Paper
INCONSTANT SUN: HOW SOLAR EVOLUTION HAS AFFECTED
COSMIC AND ULTRAVIOLET RADIATION EXPOSURE OVER
THE HISTORY OF LIFE ON EARTH
P. Andrew Karam*
sources. The effects of these parameters will be discussed
quantitatively, semi-quantitatively, or qualitatively as
appropriate.
Stars form by the gravitational collapse of clouds of
gas and dust. This collapse causes steadily increasing
pressures and temperatures at the cloud’s center, leading
to hydrogen fusion. Compression continues until the
radiation pressure emitted by the core balances the
pressure of gravitational collapse. During the early (Ttauri) stage of star formation, large quantities of gas and
dust are blown away by radiation pressure, producing
strong stellar winds. This stage is relatively short-lived
and would have been over by the time life first evolved
on Earth.
Young stars are thought to have strong magnetic
fields (Feigelson et al. 1991) that weaken over time.
These stronger magnetic fields seem linked to higher
levels of stellar activity in general (Walter and Barry
1991), including more frequent and stronger stellar
flares. Over time, as the star’s rotational velocity slows,
the internal dynamo appears to slow as well, resulting in
a generally lower level of stellar activity.
Stellar luminosity changes, too, over time, with stars
becoming more luminous. As stars fuse hydrogen into
helium, they build up a core of helium “ash.” As this
happens, the hydrogen-burning zone moves outward and
the stellar surface temperature and luminosity increase.
These changes occur steadily and predictably over a
star’s life.
At the present time in the sun’s evolution, it is about
halfway through its expected lifespan. Solar luminosity is
nearly 50% greater than when the Earth first formed, and
the sun is likely producing the majority of its energy from
hydrogen fusion with a much smaller amount coming
from other reactions.
Abstract—Four billion years ago, sea-level UV exposure was
more than 400 times as intense as today, the dose from solar
cosmic rays was five times present levels, and galactic cosmic
rays accounted for only about 10% their current contribution
to sea-level radiation doses. Exposure to cosmic radiation
accounts for about 10% of natural background radiation
exposure today and includes dose from galactic cosmic rays
and solar charged particles. There is little exposure to ionizing
wavelengths of UV due to absorption by ozone. The sun has
evolved significantly over its life; in the past there were higher
levels of particulate radiation and lower UV emissions from
the sun, and a stronger solar wind reduced radiation dose in
the inner solar system from galactic cosmic rays. Finally, since
the early atmosphere contained little to no oxygen, surface
levels of UV radiation were far higher in the past.
Health Phys. 84(3):322–333; 2003
Key words: radiation, cosmic; radiation, background; ultraviolet radiation; radiation, nonionizing
INTRODUCTION
COSMIC RADIATION accounts for about 10% of the natural
background radiation levels today (UNSCEAR 2000).
This cosmic radiation includes particulate radiation from
the sun as well as “galactic” cosmic rays (GCRs, particles that originate outside the solar system); the sun emits
ionizing ultraviolet (UV) radiation also. The sun has
evolved considerably over the history of the Earth, and
this evolution has had a significant effect on background
cosmic radiation levels on Earth. In this paper, the
manner in which several source terms have changed over
the past four billion years and the effect these changes
have had on the terrestrial radiation environment is
discussed. In many cases, terrestrial parameters have
changed in a manner that has had important effects on
radiation exposure rate from solar and other cosmic
* University of Rochester, Department of Environmental Medicine, 601 Elmwood Avenue, Box HPH, Rochester, NY 14642.
For correspondence or reprints contact P. Andrew Karam at the
above address, or email at [email protected].
(Manuscript received 7 May 2001; revised manuscript received
01 April 2002, accepted 15 September 2002)
Sources of solar radiation and their contributions to
cosmic radiation exposure
As the nearest star, the sun is the single greatest
source of cosmic radiation to Earth on a continuing basis,
although high-energy events such as supernovae or
0017-9078/03/0
Copyright © 2003 Health Physics Society
322
Inconstant sun ● P. A. KARAM
gamma ray bursts occasionally provide higher levels of
radiation for short amounts of time at infrequent intervals
(Karam 2002a, b). The sun influences the terrestrial
radiation environment in several ways:
323
this phase was well over, and the sun had settled down
into a quieter phase of its life. In its youth, the sun rotated
more rapidly than it does today. Because the solar
dynamo depends on the sun’s rotation to generate the
solar magnetic field, the early sun had a much stronger
magnetic field than is present today. Therefore, it is
thought that the early sun was more active, had more
numerous solar flares, had a stronger solar wind, and
accelerated charged particles to higher velocities than is
the case today (Bahcall et al. 2001).
Solar flares, coronal mass ejections, and similar
events briefly increase the flux of particles in the solar
wind. Zombeck (1990) reports that radiation dose in
space is about 2 Gy wk⫺1 (104 Gy y⫺1) from all
sources of solar charged particles. This includes 3
solar flares of average size (on the order of 1023—1024
J) (Tayler 1997) per week. During the 11-y solar cycle,
the number of flares and the amount of energy they
release changes; these values reflect the average
through an entire solar cycle. Cosmic radiation is
attenuated by the Earth’s magnetic field and atmosphere to a global average of 0.22 mGy y⫺1 at sea
level, which is mainly from photon and muon radiation
(UNSCEAR 2000) because neutrons contribute only
about 2 ␮Gy y⫺1 at sea level. Most solar protons lack
the energy to produce muons, so the majority of
sea-level radiation dose today is from galactic cosmic
ray-produced muons (UNSCEAR 2000; NCRP 1987;
Eisenbud and Gesell 1997). This suggests particle
radiation from the sun is not generally a significant
source of sea-level radiation, although energetic solar
flares are capable of producing high radiation levels at
the Earth’s surface (O’Brien 1978).
Because the early sun likely had a much stronger
magnetic field and a more energetic solar wind (Saar
1. Charged and uncharged particles emitted during normal solar activity (the solar wind);
2. Ultraviolet radiation emissions;
3. Modulations in galactic cosmic ray flux due to
changes in the solar magnetic field; and
4. Induced radioactivity via particle interactions in the
upper atmosphere.
Charged particles from the solar wind. The sun,
primarily in the form of the solar wind, continually emits
charged and uncharged particles. Due to extremely high
temperatures, the solar wind is a plasma, consisting of
electrons, protons, helium nuclei, and some neutrons.
