Phil. Trans. R. Soc. A (2007) 365, 1843–1866
doi:10.1098/rsta.2007.2046
Published online 18 May 2007
The stability of the stratospheric ozone
layer during the end-Permian eruption
of the Siberian Traps
B Y D AVID J. B EERLING 1, * , M ICHAEL H ARFOOT 2 , B ARRY L OMAX 1
2,3
AND J OHN A. P YLE
1
Department of Animal and Plant Sciences, University of Sheffield,
Sheffield S10 2TN, UK
2
Centre for Atmospheric Science, Department of Chemistry, and 3National
Centre of Atmospheric Science (NCAS), Atmospheric Chemistry
Modelling Support Unit, University of Cambridge, Cambridge CB2 1EW, UK
The discovery of mutated palynomorphs in end-Permian rocks led to the hypothesis that the
eruption of the Siberian Traps through older organic-rich sediments synthesized and released
massive quantities of organohalogens, which caused widespread O3 depletion and allowed
increased terrestrial incidence of harmful ultraviolet-B radiation (UV-B, 280–315 nm;
Visscher et al. 2004 Proc. Natl Acad. Sci. USA 101, 12 952–12 956). Here, we use an extended
version of the Cambridge two-dimensional chemistry–transport model to evaluate
quantitatively this possibility along with two other potential causes of O3 loss at this
time: (i) direct effects of HCl release by the Siberian Traps and (ii) the indirect release of
organohalogens from dispersed organic matter. According to our simulations, CH3Cl
released from the heating of coals alone caused comparatively minor O3 depletion (5–20%
maximum) because this mechanism fails to deliver sufficiently large amounts of Cl into the
stratosphere. The unusual explosive nature of the Siberian Traps, combined with the direct
release of large quantities of HCl, depleted the model O3 layer in the high northern latitudes
by 33–55%, given a main eruptive phase of less than or equal to 200 kyr. Nevertheless, O3
depletion was most extensive when HCl release from the Siberian Traps was combined with
massive CH3Cl release synthesized from a large reservoir of dispersed organic matter in
Siberian rocks. This suite of model experiments produced column O3 depletion of 70–85%
and 55–80% in the high northern and southern latitudes, respectively, given eruption
durations of 100–200 kyr. On longer eruption time scales of 400–600 kyr, corresponding O3
depletion was 30–40% and 20–30%, respectively. Calculated year-round increases in total
near-surface biologically effective (BE) UV-B radiation following these reductions in O3
layer range from 30–60 (kJ mK2 dK1)BE up to 50–100 (kJ mK2 dK1)BE. These ranges of daily
UV-B doses appear sufficient to exert mutagenic effects on plants, especially if sustained over
tens of thousands of years, unlike either rising temperatures or SO2 concentrations.
Keywords: atmospheric O2; atmospheric CO2; flood basalt eruptions;
genome stability; mutations, ozone photochemistry; polar stratospheric clouds
* Author for correspondence (d.j.beerling@sheffield.ac.uk).
One contribution of 18 to a Discussion Meeting Issue ‘Trace gas biogeochemistry and global
change’.
1843
This journal is q 2007 The Royal Society
1844
D. J. Beerling et al.
1. Introduction
The Permo-Triassic (P-Tr) mass extinction, 251 Myr BP, is recognized as the
largest extinction event in the Phanerozoic, with the global loss of up to 95% of
all marine species (for reviews, see Erwin (1994) and Benton & Twitchett
(2003)). The dramatic nature of the extinction itself, associated environmental
disturbance, and the recovery of ecological and climatic systems have been
extensively documented and continue to remain the focus of considerable
research effort (Erwin 1994; Benton & Twitchett 2003). One of the most
remarkable claims to emerge in recent years is that extensive volcanic activity
over a vast area of Siberia coincident with the extinctions (Renne et al. 1995) led
to a worldwide depletion of stratospheric O3 (Visscher et al. 2004). The basis for
this palaeobotanical hypothesis was the identification of unusual ‘mutated’ forms
of fossil spores and pollen (palynomorphs) in end-Permian rocks (Visscher et al.
2004; Foster & Afonin 2005; Looy et al. 2005). Mutated spores of herbaceous
lycopsids are represented by a permanent tetrad configuration indicative of a
failure of individual spores to separate and complete the normal process of
development. These mutations have been identified in end-Permian sediments
from a global suite of localities ranging from Greenland to Australia (Visscher
et al. 2004; Looy et al. 2005). Subsequently, fossil gymnosperm pollen mutations
with disrupted sacci that are indicative of a failure to form normal bisaccate
grains have been reported from localities in Russia and northwest China (Foster &
Afonin 2005).
The widespread occurrence of mutated fossil palynomorphs has been suggested
to represent a chronic mutagenesis event brought about by worldwide increases
in near-surface fluxes of ultraviolet-B radiation (UV-B, 280–315 nm; Visscher
et al. 2004; Looy et al. 2005). Visscher et al. (2004) postulated that increased
surface UV-B radiation resulted not from the direct chemical effects of the
massive Siberian flood basalt eruption, but instead from the production of large
quantities of organohalogens produced by the heating of organic-rich rocks and
hydrothermal fluids as the Siberian Traps erupted over several million square
kilometres. Halogen emissions are extremely important to atmospheric chemistry
because they have the potential to destroy stratospheric O3. Limited
experimental evidence from Glycine max (soybean) supports the contention
that moderately enhanced fluxes of UV-B radiation can induce damage to pollen
and impair germination (Koti et al. 2005). However, exposure of gymnosperm
trees to high SO2 in polluted areas in the former Soviet Union has also produced
aberrant pollen and spore mutations (Mičieta & Murı́n 1996; Tretyakova &
Noskova 2004), and substantial SO2 production from the Siberian Traps may
represent another causal agent involved in the mutated fossil spores.
We are unaware of any investigations examining the direct or indirect effects of
the Siberian Traps eruptions on the chemistry of the atmosphere, although other
modelling studies have considered the aspects of environmental change associated
with a major end-Permian carbon cycle perturbation (Berner 2002; Schmidt &
Shindell 2003; Grard et al. 2005; Lamarque et al. 2006) and anoxic oceans (Kump
et al. 2005; Berner & Ward 2006). Therefore, the central aim of the present paper is
to provide a quantitative framework for evaluating the possible effects of the
Siberian eruptions on the photochemistry of the atmosphere, with particular
emphasis on stratospheric O3 chemistry and transport.
Phil. Trans. R. Soc. A (2007)
Ozone layer stability in the end-Permian
1845
We first briefly review the nature of the eruptions of Siberian Traps and
regional geology, and detail our assessment of the masses of gases released
directly, and indirectly, by the heating of organic- and halogen-rich Neoproterozoic and Palaeozoic sediments. Section 2 outlines the Cambridge twodimensional atmospheric chemistry–transport model and the modifications
incorporated to simulate the photochemical effects of large volcanic eruptions,
including representation of the catalytic properties of polar stratospheric clouds
(PSCs) on O3 destruction (Hanson & Mauersberger 1988), the absorption and
scattering of UV-B radiation by sulphate aerosols (Kreidenweis et al. 1999; Chin
et al. 2002) and a sulphate aerosol scheme to calculate the aerosol surface area
and heterogeneous reaction rates (Bekki & Pyle 1992). We then present a series
of model sensitivity studies for the end-Permian, investigating how O3 chemistry
is influenced by (i) the HCl release from main eruptive phase of the Siberian
Traps, (ii) organohalogen (CH3Cl) emissions from the heating of coals, (iii) rocks
containing dispersed organic matter, and (iv) the influences of simultaneous
releases from sources (i) to (iii). Our analyses include a preliminary assessment of
the effect of uncertainties in the time span of the main eruptive phase of the
Siberian Traps (Kamo et al. 2003). The results are discussed in the context of
increased near-surface UV-B radiation fluxes and tropospheric SO2 loading and
their possible impacts on plant genome stability.
2. The Siberian Traps and estimated gas emissions
The Siberian Traps flood basalt province is the largest known subaerial event of
its kind and presently covers a modest 3.4!105 km2 of northwest Siberia.
Inclusion of pyroclastic and intrusive sediments increases the total area of the
province to 1.5!106 km2 although the original area was undoubtedly much
larger (Wignall 2001). The eruptions are estimated to have produced an
immense volume of basalt lava, with estimates ranging from 2 to 4!106 km3,
although only 4!105 km3 remain (Courtillot et al. 1999). Radiometric dating
(40Ar/39Ar) indicates that the main eruptive phase was broadly contemporaneous with the P-Tr boundary extinction event (Renne et al. 1995; Reichow et al.
2002), and U–Pb dating of lava flows indicates that the Siberian flood volcanism
was initiated at 251.7G0.4 Ma and waned by 251.1G0.3 Ma (Kamo et al. 2003).
These dates establish the duration of the main eruptive duration as 600 kyr and
are in line with earlier estimates derived from magnetostratigraphy (Haag &
Heller 1991); more detailed consideration of the ages of the different intrusive
complexes suggests a substantial proportion of the main Siberian flood basalt
eruption lasted only 100 kyr (Kamo et al. 2003).
