact
hypersensitivity using serum from UV-irradiated mice (Swartz 1984).  Serum
taken immediately after UV-irradiation with FS40 sunlamps actually slightly
enhanced contact hypersensitivity but serum taken 3 hours after UV irradiation
suppressed contact hypersensitivity.  More interestingly, spleen cells from
mice that had received serum that had depressed contact hypersensitivity were
able upon passage into normal mice to suppress contact hypersensitivity in
those mice, indicating the probability that the transferred serum had played a
role in the induction of a suppressor cell.

    It has also been reported that prostaglandins play a role in the induction
of hyposensitivity.  Chung et al (1986) reported that both local and systemic
suppression of contact hypersensitivity by UVR could be abrogated by treatment
of the mice with indomethacin.  Also, effector cells for contact
hypersensitivity could be recovered from draining lymph nodes of all mice
sensitized through non-irradiated skin sites and the spleen of both normal and
UV-irradiated mice contained suppressor cells after hapten application.
Therefore, the authors concluded that the effect of UV radiation on the immune
system is not only through the induction of suppressor cells, but by
modification of effector cell generation or by prostaglandin-mediated
depression of contact hypersensitivity effector cell activity.

ANTIGEN PRESENTATION IN  VITRO

    In the experiments discussed above, the effect of UVR on immune reactivity
in vivo has been assessed.  The effect of UVR on antigen presentation has also
been studied in a slightly different fashion using in vitro measures of immune
function.  It has been shown that there are epidermal cells that are capable

                          * * *  DRAFT FINAL  * * *

-------
                                   9-12
of inducing proliferation of T cells in vitro (Stingl et al 1978).  There are
two experimental systems used which will be discussed.  In one case, the
ability to stimulate the proliferation of T cells in an allogeneic (same
species, different strain or histocompatibility antigens) lymphocyte
proliferation assay is studied.  Epidermal cells are mixed with allogeneic
lymphocytes.  The non-self, alloantigens present on the epidermal cells induce
the lymphocytes to proliferate, if the antigens present on the epidermal cells
are presented to the lymphocytes in a fashion to which the lymphocytes can
respond.  Thus this is an in vitro measure of antigen presentation.  Epidermal
cells can also stimulate the proliferation of syngeneic T cells; although the
biologic significance is not understood.  In the second experimental system,
researchers take advantage of the fact that epidermal cells can also present
exogenous antigen to immune T lymphocytes and induce proliferation in those
cells (Faure et al 1984).

    Langerhans cells are the only epidermal cells which express la antigen
under normal circumstances.  Pretreatment of epidermal cells with anti-la
antibody was reported to abolish ability of epidermal cells to function in an
allogeneic stimulation assay.  Therefore, it was concluded that it was the
Langerhans cells that were necessary for stimulation of lymphocyte
proliferation (Aberer et al 1982).  A dose of UVB irradiation (50 J/m2) in
vitro was found to abolish the stimulation of allogeneic lymphocyte
proliferation by Langerhans cells.  This same dose of UVR did not affect
epidermal cell viability and only marginally decreased the protein
synthesizing ability of epidermal cells as measured by uptake of radiolabeled
leucine.  Thus it was not necessary to kill the Langerhans cell to reduce its
ability to present antigen.

    When one mixes lymphocytes from two different strains of mice, the
lymphocytes recognize that there are histocompatibility differences and
respond by proliferating.  The relative roles of the two populations can be
measured by treating one of the other with a drug such as mitomycin C that
prevents division of that cell without killing it.  Using this system, studies
somewhat similar to those performed on epidermal cells were done with
lymphocytes.  Spangrude et al (1983) reported that in vitro UVR of lymphocytes
showed that quite low doses of UVR were capable of depressing the ability of
lymphocytes to act as stimulator or responder cells in a mixed lymphocyte
culture.  Doses of 20 J/m2 of UVB or UVC were effective in reducing the
response whereas the maximum effect was seen at a dose of at 40 to 60 J/m2.
UVC was slightly more effective than UVB in depressing proliferation.
Therefore the ability of UVR to depress recognition of foreign antigens by
lymphocytes is not limited to antigens presented by Langerhans cells.

    There is a second system in which the antigen presenting capabilities of
epidermal cells has been studied (Stingl et al, 1983).  Mice were immunized in
vivo with dinitrophenylated ovalbumin (DNP6-OVA).  An epidermal cell
suspension was prepared from normal mice and pulsed with antigen.  The
antigen-bearing epidermal cells were then washed and exposed to T cells from
the immunized mice.  If the epidermal cells are able to properly present the
antigen, DNP6-OVA, to the immune T cells, the T cells will proliferate in
response.   Exposure of epidermal cells in vitro to UVB before pulsing with
                          * * *  DRAFT FINAL  * * *

-------
                                   9-13
DNP6-OVA led to a radiation dose-dependent reduction in their ability to
stimulate antigen-specific proliferation in syngeneic immune T-cells.
However, when epidermal cells were antigen-pulsed before UVR, their
stimulatory capabilities were at least partly maintained.

    A potential mechanism for the lack of stimulation of lymphocyte
proliferation was reported by Sauder et al (1983) who reported that epidermal
T-cell activating factor (ETAF), which has been found to be virtually
identical to IL-1 (Luger et. al 1983), could partially restore T cell
stimulation which had been suppressed by UVR.  In addition the differences in
T-cell proliferation induced by epidermal cells which are UV irradiated prior
to antigen pulse, compared to that induced by normal cells, correlated well
with the amount of ETAF generated in epidermal cell cultures consisting of
keratinocytes and Langerhans cells (Stingl et al 1983).  Addition of ETAF
could partially restore the decreased capacity of epidermal cells which are
pulsed with antigen after UVR but not before UVR to stimulate antigen-specific
T-cell proliferation.  However, UVR was also shown to increase the production
of ETAF in keratinocyte cultures (Ansel et al 1983) and epidermal cells from
both normal and UV-irradiated mice were found to produce equivalent amount of
ETAF (Gahring et al 1984).  Therefore, the role of ETAF is far from clear.

LANGERHANS CELLS AND ANTIGEN PRESENTATION

    As discussed above, Ia+ epidermal cells can present antigen in vitro, this
capability is depressed by UVR, and the epidermal cell which has been
implicated in antigen presentation is a Langerhans cell.  Further research has
focused on the relationship of Langerhans cells to the induction of contact
hypersensitivity and effects of UV on Langerhans cells.

    To determine the relationship between antigen presentation and the
presence of Langerhans cells, Toews et al (1980) took advantage of the fact
reported earlier (Bergstresser et al, 1980) that murine tail skin has
significantly fewer Langerhans cells than body wall skin. Sensitization could
only be effected through the body wall skin.  When a short course of UV light
at a dose of 100 J/m2 day for 4 days from FS20 lamps was used to deplete
Langerhans cells from body wall skin, sensitization could also not be effected
through UV-irradiated skin. In addition, mice whose first exposure to DNFB was
through Langerhans cell-deficient skin (tail skin or UV-treated body wall
skin) could not subsequently be sensitized through body wall skin.  There was
no hyporesponsiveness in UV-treated mice if the first antigen contact was
through unirradiated skin.  This work was confirmed by Semma and Sagami (1981)
who in addition showed that the hyporesponsiveness was due to active
suppression of contact hypersensitivity by hapten-specific T suppressor
cells.  Interestingly, it was shown that when tail skin is transplanted to the
body, attempts to sensitize the animal through the tail skin are unsuccessful
(Streilein et al 1980). Therefore, the inability to induce contact
hypersensitivity through the tail skin is due to a peculiarity of the skin and
not its anatomic location.

    UVR can deplete Langerhans cells from skin as measured both by ATPase
staining, expression of la antigen, or morphologic criteria.  The depletion of
Langerhans cells in both murine and human skin by UVB and UVA had very similar


                          * * *  DRAFT FINAL  * * *

-------
                                   9-14
dose responses in both mice and humans (Aberer et al 1981) upon exposure to
either single or cumulative doses of UVA, UVB, or UVA containing small amounts
of UVB..  Doses of UVR which completely abrogated the Langerhans cell markers
of ATPase staining and la antigens also caused cellular damage to some
Langerhans cells, as seen by electron microscopy.  However, the vast majority
of the Langerhans cells seen ultrastructurally appeared normal.   At lower
doses of UVR, such as 150-300 J/m2 of UVB and 4000-10,000 J/m2 of UVA,
Langerhans cells were the only epidermal cells to show damage at the
ultrastructural level whereas higher doses affected all epidermal cells.
Depression of la antigens and ATPase staining occurs at lower doses of UVR
than cellular damage.  There was a rough correlation between staining for la
antigens and ATPase staining, but staining for ATPase persisted in
intracellular areas at UVR doses which completely abolished staining for la
antigens.   Gilchrest et al (1982) reported that although young adult (22-26
year-old) humans have more Langerhans cells than old (62 to 68 year-old)
humans, the Langerhans cells in young adults were much more sensitive to UVR.
Obata and Tagami (1985) studied the number of Ia+ cells which remained after
UVR. After a single dose of UVB radiation (2000 J/m2 at 300 nm) they
determined that although some Langerhans cells simply lost la antigen, the
majority disappeared due to cell damage. From 4 to 14 days after UVR,
enlargement of the remaining Langerhans cells occurred and very large
Langerhans cells were seen.  The recovery rate of Langerhans cell number after
UVR was slower than for keratinocytes.

    Action spectra for depletion of Langerhans cells and keratinocyte damage
were reported by Obata and Tagami (1985) to be similar and to also be similar
for what has been reported for the induction of unscheduled DNA synthesis and
UVR-induced erythema.  To reduce the number of Ia+ cells to less than 20
percent of normal numbers, UVR at 260 nm was more effective than 280 nm and
300 nm, which gave similar results.   UVR at 320 nm and greater was much less
effective.  For example at 340 nm and 360 nm, 10,000 J/m2 did not reduce the
number of Ia+ cells and at 327 nm, 10,000 J/m2 only gave a 40 percent
reduction whereas 2000 J/m2 at 280 nm and 300 nm reduced the number of Ia+
cells to less than 20 percent of control.  This may play a role in the
suppression of local contact hypersensitivity.

    However, Noonan et al (1984) reported that the depletion of Langerhans
cells by UVR and the suppression of systemic contact hypersensitivity by UVR
are not causally related.  Langerhans cells were identified by ATPase staining
of EDTA-separated epidermal sheets.  Electron microscopy was used to confirm
the staining and it was reported that there was no evidence for Langerhans
cells which did not stain for ATPase.  Mice were irradiated with 270 nm, 290
nm, or 320 nm UVR produced by interference filtration of a Xenon-arc light
source.  UV radiation at 270 nm or 290 nm caused a loss of dendrites and a
decrease in the total number of Langerhans cells in a dose dependent fashion.
The dose to yield a 50 percent reduction in the number of Langerhans cells was
about 250 J/m2 for 270 nm and 290 J/m2 for 290 nm. Doses of 270 nm and 290 nm
UVR which decreased the number of Langerhans cells were insufficient to
suppress contact hypersensitivity to TNCB applied to unexposed skin.  A dose
of 320 nm (8200 J/m2) which caused 50 percent suppression of contact
hypersensitivity had no effect on the number or morphology of Langerhans
cells.  In addition, the number and morphology of Langerhans cells on the
                          * * *  DRAFT FINAL  * * *

-------
                                   9-15
ventral epidermis was not affected following a dose of UV to the dorsal skin
even though that UVR resulted in systemic suppression of contact
hypersensitivity when there was an attempt to sensitize mice though the
ventral epidermis.

    The same conclusion was reached by Morison et al (1984) using a different
approach.  Mice were exposed to UVB on the dorsum with a single 3 hour
exposure. Mice were also subjected to the trauma of photosensitization with
rose bengal or eosin and visible light filtered to cut off below 390 nm.
Finally, mice were also treated with UVA continuously for 72 hr.  The
Langerhans cells in the skin were examined 3 days after UVR by staining for
ATPase and by electron microscopy. The number of normal-appearing Langerhans
cells 3 days after UVB was only 28 percent whereas the number and morphology
in unexposed skin in the same mice were normal.  UVA radiation resulted in the
complete destruction of normal Langerhans cells.  Photosensitization with rose
bengal or eosin and visible radiation did not reduce Langerhans cells.
However, when contact hypersensitivity was tested, photosensitization with
rose bengal and visible light did suppress TNCB reactivity whereas
photosensitization with eosin and visible light or treatment with UVA did
not.  Thus the absence of Langerhans cells in the epidermis did not correlate
with suppression of contact sensitivity.  Also, the presence of the UVB
damaged Langerhans cells was not the cause of hyporesponsiveness since
depletion of Langerhans cells by UVA (which did not decrease contact
hypersensitivity) followed by UVB radiation also depressed contact.
hypersensitivity.   This is important because it implies that the mechanism of
the depression of immune reactivity by UVB is more specific than simple tissue
damage.

    A possible answer to the conflicting reports on the role of epidermal
cells in UV-induced suppression of contact hypersensitivity has been
proposed.  In agreement with a larger number of previous studies, Granstein
(1985a, 1985b) has reported that when mice are immunized subcutaneously with
hapten-coupled UV-irradiated epidermal cells, the mice are hyporesponsive when
delayed-type hypersensitivity is measured with a footpad assay and in addition
hapten-specific T suppressor cells are present.  However, depletion of I-J
bearing cells with anti-I-J antibody and complement from the UV irradiated
epidermal cells (prior to UVR and haptenation) prevents the appearance of
these suppressor cells.  Non-UVR treated epidermal cells which are I-J
depleted and haptenated can induce delayed-type hypersensitivity.  Therefore,
there seems to be an additional cell in the epidermis, other than the
Langerhans cell and the keratinocyte, which is UVR resistant and which
apparently presents antigen in a tolerogenic fashion and thus preferentially
induces suppressor cells.  Thus hypothetically, when antigen is presented in
the skin in the usual fashion, Langerhans cells present the antigen and
effector cells are generated.  However, when skin is UV-irradiated, the
function of the Langerhans cell is depressed and antigen can then be presented
by the I-J cell, resulting in the generation of suppressor cells.

    However, one needs to explain the relationship between induction of
tolerogenic presentation of antigen by UVR and the induction of suppressor
cells which are specific for antigens present on UV-induced tumors.  Spellman
                          * * *  DRAFT FINAL  * * *

-------
                                   9-16
and Daynes (1984) have reported that there are antigens present on
UV-irradiated skin that are cross-reactive with UV-induced tumors.  Mice were
irradiated 30 minutes per day for 6 weeks, which is well under the time of
radiation required to produce tumors.  Skin from these animals was then
transplanted as 1 cm grafts to normal syngeneic mice.  Twenty days later a
second graft of UV-irradiated skin was made.  When mice were challenged with a
UV-induced progressor tumor 15 days later, animals were protected from the
challenge.  Since grafting of UV-irradiated skin protected a mouse from
challenge with a UV-induced tumor, the conclusion was drawn that similar
antigens were present on both the UV-irradiated skin and the UV-induced
tumors. Therefore, if UVR causes the expression of neoantigens on skin and at
the same time facilitates the induction of suppressor cells to antigens, then
the immune response against those tumors might take the form of induction of
suppression rather than induction of cytotoxic cells.

HUMAN STUDIES

    There have been fewer studies of the effects of UVR on the immune response
in humans.  Morison et al (1979) irradiated human volunteers with 1.5 minimal
erythemal doses (MED) to produce what they termed asymptomatic erythema or 3
MED to produce what they termed symptomatic erythema.  Post irradiation, all
subjects showed a significant increase in the proportion of circulating
polymorphonuclear leukocytes and a corresponding increase in the proportion of
circulating lymphocytes.   Subjects who received 3 MED also had a significant
decrease in the proportion of circulating E-rosette forming cells, i.e. T
cells, and an increase in the proportion of null cells.  The changes in the
absolute numbers of these cells were not significant.  Thirty minutes and 4
hours after UV-irradiation, the response of lymphocytes to PHA was increased,
then decreased, reaching a minimum 12 hours after UVR, and returning to normal
72 hours post exposure.

    Hersey et al (1983a) reported on the effects of solarium exposure on
normal human subjects.  Normal volunteers were given a standard course of
treatment in a solarium to acquire a suntan; twelve 1/2 hr exposures.  The
energy delivered was 10 J/m2 for UVA and about 1% of total irradiation was UVB
(which was not quantitated).  Immune functions were carried out before, on
completion, and 2 weeks after the end of irradiation.  The test subjects had
reduced skin test responses to DNCB and slightly reduced numbers of blood
lymphocytes.  There was a relative increase in T (OKT3+) cells which was
attributable to an increase of OKT8+ (suppressor) cells.  There was a
significant increase in suppressor T-cell activity against pokeweed mitogen
induced IgG production in vitro and a depression of NK activity.  The changes
were still present in some subjects 2 weeks after solarium exposure.  The NK
data is a little uncertain since the control subjects had lower levels of NK
activity throughout the studies than the UV-irradiated subjects.  Some
patients had increased NK activity after treatment and some did not change.
In a second study (Hersey, 1983b), normal human volunteers were exposed to
sunlight for 12 days in a 2 week period.  Tests of immune function were made
before, on completion and 2 weeks after the end of the UVR.  There were 15
UV-irradiated subjects and 13 age- and sex-matched controls.  There were no
significant changes in hemoglobin levels, total leukocyte count, and
lymphocyte, or neutrophil counts. In test subjects there was a small but
                              *  DRAFT FINAL  * * *

-------
                                   9-17
significant drop in T cell numbers which returned to normal by 2 weeks.   There
was, however, a marked increase in the ratio of suppressor cells to cytotoxic
helper cells which had not returned to normal at 2 weeks after
UV-irradiation.  After UVR there was an increase in the activity of pokeweed
mitogen-induced gamma radiation-sensitive suppressor T cell activity against
IgG and IgM production.  Although immunoglobulin production in vitro has too
wide a range for statistical analysis of the small numbers of patients to be
of value, gamma irradiation of T cells prior to culture increased the amount
of immunoglobulin produced in patients exposed to sunlight but not in
controls.  There were 5 persons who did not show these changes, but they were
largely accounted for by persons who did not receive the full dose of UVR.
Immunoglobulin levels in vivo did not change nor did NK activity change
significantly in this study.  O'Dell et al (1980) reported that there was a
diminished immune response in sun-damaged skin.  The concentration of DNCB
required to elicit a positive patch test was greater in sun-damaged skin than
in skin which was normal.  In addition, the delayed-type hypersensitivity to
intradermal injection of Candida, mumps, and PPD antigens was decreased in
sun-damaged skin so that the differences were not due to a difference in
percutaneous absorption of antigen through sun-damaged skin.  The response to
a primary irritant was the same in both sun-damaged skin and normal skin and
there was no difference in the two tested sites (back of the neck and the
back) in volunteers without sun-damaged skin at the back of the neck.
Therefore, there is apparently a local suppression of contact hypersensitivity
in sun damaged skin.

    There is also evidence that the inability to repair UV-induced damage may
also play a role in impairment of immune function.  A syndrome xeroderma
pigmentosum has been defined in which the patients, who suffer from sun
sensitivity and increased numbers of sunlight-induced cancers, also have been
shown to have impaired ability to remove and repair lesions in their DNA
caused by UVR.  Morison et al (1985) reported that xeroderma pigmentosum
patients had impaired ability to develop contact hypersensitivity to DNCB in
sun-exposed skin when compared to normal control subjects.  The numbers and
morphology of Langerhans cells in the skin were the same in both groups.
Unfortunately, the study did not include an assessment of ability to develop
contact hypersensitivity in skin not usually exposed to sunlight.

EFFECTS OF  ULTRAVIOLET RADIATION ON INFECTIOUS DISEASES

    Upon the realization that UVB can suppress antigen presentation of
antigens applied through both irradiated and unirradiated skin, concern
developed that the immune response to diseases in which the first contact of
the immune system with the invading organism was though the skin might be
adversely affected by UV irradiation.  This was especially true since only low
doses of UV irradiation had been found to be necessary to suppress an immune
response the allergen was applied through the UV-irradiated skin.  As yet, not
a great deal of research has been published on this problem, although there
are at least two infectious diseases, Herpes virus infections and cutaneous
leishmaniasis in which research has begun to indicate that the impact of
UVR-induced immunosupression may play a role..
                            * *  DRAFT FINAL  * * *

-------
                                   9-18
    It has been reported for some time that exacerbation of Herpes simplex
virus infections can be caused by sun exposure.   The following generalizations
have been suggested (Harber and Bickers, 1981).   Each patient's
photosensitivity is site-specific and the action spectrum may be unique to a
patient.  Some have been reported to react to UVC and some to a source
emitting UVC and UVB.   Other patients may exacerbate with heat exposure.  An
exposure of 5 to 6 MED has been reported to cause exacerbation of oral herpes
(Spruance 1985).  It is common to report exacerbation following sun exposure
or UVB phototherapy although the mechanism is unknown (Morison 1984).   In
fact, exacerbation of genital herpes has also been reported to occur following
exposure to the penis to experimental UVR (Klein and Linnemann, Jr. 1986) and
there are anecdotal reports of recurrence following sun exposure which did not
include the genitalia (Dates 1986).

    Although, the mechanism of the exacerbation is unknown, recent
investigations on the effects of UV radiation on primary Herpes infection in
mice are yielding very interesting information.   Low doses of UVR from
sunlamps given 7 days prior to infection of mice with Herpes simplex virus
suppressed the delayed-type hypersensitivity to the virus (Howie et al
1986a).   Once it was induced, immune suppression to the virus persisted for at
least 3 months after the UVR.  The suppression has been shown to be due to two
different subsets of T lymphocytes (Howie et al 1986b).   Thus in this animal
model, exposure to UV radiation at an appropriate time prior to infection,
suppressed the immune response made to that organism for some time and may
have profound effects on the outcome of the infection.

    There have been a few results with regard to the potential role of
UVR-induced immunosuppression in infectious disease progression from studies
using an animal model of an important tropical and subtropical cutaneous
disease, leishmaniasis, .   Giannini (1986) has reported that suberythematous
doses of UVB irradiation to mice blocked the development of the initial skin
lesions when Leishmania major promastigotes were injected.  However, the UVB
also abrogated the development of delayed type hypersensitivity to the
leishmanial antigens.  Protective immunity in leishmaniasis is known to be
cell mediated and delayed type hypersensitivity to leishmanial antigens is
considered indicative of healing.  Therefore, UV irradiation of mice infected
with leishmanial parasites profoundly interferes with immunity.  The
possibility exists that in human leishmaniasis UV irradiation might also play
a role in the pathogenesis of the disease.  Infection of humans through
recently UV-irradiated skin might suppress the immune response to leishmanial
antigens and thus predispose to a more chronic, more severe disease.

UVR and Cutaneous Conditions Other Than  Cancer

    There are a number of cutaneous conditions in which UVB is thought to play
a crucial role.  Although the incidences of these diseases seem small, their
actual impact on human health is magnified because almost all of them are
chronic diseases for which there is no known cure.  Some of them, such as
lupus erythematosus and solar urticaria, encompass a spectrum of severity and
may indeed be several diseases of different etiology but labeled with the same
name, thus studies indicating a broad action spectrum may be misleading.
                          * * *  DRAFT FINAL  * * *

-------
                                   9-19
Because variations in thickness of the ozone layer will have the largest
effects on UVB fluxes, we address primarily the role of these wavelengths in
diseases affect by UVR, even though the greater magnitude of responses may be
elicited by the UVA wavelengths for some of these diseases.

    It is not the intent of this section to be a complete discussion of those
diseases in which UVR may play a role, but to only briefly discuss those
diseases in which there is evidence that UVB plays a role.

Lupus erythematosus

    Lupus erythematosus (LE) occurs in two main forms, discoid lupus
erythematosus (DIE) which is a limited disease with lesions primarily limited
to the skin, and systemic lupus erythematosus (SLE).  SLE is a very
heterogeneous disease with an autoimmune etiology.  Included in the disease
symptomatology is an abnormal reactivity to sunlight, which has been reported
in as many as 58 per cent of patients.  Exposure to sunlight can cause new
skin lesion, exacerbate old skin lesions, and can also induce systemic
exacerbation of the disease (Zamansky 1985).  The patients show a variety of
serum antibodies reactive to many self antigens.  However, the characteristic
antibodies are those to DNA. Patients with SLE have been reported to have
antibodies in their serum that react with UV/irradiated DNA.  Some of the
antibodies will react quite specifically only with UV-irradiated DNA (Davis et
al 1976).  Patients with LE possibly have increased sensitivity to UVB.  There
have been suggestions, although unconfirmed, that the minimal erythemal dose
is at the lower limit of normal (reviewed in Morison, 1983).  It appears that
the skin lesions of lupus were induced more efficiently by UVB than by UVA but
the studies are not definitive.  The fact that many patients are on
immunosuppressive therapy probably complicates results and may be the reason
that not all patients develop skin lesions following ultraviolet radiation.
Eight patients with LE were phototested by irradiation of involved and
uninvolved skin sites with a high pressure mercury arc source.  In 3 or 8
patients the involved skin showed extension of the lesions whereas the
uninvolved skin did not react (Everett and Olson, 1965).  Freeman et al (1969)
studied 15 patients testing the action spectrum of photosensitivity.  Although
not all patients reacted positively, the peak sensitivity for the for the
abnormal photosensitivity was 300 nm, in the UVB range.   Skin lesions were
produced in patients with DLE by UVB radiation. In disseminated OLE, the
sensitivity to ultraviolet radiation extended to 330 nm.  Cripps and Rankin
(1973) showed that UVR at 300 nm caused a long/lasting erythema in the six
patients tested.  Four of those patients had histologic evidence of lupus
erythematosus in those sites from 7 weeks to 6 months after testing.  It was
nessary to administer 8 MEDs to achieve this result; lesser doses had no
effect on the skin.  Sunlight is considered to be a factor in exacerbation of
both DLE and SLE, although all skin lesions may not be on exposed skin
(reviewed in Cripps, 1983).  Cripps (1983) stressed that although
photosensitivity occurs in SLE, under routine clinical conditions, most
patients with DLE do not exhibit photosensitivity.   Lymphocytes from patients
with SLE were found to be more sensitive to the toxic effects of UVC that
lymphocytes from normal patients (Beiglie and Teplitz 1975)
                            * *  DRAFT FINAL  * * *

-------
                                   9-20
    Also apparent is that exposure to sunlight may on occasion exacerbate not
only the skin lesions of SLE but the systemic disease.  In some patients
symptoms usually flare up following sun exposure.  Other patients may show
sensitivity at one time and not another. Exacerbation of SLE for any reason
requires additional immunosuppression for treatment and may be life
threatening (Steinberg, 1985).  However, because of the normal cycles of
exacerbation and remission of SLE, it is extremely difficult to study the
effects of UVR on systemic lesions of SLE.  Murine models of SLE have been
used to study the exacerbation of the human disease after exposure to
sunlight.  One strain, male BxSB mice died prematurely and had acceleration of
their disease after both acute and chronic treatment with FS40/sunlamps.
However, filtering the radiation through Mylar filters, which removes UVB
wavelengths, abrogated the effect.  Therefore, the accelerated autoimmunity
reported in these mice was due to a sensitivity to UVB (Ansel et al 1984,
1985).  Not all strains of mice which are used as models for autoimmune
disease are sensitive to sunlamps (Ansel et al 1985, Strickland 1984) but the
male BxSB mouse should prove to be. a valuable model to study the problem.
Therefore, although the exact interaction between UVB and systemic and skin
lesions of SLE is not clear, there is ample evidence that UVB can play a role
in the disease.

Phototoxicity

    Phototoxic reactions occur when drugs or natural substances interact with
UV radiation or light and result in toxicity.  Action spectra of phototoxic
reactions in humans show that the greatest magnitude of responses are found in
the UVA wavelengths.  However, the data also suggest that wavelengths below
UVA may also contribute to the reaction (Harber and Baer, 1972).

    Phototoxicity associated with chlorpromazine has been reported to occur
primarily with UVA  (Epstein S, 1978, Hunter et al 1970), but UVB has also been
implicated (Hunter et al 1970).  When chlorpromazine is UVB-irradiated and
then injected subcutaneously in experimental animals, it produced toxic
responses.  Therefore, it is possible that at least some part of the
phototoxic responses to chlorpromazine are due to stable photoproducts which
are toxic (Reviewed in Harber and Bickers, 1981, pg. 131).

    Sulfanilamides were the first effective antibacterial drugs to
be/developed.  Sulfanilamides were also the first drugs recognized as being
capable of inducing both phototoxic and photoallergic reactions in humans
(Epstein, S 1939).  The action spectrum has been measured following
irradiation of intradermally injected sites in humans and has been found to be
in the UVB  range (reviewed in Harber and Bickers, 1981, pg. 137).  Blum
(1941) reported that the skin reaction of a patient to irradiation from a
mercury arc source could be prevented by filtering the radiation though window
glass, which removes UVB.  In addition, testing of sulfanilamide for
phototoxicity in vitro using toxicity to human lymphocytes showed that the
phototoxocity was induced by UVB but not UVA (Morison et al 1982).  The same
in vitro testing demonstrated that phototoxocity of hexachlorophene is also
due to UVB and not UVA.
                                 DRAFT FINAL  * * *

-------
                                   9-21
Photoallergy

    Photoallergy differs from phototoxicity in that an immune response in the
cause of the lesions induced by UV radiation or light.  Induction of
photoallergy is generally considered to be due to UVA rather than UVB.  In
fact, low doses of UVB to the site of induction of photoallergy have been
shown to suppress the photoallergic response in a manner analogous to
suppression of contact hypersensitivity in that antigen-specific suppressor T
cells were generated (Morison, 1984).  Also systemic suppression of
photoallergy, not associated with suppressor cells can also be caused by UVB
radiation of cyclophosphamide-treated mice.

Solar Urticaria

    Solar urticaria can be induced by UVB, UVA, and visible light.  It can be
passively transferred by an intradermal injection of serum from a patient into
previously irradiated skin of a/normal volunteer (Epstein, S 1949) or can be
transferred when serum from a patient is injected into the skin of a normal
volunteer and then UV-irradiated (Sulzberg and Baer 1945). Patients can be
classified into 7 groups on the basis of action spectra and immunologic
parameters.  In Type I patients the active wavelengths are found in the UVB
range and it is apparently mediated by IgE whereas in Type V patients the
action spectrum crosses UVB, UVA, and visible light and it is not transferable
by the patients serum (Harber et al 1963, Ive et al, 1965).  It is a rare
disease with somewhat more than 100 patients having been reported in the
literature (Harber and Bickers, 1981).

Polymorphous Light Eruption

    Polymorphous light eruption is a disease of unknown etiology.
Phototesting has led to the realization that it can be induced by UVA, UVB, or
visible light.  However, the papular and plaque-like varieties have an action
spectrum primarily in the UVB range (Harber and Bickers 1981).  Some patients
seem to respond abnormally to UVA (Epstein J 1983).  There is disagreement on
the action spectrum which probably arises at least in part from tests on a
heterogeneous group of patients with a heterogenous group of disorders.
However, most investigators believe that the primary action spectrum which
induces papules is in the UVB range (reviewed in Harber and Bickers, Ch. 14,
1983).  It has also been reported (reviewed in Cripps, 1983) that patients
have a normal minimal erythemal dose in the UVB range with prolonged
erythema.  However, UVA was reported to be more effective than UVB in
induction of delayed erythema (reported in Cripps, 1983).

Actinic Reticuloid and Photosensitive Eczema

    These two disorders are similar in that they occur in elderly males.
Photopatch testing with known photosensitizers has yielded negative results.
In patients with photosensitive eczema, the action spectrum is limited to
UVB.  In patients with actinic reticuloid, a chronic dermatosis, the
                          * * *  DRAFT FINAL  * * *

-------
                                   9-22
sensitivity is first seen in the UVB range but later extends into the UVA
range as the disease becomes chronic.  Patients show a decreased minimal
erythemal dose, sometimes as low as 30 times less than normal (reviewed in
Harber and Bickers, 1981, Cripps, 1983).  It has been suggested that the two
diseases may actually be two phases of the same disease (Morison, 1984).

Bloom's Disease

    The skin lesions of Bloom's disease may exacerbate with sun exposure and
is thought to be due to UVB (Harber and Bickers, 1983).

Pellagra

    The cutaneous lesions of pellagra, erythema followed by desquamation and
hyperpigmentation; may be precipitated by sunlight.  Harber and Bickers (1983)
speculate that lack of NADP and NADP-H may render the patients unable to
repair damage caused by the skin by UVB.

CONCLUSIONS

    Ultraviolet radiation has been shown to have a variety of effects on the
functioning of the immune system.  Acute doses of UVR suppress some immune
parameters in a nonspecific fashion.  Thus this may be similar to the immune
suppression caused by thermal burns.  In addition, UVR can suppress the
induction of contact hypersensitivity and affect the immune response in such a
way that suppressor cells are generated to an antigen to which the animal is
exposed during that time.  One of the more interesting effects of UVR is the
generation of T suppressor cells which specifically suppress the immune
response to UV-induced tumors.  The suppressor cells have been shown to be
capable of shortening the latent period of tumors induced by UVR.  This is a
unique tumor system in which the carcinogen also affects the immune system
such that the growth of the tumors is facilitated in a specific fashion.  Two
recent pieces of information offer a possible explanation as to the mechanism
by which UVR induces such a facilitation.  UVR of mouse skin causes the
appearance of antigens which cross react with those of UV-induced tumors.
This occurs at a time when the mice are susceptible to suppression of contact
hypersensitivity.  Therefore, it is possible that mice which are UV irradiated
and whose ability to mount a contact hypersensitivity reaction is suppressed
in a way which facilitates the induction of tolerance at the same time during
which neoantigens are expressed on irradiated skin.  Therefore the mice
develop tolerance to those neoantigens which are expressed on UV-irradiated
skin as well as UV-induced tumors and are unable to mount an immune response
to the incipient tumors.  This interferes with immune surveillance and the
growth of the tumors occurs unchecked.

    We have no direct evidence that UVR of man also affects the immune
response in such a way that the growth of tumors is facilitated by a
suppressor cell.  However, we do know that UVR of humans does cause
suppression of the immune response and, more important, that induction of
contact hypersensitivity is depressed in human skin which has been
                                 DRAFT FINAL  *

-------
                                   9-23
sun/damaged.  Thus increasing dose of UVR may act to increase the incidence of
skin tumors in man both by acting as a carcinogen and also, in an indirect
manner, facilitating the growth of those tumors by interfering with immune
surveillance.

    There is at present only preliminary experimental evidence from animal
models that UVR can detrimentally affect the immune response to infectious
diseases.  However, it is well to note that these effects were successfully
predicted from the earlier studies on the effects of UVR on the immune
response and the probability is great that this is only the tip of the
iceberg.  There are a wide variety of diseases in which the primary immune
response to the offending organism occurs in the skin.  In some diseases, the
infection remains confined to the skin whereas in others the infection may
become widespread throughout the body.  In evaluating the possible effects of
increased UVR on the populations of the world that we recognize the potential
for increased morbidity that can accrue on a world wide scale.

    The spread and cure of many cutaneous diseases are controlled at least
partially by the immune response.  Some of those diseases, such as
leishmaniasis, may become systemic if they are uncontrolled.  The majority may
cause significant morbidity and loss of function if allowed to progress
without intervention from the immune system or medication.  If UVR of skin
depresses the immune reactivity in the skin/by facilitating induction of
suppressor cells, this could have major impact on the health and welfare of a
great number of people in the world, in particular the citizens of the third
world countries.  If such mechanisms do occur, then increasing doses of UVR
could cause suffering, losses of productivity, and increased health care costs
in the. ensuing years.  At present, there is little research in this area.
                          * * *  DRAFT FINAL  » * *

-------
                                   9-24
                                APPENDIX 9A
RADIATION  SOURCES

    It is appropriate to discuss briefly the sources of UVR used in
experimentation since the characteristics of these lamps play a great role in
the interpretation of results of experiments conducted with them.  A more
complete discussion is made by Spikes (1983).  Solar radiation consists of
broad spectrum irradiation with the lowest wavelengths reaching the earth
being approximately 290 nm.  There is proportionally greater energy in the UVA
range than UVB.  However, it is impractical to use sunlight for
experimentation due to both short and long term variability in intensity and
wavelength.  Therefore, in order to estimate the effects of solar radiation
investigators have used a variety of UV-emitting lamps.  These lamps are good
and usually inexpensive sources of UV radiation; but it is important to
realize the limitations in the precision and comparability of data generated
by use of each light source.   The two major types of light sources are
germicidal lamps and sunlamps.  Germicidal lamps emit about 90 percent of
their radiation at 254 nm with approximately 1 percent at 313 nm, 1 percent at
365 nm, 2 percent at 405 nm, 3 percent at 436 nm, and 0.4 percent at 577 nm.
There are major limitations in the use of such lamps in studies to evaluate
the impact that solar radiation may have on various systems as the UV
radiation produced by these lamps is not principally at 254 nm a wavelength
which does not penetrate the earth's atmosphere.  In addition, these lamps are
not a monochromatic sources since small but detectable amounts of radiation
are emitted in the UVB, the UVA, and the visible range.

    Fluorescent sunlamps emit most of their radiation in the UVB range and so
are better, yet still not perfect surrogates for solar radiation.  A small
amount of the radiation is in the UVC range and there is radiation in the UVA
and visible range also.  The proportion of radiation in the various
wavelengths is not the same as in sunlight because of the relationship of UVB
to UVA and visible light.  Therefore, although the bulbs are called sunlamps,
they are not actually a solar simulator and there is significant irradiation
at wavelengths other than UVB so that when experiments are performed using
sunlamps, such as the Westinghouse FS40 or FS20, one must be careful in
interpreting the data.  These are not pure UVB sources.  Solar simulators and
sources of monochromatic light are available but are quite expensive.  It is
difficult to perform experiments requiring monochromatic light since high
intensity monochromaters usually use a narrow band of light and diffraction
grating to select the desired wavelengths so that only small areas can be
irradiated.  Filters can also be used to control wavelengths; however filters
which allow only narrow wavelengths to pass are expensive, usually small, and
decrease the intensity of the emission significantly.  Plastics can also be
used to remove shorter wavelengths of light; however the transmission
characteristics of the plastic must be periodically monitored since UV
irradiation will change the absorbency of the plastic.  Therefore, in
interpreting the results of experimentation done with UV radiation, it is
always necessary to be aware of the characteristics of the lamp and the
filters used.
                          * * *  DRAFT FINAL  * * *

-------
                                   9-25
                              REFERENCES

Aberer, W, Schuler, G, Stingl, G, Honigsmann, H, Wolff, K.  Ultraviolet light
depletes surface markers of Langerhans cells  J. Invest. Derm.  76:202-210,  1981

Aberer, W, Stingl, G, Stingl-Gazze, LA, Wolff, K.  Langerhans cells as
stimulator cells in the murine primary epidermal cell-lymphocyte reaction:
Alteration by UV-B irradiation.  J. Invest. Derm. 79:129-135, 1982

Ansel, JC, Luger, TA, Green, I.  The effect of in vitro and in vivo UV
irradiation on the production of ETAF activity by human and murine
keratinocytes.  J. Invest. Derm. 81:519-523, 1983

Ansel, JC, Mountz, Steinberg, A, De Fabo, E, Green, I.   Effects of UV
radiation on autoimmune strains of mice:  Increased mortality and accelerated
autoimmunity in BxSB male mice.  J. Invest. Dermatol. 85:181-186, 1984

Ansel, JC, Mountz, J, Steinberg, A, Green, I.  Phototoxicity in autoimmune
BxSB mice.  J. Invest. Derm. 82:391, 1984 (abstract)

Beighlie, DJ, Teplitz, RL.  Repair of UV damaged DNA in sysemic lupus
erythematosus.  J. Rheumatol. 2:149-160, 1975

Bergstresser, PR, Fletcher, CR, Streilein, JW. Surface densities of Langerhans
cells in relation to rodent epidermal sites with special immunologic
properties.  J. Invest. Derm. 74:77-80, 1980

Blum, HF.  Studies of photosensitivity due to sulfanilamide.  J. Invest.  Derm.
4:159-173, 1941

Bowen, RG, Brody, NI. Increased susceptibility to B16 melanoma tumor takes
induced by ultraviolet light.  J. Invest. Derm. 80:370, 1983 (abstract)

Chung, H-T, Burnham, DK, Robertson, B, Roberts, LK, Daynes, RA.  Involvement
of prostaglandins in the immune alterations caused by the exposure of mice to
ultraviolet radiation.  J. Immunol. In Press, 1986

Claas, FHJ, Rothert, M, Havinga, I, Schothorst, AA, Vermeer, B-J, van Rood,
JJ.  Influence of ultraviolet radiation treatment on the survival of
heterotropic skin grafts in the mouse.  J. Invest. Derm. 84:31-32, 1985

Cooper, KD, Fox, P, Neises, G, Katz, SI.  Effects of ultraviolet radiation on
human epidermal cell alloantigen presentation:  initial depression of

Langerhans cell-dependent function is followed by the appearance of T6- Dr
cells that enhance epidermal alloantigen presentation.   J.  Immunol.
134:129-137, 1985.

Cripps, DJ.  Ultraviolet light and photosensitivity.  pp. 116-129, In:  CRC
Experimental and Clinical Photoimmunology, Vol. II, Daynes, RA, Krueger,  G,
Ed.  CRC Press, Boca Raton, FL, 1983
                              *  DRAFT FINAL  *

-------
                                   9-26
Cripps, DJ, Rankin, J. Action spectra of lupus erythematosus and experimental
immunofluorescence.  Arch. Derm. 107:563-567, 1973

Davis, P, Russel, As, Percy, JS.  Antibodies to UV light denatured DNA in
systemic lupus erythematosus.  J. Rheumatol. 3:375-379, 1976

Daynes, RA, Schmitt, MK, Roberts, LK, Spellman, CW. Phenotypic and physical
characteristics of the lymphoid cells involved in the immunity to syngeneic
UV-induced tumors.  J. Immunol. 122:2458-2464, 1979

Daynes, RA, Spangrude, GJ, Roberts, LK, Krueger, GG.   Regulation by the skin
of lymphoid cell recirculation and localization properties.  J. Invest. Derm.
85:14s-20s, 1985

De Fabo, EC, Kripke, ML.  Dose-response characteristics of immunologic
unresponsiveness to UV-induced tumors produced by UV-irradiation of mice.
Photochem. Photobiol. 30:385-390, 1979

De Fabo, EC, Kripke, ML.  Wavelength dependence and dose-rate independence of
UV radiation-induced immunologic unresponsiveness of mice to a UV-induced
fibrosarcoma. Photochem. Photobiol. 32:183-188, 1980

De Fabo, EC, Noonan, FP  Mechanism of immune suppression by
ultraviolet/irradiation in vivo I. Evidence for the existence of a unique
photoreceptor in skin and its role in photoimmunology.  J. Exp. Med.
158:84-98, 1983

De Gruijl, FR, Van Der Leum, JC.  Systemic influence of pre-irradiation of a
limited skin area on UV-tumorigenesis.  Photochem. Photobiol. 35:379-383, 1982

De Gruijl, FR, Van Der Leun, JC.  Follow up on systemic influence of partial
pre-irradiation on UV-tumorigenesis.  Photochem. Photobiol.  38:381-383, 1983

Ebbesen, P. Enhanced lymphoma incidence in BALB/c mice after ultraviolet light
treatment.  JNCI 67:1077-1078, 1981

Elmets, CA, Bergstresser, PR, Tigelaar, RE, Wood, PJ, Streilein, JW
Analysis of the mechanism of unresponsiveness produced by haptens painted on
skin exposed to low dose ultraviolet radiation. J. Exp. Med. 158:781-794, 1983

Elmets, CA, LeVine, MJ, Bickers, DR.  Suppression of dinitrofluorobenzene
(DNFB) contact hypersensitivity by in vivo low dose ultraviolet B (UVB)
radiation:  Action spectrum studies.  J. Invest. Derm. 80:308, 1983 (abstract)

Epstein, JH.  Phototoxicity and photoallergy.  pp. 108-114, In:  CRC
Experimental and Clinical Photoimmunology, Vol. II, Daynes, RA, Krueger, G,
Ed.  CRC Press, Boca Raton, FL, 1983

Epstein, JH, Brunsting, LA, Petersen, MC, Schwarz, BE.  A study of
photosensitivity occurring with chlorpromazine therapy.  J. Invest. Derm.
329-338, 1957
                          * * *  DRAFT FINAL  * * *

-------
                                   9-27
Epstein, S. Photoallergy and primary photosensitivity to sulfanilamide.   J.
Invest. Derm. 2:43-51, 1939

Epstein, S. Chloropromazine photosensitivity, phototoxic, and photoallergic
reactions. Arch. Derm. 98:354-363, 1968

Evertt, MA, Olson, RL.  Response of cutaneous lupus erythematosus to
ultraviolet light.  J. Invest. Derm. 44:133-139, 1965

Faure, M, Schmitt, D, Dezutter-Dambuyant, C, Frappaz, A, Gaucherand, M,
Thivolet, J.  In vitro studies of epidermal antigen-presenting cells.  The
mixed skin lymphocyte reaction: an in vitro model for the generation of
alloreactive cytotoxic T cells by human epidermal cells. Brit. J. Derm.  Ill,
27s:ll-17, 1984

Fisher, MS, Kripke, ML.  Systemic alteration induced in mice by ultraviolet
radiation and its relationship to ultraviolet carcinogenesis.  Proc. Natl.
Acad. Sci. USA. 74:1688-1692, 1977

Fisher, MS, Kripke, ML.  Further studies on the tumor-specific suppressor
cells induced by ultraviolet radiation.  J. Immunol. 121:1139-1144, 1978

Fisher, MS, Kripke, ML.  Suppressor T lymphocytes control the development of
primary skin cancers in ultraviolet-irradiated mice. Science 216:1133-1134,
1982

Fortner, GW, Kripke, ML.  In vitro reactivity of splenic lymphocytes from
normal and UV-irradiated mice against syngeneic UV-induced tumors.  J.
Immunol.  118:1483-1487, 1977

Fox, IJ, Sy, M-S, Benacerraf, B, Greene, MI. Impairment of antigen/presenting
cell function by ultraviolet radiation.  II. Effect of in vitro ultraviolet
irradiation on antigen-presenting cells.  Transplantation 31:262-265, 1981

Freeman, RG, Knox, JM, Owens, DW.  Cutaneousa lesions of lupus erythematosus
induced by monochromatic light.  Arch. Derm. 100:677-682, 1969

Gahring, L, Baltz, M, Pepys, MB, Daynes, R.  Effect of ultraviolet radiation
on production of epidermal cell thymocyte-activating factor/
interleukin 1 in vivo and in vitro.  Proc. Natl. Acad. Sci. 81:1198/1202, 1984

Giannini

Gilchrest, BA, Murphy, GF, Soter, NA.  Effect of chronologic aging and
ultraviolet irradiation on Langerhans cells in human epidermis.  J. Invest.
Derm. 79:85-88, 1982

Granstein, RD. Epidermal I-J-bearing cells are responsible for transferable
suppressor cell generation after immunization of mice with/ultraviolet
radiation-treated epidermal cells.  J. Invest. Derm. 84:206-209, 1985a
                            * *  DRAFT FINAL  * *

-------
                                   9-28
Granstein, RD.  UVR-sensitive and UVR-resistant antigen-presenting cells.
(Editorial).  Photoderm. 2:129-131, 1985b

Greene, MI, Sy, MS, Kripke, ML, Benacerraf, B.  Impairment of
antigen/presenting cell function by ultraviolet radiation. Proc.  Natl.  Acad.
Sci. USA 76:6591-6595, 1979

Gurish, MF, Lynch, DH, Daynes, RA.  Changes in antigen-presenting cell
function in the spleen and lymph nodes of ultraviolet-irradiated mice.
Transplantation 33:280-284, 1982

Gurish, MF, Roberts, LK, Krueger, GG, Daynes, RA.   The effect of various
sunscreen agents on skin damage and the induction of tumor susceptibility in
mice subjected to ultraviolet radiation.  J. Invest. Derm. 76:246-251,  1981

Haniszko, J, Suskind, RR.  The effect of ultraviolet radiation on experimental
cutaneous sensitization in guinea pigs.  J. Invest. Derm. 40:183-191, 1963

Harber, LC, Baer, RL.  Pathogenic mechanisms of drug-induced
photosensitivity.  J. Invest. Derm. 58:327-342, 1972

Harber, LC, Bickers, DR.  Drug-induced photosensitivity (Phototoxic and
photoallergic drug reactions). In:  Photosensitivity Diseases. /Principles of
diagnosis and management.  Ch. 11, WB Saunders, Philadelphia, 1981

Harber, LC, Hollaway, RM, Wheat ley, VR, Baer, RL.   Immunologic and biophysical
studies in solar urticaria.  J. Invest. Derm. 41:439-443, 1963
Hersey, P, Bradley, M, Hasic, E, Haran, G, Edwards, A, McCarthy, WH
Immunological effects of solarium exposure.  Lancet 1:545-548, 1983a

Hersey, P, Haran, G, Hasic, E, Edwards, A.  Alteration of T cell subsets and
induction of suppressor T cell activity in normal subjects after exposure to
sunlight. J. Immunol. 131:171-174, 1983b

Hunter, JA, Bhutani, LK, Magnus, IA.  Cdhlorpromazine photosensitivity in
mice:  its action spectrum and the effect of anti-inflammatory agents.  Brit.
J. Derm. 82:157-168, 1970

Klein, KL, Linnemann, Jr., CC.  Induction of ercurrent genital herpes simplex
Type 2 infection by ultraviolet light.  Lancet i:796-797, 1986 (Letter)

Kripke, ML.  Antigenicity of murine skin tumors induced by ultraviolet light.
J. Natl. Cancer Inst. 53:1333-1336, 1974

Kripke, ML.  Target organ for a systemic effect of ultraviolet radiation.
Photochem. Photobiol. 24:599-600, 1976
                          * * *  DRAFT FINAL  * * *

-------
                                   9-29
Kripke, ML.  Latency, histology, and antigenicity of tumors induced by
ultraviolet light in three inbred mouse strains.  Cancer Res. 37:1395/1400,
1977

Kripke, ML, Fisher, MS.  Immunologic parameters of ultraviolet
carcinogenesis.   J. Natl. Cancer Inst. 57:211-215, 1976

Kripke, ML, Lofgreen, JS, Beard, J, Jessup, JM, Fisher, MS.  In vivo immune
responses of mice during carcinogenesis by ultraviolet irradiation.  J. Natl.
Cancer Inst. 59:1227-1230, 1977

Kripke, ML, Thorn, RM, Lill, PH, Civin, CI, Pazmino, NH, Fisher, MS.  Further
characterization of immunological unresponsiveness induced in mice by
ultraviolet radiation.  Growth and induction of nonultraviolet/induced tumors
in ultraviolet-irradiated mice. Transplantation 28:212/217, 1979

Kupper, TS, Green, DR.  Immunoregulation after thermal injury: sequential
appearance of I-J+, Ly-1 T suppressor inducer cells and Ly-2 T suppressor
effector cells following thermal trauma in mice.  J. Immunol. 135:3047-3053,
1984

Lill, PH, Fortner, GW.  Identification and cytotoxic reactivity of
inflammatory cells recovered from progressing or regressing syngeneic
UV-induced tumors.  J. Immunol. 121:1854-1860, 1978

Luger, TA, Sztein, MB, Schmidt, JA, Murphy, P, Grabner, G, Oppenheim,/JJ.
Properties of murine and human epidermal cell-derived thymocyte/activating
factor.  Fed. Proc. 42:2772-2776, 1983

Lynch, DH, Daynes, RA.  Evaluation of naturally occurring cell-mediated
cytotoxic activity in normal and UV-irradiated mice.  Transplantation
35:216-223, 1983

Lynch, DH, Gurish, MF, Daynes, RA.  The effects of high-dose UV exposure on
murine Langerhans cell function at exposed and unexposed sites as assessed
using in vivo and in vitro assays.  J. Invest. Derm. 81:336-341, 1983

Lynch, DH, Gurish, MF, Daynes, RA.  The effects of ultraviolet irradiation on
the generation of antitumor cytotoxic effector cell responses in vitro.  J.
Immunol. 127:1163-1168, 1981

Milton, JD.  The effect of ultraviolet irradiation on the localization of
syngeneic and allogeneic lymphoid cells and on alloimmunogenicity.
Immunol. 44:281-287, 1981

Morison, WL.  Yearly review. Photoimmunology.  Photochem. Photobiol.
40:781-787, 1984

Morison, WL, Bucana, C, Hashem, N, Kripke, ML, Cleaver, JE, German, JL.
Impaired immune function in patients with Xeroderma Pigmentosum.  Cancer Res.
45:3929-3931, 1985
                            * *  DRAFT FINAL  * *

-------
                                   9-30
Morison, WL, Bucana, C, Kripke, ML.  Systemic suppression of
contact/hypersensitivity by UVB radiation is unrelated to the UVB-induced
alterations in the morphology and number of Langerhans cells.
Immunol. 52:299-306, 1984

Morison, WL, Kelley, SP.  Sunlight suppressing rejection of 280- to 320-nm
UV-radiation-induced skin tumors in mice.  JNCI 74:525-527, 1985

Morison, WL, Kripke, ML.  Systemic suppression of contact hypersensitivity by
ultraviolet B radiation or methoxsalen/ultraviolet A radiation in the guinea
pig.  Cell. Immunol. 85:270-277, 1984

Morison, WL, McAuliffe, DJ, Parrish, JA, Block, K.  In vitro assay for
phototoxic chemicals.  J. Invest. Derm. 78:460-463, 1982

Morison, WL, Parrish, JA, Bloch, KJ, Krugler, JI.   In vivo effect of UV-B on
lymphocyte function.  Brit. J. Derm. 101:513-519,  1979

Morison, WL, Parrish, JA, Woehler, ME, Bloch, KJ.   The influence of
ultraviolet radiation on allergic contact dermatitis in the guinea pig.  I.
UVB radiation.  Brit. J. Derm. 104:161-164, 1981

Morison, WL, Pike, RA.  Suppression of graft-vs-host reactivity in the
popliteal node by UVB radiation.  J. Invest. Derm. 84:483-486, 1985

Morison, WL, Pike, RA, Kripke, ML.  Effect of sunlight and its component
wavebands on contact hypersensitivity in mice and guinea pigs.  Photoderm.
2:195-204, 1985

Noonan, FP, Bucana, C, Sauder, DN, De Fabo, EC.  Mechanism of systemic immune
suppression by UV radiation in vivo. II. The UV effects on number and
morphology of epidermal Langerhans cells and the UV-induced suppression of
contact hypersensitivity have different wavelengths.  J. Immunol.
132:2408-2416, 1984

Noonan, FP, De Fabo, EC, Kripke, ML.  Suppression of contact hypersensitivity
by UV radiation and its relationship to UV-induced suppression of tumor
immunity.  Photochem. Photobiol. 34:683-689, 1981a

Noonan, FP, Kripke, ML, Pedersen, GM, Greene, MI.  Suppression of contact
hypersensitivity in mice by ultraviolet irradiation is associated with
defective antigen presentation.  Immunol. 43:527-533, 1981b

Norbury, KC, Kripke, ML, Budmen, MB.  In vitro reactivity of macrophages and
lymphocytes from ultraviolet-irradiated mice.  J.  Natl. Cancer Inst.
59:1231-1235, 1977

Obata, M, Tagami, H. Alteration in murine epidermal Langerhans cell population
by various UV irradiations:  Quantitative and morphologic studies on the
effects of various wavelengths of monochromatic radiation on la-bearing
cells.  J. Invest. Derm. 84:139-145, 1985
                          * * *  DRAFT FINAL  * * *

-------
                                   9-31
O'Dell, BL, Jessen, RT, Becker, LE, Jackson, RT, Smith, EB.   Diminished immune
response in sun-damaged skin.  Arch. Derm. 116:559-561, 1980

Dates, JK.  Ultraviolet and recurrent Herpes.  Lancet i:1094, 1986 (Letter)

Palaszynski, EW, Kripke ML.  Transfer of immunological tolerance to
ultraviolet-radiation-induced skin tumors with grafts of ultraviolet-
irradiated skin.  Transplantation 36:465-467, 1983

Roberts, LK, Daynes, RA.  Modification of the immunologic properties of
chemically induced tumors arising in hosts treated concomitantly with
ultraviolet light.  J. Immunol. 125:438-447, 1980

Roberts, LK, Spellman, CW, Daynes, RA.  Modulation of immunoregulatory
responses directed toward various tumor antigens within hosts possessing
distinct immunologic potentials.  J. Immunol. 125:663-672, 1980

Roberts, LK, Spellman, CW, Daynes, RA.  Establishment of a continuous T cell
line capable of suppressing anti-tumor immune responses in vivo.  J. Immunol.
131:514-519, 1983

Romerdahl, CA, Kripke, ML.  Detection of UV tumor specific T helper activity
in vitro.  J. Immunol., In press.

Sauder, DN, Noonan, FP, De Fabo, EC, Katz, SI.    Ultraviolet radiation
inhibits alloantigen presentation by epidermal cells:  Partial reversal by the
soluble epidermal cell product, epidermal cell-derived thymocyte-activating
factor (epidermal T-cell activating factor).    J. Invest. Derm. 80:485-489,
1983

Semma, M, Sagami, S.  Induction of suppressor T cells to DNFB contact
sensitivity by application of sensitizer through Langerhans cell/deficient
skin.  Arch. Derm. Res. 271:361-364, 1981

Shevach, EM and Rosenthal, AS.  Function of macrophages in antigen recognition
by guinea pig lymphocytes. J. Exp. Med. 138:1213-1229, 1973

Spangrude, GJ, Bernhard, EJ, Ajioka, RS, Daynes, RA.  Alterations in
lymphocyte homing patterns within mice exposed to ultraviolet radiation.
J. Immunol. 130:2974-2981, 1983

Spellman, CW, Woodward, CW, Daynes, RA.  Modification of immunological
potential by ultraviolet radiation.  I. Immune status of short-term
UV/irradiated mice.  Transplantation 24:112-119, 1977

Spellman, CW, Daynes, RA.  Modification of immunologic potential by
ultraviolet radiation.  II.  Generation of suppressor cells in short/term
UV-irradiated mice.  Transplantation  24:120-126, 1977

Spellman, CW, Daynes, RA.  Properties of ultraviolet light-induced suppressor
lymphocytes within a syngeneic tumor system.  Cell. Immunol. 36:383-387, 1978
                                 DRAFT FINAL  * * *

-------
                                   9-32
Spellman, CW, Daynes, RA.  Cross-reactive transplantation antigens between
UV-irradiated skin and UV-induced tumors.  Photoderra.  1:164/169,  1984

Spikes, JD. Comments on light, light sources and light measurements,  pp.  5-21,
In:  Experimental and Clinical Photoimmunology, Daynes, RA,  Spikes, JD.  Eds.,
CRC Press, Boca Raton, FL, 1983

Spruance, SL.  Pathogenesis of Herpex Simplex Labialis:  Experimental
induction of lesions with .UV light.  J. Clin. Microbiol. 22:366-368,  1985

Steinberg, AD.   Management of systemic lupus erythematosus.   In:   Textbook of
Rheumatology, Vol. 2, Second Edition.  Kelley, WN, Harris Jr., ED, Ruddy, S,
Sledge, CB.  WB Saunders, Philadelphia, 1985

Stingl, G, Katz, SI, Clement, L, Green, I, Shevach, EM. Immunologic functions
of la-bearing epidermal Langerhans cells.  J. Immunol. 121:2005-2013, 1978

Stingl, LA, Sauder, DN, lijima, M, Wolff, K, Pehamberger, H, Stingl,  G.
Mechanism of UV-B-induced impairment of the antigen-presenting capacity of
murine epidermal cells..  J. Immunol. 130:1586-1591, 1983

Streilein, JW,  Toews, GB, Bergstresser, PR.  Langerhans cells:  Functional
aspects revealed by in vivo grafting studies. J. Invest. Derm. 75:17-21,  1980

Strickland, P.  Photocarcinogenesis and influence of UV radiation on autoimmune
disease in NZB/N mice.  JNCI In press in 1984

Strickland, FT, Creasia, D, Kripke, ML.  Enhancement of two-stage skin
carcinogenesis by exposure of distant skin to UV radiation.   JNCI
74:1129-1134, 1985

Swartz, RP.  Role of UVB-induced serum factor(s) in suppression of contact
hypersensitivity in mice.  J. Invest. Derm. 83:305-307, 1984

Takigawa, M, Miytachi, Y, Toda, K, Yoshioka, A. Mechanisms of contact
photosensitivity in mice.  IV. Antigen-specific suppressor T cells induced by
preirradiation of photosensitizing site to UVB.  J. Immunol. 132:1124-1129,
1984

Thorn, RM.  Specific inhibition of cytotoxic memory cells produced against
UV-induced tumors in UV-irradiated mice.  J. Immunol..121:1920/1926,  1978

Thorn, RM, Fisher, MS, Kripke, ML.  Further characterization of immunological
unresponsiveness induced in mice by ultraviolet radiation.  Transplantation
31:129-133, 1981

Toews, GB, Bergstresser, PR, Streilein, JW, Sullivan,  S.
Epidermal Langerhans cell density determines whether contact hypersensitivity
or unresponsiveness follows skin painting with DNFB.
J. Immunol. 124:445-453, 1980
                          * * *  DRAFT FINAL  * *

-------
                                   9-33
Ullrich, SE, Kripke, ML.  Mechanisms in the suppression of tumor rejection
produced in mice by repeated UV irradiation. J. Immunol. 133:2786-2790, 1984

Zamansky, GB.  Sunlight-induced pathogenesis in systemic lupus erythematosus.
Editorial.   J. Invest. Dermatol. 85:179-180, 1985
                          * * *  DRAFT FINAL  * *

-------
Chapter 10

-------
                               CHAPTER 10

                  CATARACTS AND OTHER EYE  DISORDERS
SUMMARY

    Cataracts are opacities  that  develop  in  the  lens of the eye and impair
vision.   In the United States  and other developed  countries, cataract
operations  prevent most cataracts from causing blindness.  However, in the
U.S. cataract remains the third leading cause of legal blindness.  In
developing  countries where such operations are not always available, cataracts
often result in blindness.

    Scientific understanding of the  physical mechanisms which cause cataracts
is incomplete; it is likely that  more than one mechanism operates.
Epidemiological studies.,  laboratory  animal studies, and biochemical analysis
support  the belief that some cataracts are caused  by ultraviolet radiation B
(UV-B).   Ultraviolet radiation A and other causes  are also likely.  A change
in the amount of UV-B radiation is reasonably  likely to alter the  incidence of
cataracts.   UV-B may also play a  role in  causing or exacerbating other eye
disorders.   Ozone modification that  alters the amount of UV-B reaching the
earth's  surface is likely to change  the prevalence  (and incidence) of
cataracts.
                                 DRAFT FINAL

-------
                                   10-2
FINDINGS

    1.   THERE APPEARS TO BE A REASONABLE PROBABILITY THAT CATARACT INCIDENCE
        WILL CHANGE WITH ALTERATION'S IN THE FLUX OF UV-B CAUSED BY OZONE
        MODIFICATION.

        la)  Many possible mechanisms exist for formation of cataracts. UV-B
             may play an important role in some mechanisms.

        Ib)  Although the cornea and aqueous of the human eye screen out
             significant amounts of UV-A and UV-B radiation, nearly 50 percent
             of radiation at 320 nm is transmitted through to the lens.
             Transmittance declines substantially below 320 nm, so that less
             than one percent is transmitted below approximately 290 to 300
             nm.   However, the results of laboratory experiments on animals
             indicate that short wavelength UV-B (i.e., below 290 nm) is
             perhaps 250 times more effective than long wavelength UV-B (i.e.,
             320 nm) in inducing cataract.

        Ic)  In laboratory animal experiments, the action spectrum for
             cataracts is weighted heavily in the UV-B range.

        Id)  Human cataract prevalence appears to vary with latitude and UV
             radiation; brunescent nuclear cataracts show the strongest
             relationship.                    •  .

    2.   INCREASES IN THE AMOUNT OF UV-B THAT CAN REACH THE RETINA APPEAR
        CAPABLE OF CAUSING STABLE RETINAL DISORDERS AND RETINAL DEGENERATION,
        TCP CAUSES OF BLINDNESS.

    3.   UV-B MAY PLAY A ROLE IN DISORDERS OF THE EYES AS WELL AS IN
        DEVELOPMENTAL DISORDERS.
                          * * •-  DRAFT FINAL  * »

-------
                                   10-3
CATARACTS1

    The  longest standing  hypothesis which may account for the development of
senile cataracts is  that  radiant energy,  particularly sunlight,  is a major
factor  in the etiology of the  disease.  This concept apparently originated
from observations  reported by a  number of individuals indicating that
cataracts occurred more frequently  or earlier  in persons whose occupations
kept them  outdoors or that populations  living  in areas  with  longer hours of
sunshine have  a  higher frequency of cataract than populations from areas where
there is  less sunshine.  While the early studies were severely flawed by their
failure to consider adequately the possible effects of a variety of
socioeconomic and  other variables,  Duke-Elder (1926,  1972)  proposed that "the
fundamental cause  of  cataract in  all its  forms  may be traced to the incidence
of radiant  energy  directly on the lens  itself."

Definition of Cataract

    Cataract is defined  as an opacity  in the normally transparent lens of the
eye which  produces an  impairment of vision.   Cataracts may occur as  a  result
of a wide variety of factors  including metabolic disorders,  exposure to toxic
agents, trauma,  exposure to radiation,  and  hereditary factors.  The great
majority  of cataracts, however,  are  the so-called senile cataracts which  occur
in older  individuals and for  which  no specific causative factor can be
identified.   Cataract is  a  major cause of visual impairment and blindness
particularly in developing countries where access to modern surgical
facilities is  limited.  Even in the United States, cataract is  the third
leading cause of  legal blindness.  Some 60 percent of people aged 60 to 74
have at least  some cataractous changes  in their lenses.   The only efficacious
treatment for cataract at present is the surgical removal of the  opaque lens
and  nearly 660,000 such operations  were performed  in  1982  in the U.S.

    It  is clear therefore that senile  cataract is a very  significant health
problem, both  in terms  of its impact on the affected  individuals and on
society  at  large.    While considerable progress  has been made  in elucidating
the biochemical etiology of certain specific cataract types,  such as sugar
cataracts,  there  is  little conclusive  data on the causes of senile cataracts.
It is likely that there are a  variety of potential risk factors and that, in
general, senile cataracts have a multifactorial  etiology.  This  conclusion is
supported  by the great variability  observed clinically in the time of onset,
the rate of maturation,  and the morphological  appearance and  location  within
the lens of these  opacities.   It appears that many,  if not all,  of the
processes contributing to  senile cataractogenesis are normal aging processes
which for whatever reason are accelerated in  certain individuals.

Cataract Classification

    There  are  now rather sophisticated  systems for the classification of
cataracts.   Chylack et al.  (1978, 1983)  have devised an  in vitro system which
    1This section in boldface type is taken from Pitts, D.G. et al. (in
press).


                          * * *  DRAFT FINAL  •• * *

-------
                                   10-4
is  based on photographic documentation  of ©pacification and nuclear color.
Almost 2500 cataracts  have been studied and classified since the original
methods  were adopted (Chylack  et al.,  1983).   In addition, Marcantonio et al.
(1980) have suggested a system of classification of human  senile cataracts by
using photography for in  vivo and in vitro analysis and determining the sodium
and protein content of the extracted  lens.  The photographic method  gave
consistent  involvement of the lens nucleus but  it was not always possible to
relate sodium  changes to light scattering.   Marcantonio et  al.  (1980) argue
that a dual classification system is  necessary because most cataracts are
mixed in nature and the osmotic and  nuclear mechanisms provide quite different
changes  in protein distribution.

    The  present state of knowledge concerning the human  senile cataract
probably justifies a classification into only two  major types  in spite of the
elegant  classification  systems mentioned  above.   The first and  most common
type of human senile cataract is the  cortical cataract.   Cortical cataracts
are characterized by imbalances  in  cation levels within the lens cells which
produce  osmotic swelling and ultimate opacity in the lens cortex.   Altered
cation balance may result  from damage to the Na + ,  K*-ATPase  or from compromise
of the normal permeability  characteristics of the lens membranes.   Any of a
great variety  of potential  insults could  be the ultimate cause of such
cataracts.   The second  major type of senile cataract is  the  nuclear cataract
which is characterized primarily by very marked modifications to the
structural  proteins of the lens,  the crystalline,  in  the central  region  of the
lens.   These  two  types  of  cataracts are not at  all mutually  exclusive;  in many
instances,  senile cataracts  contain both  cortical  and nuclear opacities.   It
is  advisable for those who  are involved  in cataract research to  become
knowledgeable and use a recommended classification  system.

Epidemiological  Studies

    A few  general  statements should  be  made relative to the cataract
epidemiological studies.  Most epidemiological  studies have  reported an
increase in the prevalence  of cataracts  with  an increase in  age  and at 65
years of age  and  above there is an acceleration in  the prevalence  of
cataracts.   Women  have shown a  larger  prevalence  of cataracts  than men;
however, this difference has not been shown to be statistically significant.

    In recent years there have been  a  number  of epidemiological studies
reported which have attempted to establish that there  is an association
between  cataract prevalence and  exposure to sunlight or the ultraviolet
component of  sunlight.  Hiller, Giacometti and  Yuen (1977) utilized data on
the populations sampled in two independent health  surveys, the Model Reporting
Area  for Blindness Statistics (MRA)  and the National Health and Nutrition
Examination Survey (HANES) to compare the  prevalence of  cataract in
geographical  areas with  varying total annual hours of  sunlight.  The total
annual hours  of sunlight was obtained from U.S. Weather Bureau statistics and
ranged  from 1800 to 3800  hours.   For each health survey ocular diseases  other
than cataract were used as control and  the data reported as age-specific
ratios of cataract cases  to  control cases for  areas of differing  total annual
insolation.  The findings suggest that in the youngest  age group  studies
(20-44 yrs) there is  no greater  incidence of cataract in  areas of high  annual
sunlight, but that with  increasing age  there appears to develop an increasing

                          * * *   DRAFT  FINAL   * * *

-------
                                   10-5
association of cataract with sunlight exposure.   In the 65-74 yr  sample
population,  there was at least a doubling of the ratio of cataract to each
control  disease between  the lowest  and highest sunshine area and in persons 75
years and older this trend was even  more pronounced.   The study  did not
consider the possible effects  of genetic or socioeconomic differences among
the populations  from different areas nor did it consider the actual  sunlight
exposure of the  individuals (e.g.  indoor  vs  outdoor  occupation).

    Zigman,  Datiles, and Torczynski  (1979)  have studied populations from three
widely separated areas (Manila, the Philippines; Tampa,  Florida; and
Rochester, New York) which differ considerably in the yearly  levels of
ultraviolet  radiation in  sunlight.  This study considered not only the age of
the individuals studied, but  also considered separately those with indoor  and
outdoor occupations.  The  findings which  were based on  study of extracted
cataracts included  analysis of the  cataract type for each  lens.   The results
indicated that no correlation  existed between geographic location and
distribution of cataracts except for the brunescent cataracts; i.e.,  those
nuclear cataracts  with significantly increased pigmentation levels.  In
Manila,  the  area  with greatest UV, such cataracts accounted for 43% of total
cataracts while in Tampa 20% of cataracts were brunescent and in Rochester,
the area of  least UV,  only 9%.   In  all three  populations,  the percentage of
brunescent  cataracts extracted in  persons with outdoor occupations  was
markedly higher  than  in those persons working indoors.  Thus,  both the
latitudinal variation and the  individual differences within  geographic regions
suggested a  strong relationship  between UV exposure  and brunescent  cataract.
These data  are consistent with numerous  reports that  tropical  areas have
higher cataract  incidence than areas  at higher latitudes and that the
percentage of brunescent  cataracts is higher in tropical  areas  (Pirie 1972  and
references  in Taylor 1980).  The confounding factors which  were pointed  out by
Zigman,  Datiles,  and Torczynski  (1979) as not being controlled  in their study
included the economic,  nutritional and genetic backgrounds of  the  individuals
in the respective populations.  While these differences would probably be
substantial,  particularly between  Manila and the two U.S. sites,  this  study is
significant in that it involved some biochemical evaluation of the cataracts
as well  as the epidemiological data.

    Two recent  epidemiological studies have  concerned cataracts  in Australian
aborigines,  a rural population exposed to  relatively  intense solar radiation.
Taylor  (1980) studied 350  individuals  from which detailed personal  histories
were  obtained.   The possible role of a variety of personal and environmental
factors  in cataractogenesis were investigated.   Among the 350  individuals, all
of whom Were over  30 years  of age,  116 had lens opacities  as  determined  by
slit-lamp examination.   The major findings relative to the possible influence
of radiation exposure  to cataractogenesis  were  a trend toward  association  of
cataracts with increased hours of  sunlight and with  higher annual  mean UV-B
levels.  These trends were reflected by a strong association of cataract with
lower latitudes.   No other  environmental  factors studies appeared to be
associated with  cataract nor  did any  personal  factors other than age.  No
correlation would found between occupation  and cataract; it is not clear
whether this finding is  inconsistent with  Zigman, Datiles,  and  Torczynski
(1979) since it is not  reported whether any  of the individuals  studied  had
                                 DRAFT FINAL  * * *

-------
                                   10-6
indoor occupations.   Unfortunately,  the types of cataracts present were not
reported in this  study (Taylor 1980).

    In a second  study involving a much greater geographical  area and over 50%
(64,307) of the total Aborigine population, Hollows and  Moran (1981) also
found a statistically significant correlation between environmental UV
irradiation and the prevalence of cataract  (p  0.005).  Additionally cataracts
occurred much more frequently in the younger age group (40-59 yrs) in
geographical zones of high UV radiation than in zones of low  UV irradiation.
The  authors suggest that the  Aborigines from remote rural  Australia are a
suitable population for such studies  since  their lifestyle  tends to be  highly
uniform throughout  the country.   In contrast to these results,  a large sample
of non-Aboriginal people from  the same areas showed much  lower levels  of
cataract,  particularly  at  younger ages, and there was no correlation between
cataract prevalence  and sunlight  in  this group.  This was attributed to the
much higher standard of living and  the greater likelihood of indoor
occupations in this group.

    The data  of  Hollows and Moran (1981) may be used to estimate the number of
senile cataracts which are caused  by sunlight if we can  accept the confounding
factors  mentioned above.  Exhibit 10-1 presents the  prevalence  of cataracts
found in aborigines  by UV zone for  three different age groups.  Exhibit 10-2
compares  the  prevalence  of cataracts found in non-aborigines  and aborigines
who  live in the same regions.

    Exhibit 10-1  shows that 13.6% of the aborigines in the lowest solar region
develop cataracts while about  30%  in the more solar intense zones  develop
cataracts.   These data appear to indicate  that about  15% of the  senile
cataracts  are  due to solar exposure.   If the aborigine and non-aborigine are
compared  (Exhibit 10-2)  there was about a 13% increase  in cataracts  which  are
probably  due to  solar radiation.  More recently, Weale (1982)  has presented  a
method  of  estimating the risk  attributed to light for  the incidence of senile
cataracts  and reported a risk  factor of 5.  If the risk factor  may be
expressed in  percent  form  or  20%, the results are  not too different  from the
above.  It is  interesting that  the  two sets of diverse data present such
interesting results.   Incidentally,  the 18.3%  prevalence for  cataracts for the
non-aborigine compared quite  favorably to the Framingham Study  (Kahn et al.
1977a and  b)  and the Gisborne Study (Martinez et al. 1982).

    Thus,  there appears to be a consensus from these epidemiological studies
in support of the notion  that  senile  cataract or at  least  a particular  segment
of the heterogeneous  mixture  of opacity types which  comprise senile  cataracts,
is associated with higher exposure to  sunlight.  Substantive questions could
certainly  be raised concerning each  of the studies  cited  above;  however, taken
in aggregate they represent a variety of approaches, using different types  of
populations, different criteria  for cataract, different  sampling and
statistical  methods,  and different variables to test the same hypothesis.  It
is striking that the  general conclusions of each study are so  similar.
                                 DRAFT FINAL

-------
              10-7



         EXHIBIT 10-1

  Cataract Prevalence by UV Zone
Age
UV Zone
1
2
3
4
5
0-39
0
0
0.1%
0.1%
0.2?;
40-59
1
2
3
3
5
-TO'
. I/O
6O'
/o
7°'
. / /o
80'
/o
.1%
60+
13
24
29
30
29
.6%
.2%
CO,
. -> '0
.5%
.8%
 Source:   after  Hollows and Moran 1981,
         EXHIBIT 10-2

 Comparison of Cataract Prevalence
for Aborigines  and Non-Aborigines
                         Age
  Ethnic Group     50-59      60+
Non-aborigine
Aborigine
0.8%
4.4%
18.3%
29 . 3%
 Source:   after  Hollows and Moran
          1981.
            DRAFT FINAL

-------
                                   10-8
Ocular Transmittance

    The effect of  nonionizing radiation on a cell  depends on the specific
chemical composition within the cell, that is, on  the presence  of absorbing
molecules  or  chromophores  (Lerman 1980a).   This type of radiation must be
absorbed  in order to cause a change  in the molecule since  absorbed energy  is
required to promote a chemical  change.  Molecules in  excited electronic states
have different chemical and physical properties than their  counterparts in the
ground  state  (prior to absorption  of energy).   Thus  cells  that do not contain
chemical compounds  absorbing at certain wavelengths  will transmit these
wavelengths.   For example, the nucleic acids and most proteins in  a  cell  are
essentially transparent to and completely transmit visible light but  absorb
certain wavelengths in the UV  region (between 250  and 295 nm) and  can  be
damaged by this form of  radiation, while other macromolecules  in a cell such
as rhodopsin, which  absorbs at 498 nm,  and hemoglobin, which has absorption
peaks in the UV and visible region (275, 400,  and 540 to 576  nm), appear
colored  since  they absorb visible light.  These latter  macromolecules  can  be
damaged by high intensities of visible radiation at their  specific absorption
wavelengths.

    The transmittance data for the rabbit, primate and human  eye and ocular
media for  the wavelength  range of  200nm to 2500nm are given  in Exhibits 10-3,
10-4 and  10-5.  Exhibits  10-6 and  10-7 present transmittance  data for the 200
to 400 wavelength range  (Kinsey 1948; Boettner and Wolter 1962; Maher 1978;
Barker  1979).

The Action Spectrum

    Since  the absorption  of nonionizing radiation  is  determined by the
chemical composition of the tissue  being  exposed, the more radiation  that the
molecules  absorb the greater will be the effect of the radiation.  The term
"action  spectrum" is  used as a measure of the relative effect of different
wavelengths  of  radiation  on a  chemical compound, macromolecule, cell, or
entire organism.  The  action spectrum is a  plot  of the dose or radiant
exposure  necessary to produce the defined  effect versus the wavelength.   For
example,  the  maximum  efficiency for experimental photokeratitis has been shown
to occur at approximately 300 nm,  with a smaller peak at 295  and  320 nm (Pitts
1978 and  Pitts and Cullen 1981).  The action  spectra  for photokeratitis,
cataracts  and retinal lesions for the rabbit, primate and  human are presented
in Exhibits 10-8 and 10-9.

    The eye  is  the only organ  or  tissue  in the body  (aside from the  skin)  that
is particularly sensitive to the  non-ionizing wavelengths  of radiation  (longer
than 280  nm)  normally present in  our environment.    In addition to infrared  and
visible radiation, we are  constantly exposed  to ultraviolet radiation (solar
and  man-made)  throughout life.  It is estimated  that  approximately 8% (11
mW/cm2) of solar radiation above the atmosphere is in the ultraviolet region
(280-400 nm).  At sea level this is decreased  to 2-5  mW/cm2, depending on
geographic location and season (Lerman 1980b).
                            * *  DRAFT  FINAL

-------
                                     10-9
                               EXHIBIT 10-3

             Composite Transmittance  Curves  for  the Rabbit With
         Representative Data for  Boettner and Wolter (human,  1962),
           Weisinger  (rabbit,  1956) and Kinsey (rabbit,  1948) for
                       Comparison (after Barker,  1979)
                                       COMPOSITE:
                                       1. at the Aqueous
                                       2. at the Lens
                                       3. at the Vitreous
                                        at the Retina
10  -
        250  300   350   tOO  450   500   550  600   650   700  000   1100   1300  1500  1700   1900  2100  2300  2500
                           * * *  DRAFT FINAL  * *

-------
                                   10-10
                               EXHIBIT 10-4


                  Calculated Total Transmittance of the Human Eye

                        (after Boettner and Wolter 1962)
  100
  80
  60
o
Z
in
  40
Z
LU
O
o:

£20
                                                                     I  I II
TOTAL TRANSMITTANCE  AT THE

VARIOUS ANTERIOR SURFACES

                3   VITREOUS

                4   RETINA
                       I   AQUEOUS

                       2   LENS
         300      400    500  600

                         WAVELENGTH
                800   1000 1200

                MILLIMICRONS
1600  2000
                           * * *  DRAFT FINAL  * *

-------
                      10-11
                 EXHIBIT 10-5

Percent Transmissivity Through the Entire Rhesus Eye
                (after Maher 1978)
                      1.2    I.H   I.E    I.B    2.0    2.2   2.H
                 *  DRAFT FINAL  * *

-------
                                          10-12
                                    EXHIBIT  10-6

        Transmittance  of the  Total Rabbit  Cornea, the Total  Human Cornea,
            and  the Rabbit  Corneal Epithelium  (after Kinsey  [1948]  and
                             Boettner and  Wolter  [1962])
i.o

0.9


0.8


o.;
O RAM IT CORNEAL INTHELIIM. (I MET

X RAUIT TOTAL COME*. KIKSEY

O DAM IT TOTAL CORNEA. SACHEM

• MMAN TOTAL CORNEA. METTNER AND NOLTERI
 0"	,	1	1	1	¥—	f=	f	—1	—1	1	1	1	1	"	1	1	1	r-
    250   210   250   260   270   280  290    300 .  310   320   330   310   350   360   370   380    590   100


                                       WAVELENGTH IN KANOHETERS
                               * *  *  DRAFT FINAL  * *  *

-------
    1.0


   0.9


   0.8
_J

j| 0.7-

^J
^"\
   0.6


   0.5-


   0.4
Z

LJ
tr
   0.3-
   Q2-
   0.1-
  o.o
                                   10-13
                              EXHIBIT  10-7

        Transmittance of the Anterior Ocular  Structures of the Human
          and Rabbit Eyes.   The  Symbols  Indicate the Percentage of
               Radiant Energy Incident On a  Certain  Structure
                    (after Pitts,  Hacker,  and Parr  1977)
         OINCIOENTON AQUEOUS, HUMAN
         X INCIDENT ON LENS, HUMAN
         • INCIDENT ON VITREOUS, HUMAN
         D INCIDENT ON AQUEOUS, RABBIT
         A INCIDENT ON LENS, RABBIT
         7 INCIDENTONVITREOUS.RABBIT
      270  280  290  300 310   320  330  340  350  360   370  380  390  400

                   WAVELENGTH IN NANOMETERS
                          * * *  DRAFT FINAL  * * *

-------
                                  10-14



                               EXHIBIT  10-8

           UV  Radiant  Exposure Threshold  Data  for the Cornea H ,
      Lens H.  Cataracts,  and Retina HR  for the Rabbit and  Primate
WAVELENGTH
nm
210
220
230
240
250
260
270
280
290
295
300
305
310
315
320
325
330
335
340
345
350
355
360
365
370
375
380
385
390
395
405
441
CORNEAL LENS
Radiant Exposure Radiant Exposure
H in J/cm2 HT in J/cnn2
c
Rabbit
0
0
0
0
0
0
0
0
0

0
0
0
7
7
18
30
30
--
--
--
50
--
65
--
--
--
--
--
--
--
- -
.17
.046
.03
.033
.04]
.018
.005
.11
.012

.022
.07
.05
.25
. 5
.0
.0
.0



.0

.0








Ll
Primate Rabbit Primate
0
0
0
0
0
0
0
0
0

0
--
0
--
9

41
--
58
--
61
--
88
--
130
--
170
--
258
--
--
- -
.33
.021
.022
.012
.020
.011
.004
.006
.007 3.00
0.75
.01 0.15 0.12
0.30
.02 0.75
4.50
.6 12.60
50.00
.1

.3

.5

.4



,





RETINA
Radiant Exposure
H0 in J/cm2
K
+Rabbit -Primate










0.225 0.12



5.0





5.4





8.1



15.0
30.0
"v for 100s exposure duration  in  aphakic primate producing retinal lesion.

+ for 650s exposure duration  in  phakic rabbit eye producing changes in the
retina.
                            *
                                 DRAFT FINAL  * *

-------
    300.0  -


    250.0  -
    220.0  -
I
O  200.0  -

X
N
I
    100.0  -

     50.0  -
 LU
 cr
 D
 CO
 O  10.0
 Q_
 X   so
 UJ
 §   ••«
 <   o,
 Q
 O
 CO
 cr
 x
O.I


c.os





0.01


 .005





0.001
                                           10-15
                                        EXHIBIT 10-9

                    The  Action Spectra for Photokeratitis and Cataracts
                      for the Primate  and  Rabbit (after Pitts 1981)
                    &    COGiN AND KINSEY
                    O    KURTIN AND ZUCIICH
                    • v • » PITTS,CULLEN AND BARKER


                    »    HUMAN CORNEA
                    V    PRIMATE CORNEA
                    •    RAB3IT CORMEA
                    •    RA6SIT LENS
             —I	1	;	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	|	
             210  220  230  240 2SO  260  270 280 290  SOO HO  320 310  340  J50  360 370 380  390  400
                             WAVELENGTH   IN  MANOMETERS
                                         DRAFT FINAL  * *

-------
                                           10-15
                                        EXHIBIT  10-9


                    The Action Spectra  for  Photokeratitis  and Cataracts
                      for  the Primate  and Rabbit (after Pitts  1981)
    300.0  -



    250.0  -
 O  200.0


 X

CM
    100.0  -
 LJ
 cr

 00
 o
 CL
 X
 LJ
 0
 <
 cr
 o
 X
 00

 cr
 x
     50.0 -
10.0  -
 50  -
0.5  •








O.I  •



C.05








0.01



 .005








0.001
                    £t    COOiN AND KINSEt

                    O    KURTIN AND ZuCLlCM

                    • v • » PITTS,CULLEN AND BARKER


                    1    HUMAN CORNEA

                    V    PRIMATE  CORNEA

                    •  .  RiBBiT CORNEA

                    •    RABBIT LENS
            —I	1	:	1	
             210  220  210 240
                       -i	:    i	1	1	1	1	1	1	:	1	1    i	1    |   |—
                       :50  260  270 280 290 300 110 120 310  140  150 360 170  380  390-  400


                       WAVELENGTH  IN NANOMETERS
                                 * * *  DRAFT FINAL  * *

-------
                                   10-16
    Nature has provided us with transparent ocular media which are essentially
avascular and contain very few  visible  wavelength absorbing chromophores in
order to effectively transmit  (as well as refract)  the  specific wavelengths
required to initiate  the  visual process by photochemical reactions.   However,
these tissues do  have the ability to absorb varying amounts of  ultraviolet
radiation (particularly the ocular lens). The shorter the wavelengths  of
radiation absorbed the  greater the  potential for photic damage  since there  is
an inverse relationship  between  a wavelength and the photon energy association
with it.  Thus, UV radiation  is  the non-ionizing  portion of the
electromagnetic spectrum which  could cause the most  damage, provided that it
is  absorbed.   This  axiom applies to all  the ocular tissues including the
retina  in the very young eye where the lens has  not as yet become as effective
UV filter but,  in  particular,  the ocular lens sustains the greatest  amount of
photochemical change during  a lifetime  of exposure  to ambient UV  radiation.

Some Biochemical  Mechanisms

    There  are  several mechanisms which are biochemically  related to radiation
damage, to  the  eye and the lens  in  particular.   These mechanisms include
photo-oxidation of free  and protein bound tryptophan,  photosynthesis processes
involving the activated  species of oxygen,  disruption of the cation  transport
system  and damage  to the nucleic acids (DNA) of the lens epithelium.  Some of
these mechanisms  have only been studied only in  cultures,  some in vitro and
others in vivo.   It  is important to  be able  to project the mechanism to a
"real live" situation in  order to be  able to  evaluate the relative importance
of the factors  involved  in inducing the senile cataract.   Prior to discussing
these mechanisms  it may be desirable to review some  of the basic concepts in
the biochemical mechanisms of radiation  induced damage; therefore, some
current interpretations  of lens free radicals and  oxidation  reduction
reactions  relative  to cataract follows.   Exposure to  UV  radiation initiates
enzymatic activity involved in cellular protection  from oxidative processes
which may be due to both  light and metabolism (Exhibit 10-10).  Catalase
destroys hydrogen  peroxide,  which is  produced  in  all cells by metabolism;
superoxide dismutase (SOD)  has the  responsibility  of destroying the superoxide
radical, a  very toxic radical  due to its powerful  oxidative nature.   In the
lens cortex,  there is very  little SOD or catalase,  but they are  highly
concentrated in the epithelial cells.   Of all the ocular tissues,  the greatest
concentrations  of  these  enzymes are present in the retina.   In  the lens cortex
there are other agents  to protect against oxidation  such as glutathione  (GSH)
and  asorbic acid.   There is only a  small amount of  vitamin  E in the lens, but
attempting to protect animals  from  ocular tissue oxidative damage by feeding
them high  levels  of vitamin E or other  anti-oxidants has not succeeded.   The
sum total of  all of the anti-oxidants that are present in  the ocular tissues
still does not protect against the formation of free  radical oxidative
reactions to  near-UV radiation.   Another toxic oxidant  that can form  is
singlet  oxygen.

