University of Southern California Dissertations and Theses

How open ocean calcifiers broke the link between large igneous provinces and mass extinctions

(USC Thesis Other)

How open ocean calcifiers broke the link between large igneous provinces and mass extinctions

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content HOW OPEN OCEAN CALCIFIERS BROKE THE LINK BETWEEN LARGE IGNEOUS
PROVINCES AND MASS EXTINCTIONS

by

Jessica Lynn Stellmann

A Thesis Presented to the
FACULTY OF THE DORNSIFE COLLEGE OF LETTERS, ARTS AND SCIENCES
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
MASTER OF SCIENCE
(GEOLOGICAL SCIENCES)
December 2022

ii
ACKNOWLEDGEMENTS
Thank you to my collaborators on this project: A. Joshua West, Andy Ridgwell and
Thorsten Becker. This work benefited from discussions with Frank Corsetti, Naomi Levine, Julien
Emile-Geay, William Berelson, David Bottjer, and Lowell Stott at the University of Southern
California, Sarah Greene at the University of Birmingham, and Joyce Yager. I would also like to
thank John Yu for his support and guidance in the computational components of this project.
My advisor, Josh West, has my gratitude for his guidance, support, and patience over my
time as a graduate student at USC. I deeply appreciate the faith he expressed in my abilities, as
well as his understanding of my decisions for my future. I also want to acknowledge the grace with
which he dealt with my fairly constant, irreverent (attempts at) humor and levity.
I would like to thank Susan Harris, Jake Peters, Samantha Bernstein-Sierra, Sable Manson,
and all of the (full-time, undergraduate, and graduate) staff of USC’s Joint Educational Project
over the last several years for giving me an on-campus home and bringing me into the JEP Family.
Special thanks to Dieuwertje Kast, of JEP’s STEM Education Programs, for being a role-model,
mentor and friend who has shown me the roadmap to my life’s work.
On a personal note, I want to express my unending gratitude to Matthew Helfgott, Lottie
Morris, all of Duck and Goat Co. as well as my family for their ceaseless love and support of me
throughout my, at times tumultuous feeling, journey to this degree. Thank you all for genuinely
not caring what letters follow my name and for encouraging me to make the best choice for myself.
To my friends and fellow students in the Earth Sciences Department both current and former,
especially Maxwell Dahlquist (as well as his wife Meryn), Alex Hatem, Dylan Wilmeth, Thomas
Luckie (as well as his wife Sarah), Naomi Rodgers, Emily Tibbett and Maya Yanez: thank you for
sharing this journey with me and filling it with camaraderie and laughter. You are all incredible,
and I look forward to hearing about your triumphs for years to come!
My thanks as well to the faculty and staff of USC’s Earth Sciences Department, especially
Darlene Garza, for their help and support of the completion of this thesis.
Finally, I would like to thank my thesis committee: Josh West, Frank Corsetti and Naomi
Levine for their support in completing this work. Funding for this work was provided by National
Science Foundation’s Earth Life Transitions Grant #EAR-1338329.
iii

TABLE OF CONTENTS
ACKNOWLEDGEMENTS ………………………………………………………………………ii
LIST OF TABLES ………………………………………………………….……...…………….iv
LIST OF FIGURES …………………………………………………………….…………………v
ABSTRACT ………………………………………………………………………....…………...vi
CHAPTER ONE: INTRODUCTION ..………………………………………………….………..1
CHAPTER TWO: EXPERIMENT SETUP & METHODS …………………………….…...….11
CHAPTER THREE: RESULTS .………………………………………………………....……. 26
CHAPTER FOUR: DISCUSSION…….…………………………………………………….......28
CHAPTER FIVE: CONCLUSIONS ….………………………………………………………...36
CHAPTER SIX: SUPPLEMENT ……………….………………………………………………37
FIGURES & TABLES …………………………………………………………………………..44
REFERENCES ……….………………………………………………………………………….59

iv
LIST OF TABLES
TABLE 1: Summary of the major LIPs of the Phanerozoic …………………………………….44
TABLE 2: Key results from all model runs ……………………………………………………..49

v
LIST OF FIGURES
FIGURE 1 Phanerozoic LIPs associated with mass extinctions & kill mechanisms…………... 45
FIGURE 2 Schematic of Experimental Setup …………………..………………………………46
FIGURE 3 Response of CO2 and calcite saturation state to 10,000 GtC injected over 10kyrs ...47
FIGURE 4 Response of CO2 and calcite saturation state across experimental suite ………...…48
FIGURE 5 Comparison of DIC flux from sediments to ocean in open ocean vs. reef
carbonate production scenarios …………………………………………………………………51
FIGURE 6 Response of average global land temperature to CO2 injection ……………………52
FIGURE 7 Comparison of key time-series results for end-Permian runs with
accurate solar radiation flux vs. modern solar radiation flux ……………………………...……53
FIGURE 8 Ocean-atmosphere CO2 exchange in high initial pCO2 model runs …………..……54
FIGURE 9 Summary of diagnostic results for anomalous end-Permian run …………………...55
FIGURE 10 Comparison of key time-series results for end-Permian runs with standard
vs. random bathymetry ..……………………………..………………………..…………………56
FIGURE 11 Comparison of key time-series results for runs with altered reef
carbonate parametrization ………………………………………………………………….……57

vi
ABSTRACT
Four of the “Big Five” mass extinctions in Earth’s history have been associated with
intense volcanism from Large Igneous Provinces (LIPs), suggesting close ties between carbon
emissions and biotic crisis. Yet not all LIP events had the same environmental impacts, nor did
all cause mass extinctions. Here we use an Earth system model to address the question of why
some LIPs were associated with severe environmental collapse while others were not. We find
that the production of carbonate by calcifying organisms in the open ocean can significantly
buffer the Earth system, reducing the impact of CO2 release on the oceans and atmosphere when
compared to carbonate production only at reefs. The evolution of open ocean calcifiers in the
mid-Mesozoic (~ 200 Myrs ago) may thus at least in part explain less severe impacts from LIPs
over the past 150 Myrs compared to prior eruptions, pointing to the possibility that life and the
Earth system co-evolved to buffer carbon cycle perturbations.
1
CHAPTER ONE: INTRODUCTION
Large Igneous Provinces are defined as those igneous provinces which cover aerial
extents of 10
5
-10
7
km
2
and have total magmatic and extrusive volumes in excess of 10
6
km
3

(Marzoli, et al., 1999, McHone 2003, Bryan and Ernst 2008, Self et al. 2014). These events are
thought to be caused by the decompression melting of mantle plume heads (Campbell
and Griffiths, 1990, Ernst and Buchan, 2001, Campbell 2005, Sobolev, et al., 2011). Though LIP
eruptions may last ~1 Myrs or longer overall, most material is emplaced during series of shorter
pulses, each lasting on the order of several thousand years but some possibly as short as 1,000
years (Chenet et al. 2008, 2009; Marzolli et al. 2011; Blackburn et al. 2013). The features that
comprise LIPs when they are erupted on continental lithosphere are continental flood basalts
(Jerram and Widdowson 2005, Self et al. 2014) and often-extensive igneous complexes (Marzoli
et al. 1999, Svensen et al. 2004, Polteau et al. 2008). When erupted in marine settings they can
form oceanic plateaus and marine flood basalts (Neal et al. 1997, Bryan and Ernst 2008, Greene
et al. 2008). The majority of LIPs are mafic in nature (Jerram 2002, Bryan and Ernst 2008, Self
et al. 2014), though there are cases of LIPs dominated by silicic composition (Bryan 2007). The
Earth system’s response to LIP eruption and emplacement, in particular the effects of the
volatiles released during eruption, can be devastating (Wignall 2001, Sobolev et al. 2011, Bond
and Wignall 2014, Jones et al. 2016), making studying and understanding these events a useful
tool to investigate the Earth system’s response to the rapid addition of volatiles to the
atmosphere.

2
1.1 VOLATILE CONTENT OF LIP LAVAS
The major volcanic gasses which make up the volatile component of these eruptions are
H2O vapor (50-90%), CO2 (1-40%), Sulfur in the forms of SO2 and H2S (2-35%)
%) and small amounts of halogens such as chlorine and fluorine (Bond and Wignall, 2014, Self
et al. 2014, Jones et al. 2016). Of these, water vapor, while the largest component by volume and
a greenhouse gas, has little direct impact on the global system as it is already such a large
component of the atmosphere; its impact is relegated to providing a positive feedback to
warming or cooling caused by other species (Jones et al. 2016). Of the major volcanic gases,
CO2 and SO2 exert the largest effect on global climate through their effects on the Earth’s
radiative balance (Wignall 2001, Self et al. 2005, Sobolev et al. 2011, Bond and Wignall 2014,
Self et al. 2014, Jones et al. 2016). Measurements on LIP basalt inclusion samples indicate that
every cubic kilometer of lava erupted in a continental flood basalt could release 6∓1.7 Mt of SO2
and 14 Mt of CO2 to the atmosphere (Self et al. 2006, 2014), though these emissions could
potentially be greatly increased by interactions with surrounding host rock and sediments
(Ganino and Arndt 2009). Volatile estimates from oceanic plateau eruptions are not available due
to the difficulty in sampling these provinces.
The long-term effect of SO2 emission is global cooling due to aerosol formation when
SO2 is injected into the stratosphere. Cooling from large historic eruptions such as Huaynaputina
(1600 CE), Laki (1783-84 CE) and Tambora (1818 CE) among several others, has been
documented to last for ~1-3 years (de Silva & Zielinski, 1998, Thordarson & Self, 2003, Briffa et
al., 1998). One additional impact of SO2 -driven global cooling may be a decrease in silicate
weathering rate which could amplify the rise of pCO2 by up to 25% as CO2 drawdown via
silicate weathering is reduced (Mussard et al. 2014). Though SO2 emissions have been linked to
3
mass extinctions (Callegaro et al. 2014, Black et al. 2014a), the timescale of its effects is short.
CO2, on the other hand, is a greenhouse gas which affects Earth’s radiative balance over
hundreds of thousands of years (Walker et al. 1981, Berner and Caldiera 1987, Dessert et al.
2003, Jones et al. 2016). Upper estimates of volatile release for the eruption of the Deccan Traps
suggest CO2 release may have reached a rate of 1 GtC/yr or even higher (Chenet et al. 2009).
This is lower than the rate of anthropogenic emissions (approaching 10 GtC/yr in 2018; (Le
Quéré et al. 2018) but nonetheless represents extremely rapid CO2 release in the geologic
context.