Geological evidence indicates that the Siberian eruptions may have been
unusually explosive (Campbell et al. 1992), with pyroclastic deposits reaching a
thickness of up to 800 m (Khain 1985). The highly explosive nature of the
Siberian eruptions expelled Permian coal-bearing deposits and Devonian
sediments from a depth of 0.3–1.5 km, and occasionally from depths of 10 km
(Campbell et al. 1992). Current estimates suggest that the igneous province
contains approximately 20% pyroclastic material (by volume), with the
remainder comprising intrusive dykes and sills (50%) and basalt lava flows
(30%; Reichow et al. 2002). The unusually explosive nature of the eruptions
Phil. Trans. R. Soc. A (2007)
1846
D. J. Beerling et al.
is noteworthy because this could represent the primary transport mechanism
for injecting O3-depleting halogen gases through the troposphere and into
the stratosphere.
The potential for the Siberian eruptions to disrupt the balance between
photochemical production/destruction and transport of stratospheric O3 depends
on the mass of inorganic and organic gases released. For the present simulations,
we estimated the mass of hydrochloric acid (HCl) and sulphur dioxide (SO2)
emissions by scaling the detailed inventory of emissions drawn up for the
Columbia River Basalt (CRB) Group that erupted approximately 14.7 Ma
(Thordarson & Self 1996). According to Thordarson & Self (1996), the CRB
produced 1.3!103 km3 of basalt and liberated 12 425 Mt SO2 and 704 Mt HCl
(table 1). The approximate mass of trace organic gases released is estimated by
scaling the global average fumerolic and diffuse halocarbon source strengths for
the current annual gaseous volcanic CO2 release (78 Tg yrK1; Schwandner et al.
2004) by the mass of CO2 released during the Siberian eruptions, given a known
volume of basalt produced and a CO2 degassing rate of 1.6!104 Tg of CO2 per
1 km3 of basalt erupted (Gerlach & Graeber 1985; table 1).
A further unusual aspect of the Siberian eruptions was the nature of regional
geology of the Siberian platform. The main lithologies of the region are coals
(Czamanske et al. 1998), hydrocarbon source rocks (Kontorovich et al. 1997) and
evaporites (Melnikov et al. 1997). The juxtaposition of a volcanic province with
coal-bearing deposits, Neoproterozoic oil accumulations and salt layers
(evaporates) may have produced massive quantities of organohalogens (Visscher
et al. 2004). In particular, contact metamorphism of coals and the development
of hydrothermal systems rich in chlorine, produced from the pressure dissolution
of the surrounding evaporites, potentially synthesized large amounts of the
organohalogens methyl chloride (CH3Cl) and methyl bromide (CH3Br).
As the relative proportions of the organohalogens emitted is unknown, and
because Br availability is severely limited when compared with Cl, for the
purposes of the present simulations, we estimated the mass of CH3Cl released as
follows. The concentration of Cl in coal varies from 300 to 1100 ppmv (Vassilev
et al. 2000), but because of the high Cl concentration of Russian ‘salt coals’
(Goodarzi 1987; Bragg et al. 1991; Greive & Goodarzi 1993; Vassilev et al. 2000),
we assume the upper concentration for the Siberian coals. The thermal maturity
of the Siberian coals at the time of the eruption is unknown. Therefore, we have
assumed that the Palaeozoic Siberian coals at the time of emplacement are better
approximated by semi-anthracite than thermally immature peats. Semianthracite has a typical aliphatic content of approximately 20% (Suggate &
Dickinson 2004) and an ash content, which can dilute the aliphatic content of
coal, of approximately 30% (Medina et al. 2006) and so contains 0.0098 moles of
CH2 and up to 3.10!10K5 mol of Cl.
In the case of CH3Cl, the supply of both Cl and H may not have been limited
due to the nature of the regional geology, allowing 1 g of semi-anthracite to
release 0.49 g of CH3Cl, calculated as the product of the molar concentration
of CH2 (0.0098) and the molecular weight of CH3Cl (50.5). The generation of
a higher mass of organohalogens than the original mass of semi-anthracite
occurs owing to the high atomic weight of chlorine when compared with the
methyl group.
Phil. Trans. R. Soc. A (2007)
Phil. Trans. R. Soc. A (2007)
3.4!10K2
6.0!10K1
1.3!10K8
9.9!10K7
2.7!10K7
1.2!10K6
4.4!10K8
1.1!10K7
1.6!10K7
9.4!10K8
gas emitted
hydrochloric acidb
sulphur dioxideb
methyl bromidec
methyl chloridec
methylene chloridec
trichloromethanec
tetrachloromethanec
trichlorofluromethanec
chlorobenzenec
p-dichlorobenzenec
6.5!105
1.2!107
0.24
19.1
5.2
23.4
0.84
2.1
3.1
1.8
basaltsa total
degassed (Tg)
4.3!105
7.7!106
0.16
12.7
3.5
15.6
0.6
1.4
2.1
1.2
1.1!106
1.9!107
0.4
31.8
8.6
38.9
1.4
3.5
5.2
3.0
instrusivesa total
degassed (Tg)
2.2!106
3.8!107
0.8
63.6
17.3
77.9
2.8
7.0
10.4
6.0
total degassed
(Tg)
flux
600 kyr
time scale
(Tg yrK1)
22.0
3.7
380.0
63.3
8.0!10K6 1.3!10K6
6.4!10K4 1.0!10K4
1.7!10K4 2.9!10K5
7.7!10K4 1.3!10K4
2.8!10K5 4.7!10K6
7.0!10K5 1.2!10K5
1.0!10K4 1.7!10K5
6.0!10K5 1.0!10K5
flux 100 kyr
time scale
(Tg yrK1)
a
Total CO2 emission for each rock type calculated as volume!degassing rate (1.6!1013 g CO2 kmK3; Gerlach & Graeber 1985). We take total volume
of basalt produced by the Siberian Traps to be equal to 4.0!106 km3, with 30% basalts (1.2!106 km3), 20% pyroclastics (8.0!105 km3) and 50%
intrusives (2.0!106 km3; Kamo et al. 2003).bRelease rates calculated from total volume and estimated total masses of HCl and SO2 produced by the
Columbia River Basalt Group ( Thordarson & Self 1996).cRelease rates calculated from global estimates of volcanic fumarolic and diffuse degassing
summed and scaled by global CO2 degassing (78.0 Tg yrK1; Schwandner et al. 2004).
release rate
(Mg per Mg CO2)
pyroclasticsa
total degassed
(Tg)
Table 1. Estimated volcanic release of gases during the Siberian Traps eruptions included in model simulations: runs 1 and 4.
Ozone layer stability in the end-Permian
1847
1848
D. J. Beerling et al.
Table 2. Estimated methyl chloride (CH3Cl) emissions from contact metamorphism of Siberian
coals during the eruptions of the Siberian Traps.
gas emitted
estimated
total Siberian
coal reserves
(Gt)
for semi-anthracite
methyl
2000
chloride
emissiona
rate
(g gas gK1
coal)
total
emissions
(Tg)
0.49
9.8!105
flux 100 kyr
time scale
(Tg yrK1)
9.9
for dispersed organic matter in sediments of the Tunguska formation
103.9
methyl
21 000
0.49
1.0!107
chloride
flux 600 kyr
time scale
(Tg yrK1)
model
simulation
1.7
run 2
17.3
run 3
a
Emission rate of CH3Cl per unit of semi-anthracite are calculated assuming both H and Cl are
unlimited, in agreement with the interbedded coals, salt pans and hydrothermal systems in Siberia.
See text for details.
Finally, the calculated release of CH3Cl per unit mass of semi-anthracite
heated was scaled against two estimates of total organic carbon. The first was
based on the mass of organic carbon buried as Palaeozoic coals in the Tunguska
basin, taken to be approximately 2000 Gt (Czamanske et al. 1998; Akulov 2006;
table 2). However, a further source of organic matter is that dispersed in clastic
sediments of the region, which although not explicitly recognized by Visscher
et al. (2004), may represent a large additional source of organic precursor for
organohalogen synthesis. An estimate of the mass of organic carbon present can
be made, given the Tunguska basin is approximately 1!106 km2, and assuming
an average thickness of 0.5 km, a bulk density of shale rock of 2100 kg mK3 and a
total organic carbon content of 2% by mass. These figures yield 21 000 Gt C that
may additionally be available for the synthesis of organohalogens. Therefore, a
second estimate of the total mass of organohalogens has been obtained by scaling
the emission of CH3Cl per unit mass of semi-anthracite with this figure.
3. Two-dimensional atmospheric chemistry–transport modelling
(a ) Model outline
All the modelling experiments were undertaken using the Cambridge twodimensional atmospheric chemistry–transport model (Harwood & Pyle 1975;
Law & Pyle 1993). The two-dimensional model includes the two key transport
processes governing the global distribution of atmospheric O3 as a function of
altitude and latitude: advection due to the zonal mean circulation and eddy
transport arising from departures from the zonal mean. The zonal mean
circulation is calculated from forcing by radiative, latent and sensible heating.
Departures from the zonal mean are parametrized by eddy diffusion coefficients; see
Harwood & Pyle (1975, 1980) for further details. The inclusion of these transport
processes overcomes a major limitation in the use of one-dimensional models in
investigations of stratospheric O3 photochemistry (e.g. Kump et al. 2005).