    The content of  the  stable free  radicals previously described is highest  in
the cortex and seems to decrease in  the  nucleus, of the lens.   The process  of
aggregation of soluble proteins  (TSP) occurs both  in the cortex and in the
nucleus.   Because of the conservative  growth process of the lens,  much of the
                                 DRAFT FINAL  * * *

-------
                                  10-17



                             EXHIBIT  10-10

               Free Radials and Oxidation:  Reduction Systems
                 SOD

           EPITHELIUM
          UV
                                                         CORTEX
                                                         NUCLEUS
                                                         LOW FREE
                                                          RADICAL
                                                         GSH

                                                         ASCORBATE

                                                         HIGH  FREE
SOD represents  superoxide dismutase, GSH represents  glutathione.  Epithelial
SOD and catalase  protect these cells against oxidation  from excessive hydrogen
peroxide and superoxide anion.  High levels of ascorbic acid and glutathione
protect cortical  fibers against oxidation.   TSP represents soluble and INS
insoluble protein.  TSP converts to INS protein both in the cortex and in the
nucleus.  But INS accumulates more in the nucleus  than  in the cortex.  Free
radicals are quenched  in nucleus due to binding to proteins.
                         * * *  DRAFT FINAL  *  *  *

-------
                                   10-18
aggregated  material remains in  the  nucleus  indefinitely and  often in a form
that is associated with  the fiber cell membranes.  The lens  fraction called
the insoluble fraction contains both membranes and  the  aggregated proteins.
There  is a chemical reaction between the  free radicals and proteins so that
the protein  binds to free  radicals chemically, which quenches the ESR
(electron  spin  resonance)  signal.  Multiple  molecular species of  proteins can
be associated through new crosslinks leading to  aggregation and, therefore,
light scattering.   In the nucleus of the lens, the free radicals are quenched
by their reactions  with  proteins.  Since all enzymes are proteins,  this may
illustrate  a  universal process whereby enzyme activity is  inhibited  by near-UV
radiation.   These changes would eventually cause malfunction of all cells and
tissues.

    Exhibit  10-11  shows a scheme of a series of  enzymatic oxidation:
reduction reactions going  on  in  all  cells that  are influenced by  aging  and
exposure  to near-ultraviolet radiation.  These reactions are all  linked
together.   The co-factor NADH/NADP  system is  linked to the oxidation  and
reduction of glutathione (GSH).  When oxidation occurs,  reduced GSH becomes
oxidized  GSSG, changes which  also relate to protein oxidation involving
SS-crosslinks.   Many enzymes,  in order to remain  active, must  maintain
sulfhydril groups (SH)  of cysteine  in  reduced form in order for the protein
portion to maintain  its  activities as  an enzyme.   Many enzymes require free-SH
groups  in their active  sites to  retain their  activity.   If these enzymes are
oxidized  so  that  disulfides form, they  are inactivated.  For example,  if high
concentrations  of hydrogen peroxide are generated,  protein SH  will convert to
SS thereby  inhibiting  the activity.   Several of the  enzymes referred to in
Exhibit 10-11 are known to be  reduced with enhanced aging in  most living
cells.  In many cases  the cells turn over, so that  cells'  enzymatic  activities
are maintained, but where there are populations of cells that do not turn over
very actively,  a  cumulative process can  lead  to a great loss of  enzyme
activities.   This  may occur in the lens.

    One enzyme  that is diminished  with aging  is glutathione reductase and it
also diminishes under  in vitro circumstances in living tissues when near-UV
radiation  is provided in the presence of excess tryptophan.   There has not yet
been an aging  or photoinactivation  study of glutathione peroxidase.   The SOD
is  very important as a  protective agent because  it destroys the very  toxic
superoxide  radical.  SOD  activity does diminish  with  aging,  but is not
sensitive  to destruction from near-UV  radiation (with tryptophan as a
sensitizer).  Because the  lens  substance  contains little  oxygen to  serve as a
superoxide  source, the  site of  lens damage would be  the  epithelium if SOD
activity were to  be diminished  drastically.  Damage to the lens  epithelium
would certainly lead to  both physiological and developmental anomalies in the
lens and cataract.   The SOD activity must  be maintained  in the  retina,
however,  due  to its high  oxygen tension  and the potential for oxidation to
occur more  readily.

    Hydrogen  peroxide  (H202)  is another very powerful oxidant that  is present
in  ocular  tissues and fluids by  virtue  of  the action of SOD  on superoxide and
other metabolic systems, such as the glutathione and ascorbic acid cycles.
Recently, high levels  of H202 (mM  concentrations)  have been
                          *.* *  DRAFT FINAL  *  *  »

-------
                                     10-19



                                EXHIBIT  10-11

                Enzyme Systems  Involved  in Oxidation:  Reduction
 U2o
CATALASE
CoA-SH / PROTEIN-SH
       '
GSSG
             *

          /OX I OAT I ON on

          ll't-nOXIDATION
          \l CoA-SS-f.oA / PROT

              H202  / ROOK
                 FOOT,
NADPH
            NADPt
6-PHOSPHATE-
GLUCOHOLACTONC
                                                             TO
                                      PENTOSE
                                      SHUNT
                D-GLUCOSE-6-
                  PH03PHATC
                   GLUT ATM I ONE   GLUT ATM I ONE    GLUCOSE 6-PHOSPIIATE

     DISMUTASE      PCHOXIHASE    KEHUCTASE      DEIIYDROGENASE
                                             UCUUCEP WITH AGINQ
                                             RIJ fine tin iiy iiv •» THY
  Enzyme activities  that  are  decreased with aging are shown by solid bars and
  those  decreased  by light  exposure  (artificially induced) are shown by the open
  bars .
                           *  - *  DRAFT FINAL  * * *

-------
                                   10-20
observed  in the human aqueous  humor by Garner and Spector  (1980).  This
concentration  of H202 is suspected of being capable of causing cataracts  by
poisoning important enzymes and by crpsslinking  proteins (Spector and Garner
1982).   In eyes without lenses,  H202 from the aqueous humor could more easily
diffuse  even to the retina (Zigman 1981,  Kramer 1980),  a highly hazardous
circumstance.   Ocular tissues  utilize the enzyme catalase to destroy and
detoxify H202.  Since this enzyme is diminished with aging  and has been shown
to be inhibited by  near-UV radiation plus tryptophan (Zigman,  Yulo,  and Griess
1976),  its activity  loss would allow heightened levels of H202 to damage
ocular  tissues.  A  large imbalance in favor of H202 accumulation would lead to
extensively  altered ocular tissue  proteins and  much enzyme inactivation.

    Glutathione reductase  is another enzyme of great functional importance to
the lens and retina,  and it is  diminished with  aging  and UV exposure
(Kalustian,  Sun,  and Zigman  1978).   Its  loss of activity would estimate
oxidation  of small  molecules and  proteins  in most ocular tissues.

    Another enzyme  (not involved in oxidation: reduction, but significant with
regard  to its  photoinactivation)  is Na+K+ ATPase.  This  enzyme  is also
sensitive to near-UV radiation in the presence of tryptophan and loses its
activity accordingly.  In the  lens, it maintains the osmotic balance by
controlling the sodium and potassium cation exchange and prevents water
inhibition and  swelling.   In some genetic cataracts in mice,  it has been found
that the cause of cataract is the high  concentration of an inhibitor of ATPase
in the  lens  (Kinoshita 1974).   It is likely then that a strong inhibition of
lens epithelial  cell  ATPase by  light-sensitized action  could be a major  factor
in cataract formation osmotically.  In the retina, the ATPase activity is very
important in terms  of the chemical reactions that  support the visual process.
Should  this  enzyme be photochemically inhibited  in the retina,  the functioning
of the  visual process would be markedly  reduced.   Retinal Na*K* ATPase is  also
sensitive to inhibition by near-UV plus tryptophan.

    Before discussing the biochemical studies which pertain to the possible
cataractogenic  effects of optical  radiation, it may be pertinent  to review
some contrasts between  human lenses and those of the most commonly used
laboratory animals.   All vertebrate  lenses are  composed essentially of  a
single  cell type,  the lens fiber,  which differentiates from a single layer of
epithelial  cells present only at the bow or equator of the lens.   New fibers
are continually laid down at the periphery; thus, the oldest tissue is  located
at the  center  or  nucleus of the  lens and the  cells become progressively
younger as  one moves from the  center toward the lens capsule  in the  cortex.
Cells are  never sloughed from the lens and this makes the cells  in the lens
nucleus as old as the animal.  Furthermore, differentiated lens fibers
gradually lose  virtually  all cell organelles,  including nuclei,  and lose the
capacity to  synthesize protein as they age  and are forced toward the  nucleus.
Therefore, the proteins, primarily lens crystallins, present in the aging
human  lens  may be the  longest-lived proteins  in the organism.   This means that
unlike  most other tissues the central portion  of the lens does not have the
ability to replace damaged proteins with  newly synthesize proteins.   This may
be  an  important factor in the  presumptive long-term effects of chronic
                          * * *  DRAFT FINAL  *  *  *

-------
                                   10-21
exposure to near UV radiation.  There are  very clear differences between the
lens nucleus and cortex in terms of  their biochemistry and the loss  of the
capacity to  repair or replace damaged  proteins probably accounts for much  of
the difference.

    There is at least one very  significant difference  between human  lenses  and
the common  laboratory  animals and that is the  presence of  pigmented compounds
in the human  lens.  While the lenses of most non-primate mammals have  no
pigment,, the lenses of  diurnal primates and a  few other strongly diurnal
species are  yellow.  Cooper and Robson (1969) demonstrated that in the human
lens there are  two  classes of pigmented compounds.  One group that is  present
even  before birth is of low molecular weight,  is water soluble and absorbs
maximally at about  365  nm.  A  second  class of colored  compounds appears  later,
increases with  age,  is bound to the  lens proteins and absorbs maximally near
320 nm.  This  latter class of chromophores  is  localized primarily  in the lens
nucleus and may be responsible for  the age-related increase  in pigmentation in
human lenses.   Since these chromophores absorb in the near UV, they may be
major determinants of the  effects of  such  radiation on the  human lens.

Studies at the  Biochemical Level

    Studies  at  the  biochemical level  have generally been concerned with  the
structural modifications which crystallins undergo during aging and
cataractogenesis and have attempted to explain these reactions  in terms  of
photo-oxidative mechanisms.  It is well-documented that crystallins,
particularly in  the  lens nucleus, accumulate a  variety of modifications.
These include the formation of  disulfides and  other covalent  crosslinks,  the
development of a novel blue fluorescence, progressive pigmentation,  oxidation
of methionine,  racemization of aspartate residues, polypeptide chain
cleavages,  deamidations,  aggregation and ultimate insolubilization.  The  lack
of turnover of protein  in the lens nucleus  accounts for  the accumulation of
these modifications and they  have  been the  subject of several recent reviews
(Zigler and  Goosey 1981; Hoenders and Bloemendal 1981; Harding 1981).   It is
clear that most of the protein changes are the result of oxidative stress, and
the possible role of radiation in that stress has been the subject of  much
study over  the  last 15  years.

    Pirie (1968, 1972)  studied the effects of sunlight on  solutions of lens
proteins and of other  proteins.  The proteins  became brown  following
irradiation  and analysis of  absorbance  changes indicated  similarity to those
found in lens proteins  from cataracts.   Pirie also found decreased levels of
the oxidation sensitive  amino acids,  histidine and tryptophan in protein  from
cataracts relative to normal  lenses.  Although  the data for tryptophan was
subsequently retracted as an artifact,  Dilley and  Pirie (1974) suggested that
photooxidation  of tryptophan residues,  perhaps with  formation of N'-formyl
kynurenine, was a primary step in cataractogenesis.   These studies stimulated
a number of other  investigators.

    Kurzel  (1973) performed  fluorescence and  phosphorescence measurements  and
Weiter and Finch (1975) ESR studies on human lenses and  found  signals which
they  believed to be due to tryptophan free  radical species  in lens proteins.
                          * * *   DRAFT FINAL  * * *

-------
                                   10-22
Van  Heyningen  (1971) was  able to identify several of  the components
contributing  to  the color of human lenses as kynurenine derivatives, species
which can be derived  from  tryptophan either  metabolically or
photooxidatively.  These were part of the  low molecular weight colored
material present in human lenses  and Van  Heyningen  (1973) subsequently showed
that  lens proteins exposed  to sunlight in the  presence of these compounds were
photo-oxidized more extensively than in their absence.  The  mechanisms of this
accelerated photo-oxidation  was not determined.

    In addition  to oxidation of protein bound  tryptophan other possible
mechanisms were explored.   Zigman et al.  (1973)  and  Zigman and  Vaughan  (1974)
demonstrated  that photo-oxidation of  free tryptophan  yielded pigmented and
fluorescent species which would bind to lens crystallins in  vitro.  This
raised the possibility that the target of photo-oxidation could be either free
or protein-bound tryptophan.   Numerous investigators turned to the study of
brunescent nuclear cataracts since these lenses have  the greatest
concentration of the pigment and  of  the non-tryptophan fluorescence.   The
search for clear decreases  in the levels of tryptophan in the proteins  of such
lenses or in  their free amino acid pool has not been successful to date
(Dilley and Pirie 1974; Pirie and  Dilley 1974;  Zigler et al. 1976).  It should
be noted however that tryptophan comprises less  than 2% of total  amino acid  in
crystallins and, thus, small changes  would be difficult to  detect especially
in view of the inherent  problems  of tryptophan analysis.

    Lerman (1980b)  suggests that in  the normal lens  less than  20% of  the
protein tryptophan is  susceptible to photo-oxidative damage.   Studies  on  the
novel blue fluroescence  of  aging  crystallins, particularly the insoluble
fraction from brunescent lens  nuclei  suggest the presence of a number of
related  species  (Lerman 1980b).  The species with emissions  in the  visible are
generally more concentrated in nuclear cataracts.   There is some  evidence that
this  fluorescence may  be concentrated in certain crystallin  polypeptides
(Zigman 1981).  It has also been  demonstrated that the formation  of
non-disulfide covalent crosslinks  between crystallin polypeptides  is
associated with  the heavily pigmented protein fraction  and  can  be generated  in
vitro by irradiation  (Buckingham  and Pirie  1972).  Additional photoproducts of
tryptophan,  including B-carbolines (Dillon, Spector and Nakanishi 1976) and
anthranilic acid (Truscott,  Faull,  and Augusteyn  1977) have  been identified
from the crystallins of nuclear cataracts.   Dillon and  Spector (1980),  Dillion
et al.  (1982)  and  Borkman, Tassin,  and Lerman  (1981) have  studied the
photolysis of  free tryptophan,  or tryptophan  containing peptides, and of
isolated lens  crystallins. Analysis of these data suggests that a variety of
products are  possible  and that the microenvironment of individual tryptophan
residues is of paramount importance.  Additionally, photolysis  is much faster
in the presence of oxygen.

    Harding  and Dilley  (1976), however, raised two objections  to  the  idea that
sunlight caused brown nuclear cataracts.   First,  they pointed  out that the
lens  damage  is  in  the  nucleus whereas the shortest wavelengths reaching the
lens; i.e., those which  might be  absorbed by tryptophan  are probably absorbed
in the anterior  lens cortex.  Indeed  as noted above cataracts induced in
animals by UV-B are located in the anterior cortex.   Secondly, the  lack of
                          * * »   DRAFT FINAL  * * *

-------
                                   10-23
evidence for loss of tryptophan  in brown cataracts was cited.  While  these
arguments  are  difficult to rebut in terms of mechanisms in  which UV  oxidation
of tryptophan  is the central event, there is another  mechanism of
photo-oxidation for  which  these  objections may be less significant.

    Recently, there has been increasing interest in  the possible role  of
photosensitized processes, particularly with the involvement of activated
species of  oxygen,  in the oxidative damage observed  in the human  lens.  This
work  was spurred by data demonstrating light-mediated lens  damage with such
photosensitizing drugs as  8-methoxypsoralen,  phenothiazines,  and tetracycline
and by the studies  cited above by Pirie and Van Heyningen showing  accelerated
UV effects on  crystallins in  the  presence of kynurenines and  related
compounds.  Zigler  and Goosey  (1981) have demonstrated that several of these
compounds endogenous to  human  lenses are capable  of generating singlet
molecular oxygen, a  highly reactive  species capable of damaging proteins as
well as other biological molecules and structures.  The ability of this
oxidant to  induce the oxidative changes  characteristic of human crystallins
has been established in vitro.  It has also been demonstrated that  such
photosensitizing activity is present in the heavily altered insoluble  lens
crystallins  from brunescent cataracts and to a lesser  extent in the  soluble
crystallins  of normal human  lenses as well.  Based on experiments  in vitro,
such  a photosensitized process  could account  for the  generation of  each  of  the
oxidative modifications presently  known to occur in  lens  crystallins.
Furthermore such a process seems consistent  with nuclear  localization of
damage.  In the lens  nucleus there  is no repair or replacement of altered
molecules,  thus allowing progressive  accumulation  of  crystallins with oxidized
residues some  of which are photodynamic sensitizers.   The UV-A which these
species absorb will  readily penetrate  to the nucleus.   Additionally while  the
lens cortex is  protected by a battery of antioxidant defenses  it is known that
these  defenses are markedly  reduced in  the nucleus  (Hata  and Hockwin  1977;
Fecondo and Augusteyn 1983).

    One could  envision a system  in which there was  a slow rate of
photo-oxidation ongoing in the  lens controlled  by  antioxidant defenses and  by
the greatly reduced oxygen  tension  in the lens nucleus.  The  oxidative  stress
might  have several  components  including  direct UV oxidation of tryptophan
(primarily  UV-B),  the high  levels of H202 in  the aqueous humor  (Spector and
Gardner 1982)  and photosensitized oxidation involving activated species of
oxygen.  It  has  been recently demonstrated that even fetal human  lenses
contain low molecular  weight  chromophores which can  generate singlet oxygen
when  irradiated with UV-A.   The initial photochemical event  could  be
absorption  by  such  chromophores or  direct photo-oxidation of tryptophan to
produce N-formyl kynurenine (NFK)  a known  photodynamic sensitizer.   In  either
case,   it seems  likely that  the continued  build-up of oxidized  products in  the
nucleus is  likely due  primarily to a  sensitized  process, since such  a process
would not  require large-scale tryptophan loss  nor would  it require  penetration
of UV-B into the  nucleus.  A relatively small  number of  stable sensitizing
species such as NFK, bound  within the  long-lived nuclear  crystallins, could
continue to generate singlet oxygen  indefinitely with the gradual accumulation
of the various  oxidative protein  changes outlined above culminating in
aggregation  and  insolubilization  of much of the nuclear crystallins in
advanced brunescent  cataracts.
                                 DRAFT FINAL

-------
                                   10-24
    Several suggestions have been made that  attempt to reconcile the finding
that tryptophan is not reduced in sinile cataracts (as compared  to normal
lenses)  with  the hypothesis  that  it acts  as an important endogenous
photosensitizer.   For example,  it has been suggested that tryptophan  is
oxidized to a  reactive molecule  that simply transfers  its energy  to some other
cell component from the exicted state.  It then  returns to the ground  state
without being destroyed in the process, yet photosensitizes damage  to other
lens macromolecules (Weiter  and Finch 1975; Weiter and Subramanian 1978).
Another explanation is that  free tryptophan rather than protein bound
tryptophan may be destroyed when it acts as a photosensitizer, whereas
tryptophan incorporated within lens  proteins is  unaffected.   The level of free
tryptophan may be reduced  in  cataract  (Zigler et al. 1976).  The failure of
protein-incorporated  tryptophans to  be  photo-oxidized  might  be  explained  by
the fact that they lie in very  different  micro-environments that  render them
less susceptible to photo-oxidation (Lerman,  1980b).   A third suggestion to
explain  the finding that tryptophan  is not decreased in cataractous  lenses is
that only a small  percent of the protein-incorporated tryptophan may be
susceptible to photo-oxidation  and small losses in these susceptible
tryptophan might  not be detectable with current analytical capabilities.
Tryptophan and its photoproducts have  become  the  most frequent culprits
implicated  in  causing photo-oxidative changes in lens proteins, changes that
are hypothesized  to be responsible for  senile  cataractogenesis.

    Photo-oxidation of  lens proteins  has become a major topic in lens research
and has been shown  to be induced by photosensitizers that are  either
endogenous (Zigler and Goosey 1981) or externally applied (Goosey, Zigler,  and
Kinoshita 1980) to the  lens.   One effect frequently found to  occur  as  a  direct
effect of UV  or as an effect of UV plus a photosensitizer  is crosslinking of
lens proteins  (Buckingham and Pirie  1972; Goosey, Zigler, and  Kinoshita 1980).

    Interest  in agents that can cross-link or  otherwise  aggregate lens
proteins has  been high since Benedek (1971)  first proposed that the mechanisms
underlying senile  cataract formation  was aggregation and insolubilization of
lens crystallins.   They suggested that  aggregated lens proteins would serve as
light scattering particles in  the lens.   The  search was then  underway for
these high molecular  weight  aggregates, but unfortunately,  the research
produced very inconsistent  and conflicting results.   Some research  groups
found an increase in the amount  of high molecular weight proteins  in senile
cataractous lenses while other groups found  no difference in the amount of
high molecular weight proteins  in cataractous lenses compared to that  present
in lenses in age matched controls (Harding  and Dilley  1976).  It was also
suggested  that aggregation  was an artifact  resulting from the extraction
procedure  (Harding 1972).  Analogous  studies have more  recently been carried
out on  lenses in which the cataracts  were separated according  to the nature of
the changes  present  in the  lens.  The  results of those studies  have suggested
that nuclear  brunescent, but not cortical cataract, is associated with
formation of  increasing amounts of water insoluble proteins (Truscott and
Augusteyn 1977).   Some researchers  have attributed this  increase to
cross-linking  of lens proteins  induced by photo-oxidation.   It has been
further suggested that the  reason these changes  occur more readily in the lens
nucleus than  in the cortex is that the cortex has a higher concentration of
protective  anti-oxidants.


                          *  *  *  DRAFT  FINAL  * * *

-------
                                  10-25
    If  Benedek's (1971)  original hypothesis that the mechanism underlying
senile  cataract formation was  aggregation  and insolubilization of  lens
crystallins is  correct, then one would expect to find an increase in protein
cross-linking  and aggregation in  lenses that show light  scattering.   To the
contrary, lenses with  light scattering opacities in the cortex do not show an
increase in high  molecular weight proteins or crosslinks (Truscott and
Augusteyn 1977; Anderson, Wright,  and Spector 1979).   Instead,  reports  in the
literature suggest that there  may be an increase in  protein  aggregates  in
lenses  with substantial nuclear brunescence.  This is  rather a surprise
because  nuclear brunescence  per se  is not a light scattering change in the
lens;  rather,  the  lens turns  a yellow or a brown color that  absorbs rather
than scatters  light,  reducing its  intensity on the  retina (Lerman and Borkman
1976).   The effect of  brunescence is, therefore,  analogous to placing a
transparent filter before the  eye.

    Unfortunately, only man  and  a very few other diurnal species have been
found  to develop lens browning so that it has been  difficult to find an  animal
model  in which to test the hypothesis that browning can be  induced in  vivo
by exposure  to UV radiation.   Therefore,  those studies that have examined the
effects of chronic UV  exposure in experimental animals have not demonstrated
that UV  induces  lens  browning.  They  have, however, consistently demonstrated
that UV  exposure induces  light-scattering cortical opacities  (Bachem 1956;
Zigman and Vaughan  1974;  Pitts, Hacker,  and  Parr  1977).   Therefore,  though we
may  lack a suitable model  for  nuclear brunescence,  there are animal models
that can be used to study the  induction of cortical opacities by  UV
radiation.  These models mimic the common  and considerably more visually
disturbing senile cortical cataracts in man.

    In those  studies on  UV-induced  lens changes  in which histology of the
lenses  were examined, epithelial cell  changes were a  uniform finding (Zigman
and Vaughan  1974; Pitts,  Hacker, and Parr  1977).  In their study of
chronically exposed mice,  Zigman and Vaughan (1974) noted the similarity of
the lens changes to changes  induced  by X-irradiation.  For example, lens cells
appeared to have  lost their capacity  to  differentiate and  were found to have
migrated to the posterior pole, a situation that has  also been found to be
associated  with  senile cortical cataracts (Streeten  and Eshaghian  1978).

    While biochemical approaches  have generally concentrated on  nuclear
cataracts with  respect to UV-mediated effects,  there has been recent interest
in photosensitized reactions  in  the aqueous humor as  sources of damage to  lens
membranes and,  hence,  as a  possible initiator of cortical (osmotic)
cataracts.  Varma, Kumar, and Richards  (1979) have  been investigating the
possibility that a  photochemical conversion of molecular oxygen present in the
aqueous  humor and lens into  superoxide and subsequent derivatization of
superoxide to other potent oxidants  such  as hydrogen peroxide,  hydroxyl
radical and singlet oxygen may be involved  in  initiating a cascade of toxic
biochemical reactions  leading  to the formation of cataracts.   Thus,  according
to this hypothesis, cataractogenic influence of  light is mediated  by a
photochemical generation of superoxide  from the ambient oxygen.   Spin
restriction offered by the  molecular oxygen makes the formation of superoxide
a necessary event in most oxidation  reactions involving  oxygen.  It is
                                 DRAFT FINAL  * * *

-------
                                  10-26
commonly understood that,  like many other  free radicals,  superoxide and its
derivatives,  if allowed to remain unscavenged for any length of time in a
biological mi leu, will  initiate many  nonspecific and deleterious  reactions
such as  an upsetting of the normal redox chain, oxidation  of vitally important
protein  and nonprotein - SH,  peroxidation of membrane  and cytosolar  lipids,
and polymerization  and depolymerization of macromolecules such  as proteins and
hyaluronic acids.   The cataractogenic influence of these oxidants is  likely to
be modulated  by certain endogenous protective mechanisms.  Superoxide
dismutase (SOD),  catalase and perioxidase constitute the first line of  defense
against  the toxic effects of those oxygen  species.  Jernigan et al. (1981)
showed  that  singlet oxygen generated in the medium had similar effects.
Varma,  Beachy, and Richards (1982) have demonstrated lipid  peroxidation in
cultured lenses  irradiated  with fluorescent light.  While such studies are of
great interest in terms of  the effects of these  oxidants on  the  lens,  at the
present time  there  is no real evidence that  they are produced insignificant
quantities in the ocular  humors by light mediated processes.

    Considerable effort has been directed at elucidating mechanisms  by which
UV radiation (alone or  in combination with exogeneous photosensitizers)  might
induce cataract  formation.   As previously mentioned, it  has been, shown that  UV
radiation can  induce  cross-linking  and aggregation of proteins.   As  has also
been mentioned,  protein aggregation may  be associated with nuclear changes but
it  is not a characteristic associated with  cortical cataracts.  It may
therefore be  more relevant to consider the possibility that UV damage to lens
proteins induces local perturbations in macromolecular structure and
function.  This  might in turn result in  localized changes in lens structure of
function.  For example, transport  systems and ATPase (Varma,  Kumar,  and
Richards 1979) in the lens  have  been shown to be  inhibited by  UV  radiation.
If  the lens transport systems  are  damaged by  UV radiation, the osmotic balance
will be  disrupted which, in turn,  would produce major changes  in lens
morphology.

    The UV effects on  transport systems  may occur  due to direct UV  absorption
by and  damage to the related enzymes;  however,  other  mechanisms  may also
contributed to this inhibition.  For example, the activity of transport
enzymes could be dramatically altered secondary to  UV induced  disruption of
lipids in lens membranes;  and, lipid peroxidation has been proposed as  a
mechanisms  of cataract formation (Goosey, Allison, and Garcia 1983).  The
resulting alterations  in lipid structure could of themselves make the
membranes leaky or, as just  mentioned, such alterations could  disrupt the
structure and function of  membrane proteins.  If one can extrapolate  from .the
large amount  of research that has  been done on  other cell systems,  then it may
be predicted that fairly long wavelengths of UV  will similarly  be able  to
produce membrane  damaging effects in the lens (Moss and  Smith  1981;  Imbrie and
Murphy  1982; Sprott, Martin, and  Schneider 1976).

    If we are to determine  the role of lipid peroxidation or inactivation of
specific enzyme systems in the development  of UV cataracts, it  is critically
important that lenses be sampled  several  intervals prior to the onset and
during  the development of  lens opacities.   Once the lens  is fully opacified,
                              *   DRAFT  FINAL  * *

-------
                                   10-27
retrospectively, it  is difficult  to state with  any assurance that some
observed biochemical change was the  specific  change that initiated the
cataract.

    For example, if an ATPase were found to be inactivated after a  particular
UV-induced cataract was fully developed, one might conclude  that this
inactivation had caused  cell  swelling,  membrane lysis, and  an  "osmotic
cataract."   But it is also possible  that the primary cause of the cataract was
direct  membrane damage or inhibition of protein synthesis and, that, only
secondarily was ATPase activity involved.   Therefore,  to show that  ATPase
inactivation is  of any real significance in the  etiology of the cataract, it
is  necessary to demonstrate  that ATPase inactivation precedes development of
the lens opacities.   In the case of UV-induced cataracts, one  also needs  to
demonstrate that those wavelengths that inactivate the ATPase are those  same
wavelengths that most readily  induce  cataract formation.

    It  may be  misleading to  examine effects of UV  (with  or without added
photosensitizers) on  lens macromolecules isolated in  test tubes because  UV
effects in vitro may be  very different from the effects of UV  on  the  intact
lens.   For  example, it is possible  to cause tryptophan  destruction in lens
proteins  irradiated at a  concentration of 2.5 mg/ml in a test tube, but  the
same type  of irradiation was not found  to produce such effects at the
concentration of protein in  the lens (Dillon  et al.  1982).  Also, the action
spectrum may differ  considerably from the absorption spectrum (Turro  and
Lamola 1977).   The  action spectrum shifts  markedly in situ in the case of
the photosensitizer, psoralen.   Psoralen,  riboflavin,  methylene blue  and  rose
bengal have all been found to induce photo-oxidative changes  in  isolated lens
proteins when  these  proteins are irradiated with UV.   However,  each of  these
photosensitizers requires the presence of oxygen, otherwise lens proteins are
not damaged.   The  amount  of  oxygen in the lens  is not high  (Kwan,  Ninikoski,
and Hunt 1972; Barr and Roetman 1974).  Thus,  these  findings must be
extrapolated to the  in vivo lens with  considerable discretion and  it remains
to be determined just how much photodynamic damage (that is damage where
oxygen is  involved as an intermediate)  can  be produced in the lens  in  situ.
This is a very important research  question  because any effect that is of
significance as a mechanisms of cataract formation must be  inducible  in a Jens
that is still in  the  eye.   For example, it might  be  interesting  to  use  254  nm
radiation as a  toll in vitro to demonstrate that the lens epithelium has the
capacity to repair damaged DNA (Jose and Yielding 1977).   However, 254 nm
radiation does  not penetrate the cornea, so it will  not  be hazard  for  the  in
vivo lens and  it will not cause cataracts.  Therefore, it is a far  more
significant  finding  that  300  nm radiation can induce DNA repair synthesis in
the lens epithelium when the lens  is exposed  in situ through the intact cornea
(Jose,  Kock, and  Respondek 1982; Brenner and Grabner 1982).  This finding
shows  that SOOnm  UV radiation penetrates through the cornea and demonstrates
that epithelial  cell DNA  is a target of  its effects.

    Therefore, if  we are to  establish  mechanisms of UV damage in the lens, we
must be able to show that a given damaging reaction can be produced by UV in
the in  situ  lens.   An effect that can  be produced on  lens macromolecules
isolated in  a test tube may be interesting and stimulate further research, but
                            * *  DRAFT  FINAL  * * »

-------
                                   10-28
it must not be taken as sufficient proof that such  an effect can occur in
vitro.   This is especially true for those reactions that require oxygen.

    One photosensitizer that may provide important clues  as to mechanisms
involved in cataract formation  is  the drug psoralen  which  in combination with
UV  radiation is used in treating  psoriasis.  Two different mechanisms have
been proposed to  underly the  psoralen  cataract.  According to one hypothesis,
the cataracts  occur  as  a consequence of psoralen binding  covalently to lens
proteins which results  in formation of additional damaging photosensitizers in
the lens (Lerman, Megaw, and Willis 1980).  The second hypothesis is that the
cataracts occur as a consequence of psoralen  induced damage  to lens nucleic
acids (Jose and Yielding 1979).

    It  has  been found  that intense exposure to UV  radiation  will cause
psoralen to bind to  isolated  lens  proteins, specifically the tryptophan
moieites  (Lerman,  Megaw,  and Willis 1980);  however, the  reaction  requires
oxygen (Megaw,  Lee,  and Lerman 1980).   Therefore, it is important  to determine
whether such protein binding occurs  in  the intact  lens.   Although spectra have
been presented that demonstrate  that UV radiation  induces psoralen binding in
the in  vivo lens,  it  is  not clear  that the  binding observed represents protein
specific  binding or whether  the  binding is to some  other lens  macromolecule.
Lens protein fractions  have  been extracted from psoralen-treated  humans or
laboratory  animals and  fluorescence spectra taken from these.  These spectra
were interpreted to  demonstrate photobinding of psoralen  to lens  proteins;
however, no effort was made in  these studies to extract nucleic acids from
these "protein" preparations.  It is very possible that a significant amount
of the  binding that  was observed is accounted for  by psoralen binding to RNA
or DMA  present in the degenerating nuclei  of the lens cortex.  The  possibility
that the spectra represented binding to nucleic acids in the differentiating
fibers  was discounted by the researchers because  there is such an overwhelming
concentration of protein compared to DNA in lens fibers  (Lerman, Megaw, and
Gardner 1982a).   This  is a reasonable assumption if a  significant  percentage
of the  psoralen were actually bound to  the protein.  However, the argument
would  be nullified if the majority of the psoralen were found to be bound  to
the small amount of  nucleic acids in the fibers.  Autoradiograms  showing
psoralen binding in  the lens suggest that the latter is in  fact the case
(Lerman et al, 1981).   Examination of those autoradiograms shows pronounced
binding  of psoralen  in  the epithelium plus discrete  binding to the
degenerating fiber nuclei.  In comparison,  binding  to any other portions of
the fibers  is not distinguishable  from background grains.

    Autoradiography has also demonstrated psoralen binding  in the nucleated
layers  of the retina  and cornea but no  binding to  the  non-nucleated  layers of
these tissues (Lerman  et al, 1981).  It is rather difficult to reconcile the
failure of psoralen to bind to the non-nucleated regions of the retina and
cornea with the suggestion that it binds  significantly to cortical fiber
proteins.  A trivial  explanation for this difference  is that lens proteins  are
somehow "different" from corneal and retinal proteins, making the former
susceptible to  psoralen binding and the latter  not.   This  argument is rather
difficult to reconcile with the requirement of oxygen for photo-induction of
                          * * *  DRAFT FINAL  * *

-------
                                  10-29
psoralen linking to proteins,  a  requirement that would  lead one to expect
greater binding to the  proteins in the highly  oxygenated retina and cornea
compared  to the relatively anoxic lens.

    If  lens proteins are the  major target of psoralen's action  on the lens,
then one would expect  that the cortical fibers would  be disrupted as a  primary
event.   In fact, however, histologic observations  have  shown that the first
target  of psoralen's effects are the epithelial cells (Jose, Kock, and
Respondek 1982).   Only after very considerable damage is observable in the
epithelium  is any  damage detectable in the lens cortex which  is consistent
with DMA as a primary target of  psoralen  damage.  That psoralen plus  DVB can
induce DNA damage in  the lens is shown by the finding that the  combination
induces DNA repair in  lens epithelial  cells  (Jose and  Yielding 1979).  It will
be interesting to  resolve the question as to the relative roles of specific
macromolecular targets  in development of psoralen-induced cataracts.

    As with psoralen, most other studies of UV effects on the lens have
directed their major emphasis toward  examinations of changes in lens
proteins.   But,  as with the  psoralen  study, there are compelling  reasons to
consider other macromolecules a targets of UV damage.   Lens lipids have been
mentioned previously and interest in  lipids peroxidizing effects is  of current
interest in many  laboratories.   However, interest  in the effect of UV and
photosensitizers on lens DNA is currently  very limited.  The role  of DNA
damage in  development of lens  cortical opacities is generally overlooked or
rationalized away.  Some have assumed  that there is  so little DNA in the lens
that even if it were a target,  any effects on  it would be overwhelmed by
effects on  lens proteins.   There  are, however, good reasons to consider the
lens to be  similar  to  other biological systems in which DNA is a significant
target  of  UV damage.

    If  one  considers  the  action spectrum that  has been determined  in those
studies in  which  specific wavelengths of UV were  isolated  to  examine their
cataractogenic potential,  it is noted that these lie in  the range  of 290 to
320 nm (Bachem 1956;  Pitts,  Hacker,  and  Parr 1977).  These are the same
wavelengths that  most  readily produce thymidine dimers in skin (Pathak,
Kramer,  and Guengerich  1972).   Shorter wavelengths are  blocked  from  deeper
layers  of the  skin by the stratum corneum, quite  analogous  to corneal UV
absorption which  protects the lens.   These are the same wavelengths which most
readily induce malignant  transformation and skin  cancer (Freeman  1975).   UV
radiation up to 320 nm also can induce  DNA repair synthesis in cultured
fibroblasts (Ichihashi and Ramsey,  1976).   Longer wavelengths of UV do not
induce skin cancer, although they may,  if applied very intensely,  induce
strand breaks and other  damage  in DNA (Webb and  Peak 1981,  Harm 1978).
Endogenous photosensitizers are  likely to be involved since DNA  itself does
not absorb appreciably at that  wavelength.  Long wave UV may potentiate the
action  of short wave  UV  in cancer induction and it can also be noted that 364
nm has been found to inactivate  DNA repair mechanisms (Tyrell 1976).

    With all the interest  in tryptophan as  an  endogenous photosensitizer of
protein damage in  the lens,  it  is  interesting to point out parenthetically
that photo-oxidation  of tryptophan has  also been  found to induce DNA  damage in
                            *  *   DRAFT FINAL  * *

-------
                                  10-30
lower organisms as a consequence of production of H202 (Ananthaswamy and
Eisenstark,  1976).   The possibility  that those free radical species that have
been found  to be generated  from UV excitation of tryptophan  in the lens may
also  act on  lens DNA as a target should  not be overlooked.   Furthermore,
tryptophan  photoproducts  have also been  found to bind to DNA (Glazer,  Rincon,
and  Eisenstark 1976) and to inhibit strand rejoining in damaged DNA (Yoakum et
al. 1974).

    Another reason  to consider that DNA  is  a target of UV damage in  the lens
is the finding that 300 nm radiation can induce DNA repair synthesis in the
lens  epithelium, even when the irradiation is applied through  the  intact
cornea (Jose, Kock, and Respondek 1982;  Brenner and Grabner 1982)  and  DNA
repair  synthesis is an indication that DNA was damaged.  This point is often
misunderstood, repair is not a perfect process and  it  should  not be assumed
that  lens DNA will  be  absolutely protected.  Mutational and lethal events can
occur any time that DNA is  damaged and  repair is undertaken; thus,  anything
that  can  induce DNA repair  must be recognized for its primary damaging
effects.  Errors induced in the genome of lens cells may manifest themselves
as mutational events in  the target epithelial cell.  Such mutational events
would be cumulative over an individual's lifetime.   Those  cells that have
undergone  mutations might loose their  capacity to differentiate into normal
lens  fibers  and, for example, pile up at the posterior pole.   The  later is
seen in the senile cataract  (Streeten and  Eshaghian 1978)  as  well  as in
animals exposed to near UV  radiation (Zigman and Vaughan 1974).   The cells may
also  not carry out  their normal functions  such as  maintaining  less
osmolarity.   Then we might  develop what  appeared to be  an "osmotic"  cataract,
when in fact, the  underlying mechanism was genetic.

POTENTIAL CHANGES  IN  SENILE CATARACT  PREVALENCE
FOR  CHANGES IN  UV-B

    Epidemiological studies  have  identified  a correlation between the
prevalence  of various  types  of cataracts  in  humans and the flux of sunlight or
ultraviolet radiation  reaching the  earth's surface  (Killer,  Giacometti and
Yuen 1977,  Zigman,  Datiler,  and  Torczynski 1979; Taylor 1980,  Hollows  and
Moran 1981).  Miller,  Sperduto and  Ederer  (1983) developed a  multivariate
logistic  risk function  that  describes  the  correlation found between the
prevalence  of senile cataracts and  the flux  of UV-B and other risk factors.
The results of this study  may  indicate the magnitude of change in the
prevalence  of senile cataracts that could  be associated with  changes in UV-B
flux due  to ozone  depletion.

    Miller, Sperduto and Ederer  based  their  analysis on the 1971-1972  National
Health and  Nutrition Examination  Survey (HANES) general medical and
opthalmological  examinations of  over 10,000  persons, ages 1-74 years.   Using
the HANES data,  persons  were assigned  to one of two mutually  exclusive
groups:   (1) cataract  or aphakia  (either  eye), and  (2) neither cataract nor
aphakia (either eye).   Cataract  was defined  as "senile lens changes (cortical,
nuclear,  posterior-subcapsular,  or  other)  consistent with best corrected
visual acuity of 20/30  (6/9) or  worse" (Miller, Sperduto  and  Ederer (1983) p.
240).  This definition differs from the HANES survey,  which used a visual
                              *  DRAFT FINAL  * *

-------
                                   10-31
acuity of 20/25 (6/7.5) or worse.  Aphakia was diagnosed when "the lens had
been surgically removed and there was no history of congenital, traumatic or
secondary cataract" (Miller, Sperduto and Ederer (1983) p. 240).  The study by
Miller, Sperduto and Ederer included HANES data on a total of 2,225 persons
between the ages of 45 and 74 years who had resided at least one half of their
life in the state where the HANES examination took place.  Of these 2,225
people, 413 (18.6 percent) were placed in the cataract or aphakia outcome
category.

    The UV-B data were developed by NOAA for the 35 HANES locations used in
the study based on a statistical analysis of UV-B data collected at 10
locations using Robertson-Berger meters (RB-meters).  The statistical analysis
incorporates season, latitude, elevation, weather (clouds), and haze.
Subsequent validation of the estimates at six locations indicated that the
differences between the estimated and observed mean daily flux average about
seven percent.

    These data on UV-B arid outcome (i.e., cataract) were used in conjunction
with demographic and medical history data to estimate the following
multivariate logistic risk function:
                                   1
                   P =
exp(-a-b.X.-.
                                          . -b. X. )
                                            k k
where P is the probability (or risk) of having a cataract, and X. are risk

factors.  In addition to UV-B, the following risk factors were analyzed:  age;
race; sex; education; diabetis; systolic blood pressure; and residence  (urban,
rural).

    Exhibit 10-12 displays the standardized regression coefficients estimated
for each of the risk factors.  Positive coefficients indicate factors that are
correlated with increased risk, negative coefficients indicate factors  that
are correlated with decreased risk.  The coefficients presented in the  exhibit
are "standardized," meaning that they represent the expected change in  the
logit of P (equal to In (P/l-P)) for a one standard deviation change in the
risk factor.   Standardization of the coefficients allows the relative
importance of the risk factors to be identified by the relative size of the
standardized coefficients.

    As shown in Exhibit 10-12, three risk functions were estimated:  (1)
univariate (outcomes as a function of the risk factor);  (2) bivariate (outcome
as a function of the risk factor and age); and (3) outcome as a function of
all the risk factors simultaneously.  For all three formulations, UV-B  is
statistically significant, and positively correlated with the increased risk
of being in the cataract outcome category.

    Using the multivariate risk function coefficients, and the mean values for
all the risk factors other than UV-B, the change in the prevalence of cataract
for each 1.0 percent change in UV-B is estimated to be approximately 0.5
percent.  This relationship holds for changes in UV-B as large as minus 20
                                 DRAFT FINAL  *

-------
                                   10-32



                              EXHIBIT 10-12

             Standardized Regression  Coefficients  for  Cataract
a/
Risk Factor
Age
Race
Sex
Education

Diabetes
Systolic Blood Pressure
UV-B

Residence
V
Mean
61
1
1
2

0
146
3

1
.97
.22
.52
.87

.08
.3
.59

.37
£/
Univariate
1
0
0
-0

0
0
0

0
.22
.18
:o7
.43

.25
.33
. 19

.18
f
f

f
f

f
f .
f

Bivariate
_
0.
0.
-0.

0.
0.
0.

0.

20
08
25

23
15
20

21

f

f
f

g
f
f

Multivariate
1.
0.
0.
-0.

0.
0.
0.

0.
20
13
08
14

21
08
13

19
f
g

g
f


g
f

  a/ Values for categdrial risk factors:   race:   1 = white,  2 = black;
sex:  1 = male, 2 = female;  education:  1  = <5 grades; 2=5-8 grades;  3 =
9-11 grades; 4 = 12 grades;  5 = college;  diabetes:  0= absent,  1 = present;
residence:  1 = urban; 2 = rural.

  b/ Mean value Tor the risk factor in  the 2,225 persons in  the study.

  c/ Each risk factor analyzed separately.

  d/ Each risk factor analyzed with age only.

  e/ All risk factors analyzed simultaneously.

  f/ p(two sided) <0.005.

  £/ p(two sided) <0.05.

  Source:  Killer, R. , R. Sperduto and  F. Ederer (1983), "Epidemiologic
           Associations with Cataract in the 1971-1972 National Health and
           Nutrition Examination Survey," Americal Journal of Epidemiology,
           Vol. 118, No. 2,  pp. 239-249.
                                 DRAFT FINAL  * * *

-------
                                   10-33
percent to plus 30 percent.  Outside of this range, reductions in UV-B are
associated with less of a  reduction in cataract prevalence, and increases in
UV-B are associated with larger increases.

    Of note  is that this estimated relationship between UV-B and cataract
prevalence varies with age; UV-B has a larger effect on prevalence  (on a
percentage basis) among younger individuals.  Exhibit  10-13 displays the
percent increase in cataract prevalence expected due to increases in UV-B, for
peoples of different ages.  As shown in the exhibit, the percentage increase
in prevalence due to changes in UV-B are estimated to  be larger for 50 year
olds than for 70 year olds.

    Although the effect of UV-B on prevalence is estimated to be larger at
younger ages (on a percentage basis) using the multivariate risk function, the
prevalence of senile cataracts is known to increase substantially with age.
Leske and Sperduto (1983)  report the prevalence of senile cataracts in both
sexes found  in the Framingham Eye Study to be as follows:  52 to 64 years old
-- 4.5 percent; 65 to 74 years old -- 18.0 percent; 75 to 85 years old -- 45.9
percent.  These prevalence estimates use the same definition of cataracts as
used by Miller, Sperduto,  and Ederer.  (Larger prevalence rates are reported
by Leske and Sperduto based on HANES data.  These estimates, however, use a
definition of cataract that includes a decrease in vision to 20/25  (6/7.5),
instead of the 20/30 (6/9) used in both the Framingham study and the Hiller,
Sperduto, and Ederer risk  study.)  Because cataracts are more prevalent.in
•older individuals, increases in the actual number of cases of cataracts would
likely be larger for older individuals, even though the percentage  increase in
risk has been estimated to be larger for younger individuals.

    Using the prevalence data cited above, the prevalence of cataracts in the
U.S. population is on the  order of 9.3 million.  Using the multivariate risk
function (with all values  set to their means except age and UV-B) the
hypothetical increased prevalence for 1985 that would  have occurred had the
entire population experienced 1.0 percent ozone depletion can be estimated.
For a 1.0 percent depletion, the annual UV-B flux measured on the RB-meter has
been estimated to increase by approximately 0.83 percent (see Serafino and
Frederick (in press) and Chapter 17 for a discussion of the relationship
between ozone depletion and UV flux).  A 0.83 percent  increase in UV-B is
associated with increases  in cataract prevalence that  varies by age.  Across
all the ages, prevalence would be expected to be about 0.26 percent higher, or
about 24,000 cases, had ozone been depleted by 1.0 percent.

    Of note  is that the RB-meter measure of UV radiation may not be the
appropriate  action spectrum to use to evaluate the potential biological
effects of increased UV-B  such as cataract.  For example, the DNA action
spectrum may be preferred.  Even though the RB-meter and DNA action spectra
are highly correlated in the range of current observations, because the DNA
action spectrum is more heavily weighted toward shorter wavelengths, it
increases more rapidly with decreases in ozone levels; a 1.0 percent depletion
would lead to approximately a 2.0 percent increase in  UV-B.  Using  the 2.0
percent increase in UV-B would yield about a 0.6 percent increase in
prevalence,  or about 57,000 cases.
                                 DRAFT FINAL

-------
                                    10-34
                              EXHIBIT 10-13

               Estimated Relationship Between Risk of Cataract
                               and UV-B Flux
Percent
Increase
in
Cataract
Prevalence
                                                                            50

                                                                            60
= 70
                              Percent Increase in UV-B Flux
Increased UV-B flux (measured with an RB-meter)  is  associated  with increased
prevalence of cataract.  The percent change  in prevalence  varies  by age.

Source:  Developed from data presented  in Miller R.,  R.  Sperduto,  and F.
Ederer (1983), "Epidemiologic Associations with Cataract in  the  1971-1972
National Health and Nutrition Examination Survey,"  American  Journal of
Epidemiology, Vol. 118, No. 2, pp. 239-249.
                                 DRAFT FINAL

-------
                                   10-35
    There are various  important  limitations in the use of these estimates and
data.  The correlation between UV-B  and cataracts reported by Killer, Sperduto
and Ederer does not prove a causal connection -- other (unknown) factors could
be playing a role.   These (unknown)  factors would have to be correlated with
UV-B flux.  Also,  the  study does  not have estimates of individual lifetime
UV-B exposure, thereby limiting  the  strength of the evidence for the
association between UV-B  exposure and  cataracts.  Additionally, the sample
population analyzed may not be representative of the entire U.S. population.
Finally, the outcome category used in  the study does not differentiate between
different types of cataracts, some of  which may be more strongly related to
UV-B exposure than others.

    Confidence in  the  estimates  developed here are strengthened by several
considerations. The correlation  between UV-B flux and sunlight flux is high,
and a correlation  between sunlight and cataracts has also been found in
Australia (Taylor  1980) and in China (Mao and Hu 1982).  An association
between UV-B exposure  and cataract has also been demonstrated in laboratory
animals.  Therefore, although considerable investigation remains to be
performed, indications are that  the  association between UV-B and cataract is a
reasonable basis for evaluating  potential impacts due to increased UV-B flux
associated with ozone  depletion.


OTHER EYE  DISORDERS2

Stable Retinal  Disorders

    Evidence  has been adduced which  provides  a  chain  of evidence  on  which to
postulate that  optical  radiation may contribute  to  human  retinal  disorders
which  are not  progressive, that is,  stable retinal problems.  This chain
starts with  the knowledge that  UV-B,  UV-A and light  are absorbed by various
retinal tissues (Wolbarsht 1976;  Warner 1982).   Moreover, the amount of
radiant energy incident on the  retina  is  dependent on  the transparency of the
cornea and  lens, and  on  pupil diameter.  Next  a  large number of bioeffects of
these wavelengths  have been identified which are hazardous  to the  retina
(Marshal  1970; Williams and Baker 1980).   Further, empirical estimates have
been  made of the  domain  of time, intensity, and wavelength  which  can induce
damage to the pigment epithelium, the photoreceptors and the inner  retina.
These estimates provide  sufficient quantitative evidence to establish  safety
standards to protect people from the short-term retinal damage  which could be
induced  by lasers  and other intense sources of optical radiation  (Sliney and
Wolbarsht 1980; Pitts  1973).  There  is, moreover,  sufficient  evidence of
additive and persistent changes  in the monkey  and  human  retina to hypothesize
that some stable long-term visual impairments in humans may be related to
optical radiation.   There  are reports that optical  radiation  produces  retinal
damage hours, days,  months and years after  irradiation.  Although sparse, this
evidence  is instructive.  For example, sungazing  by humans has been  reported
2This section in boldface type is taken from  Waxier, M.  (in press).
                          * * *  DRAFT FINAL  *  '•  »

-------
                                  10-36
to produce persistent,  evolving and delayed alterations  in the visual  system.
Cavonius, Elgin, and Robbins (1974) reported that aniseikonia may persist for
years.  Pigmentary patches  and blind  spots  have been reported  to evolve in a
complicated  fashion over many months  after solar retinopathy (Lowenstein and
Steel  1941).   In addition,  late complications  resulting  in reduced visual
acuity have been described  (McFaul 1969).   Furthermore there are several
reports in  humans of prolonged, and even cumulative,  impairments in dark
adaptation following  extended viewing  by humans of skylight (Hecht et al.
1948;  Clark, Johnson,  and Dreher  1946) or bright blue,  violet  (Brindley 1953)
and ultraviolet (Wolfe 1949)  radiation.   Also there is at least one  report of
persistent visual impairment resulting  from the accidental exposure to
ultraviolet  radiation  from a welding arc (Naidoff  and Sliney  1974).  Even more
important is the description  by Tso and Woodford (1983) of  sub-RPE
neovascularization and  late impairment in  fluid transport after several years
of evolving  changes in  retinal pathology following excessive  exposure to
optical radiation.

    Some of these effects  are the result of  injury to the retina which was
ophthalmoscopically visible,  some of these effects resulted from exposure to
optical radiation which  did not produce ophthalmoscopically visible damage.
Most important are two demonstrations that  the addition of subthreshold
exposures separated by several days induces retinal damage  (Greiss and
Blankenstein 1981;  Kuwabara and Okisake 1976).  The data  strongly  suggest that
retinal damage induced by optical radiation  can cumulate over days.  This is
especially important since  the concern  about long-term visual health problems
necessarily  involves consideration of the intermittent nature of exposure to
emissions from multiple sources over an extended period of time.

Retinal Degeneration

    There are  some  reasons  to  suspect that  retinal damage  induced by
ultraviolet  radiation  can be  unstable, that is, produce progressive damage.
The retina of monkeys  can be damaged by ultraviolet  radiation  (Ham  1984; Sykes
et al.  1981).   This damage to the  retinal pigment epithelium, to the
photoreceptors and the neurons is  dose dependent,  photochemically mediated and
cumulative  over hours  and days.   Although  very few  long-term  observations have
been made  on the primate  retina following excessive exposure to  optical
radiation, there are theoretical  reasons to suspect that  a retina  compromised
by disease or age would be  more vulnerable to any  degenerative changes which
could  be induced by optical  radiation  (Young 1981).   The combination of
short-wavelength optical radiation,  oxygen,  chromophores,  and photosensitizers
in the retina is potent  with  possibilities for  producing retinal degeneration,
although the only direct evidence of the exacerbation  of retinal  degeneration
by optical radiation  has  een obtained in rodents (LaVail and T   .slle 1975).
In addition, the higher 'stinal  doses of  UV  received by aphaket and
pseudophakes make these  individuals especially vulnerable to cystoid  macular
edema  (Kraff et al.  1985).  Fluorescein leakage occurs following  exposure of
the retina of the aphakic  monkey eye  to UV-A radiation, but comparable studies
using  UV-B radiation have not  been conducted.   Therefore,  retinal degeneration
should be considered a risk of  excessive exposure to  ultraviolet  radiation.
                                 DRAFT FINAL

-------
                                   10-37
    Very  little research has been  conducted on primates, on diurnal animal
models of human retinal degenerations, or on retinas compromised  by chemicals,
age or other agents.  The striking evidence of long-delayed and persistent
changes  in the  monkey retina  following optical radiation,  coupled with  some
clinical evidence in  humans, suggests that  research  on primates needs emphasis.

Aging Disorders

    Deterioration of the visual  system as individuals age  is a fact  (Owsley et
al.  1986).  The likelihood that optical radiation contributes to this
deterioration is  high for  several reasons.   First,  the cellular pathology
observed  in older individuals is similar in appearance  to  radiation  induced
damage  (Kuwabara 1978).   Second, the ophthalmoscopic appearance of the aged
retina is sometimes  similar  to that induced by optical radiation  (Klein 1958).
Third, some  of  the  kinds  of vision loss  described  in the  aged are consistent
with the losses  which could be produced by optical radiation  damage (Jaffe,  de
Monestario,  and Podgor 1982;  Harweth and Sperling 1975).   Fourth, the
accumulation  of  lipofushin materials in the RPE  and other changes in the
retina suggest that  the aged retina has  a lessened capability of repair
(Feeney-Burns, Berman, and  Rothman 1980).

    Optical radiation may  have quite  different visual health consequences for
eyes with and without ocular disease.  Optical radiation may accelerate the
deterioration of central vision in older individuals  with central  retinal
disease  (Hyman  et al. 1983), whereas in  those free of ocular disease optical
radiation may impair paramacular and peripheral vision more than foveal
vision.   This interpretation is  offered to reconcile the facts of visual loss
outside  the macular  in disease-free aged  eyes  (Owsley et al.  1986)  with
Young's  prediction that the central retina will exhibit  the major effects of
lifetime  retinal exposure to optical radiation (Young 1981).   This ad hoc
explanation  is not very satisfactory  but it does indicate the need to
temporize theory with data.  Other interpretations  also are possible, e.g.,
those individuals with central  retinal disease  may  have had a different
exposure  history or  some  genetic  vulnerability.

    Given these facts and theoretical considerations,  exposure  of the aged
retina to ultraviolet  radiation  in excess of amounts permitted by the phakic
eye probably is hazardous.  The  retina  of  the aged individual  is exposed  to
more radiation than  phakics receive  when the crystalline  lens  is extracted
without placing  a  comparable filter in front of the retina, even when ordinary
sunglasses are used.

    Research  is needed to determine  the extent to which  optical radiation  is  a
factor in the deterioration  of the  vision  of  the elderly.   In the meantime,  we
should take practical  steps to protect our eyes from excessive  optical
radiation, and thus  prolong our visual lifetime.

Development  Disorders

    While there  are  several  reasons to suspect hazardous effects of optical
radiation on  visual development,  almost  no  research has been  conducted on the
                                 DRAFT FINAL  » •• *

-------
                                  10-38
developing primate retina.  In fact,  there  is very little  experimental
exposure  data on  the developing retina of  any species.  Glass et al.  (1985)
have  reported that the probability of retinopathy of prematurity (ROP) is
highest in infants exposed to more light during their  hospital  stay.   This
evidence confirms a prediction by Wolbarsht et al.  (1983)  about  ROP  and light
and is consistent  with the evidence for the interaction of  oxygen and  light  in
retinal damage (Ruffolo et al.  1984).   The action spectrum for this effect is
unknown  but  UV-B  and  UV-A  probably play a role.   Furthermore,  there are  a
large  variety  of sources and situations in which infants  are exposed to
optical radiation.  Some  of the same  evidence  which  gave  rise  to suspicions
about the role of  optical radiation in retinal degeneration  and visual aging
is  relevant to the potential  hazards to visual development:  short-wavelength
radiant energy incident on the retina, short-wavelength chromophores in the
retina,  dose-dependence of retinal damage,  cumulative effects, long duration
effects, and  delayed effects.   In  addition, there are special properties of
the developing visual system which  might decrease injury  thresholds,  such  as
more  transparent  ocular media and different densities  of chromophores and
screening pigments, or which might  increase the vulnerability of infants, such
as the long period of post-natal  retinal development (especially foveal)  and
early  dependence  on non-foveal visual fields.   Thus,  the  long periods required
for many  visual  processes to mature  suggests a window-of-vulnerability during
which time optical radiation might alter later visual performance, especially
if  optical  radiation alters the  spatial  and temporal summation properties of
the neural retina.

    Even  though there  is little experimental data on the effects of optical
radiation  on the developing retina, the photobiological and visual science
evidence  is sufficient to postulate that disorders of  visual development are
risks  of optical  radiation.

Retinal  Problems

    The only  human data from which one could derive an  estimate of  the radiant
exposure  of UV  which could damage  the retina is  Hecht et al. (1948)  and Clark,
Johnson,  and Dreher (1946).   These data  showed that staring at skylight for
several hours induced an abnormal retardation of  dark adaptation.  Since
rhodopsin  has an  absorption band  in the UVA  (Kurzel,  Wolbarsht, and Yananashi
1977), this deficit could have been due to the ambient outdoor UVA.   If the
                                                   -2     -2
ambient level  of UVA at the  cornea was  about 9 x 10    Jem  , as  suggested
by Ham and Mueller (1982),  then the retinal radiant exposure might have been
            -3     -2
about 9 x  10   Jem   in the Hecht and Clark  studies.   Experimental
estimates  of UV  retinal damage in monkeys have been  few.  Schmidt and Zuclich
                                                 _2
(1980) found the  threshold at 325 nm to be 10 Jem  .    Recent evidence  in
phakic rabbits (Pitts, Bergmanson, and  Chu 1983) and  rats (Rapp, Jose, and
Pitts  1985) suggest  that the radiant  exposure necessary to damage the retina
is  even lower at 300 nm.   The threshold for 300 nm damage to the monkey  retina

currently is approximately 0.6 Jem   .   Ham and Mueller (1982) found the
                                         -2                   -2
aphakic monkey has thresholds of 5.0  Jem   at 325 nm,  5.4 Jem   at  350 and
8.1 Jem"2 at 380  nm.
                          *  *  *  DRAFT FINAL  * * *

-------
                                  10-39-
REFERENCES


Ananthaswamy HN and A Eisenstark (1976)  "Near-UV-induced breaks in phage DNA:
sensitization by  hydrogen perioxide  (a tryptophan photoproduct)."  Photochem.
Photobiol. 24:439-442.

Anderson El, DD Wright and A Spector (1979)  "The state of protein  sulfydryl
groups  in normal and cataractous proteins II cortical and nucelar regions."
Exp.  Eye Res.  29:  233-243.

Augusteyn RC (1981)  In Mechanisms of Cataract Formation in the Human  Lens.
G. Duncan (ed).  Academic Press.

Augusteyne R (1974)  "Human lens albuminoid."  Jap J. Ophthalmol. 18: 127-134.

Augusteyne RC  (1975)  "Distribution of fluorescence in the human cataractous
lens."  Ophthalmic  Res.  7: 217-224.

Bachem A (1956)  "Ophthalmic ultraviolet acton spectrum."  Amer. J.  Ophthal.
41: 969-980.

Bando M, A Nakauima  and  K Satoh  (1975)   "Coloration  of human lens  protein."
Exp.  Eye Res.  20:  489-492.

Barker  FM II (1979)  The transmittance of the EMS  from 200nm to 2500nm through
the optical tissues of the eye of the pigmented rabbit.  MS Thesis University
of Houston, Houston, Texas.

Barr  RE and EL Roetman (1974) "Oxygen gradients  in the anterior chamber of
anesthetized rabbits."  Invest. Ophthal. 13: 386-392.

Benedek GB (1971) "Theory of transparency of the  eye." Applied Optics 10:
459-473.

Berler D and R  Peyser (1982)  "Light intensity and visual acuity following
cataract surgery."  Ophthalmology 89 (Suppl.)  L 117.

Bhuyan KC  and  DK  Bhuyan (1977)   "The relative functions of superoxide
dismutase and catalase in the eye."   IRCS Medical Science 5:  279.

Boettner EA and JR Wolter (1962)   "Transmission of the ocular media."   Invest.
Ophth.  1(6): 766-783.

Borkman RF (1977)  "Ultraviolet action spectrum for tryptophan destruction in
aqueous solution."   Photocchem. Photobiol.  26: 163-166.

Borkman RF and S  Lerman (1977)   "Evidence  for a  free radical mechanism  in
aging and UV irradiated ocular lenses."   Ep.  Eye Res. 25:  303-309.
                                DRAFT FINAL  *

-------
                                  10-40
Borkman RF, A Dalrymple and S Lerman  (1977)  "Ultraviolet action spectrum for
fluorogen  production  in the ocular  lens."  Photochem. Photobiol. 26: 129-132.

Borkman RF, JD Tassin and S Lerman  (1981) Exp. Eye.  Res. 32:  747-754.

Brenner W and G Grabner (1982)   "Unscheduled DNA repair  in human  lens
epithelium following in vivo and in  vitro UV  irradiation."  Ophthal. Res. 14:
160-166.

Brilliant LB, NC Grasset,  RP Pokhrel,  A Kolstad,  JM Lepkowski, GE Brilliant,
WN  Hawks and  B Pararajasegaram  (1983)  "Associations among cataract
prevalence,  sunlight  hours  and altitude in the himalayas." Amer. J. Epidem.
118: 250-264.

Brindley GC (1953)   "The effects  on colour vision  of adaptation  to  very bright
lights."  J.  Physiol.  122:  332.

Buckingham RH and A Pirie  (1972).  "The effect of light on  lens proteins in
vitro."  Exp.  Eye Res. 14: 297.

Castineiras  SG, J Dillon  and A  Spector (1979)  "Effects of reduction on
absorption and fluorescence of human lens proteins."  Exp.  Eye Res. 29:
573-575.

Cavonius, CR,  S Elgin, and DO Robbins (1974)  "Thresholds for damage to the
human retina by white light."   Exp.  Eye Res. 19:   543-548.

Chand D,  HK EI-Aguizy,  RD  Richards  and SD Varma  (1982)  "Sugar cataracts in
vitro:  implications of  oxidative stress and aldose reductase I."   Exp.  Eye
Res. 35: 491-497.

Chand D and SD Varma (1983)  "Human lens  fatty acids composition."   Invest.
Ophthal. and Vis. Sci. (ARVO  suppl.) 24:  31.

Chand D,  SD  Varma  and  RD Richards (1982)   "Lens lipid peroxidation.
Conjugated  acylpolyene and schiff bases in  human  cataracts."  Invest. Ophthal.
and  Vis. Sci.  (ARVO suppl.) 20:  218.

Chylack LT  Jr. (1978)  "Classification  of human cataracts."  Arch. Ophthal.
96:  888.

Chylack LT, MR Lee, WH Tung and  H-M Cheng (1983) "Classification of human
senile cataractrous  change by the American cooperative Cataract Research Group
(CCRG) Method." Invest  Ophthalmol. and Vis Sci,  24: 424-431.

Clark,  BA,  ML Johnson and RE Dreher (1946)  "The effect  of sunlight on  dark
adaptation." Am. J.  Ophthal. 29:   823-836.

Cooper  GF and JG  Robson (1969) J.  Physiol. 203:  411-17.
                         * -;.- *  DRAFT FINAL  * * *

-------
                                  10-41
Crylin MM, A Red vis-Lef tick and J Sugar  (1980)  "Cataract formation in
association with  ultraviolet photosensitivity."  Annals of Ophthalmol.  12:
786-790.

Dilley KJ  and A Pirie (1974)  "Changes  to the proteins of the human lens
nucleus in cataract."  Exp. Eye  Res.  19:  59-72.

Dillon J and  A Spector  (1980)  "A  comparison of aerobic  and anaerobic
photolysis of lens  proteins."  Exp. Eye  Res.  31: 591-599.

Dillon J, A Spector  and K Nakanishi  (1976) Nature  259: 422-433.

Dillon J, M Garner,  D Roy  and A Spector (1982)   "The photolysis of lens
protein:   molecular changes."  Exp.  Eye Res. 34:  651-658.

Doniger J, ED Jacobson,  K Krell and  JAD Paolo (1981)  "UV light action spectra
for neoplastic transformation and lethality of  Syrian  hamster embryo cells
correlate with spectrum for pyrimidine dimer  formation in cellular DNA."  PNAS
78: 2378-2382.

Duke-Elder S and  PA MacFaul  (1972)  System  of Ophthalmology V xiv pt 929-932.

Duke-Elder WS (1926)  Lancet  1:  1188-91.

Fecondo JV and RC  Augusteyn (1983) Exp. Eye Res. 36:  15-23.

Feeney-Burns L, ER Berman and H Rothman  (1980)  "Lipofuscin  of  human retinal
pigment epithelium."  Am. J.  Ophth.  90: 783-791.

Fraunfelder F, J Megaw and S Lerman (1983)  "Further studies  on allopurinol  in
human cataracts."  Invest.  Ophthalmol.  Vis.  Sci. 24 (Suppl.):  3.

Freeman RG  (1975)  "Data on the action spectrum for UV carcinogenesis."  J.
Nat. Cancer  Inst.  55:  1119-1121.

Garner MH and A  Spector (1980)   "Selective  oxidation of cysteine and
methionine in  normal and senile cataractous lenses."  Proc.  Natl. Acad. Sci.
77: 1274.

Girgus JS, S  Coren  and C  Porac (1977)   "Independence of in  vivo human lens
pigmentation  from  UV-light  exposure."  Vision Res.  17: 749-750.

Glass PA, G  Avery, KN Sivasubramanian, M  Keys,  AN Sostek and  DS Friendly
(1985)  "Effect of bright light in the  hospital nursery on the  incidence of
retinopathy of prematurity."  N. Engl. J. Med. 313: 401-

Glazer LD, A Rincon and A Eisenstark (1976)  "The binding of  near UV
light-induced  tryptophan photoproducts  to DNA."   Biochem. Biophys. Acta  418:
137-145.
                                DRAFT FINAL

-------
                                  10-42
Goosey JD,  JS Zigler and JH Kinoshita  (1980)  "Cross-linking of  lens
crystallins in  a photodynamic system: a process mediated by a singlet oxygen."
Science 208:  1278-1280.

Goosey JD,  ME Allison  and  CA Garcia (1983)  "Lipid peroxidative mechanism for
posterior  subcapsular cataract formation in the rabbit."   ARVO Abstracts,
Invest.  Ophthal.  Vis.  Sci.  24: 75.

Greiss GA and MF  Blakenstein (1981)  "Additivity and repair of actinic  retinal
lesions."  Invest. Ophth. and Vis.  Sci. 20: 803-807.

Ham WT Jr  (1984) "Basic mechanisms underlying  the production of photochemical
lesions in the mammalian retina."  Curr. Eye Res. 3: 165-174.

Ham WT Jr and HA Mueller  "Action spectrum for retinal injury from near
ultraviolet ratiation in the aphakic monkey."  Am. J. Ophthalmol. 93:299-

Harding JJ  (1972)  "The nature and origin of the urea-insoluble protein of
human lens."  Exp. Eye Res.  13: 33-40.

Harding JJ  (1981)  In Molecular and Cellular Biology of  the Eye Lens.   H.
Bloemendal (ed).   Wiley Interscience.

Harding JJ and KJ Dilley (1976)  "Structural proteins of  the mammalian lens: a
review with  emphasis on changes  in  development,  aging  and  cataract."  Exp.  Eye
Res. 22:  1-73.

Harm,  W  (1978)  "Evidence for photoenzymatically repairable, lethal 'non-dimer'
photoproducts formed in E.  coli cells by 265 nm radiation or by sunlight 360
nm." Mutat.  Res.  51:  301-310.

Harweth  RS and  HG Sperling (1975) "Effects of intense  visible  radiation on  the
increment-threshold spectral sensitivity of the  rhesus monkey eye."  Vis. Res.
15: 1193-1204.

Hata N and  Hockwin O  (1977)  Ophthal.  Res. 9: 256-262.

Hecht S,  CD Hendley,  S Ross, and  PV  Richmond  (1984)  "The  effect of exposure
to sunlight on night vision."  Am.  J. Ophthal. 31:  1573-1580.

Hiller R,  L  Giacometti and  K Yuen  (1977)  "Sunlight and cataract:  An
epidemiological investigation."  Amer. J. Epidemiol.  105:  450-459.

Hiller R,  RD Sperduta  and  F Ederer  (1983)  "Epidemiologic associations with
cataract  in the 1971-1972 national health and nutrition examination survey."
118: 239-249.

Hochheimer  BF (1981)  "A  possible  cause  of chronic cystyc maculopathy the
operating microscope."   Ann. Ophthalmol.  13:  153-155.
                                DRAFT FINAL  * * *

-------
                                  10-43
Hockwin  O, V Dragomirescu  and S Lerman  (1983)  "In vivo age-related changes  in
normal and cataractous human  lens density." Acta XXIV  International Cong.
Ophthalmol. In Print.

Hoenders HL and H  Bloemendal (1981)   In Molecular  and Cellular Biology of the
Eye Lens.  H.  Bloemendal (ed).  Wiley  Intersciences.

Hollows F and D Moran (1981)   "Cataract -- The Ultraviolet Risk Factor."
Lancet ii: 1249-1251.

Hymen LG, AM Lilienfeld,  et al. (1983)   "Senile macular degeneraton:  A
case-control study."   Am. J. Epidemiol.  118: 213-227.

Ichihashi  M and CA  Ramsey  (1976)  "The action spectrum and dose  response
studies of  unscheduled DNA  synthesis in normal human fibroblasts."  Photochem.
Photobiol. 23: 103-106.

Imbrie CW and TM Murphy (1982)  "UV  action spectrum  (254-405nm) for inhibition
of K-stimulated and  adenosine  triphophosphatase from the pasma  membrane  of
Rosa damascene."  Photochem.  Photobiol. 36: 537-599.

Jaffe  MJ, FM de Monestario and MJ  Podgor  (1982)  "Retinal aging and blue-
sensitivity cone functioning."   Invest. Ophth. Vis.  Sci.  22 (Suppl.): 218.

Jernigan HM,  HN Fuki, JD Goosey and JH  Kinoshita  (1981)  "Photodynamic effects
of rose bengal  or riboflavin  on carrier mediated transport  systems in rat
lens." Exp. Eye Res.  32:  461-466.

Jose JG  (1978)  "The  role of DNA damage,  its repair and its misrepair in the
etiology of cataract: a review."  Ophthal. Res.  10:  52-62.

Jose JG and  KL Yielding  (1977)   "Unscheduled DNA  synthesis in lens epithelium
following ultraviolet  irradiation."  Exp.  Eye Res. 24: 113-119.

Jose JG and  KL Yielding  (1979)   "Photosensitive cataractogens,  chlorpromazine
and methoxypsoralen cause DNA repair synthesis in  lens epithelial cells."
Invest. Opthal.  Vis. Sci.  17: 687-691.

Jose JG, HR  Kock and A Respondek (1982)   "Histological observations of the
lenses of  psoralem + UV-A treated albino rats and a  theory as  to the
underlying mechanism."   Graefe's Arch.  Clin. Exp.  Ophthalmol.  219-44-

Kahn  HA  and RC Milton,  Revised Framingham eye study prevalence  of glaucoma  and
diabetic  retinopathy.

Kahn  HA,  HM Leibowitz,  JP  Ganley,  MM  Kini, T Cotton, RS Wickerson and  TR
Dawaber  (1977a)  "The Framingham  eye  study:  I.   Outline and  major prevalence
findings." Am. J.  Epidemiol. 106: 17-32.
                                DRAFT FINAL

-------
                                  10-44
Kahn  HA, HM Leibowitz, JP  Ganley,  MM Kini, T Colton, RS Wickerson and TR
Dawaber (1977b)   The Framingham eye study:  II.  Association  of ophthalmic
pathology with single variables previously measured in the Framingham Heart
study."  Am.  J.  Epidemiol. 106:  33-41.

Kalustian A,  M Sun  and S Zigman  (1978)  "N-UV light sensitized inactivation of
glutathione reductase."  Am.  Soc.  Photobiol.  (Abstracts-June)  112-113.

Kinoshita J (1974)   "Mechanisms  of cataract  formation."  Invest.  Ophth. Vis.
Sci.  18: 713-724.

Kinsey VE  (1948)  "Spectral transmission of the eye to ultraviolet."  Arch.
Ophth.  39: 508-513.

Klien  B (1958)  "Diseases of the macula."   Arch. Ophthalmol. 60: 175-186.

Kraff MC, DR Sanders,  LM Jampol and HL  Lieberman  (1985)   "Effect of  an
ultraviolet-filtering intraocular lens on cystoid macular edema."
Ophthalmology 92:366-

Kramer  SG (1980)  "Epinephrine  distribution after topical administration  to
phakic and aphakic eyes."  Trans-Am Ophthalmol. Soc. 75: 947-982.

Kurtin WE and JA Zuclich  (1978) "Action spectrum for oxygen-dependent
near-ultraviolet  induced corneal damage."   Photochem. Photobiol. 27: 329.

Kurzel R, ML Wolbarsht and BS  Yamanashi  (1977)   "UV radiation effects on  the
human eye."   Photochem. Photobiol.  Rev.  2: 133-167.

Kurzel RB  (1973)  "Trypotophan excited states and cataracts  in the human
lens."  Nature 241:132-133.

Kuwabara T and  S Okisake (1976)   "Effect of electronic strobe flashlight on
the monkey  retina."   Jap. J. Ophthal. 20:  9-18.

Kuwabara T (1978)  "Photic  and  photo-thermal effects on  the  retinal  pigment
epithelium."   In the Retinal  Pigment  Epithelium,  Harvard  University Press,
Cambridge.

Kwan M, J Ninikoski and TK Hunt (1972)   "In vivo measurements of  oxygen
tension  in the cornea, aqueous human and anterior lens of the open eye."
Invest.  Ophthal. 11: 108-111.

Lakhanpthal V et  al. (1982)  Arch Ophthalmol 100:  1068-1070.

LaVail MM and BA Battelle (1975)  "Influence of eye pigmentation and light
deprivation on inherited retinal dystrophy in the rat."  Exp.  Eye Res. 21:167.

Lerman  S  (1980a)  "Human ultraviolet light cataracts."  Ophth.  Res.  12:
303-314.
                         * * *  DRAFT FINAL

-------
                                  10-45
Lerman  S (1980b)   "Radiant energy and the eye."   MacMillan Publishing  Company,
New York.

Lerman  S (1982)  "Ocular photoxicity and  PUVA therapy:   An experimental  and
clinical  evaluation."   FDA Photochemical Toxicity Symp.   J.  Natl. Cancer Inst.
69: 287-302.

Lerman  S (1983)  "Psoralens and ocular effects in animals and man:   In  vivo
monitoring of human ocular  and cataneous  manifestations."  Conf. on
Photobiologic,  Toxicologic and  Pharmacologic Aspects of  Psoralens, March 1-3,
1982. J. Natl. Cancer  Inst.,  In Print.

Lerman  S and  R Borkman (1976)  "Spectroscopic evaluation and  classification of
the normal, againing  and cataractous lens." Ophthal. Res.  8: 335-353.

Lerman  S, J Megaw,  and K  Gardner (1982a)  "Psoralen-long-wave ultraviolet
therapy  and human cataractogenesis." Invest.  Ophthal.  Vis.  Sci. 23:  801-803.

Lerman  S, J Megaw and K Gardner  (1982b)  "Allopurinol therapy and human
cataractogenesis." Amer.  J. Ophthalmol. 94: 141-146.

Lerman  S, JF  Kuck,  R  Borkman and E  Saker (1976)   "Acceleration  of an aging
parameter  (Fluorogen) in the ocular lens."  Ann. Ophthal. 8: 558-562.

Lerman  S, JF  Kuck,  R  Borkman and E  Saker (1976) .  "Induction, acceleration and
prevention (in  vitro)  of an  aging  parameter in the ocular  lens."  Ophthal.
Res. 8:   213-226.

Lerman  S, JM  Megaw, K Gardner,  Y Takei and I Willis (1981)  "Localization of
8-methoxypsoralen in  ocular tissues."   Opthal.  Res.  13:  106-116.

Lerman  S, K Gardner,  J  Megaw and R  Borkman  (1981)  "The prevention  of direct
and  photosensitized UV radiation damage to the ocular lens."  Ophthalmic Res.
13: 284-292.

Lerman  S, J Megaw and I Willis (1980)   "The photreactions of 8-methoxypsoralen
with tryptophan and  lens proteins."  Photochm.  Photobiol.  31: 235-242.

Leske MC and RD Sperduto (1983)  "The epidemiology of senile  cataracts:  a
review," American Journal of Epidemiology.  118: 152-165.

Lowenstein A and J Steel (1941)  "Macular burning  and  ring scotoma."   Glasgow
Med. J.  135: 73-85.

Maher EF (1978)  "Transmission and absorption coefficients for ocular media of
the rhesus monkey."   Report SAM-TR-78-32, USAF  School  of Aerospace  Medicine,
Brooks  AFB, Tx.

Marcantonio JM, G  Duncan,  PD Davis,  and AR  Bushell (1980)  "Classification of
human  senile cataracts  by nuclear color and sodium  content." Exp. Eye  Res.
31: 227-237.
                           * *  DRAFT FINAL

-------
                                  10-46
Marshal, J (1970)  "Thermal and mechanical mechanisms  in laser  damage to the
retina."   Invest. Opthalmol. 9: 97-115.

Mao W and T Hu (1982),  "An Epidemiologic Survey of  Senile Cataract in China,"
Chinese Medical Journal,  95(11): 813-818.

Martinez GS, AJ Campbell,  J Reiken and BC  Allan  (19S2)  "Prevalence of  ocular
disease in a population study of subjects 65 years old  and older."  Am. J.
Ophthalmol. 94: 181.

McCanna, P. et al  (1982) Arch  Ophthalmol  100:  1071-1073.

McFaul PA (1969)   "Visual prognosis after  solar retinopathy."  Brit. J.
Opthalmol. 53:  534-541.

Megaw J,  J  Lee and S Lerman (1980)   "NMR  analyses  of tryptophan-8-
methoxypsoralen photoreaction products formed in the  presence of oxygen."
Photochem.  Photobiol.  32:  265-269.

Moss  SH and KC Smith (1981)   "Membrane  damage can be a significant factor  in
the inactivation of  E.  coli by near UV  radiation."  Photochem. Photobiol. 32:
203-210.

Naidoff MA and DA  Sliney (1974)   "Ritinal  injury from a welding  arc."  Am.  J.
Ophthal 77: 663-668.

Owsley C, DW  Kline, JS  Werner,  V Greenstein and J Marshall  (1986)   "Optical
radiation effects on aging and visual perception."  In  Light Toxicity.  M
Wahler and VM  Kitchens  (eds.).  CRC Press, Boca Raton,  Florida.

Parrish JS,  RR Anderson,  R Urback and D Pitts (1978)   UV-A Biological  Effects
of Radiation with Emphasis  on Human  Responses to Longwave Radiation.  Plenum
Press, New York.

Pathak MA,  DM Kraemer  and U Guengerich (1972) "Formation of  thymine dimers in
mammalian skin  by  UV radiation in vivo" Photochem. Photobiol.  15: 177-185.

Pirie  A (1968)  Invest. Ophthalmol. 7: 634-649.

Pirie  A (1972)  "Photo-oxidation of  Proteins and  Comparisons of Photo-oxidized
Proteins  with Those of the  Cataractous  Human  Lens."  Israeli J.  Med Sci. 8:
1567-.

Pirie  A and KJ  Dilley  (1974)  "Photooxidation of  N-formylkynurenine and
tryptophan peptides by  sunlight or simulated  sunlight."   Photochem.  Photobiol.
19: 115-118.

Pitts  DG (1970)  "A comparative study of the effects of  ultraviolet radiation
on the eye."  Am.  J.  Optom. 47: 535-546.
                          *  •• *  DRAFT FINAL

-------
                                  10-47
Pitts DG (1973)   "The ocular ultraviolet action spectrum and  protection
criteria."  Health  Physics  15: 559-566.

Pitts DG (1974)   "The human ultraviolet action spectrum."  Am. J. Optom.  and
Physiol.  Optics 51: 946-960.

Pitts DG (1978)  "The ocular effects of ultraviolet radiation."  Amer. J.
Optom.  and Physiol. Optics 55: 19.

Pitts DG (1981)   "The threat of ultraviolet radiation to the eye - how  to
protect  against it."  J.Amer. Optom.  Assn., 52: 949-957.

Pitts DG, PD Hacker and WH Parr (1977)   "Ocular  ultraviolet effects from 295
nm  and  400 nm in the rabbit eye."  DHEW (NIOSH)  Publication  no. 77-175.

Pitts DG, JPG Bergmanson  and LWF Chu (1983)  "Rabbit eye exposure to
broad-spectrum fluorescent light."  Acta Ophthalmol (Copenhagen)  Supp. 159.

Pitts DG and KR Kay  (1969) "The photoophthalmic threshold for the rabbit."
Am. J.  Optom. 46: 561-572.

Pitts DG and AP  Cullen  (1981)   "Determination of infrared  radiation levels for
acute ocular cataractogenesis."   Graefes Arch. Klin.  Exp. Ophthalmol.
217:285-

Pitts DG and TJ  Tredici (1971)  "The effects of ultraviolet on the eye."  Am.
Ind. Hyg.  Assoc. J.  32:  235-246.

Pitts DG, et  al.  (in press), "Opticla Radiation and Cataracts,"  in Vicusal
Health and Optical Radiation, M. Waxier and V. Kitchens,  eds.,  CRC Press,
Inc., Boca Raton,  Florida.

Rapp LM, JG Jose and DG  Pitts (1985)  "DNA reapir synthesis  in the  rat  retina
following in vivo exposure to 300 nm  radiation."  Invest. Ophthalmol.  Vis.
Sci. 26:  384.

Ruffolo  JJ  Jr, WT Jr Ham,  HA  Mueller and JE Miller  (1984)   "Photochemical
lesions  in the primate  retina under conditions of elevated blood oxygen."
Invest.  Ophthalmol.  Vis.  Sci. 25: 893-898.

Satoh K, M Bando and A Nakajima (1973)   "Fluorescence  in human  lens."  Exp.
Eye Res.,  16: 167-172.

Schmidt  RE and JA Zuclich (1980)  "Retinal  lesion due to  ultraviolet laser
exposure."  Invest.  Ophthalmol.  Vis.  Sci.  19: 1166-1175.

Serafino GM and  Frederick JE  (in press)  "Global modeling of  the ultraviolet
solar flux incident  on the  biosphere," prepared for the U.S.  Environmental
Protection Agency, Washington,  B.C.
                         *  » *  DRAFT FINAL

-------
                                  10-48
Sliney,  D and M Wolbarsht (1980)  Safety with  Lasers and Other Optical
Sources.  Plenum  Press, New York.

Spector A, D  Roy and J Stauffer (1975)   "Isolation and characterization of an
age-dependent polypeptide from human lens with non-tryptophan  fluorescence."
Exp.  Eye Res. 21: 9-24.

Spector A, O  Roy and J Stauffer (1975)   "Isolation and characterization of an
age-dependent polypeptide from human lens with non-tryptophan  fluorescence."
Exp.  Eye Res. 21: 9-24.

Spector A and WH Gardner  (1982)  "Hydrogen perioxide and  human cataract."
Exp.  Eye Res. 33: 673-681.

Sprott GD, WG Martin and H Schneider (1976) "Differential effects of near UV
and visible light on  active transport and  other  membrane process in  E.  coli."
Photochem. Photobiol.  24: 21-27.

Streeten BW and J Eshaghian (1978)   "Human posterior subcapsular cataract: a
gross and flat preparation study."  Arch. Ophthal. 96: 1653-1658.

Sykes SM, WG Jr  Robison, M Waxier and  T Kuwabara (1981)   "Damage  to the monkey
retina by broad-spectrum fluorescent  light.  "Invest. Ophth.  Vis. Sci.  20:
425-434.

Taylor HR (1980)  "The Environment and  the Lens."  Brit. J.  Ophthal.  64:
303-310.

Thomas DM and  KL Schepler (1980) "Raman spectra of normal and ultraviolet-
induced cataractous  rabbit lens."  Invest. Ophthalmol. Vis.  Sci.  19: 904-912.

Trevithick JR, MO Creighton and WM  Ross (1980)   "Vitamin  E prevents
cataractogenesis in diabetic rats."  Invest. Ophth.  and Vis.  Sci.  (Suppl.) 19:
207-208.

Truscdtt RJW  and  RC  Augusteyn (1977)  "Oxidative changes  inhuman lens protein
during  senile  nuclear cataract formation."  Biochem. Biophys. Acta 492: 43-52.

Truscott RJW, K Faull and RC Augusteyn (1977) Ophthal. Res. 9: 263-8.

Turro NJ and A Lamola (1977)  Photochemistry  in the Science of Photobiology.
KC Smith (ed),  Plenum, New York, pp. 63-86.

Tso MOM  and  BJ Woodford (1983)  "Effect of photic injury on the retinal
tissues."  Ophthalmology  90: 952.

Tyrell RM  (1976)  Photochem.  Photobiol. 23:  13-20.

Van Heyningen R  (1971) Nature  230: 393-394.

Van Heyningen R'  (1973) Exp.  Eye Res. 17:  137-147.
                                DRAFT FINAL

-------
                                  10-49
Van  Heyningen R  (1976)  "What happens to the human lens in cataract."   Sci.
Amer.  233:  70-81.

Varma  SD (1981)  "Superoxide and lens of the eye: a new theory of
cataractogenesis."  Inter. J. Quantum Chem.  XX:  479-484.

Varma  SD, A Tina and  RD Richards  (1977)  "Protection against superoxide
radicals in rat  lens."  Ophthal. Res. 9: 421-431.

Varma  SD, D Chand, YR Sharma and JF Kuck (1983)  "Effect of vitamin E on
cataract development  in emory  mice."  Invest.  Ophthal. Vis. Sci.  (ARVO  Suppl.)
24:  31.

Varma  SD, D Chand, YR Sharma, JF Kuck and RD Richards (1983)   "Oxidative
stress  on  lens  and cataract formation:   role of light  and oxygen."  Presented
at the  1st International Symposium on Light and Oxygen Effects on the Eye.
Current Eye Research,  In press.

Varma  SD, NA  Beachy and RD Richards  (1982)  "Photoperoxidation of lens lipids:
prevention by  vitamin E.  Photochem. Photobiol. 36: 623-626.

Varma  SD, S Kumar and RD  Richards (1979)   "Light-induced damage to ocular len
cation  pump: prevention by vitamin C."  Proc. Natl. Acad. Sci.  76:  3504-3506.

Varma  SD, VK  Srivastava and  RD Richards (1982)  "Photoperoxidation  in lens and
cataract formation: preventive  role of superoxide  dismutase, catalase  and
vitamin C."   Ophthal. Res. 14: 167-175.

Warner, JS  (1982)  "Development of scotopic sensitivity and the absorption
spectrum of the human  ocular media."  J. Optical. Soc. Am. 72: 247-258.

Waxier, M.  (in  press), "Long-Term Visual Health Problems:  Optical Radiation
Risks," in Visual  Health and Optical Radiation, M. Waxier and V. Kitchens,
eds., CRC  Press,  Inc., Boca Raton,  Florida.

Weale RA  (1982)   "The  age variation of  senile  cataract in  various  parts of the
world."  Br. J. Ophthalmol.  66:  31-  .

Webb RB and MJ  Peak (1981)   "The effect of stray light in the 300-320 nm range
and  the role of cyclobutyl pyrimidine dimers in 365 nm  lethality." Mutat.
Res. 91:  177-182.

Weiter  JJ  and ED  Finch (1975)   "Matters arising:  UV light and human cataract."
Nature 254:  536-537.

Weiter  JJ  and S Subramanian (1978)  "Free radicals  produced  in  human lenses by
a biophotonic process."  Invest. Ophthal. Vis.  Sci. 17:  869-873.

Wiesinger  H et al. (1956)  "Transmission of light through the ocular  media of
the rabbit eye."   Am. J. Ophthalmol. 42: 907-910.
                                DRAFT FINAL

-------
                                  10-50
Williams, TP and  BN Baker eds.  (1980)  The Effects of Constant Light on Visual
Processes.  Plenum Press, New York.

Wolbarsht, ML  (1976)  "The function of  intraocular  color filters."  fed. Proc.
35: 44-50.

Wolbarsht ML,  GS Goerge, WA Jr.  Shearin,  JA Klystra  and MB  III  Landers (1983)
"Retinopathy of Prematurity:   A  new look  at an  old disease."   Paper presented
at 1st  Int. Symp. on Light and Oxygen  Toxicity, Baltimore.

