ArticlePDF Available

Phanerozoic atmospheric oxygen concentrations reconstructed from sedimentary charcoal

Springer Nature
Nature Geoscience
Authors:

Abstract and Figures

Variations of the Earth's atmospheric oxygen concentration (pO2) are thought to be closely tied to the evolution of life, with strong feedbacks between uni- and multicellular life and oxygen. On the geologic timescale, pO2 is regulated by the burial of organic carbon and sulphur, as well as by weathering. Reconstructions of atmospheric O2 for the past 400million years have therefore been based on geochemical models of carbon and sulphur cycling. However, these reconstructions vary widely, particularly for the Mesozoic and early Cenozoic eras. Here we show that the abundance of charcoal in mire settings is controlled by pO2, and use this proxy to reconstruct the concentration of atmospheric oxygen for the past 400million years. We estimate that pO2 was continuously above 26% during the Carboniferous and Permian periods, and that it declined abruptly around the time of the Permian-Triassic mass extinction. During the Triassic and Jurassic periods, pO2 fluctuated cyclically, with amplitudes up to 10% and a frequency of 20-30million years. Atmospheric oxygen concentrations have declined steadily from the middle of the Cretaceous period to present-day values of about 21%. We conclude, however, that variation in pO2 was not the main driver of the loss of faunal diversity during the Permo-Triassic and Triassic-Jurassic mass extinction events.
Content may be subject to copyright.
LETTERS
PUBLISHED ONLINE: 1 AUGUST 2010 | DOI: 10.1038/NGEO923
Phanerozoic concentrations of atmospheric
oxygen reconstructed from sedimentary charcoal
Ian J. Glasspool
1
*
and Andrew C. Scott
2
Variations of the Earth’s atmospheric oxygen concentration
(pO
2
) are thought to be closely tied to the evolution of
life, with strong feedbacks between uni- and multicellular life
and oxygen
1,2
. On the geologic timescale, pO
2
is regulated
by the burial of organic carbon and sulphur, as well as by
weathering
3
. Reconstructions of atmospheric O
2
for the past
400 million years have therefore been based on geochemical
models of carbon and sulphur cycling
4–6
. However, these
reconstructions vary widely
4–10
, particularly for the Mesozoic
and early Cenozoic eras. Here we show that the abundance
of charcoal in mire settings is controlled by pO
2
, and use
this proxy to reconstruct the concentration of atmospheric
oxygen for the past 400 million years. We estimate that pO
2
was continuously above 26% during the Carboniferous and
Permian periods, and that it declined abruptly around the time
of the Permian–Triassic mass extinction. During the Triassic
and Jurassic periods, pO
2
fluctuated cyclically, with amplitudes
up to 10% and a frequency of 20–30 million years. Atmospheric
oxygen concentrations have declined steadily from the middle
of the Cretaceous period to present-day values of about 21%.
We conclude, however, that variation in pO
2
was not the main
driver of the loss of faunal diversity during the Permo–Triassic
and Triassic–Jurassic mass extinction events.
Fire is an exothermic oxidation reaction dependent on the rapid
combination of fuel and oxygen in the presence of heat
11
. Charcoal
is a by-product of wildfire and is first documented in the latest
Silurian
12,13
, and has subsequently been recorded in all geological
periods from a range of sedimentary settings
14
. Calculation of fuel
flammability at varying oxygen concentrations enables past pO
2
to
be constrained within the range 15–35% (‘fire window’) whenever
charcoal is recovered from the fossil record
11,15,16
. To extrapolate
further and to predict a pO
2
curve from the geological abundance
of charcoal, both sources of heat and fuel must be decoupled
from this relationship.
Lightning is the pre-eminent source of heat for the ignition
of fossil wildfires
17
. Owing to its global frequency (44 ± 5
strikes/second
18
) and proven ‘fossil’ record as fulgurites
19
, it is
unlikely that a lack of lightning could ever have limited fire
initiation. However, through deep time, fuel availability will
have affected the frequency of ignition. All terrestrial vegetation
is potential fuel. As recorded by the occurrence of charcoal
throughout deep time, a source of fuel has been present for at least
419 million years
12
. However, in the fossil record the distribution
of biomass has varied both spatially and temporally. Peat-forming
environments are by definition regions of biomass accumulation. In
this environment, an absence of fire ignition cannot be attributed to
an absence of vegetation.
However, the presence of vegetation need not equate to the
presence of fuel, as vegetation may not be combustible under the
1
Department of Geology, Field Museum of Natural History, 1400 S. Lake Shore Drive, Chicago, Illinois 60605, USA,
2
Department of Earth Sciences, Royal
Holloway University of London, Egham TW20 0EX, UK. *e-mail: iglasspool@fieldmuseum.org.
prevailing environmental conditions. Vegetation is heterogeneous
in composition, where the most important heterogeneity affecting
potential flammability is moisture content
20
. For ignition, fuel must
be heated sufficiently to drive-off moisture and to liberate volatiles
that can be oxidized to generate a self-supporting exothermic
pyrolytic reaction (that is, fire). The greater the moisture content the
less flammable fuel becomes
2,20
. Peat-forming environments were
selected for this study as their formation requires that ‘groundwater
must remain throughout the whole year, above or close to the
ground surface’
21
. Therefore, these are high-moisture environments
that best minimize the potential role of climatic variation on
the accumulation of charcoal deposits through deep time. This
approach is confirmed by data from a range of modern peats
representing divergent ecological settings and vegetation types
(Supplementary Table S1). In these peats (modern–Pleistocene, 21
seams/81 samples), which include environments such as the Florida
Everglades and Kalimantan, where modern fires are frequently doc-
umented, charcoal abundance is consistently low with no sample
containing more than 11.1% (average 4.3%). Even extrapolating to
the past 50 million years, average charcoal abundance is still low
compared with earlier deposits. Of 139 seams (759 samples) from
this interval, only one contains >20% charcoal (26.5%: average
4.8%), 13 contain none and 23 contain <1.0%. Throughout this
interval, global climate and weather have fluctuated markedly,
whereas charcoal abundance demonstrably has not. Therefore,
it can be argued that, by examining charcoal abundance in
peat-forming environments, the impact of moisture content on fuel
ignition, although not eliminated, is greatly restricted.