Phil. Trans. R. Soc. A (2007)
Ozone layer stability in the end-Permian
1849
The model has been used extensively to consider the chemistry and dynamics of
the middle atmosphere (Harwood & Pyle 1980; Haigh 1984; Bekki & Pyle 1992;
Bekki et al. 1994), the region of interest for this investigation.
The two-dimensional model chemistry scheme includes 178 reactions, involving
the main species governing the chemistry of the atmosphere. The basic model used
here includes the gas phase chemistry of the Ox, HOx, NOy, ClOy, BrOy, SOx
families, as well as an oxidation scheme for methane and other representative nonmethane hydrocarbons. We have implemented a sulphate aerosol scheme as
detailed by Bekki & Pyle (1992), which provides the aerosol surface area used to
calculate heterogeneous reaction rates. Details of calculation procedures,
including the treatment of stratospheric photolysis, cloud, aerosol and tropospheric water vapour schemes implemented are given in Harfoot et al. (2007).
The model atmosphere is heated from below by prescribed mean monthly seasurface temperatures (SSTs). SSTs influence surface-to-air heat fluxes in the
lowermost level and, following the adiabatic lapse rate, the temperature up to the
tropopause. Radiative heating is calculated interactively for the stratosphere: in
the long wave from the mixing ratios of CO2, H2O, O3, SO2 and N2O (Haigh 1984);
and, in the shortwave, from O2 and O3 mixing ratios. Radiative cooling is fixed at
1.5 K dK1 in the lower troposphere and 0.75 K dK1 in the upper troposphere.
Because the heating rate of the troposphere is prescribed, increasing CO2 exerts no
changes in the heating rate there; however, imposing elevated SSTs representative
of warmer climate does influence tropospheric conditions as described in §3c.
(b ) Model development
We made two modifications to the model for the purposes of the present
investigations. First, we developed and incorporated a simple parameterization
for empirically representing the chemical effects due to the formation of HNO3–H2O
clouds, otherwise known as type I PSCs. For those grid boxes containing water
vapour and HNO3 at temperatures below 195 K, a threshold temperature for type I
PSC formation under typical stratospheric conditions (Hanson & Mauersberger
1988), heterogeneous activation of reservoir forms of chlorine (described by
reactions (4.1) and (4.2) in §4) was assumed with a time constant of 7 days. This
time constant is consistent with the time scale observed for formation of the
springtime ‘active chlorine’ layer within the present day Antarctic lower
stratosphere (Solomon et al. 2002). Second, we included a scheme to account for
the absorption and scattering of solar UV-radiation by sulphate aerosol from
Kreidenweis et al. (1999). Because calculation of the particle size and wavelengthdependent aerosol extinction coefficient, Qext, is computationally expensive, we use
here a constant mean value of 1.4, the value for sulphate aerosol reported by Chin
et al. (2002) at lZ500 nm.
In all the simulations of the Siberian Traps volcanism, gases are emitted directly
into the atmosphere in a single pulse (table 1) and distributed over different
heights, according to the eruption phase from which they originate and the water
solubility of each species (Textor et al. 2003). More highly soluble gases are likely
to be incorporated into hydrometeors formed in the rapidly ascending eruption
column and are consequently more likely to be precipitated out of the atmosphere.
HCl is the most chemically important halogenated compound, released directly
by the Siberian Traps volcanism, comprising some 5% by mass of the volcanic
Phil. Trans. R. Soc. A (2007)
1850
D. J. Beerling et al.
emissions (table 1). Owing to its high solubility in water, Textor et al. (2003)
estimated the proportion of HCl reaching the stratosphere within an explosive
volcanic eruption plume would be around 25%. However, this value is relevant for a
tropical (i.e. low-latitude) eruption and the Siberian Traps erupted at a much
higher latitude where the tropopause is lower. Based on measurements of the
volcanic plume from the 2002 eruption of Hekla volcano on Iceland made in
the lower stratosphere (Millard et al. 2006; Rose et al. 2006), we allow 75% of the
emitted HCl to reach the model stratosphere. This stratospheric injection is
distributed uniformly over all stratospheric levels below approximately 18 km.
(c ) Boundary conditions for model simulations
All simulations include the effects of the warm end-Permian climate and a low
O2/high CO2 atmospheric composition (Berner 2005) on O3 chemistry and
transport. We used zonally averaged SSTs reported for a coupled ocean–
atmosphere model simulation for the warm end-Permian climate (Kiehl &
Shields 2005). This altered the chemical composition of the troposphere,
particularly by elevating tropospheric water vapour concentrations and exerted
a strong influence on stratospheric H2O. The major source of stratospheric H2O is
transport from the troposphere through the tropical tropopause ‘cold trap’,
although the oxidation of CH4 in the stratosphere also makes a contribution.
Assuming a constant adiabatic lapse rate through the troposphere, warmer SSTs
would give rise to a corresponding change in the temperature of the cold trap,
where tropospheric water vapour is dehydrated to its local saturation pressure.
The warm end-Permian SSTs could therefore have strongly influenced stratospheric water vapour content, and hence stratospheric O3 through the chemical
cycles of hydrogen radical species. This effect has been included in our
simulations by imposing a warmer tropical tropopause cold trap temperature
value equivalent to the difference in pre-industrial and end-Permian tropical
SSTs (Kiehl & Shields 2005); at the bottom boundary of the model stratosphere,
the H2O concentration is fixed at the saturation vapour pressure over ice for the
temperature of the tropopause.
The end-Permian atmospheric composition influences O3 chemistry and
transport in different ways that are important to represent in the modelling.
Reducing the atmospheric O2 content from the present-day value of 21% to a late
Permian low of 15% (Berner 2005) can exacerbate stratospheric O3 loss over midand high latitudes (Harfoot et al. 2007). Stratospheric O3 is produced by photolysis
of O2 (Chapman 1930) with a rate determined by the product of the photolysis
frequency, J2, and the concentration of O2, [O2]. In the subtropical stratosphere a
‘self-healing’ effect occurs. Decreased [O2] leads to a reduction in O3 production in
the upper atmosphere and lower [O3]. Given that the photolysis frequency at any
altitude is itself dependent on the concentrations of O2 and O3 at higher levels in
the atmosphere, decreased [O2] and [O3] above enhance the rate of J2 at lower levels
and therefore [O3]. The net result is effectively no major changes in the total low
latitude O3 column amount. At high latitudes in the middle and upper
stratosphere, [O3] decreases with [O2] decline in a similar manner. However, in
the lower stratosphere, where the chemistry is slower and transport dominates, the
effect of self-healing is masked. Consequently, at high latitudes the O3 column thins
as atmospheric O2 falls. A CO2-rich atmosphere traps outgoing long-wave
Phil. Trans. R. Soc. A (2007)
1851
Ozone layer stability in the end-Permian
Table 3. Description of model simulations.
emission
duration (kyr)
simulation description
run 0
run 1
run 2
run 3
run 4
background end-Permian simulation: 15% O2,
2000 ppm CO2, warmer SSTs and tropical
tropopause temperatures
run 0 including emissions from Siberian Traps
volcanism (table 1)
run 0 including emissions of CH3Cl from heating of
semi-anthracite (table 2)
run 0 including emissions of CH3Cl from heating of
dispersed organic sediments (table 2)
run 0 including all emission mechanisms
Siberian Traps volcanism (table 1)
CH3Cl from semi-anthracite heating (table 2)
CH3Cl from heating of dispersed organic sediments
(table 2)
simulation
name
run 0
100
200
400
600
100
600
100
600
100
200
400
600
run
run
run
run
run
run
run
run
run
run
run
run
1A
1B
1C
1D
2A
2B
3A
3B
4A
4B
4C
4D
radiation in the lower atmosphere. In our model, a 2000 ppm CO2 atmosphere
reduces the annual mean temperature of the mid-stratosphere by as much as 10 K,
when compared to the pre-industrial atmosphere, resulting in temperatures below
195 K, especially at high latitudes in both hemispheres and persisting for much of
the year. Therefore, high CO2 effectively creates more widespread and long-lived
conditions for the formation of PSC.
(d ) Model experiments
Using the modified two-dimensional model and the above boundary conditions, we
performed the following suite of experiments to investigate the sensitivity of
stratospheric O3 to Siberian Traps volcanism and organohalogen release from coals
and rocks (table 3). Run 1: calculated the effects of HCl release from the Siberian
Traps, with different durations of the main eruptive phase (100, 200, 400 and 600 kyr)
to encompass uncertainties in the dating. These experiments were performed by
decreasing the annual flux of HCl injected into the stratosphere, as given by total
mass of HCl/duration. Runs 2 and 3 investigated the effects of CH3Cl emissions by
heating coals and dispersed organic matter, respectively, given a main eruptive phase
of either 100 or 600 kyr. Run 4 investigated the combined influences of runs 1, 2 and 3
occurring simultaneously with main eruption phases of 100, 200, 400 and 600 kyr.