1.2 THE GLOBAL CARBON CYCLE
The Earth’s carbon cycle operates through a number of reservoirs on time scales which
can vary from days to millions of years, making understanding the impact of large carbon release
events challenging. The long-term carbon cycle, operating on time scales of hundreds of
thousands to millions of years, is mainly regulated by the transfer of carbon from surface
reservoirs (the combined reservoir of carbon in the atmosphere, biosphere, oceans, soils and
sediments, which can exchange with the water column), and geologic reservoirs, comprising
crustal carbonate rocks, sufficiently buried marine sediments and the mantle (Berner et al. 1983,
Berner and Caldeira 1997). This long-term cycle is thought to be controlled primarily by the
relationship between atmospheric CO2 concentrations (in partial pressure, pCO2) and the rate of
weathering of calcium-silicate rocks (Berner et al. 1983, Berner 1991, Berner 1994). This
relationship, first described in Walker et al. (1981), generates a negative feedback loop in which
higher atmospheric CO2 concentrations cause global warming, and this increase in global
temperature then causes increased rates of silicate rock weathering. The weathering products are
4
carried to the oceans where they eventually combine to form carbonate minerals which are then
buried in ocean sediments, removing them from the surficial carbon cycle until and unless they
are subducted and released during decarbonation. The net result of the weathering of silicate
minerals (and their eventual deposition in the oceans as carbonate minerals) is the sequestration
of half of the CO2 consumed by the initial weathering reaction (see Eq. 1 and 2 below, Berner
1998). Thus, the increase in silicate weathering under warmer climate leads to an increase in the
rate of CO2 drawdown.
2CO2 + H2O + CaSiO3 ⇄ Ca
2+
+ 2HCO3
-
+SiO2 (Eq. 1)
Ca
2+
+ 2HCO3
2-
⇄ CaCO3 + H2O + CO2 (Eq. 2)
While the rate of weathering of carbonate rocks also increases with increasing temperature, there
is no net effect on pCO2 with carbonate weathering (Eq. 2). The response of silicate weathering
to a change in atmospheric pCO2 continues until the atmospheric concentration of CO2 is low
enough that temperatures fall, and silicate weathering slows. Over millions of years this brings
the silicate weathering flux and the volcanic CO2 outgassing flux into steady state (Walker et al.
1981, Berner et al. 1983, Berner 1991, Francois and Walker 1992). While this feedback is what
controls atmospheric CO2 and to a large degree global climate on multi-million year to billion
year time scales, at time scales of ~100 kyr and less this feedback mechanism is unable to keep
pace with changes in pCO2, and the processes which determine CO2 concentration in the
atmosphere are those involving CO2’s interactions with Earth’s oceans (Ridgwell and Zeebe
2005, Lord et al. 2016).
The exchange of CO2 between the surface oceans and the atmosphere is governed by the
partial pressure (pCO2) of the gas in each reservoir (Zeebe and Wolf-Gladrow 2001). When CO2
enters the ocean some of it combines with water to form H2CO3, carbonic acid. This carbonic
5
acid then quickly (and almost completely) dissociates, resulting in bicarbonate ions (HCO3
-
).
Further dissociation can occur, producing carbonate ions (CO3
2-
). This chain of reactions,
summarized in Eqs 3-5 (Ridgwell and Zeebe 2005) below, leads to an increase in the
concentration of H
+
ions (and therefore a decrease in pH) and, when this change is significant
enough to bring ocean pH below average levels, is referred to as ocean acidification (Caldeira
and Wickett 2003).
CO2(aq) + H2O ⇄ H2CO3 (Eq. 3)
H2CO3 → H
+
+ HCO3
-
(Eq. 4)
HCO3
-
→H
+
+ CO3
2-
(Eq. 5)
The total amount of all of these species (CO2(aq) + HCO3
-
+ CO3
2-
, H2CO3 is negligible) is
the dissolved inorganic carbon, or DIC, content of the ocean. The relative proportions of each
species of DIC will be affected by a few different processes, though generally (at modern
seawater conditions) HCO3
-
is the most common species. The processes of ocean acidification
above (Eqs. 3-5) have equilibrium constants:
K1= ([H
+
][HCO3
-
])/[CO2(aq)]
K2=([H
+
][CO3
2-
])/[HCO3
-
]
Where K1 corresponds to the combination of Eq. 3 and 4 and K2 to Eq. 5. These
equilibrium constants dictate that an increase of [H
+
] like that associated with ocean acidification
will favor a DIC partitioning which favors CO2(aq). Another process which will affect the
partitioning of DIC is the precipitation or dissolution of carbonate minerals (Eq. 2). Carbonate
ions (CO3
2-
) combine with calcium ions (Ca
2+
, abundant in seawater) to form calcium carbonate
(CaCO3) as either calcite (trigonal crystal structure, more common in modern oceans) or
aragonite (orthorhombic) (Ridgwell and Zeebe 2005). Whether conditions are favorable for
6
carbonate precipitation or dissolution to take place is measured by the saturation state, Ω which
is defined as:
Ω = ([Ca
2+
][CO3
2-
])/Ksp
where Ksp is the solubility constant for CaCO3 (Zeebe and Wolf-Gladrow, 2001). If this value is
above unity, precipitation of CaCO3 is favored; if it is less than unity, dissolution is favored
(Walter and Morse 1985). Both precipitation and dissolution will affect the partitioning of DIC.
When CaCO3 is precipitated, the concentration of CO3
2-
decreases, and this causes the remaining
DIC to repartition in favor of CO2(aq), increasing ocean pCO2 and leading to a flux of CO2 into
the atmosphere (Ridgwell and Zeebe 2005). This precipitation-caused repartitioning is why only
half of the CO2 consumed by silicate weathering is removed from the surface reservoirs and why
carbonate weathering has no net effect on atmospheric pCO2 at all. On the other hand, dissolution
of carbonate encourages repartitioning towards higher CO3
2-
concentrations and will lower ocean
pCO2 possibly causing a flux of CO2 out of the atmosphere (lowering atmospheric pCO2) and
into the ocean.
It should be noted that even if the ambient environment is supersaturated with respect to
CaCO3 (Ω > 1), like in the modern ocean, spontaneous precipitation of CaCO3 from the water
column is not observed due to the initial nucleation being kinetically unfavorable (Morse and He
1993). In laboratory experiments, spontaneous precipitation has not occurred until ΩCalcite > 20-
25 (Morse et al. 2003). Most of the CaCO3 precipitation that we observe in modern oceans is
associated with biotic processes: corals, plankton, etc. Many of these processes are also limited
by Ω. Corals for example do not thrive and produce carbonate skeletons unless their environment
has sufficient carbonate saturation (Greene et al. 2012 and references therein), making ocean
acidification detrimental to calcifying biota and decreasing the amount of carbon they can
7
remove from the system. While values of Ω observed in modern oceans today (~4.8, Ridgwell &
Zeebe 2005) do not favor spontaneous precipitation, they do favor preservation of CaCO3 that is
produced biotically. As CO2 concentrations in the oceans increase due to increased atmospheric
pCO2, DIC re-partitioning causes Ω to decrease (through a decrease in CO3
2-
concentration, see
Goodwin et al. 2007 Figure 1), and dissolution to be favored over preservation. This dissolution
will increase the CO3
2-
concentration of seawater and thus allow for a lower ocean pCO2 and
more uptake of CO2 from the atmosphere. This mitigation of an atmospheric pCO2 change by a
change in deep-sea carbonate preservation is a strong control on the global carbon cycle at
timescales of thousands of years and is called carbonate compensation, or seafloor neutralization.

1.3 EARTH SYSTEM RESPONSE TO LIP VOLATILE RELEASE
The rapid CO2 release from LIPs has been implicated in periods of global warming, ocean
acidification, ocean anoxia and ultimately, mass extinctions, including 4 of the “Big 5”
(Rampino and Stothers,1988, White and Saunders 2005, Courtillot and Renne 2003, Hautmann
et al. 2008, Alroy 2010, Sobolev et al. 2011, Kravchinsky 2012, Bond and Wignall 2014,
Courtillot and Fluteau 2014, Self et al. 2014, Jones et al. 2016). CO2 emissions from LIP
eruptions first enter the atmosphere, where, due to CO2 molecules’ ability to absorb infrared
radiation emitted by Earth’s surface, increased CO2 concentrations lead to increased average
global temperatures (i.e. global warming) on the decadal-centennial time scale (IPPC Fifth
Assessment Report). Higher global temperatures will persist as long as atmospheric pCO2
remains elevated and lead to an intensified hydrological cycle (Held and Soden 2006) which
delivers larger nutrient fluxes to the oceans as a consequence of increased weathering (Berner
8
1991, 1994). This higher nutrient flux combined with warmer oceans, increased respiration rates
in ocean organisms, and decreased oxygen solubility (also effects of global warming) can cause
increased productivity and a stratified ocean possibly resulting in ocean anoxia (Knoll et al.
2007). Increased ocean temperatures caused by global warming can have direct harmful effects
as well, particularly on organisms such as corals, foraminifera, cephalopods, echinoderms,
gastropods, bivalves and more (Song et al. 2014, Urban 2015). For example, an increase in
surface ocean temperature of just 4.3 degrees C can cause an increase in extinction rate of corals
of 16% (Urban 2015).
The surface ocean and atmosphere constantly exchange CO2; when the two are at
equilibrium, this flux is virtually equal in both directions but when excess CO2 is added to the
atmosphere by a LIP eruption there is a net flux of CO2 entering the surface ocean. If CO2 levels
reach high enough concentrations in the surface oceans for even as briefly as a few weeks (200
ppm above ambient, Knoll et al. 2007; ~ 550 ppm, McNeil and Sasse 2016), more complex
organisms will begin experiencing deleterious effects such as decreased growth and reproductive
rates in the lead up to the onset of hypercapnia at concentrations of 1000 ppm (Knoll et al. 2007,
McNeil and Sasse 2016). Hypercapnia increases the oxygen demand of organisms. This effect,
paired with increased ocean temperatures and decreased O2 saturation caused by global warming,
creates an unfortunate synergy which can in fact lower the threshold values at which all of these
processes are harmful, thereby increasing ecological vulnerability (Knoll et al. 2007).
As discussed in the prior section, once CO2 is dissolved into the ocean, a series of
disassociation reactions occur (Ridgwell and Zeebe 2005). This chain of reactions leads to an
increase in [H
+
] (and therefore a decrease in pH); if enough excess CO2 is added to the ocean this
process can cause ocean pH to drop below average levels in what is termed an ‘ocean
9
acidification event’ (Caldeira and Wickett 2003). If enough CO2 is added to the ocean system,
widespread carbonate dissolution can occur in the thousands of years following the addition of
excess CO2 (Ridgwell and Zeebe 2005). Ocean acidification negatively impacts calcifying biota,
leading to increased risk of extinction. Here again there is also an additional impact of increased
temperatures due to global warming—while CO2 solubility decreases with increasing
temperature, this possible mitigation is offset by the increase to dissolution rates caused by
higher temperatures (Ries et al. 2016). Indeed, in laboratory experiments at high temperatures,
dissolution of whole-shell biogenic Ca CO3 occurred even in waters oversaturated with respect to
calcite (Ries et al. 2016).
On timescales of >100 kyrs, increased levels of CO2 in the atmosphere and the associated
increase in mean global temperature will lead to higher rates of silicate rock weathering; as
silicate rock weathering draws CO2 out of the atmosphere, this “weathering feedback” should
ultimately restore carbon cycle equilibrium, reducing atmospheric pCO2 and global temperatures
to pre-eruption levels (Walker et al. 1981, Berner et al. 1983, Berner 1991, 1994).

1.3.1 VARIATION IN OBSERVED EARTH SYSTEM RESPONSE TO LIP VOLATILE
RELEASE
The record of LIPs and their environmental effects is best preserved over the Phanerozoic
era of the past 560 Myrs. As seen in Table 1, not all LIP events are associated with mass
extinctions. Many large igneous provinces, particularly the Cretaceous oceanic plateaus (some of
the largest LIPs on record), are not associated with elevated extinction rates (Bond and Wignall
2014). Similarly, the occurrence of phenomenon such as ocean acidification crises, ocean anoxia
10
and even global warming are variable, with some LIP eruptions being correlated in the rock
record to all of the above, and others being associated with only one or two (Table 1; Wignall
2005, Bond & Wignall 2014, Courtillot & Fluteau 2014).
Notably, there was a marked decrease in the environmental impacts and lethality of LIP
eruptions after 150 Ma (Figure 1; Wignall 2005, Bond and Wignall 2014, Courtillot and Fluteau
2014, Bond and Grasby 2017). This decrease cannot be tied only to size since the largest LIP
known, the Ontong Java Plateau, erupted ~120 Ma with relatively muted environmental
consequences. Nor can the change be explained simply by subaerial vs. subaqueous eruption
since the Parana-Etendeka and Deccan Traps, two major continental flood basalt eruptions,
occurred in the post-150 Ma interval. The North Atlantic Igneous Province was emplaced (at
least partially) subaerially at ~56 Mya, and has been linked to the Paleocene-Eocene Thermal
Maximum which, while undoubtedly a period of environmental changes, does not approach the
catastrophic extinction events of the first half of the Phanerozoic (Storey et al. 2007, Jones et al.
2019 and references therein).

11
CHAPTER TWO: EXPERIMENT SETUP & METHODS
In this work, we explore possible explanations for the change in response to LIP CO2
release since ~150 Ma using the Earth system model cGENIE to model CO2 emissions similar to
those expected from a LIP eruption pulse. We consider a range of different planetary
configurations, focusing on changes that may explain the apparently diminished environmental
impact of volatile release since ~150 Ma. The first of these is the change in continental
geography from an assembled supercontinent (Pangea) towards the more fragmented
configuration of the modern era; this tectonic reconfiguration has been progressing for the last
200 Ma (Seton et al. 2012, Matthews et al. 2016). There a several proposed mechanisms through
which the break-up of Pangea is thought to have affected the global carbon cycle including:
altering the global silicate weathering flux via continental rifting and associated passive margin
formation, the closure of the shallow, carbonate-rich Tethys ocean via subduction, changing
albedo (and therefore altering the balance point for the CO2-silicate weathering feedback loop),
and changing ocean circulation (Donnadieu et al. 2006a). The second global change since ~150
Ma that we consider is the shift in the locus of carbonate production from occurring only in
shallow water environments (mostly reefs) to occurring in the open ocean with the evolution of
pelagic calcifiers over the course of the Mesozoic. This change led to an ocean well-regulated
with respect to carbonate saturation state and less prone to extreme calcite over saturation
(Ridgwell 2005). Both of these shifts have been suggested as causes of the variability in Earth
system response to LIP eruptions (Bond and Wignall 2014, Courtillot and Fluteau 2014,
Eichenseer et al. 2019). Yet quantitative evaluation of their effects has remained lacking.
Given the uniqueness of each LIP eruption in terms of location, eruption dynamics and
pre-eruption climate as well as the uncertainties in determining those factors from the rock
12
record, it is difficult to conclusively compare individual LIP eruption events to test for specific
causality of the difference in their associated Earth system responses. Modeled LIP-scale carbon
injection scenarios allow for control of these other parameters, providing one possible approach
to investigating the possible impacts of continental configuration and carbonate production on
global response.