The flux of the particles is highly variable, primarily
depending on solar activity and solar magnetic field
strength. The sun loses approximately 10⫺14 solar masses
(about 1019 g) of its mass per year via the solar wind. If
we assume the solar wind is composed entirely of
hydrogen, this amounts to an average flux of about 108
protons cm⫺2 s⫺1 at a distance of one astronomical unit
(1.5 ⫻ 1013 cm) from the sun. The actual flux varies from
this value because the solar wind is not emitted isotropically, but values measured by spacecraft at a distance
of about 1 AU (and outside of the Earth’s magnetopause)
are within an order of magnitude. The properties of the
solar wind at a distance of 1 astronomical unit (the
average distance between Earth and sun) are summarized
in Table 1a and b.
As noted in the introduction, the very early sun
likely had a strong solar wind, as has been observed in
T-tauri stars. However, by the time life evolved on Earth,
Table 1a and b. Solar wind properties. The values given are for the sun during solar minimum, solar maximum, and
“average” conditions. The flux of higher-energy protons is summarized in Table 1a and galactic cosmic rays are
summarized in Table 1b. Note that galactic cosmic ray flux doubles during solar minimum because of the weaker solar
magnetic field. Most galactic cosmic rays range in energy from 40 MeV to 1013 MeV with the majority in the range of
103⫺ 107 MeV. Radiation dose in space from solar charged particles is about 104 Gy y⫺1 at a depth of 5 gm cm⫺2,
equivalent to a depth of 5 cm in water or tissue (Zombeck 1990). At the Earth’s surface, at a depth of about 1,000 g
cm⫺2, the dose rate is 0.055 mGy y⫺1.
Flux (cm⫺2 s⫺1)
a
Velocity (km s⫺1)
Particle
Ave
Max
Min
Ave
Max
Min
Years of
data used
Proton
Alpha
4.0 ⫻ 108
1.2 ⫻ 107
5.2 ⫻ 108
1.8 ⫻ 107
1.96 ⫻ 108
9.8 ⫻ 105
400
—a
527
—a
362
—a
38
38
Data not available.
Solar minimum
Solar maximum
GCR
Proton flux
(cm⫺2 s⫺1)
Annual GCR
proton fluence
(cm⫺2)
4
2
1.3 ⫻ 108
7.0 ⫻ 107
324
Health Physics
1991; Bahcall et al. 2001; Sackman et al. 1993), galactic
cosmic rays were excluded from the inner solar system
and were less important in the distant past as a source of
radiation exposure. In addition, stronger and more frequent solar flares contributed much more significantly to
sea-level radiation exposure than at present. These effects and the resulting radiation doses are described in
more detail in the following section.
Sun-like stars have been observed to have episodic
“superflares” that emit from 1026 to 1029 J of energy in a
short time (Schaefer et al. 2000). These superflares
appear to take place every few million years, although
there is some doubt as to their cause and whether or not
the sun is a likely superflare candidate. Interestingly, if
we assume that the mean interval between solar flares of
various energies is proportional to ⑀⫺0.53 (Wdowczyk and
Wolfendale 1977), where ⑀ is the sea level energy flux in
erg cm⫺2, then the mean interval between flares with a
possible biological impact is on the order of a few million
years (O’Brien 1978), in reasonable agreement with the
mean interval between superflares observed in sun-like
stars. Sea-level radiation dose from solar flares of various
sizes and the mean interval between such flares are
summarized in Table 2.
Most charged particles are trapped by the Earth’s
magnetic field before reaching the atmosphere and are
diverted into one of two radiation belts. Electrons and
protons circulate within these belts spiraling around
magnetic field lines. As the magnetic field lines approach
the magnetic poles, the distance between them decreases
and the trapped particles spiral more tightly and more
rapidly. They eventually reach a “mirror point” at which
the combined electromagnetic forces cause them to stop
and reverse direction, traveling to the opposite magnetic
pole where they again reverse direction at the opposite
Table 2. Sea-level radiation dose and mean interval for solar flares
of various sizes (O’Brien 1978). Note that the mean interval
between solar flares producing a sea-level radiation dose of 1 Gy
is only about 40,000 y, much more frequent than the interval
between supernovae and gamma ray bursts contributing a similar dose from gamma rays (about 5–10 million years; Karam
2002a, b) and on the same time scale as large bursts of UV from
supernovae (a few tens of thousands of yearsa).
a
Sea-level radiation dose (Gy)
Mean interval between events (ys)
0.001
0.1
1
10
100
500
1000
10,000
100
6,000
4 ⫻ 104
3 ⫻ 105
2 ⫻ 106
8 ⫻ 106
107
108
Personal communication; John Scalo, Dept. of Astronomy, RLM 17.220,
University of Texas at Austin, Austin, TX 78712; April 2001.
March 2003, Volume 84, Number 3
mirror point. This process traps charged particles and
helps protect the Earth from the potential biologically
adverse effects of solar and galactic cosmic radiation.
Another source of high-energy protons and electrons
are albedo neutrons; neutrons produced by charged
particle interactions in the atmosphere that decay into a
proton and electron. Albedo neutrons are a major source
of high-energy protons (⬎50 MeV) but are considered a
minor source of other protons and electrons (Walt 1994).
Several satellites have conducted charged particle
observations for more than a quarter century. These
include satellites both beyond and within the boundaries
of the Earth’s magnetic field (the magnetopause). Data
from these satellites covers more than an entire solar
cycle and are summarized in Table 3. These data were
developed as described below.
Interplanetary magnetic field and charged particle
data were obtained for analysis from NASA satellite
data, both on CD-ROM† and from the NASA Space
Science Missions web page (http://spacescience.
nasa.gov/missions/index.htm). These data were collected
from satellites (the IMP, SOHO, and WIND probes)
operating outside the influence of Earth’s magnetic field
at a distance of about 1 astronomical unit (AU) from the
sun. The following parameters were used:
1. Solar wind flow speed;
2. Proton flux for all energy ranges greater than 1 MeV;
3. Alpha particle flux for all energy ranges greater than
1 MeV; and
4. Solar magnetic field strength (Bx, By, Bz).
Parameters were reported as daily averages. The
solar magnetic field strength was reported in units of nT
(nano Tesla) along three axes. An overall magnetic field
strength was calculated by determining the length of the
resultant vector. Charged particle data were reported in
units of counts cm⫺2 s⫺1. Each data set from each satellite
was analyzed to determine the mean, median, and standard deviation for each parameter. These data are summarized in Table 3. These were compared between
satellites within and outside the Earth’s magnetopause.
Data were averaged over 1-y intervals and compared.
Averaging over shorter spans of time (e.g., monthly,
weekly) was considered and rejected because of varying
instrument availability and missing data points on shorter
intervals.