4. Results and discussion
(a ) Siberian Traps versus CH3Cl release from rocks
The simulated degree of total O3 column depletion due to HCl release from the
Siberian Traps, with a representative Late-Permian climate and atmospheric
composition (run 1), depends upon the duration of the main eruptive phase
Phil. Trans. R. Soc. A (2007)
1852
D. J. Beerling et al.
0
3.25
(a)
400
(c)
300
300
200
global annual mean
Southern Hemisphere annual mean
Northern Hemisphere annual mean
100
0
60
200
Permian
600 kyr
400 kyr
200 kyr
100 kyr
20 kyr pulse
100
0
60
(d )
(b)
40
40
20
20
0
background
600
400
200
simulation duration (kyr)
100
–60
–30
0
latitude
30
60
percentage of O3 loss
percentage of O3 loss
O3 column depth (DU)
400
O3 column depth (DU)
stratospheric HCI (Tg yr–1)
0.5417 0.8125
1.625
0
90
Figure 1. Changes in annual mean O3 column depth (DU and percentage loss, for run1) (a, b) as a
function of the duration of the main eruptive phase of the Siberian Traps (background, 100, 200,
400 and 600 kyr) and (c, d ) expressed as a function of latitude. Du, Dobson units.
(figure 1a–d ). Zonally averaged, O3 depletion is the greatest in the Northern
Hemisphere and least in the Southern Hemisphere (figure 1a,b). In high northern
latitude regions above the local injection of HCl into the stratosphere, O3
depletion reaches 55 to 33% for the 100 and 200 kyr simulations, respectively,
and compares with 26 to 14% depletion in the high southern latitudes for
the same simulations (figure 1c,d). For the two longer eruption time scales
considered, the total O3 layer column thinned by approximately 20 and 10% in
the high northern and southern latitudes, respectively. As a sensitivity test, we
also undertook a further simulation to evaluate the possible influence of a short
discrete pulse in which one-quarter of the total mass of HCl is released within a
few tens of thousands of year, in this case 20 kyr. The results indicate the
potential for greater O3 depletion shown for the 100 kyr simulation and highlight
the possibility that discrete abrupt pulses of volcanic activity could increase the
severity of atmospheric disturbance.
In all of these simulations, high latitude O3 depletion occurs because extensive
PSC formation provides suitable conditions for conversion of chlorine held as
reservoir species to active ‘Cl’ via reactions (4.1) and (4.2)
ClONO2 C H2 O/ HNO3 C HOCl;
ð4:1Þ
ClONO2 C HCl/ HNO3 C 2Cl:
ð4:2Þ
The HOCl produced via reaction (4.1) is rapidly photolysed in sunlight to form
Cl. Active chlorine (ClOx) in the form of Cl or ClO can participate in catalytic O3
Phil. Trans. R. Soc. A (2007)
1853
Ozone layer stability in the end-Permian
(b)
40
200
100
30
20
Permian control
semi-anthracite 100 kyr
semi-anthracite 600 kyr
10
0
(d )
60
50
300
40
200
30
20
100
10
0
–90
O3 column depth (DU)
50
300
0
400 (c)
percentage of O3loss
60
–60
–30
0
latitude
30
60
90 –90
–60
–30
0
latitude
30
60
percentage of O3 loss
O3 column depth (DU)
400 (a)
0
90
Figure 2. Changes (DU and percentage loss) in annual mean O3 column depth resulting from
CH3Cl release from the heating of (a, b) coals (run 2) and (c, d ) dispersed organic matter (run 3).
Du, Dobson units.
destruction cycles, the most general example being
Cl C O3 / ClO C O2 ;
ClO C O/ Cl C O2 ;
ð4:3Þ
ð4:4Þ
Net : O3 C O/ O2 C O2 :
ð4:5Þ
Under conditions of elevated active Cl and particularly in the cold, highlatitude lower stratosphere where atomic oxygen is less common, catalytic O3
destruction proceeds via the ClO dimer cycle (Molina & Molina 1987)
2ðCl C O3 / ClO C O2 Þ;
ð4:6Þ
ClO C ClO/ ClOOCl;
ð4:7Þ
ClOOCl C hn/ Cl C ClOO;
ð4:8Þ
ClOO C M/ Cl C O2 ;
ð4:9Þ
ð4:10Þ
Net : 2O3 C hn/ 3O2 :
In the model, the creation of PSCs leading to O3 loss sets up a positive
feedback because O3 absorbs incoming short-wave radiation as well as outgoing
long-wave radiation. Significant O3 depletion leads to a local cooling of the
depleted region of the atmosphere, so increasing the prevalence of PSC formation
and hence reinforcing O3 depletion.
The second set of model calculations (runs 2 and 3) directly evaluated the
suggestion (Visscher et al. 2004; Looy et al. 2005) that organohalogens released
by the heating of organic-rich sediments as the Siberian eruptions proceeded was
Phil. Trans. R. Soc. A (2007)
1854
D. J. Beerling et al.
300
200
200
global annual mean
Southern Hemisphere annual mean
Northern Hemisphere annual mean
100
100
(b)
0
(d )
80
80
60
60
40
40
20
20
0
background
600
400
200
simulation duration (kyr)
100
O3 column depth (DU)
400
(c)
300
0
percentage of O3 loss
(a)
–60
–30
0
latitude
30
60
percentage of O3 loss
O3 column depth (DU)
400
0
90
Permian
600 kyr
400 kyr
200 kyr
100 kyr
Figure 3. Changes in annual mean O3 column depth (DU and percentage loss) due to HCl release
from the Siberian Traps and CH3Cl release from the heating of coals and rocks combined (run 4) as
a function of (a,b) the duration of the main eruptive phase (background, 100, 200, 400 and 600 kyr)
and (c,d ) expressed as a function of latitude. Du, Dobson units.
a primary mechanism of stratospheric O3 depletion. In these experiments,
volcanic emissions were switched ‘off’ and the largest estimated mass of CH3Cl
released through this route (9.8!105 Tg; table 2) was introduced into the
lowermost troposphere on a time scale of 100 or 600 kyr to represent the same
time scale as the eruption of the Siberian Traps. Results indicate that neither
experiment had substantial deleterious effects on stratospheric O3, with the
100 kyr simulation (run 2A) thinning the O3 column by 20% (figure 2a,b). These
reductions in O3 are comparable to those resulting from the volcanic runs 1C and
1D and contrast with the greater losses that occurred on shorter time scales
(figure 1). The difference arises primarily because the mass of CH3Cl released
from the heating of organic-rich sediments into the bottom of the atmosphere is
around half that of HCl released by the volcanism (2.2!106 Tg; table 1) and only
three times larger than the mass of HCl emitted directly into the stratosphere
through pyroclastic volcanism (3.25!105 Tg). CH3Cl in the model end-Permian
atmosphere has a relatively short lifetime of around six months, so that most
CH3Cl is oxidized to form inorganic chlorine, ClOy (Cl, ClO, HOCl, Cl2O2,
ClOO, ClONO2, HCl), and only a small fraction is transported up into the
stratosphere. ClOy is removed from the troposphere by rainout of HCl, the only
loss process for chlorine in the model, calculated using latitudinally and
Phil. Trans. R. Soc. A (2007)
1855
Ozone layer stability in the end-Permian
altitude (km)
40
30
20
10
0.50
–0.16
–0.81
–1.47
–2.12
–2.78
–3.44
–4.09
–4.75
–5.41
–6.06
–6.72
–7.38
–8.03
–8.69
–9.34
–10.00
ozone (ppm)
(b)
50
0.50
–0.16
–0.81
–1.47
–2.12
–2.78
–3.44
–4.09
–4.75
–5.41
–6.06
–6.72
–7.38
–8.03
–8.69
–9.34
–10.00
ozone (ppm)
(a)
0
(c)
(d )
50
altitude (km)
40
30
20
10
0
–90 –60 –30
0
30
latitude
60
90 –90 –60
–30
0
30
latitude
60 90
Figure 4. Changes in the annually averaged O3 mixing ratio by altitude and latitude for four
scenarios given releases over 100 kyr for (a) the eruption of the Siberian Traps (run 1A), (b) CH3Cl
release from the heating of coals (run 2A), (c) dispersed organic matter (run 3A) and (d ) releases
from all sources combined (run 4A).
seasonally varying first-order loss rates following Giannakopoulos (1998), which
are the same for all integrations. As a result, increasing the ClOy burden of the
atmosphere, by adding large amounts of volcanic chlorine, for example, extends
its lifetime correspondingly and so increases the amount of ClOy that reaches the
stratosphere. Comparing the steady-state stratospheric burden of chlorine, in the
form of ClOy, for the different simulations highlights the impact that each
scenario has upon levels of chlorine in the stratosphere. The stratospheric burden
calculated for CH3Cl release from coal heating over 100 kyr (run 2A) is 0.87 Tg,
around twice the burden for the end-Permian background atmosphere (run 0;
0.48 Tg) and only half that calculated for 100 kyr of Siberian Traps volcanism
(run 1A) at 2.13 Tg. The comparatively small impact of CH3Cl released from the
heating of organic-rich sediments results in correspondingly reduced O3
destruction through photochemical cycles involving Cl, giving rise to the lower
O3 depletion calculated for this scenario.