2.1 GLOBAL CARBON CYCLE MODELING APPROACHES
Most of the modeling work done to date on LIP eruptions and their global impacts has
been focused on reconstructing specific LIP events; these studies have made use of a variety of
carbon cycle models (Beerling & Berner 2002, Berner & Beerling 2007, Cui et al. 2013, Black et
al. 2014, Paris et al. 2016). What follows is a discussion of a selection of carbon cycle models at
various complexities, and where possible we describe those which have been used in
investigating LIP eruptions in order to make clear our reasoning in choosing cGENIE for this
study.
Many models have been developed to investigate changes to Earth’s carbon cycle, with
different suitability depending on the timescale of interest. The long term (> 100 kyr) carbon
cycle is viewed as a balance between mantle and metamorphic degassing of CO2, from mid-
ocean ridges, subduction zones and orogens, and the consumption of CO2 by continental
chemical weathering of silicate rocks. Due to the heightened weathering of basalt as compared to
other silicate rocks (Dessert et al. 2003), on this timescale large igneous provinces can actually
act as a sink for atmospheric CO2 (Schaller et al. 2012). A widely used group of models suitable
for investigating behavior of the carbon cycle on this timescale is the GEOCARB family of
models, most recently GEOCARB III and GEOCARBSULF (Berner & Kothavala 2001, Berner
13
2006). The GEOCARB suite of models is built around the central assumption of mass balance
between the sources of atmospheric CO2 (volcanic degassing from mid-ocean ridges and
subduction zones, as well as metamorphism of both carbonates and organic material) and the
primary sinks (silicate and carbonate weathering, plus organic carbon burial) over time scales of
hundreds of thousands to billions of years. The GEOCARB models make this assumption of
steady state and calculate atmospheric pCO2 through the assertion that any differences between
other inputs and outputs are due to silicate rock weathering, with the rate of silicate weathering
primarily controlled by pCO2. This structure means that changes in volcanic degassing of CO2
caused by changes in tectonic activity are the dominant control of atmospheric pCO2 over
Earth’s history. The models are structurally fairly simple, treating the ocean and atmosphere as a
single reservoir (with the other reservoirs being organic carbon and carbonate rock) and
parametrizing all fluxes by factors such as uplift, runoff and continental area. This simplicity
makes for a computationally inexpensive, yet powerful tool for the study of how pCO2 has
evolved over Earth’s history.
In the shorter-term carbon cycle (operating on timescales of 0.1-10’s kyrs), the primary
processes involved in removal of CO2 from the atmosphere are ocean uptake (i.e. acidification)
and sedimentary carbonate dissolution (Ridgwell and Zeebe 2005, Goodwin et al. 2007,
Goodwin and Ridgwell 2010, Lord et al. 2016). The rock record supports the idea that these
processes are linked to LIP eruptions (Payne et al. 2007, Greene et al. 2012, Hönisch et al. 2012,
Martindale et al. 2012, Bond & Wignall 2014) making their inclusion in our investigation into
the range of global responses to LIP-scale emissions critical. Additionally, there is evidence from
cyclostratigraphy that the upper end of this time scale (~10’s kyrs) is the time scale of individual
pulses of magmatism associated with large igneous provinces (Olsen et al. 2003, Whiteside et al.
14
2007, Schaller et al. 2012). As many of the time scales involved in large igneous province
eruption and their aftermath are much shorter than ~100 kyrs, GEOCARB with its dependence
on an assumption of mass balance which is only applicable to timescales of 100’s kyrs or longer
is ill suited to investigating the global response to the rapid, voluminous CO2 injection associated
with LIPs.
Box models that do not include GEOCARB’s assumption of mass balance in order to
calculate silicate weathering and pCO2 allow for investigation on shorter timescales. Schaller et
al. 2012 make use of a modified version of the COPSE model (Bergman et al. 2004), which
combines the geochemical carbon cycle of GEOCARB with a feedback-based model of
atmospheric oxygen and ocean nutrients as well as a simple Sulphur cycle. While this modified
COPSE model is not reliant on the mass-balance assumption that GEOCARB is, making it
suitable for analysis of the Earth system’s reaction to LIP-scale emissions, it is still only
appropriate for the study of long-term processes such as the silicate weathering response to such
carbon emissions. In their 2012 paper, Schaller et al. use this model to investigate the long-term
evolution of atmospheric CO2 after a rapid CO2 injection of 4200 GtC as CO2 over 1000 years
(meant to simulate a single pulse of volcanism of the Central Atlantic Magmatic Province or
CAMP, a LIP that erupted in the end-Triassic; Marzoli et al. 2004, 2011; Blackburn et al. 2013)
under several silicate weathering schemes: one based solely on increased pCO2 and the
associated temperature rise, one at a 1.5x amplification of the temperature-only response, and
one at 2x amplification. These latter two scenarios were carried out to test the idea that the
presence of the CAMP basalts would accelerate weathering rates due to the increased weathering
rates of basalt as compared to other silicate rocks (Dessert et al. 2003). Their results
demonstrated that the modeled pCO2 for a 1.5x amplification of silicate weathering had returned
15
to pre-eruption levels by ~270 kyr after eruption, and eventually decreased to ~20% below pre-
eruption levels; this finding was in line with their data of pCO2 estimates from pedogenic
carbonates sourced from sediment sequences between and above CAMP basalts in the Newark
and Hartford basins (Schaller et al. 2011a, Schaller et al. 2012). While a valuable study of the
long-term weathering response to LIP eruptions, the COPSE model that they used is unable to
resolve shorter timescales, making it ill-suited to studying the shorter-term responses or the
effects of pulsed magmatism and associated volatile release.
Another box model which was used in Beerling and Berner, 2002 (and subsequently with
modifications in Berner and Beerling, 2007) to study the effects of the CAMP is a simplified
version of the BLAG model (Berner et al. 1983). Within this model, Earth’s surficial carbon is
separated into 4 reservoirs: Atmosphere, Ocean, Carbonate Rocks, and Terrestrial Biomass (this
reservoir interacts with the carbon cycle in terms of temporary carbon storage as biomass, not as
a long-term organic carbon reservoir, Beerling et al. 2002). The model also accounts for volcanic
and methane fluxes of carbon into the atmosphere and for carbon transported from the
atmosphere to the ocean via silicate weathering. Beerling and Berner 2002 apply this model to
investigating if the ~1000 ppm rise in pCO2, recorded in stomatal leaf (McElwain et al. 1999)
and paleosol (Tanner et al. 2001) proxies, and the -2.0 to -3.5‰ negative carbon-isotope
excursion (Pálfy et al. 2001, Hesselbo et al. 2002, Ward et al. 2001) recorded near the Triassic-
Jurassic boundary, could be replicated from a modeled CAMP eruption. Four total emission
amounts were tested: 1,326, 7,957, 13,262 and 21,220 GtC, chosen to cover a range of basalt
CO2 concentrations and CAMP volume estimates (Beerling and Berner, 2002). Each of the tested
emission amounts was released in a Gaussian distribution with a peak half-height time of ±100
kyr, setting the total duration of emissions to 500 kyr, in line with the best CAMP dating at the
16
time (Olsen et al. 1996, Marzoli et al. 1999). An additional set of simulations with an added
methane release of ~5000 GtC was also carried out. The results of their modeling indicated that
the observed rise in pCO2 nor the isotope excursion could be explained by CAMP CO2 emissions
alone, and that the most likely emissions scenario was the eruption of ~8000-9000 GtC as CO2
from CAMP followed by ~5000 GtC as CH4. They attributed the methane emissions to the
release of seafloor methane hydrates as a result of global warming (Beerling and Berner, 2002).
The major change made to the model in Berner and Beerling, 2007 is the modification of the rate
of CaCO3 deposition, making this dependent on the carbonate saturation state of the ocean which
is allowed to vary with atmospheric CO2. This model is then used to investigate the possibility
put forth in Hautmann, 2004 that following the eruption of CAMP the ocean became globally
undersaturated with respect to CaCO3. Total emission amounts tested were 5,317, 10,634 and
21,269 GtC as CO2 with SO2 emissions matching CO2 1:1 in Gaussian distributions with total
eruption durations of 50 and 100 kyr. They found that in order to achieve and maintain an
undersaturated ocean for 20-40 kyr, ~21,000 GtC as CO2 (and ~57,000 GtS as SO2) had to erupt
in under 100kyr (Berner and Beerling, 2007). While this model allows for rapid computation and
incorporates the idea of carbonate deposition being dependent on atmospheric pCO2 by way of
saturation state, it does not allow for carbonate dissolution nor does it account for the ability of
the ocean to remove CO2 from the atmosphere and ‘store’ it through acidification. As both of
these processes have implications for the CaCO3 saturation state of the ocean over ~10’s kyrs
(Ridgwell and Zeebe 2005, Lord et al. 2016), this modeling effort is incomplete.
At the other side of the complexity spectrum from one-box ocean models are general
circulation models (GCMs) for the ocean and atmosphere systems. GCMs offer excellent
temporal and spatial resolution with which to study the Earth’s climate response but are far too
17
computationally intensive to be effectively run for time frames long enough to capture LIP
events. GCMs have been applied to studying the environmental effects of LIP-scale carbon
emissions (Huynh and Poulson, 2005) but it has been through comparisons of steady state
scenarios under different atmospheric CO2 concentrations, not through a dynamic injection of
CO2 through time.
The GEOCLIM model (Donnadieu et al. 2006a, b; applied to CAMP in Paris et al. 2016)
seeks a middle ground between GCMs and box-modeling. GEOCLIM makes use of a 10-box (in
its original configuration, Goddéris and Joachimski 2004) ocean biogeochemical model
COMBINE (Goddéris and Joachimski 2004, Simon et al. 2007) and results from several runs of
the GCM FOAM (Jacob 1997, Donnadieu et al. 2006b) which provide estimates for temperature
and runoff under different pCO2 for a variety of paleogeographies (Donnadieu et al. 2006a, b).
These results from FOAM allow for better constrained weathering fluxes and allow a more
complete investigation into how weathering processes respond to changes in pCO2 than in other
models described above. The expansion of ocean and associated biogeochemical modeling to a
multi-box system allows for the existence of multiple ocean basins, albeit in a very simple
manner, and the inclusion of processes such as ocean acidification and biological production as
well as the separation of carbonate deposition into shallow and pelagic regimes make it a
significant improvement over more simple one-box ocean models. Paris et al. 2016 made use of
the GEOCLIM model to explore the consequences of a CAMP emissions scenario that is
geologically constrained; that is to say instead of modeling total CAMP emissions as one
Gaussian, they confine emissions to a succession of intense but short pulses as is seen in the rock
record of CAMP emplacement (Knight et al. 2004, Schaller et al. 2011). The model was run
using FOAM-derived, offline grids of temperature and runoff under pCO2 ranging from 200-
18
4200 ppm in 200 ppm increments for a Rhaetian paleogeography (Donnadieu et al. 2006a).
COMBINE is altered from its original configuration (Goddéris and Joachimski 2004) to instead
represent an 11-box ocean with two polar oceans (above 60° N and S) with photic zones and
deep ocean reservoirs; a mid-latitude ocean with a photic zone, thermocline and deep ocean
reservoirs; and two epicontinental seas each with a photic zone and a deep epicontinental
reservoir; a 12
th
box represents the atmosphere (Paris et al. 2016). All carbonate deposition is
restricted to the epicontinental seas, with the burial flux being controlled by the CaCO3
saturation state of those boxes. CAMP degassing scenarios used in the modeling efforts are
based on the ages of CAMP lava flows (Blackburn et al. 2013) resulting in 4 main emission
peaks in two different duration scenarios: one where peaks last 1kyr, the other 10kyr, though in
both scenarios the third peak, which is less stratigraphically resolved last for 60kyr (Schaller et
al. 2012, Paris et al. 2016). Two values for the total volume of CO2 released during CAMP
eruptions were tested: a low hypothesis of 1400 GtC (McHone 2003), and a high hypothesis of
21,200 GtC (Berner and Beerling 2007). A run with 21,200 GtC is emitted as CO2 in a single
Gaussian over 500 kyr was completed as a reference. Paris et al. 2016 find that their geologically
constrained degassing scenario led to modeled carbon cycle responses of repeated short-term
decreases in carbonate production contemporaneous with negative excursions in organic and
inorganic carbon isotopes as well as an increase in atmospheric CO2. The pCO2 increases are in
good agreement with pedogenic carbonate reconstructions (Schaller et al. 2011a, 2012), while
the isotope excursions only match published records if the mantle source is ~ -20‰. Like the
much simpler Berner and Beerling 2007 above, GEOCLIM does not make allowances for
carbonate dissolution in undersaturated conditions, only for the cessation of new deposition. This
lack of carbonate compensation as an active process to buffer p CO2 rise is the main drawback of
19
the GEOCLIM model along with its reliance on the results of the FOAM GCM, which are not
publicly available.