Following this averaging, ion density was multiplied
by the flow speed to determine the flux of solar wind ions
in units of ions cm⫺2 s⫺1. Ion density was the sum of all
measured ions in units of ions cm⫺3. The ion flux was
†
“A Compilation of 1963–1991 Solar Wind Field and Plasma
Data, Near Earth and Deep Space,” prepared at the National Space
Science Data Center.
Inconstant sun ● P. A. KARAM
325
Table 3. Average annual solar wind parameters measured from 1963–2000. Each value represents the annual average
for the particular parameter. Missing data, due to instrument or communications malfunctions, were excluded from the
averages and those years had fewer data points included. Data in columns 2, 3, and 4 were obtained from satellite data
archives.
Year
ⱍBⱍ net
(nT)
flow speed
(km s⫺1)
ion density
(cm⫺3)
ion flux
(⫻ 108 cm⫺2 s⫺1)
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
median
5.50
5.00
5.20
6.55
6.40
6.20
5.75
6.40
5.80
6.40
6.35
6.55
5.50
5.40
6.00
7.15
7.15
6.95
7.80
8.35
7.80
7.70
5.90
5.50
6.20
7.25
7.90
7.35
8.70
7.95
6.35
6.10
5.60
5.20
5.55
6.80
6.65
7.20
6.400
360
441
415
431
422
459
425
416
438
405
476
534
485
448
419
432
432
378
436
465
480
482
464
447
430
425
440
426
469
432
454
506
425
425
377
410
435
466
433.750
11.9
10.0
7.30
5.70
5.60
4.50
4.60
5.10
7.50
8.30
8.70
7.40
9.10
9.10
10.3
8.95
8.95
8.75
9.30
11.2
9.10
8.95
9.70
10.2
10.3
8.40
9.00
8.25
10.4
11.0
9.85
8.10
9.85
9.10
9.85
8.00
7.20
7.60
8.95
4.28
4.41
3.03
2.46
2.36
2.07
1.96
2.12
3.29
3.36
4.14
3.95
4.41
4.08
4.32
3.87
3.87
3.31
4.05
5.21
4.37
4.31
4.50
4.56
4.43
3.57
3.96
3.51
4.88
4.75
4.47
4.10
4.19
3.87
3.71
3.28
3.13
3.56
3.958
plotted against the magnitude of the net solar magnetic
field vector (兩B兩net), as shown in Fig. 1. It indicates that,
at present, the flux of solar charged particles is weakly
related to the strength of the solar magnetic field, and
there is a great deal of scatter in the data.
It is likely that, over time, the solar magnetic field
and solar wind have weakened monotonically over the
life of the sun (Bahcall et al. 2001). This suggests that the
flux of charged particles reaching the Earth has dropped
steadily with time and the energy of these particles has
likely decreased as well. In addition, the terrestrial
magnetic field parameters have changed with time,
although a discussion of the impact of the changing
terrestrial magnetic field on radiation dose from charged
particles is beyond the scope of this dissertation. Further
calculations in this paper assume that the terrestrial
magnetic field strength in the past was similar to that
observed today and that changes in terrestrial magnetic
field strength have not significantly affected radiation
dose rates at sea-level.
Calculations of radiation absorbed dose from solar
charged particles and GCRs
During the course of a solar cycle, solar magnetic
field strength and charged particle flux change as noted
in Tables 1 and 3, with corresponding changes in
sea-level cosmic radiation levels from solar and galactic
cosmic rays. Fluxes are also affected by the terrestrial
magnetic field strength, but this variation is only about
14% at sea level (Eisenbud and Gesell 1997).
Although there is a weak correlation between solar
magnetic field strength and charged particle flux, Fig. 1
shows a great deal of data scatter, and it is equally
possible that this correlation is simply an artifact of a
326
Health Physics
March 2003, Volume 84, Number 3
cycle,‡ solar magnetic field strength at a distance of 1 AU
changes by a factor of about 2 from 4 to nearly 8 nT
according to spacecraft data, while the flux of GCRs
varies by a factor of 30% between solar maximum (when
GCR flux is lowest) and solar minimum (Gombosi 1998;
Fisk 1983). These relative values are approximate and
are subject to a great deal of scatter, so, as a first
approximation, it seems acceptable to use eqn (1) to
describe the relationship between GCR flux and solar
magnetic field strength as follows:
␾ 共 GCR 兲共 B 兲 ⫽ 4 cm⫺2s⫺1
Fig. 1. Solar charged particle flux vs. solar magnetic field strength
at 1 AU. This research assumes a constant value of 4.0 ⫻ 108
particles cm⫺2 s⫺1 as the charged particle flux, the average value
noted in Table 3 at a magnetic field strength (兩B兩) of 6.5 nT.
limited data set. Accordingly, for the purposes of this
work, and until better data become available, solar
charged particle flux in the past is assumed to be
proportional to 兩Bnet兩, using the average values of both of
these parameters today as a baseline for these calculations. Future work will examine this relationship more
critically by using observations of other nearby sun-like
stars. From a practical standpoint, radiation dose from
solar charged particles is a relatively minor part of
overall background radiation exposure so this simplifying assumption will not have a profound impact on the
calculated radiation dose at any time in the past.
Models for changes in solar magnetic field strength
through time suggest that charged particle flux in space
has changed over the past four Gyr as shown in Table 4.
This table also shows sea-level radiation dose from solar
charged particles and GCRs. During the course of a solar
5.25nT
.
ⱍBⱍ
In this equation, 4 is the average GCR flux in space at
solar minimum (NCRP 1989), 兩B兩 is the magnitude of the
solar magnetic field vector, and 5.25 nT is the average
solar magnetic field strength at solar minimum (from the
data sets referenced above). Charged particle flux (and,
hence dose) from solar charged particles is assumed to
vary directly with solar magnetic field strength in the past
using the average value today as a baseline.
Absorbed radiation dose from GCRs is presently
about 0.165 mGy y⫺1 and we receive about 0.055 mGy
y⫺1 from solar cosmic rays (UNSCEAR 2000). If we
assume that the energy spectra of these sources of
radiation do not change with solar activity, then the
sea-level radiation dose from each is proportional to their
flux at Earth’s orbit at any time, calculated as shown
above. These results, shown in Fig. 2 and Table 4,
suggest that sea-level absorbed radiation dose from
‡
During a solar cycle, a variety of solar parameters will vary in
a sinusoidal manner. Although any given parameter may have a value
at “solar maximum” that is less than its value at an earlier (or later)
“solar minimum,” the value will be higher than in the solar minimum
immediately preceding or following the peak. In other words, these
terms refer to local, and not absolute maxima and minima because of
the large variability in solar activity from one sunspot cycle to the next.