Wolfe E (1949)  "Effects on visual threshold of exposure to radiation below
4000 angstroms."  Trans  Am.  Acad. Ophthalmol. Otolaryrgol.  53: 400-414.

Yoakum G, W  Ferron,  A Eisenstark  and  RB  Webb (1974)  "Inhibition of
replication gap closure  in E. coli by near  UV photoproducts of L-tryptophan."
J. Bacterial 119:62-69.

Young RW (1981)  "A theory  of central  retinal disease."  In Future Directions
of Ophthalmic  Research.  ML  Sears  (ed.).   Yale University Press,  New Haven,
Conn.

Yu N-T, JFR  Kuck and CC Askren  (1979) "Red  fluorescence in older  and
brunescent human lenses."  Invest. Ophthalmol. Vis. Sci.  18:  1278-1280.

Zigler  JS  and  JD Goosey  (1981)  "Aging of protein  molecules:  lens  crystallins
as a model system."  Trends  in Biochem.  Sci. 6: 133-136.

Zigler  JS  and  JD Goosey  (1981)  "Photosensitized oxidation  in the ocular  lens:
evidence for photosensitizers  endogenous to the  human  lens."  Photochem. and
Photobiol. 33:  869-874.

Zigler  JS, JB  Sidbury, BS  Yamanashi and M Wohlbarsht (1976)  "Studies on
brunescent cataract.   I.  Analysis of free  protein-bound  amino  acids."
Ophthal.  Res.  8: 379-387.

Zigman S  (1970)  "Near-UV light and cataracts."  Photochem. and Photobiol.  26:
437-441.

Zigman S  (1981)  "Photochemical mechanisms  in cataract formation."   In
Mechanisms of  Cataract Formation  in the Human  Lens."   G.  Duncan  (ed.).  Publ.
Acad.  Press,  Inc. London.   Pp.  117-149.

Zigman S  and  R  Vaughan (1974)   "Effects of near-UV  light on the  lenses and
retinas of mice."  Invest. Ophth. Vis.  Sci.  13:  462-465.

Zigman S,  Griess, T Yulo and J  Schultz (1973)  "Ocular protein alterations by
near UV light."  Exp.  Eye Res.  15:  255-264.

Zigman S, M Datiles and  E Torczynski (1979)  "Sunlight and Human Cataracts."
Invest. Ophthal.  Vis.  Sci.  18: 462-467.
                         * * *  DRAFT FINAL  •'•  * *

-------
                                   10-51
Zigman  S,  T Yulo and  G Griess (1976)   "Inactivation of catalase by
near-ultraviolet light and tryptophan photoproducts."  Mol. Cell.  Biochem.  11:
149-154.

Zuclich  JA (1977)  "Lenticular  effects of near-UV laser radiation."  In
Research on  the  Ocular Effects of Laser Radiation.   Part  II.  Technology
Incorporated, Final  Report,  Contrast F41609-73-C-0017.  USAF School of
Aerospace  Medicine.

Zuclich  JA (1980)  "Cumulative effects of near-ultraviolet induced corneal
damage."   Health  Phys. 38:  833.

Zuclich  JA and JS Connolly  (1976)   "Ocular  damage induced by near-ultraviolet
laser  radiation."  15:  760-764.
                          * * *  DRAFT FINAL

-------
Chapter 11

-------
                              CHAPTER 11

                     RISKS  TO CROPS AND TERRESTRIAL
               ECOSYSTEMS FROM ENHANCED  UV-B  RADIATION
SUMMARY

    In making an assessment  of  the  risk to crops and ecosystems of increased
ultraviolet-B radiation (UV-B),  it  must be recognized that existing knowledge
is in many ways deficient.   The effects of enhanced levels of UV-B have been
studied in only four of the  ten major terrestrial plant ecosystems.  Most of
our knowledge is derived from studies focused upon agricultural crops and
conducted at mid-latitudes,  not in  the tropics or at more poleward latitudes.
Trees have not been subject  to  experimentation.  Experimental protocols often
have had flaws; too often, single year studies have been done, rather than
long term ones and too much  of  the  existing data comes from growth chambers,
in which plants grow under unrealistic conditions, rather than from field
studies.  Therefore, the full extent of the potential impacts of enhanced
levels of UV-B radiation on  a global basis cannot be adequately assessed.

    Despite these limitations,  a broad range of experimental results
demonstrated that, in nearly half of the plant species examined, UV-B
radiation deleteriously affected crop yield and quality.  Data exist that
indicate that it may be reasonably  anticipated that if UV-B increases, crop
yield, and quality will decline  for  at least some cutivars.  Existing data also
suggest that increased UV-B  will alter the distribution and abundance of
plants and potentially disrupt  ecosystems.  Unfortunately, a qualitative
prediction of how these ecosystems  would be altered cannot be determined from
the current knowledge base.
                          * * *  DRAFT FINAL  * *

-------
                                   11-2
FINDINGS

1.  OF ALL CULTIVARS TESTED. APPROXIMATELY 70% WERE SENSITIVE TO UV-B;  THE
    REMAINING 30% WERE GENERALLY NOT AFFECTED.

    la.  The large intraspecific variation in response to UV-B suggests that
         tolerant genotypes exist.

    Ib.  Selective breeding techniques for resistance to UV-B may be useful in
         reducing the impact of increased UV-B radiation if it does not
         product deleterious effects for other characteristics desired in
         plants.  Presently, the genetic bases for these differences between
         sensitive and insensitive plants are not well understood.

2.  THE EFFECTS OF UV-B RADIATION HAVE BEEN EXAMINED FOR ONLY FOUR OF THE TEN
    MAJOR TERRESTRIAL ECOSYSTEMS AND FOR ONLY A THIRD OF THE PLANT GROWTH
    FORMS.

    2a.  Little or no data exist for trees, woody shrubs, vines, or lower
         vascular plants.

    2b.  Because of a lack of data it is impossible to adequately address the
         question of the effects of UV-B changes on plants on a truly global
         basis, since one would have to assume that plant groups not studied
         respond in an analogous fashion to the herbaceous agricultural
         species which have been studied.

3.  DESPITE THE LARGE UNCERTAINTIES THAT RESULT FROM IMPERFECT EXPERIMENTAL
    DESIGN OR DOSIMETRY, THE OVERRIDING TREND OF EXPERIMENTAL DATA SUGGESTS A
    POTENTIAL REDUCTION IN CROP YIELD DUE TO ENHANCED UV-B RADIATION.

    3a.  Enhanced levels of UV-B simulating between 16 and 25% ozone
         depletions caused crop yield reductions of up to 25% in 3/4 of the
         soybean cultivars tested.

    3b.  Unlike drought or other geographically isolated stresses, increases
         in UV-B due to changes in the ozone column, could effect all areas of
         the world simultaneously.  Even small reductions in crop yield on a
         global basis could lead to considerable economic consequences.
         Therefore, increases in UV-B constitute a large, but difficult to
         quantify, risk to global agricultural productivity.

4.  THE INTERACTION BETWEEN OTHER MICROCLIMATIC FACTORS AND UV-B ARE
    IMPORTANT IN DETERMINING CROP YIELD.

    4a.  Soybean (CV Essex) yield could be accurately predicted when total
         UV-B dose, daily maximum temperature, and number of days of
         precipitation are included in a regression model.
                              *  DRAFT FINAL  * *

-------
                               11-3
CROP QUALITY MAY BE REDUCED BY INCREASES IN UV-B. BUT THE RESPONSE MAY
NOT OCCUR IN A LINEAR, DOSE-DEPENDENT MANNER.

5a.  The lipid and protein content of soybean) was reduced up to 10%;
     however higher UV-B doses did not consistently result in the largest
     reductions.

THE EFFECTS OF UV-B ON FUNGAL OR VIRAL PATHOGENS VARY WITH PATHOGEN.
PLANT SPECIES, AND CULTIVAR.

6a.  Reduced vigor in UV-sensitive plants could render the plants more
     susceptible to pest or disease damage and thus result in considerable
     reductions in crop yield.

6b.  Current evidence on the issue of pathogens is very limited.

CHANGES IN UV-B  LEVELS MAY INDUCE SHIFTS IN BOTH INTER- AND
INTRASPECIFIC COMPETITION.

la.  If these shifts favor weeds over crops, enhanced UV-B radiation could
     have severe consequences, such as the increased economic cost of
     increased tilling and herbicide application.

7b.  Although competition exists in natural ecosystems, increases in UV-B
     could .alter the results of the competition.

UV-B RADIATION INHIBITS AND STIMULATES FLOWERING. DEPENDING ON THE
SPECIES AND GROWTH CONDITIONS.

8a.  The timing of flowering may also be influenced by UV-B radiation, and
     pollen may be susceptible to UV damage upon germination.

8b.  Reproductive structures enclosed within the ovary or another wall
     appear to be weel-protected from UV-B radiation.

INTERACTIONS BETWEEN UV-B RADIATION AND. OTHER ENVIRONMENTAL FACTORS ARE
IMPORTANT IN DETERMINING POTENTIAL UV-B EFFECTS ON PLANTS.

9a.  UV-B effects tend to be exaggerated under low light regimes and
     suppressed under conditions of limited nutrients or water.

9b.  Interactions with other environmental effects make extrapolation of
     data from growth chambers or greenhouses to field conditions
     difficult and often unreliable.
                          *  DRAFT FINAL  *

-------
                                   11-4
INTRODUCTION

    This chapter examines the published  and  unpublished material currently
available to assess the likely impact  of projected  increases in UV-B radiation
upon global crop productivity and the  distribution  and abundance of plants in
natural ecosystems.  The limitations to  this assessment are formidable because
of the paucity of experimental data and  the  slow development of appropriate
experimental technology.  Therefore, actual  risks may be far greater or
somewhat less than current knowledge suggests.

ISSUES AND UNCERTAINTIES  IN  ASSESSING THE  EFFECTS OF UV-B
RADIATION ON PLANTS

    Ideally, experiments should be designed  to  develop a data base which
perfectly simulates future conditions  for all plant species.  These conditions
should include all direct effects of enhanced UV-B  in addition to all of the
possible significant combinations with other effects  (interactions).  Such
interactions should include,  but not be  limited to,  the effects of increased
atmospheric levels of carbon dioxides  (C02), drought, mineral deficiency,
etc.  A perfect simulation of all possible future environments would make it
possible to accurately assess the potential  impacts of global UV-B radiation
increases with high precision.  Unfortunately,  such ideal circumstances do not
exist.  In reality, we must make assessments based  upon imperfect experimental
designs, which include only very selective and  sometimes unrealistic growing
conditions and the testing of only a few representative plant species.
Therefore, the existing data base for  this assessment will  allow us to examine
only some of the potential effects of  UV-B changes  and cannot be regarded as
conclusive.

ISSUES CONCERNING UV DOSE AND CURRENT ACTION SPECTRA FOR UV-B
IMPACT ASSESSMENT

    Total global UV-B irradiance is dependent on a  number of factors including
solar angle, latitude and altitude, stratospheric ozone concentration,
atmospheric turbidity, and cloud cover.   The earth-sun distance and minor
solar fluctuations also contribute to  annual variations in  irradiance
(Caldwell 1971).  Because of diurnal and seasonal variations in many of these
factors, the spectral composition of solar radiation also varies
substantially.  On a daily basis, solar  UV-B irradiance is  sinusoidal, peaking
at solar noon.  Annually, UV-B irradiance is maximum during summer and minimum
during winter.  Experiments evaluating the effectiveness of UV-B radiation on
plants typically do not account for such changes because of practical
difficulties in monitoring and supplementing UV-B radiation.  Generally,
supplemental UV-B radiation is provided  using filtered sunlamps as a
squarewave function by using timers.   This system provides  a proportionately
greater UV irradiance during morning and late afternoons and under cloudy
skies than.would be anticipated outdoors.  Caldwell,  Gold et al. (1983) have
designed a modulated system to monitor ambient  UV-B and provide the desired
supplemental UV-B dose.  This system provides a more realistic simulation of
anticipated ozone depletion because it modulates  lamp output in accordance
                          * * *  DRAFT FINAL  * * *

-------
                                   11-5
with actual levels of incoming solar UV radiation.  Despite substantial
expenses, it is highly recommended that such a system be utilized to improve
field simulations and sensitivity of validation studies.

    The source of UV-B radiation most commonly used in plant effects research
is the fluorescent sunlamp which emits radiation principally in the UV-B
region.  This is a low-pressure mercury vapor lamp containing a phosphor that
fluoresces primarily in the UV-B region, but includes some UV-C and UV-A
radiation.  Although the energy emitted principally comes from the fluorescing
phosphor, some emission from mercury vapor is superimposed upon this,
producing distinct lines in the spectrum.  A weighting functions is absolutely
essential to determine the biological effectiveness of the spectral energy
emitted from various types of lamps.  The biologically effective irradiance
(!__,) is given by the following relationship:
where I. is the lamp spectral irradiance and E. is the relative

effectiveness of the energy to produce a response at wavelength X.  Thus,
the biologically effective irradiance is the product of the action spectrum
and the spectral irradiance at each wavelength.

    Several UV action spectra have been developed that share the common
feature of decreasing effectiveness as wavelength increases, but have
considerable variation in the rate of this decrease.  UV-B effectiveness in an
isolated organelle can differ considerably from an intact plant because of the
cellular shielding effects in the plant and inherent repair mechanisms.
Caldwell (1971) has developed a generalized plant damage spectrum based upon
the combined responses of a number of different plant species.  Although this
represents a step forward in understanding plant responses, an action spectrum
developed on intact plants under polychromatic radiation would be preferable.
Caldwell et al. (1986) have recently attempted to develop such an action
spectrum and although there are experimental limitations, they were able to
show that it is technologically possible.

    Another essential reason for developing appropriate action spectra is the
evaluation of radiation amplification factors  (RAF), i.e., the relative
increase in biologically effective UV-B radiation associated with a specific
ozone reduction.  RAF is a complex function taking initial ozone layer
thickness, % ozone layer reduction, latitude, season, and biological weighting
function into consideration (NAS 1979).  The increase in solar UV-B radiation
as a result of ozone reduction becomes appropriate only when the biological
effectiveness of this radiation is known.  Without an RAF, the absolute
increase of total solar UV-B radiation resulting from even an appreciable
ozone reduction is insignificant.

    Since solar spectral irradiance increases by orders of magnitude with
increasing wavelength in the UV-B region, the tails of the action spectra have
a profound effect on the net RAF.  The RAF values are much higher for those
tails with steep slopes than those with shallow slopes.  Thus, the computed
                            * *  DRAFT FINAL

-------
                                   11-6
biological effectiveness of solar radiation could either be under-  or
overestimated if the action spectra are not realistic of true plant responses.

    It is evident from the above discussion that a more realistic action
spectrum is needed for a proper assessment of the possible consequences  of
ozone depletion.

ISSUES CONCERNING NATURAL PLANT ADAPTATIONS TO  UV

    Tremendous variability exists in species sensitivity to UV-B radiation
(Krizek 1978; Van, Gerrard, and West 1976; Hashimoto and Tajima 1980;  Tevini,
Iwanzik, and Thoma 1981 and 1982; Tevini and Iwanzik 1982; Teramura 1983)
(Exhibit 11-1).   Some plants show sensitivity to ambient levels of  UV-B
radiation (Teramura, Biggs, and Kossuth 1980; Bogenrieder and Klein 1978;
Sisson and Caldwell 1976), and others are apparently unaffected by  rather
massive UV enhancements (Becwar, Moore, and Burke 1982; Ambler, Rowland, and
Maher 1978).  Similarly, large differences have been reported among cultivars
of a given species (Biggs, Kossuth, and Teramura 1981; Dumpert and  Boscher
1982; Murali and Teramura 1986a,b; Murali, Teramura, and Randall 1986).
Presently the mechanisms for these inherent differences have not been  well
documented.

    Three main categories of natural UV protective mechanisms may be
considered (Beggs, Schneider-Ziebert, and Wellmann 1986).  The first includes
repair mechanisms such as photoreactivation, a light-activated enzyme-mediated
process whereby pyrimidine dimers produced by UV absorption are split  (Rupert
1984).  Indirect evidence (Beggs, Schneider-Ziebert, and Wellman 1986; Tanada
and Hendricks 1953; Bridge and Klarman 1973) suggests that photoreactivation
is a widespread phenomenon in plants.  Excision repair, the replacement  of
deleterious photoproducts by new, correct DNA sequences, has also been
documented in plant'tissues (Rowland, Hart, and Yette 1975, Soyfer  1983).
Finally, quenching and free radical scavenging of oxygen singlets produced by
photo-oxidation also alleviate some types of UV-induced damage.

    The second category of protective mechanisms include those that tend to
minimize the damaging effects of UV-B radiation.  Probably the most important
of these mechanisms in plants is growth delay.  If cell division stops or  is
slowed upon UV irradiation, other repair mechanisms could help ameliorate  the
damage before it becomes lethal.  Although many plants show a growth
inhibition upon UV-B radiation exposure (Teramura 1983), it is not  clear
whether this inhibition is the direct effect of the damage or due to this
protective mechanism.

    The third category of protective mechanism involves those which
effectively reduce the amount of UV radiation actually reaching sensitive
plant targets.  Structural attenuation by the cuticle and cell wall may  play
only a minor role since they offer little UV absorption (Caldwell,  Robberecht,
and Flint 1983;  Steinmuller and Tevini 1985).  The principal mechanism is
probably the production of UV-absorbing pigments in outer tissue layers.
Flavonoids and other related polyphenolic compounds, which occur in epidermal
                          * » *  DRAFT FINAL  * *

-------
                                   EXHIBIT 11-1


A Summary of Studies Examining Cultivars Differences  in UV-B Radiation  Sensitivity
Number of Growth
Crop Cultivars Condition £/
Glycine max









Phaseolus vulqari




Brassica oleracea

Cucumis sativus




Tritium aestivum


Zea mays


Oryza sativa

Hordeum vulqare


Spinacia oleracea
19


2

23


5

s 2

3


2
2
5

2


U
7
2
14


5

4

3
2
G.C.


G.H.

G.H.


F

G.H. & G.C.

G.H.


G.H. & G.C.
G.C.
G.H.

G.H.


G.H.
G.C.
F
G.C.


G.C.

G.H.

G.C.
G.H.
Bas i s of
Conclusion b/ Comparison c/
20% tolerant
60% intermediate
20% sensitive
Cultivar Altona more
sensitive than Bragg
8% tolerant
33% unaffected
59% sensitive
20% sensitive
80% unaffected
BBL 290 more
sensitive than Astro
Max i dor sens i t i v
Saxa , Favor i t
to lerant
no difference
( ? )
20% tolerant
80% sensitive
Poinsett extremely
sens i t i ve. Ash ley
slightly sensitive
no d i f ference ( ? )
no difference
no difference
25% extreme ly
sens! t i ve
75% sensitive
60% tolerant
40% sensitive
25% tolerant
75% sensitive
no difference
both sensitive
d.w.


d.w.

d.w.


seed d.w.

leaf
res i stance



p. s. & d .w.

d.w.

d.w.


d.w.
d.w.
d.w.
d.w.


d.w.


d.w.
d.w.
d.w.
Reference
Biggs et a 1 . 1981


Vu et a 1 . 1978

Teramura and Murali 1986




Bennett ( 1981 )

Dumpert and Boscher (1982


Van et a 1 . ( 1976)
Garrard et a 1 . ( 1976)
Murali and Teramura
(1986a)
Krizek ( 1978)


Dumpert and Boscher (1982
Biggs and Kossuth (1978)
Biggs et a 1 . ( 198U)
Biggs and Kossuth (1978)


Biggs and Kossuth (1978)


Dumpert and Boscher (1982
Biggs and Kossuth (1978)
Dumpert and Boscher (1982













)









)








)

)
                                                                                                               i
                                                                                                               vj
             * * *  DRAFT FINAL  * * *

-------
                                                EXHIBIT 11-1

             A Summary of Studies Examining Cultivars Differences in UV-B Radiation Sensitivity
                                                (cont inued)
Crop
Go s syp i urn
Penn i setum
Cucurb i ta
Number of
Cu 1 1 i va rs
h i rsutum
g 1 aucum
pepo
2
2
3
Growth
Cond i t ion a/
G.H.
G.C.
G.C.
Cone 1
no
no
no
d
d
d
us i on b/
i f Terence
i f Terence
i Tference
Bas i s of
Compa r i son c/
d.w.
d.w.
d.w.
Reference
Ambler et
Biggs and
Biggs and
a 1 .
(1978)
Kossuth
Kossuth
(1978)
( 1978)
a/ G.H.  = greenhouse;  G.C.  = growth chamber;  F = field

b/ If data presented,  sensitive means UV-B radiation reduced d.w.  by at least 10% over control
plants.   Tolerant indicates that UV-B resulted in less than 10% reduction in growth.  In some cases,
tolerant plants were even stimulated by UV-B radiation.

c/ d.w.  = total plant  dry wt;  p.s.  = net photosynthesis.
                                                                                                                            i
                                                                                                                           00
                          * * *  DRAFT FINAL  * * *

-------
                                   11-9
cells and have high UV absorption coefficients (Caldwell, Robberecht,  and
Flint 1983), are likely candidates.  It has been shown that flavonoid
concentrations in plant leaves substantially increase upon UV exposure
(Wellman 1982; Murali and Teramura 1985a and 1986a; Robberecht and Caldwell
1978; Tevini, Iwanzik, and Teramura 1981 and 1983; Flint, Jordan, and Caldwell
1985), but it has not been established whether such an increase can completely
attenuate the damaging effects of UV.   Some studies suggest that despite large
increases in flavonoid concentration,  metabolic processes, such as
photosynthesis, are still affected (Sisson 1981, Teramura et al. 1984, Mirecki
and Teramura 1984).  Also, total leaf flavonoid concentrations alone do not
account for the range of responses observed in species sensitivity.  For
example, total leaf flavonoid levels found in UV-B irradiated soybeans were
less than those found in cucumber, yet cucumber is much more sensitive to UV
(Murali and Teramura 1986a,c).  Therefore, species sensitivity to UV-B
radiation is probably the product of a number of UV-protective mechanisms
acting in concert within the plant.  We currently need more specific
information concerning plant adaptations to UV-B radiation before we can
further refine our estimates of the ability of natural plant protective
mechanisms to compensate for the projected increases in solar UV-B radiation.

ISSUES ASSOCIATED WITH THE EXTRAPOLATION OF  DATA FROM
CONTROLLED  ENVIRONMENTS TO THE FIELD

    Plant responses in growth chambers or greenhouses may neither
.quantitatively nor qualitatively resemble field responses because the
environmental conditions are unlike those found in nature.  Plants are more
sensitive to a given UV dose when grown in growth chambers compared with
field-grown plants (Caldwell 1981, Bennett 1981, Teramura 1982, Mirecki and
Teramura 1984) because in artificial environments a single factor is generally
manipulated, while all other factors are kept constant or are optimized for
growth.  Plants outdoors commonly experience simultaneous, multiple stresses
including water or nutrient limitation.  In addition to these differences in
physical factors, artificial environments lack biotic factors such as the
interactions between other plants, insects, diseases, etc.

    However, enormous complexities are inherent in field studies.  Daily
changes in the environment are superimposed upon longer scale seasonal and
annual changes.  Both temporal and spatial variability often result in
inconsistencies in plant responses between one year and the next (for examples
see Biggs et al. 1984; Gold and Caldwell 1983; Teramura 1981b; Lydon,
Teramura, and Summers 1986).

    Therefore, one useful approach in understanding the effects of UV-B
radiation on plants under more realistic conditions has been the study of the
interactions between UV and other, commonly experienced plant stresses.  These
studies revealed that exposure to enhanced levels of UV radiation may affect
the susceptibility of some plants to water stress, and alter the sensitivity
to UV by inducing flavonoid biosynthesis (Teramura, Tevini, and Iwanzik 1983;
Tevini, Thoma, and Iwanzik 1983; Teramura et al. 1984a; Teramura, Forseth, and
Lydon 1984).  The results of field experiments (Murali and Teramura 1986c)
                                 DRAFT FINAL  * *

-------
                                   11-10
revealed no additional deleterious effects of UV-B radiation when combined
with water stress.  It was hypothesized that changes in leaf anatomy,
increased flavonoid production, and reduced growth induced by water stress
masked the UV effects.

    Plant productivity (accumulation of dry matter ) became more sensitive to
UV-B radiation as total mineral supply decreased (Bogenrieder and Doute
1982).  Murali and Teramura (1985a, b) found that plant sensitivity to UV-B
radiation decreased as the phosphorus  level decreased.  This suggests that
the greatest impact of UV-B enhancement might appear in well-fertilized
(agricultural) regions, rather than in areas of low fertility.

    In most growth chambers, visible irradiances range from 10 to 40% of
average midday irradiances.  This is a cause for concern, since many of the
deleterious effects of UV radiation may be ameliorated by exposures to longer
wavelengths.  Therefore, growth chamber experiments could overestimate the
impact of UV effectiveness in the field.  Despite the wide range of plant
species and growth conditions, a clear trend emerges:  plants grown in higher
levels of visible radiation are less sensitive to UV-B radiation.  A corollary
to this conclusion is that plant sensitivity to UV-B radiation is strongly
influenced by the level of visible radiation available during growth and
development, and that shaded environments maximize this sensitivity.

    Lydon, Teramura, and Summers (1986) and Teramura and Murali (1986) have
specifically examined the differences in UV-B radiation response between, field
and greenhouse-grown soybean.  Six soybean cultivars were grown in an unshaded
greenhouse and in the field under a similar UV enhancement.  Cultivar
sensitivity was ranked based upon a combined plant response including changes
in total plant dry weight, leaf area, and plant height.  The relative ranking
for UV-B sensitivity in the greenhouse was quite similar to that found in the
field.  The major difference was that UV-B radiation produced a substantially
larger (2- to 10-fold) effect on greenhouse plants compared with field-grown
plants.  Also, in specific instances, different conclusions could be drawn
from the individual data sets.  For example, cultivar James was sensitive and
York resistant to UV-B based upon greenhouse data.  In the field, however,
these cultivars demonstrated the opposite response.  Therefore, if controlled
environment-to-field extrapolations are necessary, they must be done with the
utmost caution!  At best, general trends may be implied, but specific or
quantitative extrapolations do not yet seem plausible.

    Despite a great range of experimental growth conditions and UV doses, most
cultivars show large individual variation in response to UV-B radiation
(Exhibit 11-1).  Therefore, the potential for ameliorating the impacts of
projected increases in solar UV radiation may be present in our current crop
germplasm by selecting for UV tolerance.  However, we have little
understanding of the mechanisms responsible for these cultivar differences and
the relationship of genetic factors that provide UV tolerance to other
desirable plant characteristics.  If tolerance to UV-B radiation were linked
(in a genetic sense) to some undesirable characteristics, crop breeding may
not ameliorate the impacts of UV-B damage.
                          * * *  DRAFT FINAL  *

-------
                                   11-11
UNCERTAINTIES IN OUR CURRENT KNOWLEDGE  OF UV-B EFFECTS ON
TERRESTRIAL ECOSYSTEMS  AND PLANT GROWTH  FORMS

    Host of our knowledge of  the  biological effects of increasing solar UV
radiation stems from research focused upon agricultural crops.  Plants from
only four of ten ecosystems,  representing only about 27% of global net primary
productivity (NPP)  have been  examined (Exhibit 11-2).   In temperate forests
and temperate grasslands, only very limited preliminary data are available.
Also of 25 categories  of major plant growth forms, the effectiveness of UV-B
radiation has only  been examined  in eight (32%).  Only very limited data exist
for trees, vines,  small woody shrubs, epiphytes, or lower vascular plants.
Thus the experimental  data  base represents only 19 of the 314 plant families
in the world (Cronquest 1981).  Of  those families tested, only seven include
representative species in which harvestable yield was examined.  Therefore, we
have very little information  from which to calculate the potential impacts of
increasing levels  of solar  UV radiation on global terrestrial productivity of
ecosystem dynamics.

    Ecosystem composition and function are dependent upon the influence of
numerous biotic and abiotic factors.  Within populations, plants may possess
protective mechanisms  to UV-B,  but  the ability of higher plants to select for
those mechanisms within the timeframe of expected changes in UV-B radiation is
unclear as to the  effect of that  selection.  Differential sensitivity could
result in subtle shifts in  species  interactions such as competition.  These
shifts as well as  changes in  abiotic factors (soil nutrients, climate, etc.)
could produce significant changes in ecosystem composition.

UNCERTAINTIES WITH THE ABILITY TO EXTRAPOLATE  KNOWLEDGE TO
HIGHER  AMBIENT CO2 ENVIRONMENT AND OTHER ATMOSPHERIC POLLUTANTS

    Global atmospheric carbon dioxide (C02) concentration has been gradually
increasing from 205 ppm some  20,000 years ago to approximately 340 ppm today
(Keelings 1978, Neftel et al.  1982).  It is anticipated that sometime between
2075 and 2100, the  atmospheric C02  concentration will reach 600 ppm (Gates
1983).

    At present we  have no direct  experimental evidence on the effects of UV-B
radiation under increased  levels  of atmospheric C02.  However, in general it
is believed that increased  C02 increases photosynthesis and water use
efficiency in most  plants  (which  are reduced by UV-B radiation), especially in
C3 plants.  Exhibit 11-3 summarizes some of that research.

    Recently, however, some confounding evidence has been found for perennials,
in which high levels of C02 led to  a decrease'in biomass production (Fried,
1986).  Furthermore, most C02 work  has been done  in growth chambers, which
differs from realistic field conditions.  The net effect of an interaction
between UV-B radiation and  increased level of C02 cannot, at this time, be
predicted.  Whether the deleterious effects of UV-B will be proportionally
higher or lower in a high C02 world is uncertain.   Similarly, the combined
impacts of higher  C02 and UV-B on inter- and intraspecific competitive balance
is impossible to ascertain.  Other  changes in the environment, such as those
                            * *  DRAFT FINAL  * * *

-------
                               11-12
                          EXHIBIT  11-2

              Survey of UV  Studies  by Major Terrestrial
               Plant Ecosystems  (after Whittaker 1975)
Ecosystem
Tropical forest
Temperate forest
Savanna
Boreal forest
Agricultural
Woodland and scrubland
Temperate grassland
Swamp and marsh
Desert and semidesert
Tundra and alpine
Global NPP
(109 ton/yr)
49.4
14.9
13.5
9.6
9.1
6.0
5.4
4.0
1.7
1.1
Total Area
(10s km2)
24.5
12.0
15.0
12.0
14.0
8.5
9.0
2.0
42.0
8.0
a
Included in
UV Study
no
yes
no
no
yes
no
yes
no
no
yes
Only studies  examining  some aspect of growth.
                      * * *  DRAFT FINAL  * * *

-------
                                   11-13
                              EXHIBIT 11-3

                 Summary of UV-B and C02 Effects on Plants
                        (Lemon  1983;  Teramura  1983)
Plant Characteristic
   Enhanced UV-B
   Doubling of C02
Photosynthesis
Leaf conductance
Decreases in many C3
and C4 plants
No effect in many plants
Water use efficiency   Decreases in most plants
Dry matter produc-
tion and yield
Leaf area


Specific leaf weight

Crop maturity

Flowering
Interspecific
differences

Intraspecific
differences

Drought stress
Decreases in many plants
Decreases in many plants


Increases in many plants

No effect

May inhibit or stimulate
flowering in some plants

Species may vary in
degree of response

Response varies among
cultivars

Plants become less
sensitive to UV-B but
not tolerant to drought
In C3 plants increases
up to 100% but in C4
plant only a small
increase

Decreases both in C3 and
C4 plants

Increases in both C3
and C4 plants

In C3 plants almost
doubles but in C4 plants
only a small increase

Increases more in C3
than C4 plants

Increases

Accelerated

Flowers produced earlier
Major differences occur
between C3 and C4 plants

Response may vary among
cultivars

Plants become more
drought tolerant
                          * * *  DRAFT FINAL  * * *

-------
                                   11-14
due to global warming and climatic changes, may also play important roles in
determing effects, but cannot be assessed.  Further uncertainty stems from the
absence of knowledge on the interactions between increased UV-B levels and
that of other pollutants in the lower atmosphere.  Primarily anthropogenic in
origin, the major pollutants include ozone, sulfur oxides, and nitrogen
oxides.  According to the National Crop Loss Assessment Network, farm crop
losses in the United States in 1981 due to air pollution were between 1 and 2
billion dollars.  Many studies have demonstrated that in combination, the
deleterious effects of air pollutants were additive and in a few cases
multiplicative  (Reinert, Heagle, and Heck 1975).  At present, no data are
available on the interaction between UV-B radiation and various air
pollutants.

RISKS TO CROP YIELD RESULTING  FROM AN
INCREASE IN SOLAR UV-B RADIATION

    During the past 10 years, only nine field studies have examined the
effects of UV-B radiation on crop yield (Exhibits 11-4 and 11-5).   These nine
studies included 22 crop species.  Corn (Zea mays) was included in six of
the nine studies, soybean (Glycine max) found in four, and tomato
(Lycopersicon esculentum), bean (Phaseolus vulgaris), and potato
(Solanum tuberosum) were each found in three studies.  Yield was reduced by
UV-B in about half of these species .

    Ambler,  Rowland, and Maher (1978).grew eight species of crops in a field
at Beltsville, Maryland (30°N).   Plants were grown under a lamp irradiation
system maintained 1.6 m above them in a linear arrangement of unfiltered
Westinghouse BZS-CLG and FS-40 sunlamps.  A 2-dimensional gradient was
established:  one parallel to the lamps and another at right angles.  Only
broccoli showed a significant UV effect although the authors suggested that
sorghum and corn also were affected.   The use of unfiltered lamps, which emit
both UV-B and UV-C, make interpretation of the results difficult,  since UV-C
radiation produces qualitatively and quantitatively different effects from
UV-B (Nachtway  1975).

    Bartholic, Halsey, and Garrard (1975) conducted field exclusion studies in
Gainesville, Florida (29° 36 *N) in which ambient UV-B was filtered through
plastic films of Mylar Type S or polyethylene.  Since plants were grown under
panels covered with these films, only the direct beam component of ambient UV
was removed (approximately one-half the total ambient UV present).  Compared
with uncovered plots, total yield decreased in beans and corn growing under
both Mylar and polyethylene.  Tomato matured significantly earlier under
Mylar, resulting in an apparent decrease in fruit weight in plants growing in
uncovered plots.  There may have been a shading problem associated with the
panel framework, or a difference may have been maintained in leaf temperatures
or soil moisture as a direct result of the exclosure themselves.

    Becwar,  Moore, and Burke (1982) also conducted an exclusion study located
at 3000 m in the Colorado Rocky Mountains (39° ll'N).  Mylar, Aclar, and
cellulose acetate filters were used.   The only UV effect reported was a
                          * * *  DRAFT FINAL  * * *

-------
                                                EXHIBIT 11-4

              Summary of Field Studies Examining the Effects of UV-B Radiation on Crop Yields.
                               Values represent percent changes from controls
Amb I er
et a I .
(1978)
(1)
Ba rtho I i c
et a I .
(1975)
(2)
Becwa r
et a I .
( 1982)
(3)
B i ggs and
Kossuth
(1978)
CO
Biggs
et a I .
(198*4)
CM
E i sensta rk
et a I .
( 1985)
CM

Esser
( 1980)
(5)
Hart
et a 1 .
(1975)
(1)
Cucurbita maxima
Cucurbita pepo
Phaseolus vulgar!s
Triticum aestivum
Zea mays
Spinacia oleracea
Sorghum bicolor
Capsicum annum
Glycine max
Cynodon dactylon
Beta vulgar is
Brassica oleracea
  va r.  cap i tata
Brassica oleracea
  var.  botrytis
Lycopers icon
  esculentum
Nicotiana tobaccum
Raphanus sativus
Pennisetum glaucum
Solanum tuberosum
Brassica juncea
Vigna unguiculata
Oryza sativa
Arachis hypogaea
-24 to
  -45%
            +12 to +15

            +29 to +39
                                      to -90
              '-5
               0    -79 to -87
                                  +53 to -75
                                                                    +11 to -56
                                                                    +19 to -49
            -5 to -26
                            0

                            0
-11 to -39
                                  -9 to -43
                                  -18 to -3b
0

0

0
 i
»-•
tn
                                  -2 to -41
(1)  Unfiltered Westinghouse BZS-CLG and FS-40 sunlamps
(2)  Ambient UV filtered with Mylar Type S or poIeyethyIene
(3)  Ambient UV filtered with cellulose acetate, Aclar or Mylar
(4)  West inghouses FS-40 sunlamps f.iltered with cellulose acetate or Mylar
(5)  Unfiltered Philips TL 40/12 sunlamps and lamps filtered with Schott WG 305 (2 and 3 mm) filters
                          * * *  DRAFT FINAL  * * *

-------
                                                       EXHIBIT  11-5

                                     Details  of field  Study by Teramura  (1981-1985) a/
                                    YEAR:
CULTIVAR b/   SIMULATED OZONE CHANGE (%):
-16
   1981
-25
-16
           1982
-25
-16
                  1983 c/
-25
-16
                            198U c/
-25
-16
                                        1985
-25
Bay
Essex
James
Wi I I iams
York
Forrest
-10
-25 e/
-14 §/
+22 e/
-12
NE
NE d/
NE
NE
NE
NE
NE
+6
-12
-22
+9
-25 e/
-1U
-8
-23
-25
+ 14
-8
+27

+39 e/ +6 e/ +1U e/ -7 +6 -20

+13 e/ -11 e/ +15 e/ +10 e/ +18 e/ +U


a/  Values reported are percent  changes  in  seed yields for given simulated ozone changes (-16% and -25%) compared  to
    controls (yields measured  on a  dry weight  basis).  Irradiance by either mylar (control) or cellulose acetate (UV-B
    supplemented)  filtered  FS-40.

b/  After initially screening  23 cultivars  (16 of which were  sensitive to UV-B; >10% yield reduction)  in the greenhouse,
    these six were chosen for  field experimentation.  The six  represent the full range of UV sensitivity found  in  the
    greenhouse;  including very sensitive  and very tolerant cultivars.  Beginning in 1983 only Essex (very sensitive) and
    Williams (very tolerant) were planted  in the field to increase the experimental sample size to 200.  As shown  in the
    table, yields  for Essex were generally  reduced at the higher UV  level (ozone change of -25%), but were mixed (relative
    to controls).at the other  UV level.   Yields for Williams were generally enhanced by increased UV.  Of note  is  that  Essex
    is currently replacing  other older cultivars (including Williams) and is becoming one of the most widely planned
    soybeans in  the U.S.   In a UV enriched  environment, Essex  will be deleteriously affected.  Therefore superior  cultivars
    being developed today by crop breeders  may not be suitable for the future  should the UV environment change.

c/  Drought year.   Low yields  for both controls and experimental (i.e., dosed) plants.

d/  NE = Not evaluated.

e/  Significantly  different at p=0.05  level.
                                                * * *  DRAFT FINAL  * * *

-------
                                   11-17
decrease in wheat height (between 8 and 19% depending upon plant age), with no
corresponding change in total plant dry weight.  A second study was conducted
using filtered FS-40 sunlamps to supplement ambient levels of solar UV-B.  No
significant effect on crop yield was produced despite a calculated 52% UV
enhancement compared with sea level irradiance simulating the dose plants
received at sea level as control to make actual comparisons.

    Biggs and Kossuth (1978) reported yields for seven  crops grown in beds
filled with a synthetic soil mix in Gainesville, Florida (29° 36').  UV-B
radiation was supplemented with a linear arrangement of six FS-40 sunlamps at
a 12° angle from horizontal, producing a gradient in UV-B irradiances.
Control plants were grown without lamps above them and were adjacent to those
receiving the highest UV dose.  Although some "significant" UV effects were
reported, no indication of statistical tests nor descriptive statistics were
given.  The greatest reductions in yield were often found in plants receiving
the lowest UV irradiance and the largest UV effects were found in plants
growing adjacent to ones showing little or no effect.  Although each plant
along the gradient received a uniquely different UV dose, plants and
treatments were pooled for analysis.  These manipulations undoubtedly added a
great deal of experimental variability and resulted in difficulty in
interpretation.

    In another field experiment, Biggs et al. (1984) conducted a two-year
evaluation of crop yield in rice, wheat, corn and soybean.  Plants were again
grown in beds filled with synthetic soils, but UV irradiation was supplied by
FS-40 sunlamps filtered with either Mylar or 3, 5, or 10 ml cellulose acetate
providing 0, 32, 23 or 16% UV enhancements.  The only significant yield
reduction reported was for wheat (5% reduction) and only for one of the two
experimental years.  The data for rice, corn and soybean were highly variable
and therefore no statistical differences could be detected.

    Eisenstark et al. (1985) grew corn in large pots for three growing seasons
in Columbia, Missouri (38° 57'N).  Two levels of UV radiation were supplied by
cellulose acetate-filtered FS-40 sunlamps simulating a 7 and 21% ozone
depletion.  Controls included lamps filtered with Mylar and no overhead
lamps.  Plants were found to be particularly susceptible to UV-induced effects
during tassel development (Eisenstark and Perrot 1985).  Plants irradiated as
seedlings produced total grain yield reductions of 23 and 32% respectively
when compared with Mylar control plants.  Even larger differences were found
when these plants were compared with those grown without overhead lamps; yield
was reduced by 80% in plants filtered by Mylar and 87% in those grown under
cellulose acetate.  The authors suggested that this large lamp effect could be
due to additional UV-A emitted from the sunlamps, but this UV-A supplement was
very small relative to UV-A present in midday solar radiation (Caldwell et al.
1986).  The implications of this study for field crops (as opposed to potted
plants) is doubtful, however, since the pots allowed root structures to reach
higher temperatures than in fields.
                            -v *  DRAFT FINAL  * * *

-------
                                   11-18
    Esser (1980) conducted field studies in Frankfurt, Federal Republic of
Germany (50°N) on bean, cabbage, spinach, and potato.  He used linearly
arranged Philips TL 40/12 sunlamps suspended 3 m above the plants and produced
four UV enhancements using a combination of reflectors, Schott WG305 filters,
and unfiltered lamps.  Significant yield reductions were found only under
unfiltered lamps in the four crops tested.  Yields for spinach, cabbage, and
bean increased under filtered lamps.

    Hart et al. (1975) grew plants under linearly arranged unfiltered FS-40
sunlamps at Beltsville, Maryland (39°N).  The absence of experimental data and
details on this study limit its usefulness.  Also, the presence of UV-C
radiation from these unfiltered lamps greatly decreases the validity of these
data.

    Teramura has grown soybeans for five seasons at Beltsville, Maryland
(39°N) (Exhibit 11-5).  During the first two years, six soybean cultivars were
grown and two cultivars grown in subsequent years under filtered FS-40
sunlamps oriented perpendicularly to the soybean rows.  This arrangement
avoids the large variation in UV irradiance along the length of the bulb which
is a major problem with linearly arranged (end-to-end) lamps.  Control plants
were grown under Mylar-filtered lamps.  Field experiments simulated 0, 16, and
25% ozone depletions.  Seed yield was consistently reduced at a simulated 25%
ozone depletion in cultivar Essex.  Yield reductions of up to 20% were
reported in three of the five years.  The "UV-tolerant" cultivar Williams was
less affected but did show an 11% reduction in yield in 1983.  Also in 1983, a
substantially higher relative yield (6% increase) was found in cultivar Essex,
which was opposite to that observed in the previous two years.  One critical
caveat which must be included with this observation is that 1983 was an
extremely dry year.  Actual seed weights of control plants during 1983 were
only 20-30% of those harvested in 1981 and 1982.  Therefore, it is
questionable whether these 1983 data are representative of true (normal) field
trends.  This graphically illustrates the importance of multiyear studies and
emphasizes the need to monitor other environmental variables in addition to UV
radiation. If the results on Essex in drought years was repeated, it would
indicate that higher UV radiation actually helped plants in severe drought
years, reducing the crop loss by a few percent.

    Soybean yield is influenced by other microclimatic factors as well as
total UV-B dose.  Yield appears to be strongly influenced by the number of
days of precipitation and the number of days where the maximum temperature
exceeds 35°C.  In general, yield decreases as the number of hot days increases
and increases as the number of days with precipitation approaches 25 in the
growing season.  However, further increases in the number of days with
precipitation deceases yield.  The relative importance of UV-B dose is a
function of the cultivar and other prevailing microclimatic factors.  For
example, in Essex, total seed yield can be predicted, within 95% confidence
intervals, by including total UV-B does, number of precipitation events, and
the number of days where air temperature exceeds 35°C in a regression model.
                            * *  DRAFT FINAL  » * *

-------
                                   11-19
    Despite the broad range of experimental protocols and dosimetry used by
various investigators, it appears that increases in solar UV radiation could
potentially have a deleterious impact upon global crop yields.   Even
discounting the data from the three unfiltered lamp studies (Ambler, Rowland,
and Maher 1978, Esser 1980, Hart et al.  1975), there are still  more instances
of significant reductions in yield than reports of no effect.   While
uncertainty exists, all of these studies were imperfect validations, suffering
from one problem or another in experimental design or dosimetry, and only
three studies (Biggs et al. 1984, Eisenstark et al. 1984, Teramura 1981-1985)
include multiyear observations in which longer-scale environmental variability
has been taken into consideration and only mid-latitude regions have been
analyzed.  It is reasonably likely that increases in solar UV will decrease
global crop yields.

RISKS TO YIELD DUE TO  A DECREASE  IN  QUALITY

    Biggs and Kossuth (1978) reported that the number of abnormally shaped
tomato fruit decreased 11 to 41% under enhanced levels of UV-B  radiation in
the field.  They also reported a 6 to 23% reduction in the number of culls
(rots, cracks, sunscald,  etc.) in plants receiving a moderate UV-B dose, while
those receiving a low dose produced a 10 to 34% increase in culls.  Mean
potato weight of grade A large potatoes increased 3 to 13% under enhanced
levels of UV-B radiation, but not in any other grade categories.  In neither
case was a clear, linear relationship between UV dose and plant response
demonstrated..                                                  .

    In a field study conducted in Beltsville, Maryland, Ambler, Rowland, and
Maher (1978) reported that the sugar content in sugar beets significantly
increased between 17-21% with increasing UV-B irradiance.  However, they only
observed these significant increases in plants directly under the lamp
irradiation system and not in plants receiving a comparable UV-B dose at some
distance away.  Possibly some other uncontrolled factor (perhaps shading) was
inadvertently introduced into the experiment.

    Teramura (1982-1985)  examined the effects of UV-B radiation on seed
protein and lipid concentrations of soybean grown in the field at Beltsville,
Maryland (Exhibit 11-6).   Overall, the effects of UV radiation were relatively
small but some were nonetheless significant.  Seed protein concentrations
declined under elevated levels of UV-B radiation in two of the four years (up
to 5% reduction) in cultivar Essex.  Seed lipid concentrations  were reduced by
3 to 5% in cultivar Forrest in 1982, the only year in which this character was
measured in Forrest.  Williams showed a slight reduction in seed lipid
content, with a maximum reduction of 10% reported in 1984.

    The extent of UV-mediated alterations in yield quality cannot presently be
estimated with any degree of confidence due to the paucity of experimental
data at hand.  Furthermore, conclusions drawn from a single growing season may
be unreliable because of the annual variation in responses reported.
Nonetheless the consistent reductions in yield quality in soybean suggest that
the risk may be quite high in some crops.
                          * * *  DRAFT FINAL  * *

-------
                                                EXHIBIT 11-6

          Summary of changes in Yield Quality in Soybean Between the 1982 and 1985 Growing Seasons
                                            (Teramura 1982-1985)



                    Yield       1982    Signifi-     1983    Signifi-   198<4   Signifi-    1985    Signifi-
   Crop           Character   % Change   cance     % Change   cance   % Change  cance    % Change   cance
Glycine max
     cv Bay
          % protein
          % I i p i d
Essex     % prote i n
          % I i P i d
James     % protein
          % I i p i d
Williams  % protein
          % I i p i d
Forrest   % protein
          % I i p i d
York      % protein
          % Hpid
0 to -5
-1.U to +3
-3 to -5
-2 to +.5
+.2 to +3
-1 to -2
0 to -.2
+5 to +8
+.5 to +2
-3 to -5
-2 to -3
0 to +3
ns
ns
P=0.05
ns
ns
ns
ns
P=0.05
ns
P=C.05
ns
ns
                                                   +2
to
to
-5
-2
P=0.05
ns
                                                       to -.2 ns
                                                   -1  to +6
                                                              ns
+1
-2
             +3
             -5
to 0
to 0
ns
ns
-0.7 to 0
    0
                   to +5  P=0.05
                   to -1C P=0.05
                          0.5
                              to
                              0
                                                                                               -1
ns
ns
                               P=0.05
                               ns
                                                                                                                            i
                                                                                                                           ro
                                                                                                                           o
                          * * *  DRAFT FINAL  * * *

-------
                                   11-21
RISKS TO YIELD DUE TO  POSSIBLE INCREASES IN DISEASE OR  PEST ATTACK

    Esser (1980) reported a significant decrease in the number of  aphids  per
plant but no significant difference in the spidermite  population with
increased UV-B radiation.  Although the results suggest that UV-B  radiation
can have potentially beneficial effects on pest control,  these conclusions
must be judged with caution since the observation period was less  than two
weeks.

    Growth chamber and _in vitro studies of fungal pathogens showed that
hyaline spores are more sensitive to UV-B radiation than pigmented spores
(Cams, Grahm, and Ravitz 1978).  Owens and Krizek (1980),  however,  found that
survival of a pigmented spore (Cladosporium cucumerinum)  was decreased with
UV-B radiation due to a delay in germ tube emergence.   Other studies (Esser
1980, Biggs et al. 1984) indicate that there is no clear relationship between
spore coloration and UV-B radiation effectiveness.

    Three leaf fungal pathogens showed a significant decrease in disease
severity with UV-B radiation in a growth chamber study (Esser 1980), and  a
field study by Biggs et al. (1984) showed no significant difference in disease
severity on leaves and seeds under increased UV-B radiation.  Biggs et al.
(1984)  and Biggs (1985) found that in a leaf rust-sensitive cultivar of wheat,
disease severity increased with UV-B radiation.  In a  rust-resistant cultivar,
however, there were no differences in disease severity.
                                                                         \
    Semeniuk and Goth (1980) found significant UV-B mediated reductions  in
potato virus infection on Chenopodium quinoa.  At irradiances over 86 m-2
UV-B,,.,, no infection occurred.   In this study, virus extract was exposed  to
    Bh
UV-B radiation immediately upon its application over the leaf surface.
Intuitively, viruses should be highly susceptible to UV-B radiation, since
they consist of nucleic acids encased in proteins,  both of which have high  UV
absorption properties.  However, virus sensitivity must be viewed  in light  of
screening offered by the host tissues.

    One of the plant defense mechanisms that inhibits  final development is  the
production of a class of chemicals known as phytoalexins (Bell 1981).
Phytoalexin production in plants can be induced artificially through
mechanical injury, high temperature, application of fungicides and
antibiotics, and by UV radiation (Bridge and Klarman 1973,  Reilly  1975).  UV-B
radiation can induce isoflavonoid phytoalexin synthesis (Bakker et al. 1983),
but excess production can be toxic due to free radical formation (Beggs et  al.
1984).

    At present, it is difficult to forecast the consequences of enhanced  UV-B
radiation in terms of pest and disease damage.  From the limited information
available, it appears that in some cases UV-B radiation might decrease disease
severity while in others it might aggrevate it.  The effects vary  with
pathogen, plant species, and cultivar.  Further studies are obviously needed
to develop a better understanding of the consequences  of increased levels of
UV-B radiation on pests and plant diseases.
                              *  DRAFT FINAL  * *

-------
                                   11-22
 RISKS TO YIELD DUE TO COMPETITION WITH  OTHER PLANTS

    Many plants have been shown to exhibit a wide range of sensitivity to
 enhanced UV-B radiation  (Teramura 1983), which could lead to changes in
 competitive ability within plant communities through differential UV-B
 resistance  (Caldwell 1977).  Solar UV exclusion studies (cited in Gold and
.Caldwell 1983) using natural competing species show large differences in
 response to present levels of UV radiation.  This suggests that current levels
 of UV radiation may be partly responsible for interspecific competition among
 various native plant species.

    Gold and Caldwell  (1983) studied intraspecific competition in wheat
 (Triticum aestivum L.), wild oats (Avena fatua L.), and goat grass
 (Aegilops cylindrica Host) at various planting densities in Logan, Utah,
 under ambient and UV-B levels simulating a 16% ozone reduction.  No
 significant differences  in shoot biomass production were observed.  Therefore,
 enhanced levels of UV-B  radiation may be of little consequence in terms of
 intraspecific plant competition.  Preliminary results (Caldwell personal
 communication; 1985, unpublished manuscript) however, indicate that
 interspecific competition may be affected.  UV-B appeared to alter growth by
 the inhibition of internode elongation in wild oat but not in wheat, thus
 favoring wheat in competition for available light.

    Fox and Caldwell (1978) and Gold and Caldwell (1983) examined the effects
 of UV-B radiation on the competitive interaction of field-grown competing
 pairs from three plant associations:  agricultural crops and associated weeds,
 montane forage crops,  and disturbed weedy associates.  To measure competitive
 ability, relative crowding coefficients (RCC) were used.  RCC based upon total
 above ground biomass indicate that there were significant shifts in the
 competitive balance of some agricultural crop and weed associations.  For
 example, pigweed (Amaranthus) was more competitive under ambient UV
 conditions and alfalfa (Medicago) exhibited the competitive advantage under
 enhanced UV-B radiation.  Under a 40% ozone reduction, wheat (Triticum) had
 a competitive advantage  over wild oats (Avena) and goat grass  (Aegilops)
 but no significant differences were observed with a 16% reduction.  In the
 previous year, the competitive balance was found to be in favor of Avena.  It
 was suggested that this  difference was due to a late planting resulting in
 large water and temperature stresses during seedling development (Gold and
 Caldwell 1983).  Similarly, the weedy Geum had the competitive advantage over
 Poa grass under ambient  conditions, but the balance was shifted under
 enhanced UV-B radiation.  Preliminary field data suggest that the competitive
 ability of wheat, based  on seed biomass, increased relative to wild oats under
 enhanced UV-B radiation.

    These results demonstrate that enhanced levels of UV-B radiation can alter
 the competitive interactions of some species.  Total harvestable yield, as
 well as its quality, is  altered by the presence of weeds (Bell and Nalewaja
 1968, McWhorter and Patterson 1980).  Since there is a large number of weeds
 typically associated with various crop plants, the impact of increasing levels
 of UV-B radiation upon agricultural systems could potentially have serious
                                 DRAFT FINAL

-------
                                    11-23
 consequences  if weeds  gain a competitive  advantage  over  crops.  Even very
 subtle differences  in  sensitivity could result  in large  changes in  species
 composition over time  and possibly affect ecosystem function.  At this  time,
 however,  it is  impossible to determine  the extent of the risk.

 RISK  TO YIELD  DUE TO CHANGES  IN  POLLINATION AND FLOWERING

     In plants in which pollination and  subsequent fertilization take place
 during the  day  with flowers fully open, reproductive tissues would  seemingly
 receive an  appreciable UV-B dose.   However,  ovules  enclosed  in the  ovary are
 well-protected  against UV-B radiation.  Flint  and Caldwell  (1983) have  shown
 that the  anther wall filters out  over 98% of the incident UV-B radiation in
 six plant species.   Therefore,  before another  dehiscence, pollen  is also
 we11-protected.

     Southworth  (1969)  found UV  absorbing  compounds  with  a maximum absorbance
 in the UV-B range in the wall of  pollen . Additional studies on the UV
 absorption  profiles of stigmatic  surfaces and  exudates of many species  show
 one or more peaks in the UV-B region (Martin 1970,  Martin and Brewbaker
 1971).  Therefore,  under natural  conditions, the effectiveness of UV-B  may be
 minimal because of  UV-absorbing pigments  in the anther and pollen walls and in
 stigmatic surface exudate.

     Results of  numerous studies during  the early part of this century on the
 effects of  solar UV radiation using window glass filters indicated  an
 inhibition  of flowering by solar  UV radiation  (see reviews by Popp  and  Brown
 1936,  Caldwell  1971).   However, these experiments were generally  executed with
 insufficient  sample sizes and failed to isolate UV irradiation as the single
 contributing  factor causing the difference in  flowering. For instance,
 Caldwell  (1968) has shown that  leaf temperatures are increased sufficiently
 under window  glass  to alter the induction of flowering  (Zeevaart  1976).

     Results of  Kasperbauer and  Loomis  (1965) studying Melilotus and Caldwell
 (1968) with Trifolium dasyphyllum show  an increase in flowering with the
 exclusion of  solar  UV radiation.   A growth chamber study by  Klein,  Edsall, and
 Gentile (1965)  incorporating primarily  UV-A radiation also  shows  an increase
 in the number of flowers produced in marigold  by the exclusion of UV
 radiation.  Similarly, greenhouse trials  on beans and peas  also show an
 inhibition  of flowering and decrease in flower number due to UV-B radiation
 (Biggs and  Basiouny 1975).   However, field studies using unfiltered sunlamps
 (which emit both UV-B and UV-C) produced  no significant  effects on  flower,
 number, or  date of  flowering in marigold  and tomatoes, tasseling  in maize, or
 heading in  sorghum (Hart et al. 1975).   Biggs  and Kossuth (1978), on the other
 hand,  found an  increase in flower number  but a decrease  in  the flowering
 duration in potatoes,  and a decrease in flower number in tomato at  peak
 flowering in  UV-B irradiated plants. Apparently, some changes in flowering
. are suggested that  correlate with UV-B  radiation, but whether this  would  lead
 to an appreciable affect in harvestable yield  has not been  fully  investigated.
                             * *  DRAFT FINAL  * * *

-------
                                  11-24
REFERENCES
Ambler, J.E., R.A. Rowland and N.K.  Maher,  1978.   Response of selected
vegetable and agronomic crops to increased UV-B irradiation under field
conditions.  UV-B Biological and Climatic Effects Research (BACER),  Final
Report EPA-IAG-D6-0168, EPA, Washington, B.C.

Bakker, J., F.J. Grommers, L. Smits, A.  Fuchs  and F.W.  de Vries,  1983.
Photoactvation of isoflavonoid phytoalexins:   Involvement of free radicals.
Photochem.  Photobiol. 38:323-329.

Bartholic, J.F., L.H. Halsey, and L.A.  Garrard, 1975.   Field trials  with
filters to test for effects of UV radiation on agricultural productivity.  In
Climatic Impact Assessment Program (CIAP),  Monograph 5  (Nachtwey, D.S., M.M.
Caldwell and R.H. Biggs eds.), pp.  61-71.  U.S. Department of Transportation,
Report No. DOT-TST-75-55,  National Techn. Infor.  Serv.,  Springfield, Virginia.

Becwar, M.R., F.D. Moore III and M.J. Burke,  1982.  Effects of deletion and
enhancement of ultraviolet-B (280-315 nm) radiation on  plants grown  at 3000 m
elevation.  J. Amer. Soc.  Hort. Sci. 107:771-779.

Beggs, C.J., U. Schneider-Ziebert and E. Wellmann, 1986.   UV-B radiation and
adaptive mechanisms in plants,  I_n Stratospheric Ozone  Reduction, Solar
Ultraviolet Radiation and Plant Life (Worrest, R.C,, ed.) Springer-Verlag.
ISBN 13875-7.

Bell, A.A., 1981.  Biochemical mechanisms of disease resistance.   Ann. Rev.
Plant Physiol. 32:21-81.

Bell, A.R. and J.D. Nalewaja, 1968.   Competition of wild oats in wheat and
barley.  Weed Sci. 16:505-508.

Bennett, J.H., 1981.  Photosynthesis and gas diffusion  in leaves of  selected
crop plants exposed to ultraviolet-B radiation.  J. Environ. Qual. 10:271-275.

Biggs, R.H. and F.M. Basiouny, 1975.  Plant growth responses to elevated UV-B
irradiation under growth chamber, greenhouse,  and solarium conditions.  In
Impacts of Climatic Change on the Biosphere CIAP Monograph 5 (Nachtway, D.S.,
M.M. Caldwell and R.H. Biggs eds.),  pp.  4-197-4-248. U.S. Department of
Transportation, Report No. DOT-TST-75-55, National Techn. Info. Serv.
Springfield, Virginia.

Biggs, R.H. and S.V. Kossuth, 1978.   Effects of ultraviolet-B radiation
enhancement under field conditions on potatoes, tomatoes, corn, rice, southern
peas, peanuts, squash, mustard and radish.   UV-B Biological and Climatic
Effects Research  (BACER),  Final Report,  EPA, Washington, D.C.

Biggs, R.H., S.V. Kossuth and A.H. Teramura, 1981.  Response of 19 cultivars
of soybeans to ultraviolet-B irradiance.  Physiol. Plant.  53:19-26.
                          * * *  DRAFT FINAL  * *

-------
                                  11-25
Biggs, R.H., P.G. Webb, L.A. Garrard, T.R. Sinclair and S.H.  West, 1984.   The
effects of enhanced ultraviolet-B on rice, wheat, corn, soybean, citrus,  and
duckweed.  Year 3 Interim Report.  Evironmental Protection Agency Report
808075-03, EPA, Washington, B.C.

Bogenrieder, A. and R. Klein, 1978.  Die abhangigkeit der UV-empfindlichkeit
von der lichtqualitat bel der aufzucht (Lactuca sativa L.).  Angew.   Botanik
52: 283-293.

Bogenrieder, A. and Y. Doute, 1982.  The effects of UV on photosynthesis  and
growth in dependence of mineral nutrition (Lactuca sativa L.  and Rumex
alpinus L.).  In Biological Effects of UV-B Radiation (Bauer, H., M.M.
Caldwell, M. Tevini and R.C. Worrest, eds.), Gesellschaft fur Strahlen-und
Umweltforschung mbH Muchen, pp. 164-169.

Bridge, N.A. and W.L. Klarman, 1973.  Soybean phytoalexin hydroxyphaseollin,
induced by ultraviolet irradiation.  Phytopathol. 63:606-608.

Caldwell, M.M., 1968.  Solar ultraviolet radiation as an ecological factor for
alpine plants.  Ecol. Monogr. 38:243-268.

Caldwell, M.M-., 1971.  Solar UV irradiation and the growth and development of
higher plants.  Photophysiol. VI (Giese, A.C., ed.), pp. 131-177.  Academic
Press, New York.

Caldwell, M.M., 1977.  The effects of solar UV-B (280-315 nm) on higher
plants:  Implications of stratospheric ozone reduction.  In Research in
Photobiology (Castellani, A., ed.), pp. 597-607.  Plenum Publ. Corp., New York.

Caldwell, M.M., 1981.  Plant response to solar ultraviolet radiation.  In
Encyclopedia of Plant Physiology New Series, Volume 12A.  Physiological Plant
Ecology I (Lange, O.L., P.S. Nobel, C.B. Osmond and H. Ziegler, eds.), pp.
169-197 Springer-Verlag, Berlin, Germany.

Caldwell, M.M., R. Robberecht and S.D. Flint, 1983.  Internal filters:
Prospects for UV-allimation in higher 'plants.  Physiol. Plant. 58:445-450.

Caldwell, M.M., W.G. Gold, G. Harris and C.W. Ashurst, 1983.   A modulated lamp
system for solar UV-B (280-320 nm).  Supplementation studies in the field.
Photochem. Photobiol.  37:479-485.

Caldwell, M.M., L.B. Camp, C.W. Warner and S.D. Flint, 1986.   Action spectra
and their key role in assessing biological consequences of solar UV-B
radiation change.  Ln Stratospheric Ozone Reduction, Solar Ultraviolet
Radiation and Plant Life (Worrest, R.C., ed.), Springer-Verlag. ISBN 13875-7.

Cams, H.R., J.H. Grahm and S.J. Ravitz,  1978.  Effects of UV-B radiation on
selected leaf pathogenic fungi and on disease severity.  EPA-IAG-D6-0168.
BACER Program, EPA, Washington, D.C.
                          * * *  DRAFT FINAL  * * *

-------
                                  11-26
Cronquist, A. 1981.  An integrated system of classification of flowering
plants.  Columbia University Press, New York, p. 1262.

Dumpert, K. and J. Boscher, 1982.  Response of different crop and vegetable
cultivars to UV-B irradiance:  Preliminary results,  In Biological Effects of
UV-B Radiation (Bauer, H., M.M. Caldwell, M. Tevini and R.C. Worrest, eds.),
pp. 102-107.  Gesellschaft fur Strahlen-und Umweltforschung mbH, Muchen.  ISBN
0721-1694.

Eisenstark, A. and G.H. Perrot 1985.  Enhanced UV-B irradiation (290-320 run)
of corn (Zea mays L.).  II.  Effects of various stages of tassel formation
(greenhouse-plants).  Submitted to J. Agronomy.

Eisenstark, A., G.H. Perrot, G. Ulmer and C.D. Miles, 1985.  Enhanced UV-B
irradiation (290-320 nm) of corn (Zea mays L.) I.  Effects on growth and
yield.  Manuscript submitted to J.  Agronomy.

Esser, G., 1980.   Einfluss einer nach schadstoffimission vermehrten
einstrahlung von UV-B-licht auf kulturpflanzen, 2. Versuchsjahr.  Bericht
Battelle Institut  E.V. Frankfurt.   BF-R-63, 984-1.

Flint, S.D. and M.M. Caldwell, 1983.  Influence of floral optical properties
on the ultraviolet radiation environment of pollen.  Amer. J. Bot.
70:1416-1419.

Flint, S.D., P.W. Jordan and M.M. Caldwell, 1985.  Plant protective response
to enhanced UV-B radiation under field conditions:  Leaf optical properties
and photosynthesis.  Photochem. Photobiol. 41(l):95-99.

Fox, F.M. and M.M. Caldwell, 1978.   Competitive interaction in plant
populations exposed to supplementary ultraviolet-B radiation.  Oecologia
36:173-190.

Fried, J.S., K.A. Surano, P.F. Daley, J.H. Shinn, and P. Anderson, June,
1986.  Biomass production and nutrient responses of pondersa pine to long-term
elevated C02 concentrations, North American Forest Biological Workshop.
Lawrence Livermore National Laboratory, UCRL-94281 preprint.

Garrard, L.A., T.K. Van and S.H. West, 1976.  Plant response to middle
ultraviolet (UV-B) radiation:  Carbohydrate levels and chloroplast reactions.
Soil and Crop Sci. Soc. Florida Proc. 36:184-188.

Gates, D.M., 1983.  An overview.  In C02 and Plants (Lemon, E.R. ed.), pp.
7-20.  Westview Press Inc., Boulder, Colorado.

Gold, W.G. and M.M. Caldwell, 1983.  The effects of ultraviolet-B radiation on
plant competition in terrestrial ecosystems.  Physiol. Plant. 58:435-444.
                              *  DRAFT FINAL  * * *

-------
                                  11-27
Hart, R.H., G.E. Carlson, H.H. Klueter and H.R. Cams, 1975.  Response of
economically valuable species to ultraviolet radiation.  I_n Climatic Impacts
Assessment Program (CIAP), Monograph 5 (Nachtway, D.S., M.M. Caldwell and R.H.
Biggs, eds.), pp. 263-275.  U.S. Department of Transportation, Report No.
DOT-TST-75-55, National Tech. Info. Serv., Springfield, Virginia.

Hashimoto, T. and M.  Tajima, 1980.  Effects of ultraviolet irradiation on
growth and pigmentation in seedlings.  Plant Cell Physiol. 21:1559-1571.

Howland, G.P., R.W. Hart and M.L. Yette, 1975.  Repair of DNA strand breaks
after gamma-irradiation of protoplasts isolated from cultured wild carrot
cells.  Mutat. Res. 27:81-87.

Kasperbauer, M.J. and W.E. Loomis, 1965.  Inhibition of flowering by natural
daylight on an inbred strain of Mellotus.  Crop Sci. 5:193-194.

Keelings, C.D., 1978.  Atmospheric carbon dioxide in the 19th century.  Sci.
202:1109.

Klein, R.H., P.C. Edsall and A.C. Gentile, 1965.  Effects of near ultraviolet
and green radiations on plant growth.  Plant Physiol. 40:903-906.

Krizek, D.T., 1978.  Differential sensitivity of two cultivars of cucumber
(Cucumis sativus L.) to increased UV-B irradiance:  I.  Dose-response
studies.  Final Report EPA-IAG-D6-0168, USDA/EPA BACER Prog., EPA, Washington,
D.C.

Lemon, E.R., 1983.  C02 and Plants.  The Response of Plants to Raising Levels
of Atmospheric Carbon Dioxide.  Westview Press Inc., Boulder, Colorado.

Lydon, J., A.H. Teramura and E.G. Summers, 1986.  Effects of ultraviolet-B
radiation on the growth and productivity of field grown soybeans.  I_n
Stratospheric Oxone Reduction, Solar Ultraviolet Radiation and Plant Life
(Worrest, R.C., ed.). Springer-Verlag.

Martin, F.W., 1970. The ultraviolet absorption profile of stigmatic extracts.
New Phytol. 69:425-430.

Martin, F.W. and J.L. Brewbaker,  1971.  The nature of the stigmatic exudate
and its role in pollen germination.  In Pollen:  Development and Physiology
(Heslop-Harrison, ed.)> PP- 262-266, Appleton-Century-Crofts, New York.

McWhorter, C.G. and D.T. Patterson, 1980.  Ecological factors affecting weed
competition in soybeans.  Iji World Soybean Res. Conference II:  Proceedings
(Corgin, F.T., ed.), pp. 371-392.  Westview Press, Boulder, Colorado, ISBN
0-89158-158-899-X.

Mirecki, R.M. and A.H. Teramura,  1984.  Effects of ultraviolet-B irradiance on
soybean.  V. The dependence of plant sensitivity on the photosynthetic photon
flux density during and after leaf expansion.  Plant Physiol. 74:475-480.
                                 DRAFT FINAL  * * *

-------
                                  11-28
Murali, N.S. and A.H. Teramura, 1985a.  Effect of UV-B irradiance on soybean.
VI.  Influence of phosphorus nutrition on growth and flavonoid content.
Physiol. Plant. 63:413-416.

Murali, N.S. and A.H. Teramura, 1985b.  Effect of UV-B irradiance on soybean.
VII.  Biomass and concentration and uptake of nutrient at varying P supply.
J. Plant Nutrition.  8(2):177-192.

Murali, N.S. and A.H. Teramura, 1986a.  Intraspecific difference in Cucumis
sativa L. sensitivity to ultraviolet-B radiation.  Physiol.  Plant.   In press.

Murali, N.S. and A.H. Teramura, 1986b.  Effect of supplemental ultraviolet-B
radiation on the growth and physiology of field-grown soybean.  Env. Exp. Bot.
In press.

Murali, N.S. and A.H. Teramura, 1986c.  Effectiveness of UV-B radiation on the
growth and physiology of field-grown soybean modified by water stress.
Photochem. Photobiol. In press.

Murali, N.S., A.. H. Teramura and S.K. .Randall, 1986.  A comparative study of
ultraviolet-B sensitivity between two soybean cultivars. Env. Exp.  Bot.  In
review.

Nachtway, D.S., 1975.  Linking photobiological studies at 254 nm with UV-B.
In Climatic Impact Assessment Program (CIAP), Monograph 5 (Nachtway, D.S.,
M.M. Caldwell and R.H. Biggs, eds.), pp. 3-50 to 3-84. U.S.  Department of
Transportation, Report No.  DOT-TST-75-55, National Tech. Info. Serv.,
Springfield, Virginia.

National Academy of Sciences, 1979.  Stratospheric ozone depletion by
halocarbons:  Chemistry and transport.  National Academy Press, Washington,
B.C.

Neftel, A., H. Oeschger, J. Schwander, B. Stauffer and R. Zumbrunn.  1982.
Ice core sample measurements give atmospheric C02 content during the past
40,000 years.  Nature 295:220-223.

Owens, O.V.H. and D.T. Krizek, 1980.  Multiple effects of UV radiation
(265-330 nm) on fungal spore emergence.   Photochem. Photobiol. 32:41-49.

Popp, H.W. and F. Brown, 1936.  The effects of ultra-violet radiation upon
seed plants.  I_n Biological Effects of Radiation (Duggar, B.M., ed.), Vol. 2,
pp. 853-887, McGraw-Hill, New York.

Reilly, J.J., 1975.  The role of thymine dimers in the induction of
phytoalexin, hydroxyphaseollin, in ultraviolet irradiated soybean suspension
culture.  Ph. D. Thesis, University of Maryland, College Park, Maryland.

Reinert, R.A. ,' A.S. Heagle and W.W. Heck, 1975.  Plant responses to pollutant
combination.  In Response of Plants to Air Pollution  (Mudd,  J.B. and T.T.
Kozlowski eds.), pp. 159-178.  Academic Press, New York. ISBN 0-12-509450-7.
                          * * *  DRAFT FINAL  * * *

-------
                                  11-29
Robberecht, R. and M.M. Caldwell, 1978.  Leaf epidermal transmittance of
ultraviolet radiation and its implication for plant sensitivity to
ultraviolet-radiation induced injury.  Oecologia 32:277-287.

Rupert, C.S., 1984.  Cellular repair and assessment of UV-B radiation damage.
In Stratospheric Ozone Reduction, Solar Ultraviolet Radiation and Plant Life
(Worrest, R.C., ed.) Springer-Verlag.  ISBN 13875-7.

Semeniuk, P. and R.W. Goth, 1980.  Effect of ultraviolet irradiation on local
lesion development of potato virus S on Chenopodium quinoa cv. Valdivia
leaves.  Env. Exp. Bot. 20:95-98.

Sisson, W.B., 1981.  Photosynthesis, growth, and ultraviolet irradiance
absorbance of Cucurbita pepo L. leaves exposed to ultraviolet-B radiation
(280-315 nm).  Plant Physiol. 67:120-124.

Sisson, W.B. and M.M. Caldwell, 1976.  Photosynthesis, dark respiration, and
growth of Rumex patientia L. exposed to ultraviolet irradiance (280 to 315
nanometers) simulating a reduced atmospheric ozone column.  Plant Physiol.
58:563-568.

Southworth, D., 1969.  Ultraviolet absorption spectra of pollen and spore
walls.  Grana Palynologia 9(1-3):1-15.

Soyfer, V.N., 1983.  Influence of physiological conditions on DNA repair and
mutagenesis in higher plants.  Physiol. Plant. 58:373-380.

Steinmueller, D. and M. Tevini, 1985.  Action of ultraviolet radiation (UV-B)
upon cuticular waxes in some crop plants.  Planta 164:557-564.

Strain, B.R, and F.A. Bazzaz, 1983.  Terrestrial plant communities.  In C02
and Plants  (Lemon, E.R., ed.), pp. 177-222.  Westview Press Inc., Boulder,
Colorado.

Tanada, T. and S.B. Hendricks, 1953.  Photoreversal of ultraviolet effects in
soybeans.  Amer. J. Bot. 40:634-637.

Teramura, A.M., 1981a.  Differences  in the photosynthetic response to
UV-Bradiation between mature and immature leaves.  Plant Physiol. 67 (Suppl):
93 (Abstract).

Teramura, A.M., 1981b.  Cultivar differences in the effects of enhanced UV-B
irradiation.  Annual Report for EPA, EPA, Washington, D.C.

Teramura, A.M., 1982.  The amelioration of UV-B effects on productivity
byvisible radiation.  In The Role of Solar Ultraviolet Radiation in Marine
Ecosystems  (Calkins, J., ed.), pp. 367-382, Plenum Publ. Corp. New York.  ISBN
0-306-40909-7.
                          * * *  DRAFT FINAL  * *

-------
                                  11-30
Teramura, A.H., 1983.  Effects of ultraviolet-B radiation on the growth and
yield of crop plants.  Physiol. Plant. 58:415-427.

Teramura, A.H. and N.S. Murali, 1986.  Intraspecific differences in growth and
yield of soybean exposed to ultraviolet-B radiation under greenhouse and field
conditions.  Env. Exp. Bot. In press.

Teramura, A.H., R.H. Biggs and S. Kossuth, 1980.  Effects of ultraviolet-B
irradiances on soybean.  II.  Interaction between ultraviolet-B and
photosynthetically active radiation on net photosynthesis, dark respiration,
and transpiration.  Plant Physiol. 65:483-488.

Teramura, A.H., M. Tevini and W. Iwanzik, 1983.  Effects of ultraviolet-B
irradiation on plants during mild water stress.  I.  Effects on diurnal
stomatal resistance.  Physiol. Plant. 57:175-180.

Teramura, A.M., I.N. Forseth and J. Lydon, 1984.  Effects of ultraviolet-B
radiation on plants during mild water stress.  IV.  The insensitivity of
soybean internal water relations to ultraviolet-B radiation.  Physiol. Plant
62:384-389.

Teramura, A.M., M.C. Perry, J. Lydon, M.S. Mclntosh and E.G. Summers, 1984.
Effects of ultraviolet-B radiation on plants during mild water stress.  III.
Effects on photosynthesis recovery and growth in soybean.  Physiol. Plant.
60:484-492.

Tevini, M., W. Iwanzik, and U. Thoma, 1982.  The effects of UV-B irradiation
on higher plants.  In Nato Conference Series, Series IV.  (Chalkins, J.,
ed.), pp. 435-448 Plenum Press, New York and London.

Tevini, M. and W. Iwanzik, 1982.  The effects of UV-B irradiation on higher
plants.  In the Role of Solar Ultraviolet Radiation on Marine Ecosystems
(Chalkins, J., ed.), pp. 581-615.  Plenum Pub. Corp., New York.

Tevini, M. and W. Iwanzik, 1983.  Inhibition of photosynthetic activity by
UV-B radiation in radish seedlings.  Physiol. Plants 58:395-400.

Tevini, M., W. Iwanzik and A.H. Teramura, 1983.  Effects on UV-B radiation on
plants during mild water stress.  II.  Effects on growth, protein and
flavonoid content.  Pflanzenphysiol 110:459-467.

Tevini, M., W. Iwanzik and U. Thoma, 1981.  Some effects of enhanced UV-B
irradiation on the growth and composition of plants.  Planta 153:388-394.

Tevini, M., U. Thoma, and W. Iwanzik, 1983.  Effects of enhanced UV-B
radiation on germination, seedling growth, leaf anatomy, and pigments of some
crops.  Z. Pflanzenphysiol. 109:435-448.

Van, T.K., L.A. Garrard and S.H. West, 1976.  Effects of UV-B radiation on net
photosynthesis of some crop plants.  Crop Sci. 16:715-718.
                          * * *  DRAFT FINAL  * * *

-------
                                   11-31
Vu, C.V., L.H. Allen and L.A. Garrard, 1978.  Effects of supplemental
ultraviolet radiation (UV-B) on growth of some agronomic crop plants.  Soil
and Crop Sci. Soc. Florida Proc. 38:59-63.

Wellmann, F., 1982.  Phenylpropanoid pigment synthesis and growth reduction as
adaptive reactions to increased UV-B radiation,  In Biological Effects on
UV-B Radiation.    (Bauer, H., M.M. Caldwell, M. Tevini and R.C. Worrest, eds.),
pp. 145-149.  Gesellschaft fur Strahlen-und Umweltforschung mbH, Muchen.

Whittaker, R.H.,  1975.  Communities and ecosystems.  MacMillan Publishing Co.,
Inc.,  New York.

Zeevaart, J.A.D., 1976.  Physiology of flower formation.  Ann. Rev. Plant
Physiol. 27:321-348.
                            * *  DRAFT FINAL  * * *

-------
Chapter 12

-------
                              CHAPTER 12

                   AN  ASSESSMENT OF THE  EFFECTS OF
         ULTRAVIOLET-B RADIATION  EFFECTS ON AQUATIC ORGANISMS
SUMMARY

    Various experiments  have demonstrated that UV-B radiation causes damage
to fish larvae and juveniles,  shrimp  larvae, crab  larvae, copepods, and
plants essential to the  aquatic  food  web.  These damaging effects include
decreases in fecundity,  growth,  survival, and other functions of these
organisms.   In natural marine plant communities a  change in species
composition rather than  a decrease in net production is the probable result
of enhanced UV-B exposure.   The  change in community composition may introduce
instabilities to ecosystems and would likely have an influence on higher
trophic levels.  A decrease in column ozone  could  diminish the near-surface
season of invertebrate zooplankton populations.  Whether the population could
endure a significant shortening  of the surface season is unknown.

    The direct effect of UV-B radiation on food-fish larvae closely parallels
the effect  on invertebrate zooplankton.  Information is required on seasonal
abundances  and vertical  distributions of fish larvae, vertical mixing, and
penetration of UV-B radiation into appropriate water columns before effects
of incident or increased levels  of exposure  to UV-B' radiation can be
predicted.   However, in  one study involving  anchovy larvae, a 20% increase in
incident UV-B radiation  (which would  accompany about a 9% decrease in the
atmospheric ozone column) would  result in all of the larvae within a 10-meter
mixed layer in April and August  being killed after 15 days.  It was
calculated  that about 8% of the  annual larval population throughout the
entire water column would be directly killed by a  9% decrease in column ozone.

    Effects induced by solar UV-B radiation  have been measured to depths of
more than 20 m in clear  waters and more than 5 m in unclear water.  The
euphotic zone (i.e, those depths with levels of light sufficient for positive
net photosynthesis) is frequently taken as the water column that reaches
down to the depth at which surface of the photosynthetically active radiation
is reduced  99%.  In marine ecosystems, UV-B  radiation penetrates
approximately the upper  10% of the marine euphotic zone before it is reduced
to 1% of its surface irradiance.  Pentration of UV-B radiation into natural
waters is a key variable in assessing the potential impact of this radiation
on any aquatic ecosystem.
                          * * *  DRAFT FINAL  * *

-------
                                   12-2
FINDINGS

1.  REDUCTION IN STRATOSPHERIC OZONE WILL INCREASE SOLAR ULTRAVIOLET
    RADIATION THAT HAS THE POTENTIAL TO HARM AQUATIC LIFE.

    la.  Increases in energy in the 290-320 nm wavelengths  that would reach
         the ground if the ozone layer depleted could harm  aquatic life.

    Ib.  There is no single "correct" weighting function for determining the
         importance of different wavelengths to the damage  UV-B radiation can
         cause; a somewhat different weighting function exists for each
         biological endpoint.

    Ic.  Based on the DNA response, a 10% ozone decrease would result in a
         28% increase in biologically effective radiation,  the radiation that
         appears most relevant to aquatic organisms.

    Id.  A 10% decrease in column ozone would produce an increase in
         biologically effective radiation comparable to migrating over 30°
         toward the equator, a change of ecological significance.

2.  EXPERIMENTAL DATA DEMONSTRATES A RISK TO AQUATIC LIFE.   .

    2a.  Various experiments have shown that UV-B radiation damages fish
         larvae and juveniles, shrimp larvae; crab larvae,  copepods, and
         plants essential to the marine food web.

    2b.  Up to some threshold level of exposure, most zooplankton show no
         effect of exposure to UV-B radiation, but exposure above the
         dose/dose-rate threshold elicits significant and irreversible
         effects.

    2c.  While the exact limits of tolerance and exposure have not been
         precisely determined for any organism, estimates of these two
         properties for a wide variety of aquatic organisms show them to be
         essentially equal.

    2d.  The equality of tolerance and exposure suggests that solar UV-B
         radiation is currently an important ecological factor and the
         sunlight-exposed organisms sacrifice potential resources to avoid
         increased UV-B exposure to live in their particular niche.  Thus,
         even small increases of UV-B exposure would be likely to further
         injure the species currently under the most stress.

    2e.  A decrease in column ozone is reasonably likely to diminish the
         near-surface season of zooplankton.  Whether the population could
         endure a significant shortening of the surface season is unknown.

    2f.  Sublethal exposure of copepods produces a reduction in fecundity.
                          * * *  DRAFT FINAL  * *

-------
                                   12-3
    2g.   Of the animals tested, no zooplankton possess a sensory mechanism
         for directly detecting UV-B radiation; therefore it would be
         unlikely that they would actively avoid enhanced levels of exposure
         resulting from a reduction in column ozone.

    2h.   Exposure of a community to UV-B stress in controlled experiments has
         resulted in a decrease in species diversity, and therefore a
         probable reduction in ecosystem resilience and flexibility.

    2i.   There is a  predicted 8% annual loss of the larval anchovy
         population for a 9% reduction in column ozone in a marine system
         with a 10-meter mixed layer.

3.  SIGNIFICANT LIMITATIONS EXIST IN PROJECTING DAMAGE TO AQUATIC SYSTEMS
    FROM CHANGES IN ULTRAVIOLET RADIATION

    3a.   Because aquatic organisms are small and do not usually have fixed
         locations, it is very difficult to obtain accurate data needed to
         model the systems and verify results.  Understanding of the
         lifecycle of organisms is very limited.'

    3b.   In common with all other living organisms, the aquatic biota cope
         with solar UV-B radiation by avoidance, shielding, and repair
         mechanisms.   Uncertainty exists as to how much such behavior could
         reduce damage.

    3c.   Determination of UV-B exposure in aquatic systems is more complex
         than for terrestrial ecosystems because of the variable attenuation
         of UV-B radiation in the water column.

    3d.   Current data only indicate probable trends and are based on
         incomplete and limited observations.  Consequently they must be
         regarded as very tentative.  The possibility exists that the risks
         from UV-B radiation are smaller or larger than current experimental
         evidence appears to indicate.
                          * * *  DRAFT FINAL  * * *

-------
                                   12-4
INTRODUCTION

    As explained in Chapter 2, it is ozone,  though in very small
concentrations in the atmosphere, that is almost entirely responsible for  the
absorption of solar ultraviolet (UV) radiation in the 290-320 nm  waveband
(Exhibit 12-1).

    If Exhibit 12-1 were modified to account for a 10% ozone layer reduction,
given the scale, one would scarcely notice the difference.  The additional UV
radiation reaching the earth's surface as a result of a 10% ozone reduction
would amount to less than a 0.05% increase in the total sunlight  energy.   At
the very margin of the diagram, however,  the percentage increases would be
much larger because the base is so small  and it is in those wavelengths that
biological responses to ultraviolet radiation received from the sun appear
substantially greater.

    Based on models of depletion (see Chapter 19, Appendix A), a  10% ozone
reduction at 45°N latitude would result in a 28% increase for biologically
effective radiation for DNA damage (Setlow 1975).  The generalized plant
response would increase by 21% (plant damage) (Caldwell 1974). Exhibit 12-2
shows the relationship between ozone depletion and action spectra.  Because
of the relatively high increases in UV-B  at the lower ends of this part of
the spectrum and its relative biological  effect, ozone depletion  constitutes
a threat to aquatic life.

BACKGROUND ON MARINE ORGANISMS AND SOLAR  ULTRAVIOLET RADIATION

    Organisms in fresh water or the oceans are either swimmers (nekton),
bottom-dwellers (benthos), or drifters (plankton).  Plankton can  be either
plants (phytoplankton) or animals (zooplankton).  An important zooplankton
group, the ichthyoplankton, is comprised of the drifting eggs and larvae of
many species of fish.

    Phytoplankton provide essentially all of the chemical energy  required by
the marine food web through the photosynthetic process (primary production).
Zooplankton are consumers (grazers).  Zooplankton are usually less than about
5 cm in diameter or length, most often less than 1-2 mm.  Nearly  all groups
of aquatic animals, at least for some phase in their life-history, are
considered zooplankton -- for some species, in the egg and/or larval stage,
for others, throughout their life cycle.   Zooplankton are critical components
in typical aquatic feed webs (nutrient pathways) that lead to larger animals,
including those comprising the fisheries  (finfish and shellfish)  and
therefore affect man himself.

    In marine ecosystems, UV-B radiation penetrates approximately the upper
10 meter of the marine euphotic zone before it is reduced to 1% of its
surface irradiance.  Penetration of UV-B radiation into natural waters is a
key variable in assessing the potential impact of  this radiation on any
aquatic ecosystem (Exhibit 12-3).
                            * *  DRAFT FINAL  *

-------
                                   12-5
                              EXHIBIT  12-1


          Solar Irradiance Outside the Earth's  Atmosphere and  at the
             Surface of the  Earth  for a Solar Zenith  Angle of 60°
  (xlO1
LJ
      4
0 '

-------
                        12-6
                     EXHIBIT 12-2

  Relationship between  Ozone Depletion  and Biological
       Effectiveness of Increased UV-B  Radiation
      (Based on model  by Green, Cross, and  Smith)
                                                           Q
                                     Increase in Biological
Ozone
Decrease
10%
20%
30%
40%
Increase in
290-320 nm
8%
17%
27%
38%
UV Radiation
290-360 nm
1.1%
2.4%
3.7%
5.2%
Effectiveness
DNA
' 28%
67%
125%
213%
Plant
21%
49%
85%
132%
Action spectra are  referenced to 300 nm = 1.00.
                * *  *   DRAFT FINAL  * * *

-------
                                 12-7
                             EXHIBIT 12-3

             Solar Spectral  Irradiance at the Surface of the Ocean
                            and at Four  Depths
   600

LU
o 500
o:
   400
   300
LU

r 200
    100
                                        PERCENTAGE  OF ENERGY
                                        2   4  6 810   20  4060100,
          •£•  /«—Visible—*i
      0.2   0.4  0.6   0.8   1.0   1.2    1.4   1.6   1.8   2.0   2.2   2.4
                            WAVELENGTH
   Adapted  from Sverdrup,  Johnson, and Fleming 1942.  Insert illustrates
   percentage of energy transmitted at depth in oceanic waters.
                           - *  DRAFT FINAL  * * *

-------
                                   12-8
    Photosynthesis requires light, but adequate sunlight seldom penetrates to
the bottom of natural waters.  Thus, aquatic plants are largely confined to a
relatively thin layer at the surface of the water, which is termed the
euphotic zone, the surface layer where plants create more chemical energy by
photosynthesis than they use for their own metabolism.  This zone is where
UV-B can penetrate.  Zooplankton are found at all depths, depending on
species and season, but are most abundant in the sunlit upper 100 meter where
the bulk of their food, including phytoplankton, is found.  Many species
normally live very close to the surface, even in daylight, while others
occupy the near-surface layer during only part of their life cycle.  The
near-surface layer is a very important zone in the interactions of the
physical/chemical/biological components of aquatic systems.

    Zooplankton have apparently evolved mechanisms and behavior by which they
have adjusted to current levels of UV radiation (Damkaer 1982), but they may
not be able to adjust to relatively rapid increases in total UV exposure.  If
there are changes in abundance of Zooplankton species, those changes would
have an impact far beyond any direct effects because of the critical role
zooplankton play in energy transfer within the ecosystem.