2.1.1 cGENIE
In this study we have elected to use the medium-complexity Earth system model
cGENIE. cGENIE is a 3D ocean circulation and biogeochemical model coupled to a 2D energy-
moisture balance model of the atmosphere (Edwards and Marsh 2005) and a dynamic-
thermodynamic sea-ice model. The ocean biogeochemical cycling of DIC, ALK and a single
nutrient (PO4) is implemented on a 36x36 equal area horizontal grid, and the ocean has either 8
or 16 depth levels, depending on the model configuration selected (Lenton et al. 2006, Ridgwell
et al. 2007). It also has a fully coupled sediment model (Ridgwell and Hargreaves, 2007) and a
fairly basic parameterization of terrestrial rock weathering to deliver DIC and calcium ions to the
oceans (Coulbourn et al. 2013).
cGENIE has been used in several investigations of the fate of anthropogenic CO2 over the
next centuries, millennia and into the next million years (Lenton et al. 2006, Ridgwell and
Hargreaves 2007, Cao et al. 2009, Goodwin and Ridgwell 2010, Lord et al. 2016) and has also
been applied to carbon-release events in Earth history such as the PETM (Panchuck et al. 2008)
and the end-Permian mass extinction (Cui et al. 2013). A recent modeling project using cGENIE
that provides an interesting point of comparison for our efforts to look at the evolution of the
Earth system following LIP-scale emissions is Lord et al. 2016. This work involves a series of
simulations of a range of emissions from 1,000 to 20,000 Gt C as CO2 released as instantaneous
pulses. These model runs aimed to deconvolve the relevant timescales and efficiencies of all CO2
buffering processes and how these timescales and efficiencies vary with increasing emission
20
size. They found that the evolution of atmospheric pCO2 following a pulse of emissions could be
best represented by five separate exponential functions (Lord et al. 2016). The first three of these
exponentials have median characteristic timescales of 1.2, 36 and 730 years respectively and are
taken to all be representative of the processes of CO2 uptake by the oceans: air-sea gas exchange,
ocean circulation carrying CO2 into the deep ocean and the carbonate buffering capacity of the
ocean waters (Lord et al. 2016). Together these account for a median of ~50% of CO2 removal
from the atmosphere. This fraction of removal decreases with increasing emissions as the first
two exponential functions decrease in efficiency of removing CO2 with increasing emissions.
The timelines over which they are active decrease, possibly reflecting a decrease in the solubility
of CO2 in seawater as a result of increased global warming at higher emissions. In contrast, the
third exponential function’s efficiency and timescale both increase up to an emissions level of
4,000 GtC. At higher emissions the efficiency of this third function decreases with increasing
emissions while the timescale over which it is active continue to increase. The process captured
by the fourth exponential function has a median characteristic timescale of 11 kyr and is
responsible for removing 12-57% (median 43%) of excess CO2 from the atmosphere; both the
timescale and the fraction of CO2 removed increase with increasing emissions. This process is
taken to be seafloor neutralization (Ridgwell and Hargreaves 2005, Lord et al. 2016), i.e.,
dissolution of carbonate from seafloor sediments – importantly, a component of the carbon cycle
missing from all other models described above. The final exponential function, with a median
time scale of 268 kyr, corresponds to the increased silicate weathering caused by the emissions
and associated warming. There was very little impact on either time scale or total fraction of CO2
removed from the atmosphere (7-8%) with changing emissions, indicating that at least up to the
maximum tested CO2 emission of 20,000 GtC, the other processes are capable of compensating
21
for each other, with seafloor neutralization taking up more CO2 when the efficiency of ocean
uptake decreases, so that emissions are mostly mitigated by the time silicate weathering responds
(Lord et al. 2016). This prior work illustrates how the cGENIE model is complex enough to
capture the relevant ocean and sediment processes (i.e. biological production, ocean
acidification, carbonate compensation, carbonate deposition) while also being computationally
efficient enough to be able to capture the full duration of a LIP pulse event as well as long-term
carbon cycle processes such as the >100 kyr effects of terrestrial weathering. These features
make it a powerful tool to investigate the factors which may influence global response to LIP
eruption.

2.2 MODEL INITIATION
We developed a suite of 16 cGENIE configurations spanning different continental
configurations, initial climates and ocean carbonate ‘factories’ (reefs vs. open ocean calcifiers).
The ocean biogeochemical cycling of total dissolved inorganic carbon (DIC), alkalinity (ALK)
and a single nutrient (PO4), was implemented on a 36-by-36 equal area horizontal grid and an 8-
layer ocean with the fully coupled sediment model and basic parametrization of terrestrial rock
weathering (described above) enabled. Two basic cGENIE configurations were utilized in this
study: ‘worbe2’ (Ridgwell et al. 2007, Lord et al. 2016), a modern continental configuration, and
‘p0251a’ (Cui et al. 2015), an end-Permian continental configuration. These were selected as
they represent two endmembers of the supercontinent cycle. To isolate the effects of differing
continental configuration from differing carbonate deposition scheme, the pre-existing
configurations were modified and used as follows:
• ‘worbe2’ was used:
22
o as is; a modern continental configuration with only carbonate deposition via
planktonic biomineralization and “calcite rain” into deep ocean sediments
(‘worbe2-a’); and
o as a modern continental configuration with only shallow water carbonate
deposition in model ‘reef’ locations (‘worbe2-b’).
• ‘p0251a’ was used similarly:
o unaltered, as an end-Permian continental configuration (Pangea) with only
shallow water carbonate deposition in model ‘reef’ locations (‘p0251a-a’); and
o as an end-Permian continental configuration with only carbonate deposition via
planktonic biomineralization and “calcite rain” into deep ocean sediments
(‘p0251a-b’).
Each of these four setups was then used in combination with two different initial climate states
(cold at 300 ppm initial pCO2, warm at 2000 ppm) and, for each climate state, two different
ocean chemistries (an ocean at around average modern calcite saturation state (Ωcal ~ 7) and a
supersaturated ocean (Ωcal ~ 17)). These different scenarios were explored since both pCO2 and
ocean chemistry have varied throughout Earth history (leaving past reconstructions subject to
uncertainty) as well as to assess whether either of the two global shifts (super-continent
separation, shift in global carbonate production) would proceed differently under different
boundary conditions.
Once the initial conditions were selected the first stage of ‘spin-up’ was executed. This
first stage for the modern and end-Permian runs with open-ocean carbonate production (worbe2-
a and p0251a-b respectively) was a “closed” weathering/sediment system with a prescribed,
constant weathering flux added to the deepest ocean level; the initial value of pCO2 was also kept
23
constant through restorative forcing. In this stage of the spin-up, ocean chemistry was forced
towards the selected initial conditions via the target mean global surface ocean calcite saturation
state. If this value fell below the selected value, a flux of DIC was added uniformly to each ocean
box. This phase of the spin-up was run for 20kyr. For the second stage of ‘spin-up’, the ocean
chemistry forcing was removed, and the system was switched to an ‘open’ weathering state
(weathering flux calculated from global temperature and runoff, supplied to the surface ocean in
areas determined by the assumed topography of the landmasses). This stage was a series of short
(1kyr) runs used to bring weathering flux and CaCO3 burial into equilibrium. Once an
equilibrium was established, the pCO2 forcing was removed and the third stage of the spin-up
was run for 60kyr to ensure that steady state was maintained when no forcing was applied.
The procedure for spinning-up the modern end-Permian runs with reef carbonate production
(worbe2-b and p0251a-a respectively) was similar except for the first stage of the spin-up. Due to
the way that reef carbonate production is calculated in the model, the tuning of weathering and
CaCO3 burial used for the above two configurations does not work for reef carbonate
configurations. To produce an equilibrium state with similar ocean chemistry to their open ocean
carbonate factory counterparts, the first stage of this spin-up was a 20kyr, ‘open’ weathering
stage, run with only pCO2 being forced, with initial conditions for ocean chemistry set to the
same values as in the worbe2-a and p0251a-b counterparts and the weathering flux set to the
equilibrium value of those configurations. From this point, the spin-up procedure was the same
as outlined above. Figure 2 presents a schematic view of the 16 model configurations used in our
study.

24
2.3 EXPERIMENT RUNS
Experiment runs were completed using all 16 of our described configurations. A standard
model eruption of 10,000 GtC as CO2 over 10,000 years (injection rate: 1GtC/year) was
simulated in all 16 model configurations. Additionally, to test possible variability associated with
eruption characteristics, eruptions of 10,000 GtC as CO2 over 1,000 years (10 GtC/year) and of
1,000 GtC over 10,000 years (0.1 GtC/year) were injected into the 300 ppm, less saturated ocean
configurations for all 4 setups.
In selecting our criteria for this study, we aimed to encompass the range of previously
estimated and modeled LIP eruption durations and emissions as well as the range of climates
seen through Earth history so that the Earth system’s response could be characterized across as
complete an assemblage of scenarios as possible. Estimates of CO2 emissions from well-studied
LIPs such as the CAMP and Siberian Traps range from 1900-46,000 GtC based on direct
measurements of volatile content in lava inclusions as well as extrapolation from modern
volcanic CO2 fluxes (Beerling and Berner 2002, McHone 2003, Beerling et al. 2007, Berner and
Beerling 2007, Sobolev et al. 2011, Schaller et al. 2012, Paris et al. 2016). For context, CO2
release from remaining possibly accessible fossil fuels are estimated at 1,000-4,000 GtC
(McGlade and Ekins 2015). To encompass these ranges in our study we have selected emission
amounts of 1,000 and 10,000 GtC as CO2, excluding the highest estimates of Sobolev et al. 2011
as particular to the eruption context of the Siberian Traps and not a representation of a typical
LIP eruption pulse.
The choice of emission durations is complicated by the knowledge that many LIP
eruptions are actually comprised of discrete pulses of eruption interrupted by periods of
quiescence (Bryan and Ernst 2011, Marzoli et al. 2011, Schaller et al. 2011, Blackburn et al.
25
2013). In previous modeling efforts, simulated durations have ranged from 1,000 to 60,000 years
for individual pulses (Schaller et al. 2012, Paris et al. 2016) and from 50-500 kyr for total
eruptions (Beerling and Berner 2002, Berner and Beerling 2007, Beerling et al. 2007, Paris et al.
2016). We elected to test durations of 1kyr and 10kyr in this study. The first is in line with the
shortest reported duration for a single pulse within a LIP eruption (Schaller et al. 2011b, 2012)
and provides a point of comparison to previous studies on the fate of anthropogenic CO2 (Lord
et al. 2016). 10 kyr represents a likely duration for pulses within a LIP-eruption (Paris et al.
2016). Additionally, a handful of additional runs were conducted to eliminate or assess various
possible complications (see Discussion below for more information).
Results from cGENIE are stored in two formats: ‘time-series’ which record globally
averaged values (e.g., atmospheric pCO2, global temperature, average DIC concentration) at
specified intervals and ‘time-slices’, which are snapshots of a given parameter over either a 2-D
(e.g., Surface Temperature) or 3-D model grid (e.g. ,DIC in every ocean box). Due to the time
required to render the ‘time-slices’, they are saved less frequently than time-series. Both data
types are saved at points which decrease in frequency as the model run proceeds.