Table 4. Changes in average solar magnetic field strength, solar charged particle flux, galactic cosmic ray flux, and
sea-level cosmic radiation dose through time. The flux of solar and galactic cosmic rays were calculated as described
in the text. It was assumed that the absorbed radiation dose at sea level from each source was directly proportional to
the appropriate cosmic ray flux outside the Earth’s magnetosphere. Sea-level radiation dose from solar charged particles
is significantly less than that from GCRs because the solar charged particles deposit virtually all of their energy in the
atmosphere while GCRs can produce muons and cosmic ray showers, which penetrate to sea level.
Time
(Ga)
0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
a
ⱍBⱍ
rel.
1.0
1.25
2.0
2.2
4.0
6.3
10.0
15.8
20.0
ⱍBⱍ
(nT)
␾(solar)
(⫻ 108 cm⫺2 s⫺1)
␾(GCR)
(cm⫺2 s⫺1)
Dose (solar)
(mGy y⫺1)
Dose (GCR)
(mGy y⫺1)
Dose (total)
(mGy y⫺1)
6.5
8.13
13.0
14.3
26.0
41.0
65.0
103
130
4.0
5.00
8.00
8.80
1.60
2.52
4.00
6.32
8.00
3.23
2.58
1.62
1.47
0.81
0.51
0.32
0.20
0.16
0.055a
0.07
0.11
0.12
0.22
0.35
0.55
0.87
1.10
0.165a
0.13
0.08
0.08
0.04
0.03
0.02
0.01
0.01
0.22a
0.20
0.19
0.20
0.26
0.37
0.57
0.88
1.11
From UNSCEAR, 2000, Annex B, Figure I.
(1)
Inconstant sun ● P. A. KARAM
Fig. 2. Changes in solar magnetic field strength and cosmic
radiation dose through time. Flux and dose from solar charged
particles likely changed with changing solar activity and magnetic
field strength through time, which has dropped as the sun’s
rotation slowed. Changes in GCR flux stem from the fact that the
stronger solar magnetic field present in the early sun likely
excluded GCRs from the inner solar system. As described in the
text, the flux of solar charged particles at the present is assumed to
be the average of the values shown in Fig. 1, and this flux is
assumed to be directly proportional to the solar magnetic field at
any time in the past.
GCRs has increased with time while that from solar
charged particles has likely dropped considerably due to
decreasing solar activity.
Ultraviolet radiation emissions
Over the last 4.5 billion years the sun has become
steadily brighter, slightly larger in radius, and moderately
hotter. The increase in brightness leads to an overall
increase in the flux of all wavelengths received at the
Earth, while the increase in the effective temperature of
the solar atmosphere shifts the mean energy of the
spectrum to bluer wavelengths, increasing the fraction of
the total energy emitted as UV photons. Both of these
effects combined have resulted in a steady, monotonic
increase in the amount of UV radiation received at a
distance of 1 AU from the sun (i.e., above the Earth’s
atmosphere) over its history.
The sun emits a primarily thermal spectrum characterized by a general continuum modified by a combination of
atomic absorption lines and continuous absorption in the
solar atmosphere. This means that one cannot simply model
the emergent solar spectrum as a simple blackbody, particularly at UV wavelengths. While for stars like the sun the
first-order spectral shape is approximately that of a Planck
blackbody spectrum, at UV wavelengths the spectrum is
significantly modified by absorption lines from neutral and
ionic metal atoms in the solar atmosphere (primarily neutral
and singly-ionized iron). Further, the amount of UV opacity
327
in these lines varies as a function of the atmospheric
temperature and pressure in a complex fashion due to
changes in atomic excitation and ionization. The overall
result is that in order to model the evolution of the solar
spectrum as a function of time we need to use detailed
model atmospheres calculations that correctly account for
all of these effects.
Of primary interest to this research is the ionizing
UV flux from the sun in the 176 –300 nm band. The UV
spectrum we measure today (reported by the World
Meteorological Organization 1982) with sounding rockets represents the UV spectrum that has already escaped
the solar atmosphere to reach the Earth. Using the results
of solar evolution (Bahcall et al. 2001) and stellar
atmospheres models (Hauschildt 1999), we can compute
the relative amount of UV radiation emitted by the sun at
various times in the past, and this can be used to scale the
present day spectrum values to any given time in the past.
This will give a relatively robust estimate of the baseline
UV flux history of the quiescent sun, that is, in the
absence of episodic events (e.g., sunspots or solar flares).
To determine UV emissions from the solar photosphere we must account for changes in temperature,
luminosity, and radius that have occurred through time.
There is a robust model describing how solar radius and
luminosity have changed (Bahcall et al. 2001) and this
can be used to calculate an effective temperature of the
solar photosphere using the following eqn:
T eff ⫽
LJ
.
4 ␲␴ r 2
(2)
In this equation, Teff is an effective surface temperature,
LJ is the solar luminosity at given times in the past, and
r is the solar radius at the same time in the past. The past
solar luminosity and radius were taken from Bahcall et
al. (2001).
In order to compute the flux spectrum of a star, three
input parameters are needed: the effective temperature of
the photosphere (Teff), the “surface” gravity (GM/r2), and
the chemical composition of the atmosphere. The spectrum of the sun at different epochs was modeled using
synthetic stellar spectra computed by Hauschildt (1999).
Four stellar spectra were chosen from his spectral library
with values of Teff that bracket the range of solar Teff
computed in Table 5: Teff ⫽ (5,400, 5,600, 5,800, 6,000)
K, and which have the same surface gravity and chemical
composition as the sun. These flux spectra were numerically integrated in wavelength between 176 nm and 300
nm to compute the fraction of the total luminosity
emitted in this band, UV/L. The UV fraction in this
temperature range was found to be a slow, monotonic
function of Teff that may be evaluated in the 5,400 ⬍ Teff
328
Health Physics
March 2003, Volume 84, Number 3
Table 5. Changes in solar temperature, radius, luminosity, and integrated UV flux with time. The values in columns 2,
3, and 7 are expressed as ratios of luminosity, radius, and integrated UV emissions at times in the past compared to
present-day values. In all cases, present day values (indicated with a subscript J) are in the denominator. Values for L,
R, and Teff are taken or derived from Bahcall et al. (2001), and the total UV flux is calculated as described in the text.
Time (Ga)
L/LJ
R/R.