In the fossil record, peats are preserved as lignites or coals
21
and
because of their economic importance are routinely studied by op-
tical reflectance microscopy. Using this technique, the organic con-
stituents are described in terms of macerals
21
. One maceral group
(inertinite) is almost exclusively considered the by-product of wild-
fires and is synonymous with charcoal
22
. The percentage (by vol-
ume) of inertinite is commonly reported and provides an extensive
record of charcoal abundance
23
. A database of percentage inertinite
abundance in coal (Inert%), calculated on a mineral-matter-free
basis of all ages and geographic distribution, has been compiled
from the literature (Fig. 1a, Supplementary Table S2). Data are
binned to 10-million-year intervals (Supplementary Table S3) to
match existing geochemical models for pO
2
(refs 5–7). The concept
of comparing inertinite with pO
2
is not novel, plots of fusinite and
semifusinite, two inertinite group macerals, having been made
24
.
However, this study did not include data on other, often abundant,
inertinite macerals and made no attempt to predict pO
2
.
To calculate pO
2
from Inert%, the generation of a calibration
curve is required (Fig. 1b). From experiments, wildfire is unsustain-
able at levels of pO
2
15% (ref. 16), and therefore charcoal will not
be generated (Inert% = 0%). However, as zero values preclude the
NATURE GEOSCIENCE | VOL 3 | SEPTEMBER 2010 | www.nature.com/naturegeoscience 627
© 2010 Macmillan Publishers Limited. All rights reserved.
LETTERS
NATURE GEOSCIENCE DOI: 10.1038/NGEO923
d
Inertinite %
a
c
4
5
6
7
8
9
10
(1)
(2)
(5)
(3)
(4)
b
1/1
5/8
3/26
7/9
3/10
8/18
383/1,196
30/36
27/135
64/144
118/547
47/91
45/87
25/85
1/1
31/31
22/48
5/5
25/29
9/35
12/12
41/90
34/91
52/244
9/47
4/68
17/74
1/4
9/77
70/246
18/53
12/53
62/216
19/20
17/220
42/280
31/133
19/85
300 200 100
Age (Myr)
400 0
100
50
0
30
25
20
pO
2
(%)
35
15
10 20 30 605040
Inert%
070
30
20
10
pO
2
(%)
40
0
300 200 100
Age (Myr)
400 0
300 200 100
Age (Myr)
400 0
30
20
10
pO
2
(%)
40
0
Figure 1 | Phanerozoic inertinite distribution and predictions of pO
2
. a, Inertinite abundance. Line; bin mean. Error bars 1 s.d. from mean. Lower axis;
seams/samples per bin. b, Power-law regressions for conversion of Inert% to pO
2
. Labels refer to equations in main text. c, Biogeochemical predictions of
Phanerozoic pO
2
published to 2009 (refs 4–10). Lines: publication references in key. d, Prediction of pO
2
from Inert%. Line; best estimate based on late
Palaeozoic pO
2
maxima of 30%. Error bars 1 s.d. from mean. Shaded area; estimate of maximum error assuming Phanerozoic pO
2
maxima of 35% + 1 s.d.
(upper margin) and 25% 1 s.d. (lower margin).
application of most nonlinear trendlines, 0.2% inertinite is adopted,
based on typical coal petrographic counts of 500 points, for
maceral analyses
21
(that is one charcoal particle = 0.2%). Modern
pO
2
is 21% and, as discussed above, based on a range of ecolog-
ically, climatically and geographically differing peats, is calculated
to equate to 4.3% inertinite. These two points can be assigned
with some confidence. For calibration, Inert% data at the upper
limit of Phanerozoic pO
2
are required, but prediction of values
at above modern levels must be subjective. However, combustion
research suggests theoretical upper levels of pO
2
from 25% (ref. 2)
to 35% (refs 2,11). Previous research indicates that in the late
Palaeozoic pO
2
exceeded 25% (ref. 15), but owing to increased
plant flammability was less than 35% (refs 2,11). Inert% predictions
use the mean of these values for the late Palaeozoic (pO
2
= 30%).
Although a significant assumption, this value is in accord with other
predictions of the theoretical Phanerozoic maximum
25
.
Independent of fire, mass balance, biogeochemical and carbon
isotopic fractionation models predict maximal Phanerozoic pO
2
628 NATURE GEOSCIENCE | VOL 3 | SEPTEMBER 2010 | www.nature.com/naturegeoscience
© 2010 Macmillan Publishers Limited. All rights reserved.
NATURE GEOSCIENCE DOI: 10.1038/NGEO923
LETTERS
during the Permian at 30–35% (refs 5,7,26; Fig. 1c). The timing
of these maximal pO
2
data corresponds with the timing of maximal
inertinite abundance (that is Inert% = 44.4% (n = 547) in the
Early Permian (280 Myr)). On this basis we suggest an Inert% of
44.4% can be correlated to 30% pO
2
(derived from theoretical
calculations
25
). This datum is used as the uppermost value necessary
to enable calibration of inertinite abundance against pO
2
.