26
CHAPTER THREE: RESULTS
Across all experiment configurations and ‘eruptions,’ pCO2 rises continuously throughout
each eruption, reaching its peak in the final year of eruption before beginning to decline as CO2
is removed from the atmosphere into the ocean reservoir. When looking at the time-series results
for pCO2 from the ‘standard eruption’ (10,000 GtC over 10 kyr) in all 4 basic configurations
(initial pCO2 of 300 ppm, less well saturated ocean with respect to calcium carbonate) in Figure
3, key differences are apparent. While small differences are observed in the total rise of pCO2
between continental configurations (e.g., the 645.0 ppm rise in the end-Permian continental
configuration with planktonic calcifiers vs. 687.5 ppm rise in the modern continental
configuration with planktonic calcifiers), the major difference is between different carbonate
production regimes. The end-Permian with reef calcifiers configuration saw an increase in pCO2
that was 239.7 ppm greater than the increase in the end-Permian with planktonic calcifiers
configuration; the Modern with reef calcifiers rose 245.1 ppm more than its planktonic
counterpart. This larger change in pCO2 is tied to increased temperature changes, with the reef
configurations’ atmospheres warming 0.9°C and 1.2°C more in modern and end-Permian reef
configurations respectively. This increased warming is also present in the model ocean: the
modern reef configuration global (surface) ocean warmed 0.7°C (0.8°C) more than the
planktonic configuration, the end-Permian reef global (surface) ocean 1.2°C (0.6°C) more. While
all configurations reach peak pCO2 during the final year of eruption, the decline in pCO2 after
eruption ceases shows different behavior between the continental configurations, with end-
Permian configurations removing CO2 from the atmosphere faster than their modern
configuration counterparts. The end-Permian planktonic configuration saw the removal of half of
the added CO2 from the atmosphere by 4,020 years after the eruption ends; its modern
27
counterpart achieved this by 5,650 years post eruption. There are also differences between
carbonate production regime: for reef configurations half of the added CO2 was removed by
3,830 years and 4,459 years post-eruption for end-Permian and Modern continental
configurations respectively, faster than the planktonic carbonate production regimes.
When looking at the average surface ocean saturation state of calcite (Ωc) for these four
runs, once again there is a substantial difference between carbonate production configurations;
the modern reef configuration saw Ωc decrease by 0.85 more than the modern planktonic
configuration and the end-Permian reef configuration decreased by 0.98 more than the end-
Permian planktonic configuration. There is a significantly smaller difference between continental
configurations: 0.17 more in the end-Permian planktonic vs. modern planktonic, 0.30 more in the
end-Permian reef vs. modern reef. In all configurations Ωc reaches its minimum while eruption is
still occurring; for the end-Permian planktonic this occurs in model year 5280, while for the
modern planktonic it occurs in year 6150. In the reef configurations, the minimum occurs later,
in model year 8680 for the end-Permian reef and model year 9290 for the modern reef. In all
cases the rate of recovery of Ωc quickens once eruption ceases. Results for pCO2 increase and Ωc
decrease across the entirety of our experiment matrix are presented in Figure 4.

28
CHAPTER FOUR: DISCUSSION
4.1 CARBONATE FACTORY
The most significant differences we observe are between model runs that differ in mode
of carbonate production; the models with reef calcifiers saw increases in pCO2 that were 245
ppm and 240 ppm greater than models with planktonic calcifiers for the modern and end-
Permian geographies, respectively. In tandem, global and ocean temperatures in the reef
configurations warmed by more than their planktonic counterparts and Ωc exhibited a greater
decline (for temperature results please see Table 2). The decreased response of atmospheric
pCO2, mean global temperature, and ocean Ωc in the planktonic vs. reef models is broadly
consistent across the experimental suite, including runs with different eruption rates and initial
conditions (Figure 4), although there are some exceptions discussed below.
Of particular interest to the impact of carbonate production mechanism are our e = 10
GtC/yr experiments (where e = the eruption rate). Like the standard e = 1GtC/yr scenarios, the e
= 0.1 GtC/yr experiments show greater changes in pCO2 and Ωcalcite in the experimental setups
with reef carbonate production than those with pelagic carbonate production. The decrease in
Ωcalcite is of larger magnitude than the increase in pCO2. This result is also seen in the pCO2
changes in the e = 10 GtC/yr experiment runs; it is however not seen in the Ωcalcite response in
these higher eruption rate scenarios. Instead the response is virtually even across pelagic and reef
setups with the pelagic setups seeing very slightly larger decreases in Ωcalcite. This deviation may
indicate that the process responsible for the reef scenario’s larger response to CO2 injection is
operating at timescales greater than this experiment’s 1kyr eruption duration. This extra-
millennial timescale excludes atmospheric processes, while the fact that this behavior is only
observed in this short injection-duration run excludes processes such as silicate and carbonate
29
rock weathering which exert control on the carbon cycle on timescales longer than all of our
injection scenarios. This points in the direction of a process operating on an intermediate
timescale, such as that characteristic of global carbonate system response.
Specifically, the differences we observe between reef and planktonic models can be
attributed to differences in the buffering of atmospheric CO2 addition by carbonate dissolution in
the oceans. CO2 injected into the atmosphere is taken up by the oceans which become more
acidic, leading them to dissolve carbonate sediments which in turn lowers aqueous CO2
concentrations and allows the ocean to take up more CO2 from the atmosphere. The amount of
atmosphere CO2 taken up by the oceans is limited by the amount of carbonate sediment that
dissolves, represented by the flux of dissolved inorganic carbon (DIC) transferred from
sediments to the ocean. In models that allow open-ocean carbonate production via planktonic
calcifiers, carbonate was deposited across much of the Earth’s surface (70.2% for the modern
and 68.1% for the end-Permian geographies). On the other hand, with carbonate production
restricted to shallow reef environments, deposition was highly restricted (≈1.5% of the Earth’s
surface in the modern and 4.7% in the end-Permian), meaning that the dissolution response,
while it may be quite strong (An increase of over 10
12
mol/year is recorded in the reef sediment
grid box with the highest increase in dissolution), is limited in its ability to buffer the ocean. And
indeed, when looking at the flux of DIC from the model sediments to the model oceans across
reef and pelagic scenarios, we see that the DIC flux in reef model setups is almost an order of
magnitude lower than in their pelagic counterparts (Figure 5). Interestingly the increase in this
DIC flux from the sediments over the course of eruption in the standard 10,000 GtC over 10kyrs
case is of similar magnitude across reef and pelagic carbonate regimes, but the total flux is
nonetheless significantly less in the reef carbonate factory runs, depriving the oceans of enough
30
alkalinity to buffer the invading CO2 as efficiently. This difference in the carbonate system’s
ability to significantly mediate environmental impacts from CO2 release under a pelagic
carbonate regime such as the one present on Earth from the evolution of planktonic calcifiers in
the Mesozoic to the present may be a key factor in the decrease in mass extinction events over
the last 150 Ma as compared to the rest of the Phanerozoic.

4.2 CONTINENTAL CONFIGURATION
Differences in peak pCO2 between continental configuration were much smaller than the
differences between mode of carbonate production; pCO2 increased by 645 ppm in the end-
Permian planktonic continental configuration vs. 688 ppm in the equivalent modern runs
(additional results in Table 2). Continental configuration had a more notable effect on the timing
of recovery. All configurations reached peak pCO2 during the final year of eruption, but the end-
Permian configurations removed CO2 from the atmosphere faster than their modern counterparts
(half of excess CO2 removed by 4,020 years after the end of CO2 injection vs. 5,650 years for the
planktonic calcification runs). These timescales, effectively the “half-lives” for CO2 drawdown,
were also lower in the reef configurations, at 3,830 years (end-Permian) and 4,459 years
(modern). In all cases, Ωc reached a minimum while eruption was ongoing, but the timing was
earlier in the end-Permian than in the modern, and earlier in the planktonic than in the reef
configurations (see Table 2). The rate of recovery in Ωc quickened once eruption ceased.
Differences in recovery timescale can be attributed to the response of continental
chemical weathering, which delivers alkalinity to the oceans and thus leads to atmospheric CO2
removal. The higher peak pCO2 and temperature in reef configurations enhanced the continental
31
weathering response, leading to more rapid CO2 drawdown than in the planktonic runs.
Meanwhile, mean land surface temperature increased more over the end-Permian supercontinent
than over the modern continents (Figure 6). Though the peak temperatures reached in the end-
Permian runs were lower than in the modern due to the decreased solar flux in the end-Permian,
the change in temperature was greater, so weathering rates increased more than in the modern
experiments, leading to faster CO2 drawdown (we also tested a single run with end-Permian
geography and modern solar insolation, still seeing increased changes in surface land
temperature, larger weathering responses and faster recoveries; Figure 7).

4.3 ERUPTION RATE
Model runs with different eruption rates (e) predictably showed differing responses to
CO2 injection. The total change in pCO2 was smallest for e=0.1 GtC/yr (1,000 GtC as CO2 over
10kyrs) and largest for e=10 GtC/yr (10,000 GtC as CO2 over 1kyr). This difference was not
solely due to the addition of more total CO2; at higher eruption rates, a larger fraction of added
CO2 remained in the model atmosphere at the end of eruption (Table 2, for low initial pH, 300
ppm initial pCO2 setups). At lower eruption rates, the model oceans were better able to keep pace
with CO2 emissions, taking up a larger fraction of the additional CO2 even during the duration of
eruption. For the low eruption rate experiments (e = 0.1 GtC/yr), pCO2 and Ωcalcite changed by
more in models with reef carbonate production than in those with pelagic carbonate production,
similar to the models with e = 1 GtC/yr.
For high eruption rates (e = 10 GtC/yr), the pCO2 and Ωcalcite response was virtually even
across pelagic and reef setups (Figure 4) – likely because the brief, 1 kyr eruption duration may
have been shorter than the response time of carbonate dissolution in the oceans, thereby
32
preventing the carbonate system from buffering the CO2 significantly during eruption (see 4.1 for
more detail). Thus, we see multiple ways in which the timescales of the processes that remove
CO2 from the atmosphere influence not only the timing of recovery, but also the maximum
magnitude of change in response to perturbation.

4.4 WARM INITIAL CLIMATE
We also observed differences in model results depending on initial atmospheric pCO2.
For initial pCO2 of 2000 ppm, >37% of the released CO2 remained in the atmosphere at the end
of eruptions, compared to <20% for initial pCO2 of 300 ppm, and the half-life of atmospheric
CO2 was 2-9kyr longer for higher initial pCO2 scenarios (see Table 2 for more results). In all
cases, global mean sea-air CO2 fluxes were negative during and after eruption, reflecting transfer
of CO2 from the atmosphere to the oceans (Figure 8). For 2000 ppm initial pCO2, the fluxes were
less negative, indicating less CO2 uptake from the atmosphere. With higher initial pCO2, the
oceans had higher initial DIC concentration, making them highly saturated with carbonate – so
when additional CO2 was added to the atmosphere, the ocean could not absorb CO2 as quickly,
leading to less transfer of CO2 from air to sea, a higher fraction of added CO2 in the atmosphere,
and a longer recovery duration. Though ΔCO2 was larger for runs with an initial pCO2 of 2000
ppm, global temperatures increased by less in the lower initial pCO2 models (~2°C for high
pCO2 scenarios vs. 5-9°C for the low pCO2 scenarios), because the global temperature response
is related exponentially to atmospheric pCO2.

33
4.5 A COOL, OVERSATURATED END-PERMIAN
Generally, the results over the experimental suite (Figure 4, Table 2) indicate that oceans
with reef-bound carbonate production are less able to buffer injected CO2, leading to increased
environmental impact. The exception to this trend is the experiment run using end-Permian
geography, low initial CO2 and high ocean pH. The end-Permian low initial pCO2, high pH run
with reef carbonates saw a ΔpCO2 = 525 ppm and ΔΩcalcite=-5.5, while with pelagic carbonate
production, these values were ΔpCO2 = 690 ppm and ΔΩcalcite=-7.5. Despite having a highly
oversaturated ocean (initial average surface ocean Ωcalcite = 17.8) the model deposited CaCO3
across less surface area (and therefore less overall CaCO3 deposition) than the end-Permian low
initial pCO2, low pH run with pelagic carbonate production (Figure 9). In order to confirm that
these results were not spurious, an additional run was executed with slightly different target
ocean chemistry (initial Ωcalcite =17.9), with similar results (ΔpCO2 = 614 ppm and ΔΩcalcite=-6.8,
Figure 9). The spin-up methodology followed to achieve steady states with pre-selected pH
values in the case of the end-Permian low initial pCO2, high pH scenario requires possibly
unrealistically low carbonate deposition to maintain the target ocean pH of 8.4 under the
weathering rates reasonable under a 300 ppm pCO2 value. Because of this, when the ‘eruption’
CO2 is injected, the model shifts towards deposition of carbonate over greater surface area
(Figure 9); this deposition removes carbonate ions from the model ocean and pushes the ocean
carbonate inventory towards CO2,aq, preventing the ocean from taking up as much CO2 from the
atmosphere as in other runs. While the surface area over which the modern geography, low
initial pCO2, high pH with pelagic carbonate production deposits CaCO3 is greater than its end-
Permian counterpart, it is still less than in the modern geography low initial pCO2 low pH run,
indicating that the weathering rate used to spin-up the model remained insufficient to maintain
34
the target pH without impacting modeled carbonate deposition. This highlights the importance of
both accurate weathering rate estimates and accurate pH estimates if trying to apply model
experiments such as these to specific time periods or events.