In an additional scenario, we consider the possibility that the thermal
degradation of dispersed organic matter in rocks provided the organic precursors
for generating organohalogens (run 3). Here, the estimated total mass of CH3Cl
emitted is approximately ten times larger than that generated by heating coals
alone (table 2), giving rise to high tropospheric CH3Cl concentrations with
Phil. Trans. R. Soc. A (2007)
1856
D. J. Beerling et al.
mixing ratios of 5–10 ppb in the Northern Hemisphere. The tropospheric lifetime
of CH3Cl is governed primarily by its reaction with OH, the main source of which
is from reaction of H2O with O(1D) from the photolysis of O3 by UV-B radiation.
Significant thinning of the stratospheric O3 layer enhances the penetration of
UV-B into the troposphere, which increases the O3 photolysis frequency and
results in OH concentrations comparable to those calculated for the end-Permian
background atmosphere, despite the addition of large quantities of CH3Cl.
Consequently, the rate of CH3Cl oxidation in the troposphere remains fast, with
a steady-state lifetime of around six months, equivalent to that for the heating of
coals only, and a similarly limited fraction of the tropospheric CH3Cl burden is
transported to the stratosphere. However, given the elevated tropospheric mass
that results for the heating of dispersed organic matter, this equivalent fraction
equates to a more significant mass of CH3Cl. Levels of ClOy are enhanced
considerably in the troposphere due to its oxidation, greatly extending the
lifetime of the ClOy family, with more ClOy reaching the stratosphere as a result.
The stratospheric burden of Cl in the form of ClOy is 4.5 Tg for a 100 kyr release
of CH3Cl from heating of dispersed organic matter (run 3A)—over five times that
for the heating of coals alone (run 2A)—and produces significant O3 depletion,
with 50% losses in both the high northern and southern latitudes (figure 2c,d ).
The O3 depletion at high latitudes occurs as a result of the low stratospheric
temperatures over these latitudes, which are below the formation threshold of the
PSCs in the model over much of the year. The heterogeneous activation of such a
large mass of chlorine by PSCs provides abundant ClOx to catalyse
photochemical O3 destruction.
In the final set of model simulations (run 4), we investigated the effects of both
mechanisms of O3 depletion, HCl injection from the Siberian Traps and
organohalogen release from the heating of organic-rich sediments, occurring
simultaneously. As would be expected, this scenario produced significant O3
depletion, with average losses in the Northern and Southern Hemispheres of 20–30%
for the 600 kyr duration experiment, increasing to 60–80% depletion for the 100 kyr
simulation (figure 3a,b). In these experiments, the natural latitudinal gradient in O3
column thickness was severely disrupted, with losses of 70–90 and 55–80% in the high
northern and southern latitudes, respectively, for emission durations of 100 and
200 kyr. Even in the 400 and 600 kyr simulations, the high northern and southern
latitudes lost 30–40% and 20–30%, respectively, of total O3 column (figure 3c,d ).
(b ) Pattern and mechanism of O3 depletion
Figure 4 shows annually averaged, latitude–altitude cross-sections of O3 loss
for the 100 kyr simulations of (i) the eruption of the Siberian Traps (run 1A), (ii)
CH3Cl release from heating coals (run 2A), (iii) CH3Cl release from country
rocks (run 3A), and (iv) all three mechanisms together (run 4A), when compared
with the background end-Permian atmosphere (run 0). All simulations show O3
depletion in a band that stretches across all latitudes in the upper stratosphere
between 30 and 40 km (figure 4). The patterns of O3 loss are well explained by
the corresponding distribution of annually averaged active chlorine (ClOx) in
the atmosphere that lies in the upper stratosphere around 35 km and is enhanced
for all simulations (figure 5). The higher ClOx concentrations drive faster
O3 destruction through photochemical cycles, such as described by reactions
Phil. Trans. R. Soc. A (2007)
Ozone layer stability in the end-Permian
1857
(4.3)–(4.5). Peak O3 depletion in the upper stratosphere occurs over subtropical
latitudes, where solar irradiance is most intense and thus drives faster
photochemistry, including the Cl-catalysed O3 destruction pathways. Owing to
this, in the simulated CH3Cl release from heating of dispersed organic matter,
and of all mechanisms together, O3 is depleted by maximum values of 7–9 ppm
over the equator at approximately 30 km (figure 4c,d ). Beneath this region, at
an altitude of 15–25 km, areas of less severe O3 loss are apparent due to
increased penetration of UV-B through the depleted region above, which
increases the photolysis frequency of O2, J2, and accelerates the rate of O3
production, thus offsetting much of the O3 depletion in these regions.
For the Siberian Traps simulations, maximum O3 depletion is not as significant,
reaching 4–5 ppm in the tropical mid-stratosphere, with comparable O3 losses of
approximately 3 ppm in the lower stratosphere above the latitude of eruption
(figure 4a). This more localized depletion is driven by the direct injection of
pyroclastic volcanic emissions, giving rise to local sulphate aerosol surface area
densities that are higher than anywhere else in the atmosphere and elevated ClOy
mixing ratios up to 9 ppb that together activate high local concentrations of ClOx.
In the simulations of CH3Cl release from dispersed organic matter, and all
mechanisms together, very substantial upper stratospheric O3 depletion appears
where ClOx mixing ratios are between 4 and 6 ppb (figure 5c,d ). Even higher
ClOx concentrations occur in the polar stratosphere, in excess of 9 ppb, however,
less intense solar radiation and polar night reduce the impact of such high values
in the model. For comparison, peak ClOx values of approximately 2 ppb at
approximately 20 km are observed and predicted by chemical–transport models
in the present day springtime Antarctic polar vortex, where complete O3 column
destruction occurs (World Meteorological Organisation 2003). Simulating the
formation of the stratospheric O3 hole is problematic in a two-dimensional
model, requiring a formulation that simulates an isolated air mass and accurately
calculates the distribution of PSCs in regions that are often very close to the
critical conditions for PSC formation (Chipperfield & Pyle 1990). However, given
the extremely high chlorine loading and cold stratospheric temperatures
suggested by our end-Permian stratosphere, it is reasonable to expect that
analogous polar O3 hole events might have occurred.
Substantial depletion of stratospheric O3 occurs rapidly; within 15 years for the
100 kyr simulation of Siberian Traps volcanism and all associated rock heating
(run 4, figure 6a). Given that there is no direct removal process for chlorine in the
stratosphere, O3 recovery following cessation of emissions is governed by the timescale for transport of chlorine from the stratosphere into the troposphere where it
can then be rained out. In the two-dimensional model, hemispheric mean O3
columns rise quasi-exponentially following the termination of volcanic activity and,
after a 20 year period have recovered to around 90% of the pre-eruption value
(figure 6b). Such rapid recovery, on a geological time scale, even from an atmosphere
with severe O3 depletion, is significant because it suggests that the influence via O3
loss of the Siberian Traps volcanism on levels of UV-B radiation reaching the earth’s
surface, is only exerted for the duration of volcanic activity. This suggests that the
‘spikes’ in the abundance of mutated lycopsid spores observed in Greenland
(Visscher et al. 2004; Looy et al. 2005) may relate to discrete pulses of volcanic
activity. There is no evidence in our two-dimensional model sensitivity studies for
persistent or long-term damage to the ozone layer.
Phil. Trans. R. Soc. A (2007)
D. J. Beerling et al.
(a)
(b)
50
4.00
3.75
3.49
3.24
2.99
2.74
2.49
2.23
1.98
1.73
1.48
1.22
0.97
0.72
0.47
0.21
–0.04
altitude (km)
40
30
20
10
ClOx (ppb)
1858
0
(d )
50
>9.0
8.44
7.87
7.30
6.74
6.18
5.61
5.50
4.48
3.91
3.35
2.79
2.22
1.65
1.09
0.52
–0.04
altitude (km)
40
30
20
10
0
–90
–60
–30
0
latitude
30
60
90 –90
–60
–30
0
30
latitude
60
ClOx (ppb)
(c)
90
Figure 5. Changes in the annually averaged ClOx mixing ratio by altitude and latitude for four
scenarios given releases over 100 kyr for (a) the eruption of the Siberian Traps (run 1A), (b) CH3Cl
release from the heating of coals (run 2A), (c) dispersed organic matter (run 3A) and (d ) releases
from all sources combined (run 4A). Note the different scale used for (a,b) and (c,d ).
(c ) Near-surface increases in UV-B radiation
We calculated the resulting increases in UV-B fluxes arising from O3 depletion
for experiments in run 4 (Siberian Traps and CH3Cl release from coals and
dispersed organic matter combined) to determine if they might be sufficient to
destabilize plant genomes based on experimental evidence. To allow direct
comparison between model output and experimental investigations of plant
responses to UV-B radiation, all UV-B fluxes are expressed using a generalized
plant action spectrum normalized at 300 nm, termed BE (Caldwell 1971;
figure 7). Genomic stability is assessed by measuring changes in the frequency of
homologous recombination (HR) of DNA in somatic tissue, a central mechanism
of DNA repair that confers protection against the deleterious effects of UV-B
radiation (Schuermann et al. 2005). HR frequency is assessed by resolving HR
events within individual cells using transgenic plants carrying a visible ‘reporter’
gene that switches from an inactive to an active state whenever a DNA
rearrangement occurs and causes them to stain blue.