From these estimations, calibration of pO
2
from Inert% is based
on: 15% pO
2
= 0.2% inertinite, 21% pO
2
= 4.3% inertinite and
30% pO
2
= 44.4% inertinite. From these data, fire and pO
2
are
best described by a power relationship, this regression yielding
the highest R
2
values (30% pO
2
curve: 0.991). However, by using
nonlinear regression a diminished ability to predict variations in
pO
2
above 21% exists. The equation for the calibration curve is:
pO
2
= 18.113x(Inert%
0.1273
) or y = c
(x
b) (1)
where:
c = EXP(INDEX(LINEST(LN(y),LN(x),,),1,2))
b = INDEX(LINEST(LN(y),LN(x),,),1)
The error in the mean inertinite data is plotted to one standard
deviation (s.d.) (Fig. 1a). This error is fed-back into the calibration
to generate two further error curves for the 30% pO
2
assessment
(Fig. 1b, Supplementary Fig. S1).
pO
2
= 19.293x(Inert%
0.1376
) or y = c
(x
b) (2)
pO
2
= 17.723x(Inert%
0.1185
) or y = c
(x
b) (3)
Error bars in Fig. 1d (Supplementary Table S4) represent the
impact of these curves on pO
2
estimation. In addition, to provide
a crude assessment of the maximum variation in pO
2
predictions,
assuming late Palaeozoic pO
2
highs had fallen in the range 25% or
35%, two further calibration curves are plotted (Fig. 1b, shading in
Fig. 1d). The first is based on assumptions of a late Palaeozoic pO
2
high of 25% calibrated against Inert% 1 s.d.
pO
2
= 18.633x(Inert%
0.0985
) or y = c
(x
b) (4)
where: late Palaeozoic pO
2
= 25%.
The second is based on a late Palaeozoic pO
2
high of 35%
calibrated against Inert% + 1 s.d.
pO
2
= 17.998x(Inert%
0.1427
) or y = c
(x
b)
(5)
where: late Palaeozoic pO
2
= 35%.
As with previous predictions
5,9
(Fig. 1c), our data indicate a
marked rise in pO
2
within a 30-million-year interval following the
growth of the Middle–Late Devonian forests. Following the Early
Mississippian, pO
2
remained at or above 26% until the Mesozoic,
data that are in accord with predictions based on floral carbon
isotopic fractionation
26
. The apparent Pennsylvanian decline is
considered an artefact resulting from the unique morphology of the
dominant arborescent lycopsid coal-forming vegetation and their
ecology. These lycopsids were morphologically distinct from plants
that subsequently dominated this setting and were both highly
fire resistant
27
and grew predominantly in rheotrophic settings
(that is, mires with flowing water) without significant litter layers.
Therefore, fires were probably constrained to the canopy and
produced comparatively little charcoal.
During the Early Permian pO
2
reached a Phanerozoic maximum
of 29%. Inert% predicts a bimodal pO
2
distribution in the
Permian, similar to previous modelling
9
. Although these data
suggest a Middle Permian decline in pO
2
they do not indicate
hypoxia as a contributing factor in the end Guadalupian
(260 Myr) mass-extinction event
28
. Similarly, examination of
Changhsingian (253.8–251 Myr) age coals indicates abundant
charcoal within the last 2.8 million years of the Permian
(average 38.9%; Supplementary Table S2). Therefore, irrespective
of the Changhsingian onset of the Early Triassic superanoxic
event
29
, pO
2
must have remained elevated until at least the very
end of the Permian. It is noteworthy that in using 10-million-year
binning, Late Permian data plot at 250 Myr immediately following
the Permo–Triassic mass-extinction.
During the Mesozoic, Inert% predicts greater pO
2
volatility than
other models
5,9
and indicates almost cyclic fluctuations until the
Late Cretaceous. As early as the Middle Triassic, pO
2
is indicated
to have declined to its lowest level throughout the Phanerozoic
(18.5%). However, the timing of this event is obscure, as no Early
Triassic coals exist
23
, and although the decline in this interval may
have been even more abrupt and extreme, no data exist to test
this hypothesis either.
Among the predicted fluctuations, it is worth highlighting
elevated pO
2
through the Triassic–Jurassic mass-extinction, an
abrupt decline during the Toarcian (183–175.6 Myr) and elevated
levels from the late Middle Jurassic into the earliest Cretaceous.
Analysis of Mist Mountain Formation (150.8–140.2 Myr) coals
(n = 56) shows those in the basal 250 m contain more than double
the inertinite (45.5%) of those above (23.0%) (Supplementary
Table S5). Inert% data from unequivocal Early Cretaceous (145.5 to
121 Myr) coals are sparse, but those from the Bickford Formation
(n = 37, Supplementary Table S6) in particular suggest depressed
pO
2
immediately following the Jurassic–Cretaceous boundary, with
levels rising rapidly thereafter.
Late Early Cretaceous coals (late Aptian (121–112 Myr))
(Supplementary Table S2) indicate rising pO
2
, a trend that
continued through the Albian–Cenomanian (112–93.5 Myr), when
a Mesozoic maximum of 29% is predicted. This was the last interval
when pO
2
was at such elevated levels. It is noteworthy that many
charred flowers have been recovered from this interval, and it has
been speculated that fire, as an agent of environmental disturbance,
played a role in the early evolution of angiosperms
30
.
From the earliest Late Cretaceous, a gradual stepped decline in
pO
2
to 24% 50 Myr is predicted. This fundamental petrographic
difference between Cenozoic and older coals has been recognized
previously
21
, but no explanation has been given. Over the past
40 million years pO
2
seems to have been relatively constant at
21–22%, roughly equivalent to modern levels.
Charcoal abundance in mire settings documents the rise of
pO
2
following the diversification of early land plants and its
subsequent fluctuations throughout the Phanerozoic. From the
best estimate curve, these fluctuations are more frequent than
previously predicted, particularly during the Triassic and Jurassic,
when 20–30 million year cyclicity is detected. These data suggest an
important role for fire as a major Earth system process
13
but also
indicate that the governing feedbacks are not yet fully understood.
Although large amplitude shifts are calculated, fire activity indicates
that low pO
2
events did not drive catastrophic faunal diversity loss in
either the Permo–Triassic or the Triassic–Jurassic mass-extinction
events. Instead, charcoal indicates that pO
2
remained elevated until
at least the latest Permian, and declined precipitously only in the
Early Triassic. Similarly, earliest Jurassic charcoal data indicate
that pO
2
was not low but rather was elevated to levels among the
highest of the Phanerozoic.