Limitations of Pioneer Investigations

    As early as 1925, scientists were aware of damaging effects from the
ultraviolet component of sunlight on aquatic organisms (Huntsman 1925; Klugh
1929, 1930; Harvey 1930; ZoBell and McEwen 1935; Giese 1938; Bell and Hoar
1950; Dunbar 1959; Marinaro and Bernard 1966).  These reports are, for the
most part, of historical interest only; they cannot be strictly related to
present investigations because of a lack of precise or absolute measurements
of UV dose-rates and doses.  A lack of proper instrumentation for UV
irradiance measurements in aquatic research including laboratory studies,
persisted into at least the late 1960's.

Limitations of Recent Investigations

    There are only a few reports that were based on experimental methods
appropriate to the problem of enhanced solar UV-B radiation.  This meager
body of knowledge does not compare well to the enormous literature available
on, for example, chemical pollution.  Even for chemical pollution there are
still great uncertainties in predicting ecosystem responses.  It must be
recognized that attempts to predict ecosystem responses to enhanced UV-B
radiation are in the earliest stages.  A number of reports are difficult to
apply because the experiments used extraordinarily high UV-B dose-rates
and/or doses.  Frequently investigators increased radiation in damaging
wavelengths far beyond the amount that can be expected from ozone depletion.

    Another problem  in predicting UV-effects is that we are not yet very
close to being able to extrapolate the effects on one species to those on
another.  Nor can we say with much assurance how the same species might react
to UV radiation in a different environment.  There is simply too little data
of an acceptable quality to go beyond a general assessment of risk.  All that
follows must start off with these caveats.
                          * * *  DRAFT FINAL  * * *

-------
                                   12-9
EFFECTS OF UV-B ON RADIATION  PHYTOPLANKTON

Direct Impact on Primary  Productivity

    The amount of UV-B radiation reaching the ocean's surface has long been
suspected as a factor influencing primary production.  Research shows
convincingly that ultraviolet radiation,  at-levels currently incident  at the
surface of natural bodies of water,  has an influence on phytoplankton
productivity (Steemann Nielsen 1964; Jitts et al.  1976; Hobson and Hartley
1983).  It has been calculated that, near the surface of the ocean, a  25%
reduction in ozone concentration would cause a decrease in primary
productivity by about 35% (Smith et  al. 1980).  The estimated reduction in
production for the whole euphotic zone would be about 10%.  Because these
calculations were based on attenuation lengths, (i.e., the product of  depth
in the water column and the diffuse  attenuation coefficient of water), waters
of various turbidities and absorption characteristics can be compared.

    If one assumes that present phytoplankton populations sense and control
their average vertical position in such a way as to limit UV-B exposure to a
tolerable level, then a 10% increase in solar UV-B radiation would
necessitate a downward movement of the average position, thereby reducing the
average UV-B exposure by 10%.  There would be a corresponding reduction of
light for photosynthesis.   The loss  of photosynthetically active radiation in
many locations might not be significant.   However, in some very productive
areas, especially high latitude ocean areas, photosynthetically active
radiation is the primary limiting factor for marine productivity
(Russell-Hunter 1970).  The loss of  photosynthetically active radiation from
optical measurements has been estimated to be in the range of 2.5% to  5% for
a 10% increase in UV-B radiation (Calkins 1982).  If the photosynthetic base
of aquatic ecosystems were perturbed, one would expect ramifications to
extend up through the food web through predator-prey relationships.

    Experiments with marine diatoms  have shown significant reductions  in
biomass, protein, and chlorophyll at UV-B irradiances equivalent to ozone
reductions that would occur for a 5%-15% ozone depletion.  Laboratory studies
on chain-forming diatoms and other phytoplankton show that growth increased
when UV-B radiation was filtered out of the incident solar radiation,
indicating that existing levels of UV-B radiation depress productivity
(Thomson, Worrest, and Van Dyke 1980; Worrest 1982).

Effects  Because of Changes in Motility

    An additional effect is that UV-B radiation may endanger the survival of
microorganisms  (Euglena, slime mold, some blue-green algae) by decreasing
their motility and by inhibiting phototactic and photophobic responses
(Haeder 1986).  The inability of a population to move into favorable
environments, could result in damage that may impair development.
                                 DRAFT FINAL  * * *

-------
                                   12-10
Effects  on Photosynthesis

    Jones and Kok (1966) observed a direct inhibitory effect of UV radiation
on photosynthesis in chloroplasts.   This form of inhibition is a short-term
response and shows a different UV-B wavelength dependence from most other
forms of UV-B injury.  Smith et al. (1980) found the short-term depression of
photosynthesis by natural solar UV radiation measured in the ocean resembles
the direct photoinhibition described by Jones and Kok.  If direct
photoinhibition were the only pathway for reduction of photosynthesis by
marine phytoplankton, the effect of ozone reduction would be minor; however,
Smith and Baker (1980) point out that there could be long-term effects that
might show a different action spectrum and be of more consequence.

    In Hawaii long-term photoinhibition of growth of six algal species under
natural sunlight has been measured (Jokiel and York 1984).   Two strains could
not grow at all at the levels of irradiance for full natural surface
sunlight.  Of the species tested, those collected from tropical surface water
showed the greatest adaptive power, but it is reasonable to conclude that
resistance to solar UV radiation is achieved through expenditure of resources
that can be better applied to other needs in less exposed species.

    Worrest et al. (1981) found a short-term (4-5 h) depression of
photosynthesis in seven strains of algae upon exposure to enhanced levels of
UV radiation.  There was a large difference in strain sensitivity.  The
observed depression appeared to linearly correlate with UV-B dose using the
DNA (Setlow 1974) action spectrum.   This is in contrast to the results of
Smith et al. (1980) who found the Jones and Kok weighting function to be
appropriate.

Multi-Day Effects

    Measurements of photosynthesis that span only a single day could
underestimate the overall action of solar UV exposure by failing, to account
for effects on the next day's population level.  The effects of UV-B
radiation could result in delaying growth or direct killing.  Thus,
subsequent population could fall below the numbers that an unexposed
population would attain.  Prolonged delays (about two days) in growth of the
survivors have been observed in two strains of the diatom Thalassiosira,
which had been irradiated with simulated solar UV-B radiation at doses below
lethal levels.  If unicellular organisms are in a rapid growth phase, a
growth delay equalling the time of one growth cycle produces the same effect
on the subsequent population as would be produced by a 50% killing (without
growth delay).

Shifts in  Community Composition and  Possible Implications

    Shifts in species diversity and community composition of phytoplankton
communities have been observed in a simulated marine ecosystem exposed to
UV-B equivalent to ozone decreases of 15%, (Worrest et al. 1981).  Natural
communities appear to show changes in species composition rather than a
decreasing net production.  Experiments in simulated marine ecosystems seem
                            * *  DRAFT FINAL  * *

-------
                                   12-11
to indicate that higher UV-B radiation will not decrease biomass and
chlorophyll accumulation, (Worrest et al.  1981).   However,  a change in
community composition might result in a more unstable ecosystem and might
have influence on higher trophic levels (Kelly 1986).  For  example, one
effect of enhanced levels of UV-B radiation would be to alter the size
distribution of the component producers in a marine ecosystem.   Increasing or
decreasing the size of the representative primary producers could alter the
energy allotment that grazers use for finding and consuming food.  This could
reduce the feeding efficiency of the consumer, which would  have a variety of
effects.  In addition, the food quality of the producers is altered by
exposure to UV-B radiation.  It has been demonstrated that  the protein
content, dry weight, and pigment concentration are all depreseed or the
composition altered by enhanced levels of UV-B radiation (Doehler 1984,
1985).  Consequently, a UV-induced shift in the distribution and abundance of
different kinds of aquatic plant organisms could have important ramifications
throughout the ecosystem.

EFFECTS ON  INVERTEBRATE ZOOPLANKTON

Background

    There have been several studies on the effects of UV-B  radiation on a
variety of zooplankton.  Investigators have reported that the effect of
increased UV-B radiation on some marine zooplankton (e.g.,  copepods, shrimp
larvae, crab larvae) is to increase the mortality of the organisms and to
decrease the fecundity of the survivors.  Regardless of the species
investigated, the potential exists for a disruption of the  food web, which
would indirectly affect fish populations that are important as food.  Of
course, studies that consider economically important zooplankton species,
such as larval stages of certain shrimp, crab, and fish, would have
importance to man because damage to these species would directly reduce the
amount of food for human consumption.

Methodological  Issues

    How well photorepair mechanisms might mitigate damage is an important
issue.  It appears that up to some level (daily-dose threshold), photorepair
will mitigate damage.  However, it also appears that near-surface exposure
levels that could accompany a significant reduction in column ozone would
exceed threshold levels.

    The issue of thresholds is important.  It appears that  regardless of the
cellular-level responses to UV irradiation, it is usually noted that up to
some level of UV exposure, the organism appears to suffer no harm.
Extrapolating survival curves for different UV exposure levels would indicate
that organisms can cope with the damage produced by some levels of additional
UV radiation without negative effects.  At greater doses, where the survival
curve becomes steep, it appears that the repair systems may become
inactivated by the radiation, or that the damage to the general tissues has
gone beyond the capacity of the repair systems.
                          * * *  DRAFT FINAL  * * *

-------
                                   12-12
    A study by Damkaer et al. (1980) illustrates the importance of threshold
effect.  The authors found no alterations in activity, growth-rate,
developmental rate, and survival in a number of shrimp and crab larvae
exposed to low levels of artificial UV-B radiation.   However,  when a certain
combination of UV dose-rate and total dose was exceeded, negative effects
became quite sharply noticeable.

Lethal Effects of Enhanced UV-B Radiation

    Damkaer et al.  (1980) have done a number of studies in which enhanced
UV-B radiation produced irreversible damage and death.  UV thresholds, based
on the lowest level or irradiance where statistically significant (exceeding
the 95% confidence limits of five replicates)..  Irreversible effects occurred
compared to the controls.  Damkaer et al. (1980) have compared the effective
UV thresholds with the maximum effective solar daily dose rates at an
experimental site in Washington state (Exhibit 12-4).  The results indicated
that the maximum solar levels are currently exceeded for all tested
organisms.  Obviously, these animals can survive threshold values for maximum
solar UV levels.  To be effective, UV thresholds probably must be exceeded
during several consecutive days.  A more realistic approach would be to
compare the threshold levels with median solar UV levels (Exhibit 5).  The
threshold for all groups would then appear to be above the present median
solar incident UV levels, at least until late in the time span of natural
occurrence near the surface.  The data seem to suggest that UV levels have
exerted a considerable influence in the long-term adjustments of these
populations to those specific seasons.  Of course, many other physical as
well as biological factors operate concurrently, and the organism's life
cycle is a compromise with the total environment.  Late in the surface
season, it is possible that natural UV levels exert a detrimental effect,
particularly on the shrimp larvae, which have a lower total dose tolerance
than crab larvae.

    Damkaer et al.  (1980) compared the estimated effective UV daily doses
simulating various ozone reductions at the experimental site with the UV
survival thresholds (Exhibit 12-5).  Obviously, the probability that UV
thresholds would be exceeded increases with diminishing ozone.  With a 40%
ozone reduction, there appears only to be a short "window" of safety at the
beginning of each group's surface season.  Even a 20% ozone reduction
significantly shortens the season.  Whether or not the populations could
endure with a drastically reduced time of near-surface occurrence is not
known.  Success of any year-class depends on the timing of a great number of
other events besides UV level.  Under current stress conditions, early larvae
may do well one year, whereas only late larvae may survive in a subsequent
year.  However, an additional stress like enhanced levels of UV-B radiation
is not likely to be beneficial.  Combined with other stresses that make early
year survival impossible, it could devastate populations.

    That larval stage development appeared to be unaffected below survival
threshold UV levels is an important observation, since the survival of the
unharmed larvae is critical to their proper and timely development.  Beyond
the threshold levels, however, development and survival rapidly declined.
                            * *  DRAFT FINAL

-------
                                 12-13
                            EXHIBIT 12-4

        Lethal  Effects on Shrimp Larvae for  Various Combinations of
                      UV-B Dose-Rate and Total Dose
                                          Feb-Apr Mean Daily Dose (DNA)
Ozone
400

1
c
m
Jo 30°
i
LO
GO
CM

<200
o
CJ
j= 100
LU
00
O

• Surface 1m Reduction
1 A a 0%
! B b 16%
1
c




o


0 0
_ 0 °
o
C c 40%

*

•
at
•
9 *
• *
m *
id '•'
o r
n \j
8 a" c * 1 1
U - n
i 1 1 III II l II 1 i 1 1 1 i 1

_, 20 40 60 80 100
iS DAILY DOSE [Jm~2l A (285-315 nm)
o DNA
h-
i .
i l i i i i i
    .002
DOSE  RATE
                                    .004
                                        -2
                                            DNA
.006        .008
 (285-315 nm)
Open circles indicate maximum observed combinations for survival; closed
circles  indicate minimum observed combinations  for mortality.   Open vertical
bars indicate mean February-April surface solar UV-B daily dose-rates at
various  atmospheric ozone concentrations; closed vertical bars  indicate mean
daily dose-rates at 1 m (adapted from Damkaer et al. 1980).
                         * * *  DRAFT FINAL

-------
                                       12-14
                                  EXHIBIT 12-5

                  Estimated Effective  UV-B  Solar Daily Dose at Varoius
                  Atmospheric Ozone  Concentrations  Based on  4-Year Mean
                     of Medians, Manchester, Washington, 1977-1980
C\J
      400r
      350
300
      250
LJ   200
GO
§   150

i   100
                    SUN +  SKY
             a. Shrimp  larva,
                 euphausid larva

             b. Adult euphausid,
                 crab zoea
             C. Crab megalopa
                                                            Ozone
                                                          reduction
                                                      40%
                                                      25%
                                                      16%
                                                      7.5%
                                                       0%

                                                      4-year mean  of medians
                                                      Manchester, Wa.
                                                      1977-1980
              Mar
                          Apr  May June July  Aug  Sept  Oct  Nov
                                   MONTH
       Dose  increments are  from Gerstl, Zardeki,  and Wiser 1986.  Also shown  are
       approximate thresholds of UV-B  daily dose  for principal  experimental
       zooplankton groups,  in natural  seasonal position (adapted  from Damkaer et al.
       1981).
                               * * *  DRAFT FINAL  *

-------
                                   12-15
The apparent UV thresholds are near-current incident surface UV levels.
Nevertheless, the implications of these experiments for zooplankton in nature
remains in some doubt because of lack of information on the subsurface UV
spectrum, the effects of vertical mixing, unknown in situ behavior of
irradiated zooplankton, and unknown sublethal and pre-lethal effects.

    The degree to which other zooplankton groups would be affected by these
levels of UV-B irradiation cannot be predicted.  Although, in crustaceans the
same basic tissues would likely be involved, and therefore the response would
be qualitatively similar; species vary substantially insensitivity to UV-B
radiation.  If the accumulated level of UV-B radiation is the most important
factor in damage, then increased UV radiation through ozone depletion may
significantly affect even species that are moderately resistance to short
exposures of high UV intensities.

    Copepods are the most important zooplankton from the perspective of
ecosystems.  Karanas, Worrest, and Van Dyke (1979) reported on the survival
of developmental stages of the estuarine copepod Acartia clausii exposed to
daily UV-B dose approaching natural (Oregon coast) daily doses.  Although the
total doses in this experiment were relevant, the dose-rates exceeded natural
dose-rates by about a factor of two.

    It is known that organisms are protected to some degree by their own
pigmentation, and this suggests that organisms may have evolved pigmentation
as a consequence of the damaging effects of solar radiation.  Damkaer (1982)
reviewed several possibilities for this in regard to zooplankton.  Siebeck
(1978) found a direct relationship between pigmentation in cladocerans of
mountain and lowland lakes and their apparent natural UV exposure.  In field
and laboratory experiments, the more-pigmented cladocerans tolerated the
greater exposures.  Siebeck also noted a marked photorepair potential.

    Hairston (1980) has drawn attention to the relationship between vertical
distribution and pigmentation in some freshwater copepods.  Typically the
most pigmented forms are closest to the surface.  Because the pressure of
predation should be against the conspicuous pigmented forms, Hairston
concluded that "photodamage may be a more important selective force than
previously supposed."

    That these kinds of copepods do have a UV-tolerance related to their
pigmentation has been verified in laboratory experiments.  Ringelberg,
Keyser, and Flik (1984) studied the UV-B tolerance of a red-pigmented and a
translucent variety of the same copepod species from separate, though nearby,
lakes in France.  The red copepod tolerated higher artificial UV-B doses.
The experimental doses were obtained by exposure to a single dose-rate for
varying amounts of time.  Sharp mortality increases began in the translucent
form at about half the total dose as for the pigmented form.

    An interesting aspect of these experiments is that in all cases where
death occurs, the dose-rate needed to reach the threshold'effect for shrimp
(Damkaer et al. 1980).  The total doses tolerated by the two copepod
color-forms bracket the total dose threshold of shrimp larvae.  The tolerance
                          * * *  DRAFT FINAL  * *

-------
                                   12-16
limit appears to be the same in vulnerable zooplankton,  suggesting perhaps a
universal tolerance limit in near-surface organisms of the structure and size
of shrimp larvae and copepods.   It would be useful to conduct additional
experiments with such copepod color-forms and include exposures to lower and
higher dose-rates at more natural periodicites.   Also, Ringelberg et al.
observed no photorepair, but indicated that the intensity of visible
radiation in their experiments  was extremely low (~0.4 W m-2).

    Thomson (personal communication) has demonstrated that current levels of
UV-B radiation are of significance in the developmental life of several
species of shellfish.  For some species a 10% decrease in column ozone could
lead to as much as an 18% increase in the number of abnormal larvae
produced.  Exhibit 12-6 is a compilation by Thomson of data denoting the
estimated biologically effective UV-B doses leading to significant effects in
major zooplankton groups.

Sub lethal Effects of UV-B

    There are many examples of  sublethal stresses (of all kinds) that can
affect natural populations.  If stresses alter the reproductive capacity of
organisms, effects can be significant and rapid.  Karanas, Worrest, and Van
Dyke (1981) reported a possible mechanism through which enhanced solar UV-B
radiation could have a large impact on copepod populations without direct
killing.  The estuarine copepod Acartia clausii was irradiated with UV
comparable to mid-summer surface irradiance at 45°N.  After five days, the
surviving copepods were mated.   If either (but not both) parent had been
irradiated, the egg production  and subsequent hatch of the female was
significantly reduced.  If both parents had been irradiated, the production
of viable larvae was halved.  The subsequent survival of the hatched larvae
from irradiated crosses did not appear to be significantly affected.

    Even though Karanas, Worrest, and Van Dyke (1981) applied a natural-level
surface dose-rate of artificial UV-B radiation,  this dose-rate was still
rather high and probably represents an extreme situation.  To obtain
reasonable total doses, the copepods were exposed for only 1-2 hours.  Real
dose-rates, especially at depth, would be considerably less.  These copepods
are found at the surface, as well as some distance below, but their
distribution on the required scale of 1 m or less is not well known.
Acartia clausii are taken in a wide variety of water types, but generally
this is a coastal/estuarine species.  In its usual habitat, effective UV-B
irradiation may not extend beyond a few meters depth.  In view of the
apparently profound sublethal,  reproductive response to UV-B radiation, it
would be of great interest to have additional multi-generation observations.

Possible Photorepair of  UV-B Damage and Its  Implications on Using
Experimental Data

    The photochemistry and photobiology of UV-B-induced damage, particularly
nucleic acid disruption, as well as the several molecular repair mechanisms
that are consistently found in plants and animals, have been reviewed by
Caldwell and Nachtwey (1975).  Among the repair mechanisms, the quantitative
                          * * *  DRAFT FINAL  * *

-------
                                 12-17
                            EXHIBIT 12-6

          Estimated Biologically  Effective UV-B Doses Leading
       to Significant Effects in Major Marine Zooplankton  Groups
Zooplankton Group
Shrimp/Euphausiid larvae
Euphausiid adults
Copepod larvae
Copepod postlarvae
Copepod adults
Crab larvae
Crab postlarvae
Anchovy /Mackerel
Oyster/Mussel larvae
Dose Rate
W m-2 (DNA eff)S
6.0 x 10-2
9.9 x 10-2
3.4 x 10-1
1.6 x 10-1
3.4 x 10-1
9.9 x 10-2
2.8 x 10-1
6.0 x 10-2
1.2 x 10-1
Total Dose
J m-2 (DNA eff)
2.6 x 103
6.5 x 103
1.4 x 103
2.7 x 103
3.0 x 103
6.5 x 103
6.0 x 104
2.5 x 10*
2.2 x 103
Time For
Effect
4 days
6 days
1.0 hour
4 . 5 hours
2.5 hours
6 days
20 days
12 days
5 . 0 hours
  DNA action spectrum is referenced to 300 nm = 1.00.

Complication by Thomson.
                        * * *  DRAFT FINAL  * * *

-------
                                   12-18
                              EXHIBIT 12-7

                  Annual  Fish Production  in Coastal Waters
                      Baseline Data  for Coastal Waters

Primary production,
kJ m-2 yr-1, (P1)
Ecological efficiency per
trophic level, (e)
Number of trophic levels from plant
production to fish production, (n)
Fish production,
Baseline
4200
0.15
3
14
-5% P1
4000
0.15
3
13
-5% e
4200
0.14
3
12
   kJ m-2 yr-1, (PJen)

Total fish production, 109 kg yr-1
  (using Winberg's transformation,
  4.2 kJ = 1 g fish flesh)
121(100%)     115(95%)    103(86%)
Adapted from Ryther (1969).
                          * * *  DRAFT FINAL  * * *

-------
                                   12-19
capacity of the dark-repair systems is quite different from those systems
that require light.  Under typical experimental conditions, it is presumed
that the dark-repair mechanisms could function freely.  Dark-repair systems
appear to offset a very high proportion of the damage caused by relatively
low doses of irradiation.  However, photorepair mechanisms often appear
capable of repairing only so much damage, even over a wide UV-B dose range
(Murphy 1975).  Consequently, the restorative effects of photoreactivation
may be regarded as particularly critical in situations where organisms are
subjected to UV-B irradiation near and above tolerance limits.  There has
been an implication in some review papers (e.g., NAS 1984) that UV-B damage
will be overcome by the maximally functioning photorepair mechanism.  Up to
some level (daily dose threshold), this is probably true, but beyond that
threshold there will be some residual, irreversible damage.

    Experiments with artificial sources of UV-B radiation usually are
conducted in the laboratory under light regimes using more than an order of
magnitude less visible light than found in nature.  Near UV and/or visible
light is needed for photoreactivation, and the accuracy of some observations
may have been influenced by less-than-maximum photorepair in the laboratory.
Experiments with shrimp larvae and adult euphausiids (Damkaer and Dey 1983),
comparing survival at various UV-B doses and dose-rates combined with
different levels of near-UV and visible irradiance, suggest that photorepair
occurs in these organisms.  The dramatic sensitivity of both shrimp larvae
and adult euphausiids when deprived of photoreactivating light clearly
demonstrates the relative importance of photorepair under UV stress.
However, this apparent photorepair reaches maximum levels at relatively low
levels of visible irradiance.  Consequently, it is doubtful that photorepair
and related survival would have been greater if experimental levels had been
equal to solar levels of irradiance.

    A similar relationship has been reported between survival in UV-damaged
anchovy larvae and level of irradiance of photoreactivating light needed for
repair (Kaupp and Hunter 1981).  The level needed for maximum photorepair is
far less than ambient solar irradiation.  The photoreactivation experiments
on shrimp larvae and euphausiids suggest that the previously derived
dose/dose-rate thresholds are not altered with additional photoreactivating
light.  Clearly, a significant lack of photoreactivating light, which
normally would be concurrent with solar UV-B irradiance, would certainly
depress threshold levels.  Clearly, past experiments still have not
approached the intensity or dose of local solar photoreactivating light.
However, for UV-B exposures below threshold levels maximum photorepair
potential has been reached at the low levels (2-9 W m-2) of
photoreactivating light.  Nearly doubling the dose-rate and dose of
photoreactivating light did not increase the survival of euphausiids.  With
anchovy larvae, Kaupp and Hunter (1981) reported that the photorepair
mechanisms appeared to be "fully activated" at a photoreactivating  light
(320-500 nm) irradiance of 7 W ,m-2 and a daily dose of 77.5 kJ m-2
available over four days.  At greater irradiance levels, up to levels
equivalent to solar photoreactivating light at 45°N latitude, Kaupp and
Hunter observed no further increase in anchovy survival.  There is
extraordinary close agreement between the relatively low dose-rates and doses
                          * * *  DRAFT FINAL  * * *

-------
                                   12-20
necessary for apparent maximum photoreactivation in anchovy larvae and
euphausiids.  Thus, most of the experimental data cannot be eliminated on the
basis of failing to account for photorepair.

    Another important point is that from these observations, it is clear that
in nature there would always be sufficient photoreactivating light to provide
maximum photorepair potentials in shrimp larvae and euphausiids (and also in
anchovy larvae).  But it must be emphasized that even maximum photorepair may
not be sufficient to reverse all potential injuries caused by UV-B radiation.

Possible  Vertical Distribution and Its  Implications  on Using
Experimental Data

    The actual solar radiation to which surface-living zooplankton are
actually exposed in nature is very important in assessing risks to higher
ambient UV-B.  Given the large effects UV radiation had on zooplankton and
the apparent existence of threshold levels, understanding UV-B in the natural
environment is critical to assessing risks.  To assess the potential dangers
to zooplankton populations from increasing solar UV-B irradiaton requires
consideration of the subsurface irradiance levels.  The data for the shrimp
larvae shown in Exhibit 12-4 show that the mean of the observed daily UV-B
surface doses (current ozone thickness) between mid-February and mid-April is
within the experimentally determined safe zone.  Daily doses calculated for a
1-m depth are therefore even farther within the safe zone, so that presumably
the shrimp larvae could tolerate these mean dose-rates for long periods.
With a 16% or 40% ozone reduction, the total-dose thresholds would be
exceeded at the surface only if the mean dose-rates are attained for two or
three consecutive days (two sunny days).  Even then, however, the calculated
mean UV-B levels at 1 m would remain in the safe zone.  Understanding the
implications of higher UV-B requires knowing where the shrimp larvae spend
their time.

    The actual vertical distributions of the shrimp in nature are not known.
In the laboratory, young larvae are attracted to light and definitely prefer
the near surface.  The larvae do not appear to avoid high UV-B irradiance
(Damkaer and Dey 1982),  It is not likely they can sense this radiation in
sufficient time to prevent lethal damage.  Damkaer and Dey assumed that the
younger shrimp larvae are constantly moving toward the surface, and that
during the day the majority of the population is in the upper meter.

    Vertical motion makes it difficult to determine the probability that
particular shrimp larvae would receive lethal exposures of UV-B radiation; in
Exhibit 12-8, an approximate estimate is made.  The probability is obviously
time-dependent.  At the surface, shrimp larvae are unlikely to encounter
lethal daily or bi-daily UV-B exposures until the end of the surface season
(mid-April).  This is also about the time that the mean noon dose-rate
exceeds the dose-rate threshold.  Before then, the shrimp larvae would be
exposed daily to a one-third total dose limit for three consecutive days.
This level is in the safe exposure zone because the dose-rate would be low
(Exhibit 12-8).  At 1 m, the mean dose-rates in mid-April are still less than
the shrimp's threshold dose-rates, yet the surface probabilities for lethal
                          * * »  DRAFT FINAL  * * »

-------
                                    12-21
'E
 -7>

 LD
 CO
 LU
 CD
 O
 O
LU
O
cr
UJ
a.
                               EXHIBIT 12-8


            Percentage of Total Dose Limit Likely to be  Reached on Any
           Particular Day; Lethal  Doses Accumulated Only After Dose-Rate
                            Threshold is Exceeded
                                                             O
100


 90


 80


 70


 60


 50


 40


 30


 20


 10
            J	I
          23

          Feb
4  1
234

Mar
123
   Apr

WEEKS
4   1
2  3
May
4  1
234

June
    Circles are 4-year weekly means of surface daily UV-B doses; x's are  these
    doses calculated at 1 m (adapted from Damkaer et al. 1981).
                                  DRAFT FINAL  * *

-------
                                   12-22
exposure may be sufficiently high at that time to limit the seasonal
occurrence of the larvae.  With a 16% ozone reduction, the current mid-April
UV-B conditions would be reached near the end of March.  If UV-B radiation is
a contributing factor in regulating the seasonal occurrence of shrimp larvae,
ozone reductions may act to shorten the surface time available to the
larvae.  The seasonal relationships in the UV-B tolerance of different
zooplankton groups (Exhibit 12-8) suggest that UV irradiation is an important
environmental factor.  Hunter, Kaupp, and Taylor (1982) postulate that the
surface occurrence of eggs and larvae of a number of clupeid fishes are
already limited by UV-B.

    Knowing that shrimp larvae are limited by UV-B already, and could be
further limited, does not provide us with an assessment of damage.  Data on
population dynamics and recruitment of shrimp stocks are insufficient to
estimate the impacts of shortened surface seasons.  It is possible that
ordinarily the greatest number of successful recruits are normally derived
from early-season larvae and that increased UV-B irradiance would then have
no appreciable direct effect on shrimp population stability.  It is also
possible that late-season larvae are essential to maintaining local popula-
tions.  In fact, it is possible that the situation varies from year to year
with other environmental stresses, so that having a long season is critical
to sustaining the population in the long term.  Thus ozone reductons may lead
to no change, to a loss in productivity, or to extinction of these shrimp in
particular locations.

    It is clear that increases in UV radiation through depeletion of ozone
would create an additional stress to the seasonal near-surface stages of
these species.  It is quite possible that natural UV levels have already had
a selective role in the seasonal adaptation of these species.  The question
remains:  how much of their surface-living time could these species forego
and still maintain survivor-level populations?

Avoidance and Its Implications In Assessing Risk

    Investigation of the biological effects of UV-B radiation on aquatic
organisms have included a variety of approaches, ranging from histological
analyses to simple survival studies.- Most of this work has been concerned
with the direct effects of natural or enhanced UV-B radiation.  The
assumption implicit in some of the conclusions is that the organisms that
live in the near-surface layer will be passive recipients of ambient UV-B
radiation, regardless of its irradiance level.

    Because some of these near-surface organisms appear to be living near
their UV tolerance limits (Damkaer et al. 1980) it is possible that they have
evolved a sensitivity to fluctuations in UV-B radiation.  In addition to
photorepair mechanisms, they may possess an ability to avoid harmful levels
of UV-B radiation by appropriate defensive behavior.   In moderately
prdouctive ocean waters, the DNA-weighted irradiance at 1 m is only about 40%
incident on the surface  (Smith and Baker 1979).  Under these conditions, the
capacity to sense harmful UV-B radiation and to avoid  it by simply swimming
or sinking a meter below the surface would provide considerable protection.
                          * * *  DRAFT FINAL  * * *

-------
                                   12-23
    However, even if sensory mechanisms exist to warn an organism of
dangerous UV-B radiation, the particular stimulus that evokes the avoidance
response may not be effective under a new spectral distribution caused by
ozone depletion.  A reduction in the ozone layer would not increase UV-A and
visible radiation.  If the avoidance response is linked to high levels of
visible irradiance, for example, the avoidance response would not be
activated by increased depletion.  It seems likely that only a sensory system
that directly detects irradiance in the UV-B range and does so before
irreparable damage occurs will afford surface-living organisms a safety
potential.

    Experiments by Damkaer and Dey (1982) suggest that shrimp possess no
organ to sense UV-B.  They observed differences in behavior between
irradiated and non-irradiated shrimp larvae (Pandalus platyceros) or
copepods (Epilabidocera longipedata).   The irradiated specimens maintained
their near-surface positions in visible light (as did the controls) until the
time of decreased activity and death at about four days.  Adult euphausiids,
on the other hand, seemed to avoid this same level of visible light (with or
without UV-B radiation) but were attracted to low levels of visible light
with or without extremely high levels of UV-B radiation (Damkaer and Dey
1983).  These investigators also found that the zoea stage of the shore crab
(Hemigrapsus nudus) is extremely attracted to strong visible light.  No
difference in response-time (seconds) was noted between control larvae and
larvae receiving lethal UV-B doses until decreased activity and death of the
irradiated larvae at about 10 days.  There was no apparent reluctance of the
crab larvae to swim toward and to hold themselves within a zone producing a
lethal dose of UV-B radiation.

    These experiments suggested that there were no differences in behavior
between UV-B irradiated organisms and the control organisms until lethal
doses of radiation reduced their activity.  This is not considered to be an
active response to UV radiation, and it would have no value in avoidance.
The apparent inability to perceive potential danger from UV-B exposure
occurred at dose-rates well above established laboratory thresholds and at
exposures to total doses that, in most experiments, surpassed lethal
total-dose thresholds for similar animals (Damkaer et al. 1980).  If the
animals tested in these studies possess a behavioral mechanisms for
protection from dangerous levels of irradiation it seems unlikely that it
could be based on the direct sensing of UV-B radiation.  The zooplankton
tested seemed to be attracted to wavelengths of radiation longer than those
in the UV-B range, and the additional exposure to high  levels of UV-B
irradiance did not alter their short-term behavior.  Within the limits
examined, these animals generally positioned themselves as near the light
source as possible.  That organisms continue to seek out a strong light
source even while doomed from past UV-B exposures demonstrates not only the
strength of this photo-positive response but also, probably, their inability
to independently discriminate between safe and dangerous levels of UV-B
irradiance.  Since reductions in atmospheric ozone would allow more incident
UV-B (and create no change in visible irradiation), the currently evolved
response of some organisms to future solar irradiation  is unlikely to be
optimal.
                          * * *  DRAFT FINAL  *

-------
                                   12-24
Effects  of  UV-B  on Biological Diversity  and Subsequent
Impact on  Food Web

    Almost all life ultimately depends on plants, and in the ocean most plant
life is planktonic (phytoplankton).   To function, plants must receive
sunlight, and plants have evolved to require (and tolerate) more or less
sunlight, depending on the species.   For plants, their balance with current
levels of solar radiation is therefore critical.  Ultimately, animal life
exists because it obtains fixed chemical energy from plants.  The interaction
of life forms in obtaining energy makes them interdependent, forming food
chains,  or more frequent complex food webs.  Animals feeding directly on
plants also have evolved receptor mechanisms for solar light.  These
adaptations typically involved movements and secondary responses to visually
feeding predators.

    Even if UV-B caused an adverse effect in only a few marine organisms,
others are likely to be affected.  There have been several studies indicating
that reduced UV irradiation leads to increased productivity in
phytoplankton.  Only a very few experiments have subjected phytoplankton to
reasonable enhancements of UV radiation.  Some of these experiments (e.g.,
Thomson, Worrest, and Van Dyke 1980) suggest that phytoplankton have
resistance to above-ambient UV levels.  Also, Worrest (1982) cites a study by
Wolniakowski in which growth rates of a flagellated phytoplankton, initially
depressed by exposure to enhanced UV radiation, returned to the base rate
after a few days.  Finally, in field observations, Hobson and Hartley (1983)
found that summer UV levels did not depress phytoplankton photosynthesis.
This is not unexpected since it may have resulted from a combination of
seasonal succession of UV-resistant species and adaptation to increased UV
radiation.

    The most likely direct negative effects of higher UV radiation on
phytoplankton would be a change in community composition (Worrest et al.
1981); primary production may not suffer any decrease.  The species
composition that follows is likely to have differences in size and/or
nutritional value.   The impact of this kind of change, at the lowest level of
the food web, could be marked and severe.  One quite likely effect is a loss
of diversity.  Diversity is associated with stability in ecosystems, allowing
alternate routes or choices within food webs.  In nature, stresses are often
best measured not by changes in productivity or population size, but by
changes in species diversity.  With loss of species, an ecosystem loses some
of its natural resiliency and flexibility.  Other "shocks" to the system
become potentially more damaging.  Unfortunately, our knowledge base is
insufficient to make any definitive assessments.  All that can be done is to
argue by analogy.  Experience has shown that a loss in diversity leads to
instability in ecosystems, which can cascade to create significant change.

EFFECTS ON ICHTHYOPLANKTON (FISHERIES)

    As explained with invertebrate zooplankters, regardless of the
cellular-level responses to UV irradiation, it is usually noted that up to
some UV level there is no apparent effect on ichthyoplankton.  At greater
                          * * *  DRAFT FINAL  * * *

-------
                                   12-25
doses either the repair systems themselves may become inactivated by the
radiation, or the damage to the general tissues is beyond the capacity of the
repair systems.  To be effective, these threshold levels probably must be
exceeded during several consecutive days.  The apparent UV thresholds are
near current-incident UV levels.  In one experiment the thresholds for all
groups appeared to be above the present median solar incident UV levels at
the test location, at least until late in the time span of natural occurrence
near the surface.  An ozone depletion of 20% would significantly shorten this
season.  The effects of this shortening on annual productivity is not known.
Whether or not the population could endure with a drastically reduced time of
near-surface occurrence is not known.  Success of any year-class depends on
the timing of a great number of other events in addition to levels of
exposure to UV-B radiation.

    Hunter, Kaupp, and Taylor (1979) exposed anchovy eggs and larvae and
mackerel larvae to high doses of UV-B radiation in small closed containers.
They reported that the dose that killed one-half of the test organisms
(LD,-n) is clearly not an acceptable criterion for biological effect.  They

recognized that "it seems unlikely that any of the damaged survivors,
regardless of dosage, would be able to feed successfully."  Hunter, Kaupp,
and Taylor (1982) recognized that "none of the larvae surviving LD   doses

at age 4-days would survive to age 12-days."
                                                                       I
    Apart from the fact that laboratory experiments can never duplicate
natural responses, the subsequent experiments on UV effects by Hunter, Kaupp,
and Taylor (1981, 1982) involved some additional shortcomings that must be
kept in mind while evaluating the results.  First, it is often difficult to
maintain test organisms in good condition.  For some common animals, it is
not yet possible to keep them alive in the laboratory, much less experiment
meaningfully with them.  For most of the other animals the laboratory
conditions themselves exert a powerful negative influence, so that a true and
isolated measurement of UV stress is not attainable.  This can be seen in the
high 12-day mortalities (average 61%) in the controls for the anchovy
larvae.  These larvae were kept in small closed containers, a poor but
perhaps necessary approximation of the open ocean.  There can be no doubt
that these larvae were under significant stress even without UV radiation.
Hunter et al. refer to wide fluctuations in temperature in the solar-light
experiments; oxygen was probably also a limiting factor.  Perhaps a
flow-through seawater system or another species of fish larvae would
eliminate this serious problem.  As in some agricultural field tests
(Teramura 1986), actual field tolerance of UV radiation is greater than that
determined in the laboratory.

    The second confounding factor in the experimental design of Hunter,
Kaupp, and Taylor is that the tests apparently ended at Day 12.  Hunter,
Kaupp, and Taylor (1982) reported that at age 20-days, anchovy larvae begin
daily vertical migrations.  This larval movement into deeper water during the
day, and into shallow water at night, would coincidentally avoid significant
UV-B exposure.  After 20-days survival, then, the gauntlet of UV exposure
would have been successfully negotiated.  A more meaningful experimental
                          * * *  DRAFT FINAL

-------
                                   12-26
criterion would have been survival to at least age 20-days.  However, from
the reported mortality rates, perhaps even the control larvae would not have
survived this long.

    Related to the test endpoint is the third major problem in the
interpretation of these experiments.  When experiments are conducted to an
appropriate developmental endpoint (e.g., to the age when larvae avoid the
daytime near-surface layer), it is usually apparent that below a certain
irradiation level the organisms are not seriously affected.  On the other
hand, beyond that irradiation level there seems to be fairly general and
irreversible damage.  In experiments with planktonic invertebrates (Damkaer
et al. 1980, 1981, 1983), it can be seen that a very great range of UV-B
dose-rates and doses could lead to the same LD;.-; this depends solely on

the time chosen for the observation.  In every case, however, at the
developmental endpoint, the groups that had exhibited an LD,... were

essentially all dead.

    Hunter, Kaupp, and Taylor (1982) stated that none of the larvae surviving
the reported LD^  daily dose to age 4-days would have survived to age

12-days.  The LD^  criterion, then, was clearly inappropriate.  Hunter,

Kaupp, and Taylor admit that their LD,_0 daily dose to age 12-days may also

have underestimated the effect of UV on survival.  In their 1981 study they
predicted the effects on natural populations by assuming that all -organisms
receiving less that this LD^  value will survive, but this is not an
appropriate use of the LD^_.  To the extent that organisms receiving just

under LD,.- doses will probably all die, and that LD   values here suggest

only that a lower dose might also lead to death, the lethal daily dose
estimated by Hunter, Kaupp and Taylor could be an underestimate of UV
effects.  However, in a reinterpretation of the data considering threshold
dose-rate and dose effects (Damkaer, personal communication), the critical
daily dose can be estimated.

    Hunter, Kaupp, and Taylor (1981) noted that the lengths of anchovy larvae
in all UV treatments, even with the lowest dose, were significantly less than
the length of control larvae.  Length, then, could be a sensitive criterion
for UV-B effects, and this, and other criteria, should be explored.  A
judgement would be necessary as to what diminished length the individual
larvae could tolerate before there is a significant effect on survival.

    Hunter, Kaupp, and Taylor (1982) realized that information is required on
seasonal abundance and vertical distributions of anchovy larvae, vertical
mixing, and penetration of UV-B radiation into anchovy-populated seawater
before effects of current and increased levels of UV-B radiation can be
predicted.  There are fairly good data on seasonal abundances and vertical
distributions of anchovy larvae, as well as good calculations on average UV-B
penetration into known water types.  To detect a mortality due to UV-B
radiation, against the high natural mortality, would be a difficult problem.
But this is not the same as predicting the effect due to this factor alone.
For this, mortality rates from other causes are not required because we are

                          * * *  DRAFT FINAL  * * *

-------
                                   12-27
concerned only with how UV-B radiation would act on those that survive the
other perils of their existence.

    Hunter, Kaupp, and Taylor assumed that all larvae at depths where UV-B
exceeds the LDj.. daily dose will die.  They also assumed a static, unmixed

vertical distribution.  Finally, these authors provided data that indicated
that anchovy larvae off southern California are typically centered in
moderately productive ocean waters that have about 0.5 mg Chi a m-3.
Baker, Smith, and Green (1981) calculated surface DNA-effective UV-B
irradiance levels expected for this area, and Smith and Baker (1979)
calculated the penetration of UV-B radiation into moderately productive ocean
water.

    While Hunter, Kaupp, and Taylor calculated that 13.4% of the annual
larval anchovy population would die (presumably after 12-day exposures during
certain critical times) in a static situation, an analysis by Damkaer
(personal communication) indicates that somewhat more would die (14.7%) and
in shorter times (4 to 10 days), if there were no vertical mixing.  However,
static situations are not probable because of physical vertical mixing bf
water (through the action of wind, density currents, and tides) and the
active and passive movement of fish larvae.  If some vertical mixing did not
offer relief from the effects of UV-B radiation, it is likely that there
would be no anchovy larvae above a few meters depth; this is clearly not the
case.  Probably the surface mixed layer extends to at least 25 m from
December to February  (Sverdrup, Johnson, and Fleming 1942).  A mixed layer of
at least 10 m can be assumed for the rest of the year.

    For the mixed situation in February, the larvae would spend less than one
hour each day in the surface 1-m layer, resulting in exposures where the
threshold doses would never be reached in the larval stage.  In the February
model, with mixing to 25 m, there would have to be a six-fold increase in
incident UV-B radiation before threshold doses would be reached.  For- the
months March-October, with mixing to 10 m, the threshold doses are also not
reached within the 20 days available before the larvae become vertically
migrating.  Only in April and August, with the highest surface irradiances,
are these doses approached within this available time (~ 22 days).

    For March-October, with 10-m mixing, a 10% increase in incident UV-B
radiation would still not lead to threshold doses in less than 20 days.  With
a 20% increase in incident UV-B radiation (a result of about a 9% decrease in
the atmospheric ozone column) the depth of the threshold dose-rate is
increased.  With the 20% increase in incident irradiance, in the
dose/dose-rate threshold model, all of the anchovy larvae within the 10-m
mixed layer in April and August would be killed, the threshold doses being
achieved after 15 days.

    Apparently at all months about 36% of the larval anchovy population is
above 10 m (Hunter, Kaupp, and Taylor 1981).  The greatest numbers of anchovy
larvae are found in April (20% of annual population), so that 7.2% of the
annual population would be eliminated with a 9% ozone decrease.  Only 1.9% of
                          * * *  DRAFT FINAL  * * *

-------
                                   12-28
the annual population is present in August, so the loss of those above 10 m
would amount to 0.7% of the annual population.  The total predicted loss,
then, due to a 9% decrease in total ozone column, would be about 8% of the
larval anchovy population (Exhibit 9).

    In the static model of Hunter, Kaupp, and Taylor 13.4% of the larval
anchovy population dies through the current effects of solar UV-B radiation.
They predict only an additional 4.2% decrease if UV-B levels increase 60+%.
Presumably, in that model, a 20% increase in UV-B irradiance would lead to
only a 1.4% additional loss, so the total loss under a 20% increase would be
14.8%.  In contrast, the dose/dose-rate threshold and vertical mixing model
presented by Damkaer would predict no current solar UV-B damage, but an 8%
total larval population loss at a 20% increase in UV-B irradiance (10%
depletion) (Exhibit 9).  Because of complex interactions between
mixing-depths, vertical distribution of larval anchovy, seasonal changes in
solar irradiance and the penetration of UV into seawater, and seasonal change
in anchovy abundance, there is not a linear relationship between mixing-depth
and predicted annual loss of anchovy larvae.  However, within the range of
values represented in Exhibit 9, there is a maximum predicted loss at current
UV levels with a 5-m mixed layer, and a maximum predicted loss at a 60%
increase in UV-B irradiance with a 10-m mixed layer.

    Perhaps changes in lower levels of the food web, if they occur with
higher UV-B are likely to be more important than the direct effects upon the
fisheries changes in the composition of primary.production of organic biomass
could after the mortality experienced by larval fish.  And perhaps there
would be a synergistic effect on mortality.  Some fish die from direct
exposure, some die from lack of food, and some die from the combination of a
reduction in food and the weakening from exposure (or are weakened and become
outcompeted by other fauna for limited food reserves).

    The impact on marine fisheries as a food supply to humans 'would be
significant if the species of phyto- and zooplankton that adapted to enhanced
UV-B radiation were of different nutritional value (i.e., if they altered
growth and fecundity of the consumers or different accessibility to humans).
If the indirect impact of suppression upon consumers were linear, a 5%
reduction of primary production would result in a 5% reduction in fish
production (Exhibit 12-9).  A question still under investigation is whether
the trophodynamic relationships might be nonlinear.  For example, there may
be an amplification factor involved that results in a relatively greater
impact at higher trophic levels than at the primary-producer level.

    Levels of UV-B irradiance range latitudinally, with the highest exposures
in the tropics.  The current difference between the extremes of exposures  is
about 3- to 6-fold, but biota are presently adapted to the levels that are
normally experienced at their current locations.  An effect of ozone
depletion of ecological significance is that a 10% decrease in the ozone
concentration would produce a 25% increase of biologically effective
ultraviolet radiation  (DNA, Plant; Exhibit 2), which would correspond to
migrating over 30° toward the equator.  Continued investigations concerning
                          * * *  DRAFT FINAL  * * *

-------
                                   12-29
the range of natural ecological uncertainties, which are much larger than the
uncertainties in the particular photobiological effects, will be required to
permit assessment of the possible consequences for the many complex
ecological interactions as well as for the productivity of fisheries.
                            * *  DRAFT FINAL

-------
                             12-30
                         EXHIBIT 12-9

        Effect of Increased  Levels of Solar UV-B Radiation on the
    Predicted  Loss of Larval  Northern Anchovy from Annual Populations,
       Considering the Dose/Dose-Rate Threshold and Three  Vertical
                         Mixing Models
 30
         Larval  Northern  Anchovy
                                 10-m Mixed layer
                                           15-m Mixed layer
             10      20      30      40      50      60   •   70
             INCREASED   UV-B  RADIATION  (%)
Based on data of Hunter, Kaupp, and Taylor 1982.
                           DRAFT FINAL  •'• * ••

-------
                                   12-31
CONCLUSIONS

    At a minimum, evidence exists that both small and large changes in the
stratospheric ozone layer would affect aquatic ecosystems.  It will be very
difficult to determine the biological effects of small changes in UV-B
irradiance at the ocean surface because aquatic ecosystems have a huge
physical and biological "noise" level.  Storms, clouds, and global currents
make dramatic changes in the status of life in the ocean that are currently
unpredictable.  One should not, however, conclude that changes that are lost
in the "noise" of the system are automatically insignificant simply because
we cannot measure and define exactly what is taking place.  Changes that
occur because of small systemic events, such as higher UV-B, could accumulate
in time to produce much more significant change in aquatic organisms that
appear very sensitive to current UV-B radiation.  The possibility thus exists
that changes outside the historical range of UV-B could have implications far
greater than we are currently able to predict with confidence.  We ought not
to mistake our ignorance about what will happen if UV-B increases with the
conclusion that all is well.  At this time we can place almost no limit on
what will happen.  An increase in UV-B might have small effects, or it might
be much more significant.
                          * * *  DRAFT FINAL  * * *

-------
                                   12-32
REFERENCES

Baker K.S. (1979).  Calculation of biologically effective UV-B radiation in
aquatic systems.  SIO Ref 79-1, Scripps Institution of Oceanography.   46 p.

Baker K.S., Smith, R.C., Green, A.E.S. (1981).   Middle ultraviolet radiation
reaching the ocean surface.   Photochem Photobiol 32:367-374.

Banse K. (1982).  Cell volumes, maximal growth rates of unicellular algae and
ciliates, and the role of ciliates in the marine pelagial.  Limnol Oceanogr
27:1059-1071.

Bell G.M., Hoar, W.S. (1950).  Some effects of ultraviolet radiation on
sockeye salmon eggs and alevins.  Can J Res 28:35-43.

Caldwell, M.M. (1968).  Solar ultraviolet radiation as an ecological factor
for alpine plants.  Ecol Monogr 38:243-268.

Caldwell, M.M. (1971).  Solar UV irradiation and the growth and development
of higher plants.  Photophysiology 6:131-177.

Caldwell, M.M.,  Nachtwey, D.S. (1975).  Introduction and overview.  In:'
Nachtwey, D.S.,  Caldwell, M.M., Biggs, R.H. (eds.)  Impacts of climatic
change on the biosphere, part I - ultraviolet radiation effects, pp
(.1)3-(1)30, US Dept Transport, CIAP Monogr 5, NTIS, Springfield, Virginia.

Caldwell, M.M.,  Robberecht,  R., Billings, W.D.  (1980).  A steep latitudinal
gradient of solar ultraviolet-B radiation in the arctic-alpine life zone.
Ecology 61:600-611.

Calkins, J. (1982).  Modeling light loss versus UV-B increase for organisms
which control their vertical position in the water column.  In: Calkins, J.
(ed.) The role of solar ultraviolet radiation in marine ecosystems, pp
539-542, Plenum, New York.

Calkins, J. (1982).  Some considerations on the ecological and evolutionary
effects of solar UV.  In: Calkins, J. (ed.)  The role of solar ultraviolet
radiation in marine ecosystems, pp 685-689, Plenum, New York.

Cheng, L., Douek, M., Goring, D. (1978).  UV absorption by gerrid cuticles.
Limnol Oceanogr 23:554-556.

Cote, B., Platt, T.  (1983).   Day-to-day variation in the spring-summer
photosynthetic parameters of coastal marine phytoplankton.  Limnol Oceanogr
28:320-344.

Damkaer, D.M. (1982).  Possible influence of solar UV radiation in the
evolution of marine zooplankton.  In: Calkins, J. (ed.)  The role of solar
ultraviolet radiation in marine ecosystems, pp 701-706, Plenum, New York.
                            * *  DRAFT FINAL

-------
                                   12-33
Damkaer, D.M., Dey, D.B., and Heron, G.A., 1981, Dose/Dose-Rate Responses of
Shrimp Larvae to UV-B Radiation, Ecologia (BRL) 48:178-182.

Damkaer, D.M., Dey, D.B. (1982).  Short-term responses of some planktonic
Crustacea exposed to enhanced UV-B radiation.  In: Calkins, J. (ed.)  The
role of solar ultraviolet radiation in marine ecosystems, pp 417-427, Plenum,
New York.

Damkaer, D.M., Dey, D.B. (1983).  UV damage and photoreactivation potentials
of larval shrimp, Pandalus platyceros,  and adult euphausiids, Thysanoessa
raschii.  Oecologia 48:178-182.

Damkaer, D.M., Dey, D.B., Heron, G.A.,  Prentice, E.F. (1980).  Effects of
UV-B radiation on near-surface zooplankton of Puget Sound.  Oecologia
44:149-158.

Doehler, G. 1984.  Effect of UV-B radiation on biomass production,
pigmentation and protein content of marine diatoms.  Z Naturf 39:634-638.

Doehler, G. 1985.  Effect of UV-B radiation (290-320 nm) on the nitrogen
metabolism of several marine diatoms.  J Plant Physiol 118:391-400.

Dunbar, C.E. (1959).  Sunburn in fingerling rainbow-trout.  Progr Fish Cult
21:74.

Forsythe, W.E. (1954).  Smithsonian Physical Tables.  Smithsonian Institution
Press, Washington, D.C.  pp 721-723.

Gerstl, S.A.W., Zardecki, A., Wiser, H.L. (1986).   A new UV-B Handbook, Vol.
1.  In: Worrest, R.C., Caldwell, M.M. (eds.)  Stratospheric Ozone Reduction,
Solar Ultraviolet Radiation and Plant Life, pp 87-111, Springer-Verlag, New
York, Heidelberg, Berlin.

Giese, A.C. (1938).  The effects of ultra-violet radiations of various
wave-lengths upon cleavage of sea urchin eggs.  Biol Bull 75:238-247.

Giese, A.C. (1976).  Living with Our Sun's Ultraviolet Rays.  Plenum Press,
New York. 185 pp.

Green, A.E.S., Cross, K.R., Smith, L.A.  (1980).  Improved analytic
characterization of ultraviolet skylight.  Photochem Photobiol 31:59-65.

Haeder, D.P. (1986).  The effect of enhanced UV-B radiation on motile
microorganisms.  In: Worrest, R.C., Caldwell, M.M.  (eds.)  Stratospheric
Ozone Reduction, Solar Ultraviolet Radiation and Plant Life, pp 223-233,
Springer-Verlag, New York, Heidelberg,  Berlin.

Hairston, N. (1980)  The vertical distribution of diaptomid copepods in
relation to body pigmentation.  In: Kerfoot, W.  (ed.)  Evolution and Ecology
of Zooplankton Communities, pp 98-110,  Univ Press New England, Hanover, New
Hampshire.
                              *  DRAFT FINAL  * * *

-------
                                   12-34
Harvey, J.M. (1930).  The action of light on Calanus finmarchicus (Gunner)
as determined by its effect on the heart rate.  Contrib Can Biol 5:85-92.

Hobson, L., Hartley, F. (1983).  Ultraviolet irradiance and primary
production in a Vancouver Island fjord, British Columbia, Canada.  J Plankton
Res 5:325-331.

Hunter, J.R., Kaupp, S.E., Taylor, J.H. (1981).  Effects of solar and
artificial ultraviolet-B radiation on larval northern anchovy, Engraulis
mordax. Photochem Photobiol 34:477-486.

Hunter, J.R., Kaupp, S.E., Taylor, J.H. (1982).  Assessment of effects of UV
radiation on marine fish larvae.  In: Calkins, J. (ed.)  The role of solar
ultraviolet radiation in marine ecosystems, pp 459-497, Plenum, New York.

Huntsman, A.G.  (1925).  Limiting factors for marine animals I.  The lethal
effect of sunlight.  Contrib Can Biol 2:83-88.

Jitts, H.R., Morel, A., and Saijo, Y., 1976, The relation of oceanic primary
production to available photosynthetic irridantsm AUST.J.MAR. FRES.HWATER
RES. 27:441-454.

Jokiel, P.L., York, R.H., Jr., 1984, Importance of ultraviolet radiation in
photoinhibition of microbial growth, LIMNOL. , OCEANOG R. 29.:192-199.

Jones, L.W., Kok, B. (1966).  Photoinhibition of chloroplast reactions.  I.
Kinetics and action spectra.  Plant Physiol 41:1037-1043.

Karanas', J.J.,  Van Dyke, H., Worrest, R.C.  (1979).  Midultraviolet (UV-B)
sensitivity of Acartia clausii Giesbrecht  (Copepoda).   Limnol Oceanogr
24:1104-1116.

Karanas, J.J.,  Worrest, R.C., Van Dyke, H.  (1981).  Impact of UV-B radiation
on the fecundity of the copepod Acartia clausii.  Mar Biol 65:125-133.

Kaupp, S.E., Hunter, J.R. (1981).  Photorepair in larval anchovy, Engraulis
mordax.  Photochem Photobiol 33:253-256.

Kelly, J.R. (1986).  How might enhanced levels of solar UV-B radiation affect
marine ecosystems?  In: Proceedings of EPA/UNEP International Conference on
Health and Environmental Effects of Ozone Modification and Climate Change,
July, 1986.

Klugh, A.B. (1929).  The effect of the ultra-violet component of sunlight on
certain marine organisms.  Can J Res 1:100-109.

Klugh, A.B. (1930).  The effect of ultra-violet component of the sun's
radiation upon some aquatic organisms.  Can J Res 2:312-317.
                          * * *  DRAFT FINAL  * *

-------
                                   12-35
Lasker, R.  (1975).  Field criteria for survival of anchovy larvae:  the
relation between inshore chlorophyll maximum layers and successful first
feeding.  Fishery Bull 73:453-462.

Marinaro, J.Y., Bernard, M.  (1966).  Contribution a 1'etude des oeufs et
larves pelagiques de poissons mediterraneens I. Note preliminarie sur
l'influence  lethale du rayonnement solaire sur les oeufs.  Pelagos 6:49-55.

Murphy, T.  (1975).  Effects  of UV radiation no nucleic acids.  In:  Nachtwey,
D.S., Caldwell, M.M., Biggs, R.H. (eds.)  Impacts of climatic change on the
biosphere, part I - ultraviolet radiation effects, pp (3)21-(3)44, US Dept
Transport, CIAP Monogr 5, NTIS, Springfield, Virginia.

Nachtwey, D.S., Rundel, R.D. (1982).  Ozone change:  biological effects.  In:
Bower, F.A., Ward, R.B. (eds.)  Stratospheric Ozone and Man, Vol 2, pp.
81-121, CRC  Press, West Palm Beach.

NAS  (1984).  Causes and Effects of Changes in Stratospheric Ozone:  Update
1983.  Natl  Acad Sci, Washington, B.C.  254 pp.

Nielsen, E.S., 1964, On a complication in Maine productivity work due .to the
influence of ultraviolet light. J.CONS.CONS.INT.EXPLOR.MER., Volume
20:130-135.

.Pommeranz, T.  (1974).  Resistance of plaice eggs to mechanical stress and
light.  In:  Blaxter, J.H.S. (ed.)  The Early Life History of Fish, pp
397-416, Springer-Verlag, New York, Heidelberg, Berlin.

Ringelberg,  J., Keyser, A.,  Flik, B.  (1984).  The mortality effect of
ultraviolet  radiation in a translucent and in a red morph of
Acanthodiaptomus denticornis (Crustacea, Copepoda) and its possible
ecological relevance. Hydrobiologia 112:217-222.

Russell-Hunter, W.D. (1970).  Aquatic productivity.  MacMillan.

Ryther, J.H. (1969).  Photosynthesis and fish production in the sea.  Science
166:72-76.

Setlow, R.B. (1974).  The wavelengths in sunlight effective  in producing skin
cancer:  a theoretical analysis.  Proc Natl Acad Sci  (USA) 71:3363-3366.

Siebeck, 0.  (1978).  Ultraviolet tolerance of planktonic crustaceans.  Verh
Intern Verein  Limnol 20:2469-2473.

Smith, R.C., Baker, K.S. (1978).  Optical classification of natural waters.
Limnol Oceanogr 23:260-267.

Smith, R.C., Baker, K.S. (1979).  Penetration of UV-B and biologically
effective dose-rates in natural waters.  Photochem Photobiol 29:311-323.
                            * *  DRAFT FINAL  *

-------
                                   12-36
Smith, R.C., Baker, K.S. (1980).  Stratospheric ozone, middle ultraviolet
radiation and carbon-14 measurements of.marine productivity.  Science
208:592-593.

Smith, R.C., Baker, K.S., Holm-Hanson, 0., Olson, R. (1980).  Photoinhibition
of photosynthesis in natural waters.  Photochem Photobiol 31:585-592.

Sverdrup, H.U., Johnson, M.W., Fleming, R.H. (1942).  The Oceans, Their
Physics, Chemistry and General Biology.  Prentice-Hall, New York.  1087 p.

Teramura, A.H. (1986).   Interaction between UV-B radiation and other stresses
in plants.  In:  Worrest, R.C., Caldwell, M.M. (eds.)  Stratospheric Ozone
Reduction, Solar Ultraviolet Radiation and Plant Life, pp 327-343,
Springer-Verlag, New York,  Heidelberg, Berlin.

Thomson, B.E., Worrest, R.C., Van Dyke, H.  (1980).  The growth response of an
estuarine diatom (Melosira nummuloides  [Dillw.] Ag.) to UV-B (290-320 nm)
radiation.  Estuaries 3:69-72.
                                                        0
Worrest, R.C. (1982).  Review of literature concerning the impact of UV-B
radiation upon marine organisms.  In:  Calkins, J.  (ed.)  The role of solar
ultraviolet radiation in marine ecosystems, pp 429-457, Plenum, New York.

Worrest, R.C. (1983).  Depletion of stratospheric ozone:  impact of UV-B
radiation upon nonhuman organisms.   In:  Covello, V. et al. (eds.)  The
Analysis of Actual Versus Perceived Risks, pp 303-315, Plenum Press, New York.

Worrest, R.C., Wolniakowski, K.U.,  Scott, J.D., Brooker, D.L., Thomson, B.E.,
Van Dyke, H. (1981).  Sensitivity of marine phytoplankton to UV-B radiation:
impact upon a model ecosystem.  Photochem Photobiol 33:223-227.

ZoBell, C.E., McEwen, G.F.   (1935).   The lethal action of sunlight upon
bacteria in seawater.  Biol Bull 68:93-106.
                          * * *  DRAFT FINAL  * *

-------
Chapter 13

-------
                               CHAPTER 13

                     EFFECTS  OF UV-B ON POLYMERS
SUMMARY

    Several polymers (and largely the impurities  present  in  the polymers)
readily absorb the ultraviolet (UV)  radiation.  One  effect of  increased UV-B
radiation,  resulting from possible future depletion  of the ozone  layer in the
stratosphere, will be accelerated degradation of  the polymeric materials.
Energies associated with the UV radiation are large  enough to  initiate
reactions in polymers,  which lead to their degradation by affecting their
mechanical  and optical  properties and thus reducing  their service life.
Several methods are used to stabilize the polymers to maintain a  useful
service life in their various important applications.

    While much is known about the physical processes of degradation,  little
research has focused to date on the  potential costs  and response  to such
degradation.  Initial studies suggest that one likely response to increased
UV-B induced damage is  to increase the amount of  light stabilizer in  the
polymers.  Due to lack of relevant data,  only approximate estimation  methods
are available to determine the extent of light-induced damage  and the degree
of stabilization required to minimize it.  The effect of  increased UV-B
radiation is manifested in increased costs of production, including raw
material costs, energy costs, and maintenance costs.
                          * * *  DRAFT FINAL  * *

-------
                                   13-2
FINDINGS

1.   DEPLETION OF THE OZONE LAYER (STRATOSHPERE) WILL INCREASE THE AMOUNT OF
    ULTRAVIOLET-B (290-315 nm) RADIATION REACHING THE EARTH'S SURFACE.

2.   INCREASED UV-B RADIATION WILL ACCELERATE THE DEGRADATION OF POLYMERS.

    2a.   Several commercial polymers (e.g., polyethylene, polypropylene,
         poly(vinylchloride)), although theoretically UV transparent, contain
         sufficient chromophore impurities that absorb significant amounts of
         light in the UV-B region of the spectrum.  Other polymers (e.g.,
         polycarbonate) have structural features in the molecule which result
         in strong UV-B light absorption.
    2b.
         Several polymers have important outdoor applications (e.g., building
         industry -- siding, window profiles, glazing; packaging -- film,
         containers; housewares and toys; and paints and protective
         coatings).   Such polymers are likely to be exposed to significant
         amounts of UV-B radiation.

    2c.  Absorption of UV-B radiation in polymers causes photo-induced
         reactions and alters important mechanical and optical properties of
         the polymers (e.g., yellowing, brittleness) and thus degrade (reduce
         the useful life) the polymers.

3.  'UV-STABILIZERS ARE USED IN POLYMERS FOR PROTECTION AGINAST UV RADIATION.
    INCREASED UV-B RADIATION MAY REQUIRE INCREASED USAGE OF STABILIZERS TO
    MAINTAIN A PRODUCT'S USEFUL LIFE.

    3a.  Increased amounts of stabilizers might adversely affect the
         processing and use properties of some polymers (e.g., hardness,
         thermal conductivity, flow characteristics).  For example, increased
         amounts of titanium dioxide in poly(vinylchloride) might affect its
         processing properties.

    3b.  Changes in the amount of stabilizer (and other additives) would
         require development of new formulations.

4.   DAMAGES FOR GLOBAL STRATOSPHERIC AND CLIMATE CHANGE ARE DIFFICULT TO
    ESTIMATE.

    4a.  Due to lack of relevant experimental data, only approximate
         estimation methods are available to determine the extent of
         light-induced damage.

    4b.  Depending upon the chemical nature of polymer, the components of the
         compound and the weathering factors, both temperature and humidity
         tend to increase the rate of degradation.
                          * * *  DRAFT FINAL  * * *

-------
                                   13-3
5.  INCREASED UV-B DUE TO OZONE DEPLETION WILL HAVE ADVERSE ECONOMIC EFFECTS.

    5a.  Change in polymer processing properties can result in more equipment
         shutdown, higher maintenance costs, increased utility costs.

    5b.  Increased operating costs and material costs (stabilizers,
         lubricants, and other additives) would have adverse economic impact
         on the polymer/plastic industry and related industries.

    5c.  In a case study, based on preliminary methodology, for a given
         scenario of ozone depletion (26% depletion by 2075), undiscounted,
         cumulative (year 1984-2075) economic damage for poly(vinylchloride)
         is estimated at $4.7 billion (USA only).  Due to the lack of data,
         damage in other polymers is not assessed.  Like PVC, it is expected
         to be significant.
                          * * *  DRAFT FINAL  * *

-------
                                   13-4
PHOTODEGRADATION OF POLYMERS

Structure of Polymers

    Macromolecules in the polymers  are composed of several monomeric units
joined by chemical bonds to each other.   The monomeric units contain chemical
bonds that are either in the main chain of  the macroraolecule or connect
various atoms or side groups to it.   Side groups, if present, contain
additional chemical bonds.   All of  these bonds may be reaction sites in
polymer degradation, and various energy sources may be effective in supplying
the energy necessary to break the bonds.  The dissociation energies of
chemical bonds in common polymers range from about 65 kcal/mole (C-C1) to 108
kcal/mole (C-F) with carbon-carbon  bonds in the middle (75-85 kcal/mole).  The
most important types of energy that cause polymer degradation are heat,
mechanical energy, and radiation.  Thermal  and mechanical degradation of
polymers may occur during thermomechanical  processing.  The most common form
of radiant energy which causes degradation  is that of the UV component of
sunlight; the energy of a photon with wavelength X = 300 nm is about 95
kcal/mole, which is higher than most bond dissociation energies in polymers,
and therefore capable of breaking these bonds.

    Wavelengths of UV radiation at  which various commercial polymers have
maximum sensitivity and their corresponding photon energies are given in
Exhibit 13-1 below.
                               EXHIBIT 13-1

           Wavelengths of UV Radiation  and Polymers With Maximum
               Sensitivity and Corresponding Photon Energies

Polymer
Styrene-acrylonitrile copolymer
Polycarbonate
Polyethylene
Polystyrene
Polyvinylchloride
Polyester
Vinyl chloride-vinyl acetate copolymer
Polypropylene
Wavelength
(nm)
290,325
295,345
300
318
320
325
327,364
370
Energy
kcal/mole
99,88
97,83
96
90
89
88
87,79
77
        Source:  Kelen,  T.,  Polymer Degradation, Van Nostrand Reinhold
                Company, Inc.,  New York,  1983.
                            * *  DRAFT FINAL  *  *  *

-------
                                   13-5
Light Absorption

    A prerequisite for light-induced damage in materials is their ability to
absorb the radiation (this observation is sometimes referred to as the
Grotthus-Draper Law of Photochemistry) (Andrady 1986).  Very few materials
absorb light at all wavelengths across the spectrum.  Pure samples of
polyethylene, polypropylene, and poly(vinylchloride), for instance, are
theoretically UV transparent and are not expected to absorb light in the range
290-315 nm, the UV-B region of the spectrum.  However, completely "pure"
polymers do not exist.  During synthesis, processing, and storage, various
amounts of carbonyl and hydroperoxy groups accumulate in the commercial
polymers.  These chromophores readily absorb UV light in the 290-315 nm
wavelengths.  Initiation of degradation consists mainly of the decomposition
of these chromophores.

Photophysical and  Photochemical  Processes

    Degradation of the polymers involves a series of photophysical and
photochemical processes.  The physical processes involved in photodegradation *
include absorption of light by the material, electronic excitation of the
molecules, and deactivation energy transfer to some acceptor.  When the
lifetime of the excited state is sufficiently long, the species can
participate in various chemical transformations.  This is especially true when
the photoinduced degradation reactions are carried out in the presence of
oxygen (see equations 1-4).
                     light
        Polymer (RH) 	•* polymer excited state (RH*)                   (1)

            degradation
        RH* 	-*• polymer radicals or broken (scission)
                          chain end (R»)                                  (2)

        R« + air 	•* polymer peroxy radicals (R02»)                    (3)

                  light
        R02« + RH	> ROOH + R» (further polymer degradation)          (4)


    As mentioned earlier, hydroperoxy and carbonyl groups are the two most
common impurities present in the polymers.  In the case of hydroperoxy groups
(ROOH), the energy of UV light is sufficient to cause both of the following
decompositions (Kelen 1983):

                     light
                ROOH 	-»• R0» + »OH  (~ 42 kcal/mole)                   (5)

                     light
                ROOH 	-» R»  + »OOH (~ 70 kcal/mole)                   (6)
                          * * *  DRAFT FINAL  * * *

-------
                                   13-6
Dissociation of the 0-H bond is less facile:
                     light
                ROOH 	-» R00» + «H  (~ 90 kcal/mole)                  (7)
    Because of the low bond dissociation energy (42 kcal/mole), decomposition
according to reaction (5) is predominant in photooxidation of polymers
containing hydroperoxy groups.

    The carbonyl group absorbs UV light readily and hence is easily excited.
The excited carbonyl group decomposes via Norrish reactions of Types I, II,
and III (Kelen 1983).  The Norrish-I reaction (8) is a radical cleavage of the
bond between the carbonyl group and the o-carbon atom (a - scission).

                             0                  0
                              ||      light       ||
                        -CH2-C-CH2 -- -»• -CH2-C» + »CH2                (8)

                                               *
                                            CH2«   + CO


    The Norrish-II reaction (9) is a nonradical scission that occurs through
the formation of a six-membered cyclic intermediate to form an olefin and an
alcohol or ketone.
                      0- -H                  OH
                    /     \    light     /

                   -           " -- '"

                      CH2-CH2                  2        2


    The Norrish-III reaction (10) is also a nonradical chain scission and
leads to the formation of an olefin and an aldehyde.
                        0 CH3         0    CH2

                        II I    light   ||    ||
                       -C-CH	-* -CH + CH-                            (10)
    In the case of PVC, the photolytic scission takes place as follows:
                         light       •
                 CH2-CH	•* -CH2-CH- + »C1                          (11)


                    Cl
                          * * *  DRAFT FINAL  * * *

-------
                                   13-7
POLYMERS  IN  OUTDOOR  USES AND  POTENTIAL DEGRADATION

    A variety of polymeric materials  are used in applications where the
material is routinely exposed to solar radiation.   These include mainly the
use of plastics in the building industry and in paints/coatings.   In other
applications such as packaging, outdoor furnishing, housewares,  and toys,
exposure occurs from extended to more limited,  possibly intermittent,  periods
of time.

    According to recent estimates, of the 45,506 million pounds  of U.S.
plastic sales in 1985, the building industry above accounted for 9,428 million
pounds, 21% of the total consumption.  Packaging applications were the largest
market consuming 10,846 million pounds (Horst,  et al.  1986).

    However, bulk of the consumption  represents that of a limited number of
different plastics.  These include the so-called "commodity" plastics  and a
few other types selected on the basis of their  unique physical-chemical
properties.   Exhibit 13-2 gives a breakdown of  the types of plastics used in
each application along with the likelihood of exposure to sunlight in normal •
use.  Some plastics/products, not experiencing  direct sunlight,  can still be
exposed to UV light since the UV light is easily reflected and scattered by
terrain and buildings, resulting in significant irradiation levels at  surfaces
illuminated only by diffused or reflected light (Andrady 1986).   Thus,
vulnerability to degradation is a function of a polymer's normal lifetime and
the percent of time it is exposed to  direct or indirect sunlight.

Light-Induced  Damage

    Exposure to light and several physical-chemical processes (described
earlier) occurring simultaneously result in deterioration of polymeric
materials.

    Properties most likely to be affected by the degradation of  the polymer
include both mechanical and optical properties.  Mechanical properties such  as
tensile strength, elongation, modulus, and impact strength can all be affected
by UV degradation of polymeric materials.  Optical properties that have been
observed to change upon exposure to UV radiation include transparency, color,
chalking or cracking of surfaces, and yellowing.  These effects  are most
noticeable on the surface of plastic  materials  because of the limited
penetration of UV light through the materials.   Thus,  the material may show
cosmetic changes in appearance long before any  degradation in its bulk
mechanical property is noted.

    The mechanical weakening of plastic material occurs because  of a reduction
in the molecular weight of the polymer.  Essentially all of the  mechanical
properties of polymers are related to the molecular weight of the material.
Thus, for every bond in the polymer molecule that breaks due to  UV radiation,
the average molecular weight will be  reduced, and the corresponding mechanical
properties will be reduced in some proportion to this molecular  weight
reduction (Andrady 1986).
                          * * *  DRAFT FINAL  * * *

-------
                                   13-8
                               EXHIBIT 13-2

       Plastics Used in Applications Where  Exposure of the Material
                       to Sunlight  Might be Expected
  Type of Plastic
 Application
  Usage
1000 metric
   tons
Exposure
Poly(vinylchloride)
            (rigid)
Poly(vinylchloride)
       (plasticized)
Unsaturated
Polyester
Polycarbonate


Acrylic



Polyethylene


Polypropylene
Thermoplastic
Polyester
 Building Industry

siding                   410
door/window              284
other                     33
conduit                  493
Irrig. pipe              247
other pipe              1912

roofing                   22
liners                    25
wire-cable  .             410
weatherstrip              38
garden hose               48

glazing                   41
panels/siding            122
pipe                     237

glazing                   90
fixtures                   9

glazing      .             84

     Packaging

film                     696
containers              2990

film                     375
containers               325

containers               635
                  high
                  high
                  high
                  moderate
                  high
                  low

                  high
                  moderate
                  N/A
                  moderate
                  high

                  high
                  high
                  low

                  high
                  low

                  high
                  N/A
                  N/A

                  N/A
                  N/A

                  N/A
                          * * *  DRAFT FINAL  * * *

-------
                                   13-9
                               EXHIBIT 13-2
                                 (Continued)
Type of Plastic
Polyethylene
Polystyrene
Polypropylene
Note: Exhibit 2 is
Usage
Application 1000 metric
tons
Housewares and Toys
954
378
176
based on 1985 estimates as reported in Modern
Exposure
N/A
N/A
N/A
Plastics,
       (1986) 63(1): 69.

N/A indicates exposure information not available.   However,  based  on  the
application, it is expected to be low to moderate.
                                 DRAFT FINAL  * * *

-------
                                   13-10
    Discoloration of polymeric materials is also related to bond breaking, but
in these cases, bonds-that are broken may have little effect on mechanical
properties.  The best example of a polymer discoloring by yellowing is in
transparent flexible polyvinyl chloride (PVC) window materials used for rear
windows of convertible automobiles.  The PVC darkens from weathering because
of impurities in the resin initiating a dehydrochlorination reaction as
outlined below:
                                           H
                                light      |
                     -CH2-CH	•»• -C=CH- + HC1                    (12)

                          |    impurities
                          Cl
The color is caused by the multiple double bonds (~9 repeat units in
sequence) created between the carbon atoms, resulting in a polymer that
absorbs the light in the visible region of the spectrum and thus appears
yellow.  Often this discoloration is evident long before mechanical properties
of the polymer material begin to change.  This, again, is because mechanical
properties are related to the molecular weight of the polymer more than to the
chemical structure.

    Other changes in polymeric materials from UV exposure, similar to
yellowing, also occur mostly at the surface.  Such changes as surface
cracking, chalking or crazing (i.e., when a clear surface becomes translucent
because of formation of microcracks) are all evident in polyolefins,
polycarbonates, vinyl, acrylic, and styrenic polymers when exposed to UV.
These surface effects drastically reduce the life expectancy of materials that
require maximum transmission of visible light, such as Plexiglas and
greenhouse windows, solar energy collectors, and streetlamp globes.

    Several different modes of damage occur concurrently.  However, depending
on the application of the polymer, one or more of such properties may be
important.  Exhibit 13-3 gives the critical mode of damage and one or more
secondary modes of damage identified for selected applications of polymers.
The critical mode of damage is defined as the degree of damage which renders
the material "unacceptable" at the shortest duration of exposre, thus reducing
the polymer's service life.

Plastic Compounds

    Polymer resins are rarely, if ever, used by themselves in any
application.  To optimize the properties demanded by a given application and
to enable the product to be easily fabricated, a number of other "ingredients"
are added to the polymer to obtain a "compound."  The nature of the
non-polymer ingredients in a compound and the ratio in which they are mixed
will depend on the processing technique, the grade of polymer, and the desired
properties in the end product.
                          * * *  DRAFT FINAL  * * *

-------
                                   13-11
                               EXHIBIT 13-3

              Modes  of Damage Experienced by Polymers  Used in
                            Outdoor Applications
        Polymer
    Application
      Damage
  Poly(vinylchloride)
siding, window frames
                              pipes



                              roofing materials

                              automobile upholstery

  Unsaturated polyethylene    outdoor surfaces
  Polyethylene/polypropylene  irrigation pipe,  outdoor
                                furniture
                              synthetic turf,  stadium
properties
                                seats, packaging
  Polycarbonates
glazing material
[+ yellowing]
+ chalking
- impact properties
- tensile properties
+ surface distortion

[- burst pressure]
- impact properties
+ yellowing

[+ brittleness]

+ discoloration

[+ surface erosion]
+ discoloration
- strength

[+ brittleness]
- tensile properties
- electrical
[ + yellowing]
+ loss of transparency
         +  = increase
         -  = decrease
        [ ] = parentheses indicate critical mode of damage

        Source:  Andrady, Anthony, "Analysis of Technical Issues  Related  to
                 the Effect of UV-B on Polymers," Research Triangle Institure,
                 Rsearch Triangle Park, N.C., March 1986.
                            * *  DRAFT FINAL  * * *

-------
                                   13-12
    An example of a rigid PVC siding compound is given in Exhibit 13-4.


                               EXHIBIT 13-4

                      PVC Siding Compound  Composition
                      Component         Concentration (pph)*
                   PVC resin                     100
                   Heat stabilizer               1.5
                   Lubricant                       1
                   Titanium dioxide               12
                   Impact modifier                 6
                   Calcium carbonate               3
                   Processing aid                  1
                   * pph = parts per hundred parts resin

                   Source:  Andrady, Anthony, "Analysis of
                            Technical Issues Related to the
                            Effect of UV-B on Polymers,"
                            Research Triangle Institure,
                            Rsearch Triangle Park, N.C.,
                            March 1986.
    The fabrication of the siding involves feeding the melted compound through
a slit die, a process where the polymer is heated and sheared.  Therefore a
heat stabilizer is usually used in the compound.   The lubricant prevents
sticking of melt to metal parts in the processing machinery.   A suitable
rubber is added as an impact modifier to reduce the brittleness of PVC resin.
The processing aid is added to improve the melt processing characteristics
while the chalk is added to cheapen the compound.  Titanium dioxide is the
light stabilizer.  As explained in the previous sections, several polymers
when exposed to sunlight (even under current UV conditions),  would degrade,
some of them rapidly.  Therefore, stabilizers such as titanium dioxide are
used to reduce the damage from exposure to light.  Any depletion in the
stratospheric ozone with potential increase in the UV-B radiation would
accelerate the degradation process.

Light Stabilizers in Polymer Products

    Light stability in polymeric materials is achieved by the inclusion of an
appropriate "light stabilizer" in a polymer compound.  Light stabilizers are
substances that absorb the light in the UV-B range of the spectrum very
                          * * *  DRAFT FINAL  * * *

-------
                                   13-13
efficiently.  When incorporated into a transparent polymer matrix, the
stabilizer molecules compete with the polymer molecules for the available
light.  Light absorbed by the UV-absorber molecules is harmlessly dissipated
and does not lead to any damage of the polymer matrix.  For best results the
light absorber should be soluble in polymer, absorb light in the UV-B range of
the spectrum, and have a very high molar extinction coefficient (so that it is
effective at relatively low concentrations).

    Three different methods are used to achieve light stability.  The first
method involves use of UV-absorbing compounds such as 2-hydroxy-benzophenones,
2-hydroxyphenylbenzotriazoles, 2-hydroxyphenyl-S-triazines and derivatives of
phenyl salicylates (Heller 1969).  These compounds absorb and dissipate the
light without degrading the polymer.

    A second approach is to prevent light from reaching the bulk of the
polymer (light shielding) by opacification of the matrix.  Inclusion of a
light absorbing (or light reflecting) insoluble pigment in the matrix prevents
the light from penetrating the bulk of the polymer.  However, a fraction of
surface layer is unprotected and is therefore "affected by light.  Titanium
dioxide absorbs UV-B radiation very effectively and is used in PVC
compositions as a light screener.  Other pigments used as light screeners
include zinc oxide, magnesium oxide, lead carbonate, and barium sulfate.  One
of the most effective light screeners is carbon black.  Used in appropriate
concentrations and adequately dispersed in the matrix, the small-sized carbon
black particles impart excellent light stability in polyethylene
formulations.  Exhibit 13-5 gives the UV screening effectiveness of several
selected pigments at concentrations of 2% in polypropylene films of 10-mil
thickness (Uzelmeir 1970).  The screening power is given as the ratio of the
weathering stability (lifetime) of the pigmented to unpigmented materials
(Andrady 1986).  Higher ratio indicates longer stability of the polymer.  For
example, a polymer that is compatible with carbon black would have a lifetime
12 times longer when pigmented with carbon black and exposed to sunlight than
without carbon black.  On the other hand, a polymer that is compatible with
cadmium yellow would have a lifetime twice as long when pigmented with cadmium
yellow (at the same concentration as carbon black in the previous polymer)
than without cadmium yellow.

    The UV absorbance at specific wavelengths is also given in Exhibit 13-5.
This is an indication of how well the pigment is able to absorb the light.
Higher absorbance for titanium dioxide (0.92 at 300 nra) than for iron oxide
(0.29 at 300 nm) means that at the same pigment concentration (2%), titanium
dioxide is able to absorb more light than iron oxide.

    The third approach to achieve light stability does not prevent or control
the interaction of light with the polymer.  Instead, the light stabilizer
additive either seeks to deactivate or quench the excited state chromophores
before the degradation process occurs, or acts as a free radical scavenger,
which mops up the damaging free radicals as soon as they are formed.  Complex
chelates of nickel, for example, have been used as light stabilizers in
polyolefins and are believed to act as quenchers and peroxide decomposers.
                          * * *  DRAFT FINAL  * * *

-------
                                   13-14



                                EXHIBIT 13-5

              UV Screening Effectiveness of Selected Pigments



                                                      Absorbance
           Pigment              Screening Power     300nm   325nm  350nm
Titanium Dioxide
Iron Oxide
Cadmium Yellow
Chromium Oxide
Carbon Black
(Channel Black)
2.25
3.25
2.0
2.75
12.0

0.92
0.29
0.36
0.05
1.56

0.97
0.30
0.36
0.06
1.56

1.01
0.32
0.36
0.07
1.56

           Note:  Chemical interactions  limit  the use  of  some of the
                  above pigments in several  polymers.   For  instance,
                  iron oxide can be effectively used with
                  polypropylene but not  with PVC, as it catalyses the
                  photooxidation of the  latter polymer (King  1968).


    In recent years the use of radical scavenger molecules, particularly the
very effective Hindered Amine Light Stabilizers (HALS) has  gained wide
industrial acceptance.  Exhibit 13-6 gives  the current U.S. consumption of
light stabilizers in various types of plastics.

DAMAGE FUNCTIONS AND RESPONSE TO DAMAGE

    Depletion of the stratospheric ozone would change  the spectral
distribution as well as the solar radiation pattern, with a potential increase
in the UV-B radiation striking the earth's  surface.  The  magnitude of such
changes on the weathering of plastics and potential damage  that might occur
are difficult to assess due to lack of experimental data.   However, various
theoretical approaches are considered to estimate the  damage.

    There are three possible responses to potential damage  from increased UV-B
radiation:

        1.  Add greater amount of light  stabilizers to prevent
            further degradation of polymers.

        2.  Develop new or improved performance stabilizers to
            absorb the additional radiation.

        3.  Switch from polymers to alternative materials (e.g.
            ceramics, metal, etc.)
                          * * *  DRAFT FINAL  * * * •

-------
                          13-15



                       EXHIBIT 13-6

    Domestic Consumption of Light Stabilizers,  1984-85
                               Domestic Consumption (1000 Ib)
                                    (of UV Stabilizers)	
  Plastic Type                         1984     1985
Polyolefins                            4290     4600
Polycarbonates                          352      400
PVC and related copolymers              171      180
Acrylic                                 165      170
Polyester                               165      171
Note:  Above figures do not include light shielders such as
       titanium dioxide
The relevant statistics for pigments (colorants) are given
below.
                               Domestic Consumption (1000 Ib)
                                  (of Pigments/Colorants)
         Pigment                       1984     1985
Titanium Dioxide
Carbon Black
Iron Oxide
Cadmium Compounds
Chromium Compounds
271
80
9
5
5
272
85
10
5
5
Source:  Andrady, Anthony, "Analysis of Technical Issues
         Related to the Effect of UV-B on Polymers," Research
         Triangle Institute, Research Triangle Park, North
         Carolina, March 1986.
                     *  DRAFT FINAL  * *

-------
                                   13-16
Several attempts have been made to estimate the increased damage to materials
from potential increase in UV-B radiation.  Because of the complexity of the
analysis, each approach has focused on only the first response.  Each approach
analyzes how much additional light stabilizer would be required for a given
ozone depletion (function of UV light intensity) scenario, to maintain the
current level of useful life for various polymers.  Thus, each approach has
attempted to relate intensity of the UV light and concentration of the light
stabilizer in the polymer.

Transparent Polymers (CIAP Approach)

    Schultz et al. in a research study (Climatic Impact Assessment Program
(CIAP)) funded by the U.S. Department of Transportation were the first to
examine the potential adverse effects of the ozone depletion on materials
(Schultz et al. 1974).  Although their concern was depletion of ozone from SST
flights and not from discharge of chlorofluorocarbons, their approach is
nonetheless intuitive.

    The CIAP approach employed the Beer-Lambert law to calculate approximate
increases in UV stabilizer required to offset the effects of increased UV
light (Andrady 1986).  According to this law, for a homogeneous material with
no scattering losses at the surface or in the bulk of the polymer, the
monochromatic light absorption by the system is described as follows:

    The fraction of incident monochromatic light energy absorbed is
proportional to the concentration of the chromophores and to the path length
of the light within the substrate.
                        Log It/IQ = -ecH                                (13)

    I  = intensity of the transmitted light

    I  = intensity of the incident light

     e = molar extinction coefficient (proportionality constant)

     c = molar concentration of the light absorbing moiety

     i = thickness of the film
    An increase in the incident light intensity will result in a higher light
flux in the polymer; therefore, to maintain the same level of protection, the
concentration of the UV absorber has to be increased.  If an increase in the
light intensity by a factor x required a corresponding increase in the UV
absorber concentration by a factor y, the Beer-Lambert Law for the new system
gives
                          * * *  DRAFT FINAL  * *

-------
                                   13-17
                              Log(It/Io»x) = -eyc£                        (14)

    The two above equations can be combined to give

                                y = Log(x)/H + 1                          (15)

         where H = ec£, a constant
This simple equation predicts the variation of the factor increase in the UV
absorber concentration, y, with the factor increase in the intensity of light,
x.  An empirical observation made is that an increase in the UV light
intensity by a factor of 1.5 generally requires an increase in the UV absorber
concentration by a factor of 2.0.

    Based on this empirical observation and the damage function (equation 15),
CIAP developed the following relationship:


                        y = 1 4- (Iog10x)/(log101.5)                       (16)


    Using the above relationship and using a brash extrapolation of the world
market, CIAP developed increased stabilizer requirements for various levels of
ozone depletion.  A market period of 50 years (1970-2020) was considered in
the CIAP study to estimate the future market of plastics.  The results are
summarized in Exhibit 13-7.
                               EXHIBIT  13-7

                       Increased Stabilization  Market
                                (1970-2020)
Ozone Depletion
Percent
50
20
10
5
By 1990
million
1,040
370
172
84
By 2020
1974 dollars *
5,040
1,800
840
408
                * Values at zero discount rate.

                Source:  Schultz, A., D. Gordon, and W. Hawkins,
                         "Economic and Social Measures of
                         Biologic and Climatic Change,"  CIAP
                         Monograph 6, September 1974.
                              *  DRAFT FINAL  * *

-------
                                   13-18
Battelle Study

    This study was a follow-up to the CIAP study.   The basic assumptions
remained the same, but all economic market data used for analysis was updated
using more current information.  Also the different ozone depletion scenario
was used in the Battelle study.  Exhibit 13-8 below shows the ozone depletion
estimates used in the study.