According to our results, the 600 kyr experiment leads to the year-round UV-B
fluxes of 30–40 (kJ mK2 dK1)BE throughout a broad region in the tropics and into
the mid-latitudes of the Northern Hemisphere (figure 7). This pattern reflects a
Phil. Trans. R. Soc. A (2007)
1859
Ozone layer stability in the end-Permian
(a) 500
Northern Hemisphere
Southern Hemisphere
400
collapse phase
300
200
recovery phase
100
0
20
time (years)
30
40
90
410.0
395.0
370.0
345.0
320.0
295.0
270.0
245.0
220.0
195.0
170.0
145.0
120.0
95.0
70.0
45.0
20.0
60
latitude
30
0
–30
–60
–90
0
5
10
15
(DU)
(b)
10
20
years after cessation of volcanism
Figure 6. (a) Latitudinal averages: time-series of O3 depletion and recovery for the simulation in
figure 3d, all factors combined and released over 100 kyr (run 4A). (b) Recovery phase: spatial
pattern of O3 recovery after all emissions are switched off. Du, Dobson units.
combination of O3 loss and intensity of incoming solar energy and compares with
10–20 (kJ mK2 dK1)BE for the control Permian atmosphere (figure 7). In the
400 kyr experiment, the tropical region of high year-round UV-B broadens further
to reach 308 S and the values rise to 40–60 (kJ mK2 dK1)BE. The two shorter
simulations have a more dramatic impact resulting from the greater severity of O3
depletion (figure 3). With the 200 kyr simulation, the region of year-round high
UV-B expands considerably exposing large areas of the low- and mid-latitudes
of the Northern and Southern Hemispheres to 50–70 (kJ mK2 dK1)BE, while
the 100 kyr simulation raises the intensity of UV-B fluxes further to in excess of
100 (kJ mK2 dK1)BE.
In summary, UV-B fluxes ranged from 30 to 60 (kJ mK2 dK1)BE throughout
the growing season for the 600–400 kyr simulations and 50–100 (kJ mK2 dK1)BE
for the 100–200 kyr simulations. Current experimental evidence suggests that
these ranges are sufficient to trigger instabilities in plant genomes, especially if
sustained over tens of thousands of years. Experimental studies, for example,
indicate that HR frequency increased substantially in Arabidopsis thaliana and
Nicotiana tabacum when the daily UV-B dose rose from a control ‘background’
Phil. Trans. R. Soc. A (2007)
1860
D. J. Beerling et al.
value from 2.3 (kJ mK2 dK1)BE to 6.6 (kJ mK2 dK1)BE (Ries et al. 2000). Further
experiments increasing the daily flux up to 27.1 (kJ mK2 dK1)BE produced much
higher HR frequencies—14 times in Arabidopsis and 7 times in tobacco.
Furthermore, the genomic instability of Arabidopsis increased with each
generation; second generation progeny plants had HR rates two to five times
higher than those of their parents when exposed to the same UV-B dosage.
This suggests that UV-B fluxes induce mutations in reproductive (germ) cells
(i.e. was heritable), and highlights the possibility of long-term cumulative effects
of exposure to moderately high UV radiation.
It is also possible that rising global temperatures of approximately 3 K due to
CO2 degassing from the Siberian Traps (Grard et al. 2005) exerted an effect on
plant mutation rates. If correct, this would implicate temperature as another
possible mutagenic factor alongside UV-B. However, experiments with one of the
same transgenic lines of Arabidopsis used by Ries et al. (2000) fails to support this
possibility, with only a 1.4-fold increase in HR events in response to a 118C
temperature increase from 22 to 338C (Boyko et al. 2005). Collectively, these data
indicate that simulated increases in UV-B radiation (figure 7) might destabilize
plant genomes to a significantly greater extent than any increases in temperature.
(d ) Sulphur dioxide and mutagenesis
In addition to UV-B and temperature, there is also a possible mutagenic
influence of SO2. The evidence for this possibility is circumstantial, deriving from
the occurrence of mutated gymnosperm pollen grains in highly industrialized
locations in the former Soviet Union (Micieta & Murı́n 1996; Tretyakova &
Noskova 2004). The mutated grains show morphological similarities to those from
the P-T boundary. So, we have modelled tropospheric SO2 concentrations
resulting from large releases of SO2 (table 1) to determine how these levels
compare with those in industrial regions. Our calculations indicate SO2
concentrations were generally less than 3 ppb throughout the Northern Hemisphere and considerably lower (by a factor 10–10 000) in the Southern
Hemisphere. The only exception to this pattern was the high northern latitudes
where SO2 concentrations reached 13 ppb (figure 8). The marked latitudinal
gradient reflects the short residence time of SO2 in the atmosphere owing to its
rapid oxidation to H2SO4 and its subsequent washing out in precipitation. These
concentrations are relatively low and comparable to the range typical for rural
areas (1–30 ppb). In moderately polluted areas, SO2 concentrations reach
30–200 ppb and highly polluted regions experience reach values even higher
200–300 ppb ( Finalyson-Pitts & Pitts 1986). Moreover, even if the SO2
concentration was sufficiently high in the Northern Hemisphere, the strong
latitudinal gradient it exhibits (figure 8) means it is unable to account for the global
distribution of mutated palynomorphs (Visscher et al. 2004; Foster & Afonin 2005).
(e ) Modelling caveats
We have not included the effects of massive H2S and CH4 release in our
modelling simulations, events interpreted from various lines of palaeo-evidence
to have taken place during the end-Permian (Berner 2002; Kump et al. 2005).
Modelling studies suggest that large-scale emission of H2S (Kump et al. 2005) or
CH4 (Lamarque et al. 2006) at the end-Permian had the potential to cause
Phil. Trans. R. Soc. A (2007)
Ozone layer stability in the end-Permian
1861
significant O3 depletion. Therefore, incorporating elevated CH4 and H2S into an
investigation of the effects of the Siberian Traps is important to build an overall
picture of atmospheric response to the eruptions. Enhanced CH4 concentrations
cause depressed OH concentrations in the troposphere, but generate H2O and
consequently elevate HOx radical concentrations in the stratosphere. Since HOx is
photochemically important for both catalytic O3 destruction cycles and oxidation
of SO2 through to H2SO4, gas-phase chemistry would be influenced, as well as
heterogeneous chemistry occurring on expanded sulphate aerosol surface area.
Significantly enhanced levels of CH4 would increase the tropospheric lifetimes of
organohalogen species such as CH3Cl, the main sink for which is reaction with OH,
consequently increasing the halogen loading of stratosphere resulting from the
Siberian Traps volcanism. Elevated H2S concentrations give rise to substantially
elevated SO2 and sulphate aerosol concentrations. Since both SO2 and sulphate
aerosol have the potential to absorb UV, high levels of these species would
attenuate the amount of UV penetrating through the atmosphere to the surface, so
limiting the radiative impact of an end-Permian O3 layer collapse.
5. Conclusions
From our analyses, we conclude that (i) a 100–200 kyr main eruptive phase had
the potential to produce high latitude stratospheric O3 depletion due to large
releases of HCl (figure 1), (ii) organohalogen release from heating of organic-rich
rocks was insufficient to extensively damage the ozone layer. However, including
the possibility that dispersed organic matter provided a precursor for
organohalogen synthesis results in massive CH3Cl production that could lead
to substantial O3 depletion when released over 100 kyr (figure 2), (iii) the
potential for severe O3 depletion is greatest when the simultaneous release of HCl
from the Siberian Traps and CH3Cl release from coals and rocks are considered
together, and (iv) calculated increases in UV-B radiation resulting from this
latter scenario (figure 7) appear sufficient in destabilize plant genomes.
Finally, we point out that increases in UV-B radiation we report for the same
combined scenario (figure 7) here might have left a distinct geochemical
signature in the sedimentary record by influencing the isotopic fractionation
taking place during the biogeochemical cycling of sulphur. This may offer an
independent means for evaluating our simulations and the interpretation of a
mutagenesis event in the fossil record. Mass-independent isotope fractionation
(MIF) of sulphur has been observed in laboratory experiments subjecting SO2 to
UV radiation of less than 220 nm (Farquhar et al. 2001). This important
observation has been exploited by analysing crustal and sedimentary rocks for a
MIF signal, where its presence represents UV-transparency of the Earth’s early
(Archean) atmosphere with reduced O2 and O3 levels (Farquhar et al. 2000).
Furthermore, studies of ice from snow pits and ice cores indicate that a similar
MIF signal is recorded following major volcanic eruptions that inject SO2 into
the stratosphere, where it is exposed to high energy UV radiation and photooxidized, mainly as a result of UV photochemistry in the 248 nm wavelength
region rather than 220 nm (Savarino et al. 2003; Baroni et al. 2007). The MIF
anomaly may occur because the absorption of 248 nm produces SO2* (excited
triplet and singlet states) and eventually SO3 and H2SO4. This suggests that if
Phil. Trans. R. Soc. A (2007)
1862
D. J. Beerling et al.
90
>100.0
93.8
87.5
81.2
75.0
68.8
62.5
56.2
50.0
43.8
37.5
31.2
25.0
18.8
12.5
6.2
0.0
laitude
60
30
0
–30
–60
–90
0
60
(b)
120 180 240
day of year
(c)
300
near-surface UV flux (kj m –2 d –1)BE
(a)
360
90
60
laitude
30
0
–30
–60
–90
(d)
(e)
90
60
laitude
30
0
–30
–60
–90
0
60
120 180
day of year
240
300
360
0
60
120 180
day of year
240
300
360
Figure 7. Calculated latitudinal and seasonal daily near-surface UV-B radiation fluxes using a
generalized plant action spectrum normalized at 300 nm (kJ mK2 dK1)BE (Caldwell 1971) for the
(a) end-Permian ‘control’ atmosphere and for (b) 600 kyr, (c) 400 kyr, (d ) 200 kyr and (e) 100 kyr
simulations of releases from all sources combined (run 4).