Methods
For this study, two coal samples were collected and analysed for maceral
composition. The L’Anse-à-Brillant coal from Tar Point, Gaspé, Canada is of
NATURE GEOSCIENCE | VOL 3 | SEPTEMBER 2010 | www.nature.com/naturegeoscience 629
© 2010 Macmillan Publishers Limited. All rights reserved.
LETTERS
NATURE GEOSCIENCE DOI: 10.1038/NGEO923
Emsian age. The South Tancrediakløft coal from Jamesonland, Greenland is of
very basal Hettangian age. The coals were embedded in polyester resin and counted
following established protocols
20
. Maceral counts were assessed from 500 points of
data, including mineral matter.
Maceral data from the literature, used to determine Inert%, were included
in this analysis only where the inclusion/exclusion of mineral matter was clear.
All data were standardized to a mineral-matter-free basis. With three exceptions,
coals whose stratigraphic resolution was greater than 15 million years were
excluded (for example, Taiyuan formation = Kasimovian–Sakmarian). The
three samples included in the database derive from poorly sampled stratigraphic
intervals where they represent the only data: Givetian–Frasnian (Weatherall–Hecla
Bay–Beverley Inlet formations) and the Anisian-Carnian (Basin Creek and
Mungaroo formations).
Where not tabulated or stated in the text, data were measured from
graphics by pasting the image into CorelDRAW and overlaying guidelines to
obtain exact measurements of data point positions (data are highlighted in
Supplementary Table S2). Preference was given to literature citing named seams.
Where multiple references provide data from one seam, these data were averaged
and all references cited.
Received 10 February 2010; accepted 1 July 2010; published online
1 August 2010
References
1. Flück, M. et al. Coping with cyclic oxygen availability: Evolutionary aspects.
Integr. Comp. Biol. 47, 324–331 (2007).
2. Watson, A., Lovelock, J. E. & Margulis, L. Methanogenesis, fires and the
regulation of atmospheric oxygen. Biosystems 10, 293–298 (1978).
3. Kump, L. R. Terrestrial feedback in atmospheric oxygen regulation by fire and
phosphorous. Nature 335, 152–154 (1988).
4. Falkowski, P. G. et al. The rise of oxygen over the past 205 million years and
the evolution of large placental mammals. Science 309, 2202–2204 (2005).
5. Berner, R. A. GEOCARBSULF: A combined model for Phanerozoic
atmospheric O
2
and CO
2
. Geochim. Cosmochim. 70, 5653–5664 (2006).
6. Berner, R. A. Phanerozoic atmospheric oxygen: New results using the
GEOCARBSULF model. Am. J. Sci. 309, 603–606 (2009).
7. Berner, R. A. & Canfield, D. E. A new model for atmospheric oxygen over
Phanerozoic time. Am. J. Sci. 289, 333–361 (1989).
8. Hansen, K. W. & Wallmann, K. Cretaceous and Cenozoic evolution of seawater
composition, atmospheric O
2
and CO
2
: A model perspective. Am. J. Sci. 303,
94–148 (2003).
9. Bergman, N. M., Lenton, T. M. & Watson, A. J. COPSE: A new model of
biogeochemical cycling over Phanerozoic time. Am. J. Sci. 304, 397–437 (2004).
10. Arvidson, R. S. & Mackenzie, F. T. MAGic: A Phanerozoic model for the
geochemical cycling of major rock-forming components. Am. J. Sci. 306,
155–190 (2006).
11. Jones, T. P. & Chaloner, W. G. Fossil charcoal, its recognition and
paleoatmospheric significance. Palaeogeogr. Palaeoclimatol. Palaeoecol.
97, 39–50 (1991).
12. Glasspool, I. J., Edwards, D. & Axe, L. Charcoal in the Silurian as evidence for
the earliest wildfire. Geology 32, 381–383 (2004).
13. Bowman, D. M. J. S. et al. Fire in the Earth system. Science 324, 481–484 (2009).
14. Scott, A. C. The pre-quaternary history of fire.
Palaeogeogr. Palaeoclimatol. Palaeoecol. 164, 281–329 (2000).
15. Wildman, R. A. et al. Burning of forest materials under late Paleozoic high
atmospheric oxygen levels. Geology 32, 457–460 (2004).
16. Belcher, C. M. & McElwain, J. C. Limits for combustion in low O
2
redefine
paleoatmospheric predictions for the Mesozoic. Science 321, 1197–1200 (2008).
17. Cope, M. J. & Chaloner, W. G. Fossil charcoal as evidence of past atmospheric
composition. Nature 283, 647–649 (1980).
18. Christian, H. J. et al. Global frequency and distribution of lightning as
observed from space by the optical transient detector. J. Geophys. Res. 108,
4005 (2003).
19. Scott, A. C. & Jones, T. P. Fossil charcoal: A plant-fossil record preserved by fire.
Geol. Today 7, 214–216 (1991).
20. Whelan, R. J. The Ecology of Fire (Cambridge Univ. Press, 1995).
21. Taylor, G. H. et al . Organic Petrology (Gebruder Borntraeger, 1998).
22. Scott, A. C. & Glasspool, I. J. Observations and experiments on the origin and
formation of inertinite group macerals. Int. J. Coal Geol. 70, 53–66 (2007).
23. Diessel, C. F. K. The stratigraphic distribution of inertinite. Int. J. Coal Geol.
81, 251–268 (2010).
24. Robinson, J. M. Phanerozoic atmospheric reconstructions: A terrestrial
perspective. Palaeogeogr. Palaeoclimatol. Palaeoecol. 97, 51–62 (1991).
25. Lenton, T. M. & Watson, A. J. Redfield revisited: 2. What regulates the oxygen
content of the atmosphere. Glob. Biogeochem. Cycles 14, 249–268 (2000).
26. Beerling, D. J. et al. Carbon isotope evidence implying high O
2
/CO
2
ratios
in the Permo-Carboniferous atmosphere. Geochim. Cosmochim. Acta 66,
3757–3767 (2002).