                               EXHIBIT  13-8

                         Ozone Depletion Estimates
                          Year       Ozone Depletion
                                       (percent)
1985
1995
2005
2015
2025
2035
2045
2055
2065
2075
0.0
0.15
0.62
1.50
2.84
.4.66
7.08
10.41
15.44
26.08
                         Source:  EPA estimates based on
                                  emission scenario
                                  developed by EPA and
                                  modeled by Lawrence
                                  Livermore Lab.
    For the projection of the polymer demand to the year 2075, Battelle used a
simplified assumption that the growth rate ratio of UV stabilized plastics to
GDP (Gross Domestic Product) for the United States will remain constant at 1.7
into the 21st century.

    Historical U.S. market share of the world plastic market and projection of
the U.S. plastic market were used to project the world plastic market through
the year 2075.

    Battelle used Cutchis' formula to estimate the effect of ozone depletion
on the average UV light intensity (Cutchis 1974):
                          * * *  DRAFT FINAL  * * *

-------
                                   13-19
                                log x = 0.00954D                (17)
where x is UV light intensity factor and D is the percentage of ozone
depletion.

    To describe the relationship between light intensity (x) and the
stabilizer concentration (y), Battelle used a mathematical formula developed
by Shultz, Garden, and Hawkins (1974).
                                log x = 0.176 (y-1)             (18)


    The Combination of the Cutchis and Shultz formuli,  gave the following
relationship:


                                y-1 = 0.0542D                   (19)


    This formula was used in the economic impact model.

    Exhibit 13-9 gives cumulative added cost for the U.S.  and the world in the
presence of ozone depletion.


                                EXHIBIT 13-9

                           Cumulative Added Cost
Year
1985
2015
2045
2075
Ozone Depletion
(percent)
0.0
1.50
7.08
26.08
U.S.
(1983 $
0.0
0.13
2.3
19.4
World
billion)*
0.0
0.60
14
120
                 * Values at zero discount rate.

              Source: Hattery, G., V. McGinniss, and P.  Taussig
                      "Costs Associated with Increased Ultraviolet
                      Degradation of Polymers." BATTELLE Columbus
                      Laboratories, Columbus, Ohio, April 1985.
                              *  DRAFT FINAL  * * *

-------
                                   13-20
Extension to Filled Polymers (Andrady/Shultz)

    While the CIAP approach is sound and simple to use, its applicability is
limited to the transparent, homogeneous polymers.  With the exception of
window glazing, polymers used outdoors are generally opaque and are often
stabilized using light screeners.   Therefore a parallel model based on light
stabilization, applicable to opaque materials (filled polymer), was developed
by the Research Triangle Institute (Andrady 1986).

    A filled polymer can be described by a homogeneous resin matrix in which
essentially spherical filler particles are randomly dispersed.  The filler can
be a pigment with high light absorptivity, so that all the light reaching the
filler particle is absorbed.  For simplicity, the random distribution can be
described as spheres of radius r,  in two dimensions, within a series of layers
of thickness 2r (see Exhibit 13-10).  If the pigment volume fraction is V, the
number of spherical pigment particles within a lamina volume element of
y»y»2r cm3 is N = 2ry2V/(4ir/3)rJ.

    The fraction of light obliterated in passing through the laminar is the
cumulative cross-sectional area of the spheres per unit laminar area,
Nirr2/y2, which is equal to 3V/2.  The surface intensity, I , can then
be related to I , intensity at depth £, as


                           .  I£ = Io(l-3V/2//2r                          (20)

    For light intensity increased by a factor x and a corresponding increase
in the light stabilizer by a factor y, the above equation can be modified as


                             I£ = x«Io(l-3Vy/2)£/2r                       (21)


    Using the assumption of equal residual fractional intensity at a
characteristic depth of I units, it can be shown that
                              i
                       Inx = 	 • ln((l-3V/2)/(l-3Vy/2))                 (22)
                              2r


    This equation has the same predictive feature as the CIAP equation but
requires a knowledge of i and r for its application.

    An empirical observation relating x to y is needed in this case as well;
however, such experimental data are not presently available for the polymers
of interest.  No economic, damage estimates have been developed using this
approach.
                              *  DRAFT FINAL  * * *

-------
                                   13-21
Comprehensive Approach (Andrady 1986)

    While the two approaches described previously can be used for clear or
filled polymers to determine the amount of additional stabilizer requirements

                              EXHIBIT  13-10

        Diagrammatic  Representation  of  the Effect of Pigment/Fillers
             as Light Shield.   (Monodisperse Spherical Filler)

                                      Light
                                                      Polymer matrix
corresponding to an increase in the light intensity, neither approach takes
into account the activation spectra for degradation.

    The extent (and even the nature) of damage suffered by a polymer on
exposure to sunlight is markedly wavelength dependent.  More appropriate
methodology to obtain the damage function should include  (1) source spectral
distribution, (2) activation spectra, and (3) the functional relationship
between the ozone concentration in the stratosphere and the spectral
irradiance.

    Atmospheric attenuation of sunlight is a result of several processes.  The
following are the three main processes:  (1) Rayleigh scattering - scattering
of light due to gas molecules in the atmosphere; (2) aerosol scattering or Mie
scattering caused by fine particles and fluid droplets suspended in the
atmosphere; and (3) absorption by ozone.  These factors should be considered
in calculating the spectral distribution and intensity of light at a given
location.

    An activation spectra should be available for each specific polymer to
assess the damage.  A normalized activation spectra then can be used as a
continuous weighting function describing the spectral sensitivity of a
specific damage process in a given polymer.

    The total damage function (index), D, for a polymer exposed to sunlight
can be presented as


                          * * *  DRAFT FINAL  * * *

-------
                                   13-22 '



                                   \
                               D = / F(X)H(X)dX                           (23)

                                   X0
where   F(X) = adjusted normalized activation spectrum (ANAS)
        H(X) = spectral distribution of sunlight for natural weathering
        X , X  = wavelength range of the ANAS

    It is assumed that the effect of each wavelength in the range is additive,
with no synergism or mutual inhibition.

    If H(X) represents an actual measured terrestrial spectrum, then D will
be a measure of the damage under current ozone layer conditions, arbitrarily
defined as being deteriorated to zero extent.

    The above approach of using the actual measured spectrum of the solar
irradiance as the source of spectral distribution limits the usefulness of the
analysis to specific localities and possibly to certain times of the year (the
time when H(X) is fair representation of the solax spectrum).  This approach
avoids various scattering coefficients and their effects.  If we assume that
the depletion of the ozone layer has the sole effect on the spectral
distribution (with other effects being small), the damage index can be
modified (D*), which will relate to damage under reduced ozone level
conditions.

    This requires a knowledge of the absorption coefficient of ozone.
Reliable data have been published for the absorption coefficient, A(X)
(Griggs 1968):

                A(X) = exp (38.54 - 0.1237X)                              (24)
    The effect of reduced ozone concentration on the terrestrial solar
irradiance at a given wavelength can be expressed as


                    I(X,c',9) = I0exp(-A(X)c'sec8-T(X)sec8)               (25)


    where

        A(X) = ozone absorption coefficient, equation (20)

        c'   = ozone concentration at reduced level

        secB = approximation term for path length of light through the air
                 mass



                          * * *  DRAFT FINAL  * * *

-------
                                   13-23



        T    = Rayleigh scattering thickness

        8    = solar zenith angle at the location
    If we neglect the Rayleigh scattering, the factor increase in the UV
radiation due to ozone loss can be given by


                I(X, c', 8)/r = exp(A(X)c(l-c'/c)sec8)                   (26)


where (l-c'/c) is the fractional decrease (f) in the ozone concentration.

    The damage index under the conditions of ozone layer deterioration can be
expressed as
                              Xl
                     D(f,6) =  / F(X)H(X)exp(A(X)»c«fsec8)dX             (27)

                              X0
    The seasonal variation of ozone concentrations, the seasonal variation of
the solar zenith angle and the diurnal variation of the sun angle (time angle)
should be considered while estimating the damage index for different time
periods (i.e. daily, annual, seasonal, etc.)  For instance, the daily damage
index at a given level of ozone depletion is given by


                    D(f,8)24 = 2 Z //F(X)H(X)A(X)dXdt                    (28)


where A(X) is the factor increase defined in equation (26).  The value of
ozone concentration c is no longer a constant but is time dependent c(t).  Z
is a constant term used to convert the time angle to seconds.

Illustration  of Comprehensive  Approach  for  Polyvinylchloride:  A Case Study

    Using the solar spectrum as observed at Miami, Florida, on March 23, 1985,
at the solar noon, spectral distribution H(X) was calculated for various
wavelengths.  An adjusted normalized action spectrum , F(X), was developed
using the Miami sunlight spectrum.  The integral in equation (28) was
evaluated numerically using a computer program.  Damage function and the
                          * * *  DRAFT FINAL  * * *

-------
                                   13-24
factor increase in the damage were calculated for several solar zenith angles
at different fractional ozone depletion.  The results are shown in Exhibit
13-11.

    As might be expected, the factor increase in damage is higher at larger
solar zenith angles and hence will be greater for higher latitudes.  However,
the solar irradiation is also lower at these latitudes.

    The value of the damage index is very sensitive to (a) the action spectra
used in its calculation, (b) the zenith angle of the sun at the location in
question, and (c) the thickness of the polymer used.

    The thickness of the polymer is an important factor.  In the case of a
film of the material, a fraction of the incident light is transmitted through
it.  Often, the transmitted light is deficient in UV radiation.  In the case
of a thick polymer, the light is totally absorbed but the absorption is
limited to a thin surface layer of the polymer.  Thus in those modes of damage
involving the surface integrity, thick samples undergo as much damage as the
thin films.  This is true, for instance, where "yellowing" is the critical
mode of damage.

Mathtech  Study

    This is the most recent study of estimating the economic damage from
potential depletion of ozone.  The study utilizes the comprehensive approach
developed by the Research Triangle Institute (Andrady 1986).

    According to the comprehensive approach, the total damage D is given by


                         D =   X H(X) A (X) F(X) dX          (equation (23))


    The integral might be evaluated numerically with A(X)=1 to represent
current baseline conditions.  However, the  lack of activation spectrum for
polymers limits the usefulness of the approach.  The only action spectrum
available in literature relates to transparent PVC films and to type of damage
other than yellowing.  However, a very approximate estimate might be made
using the activation- spectra for polyene formation based on Reinisch et al.
1970.  Yellowing in PVC is a direct result  of the generation of long polyene
sequences.
                              *  DRAFT FINAL  * * *

-------
                                 13-25
                             EXHIBIT 13-11

        Relative  Damage  Indices for Yellowing of PVC  Under Miami
      (March  22nd) Conditions,  at  Different Extents of Ozone  Layer
                             Deterioration
                             Relative Damage  Index*  at  Different Solar
                            	Zenith  Angles	
Fractional Loss of Ozone
30'
60s
0
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
0.50
1
1.01
1.02
1.03
1.04
1 . 05
1.06
1.08
1.09
1.11
1.13
1
1.01
1.02
1.03
1.05
1.06
1.08
1.09
1.11
1.14
1.17
1
1.02
1.04
1.06
1.09
1.13
1.18
-1.25
1.36
1.55
1.93
Note:  Solar irradiance values used were for a surface facing  south  at 45°.

*  Relative damage index = D/D  where D  is the damage index calculated

   for current levels of ozone.

Source:  Andrady, Anthony, "Analysis of Technical  Issues  Related  to  the
         Effect of UV-B on Polymers," Research Tringle Institute,  Research
         Triangle Park, North Carolina, March 1986.
                        * * *  DRAFT FINAL  * * *

-------
                                   13-26



    Based on the data, the damage (D) can be estimated as


                                Ln(D) = a + g(S)                          (29)
        where  a, g = constants
                  S = stabilizer concentration
    For reduced ozone level, this equation can be modified further to the
following form:


                             S'    1       D'
                             — = (— • In —) + 1                    (30)
                             S    g(S)      D


    where   D_^ = factor increase in damage
            D

            S_^ = factor increase in titanium dioxide stabilizer
            S
            S ,D  = stabilizer concentration and damage at reduced ozone level


    This equation allows the determination of the required concentration of
the stabilizer of increased extent of damage.  Applying the data available for
yellowing (damage) of rigid PVC versus various stabilizer concentrations to
the above equation, ranges of factor increase in damage and stabilizer
concentration are calculated for PVC as illustrated in Exhibit 13-12.  The
same approach can be extended to other polymers and stabilizers.

    To calculate the economic damage, Mathtech has maintained a hypothesis
similar to the other studies, which is that producers of .the PVC products will
change the amount of stabilizer (titanium dioxide) in the resin compound in
order to maintain quality and lifetime characteristics of their products.
However, the earlier studies (Schwartz, Gordon, and Hawkins 1974; Battelle
1985) did not consider how changes in compound formulation would affect
production costs of PVC--which can occur in several ways.  First, resource
costs will be increased as more titanium dioxide is used per pound of plastic
produced.  Second, the change in formulation will lead to more frequent
replacements of screws and barrels since titanium dioxide is abrasive in
nature and its higher concentration will increase wear and tear of the
machinery.  Third, the change in formulation will lead to increased energy
requirements for PVC processing equipment since the melt viscosity of the
extrudate increases with titanium dioxide increase and will require higher
energy to operate the screw.  Also increased viscosity may require more
lubricant and processing for successful extrusion.
                          * * *  DRAFT FINAL  * * *

-------
                                   13-27
                               EXHIBIT 13-12

        Estimated Ranges of Factor  Increase  in Damage and  the  Factor
           Increase in Stabilizer Needed to Counter the Change  for
                    Yellowing of Rigid PVC  Compositions
                               Zenith Angle 30°          Zenith Angle 60°
                               Spring ~ Noontime         Fall ~ Noontime
   Percent Loss of Ozone       D'/D        S'/S          D'/D        S'/S
0-5
5-10
• 10-20
20-30
~ 1.01
1.01-1.02
1.02-1.05
1.03-1.08
1.01-1.02
1.01-1.05
1.03-1.11
1.03-1.18
1.01-1.02
1.03-1.04
1.04-1.09
1.07-1.18
1.01-1.05
1.03-1.09
1.05-1.20
1.08-1.38
   Note:  D'/D = factor increase in damage
          S'/S = factor increase in titanium dioxide stabilizer

          Zenith angles selected to reflect North American locations.

   Source: Horst, R., K. Brown, R. Black, and M.  Kianka,  "The Economic Impact
           of Increased UV-B Radiation on Polymer Materials:  A Case Study of
           Rigid PVC,"  Mathtech, Inc.,  Princeton, New Jersey, June 1986.


    In order to estimate the cumulative damage, demand for PVC was estimated
using historical data and related to construction activity since PVC is used
heavily in several outdoor construction activities.  The supply side of the
market for PVC products was determined through a model plant  analysis.  The
information from demand and supply analyses is used to determine aggregate
price indices for PVC products and subsequent economic damage.

  .  The ozone depletion estimates were the same as shown in Exhibit 13-8.   The
total PVC damage associated with ozone depletion is given in  Exhibit 13-13.

    The results are made for the following set of circumstances:  1) All
estimates are reported in millions of 1984 dollars; 2) the factor increase in
titanium dioxide concentration is computed for a zenith angle of 60°;  and 3)
polymer producers do not respond to the hypothesized decreases in
stratospheric ozone until 10 years after the impact is observed.

    As indicated so far, due to the lack of data, the Mathtech analysis is
limited to PVC only.  In order to provide broader coverage of potential
                          * * *  DRAFT FINAL  * * *

-------
                                   13-28
damages, it is appropriate to consider other polymers that may be adversely
impacted by increased UV-B radiation.

                               EXHIBIT 13-13

                      PVC Damage with Ozone Depletion
                          (Millions of 1984 dollars)
Discount Rate

Low Estimate of PVC Consumption
Middle Estimate of PVC Consumption
High Estimate of PVC Consumption
0%
2,440
4,716
9,158
2%
603
1,137
2,159
5%
97
174
315
10%
10
17
27
    Source: Horst, R.,  K. Brown, R. Black, and M.  Kianka,  "The Economic
            Impact of Increased UV-B Radiation on Polymer  Materials:  A Case
            Study of Rigid PVC," Mathtech, Inc., Princeton,  New Jersey,
            June 1986.
    The following equations, developed by Mathtech, Inc.,  provide future
demand projections for each polymer:
                          Q = A (1 - exp (-kT))                 (31)
For small values of parameter k, the above equation can be linearlized as
follows:
                          QT = a+b (T-T0) + c(T-TQ)2            (32)
    where       Q = consumption per person in year T

               T- = year for which the projection is made

          a, b, c = parameters to be estimated.


    Projections of future demands for selected polymers and selected years are
provided in Exhibit 13-14.  Damage to these polymers is difficult to estimate
without any activation spectra.  However, the magnitude of the damage is
expected to be similar to that for PVC.
                            * *  DRAFT FINAL  * * *

-------
                                   13-29
                               EXHIBIT 13-14

              Projections of Future Demand for Selected Years
                          (Thousands  of Metric  Tons)
                        Year     Acrylics      Polyester
1980
2000
2025
2050
2075
75.0
148.5
179.4
186.7
190.0
644.0
1130.6
1400.0
1472.8
1504.8
                    Source:  Horst,  R. ,  K.  Brown,  R.  Black,
                            and M.  Kianka,  "The Economic
                            Impact  of  Increased UV-B
                            Radiation  on Polymer  Materials:
                            A Case  Study of Rigid PVC,"
                            Mathtech,  Inc.,  Princeton, New
                            Jersey, June 1986.


EFFECT  OF TEMPERATURE  AND HUMIDITY ON PHOTODEGRADATION

    A necessary element in the discussion of damage  functions missing so far
is the possible role of temperature and humidity.  Increasing the temperature
almost always results in an increase in the rate  of  a  chemical reaction.  This
is certainly true of the chemical reactions responsible for photodegradation.
The extent of such an increase depends upon the energy of activation for the
particular reaction.  The activation energy for photodegradation of
polypropylene in air, for instance, is 77-80 kcal/mole.  That for PVC (based
on polyene formation) determined by Reinisch et al.  (1970) was 18 kcal/mole.
These values allow the estimation of approximate  factor increase in the
weathering reaction due to increased temperature.

    The general effect of humidity or  water is to increase the rate of
degradation of the polymer.   It is  generally believed  to cause slight
plasticization (or softening) of the polymer matrix.  This is hardly likely to
be the only mechanism.  Several experimental studies strongly suggest
water/pigment interactions.   The effect of irradiation of titanium dioxide in
                              *  DRAFT FINAL  * * *

-------
                                   13-30
the presence of water was shown to generate hydrogen peroxide (Hoffmann et al.
1971).  The deleterious effect of peroxides on polymers is well known.   Thus
depending upon the chemical nature of polymer, the components of the compound
and the weathering factors, both temperature and humidity tend to increase the
rate of degradation.   Generalized quantitative data are not available on the
polymers of interest to provide potential damage estimates.

FUTURE RESEARCH

    As evident in the discussion of this chapter, there are no experimental
data available on several polymers of interest to provide a good estimate of
damage in polymers under increased light intensity.  The EPA is currently
funding a research study to evaluate the available scientific and technical
information and conduct experimental as well as field studies to evaluate
effects of increased UV radiation on polymers.

    Poly(vinylchloride) (PVC) polymers are selected for the experimental study
since PVC is the most widely used thermoplastic for outdoor applications.  Two
main categories of PVC formulations, (a) rigid PVCs such as those used in
siding and window frames and (b) plasticized PVCs such as those used in
flexible roofing membranes and cable coatings, will be studied.  Typical
formulations and formulations with varying amounts of titanium dioxide will be
evaluated in a simulated solar radiation using a Xenon lamp source.

    The field research will study the question of UV-induced material damage
from a global point of view.  The types of UV radiation (in terms of spectral
quality) which North Americans would receive at 10-20% ozone depletion is
typical of the solar flux received at some equatorial locations at current
levels of ozone.  Studying the material behavior at these locations  would be
invaluable in estimating the effects of increased UV light at northern
latitudes.

    The yellowing of PVC will be the main criterion used to assess damage,
although other tests may be employed as necessary.  Data obtained from this
research will be used to assess damage to PVCs and to estimate additional
costs to stabilize PVC in the event of partial ozone depletion.
                          * * *  DRAFT FINAL  * * *

-------
                                   13-31
REFERENCES
Andrady, Anthony, "Analysis of Technical Issues Related to the Effect  of  UV-B
    .on Polymers," Research Triangle Institute,  Research Triangle Park,  North
    Carolina, March 1986.

Cutchins, 1974

Griggs, M., Journal of Chemical Physics. Vol.  49(2),  pp.  857,  1968.

Hattery, G., V. McGinniss, and P.  Taussig "Costs Associated with Increased
    Ultraviolet Degradation of Polymers," BATTELLE Columbus Laboratories,
    Columbus, Ohio, April 1985.

Heller, H.,  European Polymer Journal, Supplementary, 105, 1969.

Horst, R., K. Brown, R. Black, and M.  Kianka,  "The Economic Impact of  Increased
    UV-B Radiation on Polymer Materials:  A Case Study of Rigid PVC,"
    Mathtech, Inc., Princeton, New Jersey, June 1986.

Hoffmann, E. and A. Saracz, J. Oil Color Chem.  Association. Vol. 54, p. 450,
    1971.

Kelen, T., Polymer Degradation, Van Nostrand Reinhold Company, Inc.,
        New York, 1983.

King, A., Plastics and Polymers, 195,  1968.

Reinisch, R., R. Gloria, and G. Androes, Photochemistry of Macromolecules,
    Plenum Press, New York, 1970.

Schultz, A., D. Gordon, and W. Hawkins, "Economic and Social Measures  of
    Biologic and Climatic Change,"  CIAP Monograph 6, September 1974.

Uzelmier, C., Society of Plastic Engineers, Vol 26, p. 69, 1970.
                          * * *  DRAFT FINAL  * * *

-------
Chapter 14

-------
                              CHAPTER  14

          POTENTIAL EFFECTS  OF STRATOSPHERIC  OZONE DEPLETION
                     ON TROPOSPHERE OZONE (SMOG)
SUMMARY

    Tropospheric ozone,  03(T),  (smog) is an air pollutant formed near the
earth's surface as  a result of  photochemical reactions involving ultraviolet
radiation,  hydrocarbons,  nitrogen oxides, oxygen and sunlight.   At high
concentrations, often found during warmer months, smog can adversely affect
human health,  agricultural crops, forests and materials.   The U.S.
Environmental  Protection Agency (EPA) has established a primary standard to
prptect public health of 0.12 ppm (one-hour average) not to be exceeded more
than one day per year.   In 1979, EPA also determined that a secondary welfare
standard, more stringent than the primary standard, was unnecessary for the
protection of  vegetation.  Currently, EPA is reviewing available scientific
and technical  information to determine whether these standards are adequate to
protect health and  welfare.

    This chapter reviews preliminary scientific information which suggests
that increases in UV-B radiation may effect rate of tropospheric ozone (smog)
formation in urban  areas.  The  results from these analyses suggest that
increased ultraviolet radiation increases the rate of smog production and  acid
rain precursors. Moreover, global warming, associated with ozone depleting
substances  may enhance these reactions.  Further analyses of additional cities
are being conducted.   If these  analyses confirm these results,  it would appear
likely that in the  future more  cities and regions would violate the ambient
air standards  and that more restrictive measures to control hydrocarbons and
nitrogen oxides may be required in order to comply with current standards.
                            * *  DRAFT FINAL  * * *

-------
                                   14-2
FINDINGS

    1.   PRELIMINARY RESULTS FROM ONE STUDY SUGGEST THAT INCREASED ULTRAVIOLET
        RADIATION FROM OZONE DEPLETION WOULD INCREASE THE RATE OF
        TROPOSPHERIC OZONE FORMATION.

        la.  Increases in UV-B associated with ozone depletion would increase
             the quantity of increase in ground-based ozone associated with
             various hydrocarbon and nitrogen oxides emission levels depending
             on the chemistry of each city.

        Ib.  Global warming would enhance the effects of increased UV-B
             radiation on the formation of ground-based ozone.

        Ic.  Ground-based ozone is predicted to form earlier in the day.   This
             may cause larger populations to be exposed to peak values in some
             cities.

    2.   PRELIMINARY RESULTS ALSO PREDICT LARGE INCREASES IN HYDROGEN PEROXIDE.

        2a.  If hydrogen peroxide increases  as predicted, the oxidation
             potential of the atmosphere, including the formation of acid
             rain, would be influenced.

    3.   INCREASES IN GROUND BASED OZONE WOULD ADVERSELY AFFECT PUBLIC HEALTH
        AND WELFARE.

        3a.  There is a potential for more cities to be unable to meet health-
             based standards.

        3b.  Crops, ecosystems, and materials could be adversely affected.
                          * * *  DRAFT FINAL

-------
                                   14-3
INTRODUCTION

    Tropospheric ozone, 03, (smog) is an air pollutant formed in the ambient
air as a result of a series of complex photochemical reactions involving
ultraviolet radiation, hydrocarbons, and nitrogen oxides emitted from mobile
and stationary sources, atmospheric oxygen, and sunlight.  At ambient
concentrations, often measured during warmer months, 03 can adversely affect
human health, agricultural crops, forests, ecosystems, and materials.
Interactions of 03 with nitrogen oxides and sulfur oxides may also contribute
to the formation of acidic precipitation.   Typical short-term (1-hour) 03
levels range from 0.01 ppm in some isolated rural areas to as high as 0.35 ppm
in one of the nation's most heavily populated metropolitan areas.  Daily
daylight seasonal averages in some rural areas have been reported to be 0.06
ppm and higher.

    Using 1982-1984 data -- the most recent available -- 216 metropolitan
statistical areas (MSAs) out of 319 MSAs (68%) have enough 03 air quality data
to ascertain attainment status.  119 (MSAs) (55%) have more than one expected
exceedance per year of the current EPA health-based 03 standard of 0.12 ppm.
Thus, more than one-half of the MSAs with sufficient data violate the
standard.  Approximately 115 million people, i.e., over one-half of the total
U.S. population, live in areas which exceed the standard.  However, this does
not mean that everyone in these areas is exposed to 03 concentrations at or
above the standard (EPA 1986).

    Of MSAs with sufficient data, about 14% (34) have a characteristic highest
concentrations (CMC) of 0.16 ppm 03 or higher and over 4% (11) of MSAs have a
CMC above 0.20 ppm 03.  There is no clear temporal trend in 03 concentration
levels in MSAs around the country, although 1982 -- and to a lesser extent,
1981 -- had generally lower levels than other years during the 1979-1984 time
period.

    Generally, third quarter (July-September) and seasonal (April-September)
average of 8:00 a.m. - 4:00 p.m. daily daylight values are both in the range
of 0.043-0.050 ppm 03.  Statistically significant, but modest, relationships
exist between peak and longer-term mean indices of 03 air quality in urban
areas.

    Recently, EPA (1986) reviewed scientific and technical information on the
known and potential health effects of ozone.  The information includes
respiratory tract absorption and deposition of ozone, studies of mechanisms of
03 toxicity,- effects of exposure to 03 reported in controlled human exposure,
field, epidemiological and animal toxicology studies, as well as air quality
information.  The results of that review suggest that:

        (1)  The mechanisms by which inhaled 03 may pose health
             risks involve (a) penetration into and absorption of 03
             in various regions of the respiratory tract, (b)
             pulmonary response resulting from chemical interactions
                          * * *  DRAFT FINAL  * * *

-------
                           14-4
     of 03 along the respiratory tract,  and (c)
     extrapulmonary effects caused indirectly by effects of
     03 in the lungs.

(2)   The risks of adverse effects associated with absorption
     of 03 in the tracheobronchial and alveolar  regions of
     the respiratory tract are much greater than for
     absorption in the extrathoracic region (head).
     Increased exercise levels are generally associated with
     higher ventilation rates and increased oronasal or oral
     (mouth) breathing.  Thus, maximum penetration and
     exposure of sensitive lung tissue occurs when heavily
     exercising individuals are exposed to .03.

(3)   Factors which affect susceptibility to 03 exposure are
     activity level and environmental stress (e.g.,
     humidity, high temperature).

(4)   The major subgroups of the population at greatest risk
     to the effects of 03 include:  (a)  individuals with
     preexisting respiratory disease (e.g., asthmatics and
     persons with chronic obstructive lung disease or
     allergies), (b) "responders" who are otherwise healthy
     individuals, both adults and children, but  experience
     significantly greater group mean lung function response
     to 03 exposure, and (c) any individual exercising
     heavily during exposure to 03.

(5)   The major health concerns associated with exposure to
     03, in approximate order of strength of the data base
     include:

     (a)    alterations in pulmonary function;

     (b)    symptomatic effects;

     (c)    effects on work performance;

     (d)    aggravation of preexisting respiratory disease;

     (e)    morphological effects (lung structure damage);

     (f)    altered host defense systems (i.e.,  increased
            susceptibility to respiratory infection); and

     (g)    extrapulmonary effects (e.g., effects on blood
            enzymes, central nervous system, liver,
            endocrine system).
                  * * *  DRAFT FINAL  *

-------
                                   14-5
        (6)  The most useful exposure-response information is from
             controlled human exposure and field studies which
             .provide a quantitative relationship between alterations
             in pulmonary function and 03 exposure concentrations.

The EPA (1986) document came to the following conclusions on vegetation
impacts:

        (a)  The mechanisms by which 03 may injure plants and plant
             communities include (1) absorption of 03 into leaf
             through stomata followed by diffusion through the cell
             wall and membrane, (2) alteration of cell structure and
             function, as well as critical plant processes,
             resulting from the chemical interaction of 03 with
             cellular components, (3) occurrence of secondary
             effects including reduced growth and yield and altered
             carbon allocation;

        (b)  The magnitude of the 03 induced effects depends upon
             the physical and chemical environment of the plant, as
             well as varous biological factors (including genetic
             potential, developmental age of plant and interaction
             with plant pests);

        (c)  Effects, of 03 on vegetation and ecosystems have been
             demonstrated to occur from both short-term and
             long-term exposures.  Although there are a limited
             number of studies in which short-term (1-2 hour)
             exposures have resulted in growth and yield reduction,
             there is a growing body of evidence that repeated peaks
             above a given level are important in eliciting plant
             response;

        (d)  Concerning long-term exposures, the bulk of the
             evidence indicates that growth and yield losses occur
             in several plant species exposed to seasonal
             concentrations of 03, typically characterized as the
             daily daylight mean over the growing season.  In
             addition, evidence indicates that forests experience
             cumulative stress as a result of chronic exposure to
             03.  Exhibit 14-1 summarizes the range of 03 levels and
             exposure times required to induce 5% and 20% foliar
             injury.  Exhibit 14-2 (EPA 1986) provides a more
             complete survey; and

        (e)  Damage to materials is another effect of 03.  There
             appears to be no threshold level below which material
             damage will not occur; the slight acceleration of the
             aging processes of materials occurs at the level of the
             proposed standard.  The materials known to be most
             susceptible to ozone attack are elastomers, textile
             fibers and dyes, and certain types of paint.
                          * * *  DRAFT FINAL  * * *

-------
                              14-6
                         EXHIBIT 14-1

     Ozone  Concentrations for Short-Term Exposure That Produce
               5% or 20% Injury to Vegetation Growth
                  Under Sensitive Conditions1
                            Ozone Concentrations (ppm)
             	that may Produce 5% (20%) Injury:	
 Exposure                                                  Less
Time, Hour   Sensitive Plants    Intermediate Plants   Sensitive Plants
0.5

1.0

2.0

4.0

8.0
0
(0
0
(0
0
(0
0
(0
0
.35
.45
.15
.20
.09
.12
.04
.10
.02
to 0
to 0
to 0
to 0
to 0
to 0
to 0
to 0
to 0
.50
.60)
.25
.35)
.15
.25)
.09
• 15)
.04
0.
(0.
0.
(0.
0.
(0.
0.
(0.
0.
55
65
25
35
15
25
10
15
07
to 0
to 0
to 0
to 0
to 0
to 0
to 0
to 0
to 0
.70
.85)
.40
.55)
.25
.35)
.15
.30)
.12
>0

>0

>0

>0

>0
.70

.40

.30

.25

.20
(0.85)

(0.55)

(0.40)

(0.35)

(0.30)
1 The concentrations  in  parenthesis are for the 20% injury level.
Exhibit from U.S.  Environmental Protection Agency (1985, p. 7-225).
                     * *  *   DRAFT FINAL  * *

-------
                                                           EXHIBIT  1'4-2

                            Ozone  Concentrations  at Which  Significant  Yield  Losses  Have  Been  Noted  for
                             a  Variety  of  Plant Species  Fxposnd  Under'  Various  Experimental Conditions
  Plant Species
Exposure Duration
Yield Reduct i on .
  % of Control
0  Concentration,
 3
                                                                                               ppm
                                 Reference
AI fa I fa
Alfalfa
Pasture grass
Ladino clover
Soybean
Sweet corn
Sweet corn
Wheat
Bad i sh
Beet
Potato

Pepper
Cotton
Ca rnat ion
Coleus
Begonia

Ponderosa pine
Western white pine
Lob lolly p ine
Pitch pine
Pop la r
Hybrid poplar
Hybrid poplar
Red maple
American sycamore
Sweetgum
White ash
Green ash
WiI low oak
Sugar maple
7 hr/day, 70 days
2 hr/day, 21 day
4 hr/day, 5 days/wk, 5 wk
6 hr/day, 5 days
6 hr/day, 133 days
6 hr/day, 64 days
3 hr/day, 3 days/wk, 8 wk
4 hr/day, 7 day
3 hr
2 hr/day, 38 days
3 hr/day, every 2 wk.
120 days
3 hr/day, 3 days/wk, 11 wk
6 hr/day, 2 days/wk, 13 wk
24 hr/day, 12 days
2 hr
4 hr/day, once every 6 days
for a total of 4 times
6 hr/day, 126 days
6 hr/day, 126 days
6 hr/day, 28 days
6 hr/day, 28 days
12 hr/day, 5 mo
12 hr/day, 102 days
8 hr/day, 5 day/wk, 6 wk
8 hr/day, 6 wk
6 hr/day, 28 days
6 hr/day, 28 days
6 hr/day, 28 days
6 hr/day, 28 days
6 hr/day, 28 days
6 hr/day, 28 days
51,
16.
20,
20,
55,
45,
13,
30,
33,
40,
25,

19,
62.
74,
20,
55,

21,
9,
18,
13,
top d ry wt
top dry wt
top dry wt
slioot dry wt
seed wt/plant
send wt/p I fint
ear fresh wt
seed yield
root dry wt
storage root dry wt
tuber wt

fru it d ry wt
r i ber d ry wt
no. of flower buds
flower no;
flower wt

stem dry wt
stem dry wt
height growth
height growth
•H333, leaf abscissi~n
58.
50,
37,
9,
29,
17,
24,
19,
12,
height growth
shoot dry wt
height growth
height growth
he i ght g rowth
tota I dry we ight
height growth
height growth
height growth
0. 10
0. 10
0.09
0. 10
0. 10
0. 10
0.20
0.20
0.25
0.20
0.20
Nee ly et a 1 . , 1977
Hoffman et a I . , 1975
Horsman et a I., 1980
Blum et a I . , 1982
Heag le et a I . , 1974
Heag le et a I . , 1972
Oshima, 1973
Shannon and Mulchi, 1974
Adedipe and Ormrod, 1974
Ogata and Maas, 1973
Pel I et a 1 . , 1980
                                                          0.12
                                                          0.25
                                                        0.05-0.09
                                                          0.20
                                                          0.25

                                                          0. 10
                                                          0. 10
                                                          0.05
                                                          0. 10
                                                          0.0'41
                                                          0. 15
                                                          0.15
                                                          0.25
                                                          0.05
                                                          0. 10
                                                          0. 15
                                                          0. 10
                                                          0. 15
                                                          0.15
                                               Bennett et a I., 1979
                                               Oshima et a I., 1979
                                               Feder and Campbell, 1968
                                               Aded ipe et a I., 1972
                                               Reinert and  Nelson, 1979

                                               Wilhour and  Neely, 1977
                                               Wilhour and  Neely, 1977
                                               Wilhour and  Neely, 1977
                                               Wilhour and  Neely, 1977
                                               Wilhour and  Neely, 1977
                                               Patton, 1981
                                               Patton, 1981
                                               Dochinger and Townsend,  1979
                                               Kress and Skelly,  1982
                                               Kress and Skelly,  1982
                                               Kress and Skelly,  1982
                                               Kress and Skelly,  1982
                                               Kress and Skelly,  1982
                                               Kress and Skelly,  1982

-------
                                   14-8
 POTENTIAL EFFECTS OF ULTRAVIOLET RADIATION
 AND  INCREASED TEMPERATURES  ON URBAN SMOG

    Recently, Whitten (1986) conducted a preliminary investigation into the
 potential changes in urban ozone levels because of an increase in solar
 ultraviolet  (UV) radiation resulting from reductions in the stratospheric
 ozone  layer.  Estimates were also made of the effects on urban ozone chemistry
 resulting from a general warming of the lower atmosphere.  The focus of this
 study  was the effect of reductions in stratospheric ozone by as much as 30
 percent of  (1) peak ozone concentrations in urban areas and (2) the attendant
 control requirements predicted by a modified version of the Empirical Kinetics
 Modeling Approach (EKMA).  The model used in this study was the simple
 trajectory model used in the EKMA.  This model was developed by the EPA and is
 widely used  to assess the effectiveness of emissions control scenarios to
 abate  urban  ozone.

    Whitten  began by investigating how the ozone formation processes in the
 lower  troposphere would be affected by increased UV transmission through the
 atmosphere.  From preliminary work, it became evident that the photolysis
 channel of formaldehyde that leads to radical products is selectively enhanced
 at a higher  UV flux.  Formaldehyde emissions are products of incomplete
 combustion,  and  formaldehyde is a major intermediate oxidation product from
 virtually all organic molecules.  Radicals from formaldehyde photolysis
 provide the  main source of radicals needed to drive the chain reactions which
.generate photochemical smog (Whitten 1983), therefore, an increase in rate of
 photolysis will have a bearing on the design of future control strategies.

    The photolysis of ozone to electronically excited oxygen atoms is thought
 to be  the second most important source of the radicals that drive smog
 formation.   However, the role of ozone photolysis is different from
 formaldehyde because ozone is the principal ingredient of photochemical smog.
 At low levels of oxidation potential, the excited oxygen atoms tend to
 accelerate smog reactions, making the atmospheric chemistry more efficient in
 generating ozone from minimal precursor emissions.  However, at the higher or
 more severe  ozone levels, the excess radicals can partially suppress the ozone
 peak,  making the precursors seem less efficient in generating ozone.  Other
 photolysis rates can also be affected, but their contribution to smog
 formation is less important.

    A  specific increase in formaldehyde photolysis to radical products is
 difficult to accurately determine for a number of reasons.  Atmospheric
 photolysis rates have often been calculted in 10 nm wavelength intervals.
 However, the accurate calculation of formaldehyde photolysis requires fine
 spectral resolution for both the formaldehyde absorption cross-section and the
 surface solar flux (related to the ozone absorption cross-section).  Only in
 this manner  can a comparison of the respective fine structures be performed.
 Near-ground  solar flux data of high resolution and known stratospheric ozone
 abundance have not been readily available.  Hence, Whitten (1986) estimated
 surface solar fluxes using low-resolution information provided by the National
 Aeronautics  and Space Administration.
                          * * *  DRAFT FINAL  * * *

-------
                                   14-9
    Using the spectral data, Whitten (1986) estimated formaldehyde and ozone
photolysis rates as a function of zenith angle and ultraviolet flux changes
due to stratospheric ozone depletion.  The estimated photolysis rates for
formaldehyde, coupled with the information compiled for surface ozone
photolysis, provide the inputs needed to calculate diurnal photolysis rates
for projected future ozone column densities.  This information is used to
evaluate the future impact of stratospheric ozone changes on near-surface smog
formation.

    The preliminary results of the possible effects of increased ultraviolet
radiation on urban smog are based on simulation of atmospheric conditions for
three urban cities:

        (1)  Nashville -- because it is nearly in compliance with
             the 0.12 ppm federal ozone standard.

        (2)  Philadelphia -- to represent cities that require
             moderate control (30%-50% reduction in organic
             precursors) to achieve the 0.12 ppm standard.

        (3)  Los Angeles -- because of the severity of the
             exceedance of the ozone standard in that region.

Present-day estimates assume a total ozone column of 300 DU and future
predictions assume a total ozone column of 250 and 200 DU.  A 33% decrease in
total ozone column would increase ozone photolysis by approximately a factor
of two, and increase formaldehyde photolysis to radical products by nearly
20%.  Both of these increases are approximations because of uncertainties
regarding the spectral fine structure and other factors.

    The effects of these photolysis rate increases are given in Exhibit 14-3
for background temperatures of 298K and 302K.  For all three cities, the model
predicts higher ozone concentrations resulting either from increases in
temperature or decreases in the Dobson number.  The magnitude of the increases
in ozone is apparently a function of local .hydrocarbon-to-nitrogen oxide
ratio, reactivity, meteorology, and emission distribution.  The linearity of
the response to stratospheric ozone depletion is not general.  For the Los
Angeles case, the increases in ozone are moderate and quite linear with
decreasing Dobson number.  For the Philadelphia case, ozone is predicted to
increase progressively as the Dobson number declines; that is, only a modest
increase in ozone is seen for a decline in Dobson number from 300 to 250
units, but a more dramatic increase in ozone is seen when the Dobson number
declines from 250 to 200 units.  For the Nashville case,-the increases are all
very dramatic and the linearity exists only if temperature does not increase.
At the "current" temperature of 298K, the simulated urban ozone tends to
increase linearly with Dobson number decrease; but at the warmer 302K
temperature, the simulated ozone increases more for the first 50-unit Dobson
change than for the second 50-unit change, down to the 200 unit limit
studies.  Preliminary trajectory model results suggest that peak smog levels
are reached earlier in the day, which would expose larger populations to peak
values.  Hydrogen peroxide increased from approximately 1.7 ppb to 3.3 ppb
                          * * *  DRAFT FINAL  * * *

-------
                           14-10
                      EXHIBIT  14-3

      Ozone Concentrations (ppm)  Predicted for Changes in
        Dobson Number and Temperature for Three Cities
                             Ozone Concentrations
Temperature (K)
Dobson Number

300
298K
250

200

300
302K
250

200
Los Angeles*
Philadelphia
Nashville
0.288
0.112
0.130
0.301
0.127
0.161
0.315
0.149
0.195
0.306
0.122
0.146
0.318
0.134
0.184
0.331
0.159
0.215
* For Los Angeles, values under 298K used actual  hourly tempera-
tures and values under 302K are from simulations  using those
temperatures  increased by 4K.
                    * *  DRAFT FINAL  * * *

-------
                                   14-11
in Los Angeles and from .19 ppb to approximately 3.0 ppb in Philadelphia for a
100 unit change in Dobson.  These later changes suggest significant potential
impacts for ozone depletion on acid rain formation.

    The effects of a 4K increase in temperature are most pronounced in the Los
Angeles and Nashville simulations where 14 to 23 ppb increases in simulated
ozone peaks are predicted.  The effects of a combined temperature increase and
reduce Dobson number appear to be additive, if not synergistic,  especially for
the Nashville simulations.  When a 4K temperature increase is combined with
only a 50-unit Dobson reduction, the simulated ozone peak jumps  from a base
value of 0.13 ppm to over 0.18 ppm ozone.   See Exhibit 14-4.

CONCLUSIONS AND  FUTURE  RESEARCH DIRECTIONS

    The preliminary study (Whitten 1986) was limited in several  ways because
the primary focus was on the atmospheric chemistry of ozone formation e.g.,
potential changes in acid rain were not addressed.  (Although the effects of
local warming and increased ultraviolet radiation were addressed in the
preliminary study, increases in air pollution such as smog and acid rain can
only be addressed through more extensive modeling using more complex
atmospheric models).  However, the simple, moving box model remains applicable
in the more extensive modeling.

    Although the initial study involved the use of only one atmospheric
chemical mechanism, the Carbon Bond Mechanism (CBM), the key reactions are
virtually identical in all currently accepted atmospheric mechanisms.  Hence,
it can be expected that other mechanisms would show the same or  similar
responses to the global effects.  Thus, the use of the CBM is not considered a
serious limitation of the current work.  However, the increased  radiation
effects studied could not utilize high resolution spectral data  even though
the preliminary study seemed to show a need to use such data.

    Thus far, atmospheric modeling has addressed only three trajectories, one
each in three cities.  Three cities is not an adequate sample/size to
characterize the potential range of conditions affected by these global
effects.  Although the model is sensitive to initial and boundary conditions,
these conditions were not varied, even though the hypothesized global effects
would affect such conditions.  The data for initial and boundary conditions
were primarily composed of air from upwind cities or air recirculated from the
day before.  Emissions were not varied to account for local warming, even
though many emissions increase with temperature.  Climate and other
meteorological changes would affect important factors such as mixing height
and wind flow patterns.  These possible changes were also not addressed in the
preliminary study.

    A very important finding of the preliminary study is the increased rate of
smog formation due to increased ultraviolet radiation.  Even if  the peak smog
concentration is not affected, the increased rate of formation will move the
high concentrations of smog closer to the precursor emissions where the
population density is invariably higher.  Hence the number of people exposed
                                 DRAFT FINAL  * * *

-------
                           14-12
                        EXHIBIT 14-4

              Global Warming Would Exacerbate Effects of
             Depletion on Ground-Based Ozone in Nashville
17% Depletion
and
4°C Temperature
Rise
                          41.5%
                          17.7%
                        Increase in
                      Ground-Based
                          Ozone
 Source: Adapted from Whitten (1986).
  4°C Temperature
>  Rise
                                         17% Depletion
                                         Only
                    * * * DRAFT FINAL * * *

-------
                                   14-13
to high smog concentrations would increase.  This exposure effect is not
addressed in the preliminary study.  A grid type model is more appropriate for
such exposure studies.

    In summary, the following limitations of the preliminary study are noted:

        •   Effects on pollutants other than urban smog or ozone
            were not addressed; acid rain may be affected by global
            effects.

        •   The chemical effects were calculated using low
            resolution spectral data; high resolution data appear to
            be needed.

        •   Since only three cities were studied, the range of
            conditions most affected by global effects could not be
            elucidated.

        •   Multi-day effects such as inter- or intracity
            carryover of pollutants were not addressed.

        •   Temperature effects on emissions rates or changes in
            wind patterns, mixing heights, or frequency of episodic
            meteorological conditions were not addressed.

        •   Possible increases in population exposure due to
            increased smog levels and increased rates of smog
            formation were not addressed.

    Current studies attempt to address these limitations by (1) increasing the
number of trajectory cases, (2) using high resolution spectral data, (3)
studying the range of conditions which can possibly affect carryover of
pollutants, (4) varying the emissions with temperature, (5) studying climate
change effects, and (6) calculating population exposure changes.  The use of
an acid rain model and a grid model will be incorporated in the on-going
investigations.
                            * *  DRAFT FINAL  * * *

-------
                                   14-14
REFERENCES
Bass AM, LC Glasgow, C Miller, JP Jesson, and DL Filken.   1980  Planet.  Space
Sci. 28:675.
                                                  Itr
Bahe FC, WN Marx, U Schurath, and EP Roth.   1983   Determination of the
Absolute Photolysis Rate of Ozone by Sunlight, 03 + nv •* OC'D)  + 0

(*A ), at Ground Level."  Institut fur Physikalische Chemie der
   g
Universitat Bonn.  Bonn, W. Germany.

EPA 1986.  Review of the National Ambient Air Quality Standards for Ozone.
Preliminary Assessment of Scientific and Technical Information, Office of. Air
Quality Planning and Standards Staff Paper.   March 1986.

Whitten, G.Z.  1983.  The chemistry of smog formation:  A review of current
knowledge.  Environ. International, 9:447-463.

Whitten GZ, KR Styles, and MW Gerry.  1986  "Assessment of Existing
Tropospheric UV-Radiation Data and the Effect of its Increase on Ozone
Formation in the Troposphere."  Monthly Technical Narrative No. 2 to U.S.
Department of Interior, National Park Service, Washington, D.C.

Whitten, GZ, and M Gery.  1986  Effects of Increased UV Radiations on Urban
Ozone Presented at EPA Workshop on Global Atmospheric Change and EPA
Planning.  Edited by Jeffries, H. EPA Report 600/9-86016 July 1986.
                          * » *  DRAFT FINAL  » * *

-------
Chapter 15

-------
                              CHAPTER 15

                CAUSES AND EFFECTS  OF SEA LEVEL RISE
SUMMARY

    One of the most widely examined impacts of the projected global  warming
is the possible rise in sea level.   Researchers have identified at  least
three mechanisms that might cause a significant rise:   the warming  and
resulting expansion of the upper layers of the ocean,  the melting of alpine
glaciers, and the melting and disintegration of polar  ice sheets in  Greenland
and Antarctica.  Estimates of the rise through the year 2100 range,  in the
absence of efforts to limit the greenhouse warming, range from 50 cm to over
2 m.  Even the most conservative estimate implies a substantial acceleration
over the 10 to 15 cm rise of the last century.

    A rise in sea level in that range would permanently inundate wetlands and
lowlands, accelerate coastal erosion, exacerbate coastal flooding,  and
increase the salinity of estuaries  and aquifers.  Although wetlands  have kept
pace with sea level rise in the last several thousand  years, a 1-2  m rise
would destroy a majority of U.S. coastal marshes and swamps.  River  deltas
such as those of the Mississipi, Ganges, and Nile rivers appear to  be
particularly vulnerable.

    Along the open coast, beach erosion could reach 1  to 2 m for every 1 cm
rise in sea level, in addition to whatever erosion might be caused  by other
factors.  Because buildings are generally found within 50 m of the  shore,
even the 30 cm rise projected for the next 40 years could threaten  coastal
property and the recreational use of beaches, unless additional remedial
measures are implemented.

    Sea level rise would also increase the vulnerability of coastal  areas to
flooding from storm surges and rainwater. In the area  of Charleston, S.C.,
for example, the area now flooded by a 100-year storm  would be flooded by a
10-year storm if sea level rises 1.6 m. Protecting against increased flooding
would require improvement or construction of levees, seawalls, and  drainage
facilities.

    Higher water levels wold also increase the salinity of estuaries and
aquifers.  For example, Philadelphia's drinking water  intake on the Delaware
River would be threatened by a 73-cm rise, as would adjacent aquifers in New
Jersey that are recharged by the (currently) fresh part of the river.
Construction of additional reservoirs might be necessary to offset  salinity
increases.

    Few studies have estimated the economic significance of future  sea level
rise. One study suggests that the impacts of a 0.9 to  2.4 m rise by 2075
could be as great as 17 to 35 percent of-total economic activity in the
Charleston, South Carolina area and 5 to 16 percent of the activity around
Galveston, Texas.  No one has yet estimated the potential nationwide cost of
defending shorelines arid other resources from the projected rise in sea level.
                            * *  DRAFT FINAL  * *

-------
                                   15-2
FINDINGS

    1.    GLOBAL AVERAGE SEA LEVEL HAS RISEN 10 TO 15 CM OVER THE LAST
         CENTURY.

         la.    Ocean and glacial studies suggest that the rise is higher than
               what models of thermal expansion and alpine meltwater would
               project given the 0.6  C warming of the past century.

         Ib.    However, no cause-and-effect between warming and past sea
               level has been conclusively demonstrated.

    2.    THE  PROJECTED GLOBAL WARMING WOULD ACCELERATE THE CURRENT RATE OF
         SEA  LEVEL RISE BY EXPANDING OCEAN WATER, MELTING ALPINE GLACIERS.
         AND  EVENTUALLY INCREASING THE RATE AT WHICH POLAR ICE SHEETS MELT
         OR DISCHARGE ICE INTO THE OCEANS.

    3.    ESTIMATES OF THE RISE IN SEA LEVEL THAT COULD TAKE PLACE IF
         MEASURES  TO LIMIT THE GLOBAL WARMING ARE NOT UNDERTAKEN RANGE FROM
         10 TO 20  CM BY THE YEAR 2025, AND 50 TO 200 CM BY 2100.

         3a.    Greenland and alpine glaciers could each contribute 10 to 30
               cm  through 2100.

         3b.    The contribution of Antarctic deglaciation is likely to be
               between 0 and 100 CM; however,the possibilities cannot be
               ruled out that (1) increased snowfall could increase the size
               of  the Antarctic ice sheet and thereby offset part of the sea
               level rise from other sources: or (2) meltwater and enhanced
               calving of the ice sheet could increase the contribution of
               Antarctica to as much as 2 m.

         3c.    When added to estimated 40-cm rise due to thermal expansion,
               the 10 to 160 cm estimated glacial contribution estimated by
               the NAS Polar Research Board is consistent with the conclusion
               stated here.

    4.    DISINTEGRATION OF THE WEST ANTARCTIC ICE SHEET MIGHT RAISE SEA
         LEVEL 6 M OVER THE NEXT FEW CENTURIES, IN ADDITION TO THE SMALLER
         CONTRIBUTIONS FROM THERMAL EXPANSION, ALPINE GLACIERS, AND ICE
         SHEETS IN GREENLAND AND EAST ANTARCTICA.

         4a.    Glaciologists generally believe that such a disintegration
               would take at least 300 years, and probably at least 500 years.

         4b.    A global warming might result in sufficient thinning of the
               Ross and Filcher-Ronne Ice Shelves in the next century to make
               the process irreversible.
                          * * *  DRAFT FINAL  * * *

-------
                               15-3
5.   LOCAL TRENDS IN SUBSIDENCE AND EMERGENCE MUST BE ADDED OR
     SUBTRACTED TO ESTIMATE THE RISE AT PARTICULAR LOCATIONS.

     5a.   Most of the Atlantic and Gulf Coasts of the United States--as
           well as the Southern Pacific coast — are subsiding 10-20 cm per
           century.

     5b.   Louisiana is subsiding 1 m per century and parts of Alaska are
           emerging 10-100 cm per century.

6.   A SUBSTANTIAL RISE IN SEA LEVEL WOULD PERMANENTLY INUNDATE WETLANDS
     AND LOWLANDS, ACCELERATE COASTAL EROSION. EXACERBATE COASTAL
     FLOODING. AND INCREASE THE SALINITY OF ESTUARIES AND AQUIFERS.

7.   LOUISIANA IS THE STATE MOST VULNERABLE TO A RISE IN SEA LEVEL.
     IMPORTANT IMPACTS WOULD ALSO OCCUR IN FLORIDA, MARYLAND, DELAWARE,
     NEW JERSEY, AND IN THE COASTAL REGIONS OF OTHER STATES.

8.   A RISE IN SEA LEVEL OF 1 TO 2 M BY THE YEAR 2100 COULD DESTROY 50%
     TO 80% OF U.S.  COASTAL WETLANDS.

9.   EROSION CAUSED BY SEA LEVEL RISE COULD THREATEN U.S. RECREATIONAL
     BEACHES.

     9a.   Case studies of beaches in New Jersey, Maryland, California,
           South Carolina, and Florida have' concluded that a 30-cm rise
           in sea level would result in beaches eroding 20-60 m or more.

     9b.   Because the first row of houses is generally less than 20 m
           from the shore at high tide, and if available studies are
           representative, recreational beaches throughout the nation
           would be threatened by a 30-cm rise unless major beach
           preservation efforts are undertaken.

10.  SEA LEVEL RISE WOULD INCREASE THE DAMAGES FROM FLOODING, FLOOD
     PROTECTION, AND FLOOD INSURANCE IN COASTAL AREAS.

     lOa.  Flood damages would increase because higher water levels would
           provide a higher base for storm surges.

     lOb.  Erosion would increase the vulnerability to storm waves,and
           decreased natural and artificial drainage would increase
           flooding duirng rainstorms.

11.  INCREASED SALINITY FROM SEA LEVEL RISE WOULD CONVERT CYPRESS SWAMPS
     TO OPEN WATER AND THREATEN DRINKING WATER SUPPLIES IN LOUISIANA.
     PHILADELPHIA, AND NEW JERSEY.  OTHER AREAS, SUCH AS SOUTHERN
     FLORIDA, MAY ALSO BE VULNERABLE BUT HAVE NOT BEEN INVESTIGATED.
                      * * *  DRAFT FINAL  * *

-------
                               15-4
12.   ESTIMATES OF DAMAGE FROM SEA LEVEL RISE MUST CONSIDER HUMAN
     RESPONSES.
     12a.   The adverse impacts of sea level rise could be ameliorated
           through anticipatory land use planning and structural design
           changes.

     12b.   In two case studies in Charleston and Galveston,  accelerated
           anticipatory planning was estimated to reduce net damages by
           20 to 60%.

13.   OTHER IMPACTS OF A GLOBAL WARMING MIGHT OFFSET OR EXACERBATE THE
     IMPACTS OF SEA LEVEL RISE.

     13a.   Increased droughts might amplify the salinity impacts of sea
           level rise.

     13b.   Increased hurricanes and increased rainfall in coastal areas
           could amplify flooding from sea level rise.

     13c.   Warmer temperatures might impair peat formation of salt
           marshes and would enable mangrove swamps to take over areas
           that are presently salt marsh.

     13d.   Decreased northeasterners might reduce damage.

14.   RIVER DELTAS THROUGHOUT THE WORLD WOULD BE VULNERABLE TO A RISE IN
     SEA LEVEL, PARTICULARLY THOSE WHOSE RIVERS ARE DAMMED OR LEVEED.

15.   ECONOMIC STUDIES OF THE CHARLESTON, SOUTH CAROLINA AND GALVESTON,
     TEXAS AREAS SUGGEST THAT SEA LEVEL RISE COULD BE ECONOMICALLY
     IMPORTANT TO COASTAL AREAS.
                      * * *  DRAFT FINAL  * * *

-------
                                   15-5
CAUSES OF  SEA LEVEL RISE

    The worldwide average sea level depends primarily on (1) the shape and
size of ocean basins, (2) the amount of water in the oceans, and (3) the
average density of seawater.  Only the latter two factors are influenced by
climate.  Subsidence and emergence caused by natural factors such as
isostatic and tectonic adjustments of the land surface, as well as
human-induced factors such as oil and water extraction, can cause trends in
relative sea level at particular locations to differ from trends in global
sea level.

Past Trends in  Sea Level

    Hays and Pitman (1973) analyzed fossil records and concluded that over
the last 100 million years, changes in mid-ocean ridge systems have caused
the sea level to rise and fall over 300 m.  However, Clark, Farrell, and
Peltier (1978) have pointed out that these changes have accounted for sea
level changes of less than 1 mm per century.  No published study has
indicated that this determinant of sea level is likely to have a significant
impact in the next century.

    The impact of climate on sea level has been more pronounced.  During ice
ages the glaciation of substantial portions of the northern hemisphere has
removed enough water from the oceans to lower sea level 100 m below the
present level during the last (18,000 years ago) and earlier ice ages. (Bonn,
Farrand, and Ewing 1962; Kennett 1982; Oldale 1985).

    Although the glaciers that once covered much of the northern hemisphere
have retreated, the world's remaining ice cover contains enough water to
raise sea level over 75 m (Hollin and Barry 1979).  As Exhibit 15-1 shows,
Hollin and Barry (1979) and Flint (1971) estimate that existing alpine
glaciers contain enough water to raise sea level 30 or 60 cm, respectively.
The Greenland and West Antarctic Ice Sheets each contain enough water to
raise sea level about 8 m, and East Antarctica has enough ice to raise sea
level over 60 m.

    There is no evidence that either the Greenland or East Antarctic Ice
sheets have completely disintegrated in the last two million years.  However,
it is generally recognized that sea level was about 7 m higher than today
during the last interglacial (Moore 1982; Mercer 1968; Hollin 1962), which
was 1° to 2° warmer than today.  Because the West Antarctic Ice Sheet is
marine-based and thought by some to be vulnerable to climatic warming,
attention has focused on this source for the higher sea level.  Mercer (1968)
found that lake sediments and other evidence suggested that summer
temperatures in Antarctica have been 7° to 10°C higher than today at some
point in the last two million years, probably during the last interglacial
125,000 years ago, and that such temperatures could have caused a
disintegration of the West Antarctic Ice Sheet.  However, others are not
certain that marine based glaciers are more vulnerable to climate change than
                                 DRAFT FINAL  * * *

-------
                                   15-6



                               EXHIBIT 15-1

                          Snow and Ice  Components
                                                      Ice          Sea-Level
                                         Area         Volume       Equivalent a/
                                       (106  km2)     (10s km3)          (m)
Land ice:   East Antarctica b/             9.86        25.92           64.8
  West Antarctica c/                      2.34         3.40             8.5
  Greenland                               1.7          3.0              7.6
  Small ice caps and mountain             0.54         0.12             0.3
    glaciers d/ (Flint 1971,                                         0.6
    Hollin and Barry 1979)

Permafrost (excluding Antarctica):
  Continuous                              7.6          0.03             0.08
                                                       to            '  to
  Discontinuous                          17.3          0.7              0.17

Sea ice:  Arctic e/
            Late February                14.0          0.05
            Late August                   7.0          0.02

          Antarctic f/
            September                    18.4          0.06
            February                      3.6          0.01

Land snow cover g/
  N. Hemisphere
     Early February                      46.3          0.002
     Late August                          3.7

  S. Hemisphere
     Late July                            0.85
     Early May                            0.07
                                 DRAFT FINAL  * * *

-------
                                   15-7
                         EXHIBIT 15-1 (continued)

                          Snow and Ice Components
                                 Footnotes
    a/ 400,000 km3 of ice is equivalent to 1 m global sea level.

    b/ Grounded ice sheet, excluding peripheral,  floating ice shelves  (which
do not affect sea level).  The shelves have a total area of 1.62  x 10s
km2 and a volume of 0.79 x 10s km3 (Drewry and Heim 1983).

    c/ Including the Antarctic Peninsula.

    d/ Flint, Hollin, and Barry (1971).

    e/ Excluding the Sea of Okhotsk, the Baltic Sea, and the Gulf of St.
Lawrence (Walsh and Johnson 1979).  Maximum ice extents in these  areas  are
0.7 million, 0.4 million, and 0.2 million km2, respectively.

    f/ Actual ice area excluding open water (Zwally et al.  1983).   Ice
extent ranges between 4 million and 20 million km2.

    g/ Snow cover includes land ice but excludes  snow-covered sea ice  (Dewey
and Heim 1981).
Source:  Glaciers, Ice Sheets, and Sea Level.   National Academy Press,  p.
         242.  Modified from J.T. Hollin and R.G.  Barry, "Empirical and
         Theoretical Evidence Concerning the Response of the Earth's Ice and
         Snow Cover to a Global Temperature Increase," Environment
         International, 1979, 2:437-444.
                            * *  DRAFT FINAL  * *

-------
                                   15-8
land based glaciers;1 Robin (1985) suggests that the higher sea level
during the last interglacial period may have resulted from changes in the
East Antarctic Ice Sheet.

    Tidal gauges have been available to measure the change in relative sea
level at particular locations over the last century.2  Studies combining
these measurements to estimate global trends have concluded that sea level
has risen 10-15 cm during the last century  (Barnett 1984; Gornitz, Lebedeff,
and Hansen 1982; Fairbridge and Krebs 1962; Hicks et al. 1983 for U.S.
Coast);  Exhibit 15-2 shows the sea level curve estimated by Gornitz et al.
Barnett (1984) found that the rate of sea level rise over the last 50 years
had been about 2.0 mm/year, while in the previous 50 years there had been
little change; however, the acceleration in the rate of sea level rise was
not statistically significant.  Emery and Aubrey (1985 and n.d.) have
accounted for estimated land surface movements in their analyses of tidal
gauge records in Northern Europe and western North America and have found an
acceleration in the rate of sea level rise over the last century.3  Braatz
and Aubrey (n.d.) have found that the rate of relative sea level rise on the
east coast of North America accelerated after 1934.

    Several researchers have investiaged possible causes of current trends in
sea level.  Barnett (1984) and Gornitz, Lebedeff, and Hansen (1982) estimate
that thermal expansion of the upper layers of the oceans resulting from the
observed global warming of 0.4°C in the last century could be responsible for
a rise of 0.4 to 0.5 mm/year.  Recent analysis suggests that the warming has
been .6°C, which would account for thermal expansion of .6-.7 mm/year (Jones,
Wigley, and Wright 1986).  Roemmich and Wunsch (1984) examined temperature
and salinity measurements at Bermuda and concluded that the 4°C isotherm had
migrated 100 m downward, and that the resulting expansion of ocean water
could be responsible for some or all of the observed rise in relative sea
level.   Roemmich (1985) showed that the warming trend 700 m below the surface
was statistically significant.  However, Barnett (1984) found no significant
trend based on an examination of the upper layers of the ocean.  Neverthless,
Braatz and Aubrey (n.d.) note that long-term steric changes in the ocean are
not confined to the upper layers of the oceans, implying that the Barnett
analysis does not necessarily contradict the Roemmich and Wunsch conclusion.
Because one-dimensional models that can not capture the complexities of the
ocean have been used to estimate the past expansion of ocean water, these
authors all caution that their results should be viewed as first
approximations.

    Meier (1984) estimates that retreat of alpine glaciers and small ice caps
could currently contribute between 0.2 and 0.72 mm/year to sea level.  The
National Academy of Sciences Polar Research Board  (Meier et al. 1985)
concluded that existing information is insufficient to determine whether the
impacts of Greenland and Antarctica are positive or zero.  Thus, thermal
expansion and alpine melting would explain a rise of  .8 to 1.4 mm/yr,
compared with the observed rise of 1 to 1.5 mm/yr.  Although the estimated
global warming of the  last century appears at least partly responsible for
the last century's rise in sea level,  no study has demonstrated that global
warming might be responsible for an acceleration in the rate of sea level
rise.
                          * * *  DRAFT FINAL  * * *

-------
                                    15-9
                                EXHIBIT  15-2

                   Worldwide Sea Level in the Last Century
Sea  Level
   (cm)
            -5
             1880
1920
                                                     1980
                                       Year
  Sources:   V.  Gornitz, S. Lebedeff, and J. Hansen,  "Global Sea Level Trend in
            the Past Century," Science, 1982, p.  1611-1614.
                           * * *  DRAFT FINAL  »  *

-------
                                   15-10
Impact of Future Global Warming on Sea Level

    Concern about a substantial rise in sea level as a result of the
projected global warming stemmed originally from Mercer (1968),  who suggested
that the Ross and Filchner-Ronne ice shelves might disintegrate, causing a
deglaciation of the the West Antarctic Ice Sheet and a resulting 6- to 7-m
rise in sea level,  possibly over a period as short as 40 years (see also
Mercer 1978).

    Subsequent investigations have concluded that such a rapid rise is
unlikely.  Hughes (1983) and Bentley (1983) estimated that such a
disintegration would take at least 200 or 500 years, respectively.   Other
researchers have estimated that this process would take considerably longer
(Fastook 1985; Lingle 1985).

    Researchers have turned their attention to the magnitude of sea level
rise that might occur in the next century. The best-understood factors are
the thermal expansion of ocean water and the melting of alpine glaciers.  In
the National Academy of Sciences report "Changing Climate," Revelle (1983)
used the model of Cess and Goldenberg (1981) to estimate temperature
increases at various depths and latitudes resulting from a 4.2°C warming by
the years 2050-2060 (Exhibit 15-3).  Although his assumed time constant of 33
years probably resulted in a conservatively low estimate, he estimated that
thermal expansion would result in an expansion of the upper ocean sufficient
to raise sea level 30 cm.

    Using a model of the oceans developed by Lacis et al. (1981), Hoffman,
Wells and Titus (1986) examined a variety of possible scenarios of future
emissions of greenhouse gases and global warming.  They estimated that a
warming of between 1° and 2.6°C could result in a thermal expansion
contribution to sea level of between 12 and 26 cm by the year 2050.  They
also estimated that a global warming of 2.3 to 7.0°C by 2100 would result in
a thermal expansion of 28 to 83 cm.

    Revelle (1983)  suggested that although he could not estimate the future
contribution of alpine glaciers to sea level rise,  a contribution of 12 cm
through 2080 would be reasonable.  Meier  (1984) used glacier balance and
volume change data for 25 glaciers for which the available record exceeded 50
years to estimate the relationship between historic temperature increases and
the resulting negative mass balances of the glaciers.  He estimated that a
28-mm rise had resulted from a warming of 0.5°C and concluded that a 1.5° to
4.5°C  warming would result in a rise of 8 to 25 cm in the next century.
Using these results, the Meier et al. (1985) concluded that the contribution
of glaciers and small ice caps to a rise in sea level through 2100 is likely
to be 10 to 30 cm.   They noted that the gradual depletion of remaining ice
cover might reduce the contribution of sea level rise somewhat.  However, the
contribution might also be greater because the historic rise took place over
a sixty-year period, while the forecast period is over 100 years.  Using
Meier's estimated relationship between global warming and the alpine
contribution, Hoffman, Wells, and Titus (1986) estimated alpine contributions
through the year 2100 at 12 to 38 cm for a global warming of 2.3° to 7.0°C.
                          * * *  DRAFT FINAL  * * *

-------
                                     15-11
                                 EXHIBIT 15-3
                                   LATITUDE
      60°N      40°N     20°N
20°S     4(P3      60°S      80°S
Q.
lit
o
0



2OO







80O
mnn

7
5.




4

—

1.



3




.0
I

^^^M

8


r5"
J4.2




2.8

3

1.3

I
-5J
4.2




2.3



1.3


3.5
2.7




1.6



0.5


_2_
1.6




0.9



0.3


2.5
2.0




1.1



0.4

I
j 3.5
2.7




1.6



0.5


5
3.9




2.2



0.7


4.5
3.5




2.0
" 1



0.6


4
3.4




2.3



1.0


t
3




2



1.


»
.4




3 1



0


3.5
2.9




_2.0_



0.9


3 2
2.5 1.7




1.7 1.1



0.8 0.5

 Computed  average  increase  in  ocean  temperatures  at particular depths  and
 latitudes  for  a doubling of atmospheric  carbon dioxide  and probable increase
 in other  greenhouse  gases  by  the year  2080, based on Flohn's  (1982)
 prognosis.  Solid lines are isoplethis for temperature  changes.  Note that
 the top 100 m  warms  2°-7°, while at depths of 800-1000  m  the warming  is
 mostly less than  1°.

 Source:   Revelle  (1983).
                             * *  DRAFT FINAL  *

-------
                                   15-12
    The first published estimate of the contribution of Greenland meltwater
to future sea level rise was Revelle's (1983) estimate of 12 cm through the
year 2080.  Using estimates by Ambach (1980 and 1982) that the equilibrium
line (between snowfall accumulation and melting) rises 100 m for each 0.6°C
rise in air temperature, Revelle concluded that the projected 6°C warming
would be likely to raise the equilibrium line 1000 m.  He estimated that such
a change in the equilibrium line would result in a 12-cm contribution to sea
level rise for the next century.

    Meier et al. (1985) noted that the large ablation area makes Greenland a
"significant potential contributor of meltwater to the ocean if climatic
warming causes an increase in the rate of ablation and an upward shift of the
equilibrium line."   They found that a 1000-m rise in the equilibrium line
would result in a contribution of 30 cm through the year 2100.  However,
because Ambach (1985) found the relationship between the equilibrium line and
temperature to be 77 m per degree (C), the panel concluded that a 500 m shift
in the equilibrium line would be more likely.  Based on the assumption that
Greenland will warm 6.5°C by the year 2050 and that temperatures will remain
constant thereafter, they estimated that such a change would contribute about
10 cm to sea level through 2100, but also noted that "for an extreme but
highly unlikely case, with the equilibrium line raised 1000 m, the total rise
would be 26 centimeters."  Bindschadler (1985) treated the two cases as
equally plausible.  However, his analysis was conducted before the results of
Ambach (1985) were known and he has since indicated agreement with the
findings of Meier et al.
    Available estimates of the Greenland contribution assume that all
meltwater flows into the oceans and that the ice dynamics of the glaciers do
not change.  Meier et al. (1985) suggested that much of the water would
refreeze and therefore decrease the contribution to sea level rise.  Although
a change in ice dynamics might imply additional deglaciation and eventually
increase the rate of sea level rise, they concluded that such changes were
unlikely to occur in the next century.

    The potential impact of a global warming on Antarctica in the next
century is the least certain of all the factors by which a global warming
might contribute to sea level rise.  Meltwater from East Antarctica might
make a significant contribution by the year 2100.  This source could be
significant since global warming is expected to be amplified in polar
regions, and large parts of East Antarctica are far enough from the poles to
possibly warm up so that meltwater runoff is significant.. Unfortunately no
one has estimated the likely contribution.5  Several studies have examined
"deglaciation," which also includes the contribution of ice sliding into the
oceans.  Bentley (1983) examined the processes by which a deglaciation of
West Antarctica might occur.  The first step in the process would be
accelerated melting of the undersides of the Ross and Filchner-Ronne ice
shelves caused by warmer water circulating underneath them.  The thinning of
these ice shelves could cause them to become unpinned and cause their
grounding lines to retreat.  Revelle (1983) concluded that the available
literature suggests that the ice shelves might disappear in 100 years; the
Antarctic ice streams would then flow directly into the oceans, without the
                            * *  DRAFT FINAL  * * *

-------
                                   15-13
back pressure of the ice shelves.  Hughes (1983) and Bentley (1983) estimate
that a complete disintegration of the West Antarctic Ice Sheet could take an
additional 200 and 500 years, respectively.

    Although a West Antarctic deglaciation would occur over a period of
centuries, it is possible that an irreversible deglaciation could commence
before 2050.  Thomas, Sanderson, and Rose (1979) suggested that if the ice
shelves thinned more than about 1 m per year, the ice would move into the sea
at a sufficient speed such that even a cooling back to the temperatures of
today would not be sufficient to result in a reformation of the ice shelf.

    To estimate the likely Antarctic contribution for the next century,
Thomas (1985) developed four scenarios of the impact of a 3°C global warming
by the year 2050:

          (1)  For a shelf melting rate of 1 m/year with seaward ice
              fronts remaining at present locations — implies a rise
              of 28 cm by the year 2100.

          (2)  For a shelf melting rate of 1 m/year with ice fronts
              calving back to a line linking the areas where the
              shelf is grounded, during the 2050s--implies a rise of
              1.6 m by 2100.

          (3)  For. a scenario similar to cas.e 1 but with a melt rate
              of 3 m/year--implies a rise of 1 m by 2100.

          (4)  For a scenario similar to case 2 but with a melt rate
              of 3 m/year--implies a rise of 2.2 meters by 2100.

    Thomas concluded that the 28-cm rise implied by case 1 would be most
likely to occur.  He also stated that even if enhanced calving did occur, it
would be  likely to occur after 2050, "suggesting that probably associated
sea-level rise would be closer to the 1 m of case 3 than the 2.2 m of case
4."

    Meier et al. (1985) evaluated the Thomas study and papers by Lingle (1985)
and Fastook (1985).  Although Lingle estimated that the contribution to sea
level rise of West Antarctica through 2100 would be 3 to 5 cm, he did not
evaluate East Antarctica; Fastook made no estimate for the year 2100.  Thus,
the panel concluded that "imposing reasonable limits" on the Thomas model
yields a  range of 20 to 80 cm by the year 2100 for the Antarctic
contribution.  However, they also noted several factors that would reduce the
amount of ice discharged into the sea:  removal of the warmest ice from the
ice shelves, retreat of grounding lines, and increased lateral shear stress.
They also concluded that increased precipitation over Antarctica might
increase  the size of the polar ice sheets there.  Thus, they concluded that
the contribution from Antarctica could cause a rise in sea level up to 1 m or
a drop of 10 cm; a rise of between 0 and 30 cm is most likely.
                                 DRAFT FINAL  * * *

-------
                                   15-14
    Exhibit 15-4 summarizes the various estimates of future global sea level
rise for specific years.  Revelle (1983) estimated that the rise was likely to
be 70 cm, ignoring the impact of a global warming on Antarctica.  He also
noted that the latter contribution was likely to be 1 to 2 m per century after
2050, but declined to add that to his estimate.   Meier et al. (1985) projected
that the contribution of glaciers would be sufficient to raise sea level 20 to
160 cm, with a rise of "several tenths of a meter" most likely.   Thus, if one
extrapolates from the Revelle (1983) estimate of thermal expansion through the
year 2100, the Meier et al. report predicts a rise of between 50 and 200 cm.
Using a range of estimates for future concentrations of greenhouse gases, the
climate's sensitivity to such increases, oceanic heat uptake, and the behavior
of glaciers, Hoffman, Wells, and Titus (1986) estimated that the rise would be
10 to 21 cm by 2025 and 57 to 368 cm by 2100.

Future  Trends in Local Sea Level

    Although most attention has focussed on projections of global sea level,
impacts on particular areas would depend on local relative sea level.  As
Exhibit 15-5 shows, tidal guage measurements suggest that relative sea level
has risen ten to twenty centimeters more rapidly along much of the U.S. coast
than the worldwide average (Hicks, Debaugh, and Hickman 1983).  Important
exceptions include Louisiana, which is subsiding close to 1 m per century, and
Alaska, which is emerging 10 to 100 cm per century.

    Local subsidence and emergence are caused by a variety of factors.
Rebound from the retreat of glaciers after the last ice age has resulted in
the uplift of northern Canada, New England, and parts of Scandinavia;
emergence in Alaksa is due more to tectonic adjustments.  The uplift in polar
latitudes has resulted in subsidence in other areas, notably the U.S. Atlantic
and Gulf coasts.  Groundwater pumping has caused rapid subsidence around
Houston, Texas; Taipai, Taiwan; and Bangkok, Thailand, among other areas
(Leatherman 1984).  River deltas and other newly created land subside as the
unconsolidated materials compact.  Although subsidence and emergence trends
may change in the future, particularly where anthropogenic causes are
curtailed, no one has linked these causes to future climate change in the next
century.

    However, the removal of ice from Greenland and Antarctica would deform the
ocean floor.  Clark and Lingle have (1977) calculated the impact of a uniform
1 m contribution from West Antarctica.  They concluded that relative sea level
at Hawaii would rise 125 cm, and that along much of the U.S. Atlantic and Gulf
Coasts, the rise would be 15 cm.  On the other hand, sea level would drop at
Cape Horn by close to 10 cm, and the rise along the southern half of the
Argentine and Chilean coasts would be less than 75 cm.

    Other contributors to local sea level that might change as a result of a
global warming include currents, winds, and freshwater flow into estuaries.
None of these impacts, however, has been estimated.
                          * * *  DRAFT FINAL  * *

-------
                                   15-15
                               EXHIBIT 15-4

                     Estimates  of  Future Sea Level Rise
                               (centimeters)

Year 2100 by Cause  (Year 2085 for  Revelle 1983):
Thermal
Expansion
Revelle (1983)
Hoffman et al. (1983)
Meier et al. (1985) c/
Thomas (1985)
Hoffman et al. (1986)
30
28-115
28-83
Alpine
Glaciers Greenland Antarctica Total
12 12 2 70
b/ b/ b/ 56-345
10-30 10-30 -10-+100 50-200
0-200
12-37 6-27 12-220 57-368
Total Rise in Specific Years: d/

Hoffman et al. (1983)
Low
Mid-range low
Mid-range high
High
Hoffman et al. 1986
Low
High
2000
4.8
8.8
13.2
17.1

3.5
5.5
2025 2050 2075 2085 2100
13 23 38 - 56.0
26 53 91 - 144.4
39 79 137 - 216.6
55 117 212 - 345

10 20 36 44 . 57
21 55 191 258 368
       a/ Revelle attributes 16 cm to other factors.

       b/ Hoffman et al. (1983) assumed that the glacial contribution would
   be one to two times the contribution of thermal expansion.

       c/ NAS (1985) "estimate includes extrapolation of thermal expansion
   from Revelle (1983).

       d/ Only Hoffman et al. reports made year-by-year projections for the
   next century.

   Sources:  Hoffman et al. (1986); Meier et al. (1985);
             Hoffman et al. (1983): Revelle (1983); Thomas (1985).
                          * * *  DRAFT FINAL  * *

-------
                                   15-16



                               EXHIBIT 15-5

                           Local  Sea  Level  Rise
                                 Historic  a/
                                Relative Sea
                                Level  Trend
                                  (mm/yr)
 Historic b/
Subsidence
   Rate
  (mm/yr)
Local Rise for c/
1 Meter Global
 Rise by 2100
     (cm)
Portland, Maine
Boston, Massachusetts
Newport, Rhode Island
New London, Connecticut
New York, New York
Sandy Hook, New Jersey
Atlantic City, New Jersey
Philadelphia, Pennsylvania
Baltimore Maryland
Annapolis, Maryland
Hampton Roads, Virginia
Charleston, South Carolina
Fernandina, Florida
Miami Beach, Florida
Cedar Key, Florida
Pensacola, Florida
Eugene Island, Louisiana
Galveston, Texas
San Diego, California
Los Angeles, California
San Francisco, Califonia
Astoria, Oregon
Seattle, Washington
Juneau, Alaska
Sitke, Alaska
Worldwide
2.3
2.3
2.6
2.2
2.8
4.2
4.0
2.6
3.2
3.7
4.3
3.4
1.7
2.3
2.0
2.4
10.
6.3
1.9
0.6
1.2
-0.5
1.9
-12.9
-2.4
1.2
1.1
1.1
1.4
1.0
1.6
3.0
2.8
1.4
2.0
2.5
3.1
2.2
0.5
l:l
0.8
1.2
8.8
5.1'
0.7
-0.6
0
-1.7
0.7
-14.1
-3.6
0
112.6
112.6
116.1
111. "5
118.4
134.5
132.2
116.1
123.0
128.8
135.7
125.3
105.8 •
112.6
109.2
113.8
201.2
158.7
108.0
93.1
100.0
80.5
108.5
-62.2
58.6
100.0
    a/ Based on assumed global rise of 12 cm/century.   From Hicks,  Debaugh,
and Hickman (1983).

    b/ Relative sea level trend minus consensus estimate of 12 cm/century
for global sea level rise.

    c/ One meter plus extrapolation of historic subsidence rate.

Source:  Hicks, Debaugh, and Hickman 1983.
                          * * *  DRAFT FINAL  * *

-------
                                   15-17
EFFECTS OF SEA LEVEL RISE

    A rise in sea level of 1 or 2 m would permanently innundate wetlands  and
lowlands, accelerate coastal erosion, exacerbate coastal flooding,   threaten
coastal structures,  and increase the salinity of estuaries and aquifers.
Substantial research has been done on the implications of sea level rise  for
coastal erosion and wetlands and relatively little work has been done in  the
other areas.

    The remaining sections of this chapter summarize research on the impacts
of sea level rise.  The level of damage would depend in large measure on  when
and how people respond.  Much of the research has been conducted to assist
agencies that might have to respond to the impacts of sea level rise, and has
attempted to estimate impacts for three alternative responses: (1)  no
countermeasures are taken,  (2) measures are taken in response to sea level
rise as it occurs, and (3)  planning for sea level rise before it occurs.   We
follow this convention and examine potential impacts, possible responses, and
the relative merits of anticipation.

Submergence of Coastal Wetlands

    The most direct impact of a rise in sea level is the innundation of areas
that had been just above the water level before the sea rose. Coastal wetlands
are generally found at elevations below the highest tide of the year and  above
mean sea level.  Thus, wetlands account for most of the land less than 1  m
above sea level.

    Because a common means of estimating past sea level rise has been the
analysis of marsh peats, the impact of sea level rise on wetlands is fairly
well-understood.  For the last several thousand years, marshes have generally
kept pace with the rate of sea level rise through sedimentation and peat
formation (Emery and Uchupi 1972; Redfield 1972 and 1967; Davis 1985).  As sea
level rose, new wetlands formed inland while the seaward boundary was
maintained.  Because the wetland area has expanded, Titus, Henderson, and Teal
(1984) hypothesized that one would expect a concave marsh profile,  i.e.,  that
there is more marsh area than the area found immediately above the marsh.
Thus, if sea level rose more rapidly than the marsh's ability to keep pace,
there would be a net loss of wetlands. Moreover, a complete loss might occur
if protection of developed areas prevented the inland formation of new
wetlands.  .(See Exhibit 15-6a.)

    Kana et al. (1986 and n.d.) surveyed marsh transects in the areas of
Charleston, South Carolina and two sites near Long Beach Island, New Jersey,
to evaluate the concavity of wetland profiles and the vulnerability of
wetlands to a rise in sea level.  Their data showed that in the Charleston
area, all of the marsh was between 30 and 110 cm above current sea level, an
elevation range of  80 cm (see Exhibit 15-6b).  The area with a similar
elevation range just above the marsh was only 20% as large.  Thus,  a rise in
sea level exceeding vertical marsh accretion by 80 cm would result in an 80%
loss of wetlands.  In the New Jersey sites, the marsh was also found within an
elevation range of 80 cm; a rise in sea level 80 cm in excess of marsh
accretion would result in 67% to 90% losses.
                          .t. .J- .%.
                                 DRAFT FINAL  * * *

-------
                                                15-18




                                           EXHIBIT 15-6a


                              Evolution of  Marsh  as Sea Level  Rises
                5000 Vears Ago
                   Today
                                   -z— Sea Level
                                                  Sedimentation and
                                                  Peat Formation
                                    Current
                                   Sea Level

                                     Past
                                   Sea Level
                                                Future
Substantial Wetland Loss Where There is  Vacant Upland
                                           Future
                                         - Sea Level
                                         " Current
                                         Sea Level
Complete Wetland Loss Where House is Protected
        in Response to Rise in Sea  Level
                                   Future
                                 . Sea Level
                                   Current
                                  Sea Level
        Coastal marshes have kept pace with the slow rate of sea level rise that has characterized the last several thousand years.
        Thus, the area of marsh has expanded over time as new lands were Inundated. If in the future, sea level rises faster than
        the ability of the marsh to keep pace, the marsh area will contract. Construction of bulkheads to protect economic develop-
        ment may prevent new marsh from forming and result in a total loss of marsh in some areas.
        Coastal marshes have kept pace  with the  slow rate of sea  level rise that has
        characterized  the last  several  thousand  years.  Thus, the area of  marsh.has
        expanded over  time as new lands were inundated.   If in the future,  sea level
        rises  faster than the ability of the marsh to keep pace,  the marsh area will
        contract.  Construction of bulkheads to  protect economic  development may
        prevent new marsh from  forming  and result in a total loss of marsh in some
        areas.
                                             DRAFT FINAL  * *  *

-------
                                        15-19



                                    EXHIBIT 15-6b

                       Composite Transect -- Charleston, S.C.
                                   COMPOSITE TRANSECT-
                                     CHARLESTON, S.C.
    HIGHLAND 47%
                                                                             PCAK YEARLY TIDC

                                                                             SPRING HIGH WATER
                                                                             MEAN HIGH WATER
                                                                             NEAP HIGH WATER
-8
                        1000
                                    2000          3000
                                 TYPICAL DISTANCE (FT)
                                                             4OOO
                                                                          sooo
  Composite wetlands  transect for Charleston area.  Area above the marsh is much
  steeper than the marsh.

  Source:   Kana, Baca and William,  1986.
                              * * *  DRAFT FINAL  * * *

-------
                                   15-20
    The future ability of marshes to accrete vertically is uncertain.   Based
on field studies by Ward & Domeracki (1978), Hatton et al. (1983), Meyerson
(1972), Sterans & MacCreary (1957), and Kana et al. (1986) concluded that
current vertical accretion rates are approximately 4-6 mm/year in the two case
study areas, greater than the current rate of sea level rise but less than the
rates of rise projected for the next century.  If current accretion trends
continue, rises of 87 and 160 cm by 2075 would imply 50% and 80% losses of
wetlands, respectively, in the Charleston area.  Kana et al. (n.d.) also
estimated 80% losses in the New Jersey sites for a 160-cm rise through 2075.
However, because the high marsh dominates in that area, they concluded that
the principal impact of an 87-cm rise by 2075 would be the conversion of high
to low marsh.

    In both cases, the losses of marsh could be greater if inland areas are
developed and protected with bulkheads or levees.  Because there is a buffer
zone between developed areas and the marsh in South Carolina, protecting
development from a 160-cm rise would increase the loss from 80% to 90%.
Without the buffer, the loss would be close to 100%.

    The marshes and swamps of Louisiana, which account for 40% of the coastal
wetlands in the United States (excluding Alaska), would be particularly
vulnerable to an accelerated rise in sea level.  The wetlands in this state
are mostly less than 1 m above sea level, and are generally subsiding
approximately 1 m per century as its deltaic sediments compact (Boesch 1982).
Until the last century, the wetlands were able to keep pace with this rate of
relative sea level rise because of the sediment conveyed to the wetlands by
the Mississippi River.

    Human activities, however, have largely disabled the natural processes by
which coastal Louisiana might keep pace with sea level rise.  Dams, navigation
channels, canals, and flood protection levees have interrupted the flow of
sediment, freshwater, and nutrients to the wetlands.  As a result, over 100
square kilometers of wetlands convert to open water every year (Gagliano,
Meyer-Arendt, and Wicker 1981).  A substantial rise in sea level would further
accelerate the process of wetland loss in Louisiana (see Exhibit 15-7).

    Throughout the world, people have dammed, leveed, and channelized major
rivers, curtailing the amount of sediment that reaches river deltas.  Even at
today's rate of sea level rise, substantial amounts of land are converting to
open water in Egypt and Mexico (Milliman and Meade 1983).  Other deltas, such
as the Ganges in Bangladesh and India, are currently expanding seaward.  These
areas would require increased sediment, however, to keep pace"with an
accelerated rise in sea level.  Additional projects to divert the natural flow
of river water would increase the vulnerability of these areas to a rise in
sea level.

    Broadhus et al. (1986) examined the possible impacts of future sea level
rise on Egypt and Bangladesh, the inhabited areas of which are in the deltas
of the Nile and Ganges Rivers.  They estimated that 50 and 200 cm rises by
2100 would flood 0.35% and 0.7% of Egypt's land area, respectively.  However,
because the nation's population is concentrated in the low-lying areas, 16 and
                            * *  DRAFT FINAL  * * *

-------
                                    15-21



                               EXHIBIT 15-7

                    Louisiana  Shoreline in the Year 2030
 N
                  50
    I	1	1	1  i   i
          miles
                                   SOURCE:  COASTAL ENVIRONMENTS. INCORPORATED
                                           COASTAL ZONE WETLANDS
PREDICTED LOUISIANA COASTLINE
IN 50 YEARS AT PRESENT LAND LOSS RATES
                                        BATON ROUGE
                                                                      NEW ORLEANS
                             MORGAN CITY
                                      THIBODAUX*
        Gulf of Mexico
                                                  HOUMA
                                                               LOOP facility
The solid  line  shows  the predicted shoreline for  2030  if current trends
continue.  The  entire shaded area could be lostby 2085  if sea level rises 1
m.  However,  a  20  cm  rise by 2020 would result  in the  shore retreating to this
point by 2020 or sooner.