Teff (K)
UV/L
␾UV
(⫻ 1015 ␥ cm⫺2 s⫺1)
␾UV/␾UVJ
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
1.000
0.960
0.922
0.887
0.854
0.823
0.795
0.768
0.741
1.000
0.983
0.967
0.952
0.939
0.926
0.914
0.902
0.891
5779
5770
5758
5747
5734
5721
5708
5694
5680
0.0169
0.0166
0.0163
0.0159
0.0156
0.0153
0.0149
0.0146
0.0142
1.95
1.92
1.89
1.85
1.82
1.78
1.74
1.71
1.67
1.000
0.984
0.969
0.949
0.930
0.912
0.892
0.875
0.854
⬍ 6,000 range using a simple quadratic interpolation
formula (personal communication; Richard Pogge, Ohio
State University, Dept. of Astronomy, 140 W. 18th Ave.,
Columbus, OH 43210-1173; September 2001):
UV
⫽0.334⫺(1.38⫻10⫺4⫻T eff)⫹(1.44⫻10⫺8⫻T 2eff) .
L
(3)
The ratio of UV/L for different times is given in Table 5,
column 5, showing how the fraction of the sun’s luminosity that is emitted in the 176 –300 nm UV band has
evolved with time. The UV flux at the top of the Earth’s
atmosphere at each time is then given by:
L
UV
␾UV⫽␾0⫻ ⫻
,
LJ
L
(4)
where ␾0 is the observed present-day UV flux, (L/LJ) is
the solar luminosity relative to the present-day (Table 5,
column 2), and (UV/L) is computed for the solar Teff at
that epoch (Table 5, columns 4 and 5) using the interpolation formula in eqn (4) above. Values of ␾UV are given
in Table 5, column 6. For comparison, column 7 in Table
5 gives the total UV flux at each epoch relative to the
present-day observed value.
As can be seen in Table 5, the solar UV flux has
steadily increased by about 60% over the last 4 billion
years. This increase has been driven primarily by a
roughly 35% increase in the overall solar luminosity,
further enhanced by the roughly 100 K increase in the
effective temperature of the solar photosphere, which has
increased the fraction of UV radiation that is emitted in
the 176 –300 nm band by the quiescent sun.
Increases in emissions of ionizing UVB and UVC,
however, address only the source term. Of equal importance is to determine the amount of radiation reaching the
Earth’s surface. The primary UV shield is ozone, and
before the formation of Earth’s ozone layer, UVB and
UVC reached the Earth’s surface almost unattenuated.
Thus, UV flux at the Earth’s surface was significantly
higher in the distant past than it is today.
Cox (2000) reported on the average column density
of ozone molecules in the atmosphere and WMO (1982)
provides the absorption cross section of ozone for each
wavelength band in the UV. These data can be combined
to arrive at the fraction of incident UV radiation that
reaches sea level, assuming that scattering and absorption
by other atmospheric constituents is negligible. This is
calculated by using the relationship
␶ 共 ␭ i兲 ⫽ n 共 O 3兲 ␴ 共 ␭ i兲 ,
(5)
where ␶␭ is the optical depth for a given wavelength, nO3
is the column density of ozone molecules in units of
molecules cm⫺2, and ␴␭i is the absorption cross-section in
units of cm2. The optical depth is a measure of the
absorption when light passes through a medium, and the
incident flux is reduced by a factor of e at an optical
depth of 1. Accordingly, the transmission of UV light to
the surface is calculated as follows:
␾ SL 共 ␭ i兲 ⫽ ␾ 0 共 ␭ i兲 e⫺␶(␭ )
i
(6)
in which ␾SL and ␾0 are the UV photon flux at sea level
and above the atmosphere, respectively for a particular
wavelength band (␭i) and the subscripts “SL” and “0”
refer to the flux at sea level and above the atmosphere,
respectively. The resulting sea-level photon flux for each
wavelength is then in units of photons cm⫺2 s⫺1.
The photon flux is converted to an energy flux by
first calculating the energy of individual photons using
the relationship
F SL 共 ␭ i兲 ⫽
␾ SL 共 ␭ i兲 hc
,
␭i
(7)
where F(SL) is the photon energy flux in ergs cm2 s⫺1 in a
given wavelength band i, h is Planck’s constant
Inconstant sun ● P. A. KARAM
⫺20
(6.62606876 ⫻ 10
J s), c is the speed of light
(2.997792458 ⫻ 108 m s⫺1), and ␭ is the wavelength in
cm. Because the UV spectrum is reported in wavelength
bands, the central wavelength of each 5 nm band was
used for these calculations. Summing the energies of all
47 wavelength bands gives an integrated UV energy flux
at sea level:
F UV
SL ⫽
冘F
SL 共 ␭ i 兲
.
(8)
i
In this equation, F is the energy flux in units of J cm⫺2
s⫺1 and the final answer reflects the integrated energy
flux at sea level across all wavelength bands for which
absorption and flux were calculated. At the present time,
the solar UV flux is about 1.4 ⫻ 10⫺3 J cm⫺2 s⫺1 in space
and is attenuated to about 29 ⫻ 10⫺6 J cm⫺2 s⫺1 at sea
level by its passage through the ozone layer.
In the past, the Earth had noticeably less oxygen in the
atmosphere and correspondingly less ozone. This allowed
higher levels of UV to penetrate to sea level. To calculate
sea-level UV flux in the past, ozone concentrations were
assumed to be directly proportional to atmospheric oxygen
levels at that time (personal communication; Charles Cockell, British Antarctic Survey, High Cross, Madingley Road,
Cambridge CB3 0ET, United Kingdom; March 2001).
Today’s atmosphere contains about 21% oxygen and has a
column density of about 8 ⫻ 1018 molecules of ozone cm⫺2.
However, a billion years ago, for example, the atmosphere
had only about 15% of today’s oxygen concentrations (or
about 3% oxygen), giving it an ozone column density of
about 1.2 ⫻ 1018 molecules cm⫺2. Summing up the photon
and energy flux at sea level at that time we find that it was
about 2.2 ⫻ 10⫺4 J cm⫺2 s⫺1 and the incident photons were
primarily of wavelengths between 280 nm and 300 nm.
Another factor that must be taken into account is the
evolving solar spectrum in time. These changes, calculated as described above, were used to determine the flux
of UV photons in each wavelength band at that time in
329
the past. The photon fluxes were then used as the basis
for calculating sea level UV photon and energy flux at
times in the past. A short summary of these results and a
comparison with current UV flux is provided in Table 6.