27. Robinson, J. M. Phanerozoic O
2
variation, fire, and terrestrial ecology.
Palaeogeogr. Palaeoclimatol. Palaeoecol. 75, 223–240 (1989).
28. Retallack, G. J. et al. Middle-Late Permian mass extinction on land. Geol. Soc.
Am. Bull. 118, 1398–1411 (2006).
29. Isozaki, Y. Permo–Triassic boundary superanoxia and stratified superocean:
Records from lost deep sea. Science 276, 235–238 (1997).
30. Bond, W. J. The tortoise and the hare: Ecology of angiosperm dominance and
gymnosperm persistence. Biol. J. Linn. Soc. 36, 227–249 (1989).
Acknowledgements
We thank the Grainger Fund at FMNH for financial support, and acknowledge R. A.
Berner, W. G. Chaloner and A. J. Watson for helpful discussions. The work of A.C.S. is
financially supported by private charitable donations.
Author contributions
I.J.G. and A.C.S. designed the research. I.J.G. gathered, compiled and interpreted data
and A.C.S. contributed data. I.J.G. wrote the paper with additions by A.C.S.
Additional information
The authors declare no competing financial interests. Supplementary information
accompanies this paper on www.nature.com/naturegeoscience. Reprints and permissions
information is available online at http://npg.nature.com/reprintsandpermissions.
Correspondence and requests for materials should be addressed to I.J.G.
630 NATURE GEOSCIENCE | VOL 3 | SEPTEMBER 2010 | www.nature.com/naturegeoscience
© 2010 Macmillan Publishers Limited. All rights reserved.

Supplementary resource (1)

... Modern and historical data have provided rich and high spatiotemporal resolution of fire's variability on Earth, but the full timespan of these datasets is relatively limited (e.g., to the satellite era) (Bowman et al., 2009). However, fire has been a component of Earth's biosphere for hundreds of millions of years (Glasspool et al., 2004;Glasspool & Scott, 2010), necessitating datasets which span longer timescales to fully resolve its controls and impacts. ...
Article
Full-text available
New approaches are needed to resolve persistent geographic gaps and biases in paleofire research. Most sedimentary paleofire research relies on lake and peat sediments. We present an unconventional sedimentary charcoal record preserved in a modern, post‐bomb bat guano deposit and compare its accumulation to historical fire data. We find strong correlations between charcoal accumulation rates (CHAR) and non‐winter prescribed burns. CHAR in bat guano is more strongly correlated with prescribed fire than wildfire or total area burned, likely due to bats seeking out areas burned by prescribed fire for better foraging opportunities and/or bats avoiding wildfire. We attribute the CHAR in guano being a better recorder of area burned during non‐winter months to winter bat hibernation. Our analyses show that charcoal preserved in bat guano is a reliable paleofire proxy system, which has important implications for the paleofire field and encourages future research using bat guano as a viable archive.
... Similar to the enigmatic absence of cladoxylopsid charcoal, many Middle Devonian coals have low, or no, inertinite content (Diessel 2010;Glasspool and Scott 2010;Glasspool et al. 2015;Lu et al. 2021). Coals from this interval are frequently dominated by liptinitic macerals, particularly cutinite. ...
Article
Full-text available
Evidence of fire in the Middle Devonian remains globally scarce. Charcoalified mesofossils recovered from the Emsian–Eifelian Trout Valley and St. Froid Lake formations of Maine are direct evidence of wildfires proximal to the Acadian Orogen, formed as the Avalon terrane and the North American plate collided. These mesofossils include charred psilophytes, lycopsids, prototaxodioids, enigmatic taxa such as Spongiophyton, and coprolites. Here, fire combusted a senesced and partially decayed litter, and the intimately associated nematophytes, following a period of extended dryness. We envisage wildfires occurred during neap tide when exposure of the flora of this estuarine setting was prolonged. Herein we provide a reconstruction of this Middle Devonian landscape and its flora in which lightning generated by post-dry season storms ignited wildfires that propagated through an extensive psilophyte-dominated litter.
... Garden cress, on the other hand, evolved much more recently, around the time when plants started transitioning from the oceans onto land some 400-300 million years ago (Dunn, 2013). This time was accompanied by a comparable Solar constant as well as oxygen, ozone and carbon dioxide levels to today (Glasspool and Scott, 2010;Dunn, 2013;Mills et al., 2021). Garden cress, as a land plant, is influenced by a range of environmental cues that are intricately linked to Solar light and has evolved to thrive in such a terrestrial environment where it is exposed to more direct Solar light than the cyanobacterium. ...
Article
Full-text available
Stars with about 45 to 80% the mass of the Sun, so-called K dwarf stars, have previously been proposed as optimal host stars in the search for habitable extrasolar worlds. These stars are abundant, have stable luminosities over billions of years longer than Sun-like stars, and offer favourable space environmental conditions. So far, the theoretical and experimental focus on exoplanet habitability has been on even less massive, though potentially less hospitable red dwarf stars. Here we present the first experimental data on the responses of photosynthetic organisms to a simulated K dwarf spectrum. We find that garden cress Lepidium sativum under K-dwarf radiation exhibits comparable growth and photosynthetic efficiency as under Solar illumination on Earth. The cyanobacterium Chroococcidiopsis sp. CCMEE 029 exhibits significantly higher photosynthetic efficiency and culture growth under K dwarf radiation compared to Solar conditions. Our findings of the affirmative responses of these two photosynthetic organisms to K dwarf radiation suggest that exoplanets in the habitable zones around such stars deserve high priority in the search for extrasolar life.
... Based on previous analyses of global inertinite databases, it could be assumed that the low levels of inertinite during the Jurassic period (Diessel, 2010;Glasspool and Scott, 2010) concentrations predicted in some models (Berner, 2009;Krause et al., 2018) would suggest that the Jurassic was a less flammable period in the Earth history than the Cretaceous (He and Lamont, 2018). The Jurassic was also a typical greenhouse climate period (Scotese et al., 2021), with high temperatures and high levels of atmospheric oxygen content ( Fig. 4b and c). ...