Source:  Coastal Environments, Incorporated.
                             * *  DRAFT FINAL  * * *

-------
                                   15-22
21 percent of the people currently reside in the areas that would be lost.
Broadhus et al. (1986) also estimate that Bangladesh could lose 12% and 28% of
its land, which currently houses 9% and 27% of its population (see Exhibit
15-8).

    To develop an understanding of the potential nationwide impact of sea
level rise on coastal wetlands in the United States, Park et al. (1986) used
topographic maps to characterize wetland elevations at 52 sites comprising
4800 square kilometers (1.2 million acres) of wetlands, over 17% of all U.S.
coastal wetlands.  Using published vertical accretion rates, they estimated
the impact of 1.4 m and 2.1 m rises in sea level through the year 2100 for
each of the sites.  Weighing their results according to the coastal wetland
inventory by Alexander, Broutman, and Field (1986), Titus (n.d.) estimated a
loss of wetlands between 47% and 82%, which could be reduced to between 31%
and 70% if new wetlands are not prevented from forming inland.

Inundation

    Although coastal wetlands are found at the lowest elevations, innundation
of low land could also be important in some areas, particularly if sea level
rises at least 1 m.  Unfortunately, the convention of 10-ft contours in the
mapping of most coastal areas has prevented a general assessment of land
loss.  Nevertheless, a few case studies have been conducted.

    -Kana et al. 1984 used data from aerial photographs to assess elevations in
the area around Charleston.  They concluded that 160 and 230 cm rises would
result in 30% and 46% losses of the area's dry land, respectively.  Leatherman
(1984) estimated that such rises would result in 9% and 12% losses of the land
in the area of Galveston and Texas City, Texas, assuming that the elaborate
network of seawalls and levees were maintained.

    The only nationwide assessment of the innundation from projected sea  level
rise was conducted by Schneider and Chen  (1980).  Unfortunately, the smallest
rise in sea level they considered was a 4.5-m (15-ft) rise, in part because
smaller contours are not generally available in topographic maps.
Nevertheless, their findings suggest which coastal states would be most
vulnerable:  Louisiana (which would lose 28% of its land and 51% of its
assets); Florida (24% and 52%); Delaware  (16% and 18%); Washington, D.C.  (15%
and 15%); Maryland (12% and 5%); and New Jersey (10% and 9%).

    As with wetland loss, the responses to innundation broadly fall into  the
categories of retreat and holding back the sea.  Levees are used extensively
in the Netherlands and New Orleans to prevent areas that are below sea level
from being flooded and could be similarly constructed around-other major
cities.  In lightly developed areas, however, the cost of a levee might be
greater than the value of the property being protected.  Moreover, even where
levees prove to be cost-effective, the environmental implications of replacing
natural shorelines with manmade structures would need to be considered.
                          * * *  DRAFT FINAL  * * *

-------
                                     15-23



                                  EXHIBIT 15-8

                   Distribution of Population in  Bangladesh
                                                    BANGLADESH
                                            DISTRIBUTION OF POPULATION
                                                      1971*
                                                 o  K  «o  so  eo  100
                                                       MilBI

                                                Each dot represents 2,000 persons
The high  scenario represents  2  to  2.5 m rise with sedimentation disrupted  by
human activities.  Low scenario represents 50 cm rise  and natural
sedimentation maintained.

Source:   Broadhus et al.  (1986).
                             *  *   DRAFT FINAL  * * *

-------
                                   15-24
Coastal Erosion

    Sea level rise can also result in the loss of land above sea level through
erosion.  Bruun (1962) showed that the erosion resulting from a rise in sea
level would depend upon the average slope of the entire beach profile
extending from the dunes out to the point where the water is too deep for
waves to have a significant impact on the bottom (generally a depth of about
10 m).  (See Exhibit 15-9):

                                 r = s * p/(h+d)

where r = shorelines retreat; s = sea level rise; p = horizontal length of
the beach profile from dune to the offshore limit of significant wave action;
h = height of the dune crest; and d = depth of the offshore limit of the
beach profile.  By comparison, innundation depends only on the slope
immediately above the original sea level.  Because beach profiles are
generally flatter than the the portion of the beach just above sea level, the
"Bruun Rule" generally implies that the erosion from a rise in sea level is
several times greater than the amount of land directly inundated.

    Processes other than sea level rise also contribute to erosion.  These
include storms, structures, currents, and alongshore transport.  Because sea
level has risen slowly in recent centuries, verification of the Bruun Rule on
the open coast has been difficult; it has been difficult to distinguish
erosion due to sea level rise from erosion caused by other processes.
However, water levels along the Great Lakes can fluctuate over 1 m in a
decade.  Hands (1976, 1979, and 1980) and Weishar and Wood (1983) have
demonstrated that the Bruun Rule generally predicts the erosion resulting
from rises in water levels there.

    The Bruun Rule has been applied to project erosion due to sea level rise
for several areas.  Bruun (1962) found that a 1 cm rise in sea level would
generally result in a 1 m shoreline retreat, but that the retreat'could be as
great as 10 m along some parts of the Florida coast.  Everts (1985) and Kyper
and Sorensen (1985) however, found that along the coasts of Ocean City,
Maryland and Sandy Hook, New Jersey, respectively, the shoreline retreat
implied by the Bruun Rule would be only about 75 cm.  Kana et al. (1984)
found that along the coast of South Carolina, the retreat could be 2 m.  The
U.S. Army Corps of Engineers (1979) indicated that along the coast of San
Franciso, where waves are generally larger than those found along the
Atlantic Coast, the shore might retreat 2 to 4 m for a 1-cm rise in sea level.

    Dean and Maurmeyer (1983) generalized the "Bruun Rule" approach to
consider the "overwash" of barrier islands.  Coastal geologists generally
believe that coastal barriers can maintain themselves in the face of slowly
rising sea level through the landward transport of sand, which washes over
the island during storms, building the island upward and landard.  Because
this formulation of the Bruun Rule extends the beach profile horizonatally. to
include the entire islands as well as the active surf zone, it always
predicts greater erosion than the Bruun Rule.  However, the formulation may
not be applicable to developed barrier islands, where the common practice of
public officials is to bulldoze sand back onto the beach after a major storm.


                          * * *  DRAFT FINAL  * * *

-------
                                    15-25



                                EXHIBIT  15-9

                               The  Bruun Rule
        Initial
        Condition
        Immediate
        Inundation When
        Sea Level Rises
        Subsequent
        Erosion Due to
        Sea Level Rise
                                                                 .... J
A rise  in  sea level of S  causes  immediate inundation.   However it would
eventually require the offshore  bottom to rise  by S.   The necessary  S  and A'
would be supplied from the  upper part of the beach A.   Total shoreline retreat
r is equal to S * p/(h+d).
                            * * *  DRAFT FINAL  * * *

-------
                                   15-26
    The potential erosion from a rise in sea level could be particularly
important to recreational beach resorts, which include some of the nation's
most economically valuable and intensely used land. Although nationwide
statistics are not available, information on particular locations is
instructive.  Every weekend in the summer, approximately 250,000 people visit
Ocean City, Maryland to sunbathe on its 15-kra (9-mile) beach.6  Along most
of the Ocean City shoreline, the beach is less than 15 m wide during high
tide--a width that is typical of the most extensively used beach resorts.
Relatively few of the most intensely developed resorts along the Atlantic
Coast have beaches wider than about 30 m at high tide.  Thus, the rise in
relative sea level of 30 cm projected in the next 40 to 50 years could erode
most recreational beaches in developed areas, unless additional erosion
response measures are taken.

Flooding and Storm Damage

    A rise in sea level could increase flooding and storm damages in coastal
areas for three reasons:  erosion caused by sea level rise would increase the
vulnerability of communities; higher water levels would provide storm surges
with a higher base to build upon; and higher water levels would decrease
natural and artificial drainage.

    The impact of erosion on vulnerability to storms is generally a major
consideration in proposed projects to control erosion, most of which have
historically been funded through the U.S. Army Corps of Engineers.  The
impact of sea level rise, however, has not generally been considered
separately from other causes of erosion.

    The impact of higher base water levels on flooding has. been investigated
for the areas around Charleston, South Carolina and Galveston, Texas (Earth
and Titus 1984).  Kana et al. (1984) found that around Charleston, the area
within the 10-year flood plain would increase from 33% in 1980, to 48%, 62%,
and 74% for rises in sea level of 88, 160, and 230 cm, respectively, and that
the area within the 100-year flood plain would increase from 63% to 76%, 84%,
and 90% for the three scenarios.9  Gibbs  (1984) estimated that a rise of
about 90 cm would double the average annual flood damages in the Charleston
area (but that flood losses would not increase substantially for higher rises
in sea level because shoreline retreat would result in a large part of the
community being completely abandoned).9

    Leatherman (1984) conducted a similar analysis of Galveston Island,
Texas.  He estimated that the area within the 100-year flood plain would
increase from 58% to 94% for an 88 cm rise in sea level,9 and that for a
rise greater than 1 m, the Galveston seawall would be overtopped during a
100-year storm.  Gibbs estimated that the damage from a 100-year storm would
be tripled for a rise of about 90 cm.9

    A wide variety of shore protection measures would be available for
communities to protect themselves from increased storm surge and wave damage
due to sea level rise (Sorensen, Weisman, and Lennon 1984).  Many of the
measures used to address erosion and inundation, including seawalls,
                              *  DRAFT FINAL  * * *

-------
                                   15-27
breakwaters, levees, and rebuilding beaches also provide protection against
storms.  In Galveston, which is already protected on the ocean side by the
seawall, Gibbs hypothesized that it might be necessary to completely encircle
the developed areas with a levee to prevent flooding from the bay side;
upgrading the existing seawall might also be necessary.

    Kyper and Sorensen (1985 and n.d.) examined the implications of sea level
rise for the design of coastal protection works at Sea Bright, New Jersey, a
coastal community that is currently protected by a seawall and has no beach.
Because the seawall is vulnerable to even a 10-year storm, the Corps of
Engineers and the State of New Jersey have been considering a possible
upgrade.  Kyper and Sorensen estimated that the cost of upgrading the seawall
for current conditions would be 3.5 to 6 million dollars per kilometer of
shoreline, noting that if designed properly, the seawall would be useful
throughout the next century.  However, they estimated that a rise in. relative
sea level of 30-40 cm would be likely to result in serious damage to the
seawall during a major storm because of higher water levels and the increased
wave heights resulting from the erosion of submerged sand in front of the
seawall.  To upgrade .the seawall to withstand a 1-m rise in relative sea
level would cost 5.7 to 9 million dollars per kilometer (50% more).  They
concluded that policy makers would have to weigh the tradeoff between the
cost of designing the wall to withstand projected sea level rise and the cost
of subsequent repairs and a second overhaul.

    In addition to community-wide engineering approaches, measures can also
be taken by individual property owners to prevent increased flooding.  In
1968, the U.S. Congress created the National Flood Insurance Program to
encourage communities to avoid risky construction in flood-prone areas.  In
return for requiring new construction to be elevated above expected flood
levels, the federal government provides flood insurance, which is not
available from the private sector.  If sea level rises, flood risks will
increase.  In response, local ordinances will automatically require new
construction to be further elevated and insurance rates on existing
properties will rise unless those properties are further elevated.  As
currently organized, the National Flood Insurance Program would react to sea
level rise as it occurred.  Various measures to enable the program to
anticpate sea level rise have been proposed, including warning policy holders
that rates may increase in the future if sea level rises, denying coverage to
new construction in areas that are expected to be lost to erosion within, the
next 30 years, and setting premiums according to the average risk expected
over the lifetime of the mortgage (Howard, Pilkey, and Kaufman 1985; Titus
1984).

    Case studies in Charleston, South Carolina and Fort Walton Beach, Florida
have examined the implications of sea level rise for rainwater flooding and
the design of coastal drainage systems.  Waddell and Blaylock (n.d.)
estimated that a 25-year rainstorm (with no storm surge) would result in no
damage for the Gap Creek watershed in Fort Walton Beach.   However, a rise in
sea level of 30 to 45 cm would result in damages of 1.1-1.3 million dollars
in this community of 4,000 residents during a 25-year storm.   An upgrade
costing $550,000, however, would prevent such damages.
                          * * *  DRAFT FINAL  * * *

-------
                                   15-28
    LaRoche and Webb (n.d.) had previously developed the master drainage plan
for Charleston, South Carolina and later evaluated the implications of sea
level rise for the Grove Street watershed in that community.   They estimated
that the cost of upgrading the system for current conditions  would be $4.8
million and the cost of upgrading the system for a 30-cra rise would be $5.1
million. If the system is designed for current conditions and sea level
rises, the system would be deficient and the city would face  a retrofit cost
of $2.4 million. Thus, for the additional $300,000 necessary  to upgrade for a
30-cm rise, the city could ensure that it would not have to spend an
additional $2.4 million later.  Noting that the decision whether to design
now for a rise in sea level depends on the probability that sea level would
rise, they concluded that a 3% real social discount rate would imply that
designing for sea level rise is worthwhile if the probability of a 30-cm rise
by 2025 is greater than 30%.  At a discount rate of 10%, they concluded,
designing for future conditions is not worthwhile.

Increased Salinity in Estuaries and Aquifers

    Although most researchers and the general public have focused on the
increased flooding and shoreline retreat associated with a rise in sea level,
the inland penetration of saltwater could be important in some areas.

    A rise in sea level increases the salinity of an estuary  by altering the
balance between freshwater and saltwater forces.  The salinity of an estuary
represents the outcome of (1) the tendency.for the ocean salt water to
completely mix with the estuarine water and (2) the tendency  of freshwater
flowing into the estuary to dilute the saline water and push  it back toward
the ocean.  During droughts, the salt water penetrates upstream; during the
rainy season, low salinity levels prevail.  A 'rise in sea level has an impact
similar to decreasing the freshwater inflow.  By widening and deepening the
estuary, sea level rise increases the ability of saltwater to penetrate
upstream.

    The implications of sea level rise for increased salinity have only been
examined in detail for Louisiana and the Delaware Estuary.  In Louisiana,
saltwater intrusion is currently resulting in the conversion  of cypress
swamps--which can not tolerate saltwater--to open water lakes, as well as the
conversion of fresh and intermediate marsh to marsh with higher salinity
levels.  In response to these trends, numerous projects have  been proposed to
divert freshwater from the Mississippi River to these wetlands (Roberts et
al. 1983; U.S. Army Corps of Engineers 1982; van Beek et al.  1982).  Although
the cause of the saltwater intrusion has been primarily the dredging of
canals and the sealing off of river distributaries that once  provided the
wetlands with fresh water, relative sea level rise is gradually increasing
the saltwater force in Louisiana's wetlands; a further rise in sea level
would accelerate this process.  Haydl (1984) also concluded that subsidence
may be resulting in increased salinity of drinking and irrigation water
supplies in some parts of Louisiana.
                            * *  DRAFT FINAL  » * *

-------
                                   15-29
    The impact of current sea level trends on salinity has been considered in
the long-range plan of the Delaware River Basin Commission since 1981 (DRBC
1981).  The drought of the 1960s resulted in salinity levels that almost
contaminated the water supply of Philadelphia and surrounding areas.  Hull
and Tortoriello (1979) found that the 13-cm rise projected between 1965 and
2000 would result in the "salt front" migrating 2 to 4 km farther upstream
during a similar drought. They found that a moderately sized reservoir (57
million cubic meters) to augment river flows would be needed to offset the
resulting salinity increases.

    Hull, Thatcher, and Tortoriello (1986) examined the potential impacts of
an accelerated rise in sea level due to the greenhouse warming.  They
estimated that 73-cm and 250-cm rises would result in the salt front
migrating an additional 15 and 40 kilometers, respectively, during a repeat
of the 1960s drought.  They also found that the health-based 50 ppm sodium
standard (equivalent to 73 ppm chloride) adopted by New Jersey would be
exceeded 15% and 50% of the time, respectively, and that the EPA drinking
water 250-ppm chloride standard would be exceeded over 35% of the time in the
high scenario (see Exhibit 15-10).

    Lennon, Wisniewski, and Yoshioka (1986) examined the implications of
increased estuarine salinity for the Potomac-Raritan-Magothy aquifer system,
which is recharged by the (currently fresh) Delaware River and serves the New
Jersey suburbs of Philadelphia.  During the 1960s drought, river water with
chloride concentrations as high as 150 ppm recharged these aquifers.  Lennon
et al. estimated that a repeat of the 1960s drought with a 73-cm rise in sea
level would result in river water with concentrations as high as 350 ppm
recharging the aquifer, and that during the worst month of the drought, over
one-half of the water recharging the aquifer system would have concentrations
greater than 250 ppm.  With a 250-cm rise, 98% of the recharge during the
worst month of the drought would exceed 250 ppm, and 75% of the recharge
would be greater than 1000 ppm.

    Hull and Titus (1986) examined the options by which various agencies
might respond to increased salinity in the Delaware estuary.  They concluded
that planned but unscheduled reservoirs would be more than enough to offset
the salinity increased from a 1-ft rise in sea level, although those
reservoirs had originally been intended to meet increased consumption.  They
noted that construction of the reservoirs would not be necessary until the
rise became more imminent.   However, they also suggested that, given the
uncertainties, it might be advisable today to identify additional reservoir
sites to ensure that future generations retained the option of building
additional reservoirs if necessary.

    A rise in sea level could increase salinities in other areas, although
the importance of those impacts has not been investigated.  Kana et al.
(1984) and Leatherman  (1984) made preliminary inquiries into the potential
impacts on coastal aquifers around Charleston and Galveston, respectively.
However, they concluded that in-depth assessments were not worthwhile because
the aquifers around Charleston are already salt-contaminated because of
overpumping, and pumping of groundwater has been prohibited in the Galveston
                                 DRAFT FINAL  * * *

-------
                                     15-30
                                  EXHIBIT 15-10


             Percent of Tidal Cycles  in Which  Specified Concentration

               is Exceeded at Torresdale During a  Recurrence of the

                   1960's Drought for Three Sea Level Scenarios
Q
UJ

§100

o
X
2  80
Z  70-j
UJ
O


§  60



O

X  50-
w 40 H
ui


> 30
O
<
O
UJ
o

-------
                                   15-31
area due to the resulting land subsidence.  The potential impacts on Florida's
Everglades and the shallow aquifers around Miami might be significant, but
they have not been investigated.

Economic  Significance of Sea Level Rise

    No one has estimated a dollar value of the likely impacts of sea level
rise for the nation or any coastal state.  Schneider and Chen (1980)
estimated the economic impact of what was once (but is no longer) thought to
be a plausible scenario:  rises of 4.6 to 7.6 m (15 to 25 ft) occurring with
little or no warning during the early part of the twenty-first century.  They
estimated that this scenario would result in real property losses of 100 to
150 billion dollars, representing 6.2% to 8.4% of all real property in the
nation.

    The only comprehensive attempt to place a dollar value on the impacts of
sea level rise for particular communties was the study by Gibbs (1984) of the
Charleston and Galveston areas.  In addition to considering more realistic
scenarios ranging from 0.9 to 2.4 m through 2075, Gibbs made several
improvements on the approach of Schneider and Chen, differentiating between
the economic impact if sea level rise was expected and the impact resulting
from reacting to it as it occurs; modeling how investment decisions might
respond to floods and erosion; and explicitly considering community-wide
strategies to limit losses, including shore protection and abandonment.

    In the Charleston study,  Gibbs evaluated the implications of (1) efforts
to avoid development of some vacant suburban areas likely to be flooded in
the future; (2) a partial abandonment; and (3) elevating the existing
seawalls protecting Charleston to provide additional protection.  For a rise
of 28-64 cm through 2025, Gibbs estimated that present value of the
cumulative impact would be $280-1065 million (5% to 19% of economic activity
in the area for the period), which could be reduced to $160-420 million if
sea level rise was anticipated.  Most of this impact would result from a 10%
to 100% increase in expected storm damages, although Gibbs also estimated
losses to erosion at $7-35 million.9  For the period between 1980 and 2075,
Gibbs estimated that the economic impacts would be $1250-2510 million (17% to
35%) and could be reduced $440-1100 million through anticipatory measures.
If accurate, these estimates suggest that Charleston would be extremely
vulnerable to a rise in sea level.  However, these estimates were necessarily
based on a number of simplifying assumptions regarding storm damage and
current elevation of structures.

    Gibbs performed a similar analysis of the Galveston area and concluded
that the impacts of sea level rise through 2025 would represent $115-360
million (l.l%-3.6%) if not anticipated, and $90-140 million if anticipated.
The impacts through 2075 were estimated at $555-1840 million (4.9%-16%) if
not anticipated, and $310-730 million if anticipated.  The assumptions Gibbs
used suggest that a 100-year storm today would result in damages equal to
about 7% of the property in the area, and that average annual storm damages
are equal to $6 million, slightly less than 0.2% of the value of property in
the study area.
                              *  DRAFT FINAL  * *

-------
                                   15-32
    Other studies can be used to gain an understanding of the economic
significance of particular classes of impacts.   As discussed above,  a 30-cm
rise in sea level would erode most recreational beaches back to the first row
of houses.  The studies cited in the section on erosion indicate that the
typical beach profile extends out about 1000 m, which suggests that 300,000
cubic yards of sand per kilometer of shoreline are required to raise the
beach profile 30 cm.  If sand costs are typically $3-10/cubic meter, the
beach rebuilding costs of a 30 cm rise in sea level would be $1 to 3 million
per kilometer.  -If the United States has a few thousand kilometers of
recreational beaches, it would cost billions and perhaps tens of billions of
dollars to rebuild the beach in response to a 30-cm rise in sea level.  This
estimate considers only the beaches themselves; raising people's lots to
avoid inundation would further increase the costs.

    The U.S. Army Corps of Engineers (1971) estimated that in 1971,  25,000 km
of shoreline (exclusive of Alaska, Great Lakes, and Hawaii) were eroding, of
which 17% were "critically eroding," and would require engineering
solutions.  If 17% of all shorelines require erosion control, that would
imply protection of close to 10,000 km of shoreline.  Sorensen (1986)
describes dozens of engineering options for preventing erosion, the least
expensive of which costs $300,000 per kilometer, which would imply a cost of
at least three billion dollars for protecting all shorelines.

Relationship to Other Impacts of the Greenhouse Warming

    The impacts of sea level rise on coastal areas, as well as their
importance, is likely to depend in part on other impacts of the greenhouse
warming.  Although future sea level is uncertain, there is a general
consensus that a global warming would cause sea level to rise; by contrast,
the direction of most hydrological and storm event changes at particular
locations is unknown.

    Warmer temperatures might change the ability of wetlands to keep pace
with sea level rise.  Mangrove swamps would replace salt marshes in some
areas that are currently too cold for mangroves.  No one has assessed the
impact of this vegetation change on vertical accretion.  Peat formation is
generally greater in New England marshes than at lower latitudes.   Warmer
temperatures might reduce that rate of vertical accretion for these wetlands.

    Changing climate could alter the frequency and tracks of storms.  Because
hurricane formation requires water temperatures of 27°C or higher (Wendland
1977), a global warming might result in an extension of the hurricane season
and in the formation of hurricanes at higher latitudes.  Besides increasing
the amount of storm damage, increased frequency of severe storms would tend
to flatten the typical beach profile, causing substantial shoreline retreat
unless additional sand was placed on the beach.  A decreased frequency of
severe winter storms might have the opposite impact at higher latitudes.

    Changes in precipitation could also affect the impacts of sea level
rise.  Because warmer temperatures would intensify the hydrologic cycle, it
is generally recognized that a global warming would increase rainfall
                              *  DRAFT FINAL  * * *

-------
                                   15-33
worldwide.   If rainfall in maritime environments increases, rainwater
flooding could be increased because of decreased drainage and increased
precipitation.  The impact of sea level rise on saltwater intrusion could be
offset by decreased drought frequency or exacerbated by increased drought
frequency (Rind and"Lebedeff 1984).

CONCLUSION

    Researchers have identified a wide variety of impacts that could result
from a 50 to 200 cm rise in sea level.  These impacts include both economic
losses resulting from land loss, flooding, shore protection, and increased
costs for drinking water, as well as environmental costs such as the
disruption of coastal ecosystems resulting from wetland loss and increased
salinities.  Although there are substantial uncertainties regarding the
magnitude of both the causes and effects of future sea level rise, even
conservative estimates suggest that important impacts are likely to result.

    A high priority for future research should be the use of existing case
studies to derive estimates of the nationwide and worldwide magnitude of the
various impacts that have been identified.  Although preliminary estimates of
the potential loss of U.S. coastal wetlands exist, the implications of those
losses have not been assessed.  For most other impacts, no natioinwide
estimates are available.  That type of information will be important for
evaluating the benefits of possible actions to mitigate the greenhouse
warming and resulting rise in sea level.
                          * * *  DRAFT FINAL  * *

-------
                                   15-34
                                   NOTES
1.   Mark Meier, University of Colorado (personal communication).

2.   The only true measure of absolute sea level would be the rise of sea
    level relative to the center of the earth.   Unfortunately,  no such
    measurements are yet available.  Therefore, researchers have  had to
    combine tidal guage measurements of relative sea level trends  at various
    locations, filter out known movements of the land surface,  and take
    weighted averages to arrive at estimates of global sea level  trends.

3.   This result was reported in the North America study.  The data also show
    this to be true in the Northern Europe study, but the result  was not
    reported. David Aubrey, Woods Hole Oceanographic Institute, Woods Hole,
    Massachusetts (personal communication).

4.   Robert Bindschadler, Goddard Space Flight Center, Greenbelt.  Maryland
    (personal communication).

5.   James Hansen, Goddard Insitute for Space Studies, New York (personal
    communication).

6.   Sandy Coyman, Department of Planning and Community Development, Town of
    Ocean City, Maryland (personal communication).

7.   North Carolina Administrative Code, Chapter 7H, 1983.  Raleigh, North
    Carolina:  Office of Coastal Management.

8.   Fred Michaud, Office of Floodplain Management,  State of Maine (personal
    communication).

9.   These results are reported in an Appendix by Titus et al. 1984

10. See Note 9.
                          * * *  DRAFT FINAL  * * *

-------
                                       15-35
                                      APPENDIX A
    The sea level rise damages reported for Charleston and Galveston in Chapter 18
were derived from the following data taken from Gibbs (1984).
                 Global           Charleston Impact a/           Galveston Impact a/
               Sea Level Rise     With        Without            With        Without
   Scenario    by 2075 (cm)    Anticipation Anticipation     Anticipation  Anticipation
Trend
Low
Medium
High
11.4
75.2
146.8
219.2
b/
440
730
1100
b/
1250
1910
2510
b/
310
415
730
b/
' 555
965
1890
a/  Through 2075 in bilions of dollars discounted at 3%.

b/  All estimates are relative to the trend scenario.
The damage estimates consist of a low estimate and a high estimate.  The low estimate
is derived by linearly interpolating the "with anticipation" estimates for Charleston
and Galveston according to the amount of global sea level rise.  The high estimates
are derived by linearly interpolating the "without anticipation" estimates.
                              * * »  DRAFT FINAL  * *

-------
                                   15-36
                                REFERENCES
Alexander, C.E., M.A. Broutman, and D.W.  Field.   1986.   An Inventory of
    Coastal Wetlands of the USA.   Rockville,  MD.:   National Oceanic and
    Atmospheric Administration, National  Ocean Service.

Ambach, W. 1985.  "Climatic Shift of the  Equilibrium Line--Kuhn's Concept
    Applied to the Greenland Ice Cap." Annals of  Glaciology  6:76-78.

Ambach, W.  (translated by G.P. Weidhaas).   1980.   "increased CO

    Concentration in the Atmosphere and Climate Change:   Potential Effects on
    the Greenland Ice Sheet."  Wetter und Leben 32:135-142, Vienna.
    (Available as Lawrence Livermore National Laboratory Report
    UCRL-TRANS-11767, April 1982.)

Associated Press.  1985.  "Doubled Erosion Seen for Ocean City."  Washington
    Post, November 14th.  (Maryland Section).

Barnett, T.P.   1984.  "The Estimation of  'Global'  Sea Level Change:  A
    Problem of Uniqueness."  Journal of Geophysical Research
    89(C5):7980-7988.

Earth, M.C. and J.G. Titus (eds.)  1984.   Greenhouse Effect and Sea Level
    .Rise:  A Challenge for this Generation.  New York:   Van Nostrand Reinhold.

Bentley, C.R.   1983.  "West Antarctic Ice Sheet:  Diagnosis and Prognosis."
    In Proceedings:  Carbon Dioxide Research Conference:  Carbon Dioxide,
    Science, and Consensus.  Conference 820970.   Washington, D.C.:
    Department of Energy.

Bindschadler,  R.A.  1985.  "Contribution  of the Greenland Ice Cap to Changing
    Sea Level:  Present and Future."  In  Meier,  M.F. et al. 1985.  op.  cit.

Boesch, D.F. (ed.)  1982.  Proceedings of the Conference of Coastal Erosion
    and Wetland Modification in Louisiana:   Causes, Consequences, and
    Options.  FWS-OBS-82/59.  Slidell, Louisiana:   National Coastal
    Ecosystems Team, U.S. Fish and Wildlife Service.

Braatz, -R.V. and D.G. Aubrey, n.d.  "Recent Relative Sea-Level Movement Along
    the Eastern Coastline of North America."  Journal of Sedimentary
    Petrology, in press.

Broadhus, J. et al.  1986.  "Sea Level Rise and the Damming of Rivers."  In
    Titus, J.G. (ed.)  Effects of Changes in Stratospheric Ozone and Global
    Climate, UNEP and EPA.  Washington, D.C.

Bruun, P. 1962.  "Sea Level Rise as a Cause of Shore Erosion."  Journal of
    Waterways and Harbors Division (ASCE) 1:116-130.

Cess, R.D., and S.D. Goldenberg.  1981.  The effect of ocean heat capacity
    upon global warming due to increasing atmospheric carbon dioxide.  J_._
    Geophys. Res. 86:498-502.

                          * * »  DRAFT FINAL  * * *

-------
                                   15-37
Clark, J.A. and C.S. Lingle.  1977.  "Future Sea-Level Changes Due to West
    Antarctic Ice Sheet Fluctuations."  Nature 269(5625)206-209.

Clark, J.A., W.E. Farrell, and W.R. Peltier.  1978.   "Global Changes in
    Postglacial Sea Level:  A Numerical Calculation."  Quarternary Research
    9:265-87.

Davis, R.A.  1985.  Coastal Sedimentary Environments.  New York:  Springer
    Verlag.

Dean, R.G. and E.M. Maurmeyer.  1983.  "Models for Beach Profile Response."
    In Coastal Engineering Research Center, Handbook of Coastal Processes and
    Erosion.  Fort Belvoir:  Army Corps of Engineers.

Delaware River Basin Commission (DRBC).  1981.  The Delaware River Basin
    Comprehensive (Level B) Study:  Final Report and Environmental Impact
    Statement.  West Trenton, New Jersey:  Delware River Basin Commission.

Dewey, K.F., and R.J. Heim.  1981.  Satellite observations of variations in
    northern hemisphere seasonal snow cover.  NOAA Technical Report, NESS 87,
    Washington, D.C.

Donn, W.L., W.R. Farrand, and M. Ewing.  1962.  "Pleistocene Ice Volumes and
    Sea-Level Lowering."  Journal of Geology 70:206-214.

Edmonson, J. and R. Jones.  1985.  Terrebonne Parish Barrier Island and Marsh
    Management Program.  Houma, La.:  Terrebonne Parish Council.  (August).

Emery, K.O. and D.G. Aubrey,  n.d.  "Relative Sea Level Change from Tide
    Gauge Records of Western North America."  Journal of Geophysical
    Research, in press.

Emery, K.O. and D.G. Aubrey.  1985.  "Glacial Rebound and Relative Sea Levels
    in Europe From Tide-Gauge Records."  Tectonophysics 120:239-255.

Emery, K.O. and E. Uchupi.  1972.  Western North Atlantic Ocean Memoir 17.
    Tulsa, Oklahoma:  American Association of Petroleum Geologists.

Everts, C.H.  1985.  "Effect of Sea Level Rise and Net Sand Volume Change on
    Shoreline Position at Ocean City, Maryland."  In Potential Impacts of Sea
    Level Rise on the Beach at Ocean City, Maryland.  Washington, D.C.:
    Environmental Protection Agency.

Fairbridge, R.W. and W.S. Krebs, Jr.  1962.  "Sea Level and the Southern
    Oscillation."  Geophysical Journal 6:532-545.

Fastook, J.L.  1985.  "ice Shelves and Ice Streams:   Three Modeling
    Experiments."  In:  Meier et al. 1985.  op. cit.

Flint, R.F.  1971.  Glacial and Quarternary Geology.  New York:  John Wiley
    and Sons.
                          - * *  DRAFT FINAL  * *

-------
                                   15-38
Flohn, H.  1982.  Climate change and an ice-free Arctic Ocean.   In Carbon
    Dioxide Review:  1982, W.C. Clark, ed.  Oxford U. Press, New York,  pp.
    145-199.

Gagliano, S.M., K.J. Meyer-Arendt, and K.M. Wicker.  1981.   "Land Loss  in the
    Mississippi Deltaic Plain."  In Transactions of the 31st Annual Meeting
    of the Gulf Coast Association of Geological Societies.   Corpus Christi,
    Texas.  Pp. 293-300.

Gibbs, M.  "Economic Analysis of Sea Level Rise:  Methods and Results."  In
    Barth and Titus (eds.) 1984.  op. cit.

Gornitz, V.,  S. Lebedeff, and J. Hansen.  1982.  "Global Sea Level Trend in
    the Past  Century."  Science 215:1611-1614.

Hands, E.G.  1980.  Prediction of Shore Retreat and Nearshore Profile
    Adjustments to Rising Water Levels on the Great Lakes.   U.S. Army Corps
    of Engineers Tech. Paper 80-7.

Hands, E.G.  1979.  Changes in Rates of Offshore Retreat:  Lake Michigan.
    Fort Belovoir:  U.S. Army Corps of Engineers Coastal Engineering Research
    Center Publication TP79-4.

Hands, E.G.  1976.  Observations of Barred Coastal Profiles Under Influence
    of Rising Water Levels:  Eastern Lake Michigan.  Fort Belvoir:  U.S. Army
    Corps of  Engineers CERC Publication TR76-1.

Haydl, N.C.  1984.  "Louisiana Coastal Area, Louisiana:  Water Supply,
    Initial Evaluation Study."  New Orleans:  U.S. Army Corps of Engineers.

Hays, J.D. and W.C. Pitman III.  1973.  "Lithospheric Plate Motion, Sea Level
    Changes,  and Climatic and Ecological Consequences."  Nature 246:18-22.

Hicks, S.D.,  H.A. Debaugh, and L.H. Hickman.  1983.  Sea Level Variations for
    the United States 1855-1980.  Rockville, Maryland:   U.S. Department of
    Commerce, NOAA-NOS.

Hoffman, J.S., D. Keyes, and J.G. Titus.  1983.  Projecting Future Sea  Level
    Rise.  Washington, D.C.:   Government Printing Office.

Hoffman, J.S., J. Wells, and J.G. Titus.  1986.  "Future Global Warming and
    Sea Level Rise."  In Bruun, P. (ed.) Iceland Symposium  '85  Rejivak:
    National  Energy Authority.

Hollin, J.T., 1962.  On the glacial history of Antarctica.   Journal of
    Glaciology, 4, 273-295.

Hollin, J.T.  and R.G. Barry.   1979.  "Empirical and Theoretical Evidence
    Concerning the Response of the Earth's Ice and Snow Cover to a Global
    Temperature Increase."  Environment International 2:437-444.

Howard, J.D., O.H. Pilkey, and A. Kaufman.  1985.  "Strategy for Beach
    Preservation Proposed."  Geotimes 30:12:15-19.

                          * * *  DRAFT FINAL  * * *

-------
                                   15-39
Hughes, T. 1983.  "The Stability of the West Antarctic Ice Sheet:   What Has
    Happened and What Will Happen."  In Proceedings:   Carbon Dioxide Research
    Conference:  Carbon Dioxide, Science, and Consensus.   Conference 820970.
    Washington, D.C.:  Department of Energy.

Hull, C.H.J. and J.G. Titus (eds.)  1986.  Greenhouse Effect, Sea Level Rise,
    and Salinity in the Delaware Estuary.  Washington, D.C.:  Environmental
    Protection Agency and Delaware River Basin Commission.

Hull, C.H.J. and J.G. Titus.  1986.  "Responses to Salinity Increases."  In
    Hull and Titus (eds.) 1986.   op. cit.

Hull, C.H.J., M.L. Thatcher, and R.C. Tortoriello.  1986.   "Salinity in the
    Delware Estuary."  In Hull and Titus (eds.) 1986.  op. cit.

Hull, C.H.J. and R.C. Tortoriello.  1979.  "Sea Level Trend and Salinity in
    the Delaware Estuary."  Staff Report.  West Trenton,  N.J.:   Delaware
    River Basin Commission.

Jones, P.O., T.M.L. Wigley, and P.B. Wright.  1986.  "Global temperature
    variations between 1861 and 1984."  Nature 322:430-434.

Kana, T.W., J. Michel, M.O. Hayes, and J.R. Jensen.  1984.  "The Physical
    Impact of Sea Level Rise in the Area of Charleston, South Carolina."  In
    Barth and Titus  (eds.) 1984.  op. cit.

Kana, T.W., B.J. Baca, and M.L.  Williams.  1986.  Potential Impacts of Sea
    Level Rise on Wetlands Around Charleston, South Carolina.  Washington,
    D.C.:  Environmental Protection Agency.

Kana, T.W., W.C. Elser, B.J. Baca, and M.L. Williams,  n.d.  "Potential
    Impacts of Sea Level Rise on Wetlands Around Little Egg Harbor, New
    Jersey."  In Impacts of Sea Level Rise on Coastal Wetlands in the United
    States.  Washington, D.C.:  Environmental Protection Agency.  (In press)

Kennett, J.  1982.  Marine Geology.  Englewood Cliffs, New Jersey:
    Prentice-Hall.

Kuo, C.  n.d.  Potential Impact of Sea Level Rise on Coastal Drainage
    Systems.  Washington, B.C.:   Environmental Protection Agency.   (In press)

Kyper, T. and R. Sorensen.  1985.  "Potential Impacts of "Selected Sea Level
    Rise Scenarios on the Beach and Coastal Works at Sea Bright, New Jersey.
    In Magoon, O.T., et al. (eds.) Coastal Zone '85.   New York:   American
    Society of Civil Engineers.

Kyper, T. and R. Sorensen.  n.d.  "Potential Impacts of Sea Level Rise on the
    Beach and Coastal Structures at Sea Bright, New Jersey."   Washington,
    D.C.:  Environmental Protection Agency and Army Corps of Engineers (in
    press).

Lacis, A. et al.  1981.  "Greenhouse Effect of Trace Gases, 1970-1980."
    Geophysical Research Letters 81:10:1035-1038.

                          * * *  DRAFT FINAL  * * *

-------
                                   15-40
LaRoche, T.B. and M.K. Webb.  n.d.  In Kuo, C., (ed.) Potential Impacts of
    Sea Level Rise on Coastal Drainage Systems.  Washington,  D.C.:
    Environmental Protection Agency (in press).

Leatherman, S.P.  1984.  "Coastal Geomorphic Responses to Sea Level Rise:
    Galveston Bay, Texas."  In Earth and Titus (eds.) 1984.   op.  cit.

Lennon, G.P., G.M. Wisniewski, and G.A. Yoshioka.   1986.   "impact of
    Increased River Salinity On New Jersey Aquifers."  In Hull and Titus
    (eds.) 1986.  op. cit.

Lingle, C.S.   1985.  "A Model of a Polar Ice Stream and Future Sea-Level Rise
    Due to Possible Drastic of the West Antarctic Ice Sheet."  In Meier et
    al. 1985.  op. cit.

Meier, M.F.  1984.  "Contribution of Small Glaciers to Global Sea Level."
    Science 226:1418-1421.

Meier, M.F. et al.  1985.  Glaciers, Ice Sheets,  and Sea Level.  Washington,
    D.C.:   National Academy Press.

Mercer, J.H.   1978.  "West Antarctic Ice Sheet and CCL Greenhouse Effect:

    A Threat of Disaster?"  Nature 271:  321-325.

Mercer, J.H.   1968.  "Antarctic Ice and Sangamon Sea Level."  Geological
    Society of America Bulletin 79:471.

Milliman,  J.D. and R.H. Meade.  1983.   Worldwide Delivery of River Sediment
    to the Oceans.  Journal of Geology 91(1):1-21.

Moore, W.S.  1982.  Late Pleistocene Sea-Level History.  In Uranium Series
    Disequilibrium:  Applications to Environmental Problems,  M. Ivanovich
    and R.S.  Harmon, (eds.)  Clarendon, Oxford, p. 41.

Oldale, R.  1985.  "Late Quarternary Sea Level History of New England:  A
    Review of Published Sea Level Data."  Northeastern Geology 7:192-200.

Park, R.A., T.R. Armentan, and C.L. Cloonan.  "Predicting the Effects of Sea
    Level Rise on Coastal Wetlands."  Effects of Changes in Stratospheric
    Ozone and Global Climate.

Redfield,  A.C.  1972.  "Development of a New England Salt Marsh."  Ecological
    Monograph 42:201-237.

Redfield,  A.C.  1967.  "Postglacial Change in Sea Level in the Western North
    Atlantic Ocean."  Science 157:687-692.

Revelle, R.  1983.  "Probable Future Changes in Sea Level Resulting from
    Increased Atmospheric Carbon Dioxide."  In Changing Climate.   Washington,
    D.C.:   National Academy Press.
                          * * *  DRAFT FINAL  * * *

-------
                                   15-41
Rind, D. and S. Lebedeff.  1984.  Potential Climate Impacts of Increasing
    Atmospheric C0_ with Emphasis, on Water Availability and Hydrology in

    the United States.  Washington, D.C.:  Government Printing Office.

Roberts, D., D. Davis, K.L. Meyer-Arendt, and K.M.  Wicker.   1983.
    Recommendations for Freshwater Diversion to Barataria Basin,  Louisiana.
    Baton Rouge:  Coastal Management Section, Louisiana Department of Natural
    Resources.

Robin, G.  1985.  "Contrast in Core:  Changes in Climate or Ice Volumes?"
    Nature 316:578-589.

Roemmich, D.  1985.  "Sea Level and Thermal Variability of the Ocean."  In
    "Glaciers,  Ice Sheets, and Sea Level."  Washington, D.C.:   National
    Academy Press.

Roemmich, D. and C. Wunsch.  1984.  "Apparent Changes in the Climatic State
    of the Deep North Atlantic Ocean."  Nature 307:447-450.

Schneider, S.H. and R.S. Chen.  1980.  "Carbon Dioxide Flooding:   Physical
    Factors and Climatic Impact."  Annual Review of Energy 5:107-140.

Sorensen, R.M., R.N. Weisman, and G.P. Lennon.  1984.  "Control of Erosion,
    Inundation, and Salinity Intrusion."  In Barth and Titus (eds.) 1984.
    op. cit.

Thomas, R.H.  1985.  "Responses of the Polar Ice Sheets to Climatic
    Warming."  In Meier et al. 1985.  op. cit.

Thomas, R.H., T.J.O. Sanderson, and K.E. Rose.  1979.  "Effect of Climatic
    Warming on the West Antarctic Ice Sheet."  Nature 227:355-358.

Titus,  J.G.  1984.  "Planning for Sea Level Rise Before and After a
    Coastal Disaster."  In:  Barth and Titus (eds.) 1984.  op. cit.

Titus, J.G.  1985.  "Sea Level Rise and the Maryland Coast."  In
    Potential Impacts of Sea Level Rise on the Beach at Ocean City,
    Maryland.  Washington, D.C.:  Environmental Protection Agency.

Titus, J.G.  n.d.  "Greenhouse Effect, Sea Level Rise, and Society's
    Reponse."  In  Devoy, R.J.  Sea Surface Studies  Cork, Ireland:
    University College.

Titus, J.G.  n.d.  "Sea Level Rise and Coastal Wetland Loss."  In EPA:
    Impact of Sea Level Rise on Coastal Wetlands (in press).

Titus, J.G., T. Henderson, and J.M. Teal.  1984.  "Sea Level Rise and
    Wetlands Loss in the United States."  National Wetlands Newsletter 6:4.

U.S. Army Corps of Engineers.  1982.  Louisiana Coastal Area, Louisiana:
    Freshwater Diversion to Barataria and Breton Sound Basins.  New Orleans:
    Army Corps of Engineers.


                          * * *  DRAFT FINAL  * « *

-------
                                   15-42
U.S. Army Corps of Engineers.  1980.  Feasibility Report and Final
    Environmental Impact Statement:  Atlantic Coast of Maryland and
    Assateague Island, Virginia.   Baltimore:   Army Corps of Engineers.

U.S. Army Corps of Engineers.  1979.  Ocean Beach Study:  Feasibility
    Report.  San Francisco:  Army Corps of Engineers.

U.S. Army Corps of Engineers, Coastal Engineering Research Center.  1977.
    Shore Protection Manual.  Fort Belvoir, Virginia:   Coastal Engineering
    Research Center.

U.S. Army Corps of Engineers.  1971.  National Shoreline Study.  Washington,
    D.C.:  Army Corps of Engineers.

Van Beek, J.L., D. Roberts, D. Davis, D. Sabins, and S.M.  Gagliano.  1982.
    Recommendations for Freshwater Diversion to Louisiana Estuaries East of
    the Mississippi River  Baton Rouge:  Coastal Management Section,
    Louisiana Department of Natural Resources.

Waddell, J.O. and R.A. Blaylock.   n.d.  "impact of Sea Level Rise oh Gap
    Creek Watershed in the Fort Walton Beach, Florida Area."  In  Kuo,  C.Y.
    (ed.).   op. cit.

Walsh, J.E., and C.M. Johnson.  1979.  Interannual atmospheric variability
    and associated fluctuations in Arctic sea-ice extent.   Journal of
    Geophysical Research, 84, 6915-6928.

Weishar, L.L. and W.L. Wood.  1983.  "An Evaluation of Offshore and Beach
  '  Changes on a Tideless Coast."  Journal of Sedimentary Petrology
    53:3:847-858.

Wendland, W.M.  1977.  "Tropical Storm Frequencies Related to Sea Surface
    Temperatures."  Journal of Applied Meteorology 16:480.

Zwally, H.J., C.L. Parkinson, and J.C. Comiso.  1983.   Variability of
    Antarctic sea ice and changes in carbon dioxide.  Science, 220,
    1005-1012.
                          * - *  DRAFT FINAL  * * *

-------
Chapter 16

-------
                              CHAPTER  16

             POTENTIAL EFFECTS OF FUTURE  CLIMATE CHANGES
            ON  FORESTS AND VEGETATION,  AGRICULTURE,  WATER
                     RESOURCES, AND HUMAN  HEALTH
SUMMARY
    The greenhouse effect  resulting from increased levels  of C02,
chloroflurocarbons,  methane, N20, and other trace gases  in the atmosphere has
been recognized by the  scientific community for several  decades as a potential
cause of future climate change.  In the last few years,  estimates of the rate
of change of these gases in the atmosphere has heightened  concern about global
warming and associated  climate and environmental change.   Chapter 6 -- Global
Warming--presents  a review of recent chemical and physical evidence supporting
the greenhouse phenomenon.  From this evidence it is generally concluded that
in the relatively  short period of time of the next 50-100  years the earth's
climate will undergo important changes.  These include:  potential increases
in temperatures, changes in precipitation, humidity, windfields, ocean
currents, and the  frequency of extreme events such as hurricanes.
Furthermore, these climate parameters will induce still  other shifts in sea
levels, ice margins, the hydrologic cycle, air pollution episodes and other
phenomenon.

    Recently the World  Meterological Organization, the United Nation's
Environmental Programme and the International Council of Scientific Union
summarized current scientific data on global climate change.  These findings
are presented in Exhibit 16-1.  Similar findings have been reported by NAS
(1979 and 1982).

                             EXHIBIT 16-1

        SUMMARY OF FINDINGS FROM THE WMO/UNEP/ICSU CONFERENCE
       ON GLOBAL  CLIMATE HELD IN VILLACH, AUSTRIA, OCTOBER  1985

        •   Many important economic and social decisions are being
            made today  on  long-term projects -- major water resource
            management  activities such as irrigation and
            hydro-power; drought relief; agricultural land use;
            structural  designs and coastal engineering projects; and
            energy planning  -- all based on the assumption that past
            climatic data, without modification, are a reliable
            guide  to the future climate conditions.  This  is no
            longer a valid assumption, since the increasing
            concentrations of greenhouse gases are expected to cause
            a significant  warming of the global climate  in the next
            century.  It is a matter of urgency to refine  estimates
            of future climate conditions to improve these  decisions.
                          * *  *  DRAFT FINAL  * * *

-------
                       16-2
The amounts of some trace gases in the troposhere,
notably carbon dioxide (C02), nitrous oxide (N20),
methane (CH4), ozone (03), and chloro-fluorocarbons
(CFC), are increasing.  These gases are essentially
transparent to incoming short-wave solar radiation, but
absorb and emit longwave radiation and are thus able to
influence the earth's climate.

The role of greenhouse gases other than C02 in
changing the climate is already about as important as
that of C02.  If present trends continue, the combined
concentrations 6f atmospheric C02 and other greenhouse
gases would be (radiatively) equivalent to a doubling of
C02 from pre-industrial levels, possibly as early as the
2030's.

The most advanced experiments with general circulation
models of the climatic system show increases of the
global mean equilibrium surface temperature of between
1.5°C and 4.5°C for a doubling of the atmospheric C02
concentration or equivalent.  Because of the complexity
of the climatic system and the imperfections of the
models, particularly with respect to ocean-atmosphere
interactions and clouds, values outside of this range
cannot be excluded.  The realization of such changes
will be slowed by the inertia of the oceans; the delay
in reaching the mean equilibrium temperatures
corresponding to doubled greenhouse gas concentrations
is expected to be a matter of decades.

While other factors such as aerosol concentrations,
changes in solar energy input, and changes in vegetation
may also influence climate, the increased amounts of
greenhouse gases are likely to be the most important
cause of climate change over the next century.

Regional scale changes in climate have not yet been
modelled with confidence.  However, regional differences
from the global averages show that warming may be
greater in high latitudes during late autumn and winter
than in the tropics; annual mean runoff may increase in
high latitudes; and summer dryness may become more
frequent over the continents at middle latitude in the
Northern Hemisphere.  In tropical regions, temperature
increases are expected to be smaller than the average
global rise, but the effects on ecosystems and humans
could have far reaching consequences.  Potential
evapotranspiration probably will increase throughout the
tropics, whereas in moist, tropical regions, convective
rainfall could increase.
              * * *  DRAFT FINAL  * * *

-------
                                   16-3
        •   Based on the observed changes since the beginning of
            this century, it is estimated that global warming of
            1.5°C to 4.5°C would lead to a sea-level rise of 20 to
            140 cm.  A sea-level rise in the upper portion of this
            range would have major direct effects on coastal areas
            and estuaries.  A significant melting of the West
            Antarctic ice sheet leading to a much larger rise in sea
            level, although possible at some future date, is not
            expected during the next century.

        •   Based on analyses of observational data, the estimated
            increase in global mean temperature of between 0.3°C and
            0.7°C during the last 100 years is consistent with the
            projected temperature increase attributable to the
            observed increase in C02 and other greenhouse gases,
            although it cannot be ascribed in a scientifically
            vigorous manner to these factors alone.

        •   Based on evidence of the effects of past climate
            changes, there is little doubt that a future change in
            climate of the.magnitude obtained from climate models
            could have a profound effect on global ecosystems,
            agriculture, water resources, and sea ice.

    As noted in the WMO/UNEP/ICSU report the projected changes in climate will
have important impacts on all aspects of society.  Agriculture, forests, human
health, water resources, energy planning, and recreation are among the sectors
likely to be affected.  Moreover, all these sectors are likely to be affected
simultaneously throughout the world, but to different extents.  Today we know
a great deal from palioclimatic records about how past shifts in climate
affected the growth and development of forest systems, the location of lakes,
and development of agriculture.  But the changes that occurred to these
systems in the past took 18 to 20 thousand years to unfold as the earth warmed
approximately 4°C-5°C.  During that time, forest composition shifted, some
lake systems were lost and new ones were formed.  Most importantly, the
changes took place during a period when the earth's population was small and
civilizations were in formative stages.

    Today modern society is much more complex, but  still vulnerable to
climatic changes.  Our industrial society relies on a sustained climate to
replenish natural resources as a source of raw materials, for transport of
goods and for the food we eat.  We assume that the climate that supports our
society, while variable and difficult to predict over short periods, will not
shift appreciably.  Indeed, most decisions made by farmers, forest managers,
state and local water management officials, utility executives and government
policy makers assume that climate will be constant.  However, if current
predictions from global climate models prove to be correct, some increase in
global temperatures may be inevitable simply because of the presence of trace
gases which have already been emitted to the atmosphere.
                            * *  DRAFT FINAL  * * *

-------
                                   16-4
    Our current understanding of the effects of global climate change on the
environment is incomplete.  Moreover, several features of the greenhouse
phenomen make it unique, different from other environmental issues and
difficult to analyze.  Among these features are the following:

        •   The effects will not take place immediately, but over
            several decades.

        •   The effects will be virtually irreversible over
            several centuries.

        •   All nations of the world still experience the effects
            at the same time.

        •   There is no historical analog for the amount of global
            warming likely to occur in a relatively short period of
            the next 50 to 100 years.

    Scientists have only begun to analyze the potential impacts from global
warming and changes in other climate variables.  Insights are available from
historical data and from the application of predictive models.  In most cases
however, our understanding of the consequences both beneficial and
detrimental, is in a formative stage.  Historical analogs provide qualitative
information about likely effects, but they cannot predict the future because
the anticipated rate and increase in temperature are beyond the range of
previous warm periods.  Predictive models of both the climate system and
potential effects often do not include complete parameterizations of important
system variables.  Thus, more advanced global climate models capable of
providing regional predictions are needed, and more comprehensive and
sophisticated analyses of environmental effects are necessary to understand
fully the implications throughout the world.

    Recognizing these limitations, the following section summarizes what is
known about climate impacts on the environment with emphasis on forests,
agriculture, water resources and human health.  Perhaps of equal importance
are potential impacts which have not been reported or analyzed, including
potential impact on ports, electricity, demand and supply, population shifts,
work place absenteeism rates, hurricane frequency, air pollution emissions,
wildlife management and national security.  These potential impact areas and
many others represent the challenges to be investigated as the science
supporting predictions of global climate change improves.
                          * * *  DRAFT FINAL  * * *•

-------
                                   16-5
FINDINGS

    The following findings are summarized from Appendix B,  which provides  a
comprehensive review of potential impacts of global climate change.

1.  CLIMATE CHANGE HAS HAD A SIGNIFICANT IMPACT ON FORESTS  IN THE PAST.   IF
    CURRENT PREDICTIONS PROVE ACCURATE.  THERE IS A POTENTIAL FOR DRAMATIC
    SHIFTS IN FORESTS AND VEGETATION OVER THE NEXT 100 YEARS.

    la. Climate models predict that a global warming of approximately 1.5°C to
        4.5°C will be induced by a doubling of atmospheric  C02 and other Trace
        gases during the next 50 to 100  years.  The period  18,000 to 0 years
        B.P. is the only general analog  for a global climate change of this
        magnitude.  The geological record from this glacial to inter- glacial
        interval provides a basis for qualitatively understanding how
        vegetation may change in response to large climatic change.

    Ib. The paleovegetational record shows that climatic change as large as
        that expected to occur in response to C02-doubling  is likely to induce
        significant changes in the composition and patterns of the world's
        biomes.  Changes of 2°C to 4°C have been significant enough to alter
        the composition of biomes, and to cause new biomes  to appear and
        others to disappear.  At 18,000  B.P., the vegetation in Eastern North
        America was quite distinct from  that of the present day.  The cold/dry
        climate of that time seems to have precluded the widespread growth of
        birch, hemlock, beech, alder, hornbeam, ash, elm and chestnut, all of
        which are fairly abundant in present-day deciduous  forest.  Southern
        pines were limited to grow with  oak and hickory in  Florida.

    Ic. Available paleoecological and paleoclimatological records do not
        provide an analog for the high rate of climate change and
        unprecedented global warming predicted to occur over the next
        century.  Previous changes in vegetation have been  associated with a
        climates that were nearly 5°C to 7°C cooler and which took thousands
        of years to evolve rather than the predicted decadal changes.
        Insufficient temporal resolution e.g., via radiocarbon dates limits
        our ability to analyze the decadal-scale rates of change that occurred
        prior to the present millennium.

    Id. Limited experiments conducted with dynamic vegetation models for North
        America suggest that decreases in net biomass may occur and that
        significant changes in species composition are likely.  Experiments
        with one model suggest that Eastern North American biomass may be
        reduced by 11 megagrams per hectare (10% of live biomass) given the
        equivalent of a doubled C02 environment.  Plant taxa will respond
        individualistically rather than  as whole communities to regional
        changes in climate variables.  At this time such analyses must be
        treated as only suggestive of the kinds of change that could occur.
        Many critical processes are simplified or omitted and the actual
        situation could be worse or better.
                          * * *  DRAFT FINAL  * * *

-------
                                   16-6
    le. Future forest management decisions in major timber-growing regions are
        likely to be affected by changes in natural growing conditions.  For
        example, one study suggests that loblolly pine populations are likely
        to move north and northeast into Pennsylvania and New Jersey while its
        range shrinks in the west.  The total geographic range of the species
        may increase, but a net loss in productivity may result because of
        shifts to less accessible and productive sites.  While the extent of
        such changes is unclear, adjustments will be needed in forest
        technology, resource allocation, planning, tree breeding programs, and
        decision-making to maintain and increase productivity.

    If. Dynamic vegetation models based on theoretical descriptions of all
        factors that could influence plant growth must be improved and/or
        developed for all major kinds of vegetation.  In order to make more
        accurate future predictions, these models must be validated using the
        geological record and empirical ecological response surfaces.  In
        particular, the geological record can be used to test the ability of
        vegetation models to simulate vegetation that grew under climate
        conditions unlike any of the modern day conditions.

    Ig. Dynamic vegetation models should incorporate direct effects of
        atmospheric C02 increases on plant growth and other air pollution
        effects.  Improved estimates of future regional climates are also
        required in order to make accurate predictions of future vegetation
        changes.

2.   AGRICULTURE IS VERY DEPENDENT ON CLIMATE.  LIMITED ASSESSMENTS SUGGEST
    THAT IMPORTANT CHANGES IN AGRICULTURE AND FARM PRODUCTIVITY ARE LIKELY
    THROUGHOUT THE WORLD. IF CLIMATE CHANGE OCCURS AS PREDICTED.

    2a. Climate has had a significant impact on farm productivity and
        geographical distribution of crops.  Examples include the 1983 drought
        which contributed to a nearly 30% reduction in corn yields in the
        U.S., the persistent Great Plains drought between 1932-1937 which
        contributed to nearly 200,000 farm bankruptcies, and the climate shift
        of the Little Ice Age (1500-1800) which led to the abandonment of
        agricultural settlements in Scotland and Norway.

    2b. World agriculture is likely to undergo significant shifts if trace
        gas-induced climate warming in the range of 1.5°C to 4.5°C occurs over
        the next 50 to 100 years.  Climate effects on agriculture will extend
        from local to regional and international levels.  However, modern
        agriculture is very dynamic and is constantly responding to changes in
        production, marketing and government programs.

    2c. The main effects likely to occur at the field level will be physical
        impacts of changes in thermal regimes, water conditions, and pest
        infestations.  High temperatures have caused direct damage to crops
        such as wheat and corn; moisture stress, often associated with
        elevated temperatures, is harmful to corn, soybean and wheat during
        flowering and grain fill; and increased pests are associated with
        higher, more favorable temperatures.


                          * * *  DRAFT FINAL  * * *

-------
                                   16-7
    2d. Even relatively small increases in the mean temperature can increase
        the probability of harmful effects in some regions.   Analysis of
        historical data has shown that an increase of 1.7°C  (3°F) in mean
        temperature changes the likelihood of a five consecutive daily maximum
        temperature event of at least 35°C (95°F) by about a factor of three
        for a city like Des Moines.  In regions where crops  are grown close to
        their maximum tolerance limits, changes extreme temperature events may
        have significant harmful effects on crop growth and  yield.

    2e. Limited experiments using climate scenarios and agricultural
        productivity models have demonstrated the sensitivity of agricultural
        systems to climate change.  Future farm yields are likely to be
        effected by climate because of changes in the length of the growing
        season, heating units, extreme winter temperatures,  precipitation, and
        evaporative demand.  In addition, field evaluations  show that total
        productivity is a function of the drought tolerance  of the land and
        the moisture reserve, the availability of land, the  ability of farmers
        to shift to different crops, and other factors.

    2f. The transition costs associated with adjusting to global climatic
        change are not easily calculated, but are likely to  be very large.
        Accommodating to climate change may require shifting to new lands and
        crops, creating support services and industries, improving and
        relocating irrigation systems, developing new soil management and pest
        control programs, and breeding and introducing new heat or drought
        tolerant species.  The consequences of these decisions on the total
        quantity, quality, and cost of food are difficult to predict.

    2g. Current projections of the effects of climate change on agriculture
        are limited because of uncertainties in predicting local temperature
        and precipitation patterns using global climate models, and because of
        the need for improved research studies using controlled atmospheres,
        statistical regression models, dynamic crop models and integrated
        modeling approaches.

3.  WATER RESOURCE SYSTEMS HAVE UNDERGONE IMPORTANT CHANGES  AS THE EARTH'S
    CLIMATE HAS SHIFTED IN THE PAST.  CURRENT ANALYSES SUGGEST AN INTENSIFIED
    HYDROLOGIC CYCLE. IF CLIMATE CHANGE OCCURS AS PREDICTED.

    3a. There is evidence that climate change since the last ice age (18,000
        years B.P.) has significantly altered the location of lakes although
        the extent of present day lakes is broadly comparable with 18,000
        years B.P.  For example, there is evidence indicating the existence of
        many tropical lakes and swamps in the Sahara, Arabian, and Thor
        Deserts around 9,000 to 8,000 years B.P.

    3b. The inextricable linkages between the water cycle and climate ensure
        that potential future climate change will significantly alter
        hydrological processes throughout the world.  All natural hydrological
        processes -- precipitation, infiltration, storage and movement of soil
        moisture, surface and subsurface runoff, recharge of groundwater, and
        evapotranspiration -- will be affected if climate changes.


                          * * *  DRAFT FINAL  * * *

-------
                                   16-8
    3c. As a result of changes in key hydrological variables such as
        precipitation, evaporation, soil moisture, and runoff, climate change
        is expected to have significant effects on water availability.  Early
        hydrological impact studies provide evidence that relatively small
        changes in precipitation and evaporation patterns might result in
        significant, perhaps critical, changes in water availability.   For
        many aspects of water resources, including haman consumption,
        agricultural water supply, flooding and drought management,
        groundwater use and recharge, and reservoir design and operation,
        these hydrologic changes will have serious implications.

    3d. Despite significant differences among climate-change scenarios, a
        consistent finding among hydrological impact studies is the prediction
        of a reduction in summer soil moisture and changes in the timing and
        magnitude of runoff.  Winter runoff is expected to increase and summer
        runoff will decrease.  These results appear to be robust across a
        range of climate change scenarios.

    3e. Future directions for research and analyses suggest that improved
        estimates of climate variables are needed from large-scale climate
        models; innovative techniques are needed for regional assessments;
        increased numbers of assessments are necessary to broaden our
        knowledge of effects on different users; and increased analyses of the
        impacts of changes in water resources on the economy and society are
        necessary.

4.  WEATHER IS CLOSELY ASSOCIATED WITH MORBIDITY AND MORTALITY RATES IN OUR
    SOCIETY.

    4a. Weather has a profound effect on human health and well being.   It has
        been demonstrated that weather impacts are associated with changes in
        birth rates, outbreaks of pneumonia, influenza, and bronchitis, and
        related to other morbidity effects and linked to pollen concentrations
        and high pollution levels.

    4b. Large increases in mortality have occurred during previous heat and
        cold waves.  It is estimated that 1,327 fatalities occurred in the
        United States as a result of the 1980 heat wave and Missouri alone
        accounted for over 25% of that total.

    4c. Hot weather extremes appear to have a more substantial impact on
        mortality than cold wave episodes.  Most research indicates that
        mortality during extreme heat events varies with age, sex, and race.
        Factors associated with increased risk include alcoholism, living on
        higher floors of buildings, and the use of tranquilizers.  Factors
        associated with decreased risk include the use of air conditioning,
        frequent exercising, consumption of fluids, and living in shaded
        residences.  Acclimatization may moderate the impact of successive
        heat waves over the short-term.
                          * * *  DRAFT FINAL  * *

-------
                               16-9
4d. Threshold temperatures for cities have been determined which represent
    maximum and minimum temperatures associated with increases in total
    mortality.  These threshold temperatures vary regionally, i.e., the
    threshold temperature for winter mortality in mild southern cities
    such as Atlanta is 0°C and for more northerly cities such as
    Philadelphia, threshold temperature is -5°C.

4e. Humidity has an important impact on mortality, since it contributes to
    the body's ability to cool itself by evaporation of perspiration.  It
    also has an important influence on morbidity in the winter because
    cold, dry air leads to excessive dehydration of nasal passages and the
    upper respiratory tract and increased chance of microbial and viral
    infection.

4f. Precipitation in the form of rainfall and snow is also associated with
    changes in mortality.  In New York City, upward trends in mortality
    were noted the day after snowfalls which had accumulated 2 in. or
    more.  In Detroit where snowfall is more frequent, mortality increases
    were noted after the snowfall accumulation exceeded 6 in.

4g. If future global warming induced by increased concentrations of trace
    gases does occur, it has the potential to significantly affect human
    mortality.  In one study, total summertime mortality in New York City
    was estimated to increase by over 3,200 deaths per year for a 7°F
    trace gas-induced warming without acclimatization.  If New Yorkers
    fully acclimatize, the number of additional deaths is estimated to be
    no different than today.  It is hypothesized that, if climate warming
    occurs, some additional deaths is likely to occur because economic
    conditions and the basic infrastructure of the city will prohibit full
    acclimatization even if behavior changes.
                      * * *  DRAFT FINAL  * * *

-------
                                   16-10
REFERENCES

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

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

World Meterological Organization,  United Nations Environment  Programme  and
International Council of Scientific Unions  (1986), Report of  the  International
Conference on the Assessment of the Role of Carbon Dioxide  and of Other
Greenhouse Gases in Climate Variations and  Associated Impacts, Villach  Austria
9-15 October 1985 WHO No.  661.
                          * * *  DRAFT FINAL  * * *

-------
Chapter 17

-------
                               CHAPTER 17

          MODELS  FOR  INTEGRATING THE ANALYSES  OF HEALTH AND
                   ENVIRONMENTAL  RISKS ASSOCIATED WITH
                          OZONE MODIFICATION
SUMMARY

    Analyses of the potential risks  due  to  emissions of ozone-modifying
substances and greenhouse gases  must assess  a variety of effects on human
health and the environment.   To  assess risks numerous factors must be
examined together,  including:  economic and  population growth; chemical use and
emissions; ozone modification; changes in ultraviolet (UV) radiation flux;
climate change; and effects  on human health and  the environment.

    To evaluate risks the Integrating Analysis Model has been developed to
provide a framework for systematically evaluating the risks and costs
associated with alternative  assumptions  about the ozone modification issue and
policies for limiting the emissions  of ozone-modifying substances.  This
chapter describes the manner in  which the Integrating Analysis Model (the
model) draws from the analyses described in earlier chapters.

    The model begins with assumptions about  the  potential future use of
ozone-modifying substances and concentrations of other trace gases, including
chlorofluorocarbons, methyl  chloroform,  carbon tetrachloride, Halons
(bromine-containing compounds),  carbon dioxide,  methane, and nitrous oxide.
This potential future use and emissions  may be modified by control
strategies.  The resulting global  emissions  of the substances are computed,
followed by the atmospheric  science  module, which assesses the impacts of
these emissions on ozone. The risks of  ozone modification for human health
and the environment are evaluated  using  specified dose-response relationships
where available for each of  the  areas of interest.

    Using this overall framework,  the model  can  reflect a wide range of
assumptions and the joint implications of all the assumptions can be
identified.  Of note is that the model is not a  substitute for choosing the
assumptions, relationships,  and  data to  describe the system being modeled.  It
merely reflects these assumptions  and data  in a  consistent framework.1
    1 When used for performing policy analysis,  the model does not make
decisions, for example,  about which,  if any,  control  strategies are preferred
or about which, if any,  ozone depletion estimates  are correct.  Decisions
about the need to protect stratospheric ozone or the  best way to do so must be
based not only on model  results,  but  also on  considerations and judgments
either not reflected in  the model or  reflected implicitly in the assumptions
provided as input to the model.
                             *** DRAFT FINAL ***

-------
                                   17-2
INTRODUCTION

    Analyses of the potential risks due to emissions of ozone-modifying
substances and greenhouse gases must assess a variety of effects on human
health and the environment.  As described earlier in this document, numerous
factors influence ozone modification and its effects.  To assess risks, the
following must be characterized, and their joint implications (including their
uncertainties) examined together: economic and population growth; chemical use
and emissions; ozone modification; changes in ultraviolet (UV)  radiation flux;
climate change; and effects on human health and the environment.

    Analyzing and combining these factors is a complex task.   To make this
task tractable, two models have been developed:

        1.  Integrating Analysis Model.  The primary role of  the
            Integrating Analysis Model is to provide a framework for
            systematically evaluating the risks and costs associated
            with alternative assumptions about the ozone
            modification issue and policies for limiting the
            emissions of ozone-modifying substances.

        2.  Transient Temperature and Sea Level Rise Model.   The
            primary role of this model is to assess the implications
            of emissions of greenhouse gases in terras of increases
            in global average temperature and sea level over  time.

Together, these two models provide a comprehensive framework  for evaluating
the potential risks of emissions of potential ozone-modifying substances and
greenhouse gases.   This chapter describes the manner in which the Integrating
Analysis Model (the model) draws from the analyses described  in earlier
chapters.2  The Transient Temperature and Sea Level Rise Model is described
in Appendix A of Chapter 6.

    This chapter first describes the framework that the model provides for the
analysis.  Then, an overview of the model's analysis procedure is presented.
The third section of this chapter presents the major limitations of the
model.  The chapter concludes with a series of appendices that describe the
model's design and operation in detail.

THE MODEL AS A  FRAMEWORK

    The objective of the model is to provide a framework within which the
implications of alternative assumptions and policies can be identified.  The
framework reflects the order and structure of the analysis,  and the user may
define the assumptions and data to be used in the model for any given run.
Exhibit 17-1 displays the major parts or "modules" of the model.  The model
    2 The Integrating Analysis Model was originally described in Gibbs and
Hoffman (1986).
                             *** DRAFT FINAL ***

-------
                                                         EXHIBIT 17-1

                                                      MODULAR STRUCTURE
   Potential  Future
Use of Ozone-Doplet ing
    Substances and
  Concentrations of
  Other Trace Gases
     I           I
     I  Control
     I  Strategy I
     I	I
                                                                        I              I
                                                                        I  Atmospheric |_
                                                                            Science   j
                                                                        I	I
 Hea I tti and
Env i ronmentaI
   Effects
                                                     ***  DRAFT  FINAL  ***

-------
                                   17-4
begins with assumptions about the potential future use of ozone-modifying
substances and concentrations of other trace gases, including chlorofluoro-
carbons, methyl chloroform, carbon tetrachloride, Halons (bromine-containing
compounds), carbon dioxide, methane, and nitrous oxide.  This potential future
use and emissions may be modified by control strategies.  The resulting
global emissions of the substances are computed, followed by the atmospheric
science module, which assesses the impacts of these emissions on ozone.3
The risks of ozone modification for human health and the environment are
evaluated using specified dose-response relationships where available for each
of the areas of interest.

    Using this overall framework, the model can reflect a wide range of
assumptions within each of the modules.  When the model is run, the joint
implications of all the assumptions are identified.  Of note is that the model
is not a substitute for choosing the assumptions, relationships, and data to
describe the system being modeled.  It merely reflects these assumptions and
data in a consistent framework.  In its simplest terms, the model is a
calculator that is needed to assist in computing multiple estimates of
possible values.  Decisions about best assumptions or the need to protect
stratospheric ozone or the best way to do so may use the model and its
results, but must also be based on considerations and judgments either not
reflected in the model or reflect implicitly in the assumptions provided as
input to the model.

ANALYSIS PROCEDURE

    A model "run" is comprised of:

        •   inputs provided by the user that define the data and
            assumptions to use in the analysis;

        •   algorithms that are carried out according to the
            user's inputs and instructions contained within the
            model; and

        •   outputs prepared by the model that describe the
            results of the algorithms that are carried out.

The model is "deterministic," meaning that a unique output or result is
produced by a specified input; there are no random (i.e., stochastic) elements
within the model.

    To initiate a model run, the user must specify inputs for each module he
wishes to use.  (The user may choose to run only a portion of the model.)
    3 These emissions rates are also used in the Transient Temperature and
Sea Level Rise Model to evaluate climate change effects.
                             *** DRAFT FINAL ***

-------
                                   17-5
Exhibit 17-2 provides an overview of the input choices for each of the
following modules:

        •   Potential Use.  The user may choose one of the five
            scenarios of future use that have been incorporated into
            the model.  The compounds included are:  chlorofluoro-
            carbons (CFC-11; CFC-12; CFC-22; CFC-113); carbon-
            tetrachloride (CCL4); methyl chloroform (CH3CCL3);
            Halon-1211; and Halon-1301.   Each compound can be
            analyzed for up to six end uses, in 10 regions of the
            world.  The five scenarios reflect a wide range of
            potential future use, and are based on historical trends
            and projections.*  Alternatively, the user may specify
            his own scenario of future use.

        •   Control Strategy.  The user may specify a variety of
            controls on future use, such as kilogram limits, use
            limits, technology requirements, or fees.  The potential
            costs and benefits of controls are not evaluated in this
            document because the objective of this risk assessment
            is to evalute risks in the absence of regulation.

        •   Global Emissions.  Emissions of the chlorofluoro-
            carbons, CCL4, CH3CCL3, Halon-1211 and Halon-1301 are
            computed from the production and use estimates based on
            the release rate algorithm data described in Chapter 3.
            The release rates reflect the fact that these chemicals
            are stored in certain products for various periods of
            time.  The effects of control strategies are reflected
            in the release estimates, and the user may specify
            alternative release rates as desired.  Scenarios of
            emissions' or atmospheric concentrations are also
            specified for carbon dioxide (C02), methane (CH4) and
            nitrous oxide (N20).

        •   Atmospheric Science.   A parameterized numerical fit
            to a one-dimensional (1-D) atmospheric model is used to
            compute global ozone depletion as a function of global
            emissions of chlorofluorocarbons, CC14, CH3CC13, and
            Halons, and concentrations of C02, N20 and CH4.5  The
            uncertainty in the estimates of global depletion may be
            modeled.6  Alternatively, the user may bypass this
    *  These five scenarios are described in Chapter 3.

    5  See Connell (1986) for a description of this method.

    6  The range of uncertainty identified in Stolarski and
Douglass (1985) is currently incorporated into the model.  The user
may provide alternative uncertainty estimates.


                             *** DRAFT FINAL ***

-------
                                                          EXHIBIT  17-2

                                                   MAJOR MODEL  INPUT CHOICES
        MODULE
             CHOICES
                                                                       COMMENTS
                                                                                                            SOURCES
Potential  Future Use
or Ozone-Modifying
Substances
Future Use:

-- Scenarios based on studies of
   historical trends and projec-
   tions of economic growth and
   population growth

-- Any user-supplied estimates of
   future use can be incorporated
Future Use:

-- Scenarios reflect a wide range
   of potential future use.
Based on:  UNEP (1986).
Control  Strategies
Strategies can be defined in
terms of limits to production in
various ways,  including:

-- limits in kilograms;

-- limits in kilograms per
   capi ta;

-- bans of specified uses;

-- requirements to implement
   specified technologies;  and

-- fees or taxes.
Data on control options are based
on U.S. data.  Information for
other parts of the world may be
entered by the user, or the U.S.
data can be applied to the rest
of the world as a default is no
data are ava i(able.
Estimates of the costs and effec-
tiveness of controls are based  on
data presented  in Cawm (1986).
Analyses are underway to update
these estimates.
Global  Emissions
Several  estimates of the rate  of
release after uise have been
developed based on analyses of
historical  data.   Emissions
reflect the effectiveness of the
control  strategy in limiting
emissions.   The user may choose
several  scenarios of effective-
ness,  which reflects the level of
reductions  expected to occur as
the result  of the control strate-
gy identified.
Data on the effectiveness of con-
trol options are based on U.S.
data.   Information on other parts
of the world may be entered by
the user.
The rates of emissions following
use are based on data presented
in Quinn (1986).
                                                      *** DRAFT FINAL ***

-------
                                                     EXHIBIT  17-2  (continued)

                                                   MAJOR MODEL  INPUT CHOICES
        MODULE
Atmospheric Science
Health and Environ-
mental Effects
             CHOICES
Ozone Modification:

-- Pa rametcr ized relationship
   between global  emissions and
   global  ozone depletion.

— The user may specify his own
   estimate of ozone depletion
   for use in the model.   If this
   choice  is made,  the global
   emissions estimate is not
   used.
Ranges of relationships with UV
radiation are developed for each
effect.   The user may specify for
each effect low,  medium,  or high
est imates.
             COMMENTS
Ozone Modification:

-- Global estimates based on
   analysis of a one-dimensional
   atmospehric model.

— Estimates that vary by lati-
   tude are based on an analysis
   of a two-dimensional atmos-
   pheric mode I .
Ranges based on published esti-
mates.
                                                                                                           SOURCES
Based on:  Connell
Isaksen ( 1980).
(1986) and
                                                                                                                                  —i
                                                                                                                                  i
Based on analyses by:  Scotto,
Fears, and Fraumeni (1981),
Scotto and Fears (in press),
Miller, Sperduto, and Ederer
(1983), Horst (1986), Andrady
(1986), and Pitcher (in press),
                                                      *** DRAFT FINAL ***

-------
                                   17-8
            algorithm and use a different relationship or
            independent estimates of global ozone depletion.   Using
            the parameterization estimate of global depletion,
            estimates of ozone modification by latitude are then
            made based on an analysis of the results of a
            two-dimensional model (Isaksen 1986).  It should be
            noted that this two step method of producing latitudinal
            depletion estimates results in lower estimates of
            depletion than Isaksen (1986) because the global
            estimates based on the parameterization are lower than
            the global estimates produced by Isaksen (1986).

        •   Health and Environmental Effects.  The association
            between ozone depletion, ultraviolet (UV) radiation, and
            effects may be specified by the user.  Ranges of
            dose-response relationships may be identified for
            melanoma and non-melanoma skin cancers, and cataracts.
            Population projections (by race, age, and sex) are
            incorporated into the model for purposes of computing
            future incidence of human health effects.  Alternative
            population estimates may be identified by the model
            user.  The effects on PVC polymers are also reported.
            Estimates of effects are made for the U.S. only.

    Based on these inputs, the model performs calculations on a year-by-year
basis.  For example, at the beginning of year t, the model first computes the
potential use of each of the ozone-modifying substances, in each of its
applications, in each portion of the world being analyzed.  This estimate of
use reflects the user's choice of scenario (for the first module) as well as
the effects of any controls initiated according to the user's instructions
prior to and including year t.7  Based on this estimate of use in year t,
and estimates for all years prior to year t, the global emissions in year t
are computed for each potential ozone-depleting substance.

    The atmospheric science module computes an estimate of global ozone
depletion for year t based on the emissions estimates for year t, atmospheric
conditions due to emissions prior to year t, and user-specified assumptions
regarding the concentrations of carbon dioxide, methane, and nitrous
oxide.8  Global ozone depletion is translated into estimates that vary by
latitude.  The latitudinal estimates of ozone depletion are then translated
into associated changes in UV flux, using three action spectra for weighting
    7  The effects of controls include the direct reduction in use due to
the application of control technologies, as well as the potential reduction in
use associated with the influence that increased production costs and prices
may have on demand.

    8  Appropriate ranges of assumptions for these gases are described in
Chapter 4.
                             *** DRAFT FINAL ***

-------
                                   17-9
the different UV wavelengths:  DNA; Robertson-Berger Meter (RB-Meter) and
erythema.  The changes in UV flux in year t (and prior years) are used in
dose-response relationships to estimate increased risks of melanoma and
non-melanoma skin cancers and cataracts.  Other impacts, such as polymer
degradation, are also computed.

    After the computations for year t are performed, the model repeats the
computations for year t+1.  This annual series of computations is consequently
repeated for each year of the model run.  After all the years have been
analyzed, the following summary outputs are produced:

        •   future production of ozone-depleting substances;

        •   future emissions of ozone-depleting substances;

        •   future global ozone depletion;

        •   future ozone depletion by latitude;

        •   future increases in UV flux; and

        •   estimates of effects, including:

                melanoma skin cancers (incidence and mortality);
                non-melanoma skin cancers (incidence and mortality);
                cataracts (incidence); and
                costs of damages to polymers (PVC used outdoors).

The effects estimates are currently based primarily on data that describe the
U.S.  Consequently, only U.S. effects estimates are currently produced.

MODEL LIMITATIONS

    In using any model, it is important to understand its limitations.  The
model is being developed in a neutral fashion.  Its flexible design allows
users to evaluate wide ranges of inputs and assumptions.  The model cannot
incorporate, however, numerous factors that are as yet undefined or poorly
understood.  Consequently, the model excludes a variety of potential factors
and analyses that may be important for making decisions regarding
stratospheric ozone protection.

    The most important exclusions from the model are a variety of effects of
ozone modification.  Many effects that increased UV radiation will have on
human health and the environment are not well quantified.  For example, only
limited data are available that describe the effects of UV radiation on plants
and aquatic organisms.  These data cannot be extrapolated to all crops and
ecosystems to evaluate the net effects of ozone modification and global
warming.  The potential for effects is substantial, but current data do not
allow quantitative estimates at this time.  Similarly, the effects of UV
radiation on ground oxidants are not well quantified at this time (see Whitten
(1986) for case studies of effects on oxidants).
                             *** DRAFT FINAL ***

-------
                                   17-10
    Similarly, too little knowledge currently exists to allow immune
suppression effects of increased UV radiation to be quantified.   These effects
may be particularly important because they may affect all individuals in the
world in important ways that are not currently anticipated.

    Exhibit 17-3 summarizes the excluded effects.  The omission of these known
effects, which cannot be quantified, is a clear bias reflected in the
estimates of the risks associated with ozone depletion.  In addition, there
may be other links among UV radiation, global warming, and human health and
the environment that have not yet been identified.  The potential bias of the
results due to lack of information on these effects is unknown.

    The model is limited in its ability to address issues related to risk
management and policy analysis.9  This document does not address these
issues because it is focused on risk assessment.
    9The model currently does not address a variety of risk management
issues, including:

        •   the feasibility of alternative control strategies;

        •   potential trade of ozone-modifying substances among
            regions of the world; and

        •   the number of countries likely to participate in
            international agreements to limit the use and emissions
            of ozone-depleting substances.

Decision makers must use their own judgments regarding these factors.  The
model may be used to test the implications of these judgments (e.g., the
implications of only half the countries participating in a control strategy
may be identified by specifying the expected participation as a model input)
However, the model is not meant to be used in place of these types of
judgments, or any others, that are required of decision makers.
                             *** DRAFT FINAL ***

-------
                                   17-11



                              EXHIBIT  17-3

                           Effects  Not Quantified
              Effect
               Comments
Impact on plants
Impact on aquatic organisms
Impact on urban ground-based ozone
formation (smog)
Increased incidence of cutaneous
diseases as a result of immune
suppression

Increased incidence of serious cases
of non-melanoma skin cancer

Impacts of climate change on human
health, water resources, forestry,
and agriculture

Impacts of sea level rise on wetlands

Impacts on polymers other than PVC
Preliminary experiment data indicate
risks of decreased yields.   See
Chapter 11.

Laboratory data indicate increase UV
radiation may adversely affect
survivability and breeding.  See
Chapter 12.

Case studies of three cities indicate
potential increases in smog formation,
and formation earlier in the day.  See
Chapter 14.

See Chapter 9.
Relationship with UV-B radiation not
quantified.

See Chapter 16.
See Chapter 15.

See Chapter 13.
                             *** DRAFT FINAL ***

-------
                                   17-12
REFERENCES
Andrady, A. (1986), "Analysis of Technical Issues Related to the Effect  of  UV-B
    on Polymers," prepared for the U.S.  Environmental Protection Agency,
    Washington, D.C.

Camm, F. et al. (1986), "Social Cost of Technical Control Option to Reduce
    Emissions of Potential Ozone Depleters in the United States:  An Update,"
    The RAND Corporation, prepared for the U.S.  Environmental Protection
    Agency, Washington, D.C.

Connell, Peter S. (1986), "A Parameterized Numerical Fit to Total Column Ozone
    Changes Calculated by the LLNL 1-D Model of the Troposphere and
    Stratosphere," Lawrence Livermore National Laboratory,  Livermore,
    California.

Gibbs, M.J. and J.H. Hoffman (1986), "STRATO:  A Model for Analyzing the
    Implications of Control Strategies for Protecting Stratospheric Ozone,"
    presented at the EPA Workshop on CFCs, March 1986.

Miller, R., R. Sperduto, and F. Ederer (1983), "Epidemiologic Associations
    with Cataract in the 1971-1972 National Health and Nutrition Examination
    Survey," American Journal of Epidemiology, Vol. 118, No. 2, pp. 239-248.

Horst, R.L. (1986), "The Economic Impacts of Increased UV-B Radiation on
    Polymer Materials:  A Case Study of Rigid PVC," prepared for the U.S.
    Environmental Protection Agency, Washington, D.C.

Isaksen, I.S.A. (1986), "Ozone Perturbations Studies in a Two-Dimensional Model
    with Temperature Feedback in the Stratosphere Included," presented at UNEP
    Workshop on the Control of Chlorofluorocarbons, Leesburg, Virginia,
    September 1986.

Pitcher, H.  (1986), "Melanoma Death Rates and Ultraviolet Radiation in the
    United States," U.S. Environmental Protection Agency, Washington,  D.C.

Quinn, T.H., et al. (1986), "Projected Use, Emissions, and Banks of Potential
    Ozone-depleting Substances," The RAND Corporation, prepared for the U.S.
    Environmental Protection Agency, Washington, D.C.

Scotto, J., T.R. Fears, and J.F. Fraumeni, Jr. (1981), "Incidence of
    Nonmelanoma Skin Cancer in the United States," U.S. Department of Health
    and Human Services, (NIH) 82-2433, Bethesda, Maryland.

Scotto, J., and T. Fears (in press), "The Association of Solar Ultraviolet
    Radiation and Skin Melanoma Among Caucasians in the United States,"
    Cancer Investigations.
                             *** DRAFT FINAL ***

-------
                                   17-13
Stolarski, R.S., and A.R.  Douglas.  (1985),  "Sensitivity of an Atmospheric
    Photochemistry Model to Chlorine Perturbations Including Consideration of
    Uncertainty Propagation," NASA/Goddard Space Flight Center,  Greenbelt,
    Maryland, and Applied Research Corporation,  Landover,  Maryland.

UNEP (1986), "Summary Paper for Topic #2:   Projections of  Future Demand,"
    presented at the UNEP Workshop on the Control of Chlorofluorocarbons,
    Rome, Italy, May 1986.
                             *** DRAFT FINAL ***

-------
                              APPENDIX A

                     MODEL DESIGN  AND MODEL FLOW
INTRODUCTION

    This appendix describes the design and flow of  the  CFG Policy Model  (the
model).  The model is deterministic -- it computes  a  series of outputs from
data supplied by the model user and from data that  reside in  files associated
with the model.  A model "run" is comprised of (1)  a  set of instructions
(i.e., inputs) supplied by the model user; (2) a set  of computations performed
within the model; and (3) a set of output tables that present the results of
the computations performed by the model.   The model is  designed with a modular
structure, with each "module" performing a designated function.  The modules
are independent entities that communicate with each other.

    The modular structure facilitates the maintenance and development of the
model by:

        •   allowing desired modifications to the model to be
            localized to individual modules;

        •   allowing communication among modules to be  identified
            easily (communication among modules is  particularly
            important for modeling feedbacks from one portion of the
            model to another portion); and

        •   minimizing memory requirements and execution time.

    The remainder of this appendix is divided into  the  following sections:

        •   Level of Aggregation;
        •   Model Flow;
        •   Specifying a Model Run; and
        •   Limitations.

LEVEL OF AGGREGATION

    The smallest unit of analysis in the model includes compound, region, end
use, and time, as follows:

        •   Compound -- Up to ten compounds may be  analyzed
            separately:

            --  CFC-11;
            --  CFC-12;
            --  CFC-22;
            --  CFC-113;
            --  CFC-114;
            --  CFC-115;
            --  Carbon tetrachloride (CC14);


                             *** DRAFT FINAL ***

-------
                                   A-2
            --  Methyl chloroform (CH CC1 );
            --  Halon-1211; and
            --  Halon-1301.

            Data are currently not analyzed for CFC-114 and CFC-115.

        •   Region -- The world is divided into 10 separate
            regions (the regions can be analyzed individually or in
            groups):

            --  USA;
            --  Canada;
            --  Western Europe;
                Japan, Australia, and New Zealand;
            --  USSR and East Bloc;
                Central Planned Asia;
            --  Middle East;
            --  Africa;
                Latin America; and
                South and East Asia.

        •   End Use -- Each compound may be identified with up to
            six applications or "end uses."  Each end use can be
            analyzed separately.  End uses currently incorporated
            into the model include:

                aerosol propellant;
                flexible foam;  .
            --  rigid polyurethane foam;
                rigid nonurethane foam;
            --  refrigeration;
            --  solvent;
                fire extinguisher; and
                miscellaneous.

        •   Time -- a year is the smallest unit of time in this
            model.  The current range is 1931-2165.

Analysis is performed at these levels of information.  Major revisions to the
model are required to modify this level of disaggregation.