4.6 PALEOBATHYMETRY
Another area of uncertainty when trying to model past Earth systems states is paleo-
bathymetry. While general ocean basin shape is well-reconstructed in tandem with continental
configuration, information on age and depth of past oceanic plates is often lost through
subduction. The prescribed bathymetry of the ‘p0251a’ configurations is artificially flat due to
the lack of constraints on paleo-bathymetry; outside of shallow seas and continental shelves, the
bathymetry was set to a uniform depth of 3575 m, the second-deepest possible ocean depth level
in cGENIE. In order to determine whether the prescribed bathymetry substantially alters the
modeled ocean’s carbonate sink, a randomized bathymetry was generated for the end-Permian
continental configuration: the locations of deeper or more shallow ocean boxes was random, but
the overall hypsometric curve of the ocean basins was set to be similar to the modern
hypsometry. Using this test bathymetry, the modified configuration was spun up under 300 ppm
pCO2 atmosphere, to a relatively less saturated ocean and 10,000 GtC as CO2 injected over
10kyrs. The results of this run are presented in Figure 10. The evolution of pCO2, calcite
saturation state, temperature, and DIC flux are similar to the standard end-Permian configuration,
suggesting that bathymetry is not the primary control on patterns observed in this study.
However, the results do differ in some second order features (e.g., saturation state during mid-
recovery, and absolute values), highlighting the need for better constrained paleobathymetric
estimates for modeling the carbon cycle.
35
4.7 REEF CARBONATE DEPOSITION PARAMETERIZATION
The parametrization of reef carbonate deposition within model box i in cGENIE is:
FCaCO3=c*(Ω-1)
n
[Eq. 1]
Where c is a scaling factor and n is the order of reaction. The scaling factor is tuned to bring the
carbonate deposition flux into balance with continental weathering during the spin-up phase. The
‘process’ power, n, of this equation was chosen to be 1.0 in accordance with previous studies
using cGENIE for the end-Permian (Cui et al. 2015), but a value of 1.7 has also been suggested
(Ridgwell 2005, Cui et al. 2015, Opdyke & Wilkinson 1993). To test the impact of the chosen
reef carbonate formulation, two runs with n = 1.7 were conducted, one in the worbe2-b setup and
one in the p0251a-b setup. Each run was spun up under a 300 ppm pCO2 atmosphere with
relatively less saturated ocean with respect to calcium carbonate and the ‘standard’ eruption was
injected. These results differ from our main experimental suite; they show lower peak CO2 and
temperature changes. However, in looking at the weathering and sedimentation fluxes from these
runs (Figure 11a &b) it is not apparent that these differences are caused by an increased ability of
the model oceans to buffer against increasing pCO2 due to increased carbonate deposition.
Instead it appears that the high weathering fluxes required to bring these model runs into steady
state with such a large carbonate deposition flux allows for buffering to occur. For n = 1.7,
initial weathering flux must be very high. For comparison, modern total weathering flux is
~2.5*10
13
mols C/year (Gaillardet et al. 1999), similar (within a factor of 2) to the initial
weathering flux in other model runs. In contrast, the model run with n = 1.7 requires weathering
flux nearly 10 times higher. In turn, this would require a very high degassing flux to balance the
weathering drawdown. There is no evidence for such high fluxes, suggesting that these model
scenarios are unreasonable.
36
CHAPTER FIVE: CONCLUSIONS
Overall, our results show how many factors influence Earth-system response to the
injection of CO2 on the scale of a pulse of LIP eruption, including pre-eruption climate and
ocean chemistry, paleo-bathymetry, and eruption rate. Yet our results also reveal that the locus
and method of carbonate formation and burial exerts a primary, first-order control. Under a
pelagic carbonate regime, such as the one present on Earth since the evolution of planktonic
calcifiers in the Mesozoic, the Earth’s ocean carbonate system is able to buffer CO2 release
significantly more efficiently than under a reef carbonate regime. This increased buffering
mediates environmental impacts, suggesting that this major step in evolution was a key factor
explaining why there were fewer mass extinctions events over the last 150 Ma as compared to
the rest of the Phanerozoic.

37
CHAPTER 6: SUPPLEMENT: UNFINISHED BUSINESS - WHY DON’T OCEANIC
PLATEAU ERUPTIONS CAUSE MASS EXTINCTIONS?
6.1 INTRODUCTION
The Cretaceous oceanic plateaus (OP) such as Ontong Java, Kerguelen and the
Caribbean/Colombian Igneous Province are a puzzling anomaly when looking at the documented
links between LIP eruptions and environmental catastrophe throughout Earth history (Wignall
2005, Bond & Wignall 2014, Courtillot & Fluteau 2014, Bond & Grasby 2017). All are large
LIPs (in fact, Ontong Java is the largest expression of volcanism currently known of on Earth)
that appear to have had relatively less severe effects than to be expected for their size. Unlike
prominent continental flood basalts such as the Siberian Trap or the Central Atlantic Magmatic
Province which have been linked to global warming, ocean acidification, ocean anoxia (in the
case of the Siberian Traps) and two of the largest mass extinctions in Earth history (Alroy 2010,
Bond & Grasby 2017), the Cretaceous plateaus have only been to ocean anoxic events, and
smaller associated environmental disturbances which did not result in mass extinctions (Bond
and Wignall 2014).
6.1.1 LINKS BETWEEN OP ERUPTIONS AND OCEAN ANOXIC EVENTS
Several geochemical links have been found that tie the eruptions of oceanic plateaus to
the onset and continuation of ocean anoxic events in the sedimentary record. Work done on the
black shales of OAE1a in the early Aptian in Italy and the Pacific shows that the
187
Os/
188
Os
ratio of marine sediments sharply declines at the beginning of the stratigraphic level that defines
OAE1a, indicating a large input of mantle Os as one would expect with an oceanic plateau
eruption (Tejada 2009, Bottini et al. 2012). Pb-isotope evidence from sections in the deep Pacific
tell a similar story (Kuroda et al. 2015). There has also been recorded spikes in trace metals
38
interpreted to be related to Ontong Java volcanism in cores in the Pacific near the location of the
Ontong Java Plateau (Erba et al. 2015).
Similar Os-isotope and Pb-isotope evidence ties the Turonian Caribbean/Colombian
igneous province to OAE2 (Kuroda et al. 2007, Du Vivier et al. 2015) in addition to Nd records
which indicate a volcanogenic spike in Nd concentrations as well as regional changes to ocean
circulation near the study site in England (Zheng et al. 2013, 2016).

6.2 INTERACTIONS BETWEEN OP ERUPTIONS AND THE GLOBAL
ENVIRONMENT
As these Cretaceous plateaus are some of the largest LIPs in Earth history volumetrically
(Bryan & Ernst 2008, Bong & Wignall 2014, Bond & Grasby 2017)—the Ontong Java Plateau
alone is estimated to have had an initial volume greater than 10 million km
3
—the difference in
environmental response to these oceanic plateau eruptions is likely not just a matter of eruption
size. Some important differences may include:
(1) The mechanisms by which oceanic plateau eruptions interact with the global environment
are different than those in play with continental flood basalt eruptions. Volatile fluxes
(e.g. CO2) may be significantly lower in oceanic plateau eruptions due to the high
pressure under which they are erupted (Gregg 2013), which may reduce the degree of
global warming and ocean acidification possible during these events.
(2) Interactions between water and erupted material are far more consistent and active in
oceanic plateau eruptions than in continental flood basalts. These water-rock interactions
may lead to the release of large amounts of biologically active trace metals such as iron
39
into the water column, a factor which could lead to increased productivity and push the
oceans towards anoxic conditions (Kerr 2013 & refs therein, Erba et al. 2015).
(3) Additionally, the mantle plumes that are thought to feed these LIP events (Campbell
and Griffiths 1990, Ernst and Buchan 2001, Campbell 2005, Sobolev, et al. 2011)
should cause uplift of the seafloor which, combined with the emplacement of lavas
during eruption could displace seawater and cause sea-level rise as well as interruptions
to existing patterns of ocean circulation (Kerr 2013).
These later two eruption-environment interactions could push the Earth system towards
anoxic conditions. Another possible interaction that could do so is the heat flux from the oceanic
plateau lavas as they cool on the seafloor which could potentially heat ocean bottom water and
disrupt circulation patterns. The model experiment runs detailed herein were developed to test
whether this last possible eruption-environment interaction is feasible.

6.3 ESTIMATING HEAT FLUX FROM THE ONTONG-JAVA PLATEAU
Fresh basalt is rapidly (0.7 s) quenched when erupted into water, forming a glassy rind.
Therefore, from 0.7 s onwards, all cooling of erupted lava is done via conduction (Gregg 2013,
Harris 2013, Palmer & Ernst 1998). Therefore the heat flux from erupted oceanic plateau lavas in
contact with seawater can be modeled according to the equation:
q = -
!∆#
$∗&∗'

describing the conductive heat flux of a molten lava core through a progressively thickening
solid layer wherein k and κ are the thermal conductivity (1.69 W/mK) and diffusivity (5.34 x 10
-
40
7
m
2
/s) of basalt respectively, ∆T is the difference between lava core temperature (1100 deg C;
Gregg 2013) and the temperature of the water in contact with the lava (4 deg C) and t is the time,
in seconds, since eruption (Harris, 2013).
In order to determine what a geologically reasonable mean heat flux value for the entire
erupting OJP would be (within cGENIE, OJP occupies only one entire grid box), we look at the
flow history of the Columbia River Flood Basalts (with the caveat that the eruption tempo of a
continental flood basalt is likely to be quite different from that of an oceanic plateau) which are
well studied and age dated. Specifically, we use the eruption history and stratigraphy of the
Grande Ronde Basalt member of the CRFB, as it is the most voluminous. Since we are
concerned with the heat flux ‘up’ from the surface area of the erupting OJP, we use the surface
area of the Grande Ronde Basalt, which is 169,600 square km. Using the available age dating
and stratigraphic sequences (including estimates of numbers of individual flows within a unit),
we look at several snapshots of the eruption history of the Grande Ronde Basalt, and use these
snapshot surface areas to calculate a surface area weighted mean heat flux as if that basalt area
had been erupted at the OJP location. Snapshot periods were selected based on the formations
that had the best age dating (Kasbohm & Schoene 2018 for dates, Reidel & Tolan 2013 for
geologic details)
o Snapshot 1: Sentinel Bluffs
§ Ages: 16.052 Ma (lower) and 16.044 Ma (upper) are taken to be the starting and
ending dates of this formation for this exercise.
§ Between 10-15 flows are estimated to make up Sentinel Bluffs.
§ Flows are assumed to be equal in size and erupted at a constant tempo for lack
of further constraints.
41
§ 15 flow scenario: a flow of 11,180 sq km every 571 years for the 8000
year duration of the formation’s eruption
§ 10 flow scenario: a flow of 16,770 sq km evert 889 years
§ Surface Area: 167,600 square km
§ The remaining 1900 square km of Grande Ronde Basalt’s total surface area
was assumed to be covered by the uppermost flow of the Winter Water
formation. Using the same logic as above to space out the Winter Water flows
based on the latest available age date (Fields Spring 16.168 Ma), this oldest
1900 sq km is set to be 37,000 years old relative to the youngest flow of
Sentinel Bluffs (age = 0 for this exercise).
o Snapshot 2: Wapshilla Ridge
§ Ages: 16.306 Ma (lower) and 16.232 Ma (upper) are taken to be the starting and
ending dates of this formation for this exercise.
§ 18 flows are estimated to make up Wapshilla Ridge
§ Flows are assumed to be equal in size, and erupted at a constant tempo for
lack of further constraints. A flow of 8251 sq km every 4353 years for the
74,000 year duration of the formation’s eruption
§ Surface Area: 149,050 square km
§ The remaining 20,500 sq km of Grande Ronde Basalt’s total surface area was
assumed to be covered by the uppermost flow of the Mount Horrible
formation. Age of 83,716 years relative to the youngest flow of Wapshilla
(age = 0 for this exercise).
42
Using the above snapshots and the heat flux equations above, we find the mean heat flux
(as weighted by surface area) of each snapshot to be:
• Sentinel Bluffs, 15 flows: 26.4 W/m
2

• Sentinel Bluffs, 10 flows: 36.4 W/m
2

• Wapshilla Ridge: 17.3 W/m
2

6.4 EXPERIMENT (RE)DESIGN
To test whether the heat flux from an erupting oceanic plateau is sufficient to
substantially alter the heat content of the oceans and/or disturb existing patterns of ocean
circulation, cGENIE, an Earth systems model of intermediate complexity, would be great to use.
The model was modified to allow for a spatially resolved geothermal heat flux that flows from
the seafloor into the bottom-most ocean grid boxes.
Two geographic configurations were selected: worjh2 a modern geography with a 16-
depth level ocean (Ridgwell et al. 2007) and p0093k a Turonian geography with a 16-level
ocean. Utilizing two geographies would allow for testing whether any thermal anomalies caused
by plateau eruption may have had different impacts at times in Earth history such as the
Cretaceous, when sea-level was higher and the position of continents was significantly altered
from the modern. Both configurations of the model would be spun-up under a 300 ppm CO2
atmosphere until steady state between weathering input to the oceans and carbonate burial in
reefs and deep-sea sediments was achieved (25,000 years) with the geothermal flux set to be 0.1
W/m
2
in every grid square.
For the experiment runs, to simulate an active plateau within cGENIE, the geothermal
flux in a grid box of one’s choice (OJP is interesting cause it erupted pretty much smack dab in
43
the middle of the open Pacific, Caribbean Plateau is interesting because it erupted very close to
the at-the-time open seaway between North and South America, NAIP is interesting because it
was also on a mid-ocean ridge, take your pick!) should be set to one of the above estimates of
heat flux. A smart design would be to run 3 different plateau scenarios in each configuration!