Article
Inertinite‐rich coal is widely distributed in the Ordos Basin, represented by the No. 2 coal seam of the Middle Jurassic Yan'an Formation. This paper combined coal petrology and geochemistry to analyze the origin of inertinite, changes in the coal‐forming environment and control characteristics of wildfire. Research has shown that there are two forms of inertinite sources in the study area. Alongside typical fusinization, wildfire events also play a substantial role in inertinite formation. There are significant fluctuations in the coal‐forming environment of samples at different depths. Coal samples were formed in dry forest swamp with low water levels and strong oxidation, which have a high inertinite content, and the samples formed in wet forest swamp and limnic showed low inertinite content. Conversely, the inertinite content of different origins does not fully correspond to the depositional environment characterized by dryness and oxidation. Nonpyrogenic inertinites were significantly influenced by climatic conditions, while pyrofusinite was not entirely controlled by climatic conditions but rather directly impacted by wildfire events. The high oxygen level was the main factor causing widespread wildfire events. Overall, the combination of wildfire activity and oxidation generates a high content of inertinite in the Middle Jurassic coal of the Ordos Basin.
Article
Full-text available
Little evidence of macrofossil charcoal, a wildfire proxy, is recorded from upper Lower to lowermost Upper Devonian rocks. Coals of this age are few, and petrographic data indicate low volumes (<10% mineral-matter free) of charcoal. This paucity of data forms the basis of the “charcoal gap,” which is used to suggest an extended interval of abnormally low atmospheric oxygen (pO2). We reassess the current evidence for this hiatus using Emsian−Eifelian charcoal from the Trout Valley and St. Froid Lake Formations, Maine (northeastern United States), and integrate the microscopic charcoal record of dispersed organic matter. We conclude there is ample evidence of fire in the Middle Devonian. This interval is not innately of low pO2. Rather, it is one in which under-interpretation of available data has led to a perceived paucity of charcoal. This reconciliation indicates the Phanerozoic record of wildfire was substantially uninterrupted. Hence, we propose that pO2 achieved levels >16% and remained at such levels from the Silurian through the floral and faunal colonization of land and, from our current estimates, stayed as such until the present.
Preprint
The Chicxulub asteroid impact event at the Cretaceous-Paleogene (K-Pg) boundary ~66 Myr ago is widely considered responsible for the mass extinction event leading to the demise of the non-avian dinosaurs. Short-term cooling due to massive release of climate-active agents is hypothesized to have been crucial, with S-bearing gases originating from the target rock vaporization considered a main driving force. Yet, the magnitude of the S release remains poorly constrained. Here, the amount of impact-released S is estimated empirically relying on the concentration of S and its isotopic composition within the impact structure and a set of terrestrial K-Pg boundary ejecta sites for the first time. The value of 67 ± 39 Gt obtained is ~5-fold lower than recent numerical estimates but concurs with numerical estimates from the 1990s. The lower mass of S-released implies global average temperatures above freezing point with key implications for species survival during the first years following the impact.
Chapter
This chapter provides an introduction into Chaps. 9 and 10, focussed on abiotic processes. Earth abiotic processes are influenced by interactions between terrestrial and extra-terrestrial forces inherited form the origins of the universe. Chaotic dynamics of abiotic processes are driving forces of biological evolution, providing energetic sources together with constraints for organic life with different factor combinations in marine and terrestrial habitats. Physical properties of water together with particular dynamic characteristics and giant dimensions of the three-dimensional oceanic habitats were preconditions for the evolution of animal life on the energetic basis of microorganisms. Physical properties of the atmosphere and lithosphere were primary preconditions for the evolution of terrestrial vegetation with trenchant consequences for the evolution of terrestrial animal life forms.
Article
Full-text available
We present a new model of biogeochemical cycling over Phanerozoic time. This work couples a feedback-based model of atmospheric O2 and ocean nutrients (Lenton and Watson, 2000a, 2000b) with a geochemical carbon cycle model (Berner, 1991, 1994), a simple sulfur cycle, and additional components. The resulting COPSE model (Carbon-Oxygen-Phosphorus-Sulfur-Evolution) represents the co-evolution of biotic and abiotic components of the Earth system, in that it couples interactive and evolving terrestrial and marine biota to geochemical and tectonic processes. The model is forced with geological and evolutionary forcings and time-dependent solar insolation. The baseline model succeeds in giving simultaneous predictions of atmospheric O2, CO2, global temperature, ocean composition, δ13C and δ34S that are in reasonable agreement with available data and suggested constraints. The behavior of the coupled model is qualitatively different to single cycle models. While atmospheric pCO2 (CO2 partial pressure) predictions are mostly determined by the model forcings and the response of silicate weathering rate to pCO2 and temperature, multiple negative feedback processes and coupling of the C, O, P and S cycles are necessary for regulating pO2 while allowing δ13C changes of sufficient amplitude to match the record. The results support a pO2 dependency of oxidative weathering of reduced carbon and sulfur, which raises early Paleozoic pO2 above the estimated requirement of Cambrian fauna and prevents unrealistically large δ34S variation. They do not support a strong anoxia dependency of the C:P burial ratio of marine organic matter (Van Cappellen and Ingall, 1994, 1996) because this dependency raises early Paleozoic δ13C and organic carbon burial rates too high. The dependency of terrestrial primary productivity on pO2 also contributes to oxygen regulation. An intermediate strength oxygen fire feedback on terrestrial biomass, which gives a pO2 upper limit of ∼1.6PAL (present atmospheric level) or 30 volume percent, provides the best combined pO2 and δ13C predictions. Sulfur cycle coupling contributes critically to lowering the Permo-Carboniferous pCO2 and temperature minimum. The results support an inverse dependency of pyrite sulfur burial on pO2 (for example, Berner and Canfield, 1989 , which contributes to the shuttling of oxygen back and forth between carbonate carbon and gypsum sulfur. A pO2 dependency of photosynthetic carbon isotope fractionation (Berner and others, 2000; Beerling and others, 2002) is important for producing sufficient magnitude of δ13C variation. However, our results do not support an oxygen dependency of sulfur isotope fractionation in pyrite formation (Berner and others, 2000) because it generates unrealistically small variations in δ34S. In the Early Paleozoic, COPSE predicts pO2=0.2-0.6PAL and pCO2>10PAL, with high oceanic [PO3-4] and low [SO=4]. Land plant evolution caused a 'phase change' in the Earth system by increasing weathering rates and shifting some organic burial to land. This change resulted in a major drop in pCO2 to 3 to 4PAL and a rise in pO2 to ∼1.5PAL in the Permo-Carboniferous, with temperatures below present, ocean variables nearer present concentrations, and PO4:NO3 regulated closer to Redfield ratio. A second O2 peak of similar or slightly greater magnitude appears in the mid-Cretaceous, before a descent towards PAL. Mesozoic CO2 is in the range 3 to 7PAL, descending toward PAL in the Cretaceous and Cenozoic.