Phil. Trans. R. Soc. A (2007)
Ozone layer stability in the end-Permian
(b)
90
13.500
6.186
2.835
1.299
0.595
0.273
0.125
0.057
0.026
6.0x10–3
3.1x10–3
1.5x10–3
7.8x10– 4
4.0x10– 4
2.0x10– 4
1.0x10––54
5.1x10
60
laitude
30
0
–30
–60
–90
0
60
120 180 240 300 360
day of year
0
60
tropospheric SO2 (ppb)
(a)
1863
120 180 240 300 360
day of year
Figure 8. Calculated latitudinal and seasonal tropospheric SO2 concentrations for (a) the shortest
(100 kyr, run 4A) and (b) the longest (600 kyr, run 40) duration of the main eruptive phase.
the Earth experienced a major UV crisis at the end-Permian during the eruption
of the Siberian Traps, it may be recorded as a MIF preserved in the rock record.
However, the magnitude of the signal, its preservation and our capacity to
actually detect it, remain open questions given the high solubility and relatively
small quantities of sulphate involved.
We thank Mark Sephton (Imperial College), University of London for advice on calculating
organohalogen emissions from coals, Henk Visscher for helpful comments on an earlier draft, and
Karl Turekian ( Yale University) for pointing out the possible consequences of increased UV-B
radiation for mass-independent sulphur isotope fractionation. M.H. was supported by an NERC
studentship (NER/S/J/2003/11963) and B.H.L. by Natural Environment Research Council award
to D.J.B. and J.A.P. (NER/A/S/2002/00865) and a Leverhulme Trust Early Career Fellowship
(ECF/2006/0492).
References
Akulov, N. I. 2006 Nodules in Carbonaceous sediments of the Southern Tunguska Basin. Lithol.
Min. Res. 41, 73–84. (doi:10.1134/S0024490206010068)
Baroni, M., Thiemens, M. H., Delmas, R. J. & Savarino, J. 2007 Mass-independent sulfur isotopic
compositions in stratospheric volcanic eruptions. Science 315, 84–87. (doi:10.1126/science.
1131754)
Bekki, S. & Pyle, J. A. 1992 Two-dimensional assessment of the impact of aircraft sulphur
emissions on the stratospheric sulphate aerosol layer. J. Geophys. Res. 97, 15 839–15 847.
Bekki, S., Law, K. S. & Pyle, J. A. 1994 Effect of ozone depletion on atmospheric CH4 and CO
concentrations. Nature 371, 595–597. (doi:10.1038/371595a0)
Benton, M. J. & Twitchett, R. J. 2003 How to kill (almost) all life: the end-Permian extinction
event. Trends Ecol. Evol. 18, 358–365. (doi:10.1016/S0169-5347(03)00093-4)
Berner, R. A. 2002 Examination of hypotheses for the Permo-Triassic boundary extinction by
carbon cycle modelling. Proc. Natl Acad. Sci. USA 99, 4172–4177. (doi:10.1073/pnas.032095199)
Berner, R. A. 2005 The carbon and sulfur cycles and atmospheric oxygen from middle Permian to
middle Triassic. Geochim. Cosmochim. Acta 69, 3211–3217. (doi:10.1016/j.gca.2005.03.021)
Berner, R. A. & Ward, P. D. 2006 Positive reinforcement, H2S, and the Permo-Triassic extinction:
comment and reply. Geology, e100–e101.
Boyko, A., Filkowski, J. & Kovalchuk, I. 2005 Homologous recombination in plants is temperature
and daylength dependent. Mut. Res. 572, 73–83.
Phil. Trans. R. Soc. A (2007)
1864
D. J. Beerling et al.
Bragg, L. J., Finkelman, R. B. & Tewalt, S. J. 1991 Distribution of chlorine in United States coal.
Coal Sci. Technol. 17, 3–10.
Caldwell, M. M. 1971 Solar ultraviolet and the growth and development of higher plants. In
Photophysiology, pp. 131–177. New York, NY: Academic Press.
Campbell, I. H., Czamanske, G. K., Fedorenko, V. A., Hill, R. I. & Stepanov, V. 1992 Synchronism
of the Siberian Traps and the Permian–Triassic boundary. Science 258, 1760–1763. (doi:10.
1126/science.258.5089.1760)
Chapman, S. 1930 A theory of upper-atmosphere ozone. Mem. R. Meteorol. Soc. 3, 103–125.
Chin, M. et al. 2002 Tropospheric aerosol optical thickness from the GOCART model and
comparisons with satellite and sun photometer measurements. J. Atmos. Sci. 59, 461–483.
(doi:10.1175/1520-0469(2002)059!0461:TAOTFTO2.0.CO;2)
Chipperfield, M. P. & Pyle, J. A. P. 1990 Two-dimensional modelling of the Northern Hemisphere
high latitude lower stratosphere. J. Geophys. Res. 95, 11 865–11 874.
Courtillot, V., Jaupert, C., Manighetti, I., Tapponier, P. & Besse, J. 1999 On the causal links
between flood basalts and continental break-up. Earth Planet. Sci. Lett. 166, 177–195. (doi:10.
1016/S0012-821X(98)00282-9)
Czamanske, G. K., Gurevitch, A. B., Fedorenko, V. & Simonov, O. 1998 Demise of the Siberian
plume: paleogeographic and paleotectonic reconstruction from the prevolcanic and volcanic
record, north-central Siberia. Int. Geol. Rev. 41, 95–115.
Erwin, D. H. 1994 The Permo-Triassic extinction. Nature 367, 231–236. (doi:10.1038/367231a0)
Farquhar, J., Bao, H. M. & Thiemens, M. 2000 Atmospheric influence of Earth’s earliest sulfur
cycle. Science 289, 756–758. (doi:10.1126/science.289.5480.756)
Farquhar, J., Savarino, J., Airieau, S. & Thiemens, M. H. 2001 Observations of wavelengthsensitive mass-independent sulfur isotope effects during SO2 photolysis: implications for the
early atmosphere. J. Geophys. Res. 106, 32 829–32 839. (doi:10.1029/2000JE001437)
Finalyson-Pitts, B. J. & Pitts, J. N. 1986 Atmospheric chemistry: fundamentals and experimental
techniques. New York, NY: Wiley.
Foster, C. B. & Afonin, S. A. 2005 Abnormal pollen grains: an outcome of deteriorating
atmospheric conditions around the Permian–Triassic boundary. J. Geol. Soc. 162, 653–659.
(doi:10.1144/0016-764904-047)
Gerlach, T. M. & Graeber, E. J. 1985 Volatile budget of Kilauea volcano. Nature 313, 273–277.
(doi:10.1038/313273a0)
Giannakopoulos, C. 1998 Modelling the impact of physical and removal processes on tropospheric
chemistry. PhD thesis, University of Cambridge.
Goodarzi, F. 1987 Comparison of elemental distribution in fresh and weathered samples of selected
coals in the Jurassic-Cretaceous Kootenay group, British Columbia, Canada. Chem. Geol. 63,
21–28. (doi:10.1016/0009-2541(87)90070-2)
Grard, A., François, L. M., Dessert, C., Dupré, B. & Goddéris, Y. 2005 Basaltic volcanism and
mass extinction at the Permo-Triassic boundary: environmental impact and modeling of the
global carbon cycle. Earth Planet. Sci. Lett. 234, 207–221. (doi:10.1016/j.epsl.2005.02.027)
Greive, D. A. & Goodarzi, F. 1993 Trace elements in coal samples from active mines in the
Foreland Belt, British Columbia, Canada. Int. J. Coal. Geol. 24, 259–280. (doi:10.1016/01665162(93)90014-2)
Haag, M. & Heller, F. 1991 Late Permian to early Triassic magnetostratigraphy. Earth Planet. Sci.
Lett. 107, 42–54. (doi:10.1016/0012-821X(91)90042-G)
Haigh, J. D. 1984 Radiative heating in the lower stratosphere and the distribution of ozone in a
two-dimensional model. Q. J. R. Met. Soc. 110, 167–186. (doi:10.1002/qj.49711046312)
Hanson, D. & Mauersberger, K. 1988 Laboratory studies of the nitric acid trihydrate: implications
for the south polar stratosphere. Geophys. Res. Lett. 15, 855–858.