44
LIP (age) Global
Warming?
Ocean
anoxia?
Ocean
Acidification?
Extinction Event?
Viluy (376 Ma) CFB Cooling

Late Devonian (Big
5)
Emeishan (258 Ma) CFB ?

? end-Guadalupian
Siberian (250 Ma) CFB

end-Permian (Big 5)
CAMP (201 Ma) CFB

end-Triassic (Big 5)
Karoo-Ferrar (180 Ma) CFB

Toarcian
Parana-Etendeka (133 Ma)
CFB

Ontong Java (120 Ma) OP

Caribbean-Colombian,
Madagascar (90 Ma) OP

Deccan (65 Ma) CFB

end-Cretaceous (Big
5)
NAIP (55 Ma) CFB

Table 1: Summary of the major LIPs of the Phanerozoic and the environmental impacts they
have been correlated with in the rock record. Red indicates no evidence of that effect occurring
contemporaneously with the LIP eruption, yellow indicates some evidence, blue indicates clear
correlation. Modified and expanded from Wignall 2005, Bond & Wignall 2014. Data for Viluy:
Ricci et al. 2013.

45

Figure 1: Number of major Phanerozoic LIPs associated with mass extinctions and three major
global environmental perturbations often linked to extinction events. Dark shades are links with
solid evidence, light are cases with some evidence (Wignall 2005, Bond and Wignall 2014,
Courtillot and Fluteau 2014, Bond and Grasby 2017)

46

Figure 2: Schematic of experiment model runs. A total of 24 runs were completed spanning this
range, plus additional tests in order to determine the effect of changing the end-Permian solar
flux, end-Permian bathymetry and the model parameterization of reef calcification (all of these
tests were run for an initial pCO2 of 300 ppm and “low” pH, 10,000 GtC was injected over
10,000 years).

47

Figure 3: (Top) The model evolution of RCO2 (pCO2/pCO2,t=0) in response to the injection of
10,000 GtC as CO2 over 10,000 years into pre-injection atmospheres of 300 ppm CO2 and a
relatively less well saturated ocean. (Bottom) The modeled change in calcite saturation state
within the top 176.4m of the global ocean in response to the same eruption.

48

Figure 4: (Top): Comparison of change in pCO2 (DpCO2) across the experimental suite.
(Bottom): Comparison of change in average surface ocean calcite saturation state (Ωcalcite) across
experimental suite. In all but one case, reef calcification leads to a system more sensitive to CO2
injection, with little difference by paleogeography. In the case of the one exception (end-Permian
geography, 300 ppm initial pCO2, high initial ocean pH setup), carbonate deposition is
anomalous (see text).

49

50
Table 2: Key results from all model runs. See text for discussion.
+: GEO= Geographic configuration used in model run; M= Modern continental configuration,
P=end-Permian continental configuration. CARB= Mode of carbonate deposition; PEL=
carbonate created in the open ocean by planktonic calcifiers and buried in sediments, REEF=
carbonate created in reefs and buried in sediments. PCO2,INIT= Model run’s initial steady-state
pCO2. PHINIT= Model run’s initial steady state pH; these pH’s vary between 300 ppm (LOW
=8.2, HIGH = 8.4) and 2000 ppm (LOW=7.7, HIGH=7.9) initial pCO2 runs, but in all cases
LOW pH runs were spun up to have an initial average surface ocean WCALCITE of ~7, HIGH pH
runs were spin up to have an initial average surface ocean WCALCITE of ~17. GTC= Amount of
carbon injected into the model as CO2 in gigatons. YRS= Duration of injection; in all cases the
total amount of carbon was injected evenly over the duration (so for 10,000 GtC over 10,000
years, the rate was a constant 1 GtC/yr).
*The model run P/PEL/300/HIGH/10k/10k exhibits abnormal model response to CO2 injection
due to the required ocean chemistry during spin up phase. For more detail and discussion, see
section 4.5.

51

Figure 5: (Top) Modeled global flux of DIC (i.e., CaCO3 dissolution) from sediments to the
ocean over the course of CO2 injection. Model runs with reef carbonate factories have lower
global DIC flux throughout model runs due to the smaller area of carbonate deposition when
carbonate formation is restricted to shallow depths. The Modern and end-Permian continental
configurations differ in steady state (i.e., different starting values for DIC flux) due to the
differences in ocean chemistry and weathering fluxes that derive from the end-Permian
configuration being cooler due to decreased solar flux. (Bottom) Flux of DIC from model
sediments to model ocean normalized in each run by the initial value of that flux. Normalizing
these fluxes isolates the system response to CO2 injection from initial differences in DIC flux
due to area of carbonate deposition. In the model runs with reef carbonate factories, the rate of
CaCO3 dissolution increases more than the pelagic carbonate runs due to a greater drop in
saturation state; however, due to the deficit in carbonate depositional area, this dissolution is
insufficient to buffer the injected CO2 (see figures 3 and 4).

DIC flux from sediment to ocean
(10
14
mol/yr)
0.4
0.8
1.2
1.4
DIC flux from sediment to ocean,
normalized by initial value
1.4
1.0
0.6
10 kyrs 20 kyrs 30 kyrs 40 kyrs 50 kyrs
Modern Geography, Planktonic Calcifiers
Modern Geography, Reef Calcifiers
end-Permian Geography, Planktonic Calcifiers
end-Permian Geography, Reef Calcifiers
52

Figure 6: Average global land surface temperature over the course of CO2 injection experiments
(top); difference of global land surface temperature from pre-eruption steady state land
temperature (bottom). The model runs with end-Permian configuration begin at much lower
surface land temperatures due to the decreased solar flux 250 Myr ago. It is clear in these figures
that these end-Permian model runs see a much higher increase in surface land temperature than
their modern configuration counterparts over the course of CO2 injection, potentially due to the
presence of the super continent. This larger increase is reflected in larger weathering rates
throughout the model runs, leading to the observed faster recovery times noted in the text. It is
worth noting that when comparing the planktonic and reef carbonate regimes within continental
configurations that the reef carbonate factory runs see greater temperature change than their
planktonic counterparts; this is due to the increased DpCO2 of these runs.

53

Figure 7: Test of solar flux vs. continental configuration in determining distinct weathering
response for end-Permian configurations. Plots show model results of 10,000 GtC as CO2
injected over 10,000 years in 3 scenarios, all with planktonic carbonate factories. Shown are
responses of (A) atmospheric pCO2, (B) surface ocean calcite saturation state, (C) average global
air temperature, (D) DIC flux from sediments to the oceans, all as change from initial values.
The end-Permian geography with Modern Solar Flux (red dashed line) exhibits similar behavior
to the end-Permian geography with end-Permian solar flux, albeit at a smaller magnitude (except
in the case of DWCALCITE ). However, the behavior still fits the pattern observed in the text of end-
Permian geographies having larger temperature increases and faster amelioration time-scales
than the modern geographies. This implies that the nature of air temperature and weathering over
a supercontinent differs from over smaller, disparate continental masses such as on the modern
Earth.

54

Figure 8: Flux of CO2 from the oceans to the atmosphere for injection experiments with
different initial pCO2. All runs plotted here have a pelagic carbonate factory. Negative values
indicate net flux of CO2 is invading the oceans from the atmosphere. The runs with an initial
pCO2 of 2000 ppm exhibit lower rates of CO2 invasion of the ocean from the atmosphere than
their 300 ppm initial pCO2 counterparts. As noted in the text, this is due to the fact that under
such high initial pCO2, the oceans are already highly oversaturated with carbonate, meaning the
oceans have ‘less room’ to take up additional CO2 from the atmosphere.

55

Figure 9: (A) Time-series results for two model runs conducted with end-Permian geography,
300 ppm initial pCO2, high initial pH and planktonic calcifiers (solid lines; hence-forth referred
to as ‘high pH 1 &2’) compared to the end-Permian, 300 ppm initial pCO2, low initial pH with
planktonic calcifiers model run (dashed, henceforth ‘low pH’). High pH 1 is the only run of our
model suite that does not follow the observed pattern of runs with planktonic calcifiers more
effectively buffering the addition of CO2. High-pH 2 was run with slightly different starting
ocean chemistry, and exhibits the same behavior, indicating that the results are not solely due to
anomalous initial conditions. As you can see in the top right panel, unlike the low pH run, the
high pH runs exhibit a sediment-to-ocean DIC flux that becomes negative over the course of the
model experiment, in this case indicating that CO2 injection promotes CaCO3 deposition instead
of dissolution. (B) Spatial results for the %CaCO3 in ocean sediments for low pH and high pH 1
runs. CaCO3 is preserved over far less area in the high pH 1 run at time zero than in the low pH
run. This is a consequence of requiring a highly oversaturated ocean with respect to CaCO3
under a 300 ppm pCO2 atmosphere with a weathering flux insufficient to maintain the required
pH. In order to meet the carbonate chemistry conditions, the model is unable to preserve CaCO3
over much of the model ocean in steady state. The injection of CO2 then pushes the system
towards increased carbonate deposition (see 4.5 text for more detail).

56

Figure 10: Model results for end-Permian CO2 injection experiment with random bathymetry
compared to end-Permian, planktonic calcifier run with 300 ppm initial pCO2, low initial pH and
to the modern, planktonic calcifier run with 300 ppm initial pCO2, low initial pH runs. 10,000
GtC as CO2 was injected over 10kyr in all runs. In general, the randomized bathymetry yields
similar results to the standard end-Permian configuration, with slight difference in WCALCITE
during middle phases of the recovery.

57

Figure 11a: Comparison of modern geography runs with: planktonic calcifiers (blue) and reef
calcification (red) in the 300ppm initial pCO2, low pH scenarios to a run with a modern
geography, reef calcification with exponent value n =1.7 (other reef run n =1.0; see methods).
10,000 GtC injected as CO2 over 10kyr in each experiment. As stated in the main text, the runs
with scaling factor n = 1.7 require unreasonably high initial weathering fluxes, which then allow
for buffering of CO2 injected into the atmosphere.

58

Figure 11b: Comparison of end-Permian Geography runs with: planktonic calcifiers (blue) and
reef calcification (red) in the 300ppm initial pCO2, low pH scenarios to a run with a modern
geography, reef calcification with exponent value n =1.7 (other reef run n =1.0; see methods).
10,000 GtC injected as CO2 over 10kyr in each experiment. As stated in the main text, the runs
with scaling factor n = 1.7 require unreasonably high initial weathering fluxes, which then allow
for buffering of CO2 injected into the atmosphere.

59
REFERENCES
Alroy, John. "The shifting balance of diversity among major marine animal
groups." Science 329.5996 (2010): 1191-1194.

Beerling, D., & Berner, R. (2002). Biogeochemical constraints on the Triassic-Jurassic boundary
carbon cycle event. Global Biogeochemical Cycles, 16.

Bergman, N., Lenton, T., & Watson, A. (2004). COPSE: a new model of biogeochemical cycling
over Phanerozoic time. American Journal of Science, 304, 397-437.

Berner, R. (1991). A model for atmospheric CO2 over Phanerozoic time. American Journal of
Science, 339-376.

Berner, R. (1994). GEOCARB II: A Revised model of atmospheric CO2 over phanerozoic time.
American Journal of Science, 56-91.

Berner, R. (1998). The carbon cycle and CO2 over Phanerozoic time: the role of land plants.
Philospophical Transactions of the Royal Society, 75-82.

Berner, R. (2006). GEOCARBSULF: A combined model for Phanerozoic atmospheric O2 and
CO2. Geochimica et Cosmochimica Acta, 70, 5653-5664.

Berner, R., & Beerling, D. (2007). Volcanic degassing necessary to produce a CaCO3
undersaturated ocean at the Triassic-Jurassic boundary. Palaeogeography, Palaeoclimatology,
Palaeoecology, 244, 368-373.

Berner, R., & Caldeira, K. (1997). The need for mass balance and feedback in the geochemical
carbon cycle. Geology, 955-956.

Berner, R., & Kothavala, Z. (2001). GEOCARB III: A revised model of atmosperic CO2 over
Phanerozoic time. American Journal of Science, 301, 182-204.

Berner, R., Lasaga, A., & Garrels, R. (1983). The carbonate-silicate geochemical cycle and its
effect on atmospheric carbon dioxide over the past 100 million years. American Journal of
Science, 641-683.

Black, B., Hauri, E., Elkins-Tanton, L., & Kiehl, J. (2014a). Acid rain and ozone depletion from
pulsed Siberian Traps magmatism. Geology, 67-70.

Blackburn, T., Olsen, P., Bowring, S., McLean, N., D.V., K., Puffer, J., . . . Et-Touhami, M.
(2013). Zircon U-Pb geochronology links the end-Triassic extinction with the Central Atlantic
Magmatic Province. Science, 941-945.