Article
Full-text available
A dynamical model (MAGic) is presented that describes the elemental cycling of sedimentary materials involving sodium, potassium, calcium, magnesium, chloride, carbon, oxygen, iron, sulfur and phosphorous through much of the Phanerozoic. The model incorporates the basic reactions controlling atmospheric carbon dioxide and oxygen concentrations, continental and seafloor weathering of silicate and carbonate rocks, net ecosystem productivity, basalt-seawater exchange reactions, precipitation and diagenesis of chemical sediments and authigenic silicates, oxidation-reduction reactions involving carbon, sulfur, and iron, and subduction-decarbonation reactions. Although MAGic contains feedback and forcing functions adapted from the GEOCARB models (Berner, 1991, 1994; Berner and Kothavala, 2001), these functions are incorporated in a reservoir-reaction scheme that is considerably more detailed. Coupled reservoirs include shallow and deep cratonic silicate and carbonate rocks and sediments, seawater, atmosphere, oceanic sediments and basalts, and the shallow mantle. Model results are reasonably consistent with recently published constraints provided by fluid inclusion, isotopic, floral, and mineralogical records. We have used these results to evaluate sensitivity to uncertainties in the history of the earth-ocean-atmosphere system over the past 500 Ma: the advent of pelagic carbonate sedimentation, the importance of burial versus early diagenetic dolomite formation, the importance of reverse weathering, and the relationship of these processes to seafloor spreading rates. Results include a general pattern of dolomite abundance during periods of elevated seafloor spreading and alkalinity production, elevated atmospheric CO2 concentrations for most of the Phanerozoic similar to those predicted by GEOCARB, and covariance of seawater sulfate to calcium ratios with magnesium to calcium ratios. These trends are broadly consistent with proxies for seawater composition and the mass-age data of the rock record itself.
Article
Full-text available
The continuous charcoal record, interpreted with the aid of the results of combustion experiments, indicates that the mixing ratio of atmospheric oxygen has varied remarkably little over the past 350 Myr. We develop a dynamic feedback model of the coupled P, N, C, and O2 cycles and use perturbation analysis and a case study of the past 40 Myr to test various feedback mechanisms that have been proposed to stabilize atmospheric oxygen. These mechanisms involve alterations in nutrient driven productivity and the subsequent burial flux of organic carbon, which provides the main source of atmospheric oxygen. Suppression of the burial of phosphorus sorbed to iron minerals under anoxic conditions in ocean bottom waters tends to increase the ocean nutrient inventory and provide negative feedback against declining oxygen [Holland, 1994]. However, denitrification is enhanced by anoxia, tending to reduce the nutrient inventory and amplify declining oxygen [Lenton and Watson, this issue]. If organic phosphorus removal from the ocean is also suppressed under anoxic conditions, this improves oxygen regulation [Van Cappellen and Ingall, 1994], as does direct enhancement of organic carbon burial due to reduced oxygen concentration in bottom waters [Betts and Holland, 1991]. However, all of the ocean-based feedback mechanisms cease to operate under increases in oxygen that remove anoxia from the ocean. Fire frequency is extremely sensitive to increases in oxygen above 21% of the atmosphere, readily suppressing vegetation on the land surface. This should transfer phosphorus from the land to the ocean, causing less carbon to be buried per unit of phosphorus and providing a weak negative feedback on oxygen [Kump, 1988]. However, a new proposal that increases in oxygen suppress the biological amplification of rock weathering and hence the input of phosphorus to the Earth system provides the most effective oxygen regulation of all the mechanisms considered. A range of proxies suggests that the input of available phosphorus to the ocean may have been significantly reduced 40 Myr ago, suppressing new production and organic carbon burial in the model. With only ocean-based feedback, the atmospheric oxygen reservoir is predicted to have shrunk from ~26% of the atmosphere 40 Myr ago. However, when land plant mediated negative feedback on phosphorus weathering is added, oxygen is regulated within 19-21% of the atmosphere throughout the past 40 Myr, in a manner more consistent with paleorecords.
Article
Full-text available
Theoretical models predict a marked increase in atmospheric O2 to ∼35% during the Permo-Carboniferous (∼300 Ma) occurring against a low (∼0.03%) CO2 level. An upper O2 value of 35%, however, remains disputed because ignition data indicate that excessive global forest fires would have ensued. This uncertainty limits interpretation of the role played by atmospheric oxygen in Late Paleozoic biotic evolution. Here, we describe new results from laboratory experiments with vascular land plants that establish that a rise in O2 to 35% increases isotopic fractionation (Δ13C) during growth relative to control plants grown at 21% O2. Despite some effect of the background atmospheric CO2 level on the magnitude of the increase, we hypothesize that a substantial Permo-Carboniferous rise in O2 could have imprinted a detectable geochemical signature in the plant fossil record. Over 50 carbon isotope measurements on intact carbon from four fossil plant clades with differing physiological ecologies and ranging in age from Devonian to Cretaceous reveal a substantial Δ13C anomaly (5‰) occurring between 300 and 250 Ma. The timing and direction of the Δ13C excursion is consistent with the effects of a high O2 atmosphere on plants, as predicted from photosynthetic theory and observed in our experiments. Preliminary calibration of the fossil Δ13C record against experimental data yields a predicted O2/CO2 mixing ratio of the ancient atmosphere consistent with that calculated from long-term models of the global carbon and oxygen cycles. We conclude that further work on the effects of O2 in the combustion of plant materials and the spread of wildfire is necessary before existing data can be used to reliably set the upper limit for paleo-O2 levels.