The calculated sea-level UV spectra at intervals of one
billion years and today’s solar UV spectrum are shown in
Fig. 3a and b, and the amount of UV reaching sea level
at times in the past is shown in Fig. 4. Note that, in
addition to increases in ozone concentrations (and the
resulting drop in UV flux), there is a small net increase in
UV flux through time due to rising solar surface temperatures resulting from solar evolution. Note, too, that these
calculations are based on current models of solar and
atmospheric evolution. There is little doubt that solar
activity has dropped with time, surface temperature (and
UV flux) has increased, and that atmospheric oxygen and
ozone levels have both increased. However, the precise
manner in which these changes have taken place, and the
magnitude of these changes at specific intervals in the
past is still subject to much debate.
These calculations show that the early Earth was
subjected to significantly higher levels of ionizing UV
radiation prior to the formation of the ozone layer. This
conclusion is hardly a surprise and has, in fact, been
reached by many in the past (see, for example, Cockell
1998; Cockell 2000). They also show that UV flux on the
earliest Earth was approximately 400 times as intense as
on today’s Earth. This value compares favorably with
Cockell’s (1998) ratio of 470 times the relative DNA
damage rates to early life from UV radiation. Cockell
also notes that other factors, including other atmospheric
gases, particulates, or sulfur compounds may have
helped screen early life from UV irradiation, but their
presence cannot be known with any degree of certainty.
The final item to consider is how these elevated
levels of UV radiation may have affected DNA mutation
rates at times in the past. DNA strongly absorbs almost
Table 6. Approximate changes in sea-level UVB and UVC flux over time. Ozone concentrations used assume that ozone
levels are directly proportional to atmospheric oxygen levels. The absorption coefficients from World Meteorological
Organization (1982) were used to calculate sea-level UV flux at times in the past, which was normalized to current levels
in the final column.
Time (Ga)
0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
Ozone concentration
(molecules cm⫺2)
18
8.1 ⫻ 10
1.2 ⫻ 1018
1.2 ⫻ 1018
1.2 ⫻ 1018
1.2 ⫻ 1018
8.1 ⫻ 1016
8.1 ⫻ 1016
8.1 ⫻ 1016
8.1 ⫻ 1015
UV flux
(ergs cm⫺2 s⫺1)
Relative UV flux
29.2
2262
2248
2232
2214
10,258
10,158
13,040
12,892
1.0
77.5
77.0
76.5
75.9
352
348
447
442
330
Health Physics
Fig. 3. (a and b): Solar UV spectrum and sea-level UV spectra over
time. Note that, until about 2 Ga, the solar and sea-level UV
spectra were almost identical because virtually no ozone existed to
absorb the UV. In addition, the sea-level spectra at 1 Ga and 2 Ga
are also nearly identical. Sea-level UV flux at all wavelengths was
nearly identical to the calculated solar spectrum at 4 Ga because of
the total absence of ozone at that time. Note, too, that even the very
small levels of ozone present 3 Ga were sufficient to cause a slight
change in the sea-level UV spectrum.
Fig. 4. Sea level UV flux vs. time. This takes into account the
steady increase in solar surface temperature (and resulting increased UV emissions) with time. The sharp drops are due to the
formation of the ozone layer. Oxygenic photosynthesis evolved
about 3 Ga, large quantities of free oxygen were first seen about
2.5 Ga, and the atmosphere reached its current composition about
500 million years ago.
exactly the same wavelengths absorbed by ozone. Therefore, in the pre-ozone world, the most damaging wavelengths reached the Earth’s surface largely unattenuated.
It is instructive to note that UV repair mechanisms, some
March 2003, Volume 84, Number 3
of which are activated by exposure to visible light
(photoreactivation mechanisms), are common in most
organisms. This is thought by some to be a legacy of
early era when exposure to visible light almost certainly
carried with it exposure to genetic damage from UV
radiation (Cockell 2001).
The fact that DNA absorbs UV so strongly suggests
that the energy deposition required to produce a given
amount of damage is less than for other forms of ionizing
radiation. Using canonical values for radiation dose (i.e.,
1 Gy ⫽ 1 J kg⫺1 ⫽ 104 ergs g⫺1) and mutation doubling
dose (1 Gy), one finds that an energy deposition of 1 J of
energy per gram of tissue will yield a doubling dose. If
we assume that tissue has a density of 1 g cm⫺3, and note
that typical thicknesses of an algal mat or biofilm are on
the order of 1 cm, then a mutation doubling dose would
be 10⫺3 J cm⫺2. However, some have reported that for
UVB and UVC 10⫺8 J cm⫺2 will produce one thymine
dimer in E. coli (Hamkalo et al. 1972) and, at a
wavelength of 254 nm, a fluence of about 2x1011 J cm⫺2
in this wavelength is the calculated mutation doubling
dose (Scalo and Wheeler 2002).
The huge discrepancy in mutation doubling doses
from UV vs. gamma and other ionizing radiations is
because of the different interactions: UV is directly
absorbed by the DNA while other ionizing radiations
tend to interact with DNA indirectly. Other lesions will
undoubtedly occur in addition to thymine dimer formation, both with and without UV irradiation, so this value
is an estimate only. It is also important to note that this
value does not apply to macroscopic organisms because
of the presence of epidermal coverings (dead skin cells,
hair, scales, etc.). However, life has consisted primarily
of single-celled organisms for most of its history and
such coverings appeared after the formation of an ozone
layer.
The DNA-weighted UV irradiance at sea level today
is about 10⫺7 J cm⫺2 s⫺1, so the earliest life (if exposed at
the Earth’s surface) received UV irradiation of over 4 ⫻
10⫺5 J cm⫺2 s⫺1. This is equivalent to an E. coli mutation
doubling dose every quarter second, which is much too
high for a living organism to survive. Since life exists
today, it is evident that early life found a suitably
shielded habitat, either in the deep sea, beneath a surface
layer of sediments, or some other UV-opaque material.
However, it is likely that many dominant life forms on
the early Earth were photosynthetic, especially after
about 3 billion years ago. Accordingly, it is likely that
these organisms were still exposed to levels of UV
radiation significantly higher than those experienced by
modern organisms (Kasting et al. 1992). In fact, if we can
draw corollaries from similar organisms that are extant
Inconstant sun ● P. A. KARAM
today, it is likely that many early photosynthetic organisms lived in waters less than a few meters in depth,
placing them well within the zone into which UVB and
UVC can penetrate (Kasting and Chang 1992; Cockell
1998; Cockell 2001). Other organisms, such as stromatolites, may have evolved the ability to survive in the
photic zone by developing UV-resistant outer (and upper) layers to protect more sensitive cells beneath (Rambler and Margulis 1980; Kasting and Chang 1992).
Survival strategies used by such organisms have been
previously discussed in the scientific literature (Cockell
2000) and will not be further discussed here. In addition,
Cockell (2000) points out that the DNA-weighted UV
flux in the ancient oceans, prior to the formation of an
ozone layer, would have decreased to roughly presentday levels at a depth of 30 m, while the early oceans may
have been covered in part by UV-opaque molecules that
would have served to help protect the life beneath
(Cleaves and Miller 1998; Cockell and Knowland 1999).