MODEL FLOW

    The model is divided into two main parts:  (1) the Analysis Program reads
input data, performs computations, and creates output files; (2) the Report
Writer Program reads output files from the Analysis Program, and creates
formatted reports as specified by the user.  This section discusses each part
of the model in turn.
                             *** DRAFT FINAL ***

-------
                                   A-3
Analysis Program

    The Analysis Program begins by reading an input file that describes the
configuration of the model run desired by the user, including:

        •   which years will be analyzed;

        •   the portions of the model (i.e., modules) that will be
            used in the analysis;

        •   the data files that will be used as inputs; and

        •   the years for which output values should be computed.

Based on these user-supplied data, a series of variables are initialized and
the model run specified by the user then proceeds.

    The flow of a model run can be described in terms of the following five
primary parts or "modules" of the Analysis Program:

        1.   Production Scenarios Module;
        2.   Policy Alternatives Module;
        3.   Emissions Module;
        4.   Atmospheric Science Module; and
        5.   Effects Module.

As shown in Exhibit A-l, these five modules are all executed within a "year
loop," i.e., all the information from all the modules is computed for one
year, then the model moves on to the next year.  This structure has several
benefits, including:

        •   the results of year "t" can be used to indicate what
            would happen in year "t+l" (in other words, feedbacks
            between modules and across years can be modeled);

        •   the information required to be stored is limited to
            the current year being analyzed.

Appendices B through F describe each of the modules in turn.  Briefly, each
module performs the following computations:

        •   Appendix B:  Production Scenarios Module -- This
            module computes a scenario of production over time for
            each compound.  The scenario for each compound is
            disaggregated by region and end use.

        •   Appendix C:  Policy Alternatives Module -- This
            module identifies the implications of alternative
            control policies specified by the user.  The module's
            input includes policy specifications (from the user),
            production scenarios (from the previous module), and
                             *** DRAFT FINAL ***

-------
                                   A-4





                               EXHIBIT A-1


                      Flow of Analysis Program
                    Input
                    Files
                  Production
                  Scenarios
                   Module
                       1
                        I	
                    Policy
                 Alternatives
                   Module
                        i	:.
Year
Loop
Emissions
 Module
                        \
                        I	
                 I	
                 Atmospheric
                   Science
                   Module
                        |	
                   Effects
                   Module
                        I	
Summary
File
^^^


Report
Writer
Program


Output
File.
^ ^^
                             ***  DRAFT FINAL ***

-------
                                   A-5
            candidate technical options for controlling the use of
            potential ozone-depleting substances (identified in an
            input file).  The final results of this module are (1)
            the incurred costs of control attributable to the
            policy; and (2) the reductions in production realized as
            a consequence of control.

        •   Appendix D:  Emissions Module -- This module
            calculates global emissions of each compound by applying
            a release algorithm to the production scenarios (which
            have been modified by the Policy Alternatives Module).
            The release algorithm specifies the time it takes (after
            production) for the compound to be released for each end
            use of each compound.  The algorithm reflects the fact
            that the chemicals are stored in products for many years.

        •   Appendix E:  Atmospheric Science Module -- This
            module uses the global emissions estimates from the
            Emissions Module and estimates of atmospheric
            concentrations of carbon dioxide (C02), methane (CH4)
            and nitrous oxide (N20) to calculate global and
            latitudinal ozone depletion.  Ozone depletion is
            calculated using parameterized equations that represent
            one-dimensional and two-dimensional models of the
            atmosphere.  Uncertainty is also modeled based on
            investigations of uncertainty propagation through an
            atmospheric model.

        •   Appendix F:  Effects Module -- This module projects
            the effects that ozone depletion may have on human
            health and the environment.  Human health effects are
            focused on melanoma and non-melanoma skin cancers and
            cataracts.  Other effects include degradation of
            materials.

    The results of each module are written to a "Summary File."  This Summary
File is used by the Report Writer Program to create detailed tables as
specified by the user.

Report Writer Program

    The Report Writer Program reads in the data stored in the Summary File,
and creates reports for the model run as specified by the user.  The reports
for a given model run may show results disaggregated by region, compound, end
use, and time.  Results across runs may also be displayed and compared.  Of
note is that once a Summary File is created by the Analysis Program, the
Report Writer Program can be run" several times to generate reports at various
levels of aggregation without re-running the Analysis Program.
                             *** DRAFT FINAL ***

-------
                                   A-6
SPECIFYING A MODEL RUN

    To specify a model run the user fills out a set  of data files  that
describe the assumptions and data to be used.   The names  of these  data  files,
and their purposes, are listed in Exhibit A-2.   Each is described  in  turn.

        •   The Run File identifies the overall model flow,
            including which modules are to be executed.  This  file
            includes:

            --  the range of years for the model run;

            --  the names of modules to be executed;

                the names of the data files to be used;

                the years for which summary results  will  be
                retained; and

                the discount rates for cost/benefit  analysis.

        •   The Production Scenarios File specifies  the scenario
            of future production of potential ozone-depleting
            compounds to use in the analysis.   The user may choose
            one of a set of defined scenarios,  or may create his
            own.   The user also chooses (or creates)  a scenario of
            population and economic growth.

        •   The Policy Alternatives File defines the control
            strategy (if any) that the user is evaluating.   In
            addition to defining the level of controls or the  policy
            objective, the user must choose among alternative
            scenarios that define the technical options available
            for reducing the use of potential ozone-depleting
            compounds.

        •   The Atmospheric Science File allows the  user  to
            choose among alternative methods for computing
            ozone-depletion.  The user may specify a range of
            uncertainty surrounding the ozone-depletion estimate.
            The assumptions regarding the concentrations  of other
            trace gases (carbon dioxide, methane, and nitrous  oxide)
            are also specified in this file.

        •   The Effects File identifies the assumptions used to
            (1) estimate effects (in their natural units);  and (2)
            evaluate the effects in monetary terms.   The  user
            chooses (or supplies) dose-response coefficients for
            each effect.  The value (in dollars) of  each  effect may
            also be specified.
                             *** DRAFT FINAL ***

-------
                                   A-7



                               EXHIBIT A-2

                      Files  Required to Specify a  Run
        File Name
             Purpose
Run File
To identify modules included in
the run and data files to be
used, specify years for which
outputs are desired, and define
the discount rates used to
calculate costs and benefits.
Production Scenarios File
To choose a scenario of future
production for each compound and
of future population and economic
growth.
Policy Alternatives File
To define the control strategy to
be evaluated and the scenario of
available technical controls.
Atmospheric Science File
To choose a method for estimating
ozone depletion and specify a
range of uncertainty surrounding
the method.  Also to specify
scenarios of concentrations of
other trace gases.
Effects File
To specify the dose-response
coefficients for each human
health effect, the action
spectrum to be used for each
human health effect, and the
economic value to be placed on
each human health effect.  Also,
to specify the values used to
estimate polymer degradation.
Report Writer File
To identify runs for which
reports are to be written and the
reports desired.
                                 DRAFT FINAL ***

-------
                                   A-8
        •   The Report Writer File identifies the runs for which
            reports are to be written and the reports desired.

Of note is that once these files have been specified, and the model is run,
the entire model need not be re-run to perform additional analyses.  For
example, to analyze an alternative set of effects assumptions, only the
Effects File needs to be filled out, and only the Effects Module needs to be
executed.  The results from the previous modules can be stored for future
use.  This flexibility in the modular design of the model helps to facilitate
numerous comparison runs and reduces the time and expense of the analysis.

LIMITATIONS

    Although the model was implemented in a modular fashion to allow for
flexibility in development and use, it has several important limitations.

        •   Size.  The size of the model was kept manageable by
            limiting the units of analysis to 10 regions, 10
            compounds, and six end uses per compound.  More regions
            (e.g., for analyses of individual countries) cannot be
            accomodated easily.  The number of compounds and end
            uses per compound are also fixed.

        •   Feedback between compounds.  Each compound is
            analyzed independently.  The control of one compound
            (e.g., CFC-12) does not influence the use of other
            compounds.  These feedbacks may be analyzed by
            specifying alternative production scenarios that reflect
            the expected substitutions.

        •   Feedback between years.  Each year is analyzed
            independently, and feedbacks are handled by making
            previous years' data available through special data
            structures.  To minimize memory requirements, special
            data structures are established only for those data
            needed to model feedbacks.  Consequently, the addition
            of new feedbacks between years would likely require
            modifications to the data structure.
                             *** DRAFT FINAL ***

-------
                              APPENDIX  B

        SCENARIOS OF CHEMICAL PRODUCTION,  POPULATION, AND  GNP
INTRODUCTION

    This portion of the model  generates  scenarios of  (1) historical and future
production of potential ozone  depleting  substances,  (2) population, and (3)
economic activity (i.e., gross national  product, or GNP).  Ten potential
ozone-depleting compounds may  be specified  for analysis:  CFC-11; CFC-12;
CFC-22; CFC-113; CFC-114; CFC-115;  Carbon Tetrachloride (CC14); Methyl
Chloroform (CH3CC13);  Halon-1211;  and Halon-1301.  The production of each
compound is disaggregated by up to 10 regions of the world, and up to six
different end uses or  applications.   The population and GNP scenarios may also
be disaggregated by up to 10 regions  of  the world.

    The remainder of this appendix is organized as follows:

        •   Production Scenarios describes  the standard procedure
            for specifying scenarios  of  production (historical and
            future);

        •   Population and GNP Scenarios describes the
            specification of these scenarios;

        •   User-Modified Scenarios describes how the user may
            override portions  of the  standard procedures for
            specifying the production, population, and GNP
            scenarios; and

        •   Limitations presents the  limitations of this portion
            of the model.

PRODUCTION SCENARIOS

    Production scenarios may cover any period from 1931 to 2100.  Historical
data are used through 1984, and a variety of scenarios are specified to
reflect potential future production from 1985 onward.  To run the model, the
user may (1) choose an existing scenario included in  the Production Scenario
File,1 (2) create a new scenario in the  Production Scenario File,  (3) modify
an existing scenario using the procedure outlined below in the section on
user-modified scenarios, or (4) omit  the specification of production from the
model by specifying emissions  directly (see Appendix  D) or ozone depletion
directly (see Appendix E).  This section describes the format of the existing
    1  The five scenarios included in the Production  Scenario File are
described in Chapter 3.
                             *** DRAFT FINAL ***

-------
                                   B-2
scenarios specified in the Production Scenarios File.   These scenarios are
defined in three parts:  global production; regional shares; and end use
shares.  Each part is described in turn.

Global Production

    World production for each of the 10 potential ozone-depleting compounds is
reported in units of millions of kilograms per year.  Exhibit B-l shows the
middle production scenario for all eight compounds currently analyzed.
Intermediate values between the years reported are calculated by the model
(using either linear or exponential interpolation, as  specified by the user).
Production is assumed to be constant following 2050.

Regional Shares

    Regional shares are used to divide global production into the following 10
regions:

            U.S.;
        •   Canada (CAN);
        •   Western Europe (WEUR);
        •   Japan, Australia, and New Zealand (JANZ);
        •   USSR and East Bloc (EUSSR);
        •   Centrally Planned Asia (ASCEN);
            Middle East (HIDE);
            Africa (AFR);
        •   Latin America (LA); and
        •   South and East Asia (SEASIA).

Exhibit B-2 shows a scenario of regional shares for CFC-11.  The regional
shares are multiplied by the global production to produce regional production
estimates.  For example, to compute 2000 U.S. CFC-11 production, the model
multiplies CFC-11 global production for that year, 543.0 million kg, by the
regional share for the U.S. for that year, 0.212, to calculate 115.1 million
kg of production.  For each year, the shares add to one, indicating that all
global production is accounted for when disaggregating across regions.  Also
of note is that the regional shares may change over time (e.g., in this
scenario the U.S. share of world CFC-11 production shrinks from 73.7 percent
[0.737] in 1958 to 19.9 percent [0.199] by 1980).

End Use Shares

    An "end use share" is the fraction of regional production that is
allocated to a particular application or end use.  Each compound may have up
to six different end uses.  End use shares may vary by the 10 regions of the
world and may change over time.  The model currently includes the following
end uses:

        •   aerosol propellant;
        •   flexible urethane foam;
        •   rigid polyurethane foam;
                             *** DRAFT FINAL ***

-------
                                   B-3
                              EXHIBIT B-1

            Future Global Production Scenarios:  Middle Scenario
                          (millions of kilograms)
      CFC-ll   CFG-12   CFC-22  CFG-113  CC14   CH3CC13   Halon-1211   Halon-1301

1985    375.0    475.0  111.0    163.0    71.2    545.0      10.8         10.8

2000    543.0    688.0  161.0    283.0   103.0    848.0      19.7         19.7

2025  1,006.9  1,275.6  298.1    525.0   191.0  1,534.6      36.5         32.3

2050  1,867.0  2,365.0  552.0    974.0   354.2  2,777.0      67.8         53.1

2075  1,867.0  2,365.0  552.0    974.0   354.2  2,777.0      67.8         53.1
                             *** DRAFT FINAL ***

-------
                                   B-4
                                EXHIBIT B-2




                       Regional Use Shares  for CFC-11






Year  U.S.    CAN   WEUR   JANZ   EUSSR  ASCEN  HIDE    AFR    LA    SEASIA  TOTAL




Historical Data




1931  1.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000    1.000




1958  0.737  0.023  0.165  0.075  0.000  0.000  0.000  0.000  0.000  0.000    1.000




1965  0.577  0.037  0.263  0.120  0.001  0.000  0.000  0.000  0.000  0.000    1.000




1970  0.410  0.052  0.364  0.166  0.003  0.001  0.001  0.001  0.001  0.001    1.000




1975  0.333  0.057  0.406  0.185  0.007  0.003  0.002  0.002  0.003  0.002    1.000




1980  0.199  0.067  0.470  0.215  0.019  0.007  0.004  0.004  0.009  0.006    1.000






Scenario Of Future Regional  Shares




1985  0.211  0.035  0.251  0.114  0.149  0.057  0.034  0.034  0.069  0.046    1.000




2000  0.212  0.033  0.231  0.118  0.133  0.066  0.037  0.037  0.081  0.052    1.000




2025  0.212  0.030  0.214  0.116  0.113  0.069  0.046  0.042  0.097  0.061    1.000




2050  0.212  0.028  0.197  0.115  0.093  0.072  0.054  0..046  0.113  0.070    1.000




2100  0.212  0.028  0.197  0.115  0.093  0.072  0.054  0.046  0.113  0.070    1.000
                             *** DRAFT FINAL ***

-------
                                   B-5
        •   rigid nonurethane foam;
        •   refrigeration;
        •   solvent;
        •   fire extinguisher; and
        •   miscellaneous.

Exhibit B-3 shows an example of end  use shares for CFC-11 in the U.S.2   All
end use shares for a year add to one (i.e.,  all production is assigned  to some
type of end use).  In the module, end use shares are multiplied by estimated
regional production to estimate the  amount of production going into each end
use.  For example, U.S. 2000 CFC-11  production used to make rigid polyurethane
foam would be calculated as follows:
       543.0
    (2000 CFC-11
       Global
    Production)
    0.212
(2000 CFC-11
U.S. Regional
   Share)
   0.588
(2000 CFC-11
 U.S. Rigid
Polyurethane
Foam Share)
  67 . 7 million kg
(2000 CFC-11 U.S.
Rigid Polyurethane
    Foam Use)
    Using these three factors (global production,  regional shares,  and end use
shares), the historical and potential future production of each of  the 10
compounds can be specified in detail.  For ease of use, the model incorporates
a range of scenarios that reflect a large range of potential future growth in
production (0 to 5 percent annual growth).  The user may choose one of these
scenarios, create a new scenario, or modify an existing scenario as described
below in the section entitled User-Modified Scenarios.

POPULATION AND GNP SCENARIOS

    A series of population and GNP scenarios are provided in the model.3
The user has the option of choosing one of these scenarios, or creating his
own scenario.  Existing scenarios reflect a wide range of projected growth,
ranging from 0.5 percent to 1.1 percent growth per year (population from 1985
to 2075) and 1.0 percent to 3.6 percent growth per year for GNP (1985 to
2075).  The model does not use historical population or GNP in its  analysis.

    The population scenario is specified for each of 10 regions in  terms of a
base year value (1975) and a list of growth indices.  Each index value is
multiplied by the base year value to produce a population projection for a
given year for a region.  Exhibit B-4 presents the middle scenario  for the
U.S. population.  To compute the U.S. population projection for 2050, the base
year population for the U.S. (213,925,000) is multiplied by the index for 2050
(1.373), which results in a U.S. population projection of 293,719 in 2050.
This method allows for the rates of growth of population to vary by region,
and to change over time.
    2  The sources for the end use shares are described in Chapter 3.

    3  Scenarios were presented in Gibbs (1986).
                             *** DRAFT FINAL ***

-------
         EXHIBIT B-3

U.S.  End Use Shares for CFC-11
HISTORICAL DATA



SCENARIO OF FUTURE
END USE SHARES


1931
1958
1970
1980
1985
2000
2050
Aeroso I
Prope 1 lant
0.000
0.869
0.542
0.132
0.055
0.038
0.011
Flexi b le
Urethane
Foam
0.000
0.045
0.157
0.284
0.290
0.275
0.283
Rigid
Po lyurethane
Foam
0.
0.
0.
0.
0.
0.
0.
000
046
162
371
490
588
604
Rigid
Nonurethane
Foam
0
0
0
0
0
0
0
.000
.003
.009
.020
.026
.031
.031
Ref r iqerat ion
1.000
0
0
0
0
0
0
.011
.037
.064
.061
.053
.055
M i SCG 1 laneous Total
0.000
0.026
0.093
0.129
0.078
0.015
0.016
1 . 000
1 . 000
1 . 000
1 . 000
1 . 000
1 . 000
1 .000
                                                                                 CD
                                                                                 I
     *** DRAFT FINAL ***

-------
                B-7



             EXHIBIT B-4

   Middle U.S.  Population Scneario
                              Population
Year     Population Index     (thousands)

1975          1.000             213,925
1985          1.072             229,328
2000          1.192             254,999
2025          1.322             282,809
2050          1.373             293,719
2075          1.376             294,361
             EXHIBIT B-5

       Middle U.S. GNP Scenario
                                  GNP
                              (millions  of
Year        GNP Index         (1975  U.S.$)

1975          1.000            1,519,890
1985          1.348            2,048,812
2000          2.040            3,100,576
2025          3.242            4,927,483
2050          4.899            7,445,941
2075          7.830           11,900,739
          *** DRAFT FINAL ***

-------
                                   B-8
    Regional GNP scenarios are specified in a similar manner.   Exhibit B-5
shows the middle scenario for U.S. GNP.   The base GNP in 1975  is in 1975  U.S.
dollars and the indices correspond to projected years.   For example,  the  model
would calculate projected GNP for the U.S.  in 2050 by multiplying the base
year GNP, $1,519,890 million, by the corresponding index for the U.S. in  the
year 2050, 4.899.   This computation results in a projected 2050 GNP for the
U.S. of $7,445,941 million (in 1975 dollars).  GNP for other regions  is also
calculated in 1975 U.S. dollars.

USER-MODIFIED SCENARIOS

    The user has the option of overriding portions of the scenarios described
in the previous sections.  This capability expands flexibility by allowing
modified scenarios to be analyzed without creating complete scenario  data
files for each.  This option is especially useful for performing sensitivity
analysis.  These are six options for modifying the scenarios:

        •   Population -- The user may specify population
            projections over time for individual regions.  The
            specified production(s) override(s) the appropriate
            portion(s) of the standard scenario without changing the
            projections for other regions.

        •   GNP -- The user may specify GNP projections over time
            for individual regions, thereby overriding the
            projections from the standard scenario without changing
            the projections for other regions.

        •   Production by End Use -- The user may specify
            production (by end use) over time for individual
            regions.  Projections for end uses and regions not
            explicitly specified are drawn from one of the standard
            production scenarios.

        •   Production as a Function of Population -- The user
            may specify the production of each compound as a
            function of population for individual regions.

        • .  Production as a Function of GNP -- The user may
            specify the production of each compound in each region
            as a function of GNP.

        •   Production as a Function of GNP per Capita -- The
            user may specify the production of each compound in each
            region as a function of GNP per capita.  GNP per capita
            is calculated within the model from the GNP and
            population scenarios for each region.

    For each user-modified scenario the user must specify in a table the  type
of projection (population, GNP, etc.), the applicable region,  the compound (if
appropriate), and the end use (if appropriate).  For example,  to specify  2.0
                             *** DRAFT FINAL ***

-------
                                   B-9
percent annual growth in the production of CFC-11 in the U.S.  from 1985 to
2050, and zero growth thereafter, the user would fill out the  table shown in
Exhibit B-6.  The values in this table would override the values for CFC-11 in
the U.S. reported in the standard scenarios.

    Another example of a user-specified projection type would  be to have U.S.
CFC-11 production vary with U.S. population.  Exhibit B-7 shows an example of
how to specify this production versus population information.   The model
utilizes this information by (1) determining the level of population for that
region, and (2) identifying (by interpolation) the associated  production for
that population.  Because the information specifying the user-modified
scenarios is entered in the form of tables, the user is free to adopt any
arbitrary function or "shape" for his data.  This capability enhances the
ability to use the model for sensitivity analysis.

LIMITATIONS

    The primary limitations of this scenarios module are described below.

        •   Range of Standard Scenarios.   A set of standard
            (i.e., pre-defined) scenarios are included in the
            module, and the user may choose among them.  To perform
            an analysis with scenarios outside the range reflected
            by the standard scenarios, the user must define his own
            scenario or modify a standard scenario.  The limitation
            of this approach is that the scenarios are comprised of
            a significant amount of data, thereby making the
            establishment of an additional scenario somewhat
            cumbersome.  This limitation is alleviated to some
            extent by allowing the user to make selective
            modification to standard scenarios to allow sensitivity
            analysis.

        •   Sub-regional Scenarios.  Scenarios for areas other
            than the 10 regions in the model cannot be specified.

        •   Consistency.  The user is free to choose or create
            individual production, population, and GNP scenarios.
            In fact, the future production of the different
            compounds was generally projected under various
            assumptions regarding population and GNP growth.
            Therefore, the user should be careful to choose
            scenarios that are consistent with each other.  The
            model does not check for consistency among the scenarios
            chosen by the user (i.e., the user may choose a high
            production scenario and a low population scenario
            simultaneously).
                             *** DRAFT FINAL ***

-------
                   B-10
                EXHIBIT B-6

User-Modified Scenario Specifying a Growth Rate
 of 2 Percent Annually for CFC-11 in the  U.S.
       Year     Millions of kg of CFC-11

       1985                  78
       2000                 105
       2025                 172
       2050                 283
       2075                 283
              EXHIBIT B-7

 A USER-MODIFIED SCENARIO:   PRODUCTION
      AS A FUNCTION OF POPULATION
     Population               Production
 Millions of People     Millions of kg of CFC-11

        200                        50
        250                       100
        300                       200
             *** DRAFT FINAL ***

-------
                                  B-ll
REFERENCES

Gibbs,  Michael J.  (1986),  "Scenarios of CFC Use:  1985 to 2075," prepared for
    the U.S.  Environmental Protection Agency, Washington,  D.C.
                            *** DRAFT FINAL ***

-------
                               APPENDIX C

                  EVALUATION OF POLICY ALTERNATIVES
    This module evalutes  control policy alternatives.   If a policy  specified
by the model user requires  reductions in the production of potential
ozone-depleting substances,  the model simulates these reductions  by choosing
from a list of possible technical control options.  Each technical  control  is
defined by its cost  and the  percentage reduction that it can achieve.

    Because this risk assessment is focused on evaluating risks  in  the  absence
of regulation, this  module  is not employed in the analysis at this  time.
Therefore, this module is not described at this time.
                             ***  DRAFT FINAL ***

-------
                              APPENDIX D

          EMISSIONS OF POTENTIAL OZONE-DEPLETING COMPOUNDS
INTRODUCTION

    Emissions are estimated on a global  basis  for  each  compound being analyzed.
Two approaches are available for modeling emissions:

        •   Internal estimates -- Emissions  are modeled based on
            the production scenarios  described in  Appendix B and
            modified through the implementation of policies
            described in Appendix C.   A  release algorithm is used to
            describe how fast the compounds  are released into the
            atmosphere after production.   This algorithm is based on
            "release tables" which are specific for  each compound
            and end use.  Each release table shows the  annual
            fraction of the chemical  that is released over time.

        •   Exogeneous specification  --  Emission scenarios may be
            defined directly by the user.

INTERNAL ESTIMATES OF EMISSIONS

    The model calculates emissions from  release tables  and estimates of
production developed by the Production Scenario Module  (Appendix B) and
modified by the Policy Alternatives Module (Appendix C).  This section first
defines a release table, then describes  how  releases are estimated from one
year's production, and, finally, presents how  emissions are calculated from
production over a series of years.

Release Tables

    Compounds are released into the atmosphere at  different rates following
production, depending on their uses.   Some uses (such as aerosol propellant)
release compounds into the atmosphere very soon after production.  However,
some uses (such as refrigeration and  rigid foam production) store compounds
for many years, releasing them gradually over  time.  This storage is a called
a "bank."

    A release table represents the fraction  of production emitted over time
for a particular compound and end use.   Release tables  have been defined for
each of the end uses in the model based  on Quinn  (1986), and the user may
specify alternative release tables as desired. Exhibit D-l presents the
release tables for CFC-11 in its end  uses.  A  total  of  20 years is shown, and
the fraction of the production from year 1 released  into the atmosphere in the
following years is listed.  For example,  during the  year of production of
refrigeration uses 19.0 percent of the compound is released.  By the end of
the second year a total of 27.1 percent  of the production from year 1 is
released.  Releases continue through  the fourth year, after which all of the
                             *** DRAFT FINAL ***

-------
                                                           EXHIBIT D-1

                                                    Release Tables for  CFC-11
                                                       Rigid
Rigid
Aerosol Propel Innt
Cumulative Annual
Fraction Fraction
Year Released Released
1 1 . 000 1 . 000
2
3
4
5
6
7
8
9
10
1 1
12
13
14
15
16
17
18
19
20
Flexible Foam Polyurethane Foam Nonurethane Foam Refrigeration Miscellaneous
Cumulative Annual Cumulative
Fraction Fraction Fraction
Released Released Released
1.000 1.000 0.145
0.190
0.235
0.280
0.325
0.370
0.415
0.460
0.505
0.550
0.595
0.640
0.685
0.730
0.775
0.820
0.865
0.910
0.955
1.000
Annual Cumulative Annual Cumulative Annual Cumulative Annual
Fraction Fraction Fraction Fraction Fraction Fraction Fraction
Released Released Released Released Released Released Released
0.145 0.600 0.600 0.190 0.190 1.000 1.000
0.045 1.000 0.400 0.271 0.081
0.045 0.344 0.073
0.045 1.000 0.656
0.045
0.045
0.045
0.045
0.045 V
t-o
0.045
0.045
0.045
0.045
0.045
0.045
0.045
0.045
0.045
0.045
0.045
Source:   Quinn,  T.,  et al.  (1980),  "Projected  Use,  Emissions,  and  Bank of Potential  Ozone Depleting Substances," The
         RAND Corporation,  prepared for the  U.S.  Environmental  Protection Agency,  Washington,  D.C.

                                                       ***  DRAFT FINAL ***

-------
                                   D-3
production from year 1 is released.  Annual release rates may be calculated by
subtracting each year's cumulative fraction from the previous year's
cumulative fraction.  For example, 8.1 percent (27.1 minus 0.190) is released
in year 2, 7.3 percent (0.344 minus 0.271) is released in year 3.

    The release tables vary in length (some end uses store their compounds for
many years).  Also, the release table need not reflect complete emissions.  If
the cumulative release fraction in the table reaches 1.0, then all of the
compound that is produced is eventually released into the atmosphere.  If some
of the compound is expected to be retained forever within a product, or
destroyed, the release table would have as its highest value a number less
than 1.0.  The difference between the reported value and 1.0 would be the
portion of the compound that is never emitted.

Estimating Emissions From One Year's Production

    Exhibit D-2 shows the emissions from a hypothetical production of 100
million kg in 1985.  The cumulative fraction of production released over time
is presented in column 2.  The annual fraction of production released is shown
in column 3.  The model calculates annual emissions by multiplying the annual
fraction by the amount of production in the initial year (100 million kg in
this example).  The projected emissions per year are shown in column 4.  In
this example the total emissions sum to the initial production amount,
indicating that all of the production is eventually emitted into the
atmosphere.

Estimating Emissions From Production  Over a Series of Years

    Exhibit D-3 presents how the model calculates emissions from production
over a series of years.  Columns 1 and 2 show years and production in that
year, respectively.  In this example, production grows by 10 million kg per
year, and, for simplicity, emissions from production before 1985 are omitted
(emissions from past production are not omitted in the model).  Columns 3
through 8 show emissions over time due to the production shown in column 2,
each row showing how releases of its production are spread over the course of
years.

    The first row shows the emissions from 1985 production.  The second row
shows emissions from the .1986 production, and so on.  The total emissions for
a year are calculated by summing emissions values in the columns to compute a
total.   For example, emissions in 1985 are 19.0, emissions in 1986 are 29.0,
emissions in 1987 are 39.0, and so on.  By 1990, a total of 441.1 million kg
has been emitted, and a total of 750 million kg has been produced.  Therefore,
750 - 441.1 or 308.9 million kg remains to be emitted after 1990.

    This method of estimating emissions reflects the delays between the
production and release of many potential ozone-depleting substances.  The
implication of the method are that releases in a given year are caused not
only by production in that year, but also by production and use in previous
years.
                             *** DRAFT FINAL ***

-------
                           D-4
                        EXHIBIT D-2

                Emissions from a Hypothetical
        100 Million Kilograms of Production in 1985
 12                   3                  4
          Cumulative            Annual         Annual Emissions
Year   Fraction Released   Fraction Released   (millions  of kg)

1985         0.190               0.190               19.0

1986         0.271               0.081                8.1

198.7         0.344               0.073                7.3

1988         1.000               0.656               65.6

  TOTAL                          1.000              100.0
                     *** DRAFT FINAL ***

-------
                            D-5
                         EXHIBIT D-3

       Emissions from Production Over a Series of Years
                   (millions  of kilograms)
Year    Production    1985     1986     1987    1988    1989    1990
1985
1986
1987
1988
1989
1990
100
110
120
130
140
150
19.0 8.1 7.3 65.6
20.9 8.9 8.0 72
22.8 9.7 8
24.7 10
26
_- __ - - - - - -

.2
.8
.5
.6

--
--
78.
9.
11.
28.


7
5
3
5
TOTAL EMISSIONS
  PER YEAR            19.0    29.0     39.0     108.0   118.1   128.0
                      *** DRAFT FINAL ***

-------
                                   D-6
EXOGENOUS SPECIFICATION OF EMISSIONS

    The user has the option of  directly specifying the  level of emissions over
time.  This is accomplished by  filling out  an  emissions table; a sample table
is presented in Exhibit D-4.  This  table shows emissions estimates for eight
chemical compounds over a time  period of 90 years.  The module can interpolate
intermediate emissions values by either linear or exponential interpolation.
These specified emissions override  the estimates of emissions that the model
would have generated from production data as discussed  in the previous
section.  This ability to specify emissions directly to the model allows the
flexibility to evaluate the ozone depletion and other effects that may be
associated with given emission  levels, without having to specify a policy that
would result in such emissions.

LIMITATIONS

    The primary limitation of this  module is that the emission tables are not
designed to change over time.   Therefore, changes in technology that influence
emissions rates of existing end use products cannot be  modeled.
                             *** DRAFT FINAL ***

-------
                                D-7
                             EXHIBIT D-4

        SAMPLE TABLE OF  EXOGENOUSLY SPECIFIED EMISSIONS
                        (millions of kilograms)
Year  CFC-11  CFC-12  CFC-22  CFC-113  CC14  CH3CC13  Halon-1211 Halon-1301
1985
2000
2025
2050
298
462
897
1,650
438
668
1,237
2,281
81
151
279
516
138
240
446
828
71
103
191
355
500
721
1,303
2,356
3
9
25
48
3
9
22
38
                           *** DRAFT FINAL ***

-------
                                  D-8
REFERENCES

Quinn,  T.,  et  al.  (1986),  "Projected Use, Emissions, and Bank of Potential
    Ozone Depleting Substances," The RAND Corporation, prepared for the U.S.
    Environmental  Protection Agency, Washington, D.C.
                             *** DRAFT  FINAL ***

-------
                              APPENDIX  E

                     ATMOSPHERIC SCIENCE MODULE
INTRODUCTION

    This portion of the model computes  the  level of global average ozone
depletion as a function of the emissions  of the 10 potential ozone-depleting
substances.  The emissions are computed in  the previous module (presented in
Appendix D).  The user has three options  for estimating ozone depletion:

        •   Use a relationship between  emissions and ozone
            depletion developed to reflect  the results of a
            one-dimensional (1-D)  atmospheric model;

        •   Specify his own relationship; and

        •   Specify the level of ozone  depletion directly.

Each of these options is described in turn.   Then, the method used to reflect
uncertainty is presented.   Finally,  the limitations of the atmospheric science
module are presented.

PARAMETERIZED  RELATIONSHIP

    The module incorporates a series of equations that describe the change in
total column ozone as a function of:

        •   emissions of potential ozone-depleting substances; and

        •   abundances of three trace gases (C02 in ppra, N20 in
            ppb, and CH4 in ppm).

This relationship was developed by Connell  (1986), and it computes global
total column ozone change (in percent)  relative to 1985.

    To use this relationship, the user  must specify scenarios of trace gas
abundances to use in conjunction with the emissions of potential
ozone-depleting substances.  A range of scenarios has been developed from
which the user may choose (see Chapter  4),  or the user may specify his own
scenarios.  The reference case set of abundances used in the development of
the relationship is shown in Exhibit E-l.  Exhibit E-2 compares the results of
the LLNL 1-D model to the relationship  for  the base case of emissions of
ozone-depleting substances used by Connell  to develop the relationship.

    The following six steps are performed to compute global ozone depletion in
year t using the parameterization:

    1.  Compute the change in the chlorine  concentration in year t relative to
        1985 as:
                             *** DRAFT FINAL ***

-------
                     E-2
                 EXHIBIT E-1

    TRACE GAS ABUNDANCES USED TO  DEVELOP
     THE OZONE-DEPLETION  RELATIONSHIP*
Year
1985
1995
2005
2015
2025
2035
2045
2055
2065
2075
Carbon Dioxide
(ppm)
344.5
362.0
382.8
407.4
436.4
470.3
509.9
555.9
609.1
670.7
Nitrous Oxide
(ppb)
303.1
310.8
318.6
326.7
334.9
343.4
352.1
361.0
370.1
379.5
Methane
(ppm)
1.756
1.939
2.142
2.366
2.614
2.887
3.189
3.523
3.892
4.299
   * This  base case set of abundances was one of several
used to develop the relationship between emissions of
ozone-depleting substances and ozone depletion.

Source: Connell, Peter S. (1986),  "A Parameterized Numerical
        Fit to Total Column Ozone  Changes Calculated
        by the LLNL 1-D Model of the Troposphere and
        Stratosphere," Lawrence Livermore National
        Laboratory, Livermore, California.
               *** DRAFT FINAL ***

-------
                    E-3
                EXHIBIT E-2

COMPARISON OF TOTAL COLUMN OZONE DEPLETION
    RESULTS FROM THE 1-D MODEL AND THE
      PARAMETERIZED NUMERICAL FIT*
         Ozone Depletion (percent)
Year
1985
1995
2005
2015
2025
2035
2045
2055
2065
2075
Numerical Fit
0.0
-0.18
-0.74
-1.64
-2.86
-4.48
-6.63
-9.60
-14.0
-21.6
1-D Model
0.0
-0.18
-0.61
-1.42
-2.65
-4.30
-6.41
-9.10
-12.9
-19.9
Difference
0.00
0.00
0.13
0.22
0.21
0.18
0.22
0.50
1.10
1.70
   * This base case was one  of many cases used to develop
the relationship between emissions of ozone-depleting
substances and ozone depletion.  This result does not
necessarily reflect the base case estimates from this
model.

Source:  Connell, Peter S.  (1986), "A Parameterized Numerical
        Fit to Total Column Ozone Changes Calculated
        by the LLNL 1-D Model of the Troposphere and
        Stratosphere," Lawrence Livermore National
        Laboratory, Livermore, California.
              *** DRAFT FINAL ***

-------
                               E-4
        dCl(t) = dCl(») + I [CF(j)remissions(i,j)
                            * exp(-(t-i)/lifetime(j))
                            * (l-exp(-(-t-i)/raixing time))]

where:  dCl(t) equals the change in chlorine concentrations in year t;

    dCl(») equals the change in chlorine concentration in year t due
           to emissions prior to 1985;

    CF equals a conversion factor for each compund j;

    emissions are the annual emissions for compound j for years 1985
           to year t;

    lifetime equals the average atmospheric lifetime for each compound j;
           and

    mixing time equals the time it takes for source emissions to become
           well mixed.

2.  Compute the changes in C02, N20 and CH4 for year t as follows:

        dC02 = C02(t)/C02(1985);
        dN20 = N20(t)/N20(1985); and
        dCH4 = CH4(t)/CH4(1985),

where the values for C02, N20 and CH4 over time are model inputs as
discussed above.

3.  Compute ozone depletion due to CFCs as:

        d03(t) = 14.58[asinh(.332(21.05(l+(dN20-l)/(dCH4)2)
                            - dCl(t)*exp(-.15(dCH4-l))))
                            - asinh(.332*21.05(l+(dN20-l)/(dCH4)2))]

where d03(t) equals the ozone depletion in year t (in percent).

4.  Compute ozone depletion- (in percent) from C02, N20, and CH4 as follows:

        d03(t) = 3.61n(dC02);
        d03(t) = 3.75(dCH4-l); and
        .d03(t) = -7.0(dN20-l).

5.  Compute ozone depletion (in percent) from Halons as follows:

        d03(t) = -0.0302[concentration of Halon-1301]
                 -0.0618[concentration of Halon 1211]

6.  Compute total ozone depletion (in percent) as the sum of the
    individual estimates computed in Steps 3, 4, and 5.
                         *** DRAFT FINAL ***

-------
                                   E-5
These equations are used to represent  the potential  for ozone depletion as a
function of emissions and concentrations  of the  various substances analyzed.
The detailes of this method are described in Connell  (1986).

USER-SPECIFIED  RELATIONSHIP

    The user has the option of specifying his own  relationship between
emissions and ozone depletion.  For each  substance,  the user specifies a
"conversion factor" and a "lifetime."   These values  are used in the following
equation to compute an atmospheric abundance for each compound, over time:


                                       T                   -(T-t)/lifetime
    Abundance (T)  = Conversion Factor   I    [emissions (t)e              ].
                                    t=t
The abundances are summed across the compounds  to  create  a total abundance.
The amount of ozone depletion associated with each level  of abundance is then
specified by the user in a table.

    This formulation of the user-specified ozone-depletion relationship is
similar to the parameterized numerical fit developed  by Connell  (1986).  It
can produce abundances of the different compounds  that exhibit behavior
similar to the results produced by 1-D models.   The value of this option is
that the user may identify, in tabular form,  his best understanding of the
manner in which ozone depletion will vary with  the abundances of the compounds
in the atmosphere.

USER-SPECIFIED  OZONE DEPLETION

    The user has the option of bypassing the  computation  of ozone depletion
based on emissions of potential ozone-depleting substances.  Instead, the user
may provide estimates of global ozone depletion directly  to the model.  This
option allows the user to evaluate the impacts  of  such a  level of depletion
(possibly computed via other means) without identifying a level of emissions
or a control policy that would produce such depletion.

    To specify levels of ozone depletion, the user fills  out the table shown
in Exhibit E-3.  For each year, a level of total column ozone depletion (or
increase) is identified.  The model interpolates between  the values supplied
in the table.

UNCERTAINTY

    The user has the option of recognizing the  uncertainty inherent in current
models of ozone depletion by specifying a range within which the "true" level
of ozone depletion is expected to fall.  This range is specified using
multiplication scaling factors, as shown in Exhibit E-4.  The scaling factors
are multiplied by the estimate of ozone depletion  produced within the
                             *** DRAFT FINAL ***

-------
                          E-6
                       EXHIBIT E-3

HYPOTHETICAL TABLE OF USER-SPECIFIED OZONE DEPLETION*
                             Total Column
                         Global Ozone Depletion
                Year     	(percent)
1985
1990
2000
2025
2050
2075
2100
0.00
0.01
0.10
1.00
5.00
10.00
20.00
               * Values are  illustrative only.
                     *** DRAFT FINAL ***

-------
                    E-7
                 EXHIBIT E-4

EXAMPLE OZONE-DEPLETION SCALING FACTORS*
        Ozone Depletion     Scaling Factors
           (percent)         Low      High

             <15.0           0.4

             >20.0           0.5
           * These  factors represent the 10th and
        90th fractile  estimates from a lognormal dis-
        tribution of uncertainty developed from
        values  reported  in:  Stolarski, R.S., and
        A.R.  Douglass  (1986), Sensitivity of an
        Atmospheric Photochemistry Model to Chlorine
        Perturbation Including Consideration of
        Uncertainty Propagation, Draft Report to
        the U.S.  Environmental Protection Agency,
        Washington, D.C.
               *** DRAFT FINAL ***

-------
                                   E-8
model.1  These multiplications produce a range of ozone depletion values
for each year; the end points of the range (as well as the middle value
estimated within the model) are passed to the effects module for evaluation.

    As shown in Exhibit E-4, the scaling factors may change at different
levels of ozone depletion.  The values shown in the exhibit are based on a
Monte Carlo uncertainty analysis performed by Stolarski and Douglass (1985).
Each Monte Carlo iteration draws from a distribution of estimated reaction
rates and cross sections to generate predictions of atmospheric abundances.
These predicted abundances can be compared to actual measurements in the
atmosphere; those iterations that do not compare favorably with actual
measurements can be screened out.  The values reported in Exhibit E-4 are
based on those iterations that compared favorably with actual measurements and
were not screened out.

    Because the ozone-depletion models are not considered reliable at high
levels of ozone depletion, the user may also specify a "cap" on ozone
depletion.  If the computed level of ozone depletion (after scaling) exceeds
the user specified cap, then the ozone depletion used to evaluate effects (in
the next module) is reduced to the level of the cap.

LIMITATIONS

    The Atmospheric Science Module is a simplified representation of a complex
set of atmospheric reactions.  In using the module, the user should become
aware of the range of emissions and trace gas values for which the
parameterized relationship is considered valid.  Outside this range, the
equations do not necessarily reflect current estimates from 1-D models of the
atmosphere.  The range of production scenarios discussed in Appendix B and
Chapter 3 are well within the range for which these equations are valid.
    1 The scaling factors are not used when ozone depletion is specified
directly by the user.
                             *** DRAFT FINAL ***

-------
                                   E-9
REFERENCES

Connell, Peter S.  (1986),  "A Parameterized  Numerical Fit to Total Column Ozone
    Changes Calculated by  the LLNL 1-D Model  of  the Troposphere and
    Stratosphere," Lawrence Livermore National Laboratory, Livermore,
    California.

Stolarski, R.S., and A.R.  Douglass (1985),  "Sensitivity of an Atmospheric
    Photochemistry Model to Chlorine Perturbation  Including Consideration of
    Uncertainty Propagation," Draft Report  to the  U.S. Environmental
    Protection Agency, Washington, D.C.
                             *** DRAFT FINAL ***

-------
                               APPENDIX  F

        HEALTH AND ENVIRONMENTAL IMPACTS OF OZONE DEPLETION
INTRODUCTION

    This appendix presents the methods  used  to model the health and
environmental effects of ozone depletion  in  the CFC Policy Model.  This module
incorporates analyses performed by a variety of researchers, including
analyses of:

        •   the relationship between global  ozone depletion and
            latitudinal-dependent  ozone depletion (Isaksen 1986);

        •   the relationship between ozone depletion and
            ultraviolet radiation  (UV);l

        •   the relationship between UV and  the incidence (and
            mortality) of non-melanoma  skin  cancers;2

        •   the relationship between UV and  the incidence (and
            mortality) of melanoma skin cancers;3

        •   the relationship between UV and  the prevalence (and
            incidence) of senile cataracts;* and

        •   the relationship between ozone depletion and the costs
            of modifying polymers  to prevent degradation of polymer
            appearance and/or performance due to increased UV.5

    These analyses pertain primarily to the  U.S., consequently, this module
focuses on the risks in the U.S. associated  with ozone depletion.  The module
is designed to allow risk estimates to  be performed for other portions of the
    1 See Serafino and Frederick (1986)  for  a  description of the model used
to estimate the relationship between ozone depletion  and UV flux.

    2 See Chapter 7 for a discussion of  the  relationship between
non-melanoma skin cancer and UV.

    3 See Chapter 8 for a discussion of  the  relationship between melanoma
skin cancer and UV.

    * See Chapter 10 for a discussion of the relationship between senile
cataracts and UV.

    5 See Chapter 13 for a discussion of the relationship between polymer
degradation and UV.
                             *** DRAFT FINAL ***

-------
                                   F-2
world if:  (1) the model user supplies data needed to quantify effects in
these other regions; or (2) the model user assumes that the data for the U.S
applies to other portions of the world.   The descriptions below relate
exclusively to the analysis of impacts in the U.S.

    This appendix is divided into the following sections:

        •   Ozone Depletion by Latitude presents the method used
            to translate the estimates of global ozone depletion
            obtained from the Atmospheric Science Module (see
            Appendix E) into estimates of ozone depletion that vary
            by latitude;

        •   Changes in UV Flux presents the method used to
            estimate changes in the flux of UV for given levels of
            ozone depletion;

        •   Skin Cancers and Cataracts presents the methods used
            to model melanoma and non-melanoma skin cancers and
            cataracts;

        •   Materials presents the methods used to model the
            costs associated with the degradation of polymers; and

        •   Limitations summarizes the major limitations of the
            methods used, including the risks believed to be
            associated with increases in ambient levels of UV that
            are not modeled at this time in the CFC Policy Model.

OZONE  DEPLETION BY LATITUDE

    The Atmospheric Science Module (Appendix E) uses a parameterized numerical
fit to a 1-dimensional model of the stratosphere to estimate average global
changes in total column ozone on an annual basis (Connell 1986).
Two-dimensional models indicate that ozone changes due to increases in
stratospheric concentrations of potential ozone depleters are not likely to be
uniform across the globe (Isaksen 1986).  A strong latitudinal gradient is
expected with larger than average levels of depletion occurring at higher
latitudes (e.g., above 40 degrees north latitude).

    This portion of the module translates the estimates of average global
ozone change from the Atmospheric Science module into latitudinal-dependent
estimates based on analysis of data presented by Isaksen (1986).  A linear
relationship is used to estimate latitudinal ozone depletion as a function of
global ozone changes and changes in methane concentrations.  Methane
concentrations are used in the equation because the effects of methane on the
ozone column are latitudinally dependent.  The general form of the linear
relationship used in this module is:


                 °Zi = Ai,GOZGOZ + Bi,GOZGM + Ei,GOZ
                             *** DRAFT FINAL ***

-------
                                   F-3
where OZ.  = total column ozone depletion at latitude i,  GOZ = the average

global ozone depletion, GM = the average global methane  concentration,  and A,
B, and E are constants for each latitude i and range of  GOZ.   The average
global methane concentration may be used to estimate latitudinal effects
because the latitudinal methane concentrations show a high degree of
correlation with their corresponding global average.

    Note that in the above equation A,  B, and E vary by  average global  ozone
depletion as well as by latitude.   These coefficients have been estimated for
three ranges of ozone depletion.  In each range, the relationship described by
the above equation is nearly linear.  These ranges are:

        •    Global ozone depletion greater than three percent.
            Latitudinal results were found to be highly  correlated
            with the global average.

        •    Global ozone depletion greater than 1.5 percent but
            less than three percent.  Methane concentrations were
            found to have an important role in estimating
            latitudinal ozone changes.

        •    Global ozone depletion less than 1.5 percent
            (including ozone abundance).  No data existed for this
            range, linearity was assumed.

Range 1:   Global Ozone Depletion Greater Than 3.0 Percent

    Within this range, latitudinal ozone depletion was found to be highly
correlated with the global average of ozone depletion, across all the
latitudes.  Methane was found to have a negligible role  in this range (i.e., B
is zero in the equation).  Exhibit F-l shows the values  used for the global
ozone coefficient (A) and the intercept term (E) for this range of global
ozone values.  These coefficients were estimated using ordinary least squares
regression analysis with data from Isaksen (1986).  Exhibit F-2 compares the
results from Isaksen with those generated by the linear  relationship, using a
scenario where all CFCs grow at 3.0 percent per year. The global ozone
depletion values reported by Isaksen (shown at the top of the exhibit)  was
inserted into the equation using the coefficients listed in Exhibit F-l.  The
latitudinal values estimated with the equation are very  close to those
reported by Isaksen.

Range 2:   Global Ozone Depletion Greater Than 1.5 Percent
and  Less  Than 3.0  Percent

    Within this range, global average methane concentrations  (measured in ppm)
(in addition to average global ozone depletion) were found to have an
important effect on estimates of latitudinal depletion.   Coefficients for the
linear relationship (A, B and E) were determined using bivariate regressions
for each latitude.  Exhibit F-3 shows the coefficients obtained from these
regressions.  Exhibit F-4 compares the results from Isaksen with those
generated by the linear relationship, using a scenario where all CFCs grow at
3.0 percent per year.

                             *** DRAFT FINAL ***

-------
                F-4
           EXHIBIT F-1

  Coefficient used to Estimate Ozone
  Depletion  at Latitudes when Global
Average  Depletion Exceeds 3.0 Percent
   LATITUDE
60 N
50 N
40 N
30 N
20 N
10 N
0
10 S
20 S
30 S
40 S
50 S
60 S
1.32185
1.12836
0.96756
0.91823
0.92589
0.90047
. 0.88249
0.86839
0.81941
0.80351
0.85886
0.99591
1.12253
1.08857
0.60972
0.15419
-0.02698
-0.02349
-0.05004
-0.06526
-0.10497
-0.33460
-0.57032
-0.65536
-0.49032
-0.37618
          *** DRAFT FINAL ***

-------
                       F-5
                   EXHIBIT F-2
     Comparison of Estimate of Latitudinal  Ozone
    Depletion  when Global Ozone Depletion Exceeds
3.0 Percent:  Isaksen versus the Linear Relationship
    LATITUDE   YEAR
ISAKSEN
LINEAR MODEL
Global


60 N


50 N


40 N


30 N


20 N


10 N


0


2010
2020
2030
2010
2020
2030
2010
2020
2030
2010 .
2020
2030
2010
2020
2030
2010
2020
2030
2010
2020
2030
2010
2020
2030
3.73
5.75
8.73
5.90
8.68
12.50
4.77
7.11
10.39
3.76
5.76
8.59
3.41
5.31
7.98
3.46
5.34
8.07
3.34
5.16
7.82
3.25
5.05
7.64
..
--
--
6.02
8.69
12.62
4.82
7.10
10.46
3.76
5.72
8.60
3.40
5.26
7.99
3.43
5.30
8.06
3.31
5.13
7.81
3.23
5.01
7.64
                 *** DRAFT FINAL  ***

-------
                       F-6
                    EXHIBIT F-2
                     (continued)
     Comparison of  Estimate of Latitudinal Ozone
    Depletion when  Global Ozone Depletion Exceeds
3.0 Percent:  Isaksen versus  the  Linear  Relationship
    LATITUDE   YEAR
ISAKSEN
LINEAR MODEL
10 S


20 S


30 S


40 S


50 S


60 S


2010
2020
2030
2010
2020
2030
2010
2020
2030
2010
2020
2030
2010
2020
2030
2010
2020
2030
3.17
4.91
7.50
2.77
4.40
6.86
2.47
4.07
6.51
2.61
4.29
6.90
3.22
5.20
8.24
3.79
6.00
9.45
3.13
4.89
7.47
2.72
4.38
6.82
2.43
4.05
6.44
2.55
4.28
6.84
3.22
5.24
8.20
3.81
6.08
9.42
    Source:  Isaksen values reported in Isaksen,
    I. (1986), "Ozone Perturbations Studies in a
    Two-Dimensional Model with Temperature
    Feedback in the Stratosphere Included,"
    presented at UNEP Workshop on the Control of
    Chlorofluorocarbons, Rome, Italy, May 1986.
                 *** DRAFT FINAL ***

-------
                     F-7
                 EXHIBIT F-3

       Coefficients used to Estimate Ozone
  Depletion of Latitudes when Global Depletion
Exceeds 1.5 Percent and is Less Than 3.0 Percent
   LATITUDE
60 N
50 N
40 N
30 N
20 N
10 N
0
10 S
20 S
30 S
40 S
50 S
60 S
1.29733
1.14571
1.01811
0.98513
0.94876
0.89818
0.88837
0.89806
0.83801
0.82972
0.84437
0.91496
1.00523
1.14576
0.47717
-0.10520
-0.30283
-0.23606
-0.17885
-0.13225
-0.24961
-0.48130
-0.67765
-0.55604
-0.14140
0.13850
-1.46820
-0.56695
0.19535
0.43503
0.39426
0.32953
0.20192
0.37560
0.66795
0.89665
0.67525
0.12097
-0.37618
               *** DRAFT FINAL ***

-------
                   F-8
               EXHIBIT F-4
Comparison of Estimates of Latitudinal Ozone
Depletion  when Global Ozone Depletion  Exceeds
  1.5 Percent and is Less than 3.0 Percent:
   Isaksen versus the Linear Relationship
LATITUDE   YEAR
ISAKSEN
LINEAR MODEL
Global
60 N
50 N
40 N
30 N
20 N
10 N
0
10 S
20 S
30 S
40 S
50 S
60 S
2000
2000
2000
2000
2000
2000
2000
2000
2000
2000
2000
2000
2000
2000
2.33
3.85
3.05
2.33
2.11
2.15
2.10
2.01
1.98
1.66
1.46
1.52
1.94
2.33
..
3.90
3.08
2.36
2.12
2.13
2.06
2.00
1.96
1.64
1.45
1.51
1.97
2.36
             *** DRAFT FINAL ***

-------
                                   F-9
Range 3:   Global  Ozone Depletion Less Than 1.5 Percent

    Data from Isaksen (1986) did not cover this range of ozone depletion
extensively.  To estimate the coefficients, it is assumed that a value of zero
global ozone depletion implies zero latitudinal ozone depletion (rather than
some abundance and some depletion.  This assumption implies that in this range
of global depletion, both the methane coefficient (B) and the intercept term
(E) in the linear relationship must be set to zero, and latitudinal ozone
depletion is obtained by multiplying the average global ozone depletion by a
constant.   Exhibit F-5 shows the coefficients used by the model for this
depletion range.  The values were estimated as the ratio of latitudinal to
global depletion values from Isaksen1s data.  By averaging over all results in
each latitude, a latitudinal-dependent scaling factor was be calculated.
Exhibit F-6 compares the results from Isaksen with those generated by the
linear relationship, using a scenario where all CFCs grow at 3.0 percent per
year.

    The module used the relationships described above to estimate ozone
depletion by latitude.  The global ozone depletion and methane concentrations
are estimated in the Atmospheric Science Module (Appendix E).  The equation
for the global ozone depletion ranges are used as appropriate.  Because these
equations were developed from a limited data set from a single 2-D model, the
latitudinal estimates should be viewed as uncertain.  In particular, the
estimates of the effects of methane concentrations may be valid for a fairly
narrow range of potential future values because the estimates of the
coefficients are based on a single methane concentration scenario.  The
equation and their coefficients do not necessarily correspond to specific
physical properties and processes.

CHANGES  IN  UV  FLUX

    This portion of the module evaluates the expected changes in UV flux
associated with changes in total column ozone.  The estimates of changes in UV
flux vary by latitude because:

    1.  the change in UV flux (measured in percent) for a given
        change in total column ozone (also measured in percent)
        varies as a function of the initial total abundance of ozone
        (i.e., the level of ozone abundance in the absence of ozone
        depletion); and

    2.  the initial level of total column ozone varies by latitude.

    The relationship between ozone change and UV flux will also vary for
different measures of UV energy.  Three measures of UV, also referred to as
action spectra, are currently incorporated into the model:  Robertson-Berger
Meter (RB-Meter); DNA Action Spectrum; and Erythema Action Spectrum.  Each
measure has a unique weighting of the wavelengths in the UV spectrum.  Because
ozone selectively screens out radiation at given wavelengths, the different
weighting schemes will produce different estimates of increased radiation
levels when total column ozone abundance changes.
                             *** DRAFT FINAL ***

-------
                           F-10
                       EXHIBIT F-5

        Coefficients Used  to Estimate Ozone Depletion
at Latitudes when Global Depletion is Less  than  1.5 Percent
                  LATITUDE
                    60 N        1.87118
                    50 N        1.40759
                    40 N        0.99645
                    30 N        0.85819
                    20 N        0.87931
                    10 N        0.86197
                     0          0.83994
                    10 S        0.80330
                    20 S        0.62353
                    30 S        0.50484
                    40 S        0.54594
                    50 S        0.81227
                    60 S        1.02931
                     *** DRAFT FINAL ***

-------
                      F-ll
                  EXHIBIT F-6

           Comparison of Latitudinal Ozone
 Depletion  when Global Ozone Depletion  is Less Than
1.5 Percent:   Isaksen versus  the Linear Relationship
LATITUDE
Global
60 N
50 N
40 N
30 N
20 N
10 N
0
10 S
20 S
30 S
40 S
50 S
60 S
YEAR
1980
1990
1980
1990
1980
1990
1980
1990
1980
1990
1980
1990
1980
1990
1980
1990
1980
1990
1980
1990
1980
1990
1980
1990
1980
1990
1980
1990
ISAKSEN
0.65
1.33
1.13
3.85
0.86
1.80
0.62
1.34
0.60
1.20
0.65
1.23
0.64
1.21
0.62
1.19
0.62
1.15
0.44
0.91
0.36
0.76
0.36
0.79
0.50
1.10
0.64
1.33
LINEAR MODEL

1.22
3.90
0.92
1.88
0.65
1.33
0.56
1.15
0.57
1.17
0.56
1.15
0.55
1.12
0.52
1.07
0.41
0.83
0.33
0.67
0.35
0.73
0.53
1.08
0.67
1.37
                *** DRAFT FINAL ***

-------
                                   F-12
    The UV Model developed by Serafino and Frederick (1986) was used to
evaluate the expected changes in UV flux for given levels of total column
ozone change.  As described above in Appendix A of Chapter 7,  the estimates of
UV flux from the UV model reproduce field measurements taken with RB-meters.
The correlation between annual RB-meter values from the UV model and field
instruments is 0.979.  Pitcher (1986) performed a more complete validation of
the UV model, comparing monthly model estimates to monthly field
instrumentation readings for three cities:  El Paso, San Francisco, and
Minneapolis.  Although the model was found to overestimate UV flux in the
months of September through March, and underestimate UV flux in May, June, and
July, the annual estimates are generally within 10 percent of the RB-meter
readings in the field.

    Changes in UV flux (as measured by the three action spectra) as a function
of changes in ozone abundance were evaluated using the UV model for 24 cities
across the U.S.  These cities are divided into three regions of the U.S.:
north, middle, and south (see Exhibit F-7).  These three regions are used to
evaluate potential changes in the incidence and mortality due to skin
cancers.  The states included in each region are shown in Exhibit F-8.  The
population weighted average latitudes for the three regions are also shown.

    Exhibit F-9 shows the percent change in UV (measured with the DMA action
spectrum) in the three regions of the U.S. associated with a range of percent
changes in total column ozone (from a 10 percent increase in ozone (listed as
a -10 percent ozone depletion) to a 30 percent ozone depletion).  As expected,
the change in UV for a given level of ozone change is not constant across the
three regions of the U.S., although they are very similar.  Statistical
analysis indicates that the estimates for the three regions are statistically
different from each other.

    Similar tables have been developed for the RB-Meter and the Erythema
action spectra.  The values for the Erythema Action Spectrum are approximately
80 to 85 percent of the DNA Action Spectrum values; the RB-Meter values are 35
to 45 percent.  The user may define additional action spectra and prepare
additional tables that relate changes in UV flux to changes in ozone depletion
if he wishes to evaluate effects as a function of different action spectra.

SKIN CANCERS AND  CATARACTS

    The objective of this portion of the module is to evaluate the potential
changes in risks of melanoma and non-melanoma skin cancers and cataracts
associated with changes in total column ozone abundance over time.  The
purpose of the evaluation is to provide information for assessing the
implications of alternative emissions rates of potential ozone modifying
substances.  The inputs to this portion of the module include estimates of
changes in UV flux associated with changes in total column ozone abundance.
The output of this portion of the module includes estimates of the increased
number of cases and deaths associated with skin cancers over time and an
estimate of the increased number of cases of cataracts.
                             *** DRAFT FINAL ***

-------
                                F-13
                           EXHIBIT F-7

              Cities Used to Evaluate Changes in UV  Flux
                  for the Three Regions of the U.S.
REGION 1:  NORTH             REGION  2:  MIDDLE           REGION 3:  SOUTH
  New York                   Chicago                     Los Angeles
  Detroit                    Philadelphia                San Diego
  Milwaukee                  Baltimore                   Houston
  Boston                     San Francisco               Dallas/Fort Worth
  Seattle                    Washington                  Phoenix
  Minneapolis                Denver                      New Orleans
  Portland                   Salt Lake City              Miami
  Buffalo                    Kansas City                 Atlanta
                          *** DRAFT FINAL ***

-------
                                F-14
                           EXHIBIT F-8

       States Includes  in the Three Regions of the U.S.
REGION 1:  NORTH
REGION 2:  MIDDLE
REGION 3:  SOUTH
  Alaska
  Connecticut
  Idaho
  Maine
  Massachusetts
  Michigan
  Minnesota
  Montana
  New Hampshire
  New York
  North Dakota
  Oregon
  Rhode Island
  South Dakota
  Vermont
  Washington
  Wisconsin
  Latitude = 43.3 N
California (N) a/
Colorado
Delaware
District of Columbia
Illinois
Indiana
Iowa
Kansas
Kentucky
Maryland
Missouri
Nebraska
Nevada
New Jersey
North Carolina
Ohio
Oklahoma
Pennsylvania
Tennessee
Utah
Virginia
West Virginia
Wyoming

Latitude = 39.1 N
Alabama
Arizona
Arkansas
California (S) a/
Florida
Georgia
Hawaii
Louisiana
Mississippi
New Mexico
South Carolina
Texas
Latitude = 31.8 N
a/  California is divided in half,  one half being included in the  Middle
    Region, and one half included in the South Region.

Source:  Latitude estimates based on population centroids  for each state
         from the 1980 U.S. census, Master Area Reference  File #2,
         Geography Section, U.S.  Bureau of Census,  Department of Commerce.
                          *** DRAFT FINAL ***

-------
                   F-15
               EXHIBIT F-9

Percent Change in UV as a  Function of Change in
    Ozone Abundance for Three U.S. Regions
             (DNA Action Spectrum)
OZONE DEPLETION (%)
-10
-5
-2
0
2
5
10
20
30

North
-17.3
-9.3
-3.8
0.0
4.2
10.8
22.9
53.8
96.0
CHANGE IN UV (
Middle
-17.2
-9.1
-3.8
0.0
4.3
10.6
22.8
53.2
94.8
'0/\
,/oJ
South
-16.7
-8.9
-3.8
0.0
4.2
10.5
22.2
51.0
90.4
 Source:   Based on analyses using the UV Model
          developed by Serafino and Frederick (1986).
             *** DRAFT FINAL ***

-------
                                   F-16
    This analysis is organized into four parts as follows:

        1.  characterize the population at risk;

        2.  estimate the baseline incidence of the conditions being
            evaluated;

        3.  apply dose-response relationships to estimate the change
            in the risk of the conditions as a function of changes
            in ozone abundance over time; and

        4.  summarize the results.

    Data have been developed primarily for the three regions of the U.S.  for
purposes of evaluating these skin cancer and cataract risks (see above for the
definition of the three regions in the U.S.).  The risks of these cancers in
other portions of the world can be evaluated by the model user if:  (1) the
user supplies data to characterize the populations at risk in the other
regions, the baseline incidence in the other regions, and the dose-response
relationships that apply to the other regions; or (2) the user assumes that
the information provied for the U.S. applies to the other regions.  Each of
the four steps is presented for the U.S.

Step 1  -- Characterize the Population at Risk

    The incidence of skin cancers and cataracts has been found to vary by
race, age, sex, and the ambient level of UV in the area of residence of the
individual.6  The population in the U.S. is characterized by age, sex, and
race using data from the U.S. Bureau of the Census.7  Each of the three
regions in the U.S. are characterized separately.

    For each region, the current and expected age distribution of the
population is defined in a series of tables as shown in Exhibit F-10.  These
age distributions were obtained by summing the age distributions of all the
states in the region (in this case the North region), and dividing by the
total population in the region.

    The age and regional distributions of the U.S. population are expected to
change over time due to migration within the U.S. and due to changing survival
and birth rates.  These changes in the distribution through 2000 are included
in the characterization of the population, as shown in Exhibit F-10.  Changes
beyond 2000 have not been estimated, and consequently the age and regional
    6 See Chapter 7 for a discussion of non-melanoma skin cancer, Chapter 8
for a discussion of melanoma skin cancer, and Chapter 10 for a discussion of
cataracts.

    7 The Census Data were obtained from NPA Data Services.
                             *** DRAFT FINAL ***

-------
                                   F-17
                              EXHIBIT F-10

                   Age  Distribution of the  U.S. Population
                       Over Time in  the  North Region
                 (fraction of  total population in  the  region)
                                AGE GROUP
FRACTION OF
YEAR   0-14  15-24  25-34  35-44  45-54  55-64  65-74  75-84    85+  TOTAL U.S. POP.*


1985   .2074 .1667  .1842  .1288  .0892  .0947  .0716  .0385   .0189       .2486

2000   .2119 .1299  .1314  .1645  .1425  .0892  .0642  .0445   .0219       .2439
* The fraction of the total U.S.  population represents  the portion  of  the  U.S.
population that is estimated to reside in this  region (in this  case the  North
region).  This fraction changes over time to reflect expected migration  patterns.

Source:  Based on analyses of data supplied by  NPA Data Services, Inc.
                             *** DRAFT FINAL ***

-------
                                   F-18
distributions are assumed to remain fixed beyond 2000.   Because the U.S.
population may continue to become older on average into the 21st century,  this
fixed population distribution under-estimates the population of older
individuals, and may consequently under-estimate risks  of ozone depletion.
The user may wish to relax this assumption by inserting alternative data  for
years beyond 2000 as appropriate.

    The portions of the population in each region that  are male and female,
black and while are also defined.  These portions may change over time within
the model to reflect changing demographic patterns.

    From these data the fraction of the U.S.  population is characterized  by
age, race, and sex for each of the three regions of the country.  The total
U.S. population over time is provided as an input to the Scenarios Module (see
Appendix B), and is used as the basis of estimating the number of people  in
each age, race, and sex grouping.

Step 2 -- Baseline Incidence

    To evaluate the potential future risk of skin cancers in the U.S. the
baseline risk (i.e., the risk in the absence of changes in the level of total
column ozone) is required.  Incidence of melanoma and non-melanoma skin
cancers has been found to vary by latitude.  Therefore, baseline incidence
values have been developed for each of the three U.S. regions separately.   The
age-sex specific baseline rates used for melanoma and non-melanoma are
described separately.  The baseline data for cataracts  follow.

    Non-melanoma Skin Cancer

    There are two major types of non-melanoma skin cancers:  basal cell and
squamous cell (Scotto, Fears, and Fraumeni 1981; and see Chapter 7).  Because
the two types may respond differently to changes in UV, their baseline
incidences are estimated separately (Scotto,  Fears, and Fraumeni 1981).  Data
describing the age-sex specific incidence for whites are available for 10
locations from Scotto, Fears, and Fraumeni (1981).  The incidence rates by age
and sex were estimated for each of the three U.S. regions by taking the
population-weighted averages of the rates for the cities reported in Scotto,
Fears, and Fraumeni (1981) that are in each of the regions.  The locations in
each region are:

        North:  Seattle, Minneapolis-St. Paul, Detroit;
        Middle:  Utah, San Francisco, Iowa; and
        South:  Atlanta, New Orleans, New Mexico, Dallas-Forth Worth.

Exhibit F-ll displays an example of the age specific incidence rates for basal
and squamous cell skin cancers in the North region of the U.S.  These data
apply only to whites.

    The baseline incidence rates for Blacks and members of other races are not
well characterized.  The rates are believed to be approximately one order of
magnitude or more below the rates for whites, and are consequently not
                             *** DRAFT FINAL ***

-------
                                                        EXHIBIT  F-11

                                      Baseline  Incidence  for Non-Melanoma  Skin  Cancers
                                    (North  Region of  the  U.S.  — White  population  only)
                                                     (Rate  per 100,000)

NONMELANOMA SKIN CANCER
Basal Cell Skin Cancer
Ma le
Fema 1 e
Squamous Cell Skin Cancer
Ma le
Fema le
AGE GROUP
15 15-24 25-34 35-44 45-54 55-64 65-74


0.1 2.9 22.1 91.1 259.2 465.8 761.0
0.5 5.6 22.2 91.0 201.8 287.4 465.9

0.2 0.3 1.6 7.4 32.6 87.4 147.4
0.0 0.1 1.4 4.1 10.5 27.5 54.8

75-84 85+


1 162.8 1311.3
638.2 754.1

349.7 431.8
112.5 167.7
Source:   Derived from Scotto,  Fears,  and  Fraumeni  (I98I).
                                                   * *  DRAFT  FINAL * * *

-------
                                   F-20
expected to play a major role in evaluating risks of skin cancers in the U.S.
(they may play a more important role in the risks in other countries).   For
purposes of this analysis, risks for Blacks and members of other races  are not
evaluated.

    The mortality associated with each of these cancers is believed to  be
quite low.  Approximately 1.0 percent of all cases of nonmelanoma skin  cancer
result in death (Scotto, Fears, and Fraumeni 1981).  The rates for basal and
squamous cell cancer differ, and have been estimated from data in Scotto,
Fears and Fraumeni to be:  basal cell -- 0.3 percent; and squamous cell --
3.75 percent.  These percentages are multiplied by the baseline incidence
rates to estimate the baseline rates of death associated with each cancer type
over time.  The model user may modify these mortality values to explore the
implications of alternative assumptions.

    The model user may also specify that the baseline incidence and mortality
ratio may change over time (increase or decrease).  The purpose of this
specification is to allow the user to include the implications of potential
changes in the baseline incidence of these skin cancer that are not related to
changes in UV flux.

    Melanoma Skin Cancer

    There are four principal classes of melanoma skin cancers:  Hutchinson's
melanotic freckle, superficial spreading melanoma, nodular melanoma and
unclassifiable melanoma (Elder et al. 1980).  For purposes of modeling, the
baseline incidence of these cancers, two groups are defined that are based on
the location of the cancer on the body:

        •   Group 1: Face, Head, Neck, and Upper extremities; and
        •   Group 2: Trunk and Lower Extremities.

These groups have been defined by Scotto and Fears (in press), and incidence
data have been collected for seven U.S. cities that divide the occurrence of
melanoma skin cancers into these groups.  The division of these cancers into
these groups appears to be warranted because Scotto and Fears have identified
different correlations with UV for each of the two groups.

    The incidence data for these groups of melanoma skin cancer were developed
for the three regions of the U.S. by taking population weighted averages of
the incidence data reported for the following seven locations in the three
regions:8
    8 Although the National Cancer Institute SEER Report (1984) reports data
for ten locations, incidence data for only seven locations were used.  These
seven locations correspond to the locations for which Scotto and Fears (in
press) reported body site percentages (face, head, neck, and upper extremities
vs. trunk and lower extremities) for melanoma incidence.  The three locations
that were not included in this analysis are:  Connecticut, Hawaii, and Puerto
Rico.
                             *** DRAFT FINAL ***

-------
                                   F-21
        •   North:  Seattle, Detroit;
        •   Middle:  Utah, San Francisco, Iowa; and
        •   South:  Atlanta, New Mexico.

The incidence data were taken from the National Cancer Institute,  Cancer
Incidence and Mortality in the United States:   SEER Report (1984), and the
data apply to whites only.  Because the SEER data do not divide the incidence
by the two groups of melanoma skin cancer defined above, the fractions of age
specific incidence associated with each type of cancer were taken from Scotto
and Fears (in press).

    Exhibit F-12 displays sample baseline incidence data for melanoma skin
cancer by sex and age for whites in the North U.S. region.  As shown in the
exhibit, the incidence rates for the two types of melanoma were estimated
separately.

    Mortality data were taken from the SEER Report (1984) for whites only.
Because the SEER data do not divide the mortality rates by the two groups of
melanoma skin cancer defined above, mortality is modeled without dividing the
data into the two cancer types.  Exhibit F-13 displays the mortality rates
currently used in the model.

  •  The incidence and mortality associated with melanoma skin cancers in
nonwhite populations in the U.S. are less well defined.  The rates for
nonwhites are much lower than the rates for whites, and are consequently not
modeled.

    The model user may also specify that the baseline incidence rates may
change over time (increase or decrease).  The purpose of this specification is
to allow the user to include the implications of potential changes in the
baseline incidence of melanoma skin cancer that are not related to changes in
UV flux.

    Senile Cataracts

    The prevalence of senile cataracts has been found to be correlated with UV
radiation (Killer, Sperduto, Ederer 1983).  The baseline prevalence, however,
is not well characterized by the three regions of the U.S.  Therefore, the
prevalence (and associated incidence) of senile cataracts is modeled for the
entire U.S. as one region.

    Several diagnostic criteria have been used to define the occurrence of
senile cataracts.  The definition used in this module was reported by Leske
and Sperduto (1983) and used by Killer, Sperduto and Ederer (1983) to estimate
a multivariate logistic risk function that includes UV-B.  The diagnostic
criteria include:  cataract, defined as "senile lens changes consistent with
best corrected visual acuity of 20/30 (6/9) or worse;" or aphakia, defined as
"the lens had been surgically removed and there was no history of congenital,
traumatic or secondary cataract" (Killer, Sperduto, and Ederer 1983).  Exhibit
F-14 displays the baseline prevalence values used in the module.  Baseline
incidence (as well as increased incidence due to ozone depletion) is estimated
by computing the rate of change of prevalence overtime.


                             *** DRAFT FINAL ***

-------
                                   F-22
                                EXHIBIT F-12

                Baseline  Incidence for Melanoma  Skin  Cancers
             (North Region of the U.S.  --  White Population Only)
                             (Rate per 100,000)
                                              Age Group
Melanoma Skin Cancer  10-14  15-24   25-34   35-44  45-54  55-64  65-74  75-84  85+
Face, Head, Neck and
Upper Extremities
    Males              0.0    0.8     2.4    5.3    5.5    8.4   10.4    9.2    9.1
    Females            0.1    0.9     3.2    4.2    5.3    5.4    4.5    4.6    4.9
Trunk and Lower
Extremities
Males
Females

0
0

.0
.1

1.2
1.4

3.4
5.3

7.5
6.9

7
8

.9
.6

12.0
8.9

14.9
7.4

13.2
7.4

13.0
8.0
Source:  Derived from Scotto and Fears  (in press) and National Cancer Institute
         SEE Report (1984).
                             *** DRAFT FINAL ***

-------
                                   F-23
                                EXHIBIT F-13

                  Mortality Rates for Melanoma  Skin  Cancers
                           (White  Population Only)
                             (Rate per 100,000)
                                      Age  Group
         10-14   15-24   25-34   35-44  45-54    55-64   65-74   75-84   85+
Male      0.0     0.5     1.0     2.9      4.6      5.5     8.0     7.7    10.3

Female    0.0     0.0     0.8     1.6      2.8      2.9     3.3     5.3    5.5
Source:  Derived from Scotto and Fears  (in press)  and National Cancer
         Institute SEE Report (1984).
                             *** DRAFT FINAL ***

-------
                         F-24
                      EXHIBIT F-14

         Baseline Prevalence of Senile Cataracts
                   (Rate per 100,000)


Male
Female

<54*
0.0
0.0

55-64
4,300
4,700
AGE
65-74
16,000
19,300

75-84
40,900
48,900

85+
40,900
48,900
* Prevalence for individuals under 54 years of age assumed
to be zero.

Source:  Leske C.L. and R.D. Sperduto (1983),  "The
         Epidemiology of Senile Cataracts:   A  Review,"
         American Journal of Epidemiology,  Vol.  118,
         No. 2, pp. 152-165.
                   *** DRAFT FINAL ***

-------
                                   F-25
Step 3 -- Apply Dose-Response Relationships

    The purpose of applying the dose-response relationships is to estimate how
the risks of melanoma and non-melanoma skin cancers and cataracts may change
as the flux of UV reaching the earth's surface changes due to changes in total
column ozone abundance.  Two factors are required to apply dose response
relationshps:   (1) a measure of exposure; and (2) a dose-response equation.
Each is described in turn.

    The measure of exposure used to model skin cancers and cataracts is
related to the level of ambient UV flux estimated to reach the earth's surface
in the three regions of the U.S.  To use the dose-response relationships
incorporated into the model, the change in UV flux (in percent) associated
with changes in ozone abundance is used -- the change in the absolute level of
UV energy reaching the earth's surface (in units such as millijoules per meter
squared) is not used.  (The method used to estimate changes in UV flux (in
percent) along several scales of UV measurement was described above.)

    Given these estimates of percent changes in UV over time for given
locations in the U.S. (i.e., the three regions), the change in exposure for
individuals estimated to live in those areas is estimated on an age specific
basis.  In year T of the analysis, the change in UV flux for a person who is
40 years old is a function of the change in UV estimated from years T-40 to
year T.  Changes in UV for years prior to 1985 are assumed to be zero.

    The current understanding for non-melanoma skin cancer is that the
cumulative exposure received during a person's life is the appropriate measure
to use for modeling the dose-response relationship.  Insufficient evidence
exists to define the appropriate measure of exposure for purposes of modeling
cataract prevalence.  Similarly, a major question remains regarding the most
appropriate measure of exposure for purposes of modeling the incidence and
mortality of melanoma skin cancer.  Recent analyses of melanoma incidence by
Scotto and Fears (in press) have used annual RB-meter values as estimates of
exposure.  Pitcher (1986) used peak values to model mortality.  Although some
evidence indicates that peak or intermittent exposures to UV may best explain
melanoma incidence patterns (see Chapter 8), significant research remains to
be done on this issue.  Of note is that the UV Model indicates that changes in
UV flux (in percent) as a function of changes in ozone levels are about the
same for the annual values and peak values for a given action spectrum.
However, the choice of peak UV flux versus annual UV flux (or flux measured
over some other time period) influences the estimate of the relationship
between mortality risks and UV significantly (by as much as 60 percent).
Therefore, the choice of annual or peak UV flux as an exposure estimate for
purposes of modeling risk influences the risk estimates substantially.

    A related issue regarding modeling exposure for purposes of estimating
melanoma risks is the relative weight to put on exposures (annual or peak)
experienced throughout a person's life.  Migration studies indicate that
exposure before the age of 15 may be most important and should be given more
weight than subsequent exposures  (see Chapter 8).  These relative weights have
yet to be quantified, however.  To allow alternative assumptions regarding the
relative importance of exposure during a person's life to be modeled, relative


                             *** DRAFT FINAL ***

-------
                                   F-26
weights can be assigned to different ages during a person's life for purposes
of estimating changes in exposure.  Exhibit F-15 presents a simplified version
of the tables that the user can use to define these weights.

    As shown in the exhibit, for a person who is currently 40 years old, the
relative weight to give to the exposure that a person of that age has received
during his life is:  2.0 times the exposure received from ages zero to 10, 2.0
times the exposure received from ages 10 to 20; 1.0 times the exposure
received from ages 20 to 30; and 1.0 times the exposure received from ages 30
to 40.  With these assumptions, the childhood exposures of individuals that
are currently 40 years old are weighted twice as heavily as subsequent
exposures.  The assumptions used in the model for both melanoma skin cancers
and cataracts are equal weights across all ages.  The user may modify these
assumptions as desired to evaluate alternative formulations.

    The dose-respose relationships currently included in the model are called
"power functions" and are generally of the form:

                    In(incidence) = a + b * ln(E),

    where:

        incidence = the age specific incidence for a given race and sex;

        a = a constant that varies by age, race, and sex;

        b = a constant that varies by age, race and sex; and

        E = measure of lifetime exposure to ultraviolet radiation.

Using this equation, the fractional change in incidence as a function of the
fractional change in UV can be expressed as:

    ..     .   n  ,            .  ,        ..fraction change in exposure  ,,.b  ,
    fractional change in incidence = (	n	c	 +1)  -1
                                               exposure


The value of this formulation is that:

        •   the change in incidence is expressed as a fractional
            change that can be multiplied by the baseline incidence
            to compute the increased age specific incidence; and

        •   the change in exposure can be expressed as a
            fractional change, thereby avoiding potential
            difficulties in specifying the absolute levels of
            changes in exposure to UV (see Serafino and Frederick
            (1986) for a description of these potential
            difficulties).

    To specify the dose-response relationship, the user must choose a measure
of UV (e.g., RB-Meter, DNA, or Erythema action spectrum) and provide estimates
of b in the above equations.  Scotto, Fears, and Fraumeni (1981) present a

                             *** DRAFT FINAL ***

-------
                         F-27
                      EXHIBIT F-15

          Sample Table for Specifying  Relative
     Weights for Exposure During a Person's  Lifetime
Current Age
(years)
10
20
30
40
50
60

0-10
2.0
2.0
2.0
2.0
2.0
2.0
Age
10-20
-
2.0
2.0
2.0
2.0
2.0
During Exposure (years)
20-30 30-40 40-50
-
-
1.0
1.0 1.0
1.0 1.0 1.0
1.0 1.0 1.0

50-60
-
-
-
-
-
1.0
All values are illustrative.
                    *** DRAFT FINAL ***

-------
                                   F-28
range of estimates of b for white males and females for use with the RB-Meter
action spectrum.9  Their analysis indicates that these coefficients do not
vary by age, indicating that the incidence at each age is correlated to
changes in exposure to UV in the same manner.

    As described in Chapter 7, an analysis of the data in Scotto, Fears, and
Fraumeni (1981) was performed using the DNA action spectrum.  The estimated
coefficients did not differ significantly from the coefficients reported by
Scotto, Fears and Fraumeni for the RB-Meter action spectrum.  This result is
expected because at current ozone levels the RB-Meter and DNA UV flux values
are highly correlated.  Exhibit F-16 displays the estimated coefficients for
the incidence of basal and squamous cell non-melanoma skin cancer used with
the DNA action spectrum.  Recent experimental data with mice indicate that an
action spectrum that weights the short-wavelength UV-B use more heavily than
the DNA action spectrum may be preferred; such as the Erythema Action Spectrum
(see Cole, Forbes, and Davies 1986).  Although the risk estimates in Chapter
18 are based on the DNA Action Spectrum, work is ongoing to evaluate the
implications of using the Erythema Action Spectrum in its place.

    The coefficients for the dose-response relationships for each of the types
of melanoma were derived from data presented in Scotto and Fears (in
press).10  These coefficients vary by type of melanoma, and sex.  As with
the coefficients for non-melanoma, the values do not vary by age, and are for
whites only.  These values are also presented in Exhibit F-16.  These
coefficients are assumed to be applicable for the DNA action spectrum.
     9 The coefficients for non-melanoma were derived from the regression
coefficients presented in Scotto, Fears, and Fraumeni (1981, p. 10).  The
values reported were for an exponential formulation of the risk model.  To
translate the coefficients into the power formulation, the coefficients were
multiplied by 135 UV units, the ambient UV exposure level at which the power
function and exponential function models predict equal incidence results.

    10  The information reported in Scotto (in press) reported the percent
increase in the incidence in each of the different types of melanoma for a 10
percent increase in UV.  To translate these data into coefficients in the
power formulation of the risk model, the power function was rearranged to
solve for the coefficient (b) as follows:

                              Change in incidence
                   b = ln(l + 	)
                                   incidence
                                 Change in UV
                                 	)
                                      UV

For example, an 8 percent increase in incidence resulting from a 10 percent
increase in UV yields the following coefficient:  ln(l.08)/ln(l.1) = 0.807.
                             *** DRAFT FINAL ***

-------
                          F-29
                     EXHIBIT F-16

               Coefficients Relating  Percent
       Change in  UV to Percent Change in Incidence

   (For use with the DNA Action  Spectrum  -- Whites only)
                                  Low
Middle
High
NON-MELANOMA SKIN CANCER
Squamous
Male
Female
Basal
Male
Female
MELANOMA SKIN CANCER
Face, Head and Neck
Male
Female
Trunk and Lower Extremities
Male
Female


1
1

0
0


0
0

0
0


.42
.47

.932
.316


.661
.798

.421
.341


2.
2.

1.
0.


0.
1.

0.
0.


03
22

29
739


846
019

651
522


2
2

1
1


1
1

0
0


.64
.98

.65
.16


.029
.236

.875
.700
a/ Middle value minus one standard error.

b/ Middle value plus one standard error.

Sources:  Melanoma coefficients  derived  from  Scotto and Fears
          (in press).  Non-melanoma coefficients presented in
          Chapter 7.
                    *** DRAFT FINAL ***

-------
                                   F-30
    A dose-response relationship for non-melanoma mortality has not been
estimated.  For purposes of modeling, the relationships for non-melanoma
incidence (basal and squamous) are assumed to apply.

    A separate analysis of melanoma mortality has been performed by Pitcher
(1986).  The coefficients estimated as the result of that investigation are
used to model the potential increase in mortality due to melanoma skin
cancer.  The coefficients estimated using peak UV flux values (weighted using
the DNA action spectrum) as the measure of exposure are reported in Exhibit
F-17.  Analysis is continuing on alternative measures of exposure.
Preliminary results indicate that using annual UV flux as the exposure measure
would result in coefficients that are as much as 60 percent smaller.

    Coefficients for dose-response relationships for Blacks and members of
other races have not been developed.  Skin cancer risks for blacks  and members
of other races are not modeled.

    The correlation between UV-B and cataract prevalence estimated  by Killer,
Sperduto and Ederer (1983) is used to model the potential changes in the
incidence of cataracts as a function of changes in UV-exposure.  The risk due
to UV in the multivariate logistic risk equation developed by Killer, Sperduto
and Ederer varies with age (see Chapter 10).  For purposes of modeling, a
simplified risk equation that is age-invariant was developed.  A single
coefficient was estimated by weighting the UV risk across the age groups by
the current estimate of increased prevalence due.to UV.  The resulting
coefficient (which is used in the "power formulation") provides an
approximation of the expected increased risk across all age groups.  The
estimated coefficients are displayed in Exhibit F-18.  As shown in  the
Exhibit, the same coefficients are used for all races and males and females.

    To compute increased incidence of senile cataracts, the changes in
prevalance over time are computed.  Incidence in a given year is equal to
change in prevalence (relative to the previous year), plus the number of cases
(counted in the previous time year) that were associated with people who have
died since last year.  It is assumed that the prevalence of cataracts is the
same among people who die or survive from one year to the next.

Step 4 -- Summarize the Results

    The dose-response relationships are applied to each of the age  groups of
the population over time using estimated changes in UV exposure as  outlined
above.  The results of the method include:

        •   changes in the total population incidence rates  (e.g.,
            per 100,000 people) for a standard population age, sex,
            and race distribution;

        •   the total number of cases of each type of cancer over
            time, and the change in the number of cases from the
            number that would exist in the absence of changes in UV
            flux associated with changes on total column ozone; and
                             *** DRAFT FINAL ***

-------
                     F-31
                  EXHIBIT F-17

   Coefficients Relating Percent  Change in UV  to
       Percent Change  in Melanoma  Mortality
(For Use with DNA Action Spectrum -- Whites  Only)
                     a/                    b/
                  Low       Middle     High
       Males       0.78       0.85       0.92

       Females     0.50       0.58       0.66
       a/  Middle estimate minus one standard
          error.

       b/  Middle estimate plus one standard
          error.

       Source:  Pitcher  (1986).
               *** DRAFT FINAL ***

-------
                         F-32
                      EXHIBIT F-18

         Coefficients Relating  Percent Change in
Senile  Cataract Prevalence for a One Percent Change in  UV
                      (all races)
                         a/                    b/
                      Low       Middle      High
          Males and   0.171     0.311       0.419
          Females
          a/ Middle estimate minus  one  standard
             error.

          b/ Middle estimate plus one standard
             error.

          Source:   Derived from Killer,  R.,
                   R.  Sperduto, and F.  Ederer  (1983),
                   "Epidemiologic Associations
                   with Cataract in the 1971-1972
                   National Health  and  Nutrition
                   Examination Survey," American
                   Journal of Epidemiology,
                   Vol. 118, No. 2,  pp.  239-248.
                   *** DRAFT FINAL ***

-------
                                   F-33
        •   the total number of deaths due to each type of cancer over time,
            and the change in the number of deaths from the number that would
            exist in the absence of changes in UV flux associated with changes
            on total column ozone.

    These results are reported for the U.S. as a whole (the calculations are
performed separately for each of the three U.S. regions).  The increased
number of cases and deaths are reported separately for those individuals that
are alive today and for those individuals who are expected to be born in the
future.

DEGRADATION  OF POLYMERS

    The purpose of this portion of the module is to evaluate the potential
damages and costs associated with polymer degradation that may be caused by
increases in the flux of ultraviolet radiation (UV) reaching the earth's
surface.  As described in Chapter 13, UV can degrade the physical and
appearance characteristics of polymers.  This module evaluates the economic
impacts that increased UV flux may have on polymer performance based on
studies by Andrady (1986) and Horst (1986).

    There are several types of polymers that may be adversely affected by
increases in UV, including:  polyvinylchloride (PVC); acrylics; polycarbonate;
polypropylene;  and polyester (Andrady 1986, p. 22).  To date, data for
assessing the potential future markets and potential damages due to UV have
only been developed for the portion of the PVC market used in construction
(siding, window profiles, rainwater systems, and pipe and conduit).  This PVC
market accounts for about 26 percent of all polymers subject to exposure to UV
(measured by production volume), and consequently the estimates from this
module are underestimates of the potential risks of polymer degradation due to
ozone depletion.

    Recognizing that the current data only account for a portion of the
polymer market, the module is designed to evaluate up to five different types
of polymers using the method outlined for PVC.  As data for other polymers
become available, it will be integrated into the module as appropriate.