Harfoot, M. B. J., Beerling, D. J., Lomax, B. H. & Pyle, J. A. 2007 A two-dimensional atmospheric
chemistry modelling investigation of Earth’s Phanerozoic O3 and near-surface ultraviolet
radiation history. J. Geophys. Res. 112, DO7308. (doi:10.1029/2006JD007372)
Phil. Trans. R. Soc. A (2007)
Ozone layer stability in the end-Permian
1865
Harwood, R. S. & Pyle, J. A. 1975 A two-dimensional mean circulation model for the atmosphere
below 80 km. Q. J. R. Met. Soc. 101, 723–747. (doi:10.1002/qj.49710143003)
Harwood, R. S. & Pyle, J. A. 1980 The dynamical behaviour of a two-dimensional model of the
stratosphere. Q. J. R. Met. Soc. 106, 395–420. (doi:10.1002/qj.49710644903)
Kamo, S. L., Czamanske, G. K., Amelin, Y., Fedorenko, V. A., Davis, D. W. & Trofimov, V. R.
2003 Rapid eruption of Siberian flood-volcanic rocks and evidence for coincidence with the
Permian–Triassic boundary and mass extinction at 251 Ma. Earth Planet. Sci. Lett. 214, 75–91.
(doi:10.1016/S0012-821X(03)00347-9)
Khain, V. E. 1985 Geology of the USSR, Beiträge zur Regionalen Geologie der Erde. BerlinStuttgart, Germany: Gebrüder Bornträger.
Kiehl, J. T. & Shields, C. A. 2005 Climate simulation of the latest Permian: implications for mass
extinction. Geology 33, 757–760. (doi:10.1130/G21654.1)
Kontorovich, A. E., Khomenko, A. V., Burshtein, L. M., Likhanov, I. I., Pavlov, A. L.,
Staroseltsev, V. S. & Ten, A. A. 1997 Intense basic magmatism in the Tunguska petroleum
basin, eastern Siberia, Russia. Petrol. Geosci. 3, 359–369.
Koti, S., Reddy, K. R., Reddy, V. R., Kakani, V. G. & Zhao, D. L. 2005 Interactive effects of
carbon dioxide, temperature, and ultraviolet-B radiation on soybean (Glycine max L.) flower
and pollen morphology, pollen production, germination, and tube lengths. J. Exp. Bot. 56,
725–736. (doi:10.1093/jxb/eri044)
Kreidenweis, S., Tyndall, G., Barth, M., Dentener, F., Lelieveld, J. & Mozurkewich, M. 1999
Aerosols and clouds. In Atmospheric chemistry and global change (eds G. P. Brasseur, J. J.
Orlando & G. S. Tyndall), pp. 117–158. Oxford, UK: Oxford University Press.
Kump, L. R., Pavlov, A. & Arthur, M. A. 2005 Massive release of hydrogen sulfide to the surface ocean
and atmosphere during intervals of oceanic anoxia. Geology 33, 397–400. (doi:10.1130/G21295.1)
Lamarque, J.-F., Kiehl, J. T., Shields, C. A., Boville, B. A. & Kinnison, D. E. 2006 Modeling the
response to changes in tropospheric methane concentration: application to the Permian–
Triassic boundary. Paleoceanography 21, PA3006. (doi:10.1029/2006PA001276)
Law, K. S. & Pyle, J. A. 1993 Modeling trace gas budgets in the troposphere 1. Ozone and odd
nitrogen. J. Geophys. Res. 98, 18 377–18 400.
Looy, C. V., Collinson, M. E., Van Konijnenburg-Van Cittert, J. H. A., Visscher, H & Brain,
A. P. R. 2005 The ultrastructure and botanical affinity of end-Permian spore tetrads. Int.
J. Plant Sci. 166, 875–887. (doi:10.1086/431802)
Medina, G., Tabares, J. A., Pérez Alcázar, G. A. & Barraza, J. M. 2006 A methodology to evaluate
coal ash content using Siderite Mössbauer spectral area. Fuel 85, 871–873. (doi:10.1016/j.fuel.
2005.08.034)
Melnikov, N. V., Khomenko, A. V., Kuznetsova, E. N. & Zhidkova, L. V. 1997 The effect of traps
on salt redistribution in the Lower Cambrian of the western Siberian Platform Russian. Geol.
Geophys. 38, 1378–1384.
Mičieta, K. & Murı́n, G. 1996 Microspore analysis for genotoxicity of a polluted environment.
Environ. Exp. Bot. 36, 21–27. (doi:10.1016/0098-8472(95)00050-X)
Millard, G. A., Mather, T. A., Pyle, D. M., Rose, W. I. & Thornton, B. 2006 Halogen emissions from
a small volcanic eruption: modeling the peak concentrations, dispersion, and volcanically induced
ozone loss in the stratosphere. Geophys. Res. Lett. 33, L19815. (doi:10.1029/2006GL026959)
Molina, L. T. & Molina, M. J. 1987 Production of Cl2O2 from the self-reaction of the ClO radical.
J. Phys. Chem. 91, 433–436. (doi:10.1021/j100286a035)
Reichow, M. K., Saunders, A. D., White, R. V., Pringle, M. S., Al’Mukhamedov, A. I., Medvedev,
A. I. & Kirda, N. P. 2002 Ar40/Ar39 dates from the West Siberian Basin: Siberian flood basalt
province doubled. Science 296, 1846–1849. (doi:10.1126/science.1071671)
Renne, P. R., Zhang, Z. C., Richards, M. A., Black, M. T. & Basu, A. R. 1995 Synchrony and
causal relations between Permian–Triassic Boundary crises and Siberian flood volcanism.
Science 269, 1413–1416. (doi:10.1126/science.269.5229.1413)
Ries, G., Heller, W., Puchta, H., Sandermann, H., Seidlitz, H. K. & Hohn, B. 2000 Elevated UV-B
radiation reduces genome stability in plants. Nature 406, 98–101. (doi:10.1038/35017595)
Phil. Trans. R. Soc. A (2007)
1866
D. J. Beerling et al.
Rose, W. I. et al. 2006 Atmospheric chemistry of a 33–34 hour old volcanic cloud from Hekla
Volcano (Iceland): insights from direct sampling and the application of chemical box modelling.
J. Geophys. Res. 111, D20206. (doi:10.1029/2005JD006872)
Savarino, J., Romero, A., Cole-Dao, J., Bekki, S. & Thiemens, M. H. 2003 UV induced massindependent sulfur isotope fractionation in stratospheric volcanic sulfate. Geophy. Res. Lett. 30,
2131. (doi:10.1029/2003GL018134)
Schmidt, G. A. & Shindell, D. T. 2003 Atmospheric composition, radiative forcing, and climate
change as a consequence of a massive methane release from gas hydrates. Paleoceanography 18,
1004. (doi:10.1029/2002PA000757)
Schuermann, D., Molinier, J., Fritsch, O. & Hohn, B. 2005 The dual nature of homologous
recombination in plants. Trends Genet. 21, 172–181. (doi:10.1016/j.tig.2005.01.002)
Schwandner, F. M., Seward, T. M., Gize, A. P., Hall, A. & Dietrich, V. J. 2004 Diffuse emission of
organic trace gases from the flank and crater of a quiescent active volcano (Vulcano, Aeolian
Islands, Italy). J. Geophys. Res. 109, D04301. (doi:10.1029/2003JD003890)
Solomon, P., Connon, B., Barrett, J., Mooney, T., Lee, A. & Parrish, A. 2002 Measurements of
stratospheric ClO over Antarctica in 1996–2000 and implications for ClO dimer chemistry.
Geophys. Res. Lett. 29, 1708–1711. (doi:10.1029/2002GL015232)
Suggate, R. P. & Dickinson, W. W. 2004 Carbon NMR of coals: the effects of coal type and rank.
Int. J. Coal Geol. 57, 1–22. (doi:10.1016/S0166-5162(03)00116-2)
Textor, C., Graf, H.-F. & Herzog, M. 2003 Injection of gases into the stratosphere by explosive
volcanic eruptions. J. Geophys. Res. 108, 4606–4624. (doi:10.1029/2002JD002987)
Thordarson, T. & Self, S. 1996 Sulfur, chlorine and fluorine degassing and atmospheric loading by
the Roza eruption, Columbia River Basalt Group, Washington, USA. J. Volcanic Geotherm.
Res. 74, 49–73. (doi:10.1016/S0377-0273(96)00054-6)
Tretyakova, I. N. & Noskova, N. E. 2004 Scotch pine pollen under conditions of environmental
stress. Russian J. Ecol. 35, 20–26. (doi:10.1023/B:RUSE.0000011105.90297.07)
Vassilev, S. V., Eskenazy, G. M. & Vassileva, C. G. 2000 Contents, modes of occurrence and origin
of chlorine and bromine in coal. Fuel 79, 903–921. (doi:10.1016/S0016-2361(99)00236-7)
Visscher, H., Looy, C. V., Collinson, M. E., Brinkhuis, H., Van Cittert, J. H. A., Kurschner, W. M.
& Sephton, M. A. 2004 Environmental mutagenesis during the end-Permian ecological crisis.
Proc. Natl Acad. Sci. USA 101, 12 952–12 956. (doi:10.1073/pnas.0404472101)
Wignall, P. B. 2001 Large igneous provinces and mass extinctions. Earth Sci. Rev. 53, 1–33.
(doi:10.1016/S0012-8252(00)00037-4)
World Meteorological Organisation (WMO) 2003 Scientific Assessment of Ozone Depletion: 2002,
Global Ozone Research and Monitoring Project-Report No. 47, Geneva.
Phil. Trans. R. Soc. A (2007)