It is worth noting that there is a high probability that
nearby supernovae have frequently produced sufficient
UV flux to have had a major impact on terrestrial life.
Due to intense UV emissions from supernovae, it is
likely that Earth will experience a sea-level UV dose
sufficient to cause a doubling of the background mutation rate with a mean interval of about 200,000 y, and the
mean interval between events that would contribute a UV
flux that is greater than that of the sun is considerably
shorter (a few thousand years, personal communication;
John Scalo, Dept. of Astronomy, RLM 17.220, Univ. of
Texas at Austin, Austin, TX 78712; April 2001).
DISCUSSION
Radiation dose rates at the earth’s surface from
cosmic sources of ionizing radiation can be broken down
into two major categories: 1) steady and relatively
predictable changes due to solar evolution and changes in
terrestrial parameters and 2) high-intensity events that
take place randomly in space and time.
Effects of solar and terrestrial evolution
It is apparent that changes in atmospheric ozone
concentrations and the concomitant reduction in DNAdamaging UV has had a far more profound impact on
terrestrial radiation levels than has the decrease in solar
wind over time. This is because neither solar charged
particle flux nor solar UV emissions have changed
dramatically with time, but the rise in atmospheric ozone
concentrations with the advent of free oxygen in the
atmosphere has resulted in a profound reduction in the
sea-level flux of ionizing wavelengths of UV reaching
the earth’s surface. In addition, the increased solar
331
charged particle emissions early in the history of life
were partially offset by the reduction in GCRs as a result
of the stronger solar magnetic field and more energetic
solar wind at that same time. In the absence of a
terrestrial magnetic field (which may have formed
shortly after the Earth, although this is still not known
with any degree of certainty) it is likely that radiation
dose from solar charged particles would have been much
greater than at present, although the atmosphere would
have still provided shielding, just as it provides shielding
from more energetic GCRs today. Due to the very short
ranges of protons in water, any organisms more than
about 1 cm deep would have received minimal dose from
these protons. In general, it is safe to say that, over the
history of life on Earth:
1. Dose from UV radiation has dropped dramatically due
to the formation of an ozone layer;
2. Radiation dose from solar charged particles and their
interaction products has dropped steadily due to
decreasing solar activity with time;
3. Radiation dose from GCRs has increased steadily
because of decreasing solar activity; and
4. Combined radiation dose from solar radiation and
GCRs dropped steadily for about 2 billion years and
has increased since that time as a result of the
combination of these two effects.
Random events
Solar flares can generate sufficient energy to produce a measurable radiation dose at the Earth’s surface,
and occasional very large flares can even deliver a
harmful dose to the Earth’s surface (O’Brien 1978). In
fact, such events were probably far more common during
the sun’s youth than today, and it is likely that the Earth’s
surface was bombarded with damaging amounts of radiation more frequently in the past than today because of
the higher levels of activity found in young stars.
O’Brien calculated that the mean interval between solar
flares delivering sea-level radiation absorbed dose of
about 1 Gy was on the order of a few million years; it is
likely that this interval was shorter in the past. Unfortunately, it is not obvious that solar flare activity and
strength scale directly with the solar magnetic field
strength; it is possible that very large flares remained
exceedingly rare and that the sun simply had more small
flares. In any event, the mean interval between solar
flares capable of generating radiation levels of about 1
Gy or more is comparable to the mean interval between
extra-solar events (such as supernovae and gamma ray
bursts) that can generate similar radiation levels (Scalo et
al. 2003).
332
Health Physics
In addition, Scalo et al. (2003) noted that supernovae can produce very high UV fluxes at the Earth’s
surface with mean intervals on the order of tens of
thousands of years, and others have noted that supernovae and large impacts are also capable of destroying a
large fraction of the Earth’s ozone layer as well (see, for
example, Ellis et al. 1996; Cockell 2000), resulting in
long-term elevations in solar UV flux at the Earth’s
surface.
The likelihood that any individual organism will be
exposed to the very high radiation levels from a supernova, “superflare,” or other long-interval event is exceedingly low. For the vast majority of individuals, the
controlling factor in their radiation environment is solar
or terrestrial evolution. However, over the lifetime of a
species (a few million years, Raup 1991, 1994; Sepkoski
1997), larger and less frequent events are likely to occur
at least once, and these events may be sufficient to place
a selection pressure on the ability of organisms to
respond favorably to high absorbed doses of radiation.
However, in the absence of a large solar flare or nearby
supernova, we are unlikely to have the opportunity to
find out from direct observation of natural events.
On the other hand, organisms in the vicinity of the
Chernobyl nuclear power plant and the Mayak nuclear
facility were exposed to very high levels of ionizing
radiation with a wide range of doses, and the high natural
background radiation levels in Ramsar, Iran, place similar (albeit lower) stresses on organisms in that region.
These areas could serve as laboratories in which to study
such effects. It may be possible to conduct similar
long-term studies with laboratory organisms. Such studies could be instructive and may be able to provide an
idea as to the effects that brief exposure to very high
levels of radiation from cosmic events might have had on
early life. In the absence of such observations, we are
free to speculate, but these speculations cannot be easily
confirmed or refuted.
CONCLUSION
The sun has evolved over its life and, because of this
evolution, the amounts of UV and charged particles
emitted have changed continuously over the history of
the solar system. In addition, changes in solar activity
have led to a steady increase in the number of galactic
cosmic rays that reach the inner solar system, resulting in
an initial decrease and subsequent increase in sea-level
radiation dose from charged particles over time.
Although the sun has emitted steadily more UV over
its history, the formation of Earth’s ozone layer 2 billion
years ago has resulted in a net decrease in sea-level UV
since that time. As a result, the Earth’s surface has
March 2003, Volume 84, Number 3
become steadily more hospitable to life as atmospheric
oxygen levels have increased. While early life likely did
not survive unshielded if it was at the Earth’s surface or
in the top several meters of the water column, it could
have found refuge in deep waters, beneath shallow
sediments, or elsewhere. Regardless of where early life
found shelter from UV and other radiations, it is obvious
that it was sheltered because life survived that period of
Earth’s history. In addition, some forms of life, such as
stromatolites, must have had strategies of surviving high
UV exposure in even shallow-water environments.
Finally, aperiodic events such as solar “superflares,”
large impacts, and other events may have also led to
elevated radiation levels at the Earth’s surface at mean
intervals that are comparable to species’ lifetimes. This
suggests that the average species would have been
exposed to high levels of radiation at least one time, and
so the ability to survive such events may have provided
an evolutionary advantage.