    The effects of UV on polymers are evaluated in three steps:

        1.  Project the size of the market for polymers that may be
            subject to degradation due to UV.

        2.  Assess the damage that increases in UV (due to ozone
            depletion) may have on the polymers.

        3.  Assess the costs of the damage to the polymers.

Each step is described in turn for evaluating PVC in construction applications.
                             *** DRAFT FINAL ***

-------
                                   F-34
Step 1 -- Market for  Polymers

    The future market for PVC polymers in construction will be influenced by
construction activity and the costs and performance of substitute materials.
Horst (1986) developed several approaches for projecting the expected
production and use of PVC through 2075.  The approach used here reflects
expected market saturation and is of the following form:
                       Q = A ( 1 - exp (-k(T - TQ)))
    where:
        Q = quantity per person per year in pounds;
        A = a constant;
        k = a constant;
        T = the current year of the projection; and
        T  = a base year identified for the polymer type.

    The constants and base year computed for PVC by Horst (1986,  p.  4A-3)
using statistical analyses of historical data are as follows:   A = 63.0308;  k
= 0.01265; and T  = 1966.  These values are used along with the U.S.

population scenario (see Appendix B) to compute a middle estimate of the size
of the future market for PVC in pounds.  By using the population scenario from
the scenarios module, the estimates of the PVC market will be consistent with
the values used to develop the scenarios of production of CFCs and the other
trace gases, and the evaluation of human health effects.

    Because there is uncertainty in the expected future market for PVC, the
module also produces low and high estimates of future demand.   The low and
high estimates are based on the statistical uncertainty in the estimates of  A
and k.  By approximating the future demand, Q, with the first three terms of
its expansion series, Horst (1986) showed that he variance of Q could be
estimated from the covariance matrix of A and k.  Specifically,

                    Var (Q) « Var(Ak) (T-T )2.

                              - Covar(Ak,Ak2) (T-T )3 +

                              Var(Ak2/4) (T-T )*.

The model computes the variance of demand in each year.  The low estimate is
set at one standard deviation below the middle estimate; the high estimate is
set at one standard deviation above.  These low and high values (along with
the middle value described above) are used in step 3 to evaluate the costs
associated with the degradation of these polymers.

Step 2 -- Polymer Damage

    A damage index that indicates the level of degradation of polymer
characteristics was developed by Andrady (1986).  When there is no
degradation, the index value equals 1.0.  Degradation is indicated by
increasing values of the index, such as 1.1.  This index is used here to


                             *** DRAFT FINAL ***

-------
                                   F-35
describe the potential damage to polymers due to increased UV flux associated
with ozone depletion.

    Andrady (1986) also developed an estimate of how the damage index may vary
as a function of ozone depletion.  Due to uncertainties in the relationship
between ozone depletion and UV, and UV and polymer characteristics, the
relationship between ozone depletion and the damage index is represented as a
range.  These ranges are displayed in Exhibit F-19.

    To evaluate the damage index in each year, the ozone depletion estimate
for that year is used get a range of values in the damage index table.  The
ozone depletion estimate for the U.S. is taken as the estimate for the middle
U.S. region.  The result of this method is a range of damage index values for
each year (low, middle, and high).

Step 3 -- Assess  Damage  Costs

    The costs of polymer degradation depend on the manner in which the
characteristics of the polymers degrade, and the steps that are taken in
response to the degradation.  Horst (1986) developed an approach for
evaluating the costs associated with the production of new PVC each year.
These costs represent the implications of changing the formulation of PVC
during manufacture in order to maintain its characteristics in light of
increased UV exposure.  These costs do not include the potential damages to
PVC already in place.  Additionally, these estimates do not reflect the losses
that may be associated with polymer degradation in the absence of changes in
the polymer formulation.  Therefore, these estimates are an underestimate.

    Horst (1986) identified the costs of changing the formulation of PVC as a
function of the increased amount of stabilizer that needs to be added to the
polymer to maintain its characteristics.  A 25 percent increase in stabilizer
was estimated to lead to a 1.86 percent increase in the price of PVC, from its
current cost of $0.604 per pound (Horst 1986, p. 5-16 and 6-8).  The increased
amount of stabilizer required as a function of ozone depletion was identified
by Andrady (1986) and is presented in Exhibit F-19.

    To estimate the costs of the increasing the stabilizer in response to
ozone depletion, Horst recommends using a ten year lag between the increase in
stabilizer required and the increase in the price of PVC.  This lag reflects
the time needed for the industry to respond to changing environmental
conditions and the fact that the current stabilizer concentrations include a
margin of safety.  This ten year lag is used in the module, although the model
user may override it with an alternative assumption.

    To compute the costs in year T, the following computations are performed:

        •   compute the increased amount of stabilizer required,
            based on the ozone depletion in year T - 10;

        •   compute the increase in price by interpolating between
            no increase in price, and a 1.86 percent increase in
            price for a 25 percent increase in stabilizer;
                             *** DRAFT FINAL ***

-------
                            F-36
                        EXHIBIT F-19

            Damage  Index  and Increase  in Stabilizer
                 For Ranges of Ozone Depletion
Ozone Depletion
(percent)
0-5
5-10
10-15
15-20
Damage Index
Low
1.01
1.01
1.02
1.03
Middle
1.015
1.025
1.055
1.105
High
1.02
1.04
1.09
1.18
Stabilizer Increase (%)
Low
1.0
1.0
3.0
3.0
Middle
3.0
5.0
11.5
20.5
High
5.0
9.0
20.0
38.0
Source:   Derived  from Horst (1986),  p,  6-10.
                      *** DRAFT FINAL ***

-------
                                   F-37
        •   given the increase in price, compute the cost as
            follows:

            C(T) = (D(T)/PQb)/(l+b) * [PQ(1+b) - P^1"1"^]

            where:

                C(T) = cost in year T;

                D(T) = demand in year T]

                P  = price of PVC in the absence of changes in the

                     formulation of the polymer;

                b = price elasticity of demand for PVC;

                P  = the new price for PVC given the change in the

                     formulation of the polymer.

    This equation represents the annual loss in consumer surplus associated
with the the estimated changes in price.  The price elasticity of demand
estimated by Horst (1986) is -1.956.  This value represents (in part) the
estimated availability of appropriate substitutes for PVC.

    This equation can be evaluated three times each year to generate low,
medium, and high estimates of the costs.  The low estimate uses the low demand
value  (from step 1) and the estimate of the change in price associated with
the low estimate of the increase in stabilizer associated with ozone
depletion.  The middle and high estimates use the middle and high values for
these  components respectively.  If there is an increase in ozone abundance,
the damage index is set to 1.0 and there are no estimated costs.  The maximum
price  increase (1.86 percent) estimated by Horst is associated with a 25
percent increase in stabilizer.  Price increases are capped at this value
(1.86  percent) even when an increase in stabilizer exceeding 25 percent may be
indicated as being required.  The present value of the costs over time can be
evaluated by discounting these annual costs at user-supplied discount rates.

F.6  LIMITATIONS

    The Effects Module draws together a diverse set of analyses and applies
these  analyses to estimates of global ozone depletion  (from the Atmospheric
Science Module) to evaluate potential risks to human health.  The analyses
require that a chain of events from ozone depletion to health outcomes be
modeled.  There are limitations in each link of this chain, creating
uncertainties in the estimates of the outcomes.

    First global ozone depletion is translated into latitudinal values.  The
current algorithm used for these latitudinal estimates is based on a limited
set of data from a single two-dimensional (2-D)  model.  Additional 2-D
analyses should be performed.
                             *** DRAFT FINAL ***

-------
                                   F-38
    The estimates of changes in UV as a function of changes in the ozone level
are based on a model.  The model is in its early stages of field validation,
and additional analyses remain to be performed.

    The dose-response relationships for skin cancers and cataracts all depend
on epidemiologic studies of humans in the U.S.  Because the studies are based
on humans in the U.S., there are no extrapolation problems (e.g., from animals
to humans, or from one population of humans to the U.S. population).  However,
the epidemiologic studies are generally limited by their measures of
environmental exposure to UV, as opposed to individual lifetime exposures.  In
addition to the limitations of the estimates of the dose-response relationship
themselves, the characterization of baseline evidence (in the absence of ozone
depletion) and baseline population characteristics are subject to uncertainty.

    In addition to uncertainties and limitations in the estimates produced by
the module, the module is limited because there are a variety of important
effects that are not evaluated.  These omitted effects include:

        •   polymers other than PVC;

        •   serious  (non-fatal) cases of non-melanoma skin cancers;

        •   impacts on plants;

        •   impacts on aquatic organics;

        •   impacts on urban smog formation;

        •   potential immune suppression in humans; and

        •   possible other diseases.

The omission of these effects biases the estimates downward.  In Chapter 18,
qualitative estimates of some of these effects are presented based on
interpolations and extrapolations from case studies and research in early
stages.  Data used to prepare these estimates include:

    •   Imacts on soybeans.  For depletion of 25 percent, a 19 percent
        reduction in yield is used (average of results reported in Chapter
        11 for Essex cultivar).  For depletion between zero and 25 percent,
        a value is interpolated from between zero and 19 percent.  Values
        are not extrapolated to outside this range.

    •   Impacts on Ground-Based Ozone (Smog).  Estimates of increased
        smog levels are interpolated from case study estimates (Chapter 14)
        as follows:

            Los Angeles:  16.6% depletion yields a 4.5 percent increase;
                          33.3% depletion yields a 9.4 percent increase;
                             *** DRAFT FINAL ***

-------
                                   F-39
        --  Philadelphia:  16.6% depletion yields a 13.4 percent increase;
                           33.3% depletion yields a 33.0 percent increase;

        --  Nashville:  16.6% depletion yields a 23.8 percent increase; and
                        33.3% depletion yields a 50.0 percent increase.

        Impacts on Aquatic Organisms.  Estimates for reduced anchovy
        populations are interpolated from laboratory tank experiments
        (described in Chapter 12) as follows:

                    Percent Change      Percent Annual
                       in UV-B         Population Death
                         10%               0%
                         20%               0% to 4.8%
                         30%               0% to 11.5%
                         40%             2.5% to 18.0%
                         50%             6.0% to 23.0%
                         60%            11.0% to 25.0%
Because these data are based on case studies and research in early stages, all
the estimates of these effects are very preliminary, and should be viewed with
caution.
                              *** DRAFT FINAL ***

-------
                                   F-40
                                REFERENCES
Andrady, A. (1986), "Analysis of Technical Issues Related to the Effect of UV-B
    on Polymers," prepared for the U.S.  Environmental Protection Agency,
    Washington, B.C.

Cole, C.A., P.D. Forbes, and R.E. Davies (1986),  "An Action Spectrum for UV
    Photocarcinogenesis," Photochemistry and Photobiology,  pp.  275-284.

Connell, P.S.  (1986), "A Parameterized Numerical  Fit to Total Column Ozone
    Changes Calculated by the LLNL I-D Model of the Troposphere and
    Stratosphere," Lawrence Liverpool National Laboratory,  Livermore,
    California.

Elder, D.E., Jacoby P.M., Tuthill, R.J.  and Clark W.H.  Jr.  (1980),  "The
    Classification of Malignant Melanoma," American Journal of
    Dermatopathology, Vol. 2, pp. 315-320.

Killer, R., R. Sperduto and F. Ederer (1981), "Epidemiologic Association with
    Cataract in the 1971-1972 National Health and Nutrition Examination
    Survey," American Journal of Epidemiology, Vol. 118, No. 2, pp. 239-298.

Horst, R.L. (1986), "The Economic Impacts of Increased UV-B Radiation on
    Polymer Materials:  A Case Study of Rigid PVC," prepared for the U.S.
    Environmental Protection Agency, Washington,  B.C.

Isaksen, I.S.A. (1986), "Ozone Perturbations Studies in a Two Bimensional
    Model with Temperature Feedbacks in the Stratosphere Included," presented
    at UNEP Workshop on the Control of Chlorofluorocarbons, Leesburg,
    Virginia,  September 1986.

Leske, C.L., and R.B. Sperduto (1983), "The Epidemiology of Senile  Cataracts:
    A Review," American Journal of Epidemiology,  Vol. 118,  No.  2, pp.  152-165.

NPA Bata Services, Inc. (1986), Regional Economic Projections Series.

Pitcher, H. (1986), "Melanoma Beath Rates and Ultraviolet Radiation in the
    United States 1950-1979," U.S. Environmental  Protection Agency,
    Washington, B.C.

Scotto, Fears, and Fraumeni (1981), "incidence of Nonmelanoma Skin  Cancer in
    the United States," U.S. Bepartment of Health and Human Services,  (NIH)
    82-2433, Bethesda, Maryland.

Scotto J. and T. Fears (in press), "The Association of Solar Ultraviolet
    Radiation and Skin Melanoma Among Caucasians  in the United States,"
    Cancer Investigation.

Serafino, G. and J. Frederick (1986), "Global Modeling of the Ultraviolet
    Solar Flux Incident on the Biosphere," prepared for the U.S. Environmental
    Protection Agency, Washington, B.C.


                             *** BRAFT FINAL ***

-------
Chapter 18

-------
                               CHAPTER 18

              HUMAN HEALTH AND ENVIRONMENTAL EFFECTS
SUMMARY

    If no actions are taken to regulate future emissions  of  ozone-modifying
substances and greenhouse gases,  damage to human health and  the  environment  is
likely to occur.  The purpose of  this chapter is to present  a  range of
quantitative estimates of the potential risks to human health  and the
environment arising from emissions of ozone-modifying substances.

    The risk estimates draw on the data and methods described  in earlier
chapters.  These chapters describe:   (1) potential  future emissions and
concentrations of ozone-modifying substances; (2) the effects  that these
substances may have on stratospheric ozone and global climate;  (3) the
potential damage to human health  and the environment arising from ozone
modification and climate change;  and (4) a comprehensive  modeling framework
for estimating these risks.

    The major types of effects for which quantitative estimates  are provided
include ozone depletion; additional cases and deaths related to  skin cancers;
additional cases of senile cataract; damage to polyvinyl  chloride materials;
rise in global equilibrium temperature; sea level rise; cost of  sea level
rise; reduction in soybean yield; increases in ground-based  ozone; and effects
on the survivability of a selected aquatic organism.   Certain  important risks
discussed in earlier chapters are not evaluated because of data  limitations
and the absence of relevant scientific information.  In most instances, the
estimates cover the U.S. only.

    Estimates of the potential risks due to emissions of  ozone-modifying
substances are necessarily uncertain for two reasons.  First,  estimating
future risks requires projections of key factors, such as population and
economic growth.  Second, the many relationships that define how emissions
result in risks are themselves uncertain.

    To investigate the significance of key assumptions and data  used to model
risks, the chapter presents a range of risk estimates varying  the most
important assumptions, one at a time.  The range of results  obtained by
varying each assumption is used to determine which  factors used  to model risks
are most important.
                          * * *  DRAFT FINAL  * * *

-------
                                     18-2
  FINDINGS

1/1 1.   FOR THE  MOST LIKELY  ESTIMATES  OF  CFC  GROWTH,  ATMOSPHERIC MODIFICATION AND
      DOSE-RESPONSE,  AN  INCREASE  IN  SKIN  CANCER  CASES AND  DEATHS, CATARACT
      CASES, MATERIALS DAMAGE,  GLOBAL TEMPERATURE AND SEA  LEVEL  IS EXPECTED.

      la.   An  additional 40  million  skin  cancer  cases and  800 thousand deaths
           from  skin  cancer  are projected for people alive today and born in the
           next  88 years,  along with 12 million  cases of cataracts  (U.S. only).

      Ib.   The earth  would be  committed to  global warming  from 3.0°C to 9.5°C by
           2075,  sea  level rise would be  from 80 cm to  94  cm.

  2.   FOR GASES  THAT  COUNTER DEPLETION, SOME ASSUMPTION IS NEEDED ABOUT FUTURE
      LIMITS TO  GREENHOUSE WARMING.  THE  STANDARD ASSUMPTION HAS BEEN THAT NO
      LIMIT WILL EVER BE PLACED ON GLOBAL TEMPERATURE INCREASES.  ESTIMATES OF
      THE RISK OF OZONE  DEPLETION FROM  CFC  AND HALON GROWTH ARE  HIGHLY SENSITIVE
      TO ASSUMPTIONS  ABOUT GREENHOUSE WARMING.   IF  GLOBAL  WARMING IS EVENTUALLY
      LIMITED', THE RISK  OF COLUMN OZONE DEPLETION FROM  CFCS AND  HALONS WILL
      RISE.

      2a.   A limit of global warming of 3°C would more  than double the risks of
           CFC and HaIon growth.

      2b.   A more stringent  global warming  limit would  have even greater effect.

      2c.   Model results are sensitive  to the rate  of methane growth, which is
           highly uncertain  even  without  regard  to  greenhouse warming
           considerations.

  3.   ALTERNATIVE ESTIMATES  OF CFC GROWTH PRODUCE DIFFERENT LEVELS OF RISKS.

      3a.   The low case  of 1.2 percent  CFC  growth  (about half of the 2.5 percent
           central case  with greenouse  gases that counter  depletion growing)
           would lower the estimates of increased risks of skin  cancer cases and
           deaths and cataracts cases by  more than  90 percent.

      3b.   The high case of  3.8 percent CFC growth  (with greenhouse gases that
           counter depletion growing) would increase the estimates of increased
           risks of skin cancer cases and deaths and cataracts cases by over 400
           percent.

      3c.   Lowering growth rates  to  zero  (with greenhouse  gases  that counter
           depletion  growing)  would  have  the effect of  reducing  skin cancer
           cases and  deaths  by more  than  an additional  10  percent,  and cataracts
           by  a  similar  amount.

      3d.   The highest growth  rate,  which is unrealistically high for the whole
           period, would add a tremendous number of cases  and deaths.
                            * * *  DRAFT FINAL  * * *

-------
                                   18-3
4.  RISK ESTIMATES ARE SENSITIVE TO UNCERTAINTIES ABOUT ATMOSPHERIC RESPONSE
    AND DOSE-RESPONSE.

    4a.  Uncertainty about atmospheric response introduces a greater
         uncertainty than uncertainty about dose-response parameters for skin
         cancers or cataracts.  Both uncertainties affect the risk estimates
         by more than 35 percent.

    4b.  Preliminary results of ongoing analyses of melanoma mortality risks
         indicate that significant uncertainties remain in the specification
         of the dose-response relationship.  The magnitude of this
         specification uncertainty is still being identified, and as a result
         the full implications of this uncertainty cannot be quantified at
         this time.

5.  WHILE NATIONAL QUANTITATIVE ESTIMATES OF AQUATIC, CROP. GROUND-BASED
    OZONE. AND SEA LEVEL RISE DAMAGE CANNOT BE MADE AT THIS TIME. CASE STUDY
    RESULTS INDICATE THAT SIGNIFICANT INCREASES IN GROUND-BASED OZONE, LOSS OF
    AQUATIC LIFE. SEA LEVEL RISE DAMAGE AND LOSS OF CROP YIELD ARE POSSIBLE.
                          * * *  DRAFT FINAL  * * *

-------
                                   18-4
INTRODUCTION

    The previous chapters of this document have described:   (1) potential
future emissions and concentrations of ozone-modifying substances; (2) the
effects that these emissions and concentrations may have on stratospheric
ozone and global climate; (3) the potential risks to human health and the
environment arising from ozone-modification and global climate change; and (4)
a comprehensive modeling framework for estimating these risks.  This chapter
integrates the data and methods presented in earlier chapters to develop
quantitative estimates of risks for human health and the environment due to
ozone modification and climate change.

    The risk estimates are based on the assumption that no action is taken to
regulate CFCs or other ozone-depleting substances.  However, the data and
analyses used here could be applied to alternative regulatory strategies.
Additionally, all of the effects, with the exceptions of global sea level
rise, global ozone depletion, and change in global equilibrium temperature,
are estimated for the U.S. only.1

    The following ozone-depleting substances are considered in estimating
risks:2

        •   CFC-11, 12, 22, and 113;
        •   methyl chloroform (CH3CC13);
        •   carbon tetrachloride (CCL4); and
            Halon-1211 and 1301.

The following trace gases that add ozone or counter ozone depletion (when
chlorine concentrations rise) are considered:3

        •   carbon dioxide (C02);
        •   methane (CH4); and
        •   nitrous oxide (N20).

    Exhibit 18-1 displays the types of effects that are estimated in this
chapter along with the chapter that describes the data and methods used to
derive the effects.  The effects are divided into two sets:   (1) effects that
can be estimated quantitatively for the U.S.; and (2) effects estimated on the
basis of case studies and research in early stages.  The first set is
    1 The estimates presented in this chapter are for the entire U.S.  As
described in Chapter 17, the U.S. estimates combine analyses of three
geographic regions (by latitude):  north, middle, and south.

    2 See Chapter 3 for a description of the scenarios of future emissions
used.

    3 See Chapter 4 for a discussion of the future concentrations of these
greenhouse gases.  These greenhouse gases also counter ozone depletion, at
least in high chlorine cases.
                          * * *  DRAFT FINAL  *

-------
                           18-5
                      EXHIBIT 18-1

 Types of Human Health and Environmental Effects  Estimated
  TYPE OF EFFECT
(Reference Chapter)
          UNITS
  Effects  Estimated Quantitatively  for the U.S.
  Ozone Depletion
  (Chapter 5)

  Non-melanoma Skin Cancer
  (Chapter 7)
  Melanoma Skin Cancer
  (Chapter 8)
  Senile Cataracts
  (Chapter 10)
  Damage to Polyvinyl  Chloride
  Materials (Chapter 13)

  Rise in Equilibrium  Temperature
  from 1980 to 2075  (Chapter 6)

  Sea Level Rise from  1980 to 2075
  (Chapter 15)
Percentage Depletion from
1985 to 2075

Additional Cases and Deaths
for People Alive Today;
People Born 1985-2029; and
People Born 2030-2074.

Additional Cases and Deaths
for People Alive Today;
People Born 1985-2029; and
People Born 2030-2074.

Additional Cases for People
Alive Today; People Born
1985-2029; and People Born
2030-2074.

Present Value Costs in
1985 U.S. Dollars

Degrees Centigrade
Centimeters
                  * * *  DRAFT FINAL  * * *

-------
                           18-6
                       EXHIBIT 18-1
                         (Continued)

 Types of Human Health and  Environmental Effects Estimated
  TYPE  OF EFFECT
(Reference Chapter)
                                             UNITS
  Effects  Based  on  Case Studies and Research in Early Stages;

                                     Percent in 2075
Reduction in Soybean Yield
(Chapter 11)

Increase in Fatality Rate
of Northern Anchovy (Chapter 12)

Increase in Ground-Based Ozone
(Chapter 14)

Cost of Sea Level Rise in
Charleston, SC and Galveston,
TX (Chapter 15)
                                     Percent in 2075
                                     Percent in 2075
                                     Present Value Costs
                                     in 1985 U.S. Dollars
                  * * *  DRAFT FINAL * * *

-------
                                   18-7
estimated using detailed epidemiologic, economic, demographic,  and other
scientific information within the modeling framework described in Chapter 17.
Effects estimated in this first set are representative of risks to the entire
U.S., and include:

        •   global average ozone depletion and associated ozone
            depletion by latitude j1*

        •   non-melanoma skin cancer (cases and deaths);

        •   melanoma skin cancer (cases and deaths);

        •   senile cataracts (cases);

        •   damage to polyvinyl chloride (PVC) materials;

        •   increased global average temperature; and

        •   global sea level rise.

The second set of estimates is based either on research in early stages or on
a small number of case studies the results of which cannot be generalized to
the entire U.S.  Effects estimates included in the second set should be
considered preliminary and are included here only to indicate the potential
order of magnitude of these effects.  The second set of estimates includes the
following types of effects:

        •   costs of sea level rise;

        •   reductions in soybean yields (for one cultivar; 2 out
            of 3 cultivars appear sensitive);

        •   increases in ground-based ozone (smog); and

        •   effects on the survivability of a selected aquatic
            organisms (the Northern Anchovy).

    Because of data limitations and the absence of relevant scientific
information, certain important effects discussed in earlier chapters were not
evaluated.  The major effects not analyzed in this chapter are:

        •   overall impacts on agricultural and natural plants
            (Chapter 11);
    a Estimates of ozone depletion are generated using a one-dimensional
model of the atmosphere that is modified to emulate the results of a
two-dimensional model.  As discussed below, the ozone depletion estimates
would be higher if estimated using a two-dimensional model.
                            * *  DRAFT FINAL  *

-------
                                   18-8
        •   impacts of climate change on human health, water
            resources, forestry, and wetlands (Chapter 16);

        •   increased incidence of cutaneous diseases from immune
            suppression (Chapter 9);

        •   impacts of sea level rise on wetlands (Chapter 15); and

        •   impacts on materials other than polyvinyl chloride
            (Chapter 13).

    Estimates of human health and environmental effects are necessarily
uncertain for two reasons.  First, effects are estimated into the future,
requiring projections of such factors as population growth; economic growth;
and production, use, and emissions of ozone-modifying substances and
greenhouse gases.  Second, the linkages from emissions to risks are themselves
uncertain.  For example, the models and analyses used to evaluate ozone
modification and global warming continue to be developed.

    Critical to understanding the risks of ozone depletion and climate change
is a consideration of the significance of uncertainties in the numerous
factors and assumptions used to model the chain of events that starts with
emissions and results in risks.  Therefore, this chapter starts with "central
case" estimates of risk that are based on the most likely estimate of each
factor.  Following the presentation of the central case, the sensitivity of
the risk estimates to the factors linking emissions to risks are explored
systematically across the most important factors.

    The remainder of this chapter is organized as follows:

        •   Methods for Estimating Health and Environmental Risks
            summarizes the major steps used to estimate risks.

        •   Health and Environmental Risks:  Central Case
            presents estimates of risks associated with the most
            likely set of assumptions and parameters, the "central
            case."

        •   Comparison of Central Case with Results Using
            Alternative Assumptions explores the sensitivity of the
            risk estimates to assumptions about:  (1) emissions of
            ozone-modifying gases;  (2) the ozone-depletion model
            used; (3) sensitivity of global warming to concentrations
            of greenhouse gases; and (4) future concentrations of
            greenhouse gases.

        •   Sensitivity of Effects to Parameter Uncertainty
            identifies how the risk estimates are influenced by
            statistical uncertainty in estimates of:  (1) dose-
            response relationships; and (2) ozone level response to
            emissions.
                            * *  DRAFT FINAL  * *

-------
                                   18-9
        •   Relative Importance of Key Uncertainties  summarizes
            the results of previous sections and compares  the
            significance of the different uncertainties  analyzed in
            this chapter.

METHODS  FOR  ESTIMATING  HEALTH  AND ENVIRONMENTAL  RISKS

    Emissions of ozone-modifying substances and greenhouse gases lead to
effects on human health and the environment by means  of  the following
principal mechanisms:

        1.   Emissions  and ozone depletion.  Emissions of
            ozone-modifying substances deplete total  column ozone.

        2.   Ozone depletion and ultraviolet radiation.   Depletion
            of total column ozone increases the amount of
            ultraviolet radiation (UV) reaching the earth's surface.

        3.   Ultraviolet radiation and health effects. Melanoma
            skin cancer, non-melanoma skin cancer,  and cataracts are
            related to increases in UV.   Changes in the  expected
            number of  cases and/or deaths associated with  increases
            in UV are  described by dose-response relationships.

        4.   Emissions  and global warming.  Emissions  of  greenhouse
            gases cause the earth's equilibrium temperature to rise.

        5.   Global warming and human health, environmental, and
            other effects.  Changes in the global equilibrium
            temperature affect the sea level, human health, aquatic
            organisms, agricultural yields, energy demand  and  other
            factors.

    To quantify risks  for human health and the environment, each of these
mechanisms was evaluated using mathematical relationships.  For  example,
estimates of additional cases of melanoma skin cancers rely on mathematical
relationships between:  (1) emissions and ozone depletion, (2) ozone  depletion
and UV, and (3) UV and cancer incidence.   The sources and  derivation  of the
mathematical relationships are described in earlier chapters.

    The mathematical relationships were integrated in the  modeling  framework
described in Chapter 17.  The framework provides a method  of estimating
expected risks as well as analyzing the joint implications of  key assumptions
and uncertainties in each mathematical relationship.   The  primary period  of
analysis used in the models is 1985 to 2075.  Emissions  of ozone-modifying
substances and concentrations of greenhouse gases are analyzed during this
period.  Risks to human health are evaluated for individuals born before  2075,
in three discrete cohorts (i.e., groups):

        •   those individuals alive today;

        •   those individuals born between now and 2029; and
                          * * -  DRAFT FINAL  * * *

-------
                                   18-10
        •   those individuals born between 2030 and 2074 (the end
            of the analysis period).

Risk estimates for these individuals are based on simulated lifetime exposures
to increased UV.  For example, a 50 year old person who is alive today is
assumed to experience no increase in UV prior to 1986.  During his remaining
life, however, this 50 year old may experience increased exposure to UV as
total column ozone depletes.  Exposure histories for each population cohort
are analyzed separately as a basis for estimating risks.

    The key exposure variable for non-melanoma skin cancer can be approximated
using cumulative lifetime exposure (see Chapter 7).  For melanoma skin
cancers, a major question remains regarding the appropriate measure of
exposure for purposes of modeling risks (see Chapter 8 and Chapter 17).
Evidence suggests that peak or intermittent exposures may be preferred
indicators of risk.  Additionally, exposures to a person prior to the age of
15 may be more important than subsequent exposures.  Insufficient evidence is
available to define conclusively the preferred exposure measure, and work is
continuing in this area.  Similar uncertainty exists regarding the preferred
exposure measure for modeling potential cataract risks (see Chapter 10).
Also, the quantification of the relationship between ultraviolet radiation and
cataracts is in the early stages.  The risk estimates presented below for
non-melanoma (incidence and mortality) and melanoma (incidence) and cataracts
are based on cumulative lifetime exposure.  Risk estimates for melanoma
mortality are based on an equation that relates peak UV flux (as a measure of
exposure) to mortality rates.  Although work is continuing in this area,
preliminary results indicate that this relationship (based on peak UV flux)
may result in larger risk estimates than would equations based on exposures
averaged over longer periods (e.g., annual UV flux).  All the risk estimates
presented below assume constant baseline incidence and mortality rates; recent
secular trends in these rates are not extrapolated into the' future.

    Of note is that some of the individuals in the second and third cohorts
(born after 1985) will live beyond the end of the primary period of analysis
(2075).  To estimate the lifetime risk for these individuals, the time horizon
was extended to 2165 holding emissions and concentrations constant after
2075.  Risk estimates for the second two population cohorts are affected by
the assumption that emissions and concentrations are constant after 2075.  In
most of the scenarios investigated, depletion, UV radiation, and temperature
are still increasing in 2075.

    Non-health risks (e.g., global warming, sea level rise, and PVC damages)
are reported for 2075.  Risk estimates based on preliminary research and case
studies are also reported for 2075.

    It is important to note that the models used to estimate risks do not
include feedbacks among factors such as an emissions growth rate (2.5 percent
for CFC emissions, for example) and the consequences of that growth rate
(6.0°C warming, for example).  To the extent that there are feedbacks from
effects to subsequent economic growth, that is, to the extent that effects may
reduce economic growth, the system will in fact be self-limiting in the growth
of CFCs.  As now conceived, the risk models implicitly assume that the damage
from effects is not large enough to alter the growth of society.

                          * * *  DRAFT FINAL  * * *

-------
                                   18-11
HEALTH AND ENVIRONMENTAL RISKS:  CENTRAL CASE

    This section presents central case estimates  of human  health  and  environ-
mental risks due to ozone modification and climate change.5   The  central  case
assumes that no actions are taken to limit:   (1)  emissions of ozone modifyers;
or (2) future global warming due to the greenhouse effect.  The significance
of these and other assumptions is investigated in later sections.
    5 The central case estimates reflect the most  likely values  of  key
assumptions and inputs used to model risks:

    •   Annual production of CFC 11 and 12 grow at an annual  average  rate  of
        2.5 percent from 1985 to 2050,  and remains constant  following 2050;
        others as described in Chapter  3.   The compounds analyzed include:
        CFC-11, CFC-12, CFC-22, CFC-113, methyl chloroform,  carbon  tetra-
        chloride, Halon-1211, and Halon-1301.   Emissions estimates  reflect the
        storage of some substances in their end-use products  for many years.

    •   Consensus estimates of the annual rates of increases  in  atmospheric
        concentrations of other trace gases are used:   carbon dioxide (C02) at
        0.6 percent; methane (CH4) at 1.0 percent; and nitrous oxide  (N20) at
        0.25 percent.  Trace gas assumptions are discussed in Chapter 4.

    •   A parameterized relationship between emissions of ozone  modifyers,
        trace gas concentrations and global ozone  depletion  is used that
        reflects the results of a one-dimensional  model of the atmosphere
        using the most recent estimates of reaction rates.   The  parameterized
        atmospheric model (described in Chapter 17) represents the  results of
        the Wuebbles Model (reported in Chapter 5).

    •   The latitudinal distribution of ozone depletion is evaluated  using
        the results of a time-dependent two-dimensional model of the
        atmosphere.  The latitudinal analysis of ozone depletion is presented
        in Chapter 17.

    •   The relationship between changes in ozone  abundance  and  changes  in
        UV flux reaching the earth's surface is based on best estimates  of a
        radiation model of the atmosphere.  The estimates of UV  flux  are
        described in Chapter 17.

    •   The risks to human health due to increases in UV are evaluated using
        middle estimates of dose-response coefficients developed in
        epidemiologic analyses in the U.S.  The quantifiable risks  to human
        health are described in Chapters 7, 8, and 10.

    •   The middle estimate by the National Academy of Sciences  (NAS) for
        the sensitivity of the global climate to greenhouse gas  forcings  is
        used -- 3.0°C equilibrium warming for a doubling of the  concentration
        of C02.  The National Academy of Sciences  estimates  are  presented  in
        Chapter 6.  Recent analyses of climate sensitivity indicate that  a 4°C
        sensitivity may be a preferred central case assumption (Manabe and
        Wetherald 1986; Washington and Meehl 1984; Hansen et al. 1984).


                          * * *  DRAFT FINAL  * *  *

-------
                                   18-12
    Exhibit 18-2 shows projections of average global ozone depletion for 1985
through 2100 for central case assumptions (see footnote 5).  As shown in the
exhibit, average global ozone levels are projected to have declined by about 2
percent by 2020, about 19 percent by 2075, and about 30 percent by 2100.  The
actual depletion will vary by latitude.  In the northern hemisphere, depletion
will be greater than average at latitudes near 40°N and above and lesser at
latitudes below 40°N.

    Of note is that the estimates show continued ozone depletion beyond 2050
when production is assumed to level off and remain constant.  Because ozone-
modifying substances are stored in products, emissions can lag production by
many years.  In the central case, emissions do not become constant until
approximately 2070.  Even after emissions level off in 2070, ozone depletion
will continue as atmospheric concentrations of chlorine continue to
increase.6  For estimating health effects beyond 2100, the modeling
assumption is that ozone levels are constant after 2100.

    Estimates of human health risks in the U.S. are presented in Exhibit 18-3.
Health risks were estimated as the additional number of cases and/or deaths
related to skin cancer and senile cataracts for three population cohorts:  (1)
people alive today, (2) people expected to be born between today and 2029, and
(3) people expected to be born from 2030 through 2074.  jeople alive today can
expect to experience the following additional estimated risks due to ozone"
depletion over their lifetimes:(1) approximately b30 thousand cases of basal
cell skin cancer; i^j Ji)0 thousand cases of squamous cell skin cancer;  (3) 16
thousand deaths from both basal and squamous cell skin cancer; (4) 12 thousand
cases of melanoma skin cancer; (5) 4 thousand deaths due to melanoma skin
cancer; and (6) 600 thousand cases of senile cataracts.  These estimates are
of additional cases and deaths over and above what would have occurred  in the
absence of ozone depletion.

    The number of additional cases and deaths due to ozone depletion is
greater for the latter two population cohorts.  Two factors account for
increased risks in future generations.  First, people alive today were  assumed
to experience increased exposure to UV only during the remainder of their
lives.  For example, the increase in cumulative lifetime exposure for a person
who is 40 years old today results only from increases in UV flux during the
remaining years of his life.  People who are born after today (i.e., the
latter two cohorts) experience increased exposure to UV during a larger
fraction of their lives.  Therefore, the impact of increases in UV flux over
time on the incidence and mortality of these conditions is greater for  people
exposed for longer periods.
    6 Concentrations rise even though emissions are constant because the
atmosphere is out of equilibrium.  The removal of CFCs and chlorine from the
atmosphere is slow, meaning that it takes a long time for equilibrium  (i.e.,
steady state) to be reached even though emissions are constant.
                            * *  DRAFT FINAL  * * *

-------
                                     10-13
                                EXHIBIT 18-2

                  Global Average Ozone Depletion:  Central Case
        o -a
z
o
b
U
Z
o
N
O
                                                        2065
2085
                             * * *  DRAFT FINAL  * * *

-------
                                  18-14
                             EXHIBIT  18-3

                     Human  Health Effects:   Central  Case
       (Additional Cumulative Cases  and Deaths by Population Cohort)
HEALTH EFFECT ALIVE TODAY3
£.-0? fU^
Non-Melanoma Skin Tumors '
Additional Basal Cases 630,600
Additional Squmaous Cases 386,900
Additional Deaths 16,500
Melanoma Skin Tumors
Additional Cases 12,300
Additional Deaths 3,900
Senile Cataract
Additional Cases 593,600
BORN 1985-2029b BORN 2030-2074'
Wv.y-f^f /

5,012,900
3,185,800
135,000

109,800
32,200

3,463,400
wO

17,630,500
12,122,400
509,300

430,500
115,100

8,295,800
a  Analysis period for health effects:  1985-2074.

   Analysis period for health effects:  1985-2118.

°  Analysis period for health effects:  2030-2164.
                          *  * *  DRAFT FINAL  * * *

-------
                                   18-15
    Second, ozone depletion, and hence exposure to UV radiation, is simulated
to increase over time (except after 2100 when ozone depletion -- and
consequently UV flux -- is held constant).   As a result, persons in the later
two cohorts are exposed to higher levels of UV flux.

    Exhibit 18-4 presents estimates of two groups of non-health effects (1)
estimated quantitatively for the U.S.  -- including effects on PVC materials,
global equilibrium temperature, and sea level; and (2) based on case studies
and research in early stages -- including impacts on agricultural output,
aquatic life,  and ground-based ozone (smog).  Each effect is as follows:

        •   PVC materials damage represents the present value of
            costs of producing PVC materials in a manner that will
            prevent damages due to increased UV flux.  Damages are
            estimated at $550 million (1985 U.S. dollars)^using a
            real discount rate of 3 percent.  Estimates of PVC
            damage were derived from case studies described in
            Chapter 13.  The damage estimate represents the
            additional cost of producing PVC polymers in a manner
            that prevents deterioration in their physical properties
            and appearence that would otherwise result from
            increased exposure to UV flux.

        •   Rise in equilibrium temperature represents the
            increase in global average annual temperature that would
            be expected under equilibrium conditions with central
            case estimates of the concentrations of the greenhouse
            gases in 2075.  The middle NAS estimate of climate
            sensitivity to a doubling of C02 is used (3°C),
            resulting in an estimate of a global equilibrium
            temperature increase of about 6.2°C between 1980 and
            2075.  Of note is that the actual warming that may be
            realized will lag by several decades or more.  Also of
            interest is that recent analyses indicate that a climate
            sensitivity of 4°C may be a preferred central case
            assumption (see Chapter 6).  Assuming a doubled C02
            climate sensitivity of  4°C implies an equilibrium
            warming of over 8°C by 2075.  (These estimates do not
            include the effect of additional stratospheric water
            vapor due to methane.  The effect of vertical ozone
            depletion is taken from Ramanathan et al. (1985), and is
            not based on the ozone depletion estimates presented
            here.)

        •   Sea level rise occurs as global warming leads to
            thermal expansion of ocean waters and to increased
            discharge of ice and meltwater from Antarctica and from
            glaciers (see Chapter 15).  Using an estimate of the
            transient change in temperature by 2075, a sea level
            rise of 101 cm is estimated.
                          * * *  DRAFT FINAL  * * *

-------
                             18-16



                        EXHIBIT  18-4

      Materials, Climate and  Other Effects:   Central  Case



     TYPE OF EFFECT                 EFFECT           UNITS


Effects Estimated Quantitatively for the U.S.

   Materials Damage a/                550      Present Value
                                               (millions of 1985
                                               dollars)

   Rise in Equilibrium                6.2      Degrees Centigrade
   Temperature by 2075 b/

   Sea Level Rise by 2075             101      Centimeters

Effects Based on Case Studies and Research in Early Stages

   Cost of Sea Level Rise         1,145-2,807  Present Value
   in Charleston and Galveston c/              (millions of 1985
                                               dollars)

   Reduction in Soybean Seed         14.3      Percent in Year 2075
   Yield d/

   Increase in Ground-Based       5.1-27.2     Percent in Year 2075
   Ozone e/

   Loss of Northern Anchovy       3.8-19.8     Percent in Year 2075
   Population f/

a/  Discounted over 1985-2075 using a real discount rate of 3
    percent.

b/  Estimated using an assumed climate sensitivity of 3°C (middle
    NAS estimate).  Recent analysis indicates that 4°C may be a
    preferred central case assumption.  Using a 4°C sensitivity, the
    estimated equilibrium warming in 2075 is about 8.4°C.

c/  Lowest estimate with anticipation of sea level rise; highest
    estimate without anticipation.

d/  Essex cultivar only in normal years.

e/  Lowest estimate is for Los Angeles, California; highest
    estimate is for Nashville, Tennessee.

f/  Lowest estimate assumes  15-meter vertical mixing of the top
    ocean layer; highest estimate assumes 10-meter vertical mixing.

                    * * *  DRAFT FINAL  * * *

-------
                                   18-17
        •   Costs of sea level rise impacts  were estimated  from
            two case studies (see Chapter 15).   At  a global rise  of
            101 cm, costs for Charleston, S.C.  and  Galveston TX
            range from about $1,145 to $2,800  million (1985 dollars,
            present value at a real 3 percent  discount  rate).  This
            range reflects different assumptions about  the  rate and
            type of adjustments made to prepare for or  respond to
            the impacts of sea level rise.   The major impacts
            include increased storm damage and land lost  to
            increased shoreline retreat (due to inundation  and
            erosion).

        •   Reductions in soybean yield are  based on preliminary
            experiments described in Chapter 11. At the  central
            case global ozone depletion in 2075 of  about  19 percent,
            the reduction in yield is estimated at  of 14.3  percent
            in the year 2075.  This estimate is very uncertain and
            is based on a series of experiments with a  sensitive
            cultivar.   Two out of three cultivars tested  in the
            greenhouse appear sensitive to UV-B. The cultivar
            tested is  one that is becoming preferred in many areas.
            It is included only to indicate  the potential magnitude
            of effects on plants and cannot  be extrapolated to other
            cultivars  or species.

        •   Increase in Ground-based Ozone (smog) is based  on the
            range reported for three urban case studies (see Chapter
            14).  Using the central case estimate of global ozone
            depletion in 2075 (19 percent),  ground-based  ozone will
            rise an estimated 5 to 27 percent  in 2075.  This
            estimate is preliminary and cannot be extrapolated to
            the entire U.S.

        •   Loss of Anchovy Population of about 4 to 20 percent
            is based on experiments in laboratory tanks.  The range
            represents alternative assumptions about the  depth of
            the mixed layer of the oceans and  incorporates  the 2075
            global ozone depletion value (19 percent).  The results
            of the laboratory tank studies cannot be extrapolated to
            natural ecosystems (see Chapter  12).

COMPARISON OF CENTRAL CASE WITH RESULTS USING
ALTERNATIVE ASSUMPTIONS

    This section identifies the sensitivity of central  case risk  estimates  to
assumptions regarding:  (1) the future production,  use  and  emissions  of
ozone-modifying substances; (2) the type of model used  to characterize ozone
depletion; (3) the sensitivity of the global climate to greenhouse gas
forcings; and (4) the concentration of greenhouse gas emissions.
                          * * *  DRAFT FINAL  * * *

-------
                                   18-18
Alternative Scenarios of  Production,  Use, and Emissions

    Four alternative scenarios of the production, use, and emissions of
ozone-modifying substances were explored.  The scenarios (discussed in Chapter
3) reflect a wide range of uncertainties and estimates made by authors from
around the world.  Key factors varied in the development of the scenarios
include:

        •   projected future population;

        •   economic growth;

        •   the relationship between economic growth and
            population; and

        •   technological innovation that would affect the
            intensity of use of ozone-modifying substances.

To investigate the significance of alternative assumptions about future
production and emissions, all other assumptions were held fixed.

    The scenarios of future production and emissions range from 0.0 percent
growth to 5.0 percent growth per year through 2050.  This wide range
represents the results of discussions at the United Nations Environment
Programme (UNEP) Workshop on the Control of Chlorofluorocarbons held in Rome,
Italy in May 1986.  Discussion at the workshop underscored the uncertainty
about potential future production, use, and emissions of ozone-modifying
substances.  The Workshop discussion indicated that it is unlikely that future
production and emissions growth would fall outside this range.

    A narrower range of uncertainty was also presented at the Workshop.  This
range, from 1.2 to 3.8 percent annual growth through 2050, reflects the likely
range of future emissions.  These rates are referred to as the low and high
scenarios respectively.  The extreme rates of 0.0 and 5.0 percent annual
growth are referred to as the lowest7 and highest scenarios respectively.

    Exhibits 18-5 through 18-9 illustrate the sensitivity of risk estimates to
this wide range of assumptions about emissions.  Exhibit 18-5 displays
estimates of ozone depletion for the central case and for the emissions
scenarios.  The lowest emissions scenario (zero percent annual growth) shows a
slight decline in ozone levels until 2030, after which ozone levels increase.

    The high and highest emissions scenarios result in greatly accelerated
ozone depletion.  In contrast to the central case, ozone depletion reaches 30
percent well before 2100 in the high and highest scenarios.  The sensitivity
of ozone depletion to emissions estimates carries over into health risks.  The
    7 The zero percent emissions growth scenario describes what would happen
if production was totally frozen at today's levels.
                          * * *  DRAFT FINAL  * * *

-------
z
o

b

Q.
LJ
O
bJ
Z
o
N
O
-10 -
-15 -
      -20 -
      -25 -
      -30
                                         18-19





                                    EXHIBIT 18-5


                  Global Average  Ozone Depletion:   Emission Scenarios
         1985
               2005
2025
2045
2065
                                    2085
                                *  *  *  DRAFT FINAL  * * *

-------
                                18-20
                            EXHIBIT 18-6

               Human Health Effects:  Emissions Scenarios
Additional Cumulative Cases and  Deaths Over Lifetimes of People Alive Today
      (Figures  in Parentheses are Percent Changes  from  Central Case)
EMISSIONS SCENARIOS
HEALTH EFFECT
Non-Melanoma Skin Tumors
Additional Basal Cases
Additional Squamous Cases
Additional Deaths
Melanoma Skin Tumors
Additional Cases
Additional Deaths
Senile Cataract
Additional Cases
Low Central
265,000 630,600
(-58)
152,600 386,900
(-61)
6,500 16,500
(-61)
5,700 12,300
(-54)
1,700 3,900
(-56)
210,500 593,600
(-65)
High
1,947,700
(209)
1,355,800
(250)
56,800
(244)
33,900
(176)
11,400
(192)
2,124,800
(258)
EXTREME
Lowest
78,700
(-88)
40,900
(-89)
1,700
(-90)
2,100
(-83)
600
(-85)
38,400
(-94)
CASES
Highest
5,167,500
(719)
3,868,800
(900)
161,100
(876)
84,200
(584)
29,000
(644)
5,133,100
(765)
                         * *  DRAFT FINAL * * *

-------
                                     18-21
                                 EXHIBIT 18-7

                    Human Health Effects:  Emissions Scenarios
   Additional Cumulative Cases and Deaths Over Liftimes of People Born  1985-2029
          (Figures in Parentheses are Percent Changes from Central Case)
                                   EMISSIONS  SCNEARIOS             EXTREME CASES
       HEALTH EFFECT           Low       Central      High       Lowest*     Highest


Non-Melanoma Skin Tumors

  Additional Basal Cases       713,800   5,012,900   23,325,300   -265,300   45,210,400
                                 (-86)                   (365)     (-105)        (802)

  Additional Squamous Cases    369,100   3,185,800   18,418,500   -161,100   37,279,900
                                 (-88)                   (478)     (-105)      (1,070)
Additional Deaths
Melanoma Skin Tumors
Additional Cases
Additional Deaths
Senile Cataract
Additional Cases
16,000 135,000
(-88)
20,500 109,800
(-81)
5,300 32,200
(-84)
389,600 3,463,400
(-89)
763,000
(465)
466,500
(325)
137,200
(326)
12,243,400
(254)
-6,600
(-105)
-4,200
(-104)
-1,500
(-105)
-262,800
(-108)
1,538,200
(1,039)
948,400
(764)
266,800
(729)
20,285,300
(486)
     * Negative numbers indicate that  fewer  cases  and/or deaths are estimated to
occur compared to the expected number  of  cases  and/or deaths that would arise if
total column ozone levels are unchanged.
                            * * *  DRAFT FINAL  *  *  *

-------
                                         18-22
                                   EXHIBIT 18-8

                      Human Health Effects:   Emissions  Scenarios
    Additional  Cumulative Cases and  Deaths Over  Lifetimes of  People  Born 2030-2074
            (Figures in Parentheses are Percent Changes from Central Case)
                                    EMISSIONS SCENARIOS               EXTREME CASES
       HEALTH EFFECT           Low       Central       High        Lowest*      Highest


Non-Melanoma Skin Tumors

  Additional Basal Cases      620,700   17,630,500   58,482,700    -1,429,500   84,939,900
                                 (-96)                      (232)        (-108)         (382)

  Additional Squamous Cases   301,700   12,122,400   52,033,800      -736,400   76,661,100
                                 (-98)                      (329)        (-106)         (532)

  Additional Deaths            13,200      509,300    2,132,600       -31,800    3,138,100
                                 (-97)                      (319)        (-106)         (516)

Melanoma Skin Tumors

  Additional Cases             21,100      430,500    1,397,000       -38,400    2,046,900
                                 (-95)                      (225)        (-109)         (375)

  Additional Deaths             5,100      115,100      354,200       -10,300      513,900
                                 (-96)                      (208)        (-109)         (346)

Senile Cataract

  Additional Cases            284,900    8,295,800   19,858,700      -887,700   28,011,400
                                 (-97)                      (139)        (-111)         (238)


     * Negative numbers indicate that fewer cases and/or deaths  are estimated  to  occur
compared to the expected number of cases and/or deaths that would arise if  total  column
ozone levels are unchanged.
                                * * *  DRAFT FINAL  * * *

-------
                                   18-23
highest emissions scenario, for example, leads to a larger increase in UV
radiation reaching the earth's surface and to correspondingly larger health
risks.  In contrast, the lowest scenario results in an ozone abundance after
2035 and thus a reduction in UV radiation.

    For purposes of evaluating potential health effects, the maximum ozone
depletion was (arbitrarily) set to 50 percent.  Depletion exceeding 50 percent
was considered to be outside the range for which the atmospheric models are
valid.  This arbitrary cutoff affected ozone depletion estimates in the high
and highest emissions scenarios only.  Once 50 percent depletion is reached,
depletion is held constant for the remainder of the analysis period.

    Exhibits 18-6, 18-7, and 18-8 present estimates of the health risks across
the emissions scenarios for the three population cohorts.  The exhibits
provide a range of risk estimates reflecting only the differences in the
emission scenarios.  The exhibits illustrate how risk estimates respond to
assumptions about emissions.

    The low and lowest emissions scenarios result in fewer additional
cumulative cases and deaths related to non-melanoma skin cancer, melanoma skin
cancer, and senile cataract.  The reduction in risk results from lower ozone
depletion (and UV radiation) in these scenarios.  For example, comparing the
lowest and central emissions scenarios in Exhibit 18-6, people alive today
would experience an estimated 88 percent fewer additional cases of basal cell
skin cancer if the lowest scenario is used.

    The high and highest emissions scenarios produce parallel increases in
cases and deaths.  People alive today would experience an estimated 7-fold
increase in additional cases of basal non-melanoma skin cancers if emissions
grow at 5 percent (highest emissions scenario) instead of 2.5 percent (central
case).

    Exhibits 18-7 and 18-8 show the results for the latter two cohorts
respectively.  Compared to the results for the cohort of people alive today
and the cohort of people born from 2030 to 2075 (last cohort), the results for
the people born between now and 2029 (middle cohort) indicate a larger
sensitivity.  For example, using the highest emissions scenario would increase
cases of basal cell skin cancers by 800 percent for the middle cohort but by
only 720 percent for people alive today and 380 percent for the last cohort.

    The increased sensitivity of the risk estimates to emissions for the
middle cohort stems from the exposure of this cohort to larger changes in UV
radiation and the longer exposure of individuals in the cohort to increased
UV.  The reduced sensitivity for the last cohort is an artifact of the method
used to model risks.  The risk estimates for this cohort are affected by the
arbitrary 50 percent cap placed on ozone depletion.  This assumption dampens
the significance of emissions assumptions particularly for individuals born
near 2075.
                              *  DRAFT FINAL  * * *

-------
                                   18-24
    The risk estimates for the high and highest emissions scenarios should be
treated with caution because the dose-response coefficients used to derive
risk estimates were estimated over a limited range of UV radiation.  The
dose-response parameters may not be valid for the extreme changes in UV
radiation expected in these scenarios.

    Exhibit 18-9a displays estimates of non-health risks for the emissions
scenarios.  As with the health risks, the range of effects is very large.   Of
note is that the global warming estimates for 2075 range from 4.9°C in the
lowest scenario to over 12°C in the highest.  Exhibit 18-9b shows the
greenhouse equilibrium temperature increase over time for the five scenarios.

Comparison with Results from a 2-Dimensional Atmospheric Model

    A variety of models have been developed to assess potential changes in the
stratosphere due to emissions of CFCs and other ozone-modifying substances
(see Chapter 5).  The central case risk estimates rely on a parameterized
representation of a one-dimensional model of the atmosphere with latitudinal
gradients generated to conform to Isaksen's latitudinal distribution (see
Chapter 17).  For comparison purposes,  the results of a time-dependent
two-dimensional model developed by Isaksen (1986) are presented here.
Two-dimensional models are believed to be preferable for risk assessment (see
Chapter 5).

    The results of Isaksen's analysis are currently only available through
2030.  Therefore, a complete comparison of risks is not possible at this time
(as additional ,data become available, a full comparison of risks will be
performed).

    Exhibit 18-10 displays estimates of ozone depletion from the one- and
two-dimensional atmospheric models for two scenarios of emissions.  The first
scenario is a 3.0 percent growth rate of CFCs-11, 12, and 22, methyl
chloroform, and carbon tetrachloride.  This scenario is not directly
comparable to the central case estimates described above  (which includes more
compounds, and grows at 2.5 percent per year).  The exhibit shows that the
two-dimensional model results in larger estimates of ozone depletion than
does the parameterized representation of the one-dimensional model for this
scenario.

    The second scenario shown in the exhibit is a 0.0 percent per year growth
scenario.  Comparing the one-and two-dimensional model estimates, the results
from Isaksen's two-dimensional model show more depletion.

    In general, two-dimensional models appear to produce  larger estimates of
ozone depletion than do one-dimensional models (see Chapter 5).  Therefore,
the use of a one-dimensional model in the central case will bias risk
estimates downward, perhaps by as much as one-half.
                          * * *  DRAFT FINAL  * * *

-------
                                                 EXHIBIT 18-9a
                          Materials,  Climate,  and Other Effects:   Emissions Scenarios
                       (Figures in Parentheses are Percentage' Changes from Central  Case)
TYPE OF EFFECT
Effects Estimated Quantitatively
for the U.S.
Materials Damage a/
Ri se in Equi 1 ibr ium
Temperature by 2075
Sea Level Rise by 2075
Effects Based on Case Studies and
Research in Early Stages

Low
268
(-51)
5.3
(-14)
94
(-7)

EMISSIONS SCENARIOS
Centra 1
550
6.2
101


High
907
(65)
8.2
(32)
115
(14)

EXTREME
Lowest
188
(-66)
4.9
(-42)
91
(-10)

CASES
Highest
1336
(143)
12.2
(97)
140
(39)

UNITS
Present Va I ue
(mill ions of
1985 dol lars)
Degrees Centi-
grade
Cent (meters

   Cost of Sea Level  Rise in
   Charleston and Galveston b/
Reduction in Soybean Seed
Yield c/
Increase in Ground-Based
Ozone d/
Loss of Northern Anchovy
Population e/
1094-2673   1145-2807   1243-3076    1073-2616

    1.4        14.3        >19           £/

  0.5-2.6    5.1-27.2   >9.4->50.0       £/

    0.0      3.8-19.8   >11.0->25.0      f/
                                                                                   1420-3556
                                                                                      >9.4->50.0
Present Value
(mill ions of
1985 dollars)
Percent in Year
2075
Percent in Year
2075
                                                                                      >11.0->25.0 Percent in Year
                                                                                                   2075
a/ Discounted over 1985-2075 using  a  real  discount  rate  of 3  percent.
b/ Lowest estimate with anticipation  of sea  level  rise;  highest  estimate  without anticipation.
c/ Essex cultivar only in normal  years.
d/ Lowest estimate is for Los Angeles,  California;  highest estimate is for Nashville,  Tennessee.
e/ Lowest estimate assumes 15-meter vertical  mixing of the top ocean  layer;  highest  estimate  assumes 10-meter
   ve rt i ca I  mixing.
£/ Impact not evaluated for increased ozone  or decreased UV-B radiation projections.
oo
 i
N3
                                           * * *  DRAFT  FINAL  *  * *

-------
                                         18-26







                                    EXHIBIT 18-9b



                 Equilibrium Temperature Change:  Emissions Scenarios*
o


13
UJ
cc
o
uj
Q
Ul
o
o

LJ

K.
UJ

Q.



U

I-
                                                                             Central
                                                                Low
         1985
1995
2005
2015
2025
2035
2045
2055
2065
2075
          * Computed assuming that the climate sensitivity to a doubling of carbon

      dioxide is 3°C.   This assumption is in the middle of the NAS range of 1.5°C to

      4.5°C (see Chapter 6).   Note that the actual warming that may be realized will

      lag by several decades  or more.
                                * * *  DRAFT FINAL  * * *

-------
                                           18-27
                                      EXHIBIT 18-10


                      Global Average Ozone Depletion:   Comparison to
                      Results with  a  2-Dimensional Atmospheric Model
z
o
Q.
u
a
LJ
z
o
N
O
                                                         e	B-
                                                            2-D Constant Emissions Growth
                                      1-D  Constant Emissions Growth
                                               2-D 3 Percent  Emissions Growth
                                                   1-D 3 Percent
                                                  Emissions  Growth
         1985
1995
                                        2005
                                                       2015
                                                                       2025
                                 * * *  DRAFT FINAL   *  * *

-------
                                   18-28
Alternative Assumptions about Sensitivity of Atmosphere to
Greenhouse Gases

     The central case assumes that the sensitivity of the global climate
system is such that a doubling of atmospheric C02 would result in a 3°C
increase in the average global temperature.  This climate sensitivity is the
middle of a range published by the NAS in 1983 of 1.5°C to 4.5°C.  Exhibit
18-11 displays estimates of temperature-sensitive effects for the limits of
this NAS range.  As expected, the warming and sea level rise estimates are
lower for the 1.5°C sensitivity and higher for the 4.5°C sensitivity.

     Also shown in the exhibit are estimates  for a 4°C sensitivity.  Recent
analyses indicate that 4°C may be a preferred central case assumption for
climate sensitivity (see Chapter 6).

Alternative Assumptions about Emissions  of Trace Gases

     Uncertainty exists about potential future changes in atmospheric
concentrations of the greenhouse gases:  carbon dioxide (C02); methane (CH4);
and nitrous oxide (N20).  This uncertainty arises from at least two sources.
First, to limit the potental for future global warming, governments may agree
to take steps to limit growth in the concentrations of these gases.  However,
the timing and stringencies of such potential steps are not known.  Second,
future concentrations of methane are influenced by numerous natural as well as
anthropogenic factors that have not been well characterized to date (see
Chapter 4).  Therefore, the rate of future increases in methane concentrations
is particularly uncertain even in the absence of potential government
regulations to limit global warming.

     Through their roles in chemical reactions in the stratosphere, trace
gases affect both total column ozone and the potential for future global
warming.  Exhibit 18-12 summarizes the implications of changes in
concentrations of the gases for ozone depletion and global warming.  In
general, each greenhouse gas increases global temperature but counters ozone
depletion.

     Any risk assessment requires an assumption about future global warming
and whether governments take actions to limit warming.  The standard
assumption generally has been that warming is never limited.  To examine the
sensitivity of the risk estimates to this issue, the implications of reducing
the growth in concentrations of the trace gases to limit warming to 3°C by
2075 was examined.  In this case the growth rate of each of the three gases
was reduced to approximately 20 percent of its central case growth rate after
2000.  Because the trace gases counter depletion, lowering rates of growth in
concentrations to limit warming to 3°C by 2075 results in estimates of more
ozone depletion than in the central case.

     Exhibits 18-13 to 18-17 compare the central case results to the results
if global warming is limited to 3°C.  In Exhibit 18-13 the increase in ozone
depletion over the central case is substantial; by 2050 the difference exceeds
5 percent depletion.  Exhibits 18-14 to 18-16 show the increased health risks
                          * * *  DRAFT FINAL  * * *

-------
                                              EXHIBIT 18-11

                                  Materials,  Climate, and Other Effects:
                   Sensitivity to Relationship Between Climate Change and C02 Emissions
                    (Figures in Parentheses are Percentage Changes from Central  Case)
Type of Effect
Effects Estimated Quantitatively
for the U.S.
Rise in Equi I ibrium
Temperature by 2075
Sea Level Rise by 2075

1.5 C
3.0
(-52)
80
(-21)
DOUBLED C02
Centra 1 (3 C) (U C)
6.2 8. U
(35)
101 110
(9)

U.5 C
9.5
(53)
11U
(13)
UNITS
Degrees Centigrade
Cent imeters
Effects Based on Case Studies
and Research In Early Stages

   Cost of Sea Level  Rise in
   Charleston and Galveston a/
985-2405
11U5-2807
1208-2980
1236-3057
Present VaIue
(mill ions of 1985
dollars)
                                                                                                                    oo
                                                                                                                    i
a./ Lowest estimate with anticipation of sea  level  rise;  highest estimate without anticipation.
                                        * * *  DRAFT FINAL  * * *

-------
                              18-30
                        EXHIBIT 18-12

    Summary of Effects  of Greenhouse Gases on Ozone Depletion
                and Global  Equilibrum Temperature
   GREENHOUSE GAS         STRATOSPHERIC OZONE       GLOBAL TEMPERATURE
Methane (CH4)            Counters Depletion              Increases

Nitrous Oxide  (N20)      Counters Depletion              Increases
                        in  High Chlorine Cases

Carbon Dioxide (C02)     Adds Ozone                      Increases
                       * *  DRAFT  FINAL  * * *

-------
                                        18-31
                                   EXHIBIT  18-13


                  Global Average Ozone Depletion:   Scenario of  Limits
                               to Future Global Warming
z
o

uj
_j
Q.
U
Q
UJ
z
0
N
O
         1935
2005
2025
2045
2065
2085
                                * * *  DRAFT FINAL  *  *  *

-------
                                  18-32
                              EXHIBIT 18-14

    Human Health Effects:   Scenarios of Limits  to Future Global Warming
Additional  Cumulative Cases and Deaths Over Lifetimes of People Alive Today
    (Figures in Parentheses are Percentage Changes from Central Case)
              HEALTH EFFECT
                                      LIMITS TO FUTURE GLOBAL WARMING
   3°C
Central Case
 (no limit)
         Non-Melanoma Skin Tumors
            Additional Basal  Cases
            Additional Basal  Cases
            Additional Deaths
         Melanoma Skin Tumors
            Additional Cases
            Additional Deaths
         Senile Cataract
            Additional Cases
1,001,800
   (59)

  632,700
   (64)

   26,800
   (62)
   18,900
   (54)

    6,200
   (59)
  970,900
   (64)
  630,600


  386,900


   16,500
   12,300


    3,900
  593,600
                         * * *  DRAFT FINAL  * * *

-------
                                   18-33
                              EXHIBIT 18-15

     Human Health Effects:  Scenarios of Limits to Future  Global Warming
Additional  Cumulative Cases and Deaths  Over Lifetimes of People Born  1985-2029
     (Figures in Parentheses are Percentage Changes from Central Case)
                                      LIMITS TO FUTURE GLOBAL WARMING
                                                      Central Case
               HEALTH EFFECT               3°C         (no limit)
          Non-Melanoma Skin Tumors

             Additional Basal  Cases       10,679,700      5,012,900
                                           (113)

             Additional Squamous  Cases     7,781,500      3,185,800
                                           (144)

             Additional Deaths             324,900        135,000
                                           (141)

          Melanoma Skin Tumors

             Additional Cases               215,000        109,800
                                           (96)

             Additional Deaths              64,500         32,200
                                           (100)

          Senile Cataract

          .   Additional Cases              6,899,400      3,463,400
                                           (99)
                          * * *  DRAFT  FINAL  * * *

-------
                                   18-34
                               EXHIBIT 18-16

     Human Health Effects:  Scenarios of Limits  to Future  Global Warming
Additional  Cumulative Cases and Deaths  Over Lifetimes of  People Born  2030-2074
     (Figures in Parentheses are Percentage Changes from Central Case)
                                      LIMITS TO FUTURE GLOBAL WARMING
                                                      Central Case
               HEALTH EFFECT                3°C         (no limit)
          Non-Melanoma Skin Tumors

             Additional Basal  Cases       39,303,000      17,630,500
                                           (123)

             Additional Squamous  Cases    33,400^200      12,122,400
                                           (176)

             Additional Deaths .           1,374,300         509,300
                                           (170)

          Melanoma Skin Tumors

             Additional Cases                903,700         430,500
                                           (110)

             Additional Deaths              237,800  .       115,100
                                           (107)

          Senile Cataract

             Additional Cases             14,947,300       8,295,800
                                            (80)
                          * * *  DRAFT FINAL  *  * *

-------
                                       EXHIBIT 18-17

    Materials, Climate, and Other Effects:  Scenarios of Limits to Future Global Warming
             (Figures  in Parentheses are Percentage' Changes from Central Case)
      TYPE OF EFFECT
                               LIMITS TO FUTURE GLOBAL WARMING
                                     3 C      Central  Case
                                               (no limit)
                                                                       UNITS
Effects Estimated Quantitativetv for the U.S.

   Materials Damage a_/             726
   Rise in Equilibrium
   Temperature by 2075

   Sea LeveI  R i se by 2075
                                   3.0
                                  (-52)
                                    76
                                  (-25)
Effects Based on Case
Studies and Research in Early Stages
   Cost of Sea Level Rise in
   Charleston and Galveston b/
   Reduction in Soybean Seed
   Yield c/

   Increase in Ground-Based
   Ozone d/

   Loss of Northern Anchovy
   Populat ion e/
                                 967-2328
                                  >19.0
   550


   6.2


   101




1145-2807



It.3
                                >9.U->50.0    5.1-27.2
                                >11.0->25.0   3.8-19.8
Present Value
(millions of 1985 dollars)

Degrees Centigrade


Cent imeters
Present Value
(mill ions of 1985
dotlars)

Percent in Year 2075
              Percent in Year 2075
              Percent in Year 2075
03
 I

l/l
a./ Discounted over 1985-2075 using a real discount rate of 3 percent.

b/ Lowest estimate with anticipation of sea level  rise; highest estimate without
   antic ipat ion.

c/ Essex cultivar only in normal  years.

d_/ Lowest estimate is for Los Angeles,  California; highest estimate  is for Nashville,
   Tennessee.

e/ Lowest estimate 15-meter vertical mixing of the top ocean layer; highest estimate
   10-meter vertical  mixing.
                                 * * *  DRAFT FINAL  * * *

-------
                                   18-36
expected for the three population cohorts.   The people alive today may
experience an increase in health risks on the order of 60 percent relative  to
the central case values.  The latter cohorts are more sensitive.

     Exhibit 18-17 displays non-health risks.  As shown in the exhibit,
warming is reduced from 6.2°C in the central case to 3°C.  As a result,  sea
level rise impacts are also reduced.  The ozone-sensitive risks are increased,
however.  If global warming were held to less than 3°C by 2075 the risks from
ozone depletion would be even greater.

     A second sensitivity case is examined in which several alternative
scenarios of methane concentrations are examined.  To evaluate the sensitivity
of the results to the methane growth rates, three methane scenarios were
examined in addition to the central case.  Other trace gases were kept at
constant values consistent with the central case.  In the lowest  scenario,  the
methane concentration grows through 2010 at 1 percent per year, and then is
held constant.  In the low scenario, the methane concentration grows at  1
percent through 2000, then grows at about 0.6 percent through 2050, after
which it remains constant.  In the highest scenario, the methane
concentration grows at 1 percent per year through 2020, after which it grows
more rapidly than 1 percent.8

     Exhibit 18-18 shows that higher concentrations of atmospheric methane
would reduce ozone depletion, while lower concentrations result in more
depletion.  Contrasting the highest methane scenario to the central case,
ozone depletion is roughly equal until 2025, when the contributions to ozone
from increased methane concentrations begin to dampen depletion.   The
influence of increased methane concentration becomes more pronounced after
2025, and depletion reaches an inflection point in 2085 after which ozone
levels rise.

     The lower methane scenarios cause the ozone depletion estimates to
increase.  The implications of the resulting differences in ozone depletion
for the risk estimates is shown in Exhibits 18-18 through 18-22.   Of note in
Exhibit 18-22 is that the lower methane scenarios reduce the estimate of
global warming, while the high scenarios increase it.

SENSITIVITY OF  EFFECTS TO PARAMETER UNCERTAINTY

     This section analyzes the sensitivity of risk estimates to statistical
uncertainties in two areas:  (1) estimates of dose-response between UV
radiation and health effects, and (2) the relationship between ozone depletion
and atmospheric concentrations of chlorine (i.e., the sensitivity of the
atmosphere to total column ozone depletion from chlorine sources).  These
uncertainties are investigated together because the parameters describing the
dose-response and the depletion-emissions relationships were estimated using
statistical methods the results of which included statistical estimates  of
lower- and upper-bound parameters.
       See Chapter 4 for a discussion of the sources of these scenarios,


                          * * *  DRAFT FINAL  * * *

-------
                                      18-37




                                 EXHIBIT 18-18

             Global Average  Ozone Depletion:   Methane Emissions  Cases
        o -»
z
o
OL
LJ
Q
UJ
Z
o
N
0
         1985
2005
2025
2045
2065
                                    2085
                             * * *  DRAFT FINAL  * * *

-------
                                 18-38
                            EXHIBIT 18-19

              Human  Health  Effects:   Methane Emissions Cases
Additional Cumulative Cases  and Deaths Over Lifetimes of  People Alive Today
    (Figures in  Parentheses  are Percentage Changes from  Central Case)
                                       METHANE EMISSIONS  SCENARIOS
            HEALTH  EFFECT
Lowest
Low
Central
 Case
     Senile Cataract
       Additional  Cases
High
Non-Melanoma Skin Tumors
Additional Basal Cases
Additional Squamous Cases
Additional Deaths
Melanoma Skin Tumors
Additional Cases
Additional Deaths
818,200
(30)
512,100
(32)
21,700
(32)
15,500
(26)
5,000
(28)
757,000 630,600
(20)
469,200 386,900
(21)
19,900 16,500
(21)
14,600 12,300
(19)
4,700 3,900
(21)
611,600
(-3.0)
372,500
(-3.7)
15,900
(-3.6)
12,100
(-1.6)
3,800
(-2.6)
791,700   720,400    593,600   562,700
 (33)      (21)               (-5.2)
                          * *  DRAFT FINAL  * * *

-------
                                   18-39
                              EXHIBIT 18-20

               Human Health  Effects:  Methane  Emissions Cases
Additional Cumulative Cases and Deaths Over Lifetimes of People Born 1985-2029
     (Figures in Parentheses are Percentage Changes from Central Case)
                                        METHANE EMISSIONS SCENARIOS
                                                        Central
         HEALTH EFFECT          Lowest        Low        Case        High
  Non-Melanoma Skin Tumors

    Additional Basal Cases      8,055,800   7,421,300   5,012,900   3,825,400
                                  (61)        (48)                    (-24)

    Additional Squamous  Cases   5,600,900   4,140,200   3,185,800   2,284,800
                                  (76)        (61)                    (-28)

    Additional Deaths           .  235,000     215,800     135,000      97,600
                                  (74)        (60)                    (-28)

  Melanoma Skin Tumors
Additional Cases
Additional Deaths
Senile Cataract
Additional Cases
166,500
(52)
49,700
(54)
5,492,900
(59)
152,500
(39)
45,700
(42)
5,192,200
(50)
109,800 89,600
(-18)
32,200 25,500
(-21)
3,463,400 2,450,700
(-29)
                          * *  *   DRAFT FINAL  * * *

-------
                                   18-40
                              EXHIBIT 18-21

               Human Health Effects:  Methane Emissions Cases
Additional  Cumulative Cases and  Deaths  Over  Lifetimes  of People Born 2030-2074
     (Figures in Parentheses are Percentage Changes from Central Case)
                                         METHANE EMISSIONS SCENARIOS
                                                        Central
          HEALTH EFFECT           Lowest        Low        Case        High


   Non-Melanoma Skin Tumors

     Additional Basal Cases     33,293,900   32,360,400   17,630,500  8,848,100
                                   (89)         (84)                   (-50)

     Additional Squamous Cases  27,448,800   26,581,900   12,122,400  5,231,900
                                  (126)        (119)                   (-59)

     Additional Deaths           1,132,500    1,097,100     509,300    223,600
                                  (122)        (115)                   (-56)

   Melanoma Skin Tumors

     Additional Cases              751,500      726,400     430,500    243,300
                                   (75)         (69)                   (-43)

     Additional Deaths             201,800      196,000     115,100     62,600
                                   (75)         (70)                   (-46)

   Senile Cataract

     Additional Cases           13,686,000   13,479,800   8,295,800   4,247,100
                                   (65)         (62)                   (-49)
                                 DRAFT FINAL  * * *

-------
                                           EXHIBIT 18-22
                  Materials,  Climate,  and  Other Effects:   Methane Emissions  Cases
                 (Figures in  Parentheses are  Percentage Changes  from Central  Case)
                                            METHANE  EMISSION SCENARIOS
        TYPE OF EFFECT
 Lowest
 Low
           Centra I
            Case
  High
                                                                                          UNITS
Effects Estimated Quantitatively
for the U.S.
Materials Damage a/
Rise in Equilibrium
Temperature by 2075 b/
Sea Level Rise by 2075
Effects Based on Case Studies
and Research in Early Stages
661
(20)
5.9
(-5)
98
(-3)

617
(12)
6.0
(-3)
99
(-2)

550 529
(-4)
6.2 6.4
(3)
101 102
(D

Present Value
(mill ions of 1985
do I lars )
Degrees Cent
Cent i meters

i grade


   Cost of Sea Level  Rise in
   Charleston and Galveston c/
   Reduction in Soybean
   Seed Yield d/
   Increase in Ground-Based
   Ozone e/
   Loss of Northern Anchovy
   Population f/
1123-2750   1123-2770   1145-2807

                           14.3

                        5.1-27.2
  7.7-41.1
7.4-39.1
>11.0->25.0 >11.0->25.0 3.8-19.8
 1151-2827   Present Value
             (mill ions of 1985
             dotlars)
    10.6     Percent in Year 2075

 3.8-20.0    Percent in Year 2075

0.2-12.1    Percent in Year 2075
                                                                                                                 oo
                                                                                                                 i
a/ Discounted over 1985-2075 using  a  real  discount  rate  of  3  percent.
Jb/ Estimates of the rise in equilibrium temperature  neglect increases  and  decreases  in  tropospheric
   that would occur with deviations from the  central  case of  methane growth  (1%)  and  changes  in
   stratospheric water vaor.
c/ Lowest estimate with anticipation  of sea  level  rise;  highest  estimate without  anticipation.
d/ Essex cultivar only in normal  years.
e/ Lowest estimate is for Los Angeles,  California; highest  estimate  is for Nashville, Tennessee.
f/ Lowest estimate 15-meter vertical  mixing of  top ocean layer;  highest estimate  10-meter vertical
   mixing.
                                     *  * * DRAFT  FINAL  *  *  *

-------
                                   18-42
     The statistical uncertainty reported here does not reflect uncertainties
in the underlying specifications of the dose-response and ozone-depletion
relationships.  For example, as described in Chapter 8 there remains a major
question regarding the appropriate measure of exposure to use for modeling
melanoma skin cancer risks.  The uncertainty estimates reported below reflect
the implications of the statistical variation reported using one particular
measure of exposure.  Using alternative exposure measures would increase the
range of risk estimates; preliminary results from ongoing work indicate that
melanoma mortality risks may be as much as 60 percent lower than the values
reported here if an alternative specification of exposure is used (annual UV
flux in the place of peak UV flux).  Consequently, the full range of
uncertainty in these relationships (including specification uncertainty) is
larger than the values reported here, which reflect only parameter uncertainty.

Effects  at  High and Low Estimates  of  UV  Radiation Impact on
Human Health

     The dose-response relationships between increases in UV radiation and
associated health risks were obtained using a variety of statistical
techniques (see Chapters 7, 8, 10 and 17).  In these studies, the
responsiveness of health effects to changes in UV radiation were characterized
by dose-response parameters estimated using mathematical functions fitted
statistically to detailed epidemiologic data.  The results of the statistical
analysis included estimates of the standard errors of dose-response parameters.

     The standard errors of the dose-response parameters were used to develop
high and low estimates of health risks.  Low risk estimates use the central
case dose-response coefficients minus one standard error; high risk estimates
use central case coefficients plus one standard error.  All other assumptions
are held fixed.

     Health risks estimated using the low, central, and high coefficients .are
presented in Exhibits 18-23 through 18-25.  For people alive today, the low
dose-response coefficients produced approximately 30 to 45 percent fewer cases
of non-melanoma skin tumors, melanoma skin tumors, and senile cataract
(Exhibit 18-23).  Deaths due to non-melanoma skin tumors, however, decrease by
81 percent in the low case.  The larger reduction in deaths results from a
combination of two factors:  (1) a lower dose-response coefficient for
non-melanoma tumors, and (2) an assumption in the low case that a smaller
fraction of individuals contracting tumors would die (0.28 percent in place of
the 1.0 percent used in the central case  -- see Chapters 7 and 17).  About 33
percent more cases among people alive today would occur if the high estimates
of dose-response coefficients were used.

Effects  at  High and Low Sensitivities  of Ozone Depletion
to Emissions

     A previous section analyzed the sensitivity of risk estimates to
assumptions about the levels of emissions of ozone-modifying substances.
This section explores another source of uncertainty in the risk estimates:
the relationship .between emissions (and their resulting chlorine
concentrations) and ozone depletion.  This uncertainty was quantified using
Monte Carlo modeling techniques to reflect the implications of the convolution

                          -•• * *  DRAFT FINAL  * * *

-------
                                 18-43
                            EXHIBIT 18-23

     Human Health  Effects:  Sensitivity to Dose-Response Relationship
Additional Cumulative Cases  and Deaths Over Lifetimes of People Alive  Today
    (Figures in  Parentheses  are  Percentage Changes from Central Case)
HEALTH EFFECT
Non-Melanoma Skin Tumors
Additional Basal Cases
Additional Squamous Cases
Additional Deaths
Melanoma Skin Tumors
Additional Cases
Additional Deaths
Senile Cataract
Additional Cases
SENSITIVITY
Low
407,500
(-35)
262,500
(-32)
3,100
(-81)
8,900
(-28)
3,500
(-10)
325,600
(-45)
OF EFFECT TO UV DOSE
Central High
630,600 857,400
(36)
386,900 515,400
(33)
16,500 22,000
(33)
12,300 15,700
(28)
3,900 4,300
(10)
593,600 801,900
(35)
                        * * *  DRAFT FINAL  * *  *

-------
                                   18-44
                              EXHIBIT 18-24

      Human Health Effects:   Sensitivity to Dose-Response Relationship
Additional  Cumulative Cases and Deaths Over Lifetimes of People Born  1985-2029
     (Figures in Parentheses are Percentage Changes from Central Case)
                                       SENSITIVITY OF EFFECT TO UV DOSE
            HEALTH EFFECT               Low         Central        High
     Non-Melanoma Skin Tumors

       Additional Basal Cases         3,142,500     5,012,900     7,044,100
                                       (-37)                       (41)

       Additional Squamous  Cases      2,028,100     3,185,800     4,540,000
                                       (-36)                       (43)

       Additional Deaths        .        24,000        135,000       192,100
                                       (-82)                       (42)

     Melanoma Skin Tumors

       Additional Cases                 78,000        109,800       142,200
                                       (-30)                       (30)

       Additional Deaths                28,800        32,200        35,600
                                       (-11)                       (ID

     Senile Cataract

       Additional Cases              1,872,500     3,463,400     4,732,800
                                       (-46)                       (37)
                            * *  DRAFT FINAL  * * *

-------
                                   18-45
                              EXHIBIT 18-25

      Human Health Effects:   Sensitivity  to  Dose-Response Relationship
Additional  Cumulative Cases and  Deaths Over Lifetimes  of People Born 2030-2074
     (Figures in Parentheses are Percentage Changes from Central Case)
                                        SENSITIVITY OF EFFECT TO UV DOSE
           HEALTH EFFECT               Low          Central          High
    Non-Melanoma Skin Tumors

      Additional Basal Cases         10,496,400      17,630,500     26,229,000
                                      (-40)                          (49)

      Additional Squamous Cases       6,951,400      12,122,400     19,329,900
                                      (-43)                          (59)

      Additional Deaths                 82,100        509,300        806,200
                                      (-84)                          (58)

    Melanoma Skin Tumors

      Additional Cases                 297,400        430,500        574,200
                                      (-31)                          (33)

      Additional Deaths                102,000        115,100        128,600
                                      (-11)                          (12)

    Senile Cataract

    .  Additional Cases               4,402,100       8,295,800     11,503,600
                                      (-47)                          (39)
                          * * *  DRAFT FINAL  * * *

-------
                                   18-46
of the individual ranges of uncertainty that surround estimates of rate
constants for atmospheric reactions.  The range analyzed here is the tenth and
ninetieth percentile estimates.

     The effect of varying the sensitivity of ozone depletion to emissions is
illustrated in Exhibit 18-26.  The exhibit shows that the tenth percentile
estimate of ozone depletion is about one half of the central case value.   The
ninetieth percentile estimate is about twice the central case value and
reaches 30 percent depletion in approximately 2067.

     The implications for health risks are displayed in Exhibits 18-27  to
18-29 for each of the three population cohorts.  In general, increased  health
risks would be 60 percent lower than central case in the tenth percentile
case, and roughly double the increased risks in the ninetieth percentile  case.

RELATIVE IMPORTANCE OF  KEY  UNCERTAINTIES

     The analysis presented in this chapter suggests that the estimates of
risks to human health and the environment are sensitive to the assumptions,
methods, and data used to model risks.  Changes in the assumptions, methods,
and data were found to alter the risk estimates significantly.

     It appears that all risks are underestimated by using the 1-D
parameterization to model ozone depletion instead of a 2-D model; the risks
may be underestimated by as much as one-half.

     The largest range of risk estimates was produced by varying assumptions
about future production, use, and emissions of ozone-modifying substances.
Estimates of ozone depletion were particularly sensitive to these
assumptions.  The lowest case emissions scenario leads to a projected
increase in ozone levels; the highest case scenario results in extremely
rapid depletion.  In addition, the high and highest emissions scenarios
caused ozone depletion estimates to reach an amount believed to be outside the
valid range of the atmospheric models used to analyze changes in total  column
ozone.

     Because estimates of human health effects are driven by projections  of
ozone depletion, health risks were similarly affected by this range of
emissions assumptions.  Increased risks to people alive today were found  to
vary by as much as 500 percent.  The sensitivity of health risks to emissions
is greater in future generations; individuals yet to be borne were simulated
to experience greater changes in exposure to UV radiation over a larger
fraction of their lives.  Non-health effects were also shown to be most
sensitive to emissions assumptions; however, compared to health effects the
range of results was narrower.

     After assumptions about emissions, assumptions about potential actions
taken to limit global warming were found to be most important for estimating
health risks.  Actions aimed at limiting future warming to 3°C would reduce
future potential concentrations of greenhouse gases.  Because these gases also
counter ozone depletion, an increase in ozone depletion and associated UV
                                 DRAFT FINAL  * * *

-------
                                       18-47
                                   EXHIBIT 18-26


                    Global Average Ozone Depletion:   Sensitivity
               to Relationship  Between Ozone  Depletion and Emissions
z
o

Id

Q.
UJ
Q
u
z
0
N
0
         1985
2005
2025
2045
2065
2085
                               * * *  DRAFT  FINAL  * *- *

-------
                                 18-48
                            EXHIBIT 18-27

                          Human Health Effects:
     Sensitivity to  Relationship  Between Ozone  Depletion and Emissions
Additional Cumulative Cases  and Deaths Over Lifetimes of People Alive Today
    (Figures in Parentheses  are  Percentage Changes from Central Case)
SENSITIVITY OF OZONE DEPLETION TO EMISSIONS
HEALTH EFFECT
Non-Melanoma Skin Tumors
Additional Basal Cases
Additional Squamous Cases
Additional Deaths
Melanoma Skin Tumors
Additional Cases
Additional Deaths
Senile Cataract
Additional Cases
Low
245,000
(-62)
147,700
(-62)
6,300
(-62).
4,800
(-61)
1,500
(-62)
232,400
(-61)
Central High
630,600 1,327,600
(111)
386,900 838,600
(117)
16,500 35,500
(115)
12,300 25,500
(107)
3,900 8,200
(110)
593,600 1,242,100
(109)
                               DRAFT FINAL  * * *

-------
                                  18-49
                              EXHIBIT 18-28

                           Human Health  Effects:
      Sensitivity to  Relationship  Between Ozone  Depletion and Emissions
Additional Cumulative Cases and  Deaths Over Lifetimes of People Born  1985-2029
     (Figures in Parentheses  are Percentage Changes from Central Case)
SENSITIVITY OF OZONE DEPLETION TO EMISSIONS
HEALTH EFFECT
Non-Melanoma Skin Tumors
Additional Basal Cases
Additional Squamous Cases
Additional Deaths
Melanoma Skin Tumors
Additional Cases
Additional Deaths
Senile Cataract
Additional Cases
Low
1,938,500
(-61)
1,157,300
(-64)
49,400
(-63)
43,200
(-61)
12,700
(-61)
1,449,600
(-58)
Central High
5,012,900 11,343,900
(126)
3,185,800 8,099,600
(154)
135,000 338,900
(151)
109,800 237,700
(116)
32,200 69,700
(116)
3,463,400 6,984,800
(102)
                         * * *  DRAFT FINAL  * *  *

-------
                                   18-50
                              EXHIBIT 18-29

                            Human Health  Effects:
      Sensitivity to Relationship Between Ozone Depletion and Emissions
Additional  Cumulative Cases and Deaths  Over Lifetimes of People Born  2030-2074
     (Figures in Parentheses are Percentage Changes from Central Case)
                                  SENSITIVITY OF OZONE DEPLETION TO EMISSIONS
         HEALTH EFFECT                 Low         Central         High
  Non-Melanoma Skin Tumors

    Additional Basal Cases           6,983,100      17,630,500    37,901,400
                                      (-60)                       (115)

    Additional Squamous Cases        4,164,000      12,122,400    31,848,500
                                      (-66)                       (163)

    Additional Deaths                 177,800        509,300     1,311,800
                                      (-65)                       (158)

  Melanoma Skin Tumors

    Additional Cases                  178,900        430,500       875,800
                                      (-58)                       (103)

    Additional Deaths                  48,000        115,100       230,500
                                      (-58)                       (100)

  Senile Cataract

    Additional Cases                3,726,100      8,295,800     14,654,400
                                      (-55)                        (77)
                          * * *  DRAFT FINAL  *  *  *

-------
                                   18-51
exposure would accompany such actions.  People alive today would experience
about 60 percent more cases of skin cancer and senile cataract compared to the
number of cases estimated to occur in the absence of actions to limit
warming.  The sensitivity of health risks to assumptions about potential
limits to global warming is greater for future generations; up to 175 percent
more cases of skin cancer are estimated for people born during 2030 through
2074.

    A comparable sensitivity in the health risk estimates was obtained by
varying the relationship between ozone depletion and emissions.  Changing the
responsiveness of ozone depletion to emissions resulted in up to 60 percent
fewer additional cases of skin cancers and senile cataract among people alive
today (low sensitivity) and up to 120 percent more additional cases for these
individuals (high sensitivity).  For reasons discussed above, the range of
health risk estimates widens for the two future population cohorts
investigated in this chapter.

    Health risk estimates also were sensitive to assumptions about future
concentrations of methane, but to a lesser extent.  The analysis suggests that
people alive today will contract about 30 percent more cases of skin cancer
and senile cataract compared to central case methane concentrations if lowest
case assumptions are used.  Note that lowering methane concentrations
increases ozone depletion and hence increases UV radiation damaging to human
health.

    Health risks were also sensitive to the uncertainty of the parameters used
to characterize dose-response.  However, low and high estimates of cases and
deaths associated with low and high dose-response coefficients (plus and minus
one standard error) fell well within the ranges of risk estimates produced by
varying assumptions about emissions and other factors.  Nonetheless, health
risks were found to vary by as much is 40 percent in either'direction if
statistical uncertainty in dose-response coefficients is taken into account.
The uncertainty in the specification of the risk models is not reflected in
these estimates.  Work is ongoing to evaluate the implications of using
alternative specifications, including alternative exposure measures.
                          * * *  DRAFT FINAL  * *

-------
                                   18-52
REFERENCES
Hansen, J.,  A. Lacis, D.  Rind,  G.  Russell,  P.  Stone,  I.  Fung,  R.  Ruedy,  and
    J. Lerner (1984), "Climate Sensitivity:   Analysis of Feedback Mechanisms,"
    in Hansen, J.E., and  T.  Takahashi,  (eds.)>  Climate Processes  and  Climate
    Sensitivity, Geophysical Monograph  29,  Maurice Ewing Volume 5,  American
    Geophysical Union, Washington,  DC.

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

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

Ramanathan,  V., R.J. Cicerone,  H.B.  Singh,  and J.T. Kiehl (1985), "Trace Gases
    and Their Potential Role in Climate Change," National Center  for
    Atmospheric Research, NCAR/0304/84-9, Boulder, Colorado.

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

-------