Article
Full-text available
The end-Permian mass extinction has been envisaged as the nadir of biodiversity decline due to increasing volcanic gas emissions over some 9 million years. We propose a different tempo and mechanism of extinction because we recognize two separate but geologically abrupt mass extinctions on land, one terminating the Middle Permian (Guadalupian) at 260.4 Ma and a later one ending the Permian Period at 251 Ma. Our evidence comes from new paleobotanical, paleopedological, and carbon isotopic studies of Portal Mountain, Antarctica, and comparable studies in the Karoo Basin, South Africa. Extinctions have long been apparent among marine invertebrates at both the end of the Guadalupian and end of the Permian, which were also times of warm-wet greenhouse climatic transients, marked soil erosion, transition from high- to low-sinuosity and braided streams, soil stagnation in wetlands, and profound negative carbon isotope anomalies. Both mass extinctions may have resulted from catastrophic methane outbursts to the atmosphere from coal intruded by feeder dikes to flood basalts, such as the end-Guadalupian Emeishan Basalt and end-Permian Siberian Traps.
Article
Charcoal is produced by pyrolysis of plant material and its occurrence in the fossil record can be broadly equated with the incidence of palaeowildfire. The past record of such naturally occurring fire, and tha availability of the biomass which represents its fuel, put two constraints on oxygen levels. For combustion of plant material to occur at all requires that the atmospheric oxygen did not drop below a threshold of 13%. Increasing inflammability of plant material at higher oxygen levels suggests that 35% would be a ceiling above which plant biomass would ignite and burn so readily as to be incompatible with sustained forest growth. As we have more or less continuous fossil evidence of forest trees from the Late Devonian onwards, and a similarly sustained record of fossil charcoal from that time to the present (Cope, 1984), this constraints oxygen levels between 13% and 35% over that period (Rabash and Langford, 1968; Watson et al., 1978). However, further experimental work is required to establish the validity of these oxygen values under appropriate conditions and also to sharpen the certainty by which we can discriminate between fusain produced by pyrolysis, and inert wood degradation products produced by other (? biogenic) means. We discuss experiments directed at attempting to establish the validity of physical parameters by which pyrolytically produced fusain can be characterized. The most convincing evidence of pyrolysis hitherto recognised is the apparent homogenization of xylem cell walls, as seen under SEM. Work on charcoal from both wildfires and laboratory wood charring under controlled conditions confirms the homogenization as seen under both SEM and TEM. Controlled temperature experiments show that a further rise in temperature causes the cell walls, initially homogenized, to crack and separate along the site of the middle lamella, giving the charcoal a characteristic fibrous texture. Both of these distinctive phases of response to pyrolysis can be observed in fossil charcoals.
Article
Modifications of the GEOCARBSULF model for Phanerozoic atmospheric oxygen have been made to account for new carbon isotopic data, reconsideration of the fractionation of carbon isotopes between carbonate and organic matter deposited in sediments, and different rates of weathering of volcanic rocks versus granitic rocks. Results indicate distincdy higher O2 values than GEOCARBSULF for the Mesozoic and Cenozoic, lack of an appreciable drop in O 2 below 15 percent at any time, and a small late Cenozoic decline of O2 to the present.
Article
Rare basal Prídolí plant fossils, which resemble the rhyniophytoid Hollandophyton colliculum and have exceptional three-dimensional cellular anatomy, are preserved as charcoal. As such, these fossils from Ludford Lane in the Welsh Borders are evidence of the earliest recorded wildfire and are the first documented from before the Devonian. The fossils represent the charred remains of a low-growing, low-diversity, in situ rhyniophytoid or cryptospore-bearing plant vegetation charred during a low-temperature fire.
Article
Charcoal is produced by pyrolysis of plant material and its occurrence in the fossil record can be broadly equated with the incidence of palaeowildfire. The past record of such naturally occurring fire, and tha availability of the biomass which represents its fuel, put two constraints on oxygen levels. For combustion of plant material to occur at all requires that the atmospheric oxygen did not drop below a threshold of 13%. Increasing inflammability of plant material at higher oxygen levels suggests that 35% would be a ceiling above which plant biomass would ignite and burn so readily as to be incompatible with sustained forest growth. As we have more or less continuous fossil evidence of forest trees from the Late Devonian onwards, and a similarly sustained record of fossil charcoal from that time to the present (Cope, 1984), this constraints oxygen levels between 13% and 35% over that period (Rabash and Langford, 1968; Watson et al., 1978). However, further experimental work is required to establish the validity of these oxygen values under appropriate conditions and also to sharpen the certainty by which we can discriminate between fusain produced by pyrolysis, and inert wood degradation products produced by other (? biogenic) means. We discuss experiments directed at attempting to establish the validity of physical parameters by which pyrolytically produced fusain can be characterized. The most convincing evidence of pyrolysis hitherto recognised is the apparent homogenization of xylem cell walls, as seen under SEM. Work on charcoal from both wildfires and laboratory wood charring under controlled conditions confirms the homogenization as seen under both SEM and TEM. Controlled temperature experiments show that a further rise in temperature causes the cell walls, initially homogenized, to crack and separate along the site of the middle lamella, giving the charcoal a characteristic fibrous texture. Both of these distinctive phases of response to pyrolysis can be observed in fossil charcoals.