Acknowledgments—I would like to offer heartfelt thanks to the many
people who helped me in various stages of this work. In particular, I would
like to acknowledge the effort that Rick Pogge, Kris Sellgren, Keran
O’Brien, Charles Cockell, Marc Pinsonneault, Juliana Sackmann, Arnold
Boothroyd, John Scalo, Craig Wheeler, and Audeen Fentiman spent in
helping me to better understand and communicate the materials presented
above. I am also grateful for the suggestions of the two reviewers and
editor who made several especially helpful suggestions. My understanding
of the subject matter and the quality of this paper has benefited immensely
from their time and patience.
REFERENCES
Bahcall JN, Pinsonneault MH, Basu S. Solar models: Current
epoch and time dependences, neutrinos, and helioseismological properties. Astrophysical J 555:990 –1012; 2001.
Cleaves HJ, Miller SL. Oceanic protection of prebiotic organic
compounds from UV radiation. Proceedings of the National
Academy of Sciences 95:7260 –3; 1998.
Cockell CS. Biological effects of high ultraviolet radiation on
early Earth—a theoretical evaluation. J Theoretical Biol
193:717–29; 1998.
Cockell CS. The ultraviolet history of the terrestrial planets—
implications for biological evolution. Planetary and Space
Science 48:203–14; 2000.
Cockell CS. The photobiological history of the Earth. In:
Cockell CS, Blaustein AR, eds. Ecosystems, evolution, and
UV radiation. New York: Springer Verlag; 2001:1–35.
Cockell CS, Knowland J. Ultraviolet radiation screening compounds. Biological Reviews 74:311– 45; 1999.
Cox AN. Allen’s astrophysical quantities. New York: Springer
Verlag AIP Press; 2000.
Eisenbud M, Gesell T. Environmental radioactivity from natural, industrial, and military sources. San Diego, CA:
Academic Press; 1997.
Ellis J, Fields BD, Schramm DN. Geologic isotope anomalies
as signatures of nearby supernovae. Astrophysical J
470:1227; 1996.
Feigelson ED, Giampapa MS, Vrba FJ. Magnetic activity in
pre-main sequence stars. In: Sonnett CP, Giampapa MS,
Inconstant sun ● P. A. KARAM
Matthews MS, eds. The Sun in time. Tucson: University of
Arizona Press; 1991: 658 – 681.
Fisk LA. Solar modulation of galactic cosmic rays. Solarterrestrial physics: Principles and theoretical foundations.
In: Proceedings of the Theory Institute. Chestnut Hill, MA:
Reidel Publishing Co.; 1983: 217–30.
Gombosi TI. Physics of the space environment. New York:
Cambridge University Press; 1998.
Hamkalo BA, Miller OL Jr., Bakken AH. Genetics and the
cancer cell: visualization of RNA synthesis. Proceedings of
the National Cancer Conference 7:45–54; 1972.
Hauschildt PH, Allard F, Baron E. The NEXTGEN model
atmosphere grid for 3000ⱕTeffⱕ10,000 K. Astrophys J
512:377–385; 1999.
Karam PA. Gamma and neutrino dose from supernovae and
gamma ray bursts. Health Phys 82:491– 499; 2002a.
Karam PA. Gamma radiation dose from supernova-produced
radioactivities. Radiat Phys Chem 64:77– 87; 2002b.
Kasting JF, Holland HH, Kump LR. Atmospheric evolution:
The rise of oxygen. In: Schopf JW, Klein C, eds. The
Proterozoic biosphere: An interdisciplinary study. New
York: Cambridge University Press; 1992: 159 –163.
Kasting JF, Chang S. Formation of the Earth and the origin of
life. In: Schopf JW, Klein C, eds. The Proterozoic biosphere. New York: Cambridge University Press; 1992:
9 –12.
NCRP. Exposure of the population in the United States and
Canada from natural background radiation. Bethesda, MD:
National Council on Radiation Protection and Measurement; Report 94; 1987.
NCRP. Guidance on radiation received in space activities.
Bethesda, MD: National Council on Radiation Protection
and Measurement; Report 98; 1989.
O’Brien K. Radiation dose from solar flares at ground level.
Health Phys 36:63– 65; 1978.
Rambler MB, Margulis L. Bacterial resistance to ultra-violet
irradiation under anaerobiosis: implications for prePhanerozoic evolution. Science 210:638 – 640; 1980.
333
Raup D. A kill curve for Phanerozoic marine species. Paleobiology 17:37– 48; 1991.
Raup DM. The role of extinction in evolution. Proceedings of
the National Academy of Sciences 91:6758 – 6763; 1994.
Saar SH. The time evolution of magnetic fields on solar-like
stars. In: Sonnett CP, Giampapa MS, Matthews MS, eds.
The Sun in time. Tucson, AZ: University of Arizona Press;
1991: 848 – 858.
Sackman IJ, Boothroyd AI, Kraemer KE. Our Sun III: Present
and future. The Astrophysical J 418:457– 68; 1993.
Scalo J, Wheeler JC. Astrophysical and biological implications
of gamma ray burst properties. Astrophys J 566:723–737;
2002.
Scalo J, Wheeler JC, Williams P. Intermittent jolts of galactic
UV radiation: Mutagenic effects. In: Celnekier LM, ed.
Frontiers of life; 2003 (in press).
Schaefer BE, King JR, Deliyannis CP. Superflares on ordinary
solar-type stars. Astrophysical J 529:1026 –1030; 2000.
Sepkoski JJ. Biodiversity: Past, present, and future. J Paleontology 71:533–9; 1997.
Tayler RJ. The Sun as a star. New York: Cambridge University
Press; 1997.
United Nations Science Committee on the Effects of Atomic
Radiation. Vienna, Austria: United Nations; Report to the
General Assembly; 2000.
Walter RM, Barry DC. Pre- and main-sequence evolution of
solar activity. In: Sonnett CP, Giampapa MS, Matthews
MS, eds. The Sun in time. Tucson, AZ: University of
Arizona Press; 1991: 633– 657.
Walt M. Introduction to geomagnetically trapped radiation.
Cambridge: Cambridge University Press; 1994:59 –91.
Wdowczyk J, Wolfendale AW. Cosmic rays and ancient
catastrophes. Nature 268:510; 1977.
World Meteorological Organization. Washington DC: National
Aeronautics and Space Administration; 1982.
Zombeck M. Handbook of space astronomy and astrophysics.
New York: Cambridge University Press; 1990.
f f