Bond, D.P.G., Grasby, S.E. (2017) On the causes of mass extinctions. Palaeogeogr.
Palaeoclimatol. Palaeoecol. 478, 3–29.

60
Bond, D., & Wignall, P. (2014). Large igneous provinces and mass extinctions: an update. The
Geological Society of America Special Paper 505, 29-55.

Bottini, Cinzia, et al. "Osmium-isotope evidence for volcanism, weathering, and ocean mixing
during the early Aptian OAE 1a." Geology 40.7 (2012): 583-586.

Briffa, K., Jones, P., Schweingruber, F., & Osborn, T. (1998). Influence of volcanic eruptions on
Northern Hemisphere summer temperature over the past 600 years. Nature, 450-455.

Bryan, S. (2007). Silicic large igneous provinces. Episodes, 20-31.

Bryan, S., & Ernst, R. (2008). Revised definition of large igneous provinces (LIPs). Earth-Science
Reviews, 175-202.

Caldeira, K., & Wickett, M. (2003). Oceanography: Anthropogenic carbon and ocean pH. Nature,
365.

Callegaro, S., Baker, D., De Min, A., Marzoli, A., Geraki, K., Bertrand, H., . . . Nestolda, F. (2014).
Microanalyses link sulfur from large igneous provinces and Mesozoic mass extinctions. Geology,
895-898.

Campbell, I. (2005). Large Igneous Provinces and the mantle plume hypothesis. Elements, 265
269.

Campbell, I., & Griffiths, R. (1990). Implications of mantle plume structure for the evolution of
flood basalts. Earth and Planetary Science Letters, 99, 79-93. doi:10.1016/0012-821X(90)90072
6

Cao, L., Eby, M., Ridgwell, A., Caldeira, K., Archer, D., Ishida, A., . . . Orr, J. (2009). The role of
ocean transport in the uptake of anthropogenic CO2. Biogeosciences, 6, 375-390.

Chenet, A., Fluteau, F., Courtillot, V., Gerard, M., & Subbarao, K. (2008). Determination of rapid
Deccan eruptions across the Cretaceous-Tertiary boundary using paleomagnetic secular variation:
results from a 1200-m thick section in the Mahabaleshwar escarpment. Journal of Geophsycical
Research-Solid Earth.

Chenet, A., Fluteau, F., Courtillot, V., Gerard, M., Quidelleur, X., Khadri, S., . . . Thordarson, T.
(2009). Determination of rapid Deccan eruptions across the Cretaceous-Tertiary boundary using
paleomagnetic secular variation: 2. Constraints from analysis of eight new sections and synthesis
for a 3500-m thick composite section. Journal of Geophysical Research-Solid Earth.

Coulbourn, G., Ridgwell, A., & Lenton, T. (2013). The Rock Geochemical Model (RokGeM)
v.0.9. Geoscientific Model Development, 6, 1543-1573.

Courtillot, V. & Fluteau, F. (2014). The Geological Society of America Special Paper 505, p.301-
317
61
Courtillot, V., & Renne, P. (2003). On the ages of flood basalt events. Comples Rendus Geosoence,
113-140.

Cui, Y., & Kump, L. (2014). Global warming and the end-Permian mass extinction event: Proxy
and modeling perspectives. Earth Science Review.

Cui, Y., Kump, L., & Ridgwell, A. (2013). Initial assesment of the carbon emission rate and
climatic consequences during the end-Permian mass extinction. Palaeogeography,
Palaeoclimatology, Palaeoecology, 389, 128-136.

Cui, Y., Kump, L. (2015) Global warming and the end-Permian extinction event: Proxy and
modeling perspectives. Earth-Science Rev. 149, 5–22.

de Silva, S., & Zielinski, G. (1998). Global influence of the AD 1600 eruption of Huaynaputina,
Peru. Nature, 455-458.

Dessert, C., Dupre, B., Gaillardet, J., Francois, L., & Allegre, C. (2003). Basalt weathering laws
and the impact on the global carbon cycle. Chemical Geology, 202, 257-273.

Donnadieu, Y., Godderis, Y., Pierrehumbert, R., Dromart, G., Fluteau, F., & Jacob, R. (2006b). A
GEOCLIM simulation of climatic and biogeochemical consequences of Pangea breakup.
Geochemistry, Geophysics, Geosystems, 7(11).

Donnadieu, Y., Pierrehumbert, R., Jacob, R., & Fluteau, F. (2006a). Modelling the primary control
of paleogeography on Cretaceous climate. Earth and Planetary Science Letters, 248, 426-437.

Du Vivier, A. D. C., et al. "Pacific 187 Os/188 Os isotope chemistry and U–Pb geochronology:
Synchroneity of global Os isotope change across OAE 2." Earth and Planetary Science
Letters 428 (2015): 204-216.
Edwards, N., & Marsh, R. (2005). Uncertainties fue to transport-parameter sensitivity in an
efficient 3-D ocean-climate model. Climate Dynamics, 24, 415-433.

Eichenseer, K., Balthasar, U., Smart, C.W., Stander, J., Haaga, K.A., Kiessling, W. (2019)
Jurassic shift from abiotic to biotic control on marine ecological success. Nat. Geosci. 12, 638–
642.

Erba, Elisabetta, et al. "Environmental consequences of Ontong Java Plateau and Kerguelen
Plateau volcanism." The origin, evolution, and environmental impact of oceanic large igneous
provinces. Geological Society of America Special Paper 511 (2015): 271-303.
Ernst, R., & Buchan, K. (2001). Large mafic magmatic events through time and links to mantle
plume heads. Geological Society of America Special Paper 352, 483-575.
Francois, L., & Walker, J. (1992). Modeling the Phanerozoic carbon cycle and climate
constraints from the 87Sr/86Sr isotopic ratio of seawater. American Journal of Science, 81-135.

62
Gaillardet, J., Dupré, B., Louvat, P., Allègre, C.J. (1999) Global silicate weathering and CO2
consumption rates deduced from the chemistry of large rivers. Chem. Geol. 159, 3–30.

Ganino, C., & Arndt, N. (2009). Climate changes caused by degassing of sediments during the
emplacement of large igneous provinces. Geology, 323-326.

Godderis, Y., & Joachimski, M. (2004). Global change in the Late Devonian: modelling the
Frasnian-Famennian short-term carbon isotope excursions. Palaeogeography, Palaeoclimatoligy,
Palaeoecology, 202, 309-329.

Goodwin, P., & Ridgwell, A. (2010). Ocean-atmosphere partitioning of anthropogenic carbon
dioxide on multimillennial timescales. Global Biogeochemical Cycles, 24.

Goodwin, P., Williams, M., Follows, M., & Dutkiewicz, S. (2007). Ocean-atmosphere
partitioning of anthropogenic carbon dioxide on centennial timescales. Global Biogeochemical
Cycles.

Greene, A., Soates, J., & Weis, D. (2008). Wrangellia flood basalts in Alaska: A record of
plume-lithosphere interaction in a Late-Triassic oceanic plateau. Geochemistry, Geophysics,
Geosystems.

Greene, S., Martindale, R., Ritterbush, K., Bottjer, D., Corsetti, F., & Berelson, W. (2012).
Recognizing ocean acidification in deep time: An evalutation of the evidence for acidification
across the Triassic-Jurassic boundary. Earth-Science Reviews, 72-93.

Gregg, T. K. P. “Deep-sea eruption.” Modeling volcanic processes: the physics and mathematics
of volcanism. Cambridge University Press, New York (2013): 258-274.

Harris, A. J. L. "Lava flows." Modeling volcanic processes: the physics and mathematics of
volcanism. Cambridge University Press, New York (2013): 85-106.

Hautmann, M., Benton, M. J.,Tomašových, A. (2008) Catastrophic ocean acidification at the
Triassic-Jurassic boundary. Neues Jahrb. für Geol. und Paläontologie - Abhandlungen. 249,
119–127.

Held, I.M., Soden, B.J. (2006) Robust Responses of the Hydrological Cycle to Global Warming.
Journal of Climate, 19, 5686-5699

Hesselbo, S., Robinson, S., Surlyk, F., & Piasecki, S. (2002). Terrestrial and marine extinction at
the Triassic-Jurassic boundary synchronized with major carbon-cycle perturbation: A link to
initiation of massive volcanism? Geology, 30, 251-254.

Honisch, B., Ridgwell, A., Schmidt, D., Thomas, E., Gibbs, S., Sluijs, A., . . . Williams, B.
(2012). The geological record of ocean acidification. Science, 335, 1058-1063.

Huynh, T., & Poulson, C. (2005). Rising atmospheric CO2 as a possible trigger for the end
63
Triassic mass extinction. Palaeogeography, Palaeoclimatology, Palaeoecology, 217, 223-242.

IPCC, 2014: Climate Change 2014: Synthesis Report. Contribution of Working Groups I,II and
III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Core
Writing Team, R.K. Pachauri and L.A. Meyer (eds.)]. IPCC, Geneva, Switzerland, 151 pp.

Jacob, R. (1997). Low frequency variability in simulated atmosphere ocean system . Type Thesis,
University of Wisconsin.

Jerram, D. (2002). Volcanology and facies architecture of flood basalts. Geological Society of
America Special Paper 362, 119-132.

Jerram, D., & Widdowson, M. (2005). The anatomy of continental flood basalt provinces:
Geological constraints on the process and products of flood volcanism. Lithos, 385-405.

Jones, M., Jerram, D., Svensen, H., & Grove, C. (2016). The effects of large igneous provinces
on the global carbon and sulphur cycles. Paleogeography, Palaeoclimatology, Palaeoecology, 4
21.

Jones, S.M., Hoggett, M., Greene, S.E. et al. Large Igneous Province thermogenic greenhouse
gas flux could have initiated Paleocene-Eocene Thermal Maximum climate change. Nat
Commun 10, 5547 (2019). https://doi.org/10.1038/s41467-019-12957-1

Kasbohm, J., & Schoene, B. (2018). Rapid eruption of the Columbia River flood basalt and
correlation with the mid-Miocene climate optimum. Science advances, 4(9), eaat8223

Kerr, Andrew. (2013). Oceanic Plateaus. 10.1007/978-94-007-6644-0_21-1.
Knight, K., Nomade, S., Renne, P., Marzoli, A., Betrand, H., & Youbi, N. (2004). The Central
Atlantic Magmatic Province at the Triassic-Jurassic boundary: Paleomagnetic and 40Ar/40Ar
evidence from Morocco for brief, episodic volcanism. Earth and Planetary Science Letters, 228,
143-160.

Knoll, A. H., Bambach, R. K., Payne, J. L., Pruss, S., & Fischer, W. W. (2007). Paleophysiology
and end-Permian mass extinction. Earth and Planetary Science Letters, 256(3-4), 295-313.

Kravchinsky, V.A. (2012) Paleozoic large igneous provinces of northern Eurasia: Correlation
with mass extinction events. Global and Planetary Change, 86-87, 31-36

Kuroda, Junichiro, et al. "Contemporaneous massive subaerial volcanism and late cretaceous
Oceanic Anoxic Event 2." Earth and Planetary Science Letters 256.1-2 (2007): 211-223.
Kuroda, Junichiro, et al. "Lead isotopic record of Barremian–Aptian marine sediments:
Implications for large igneous provinces and the Aptian climatic crisis." Earth and Planetary
Science Letters 307.1 (2011): 126-134.

64
Le Quéré, C., Andrew, R.M., Friedlingstein, P., Sitch, S... Zheng, B. (2018) Global Carbon
Budget 2018. Earth Syst. Sci. Data. 10, 2141–2194.

Lenton, T., Williamson, M., Edwards, N., Marsh, R., Price, A., Ridgwell, A., . . . Cox, S. (2006).
Millennial timescale carbon cycle and climate change in an efficient Earth system model.
Climate Dynamics, 26, 687-711.

Lord, N., Ridgwell, A., Thorne, M., & Lunt, D. (2016). An impulse response function for the
"long tail" of excess atmospheric CO2 in an Earth system model. Global Biogeochemical Cycles,
2-17.

Martindale, R., Berelson, W., Corsetti, F., Bottjer, D., & West, A. (2012). Constraining
carbonate chemistry at a potential ocean acidification event (the Triassic-Jurassic boundary)
using the presence of corals and coral reefs in the fossil record. Palaeogeography,
Palaeoclimatology, Palaeoecology, 350-352, 114-123.

Marzoli, A., Bertrand, H., Knight, K., Cirilli, S., Buratti, N., Verati, C., . . . Bellieni, G. (2004).
Synchrony of the Central Atlantic magmatic province and the Triassic-Jurassic boundary
climatic and biotic crisis. Geology, 32, 973-976.

Marzoli, A., Jourdan, F., Puffer, J., Cuppone, T., Tanner, L., Weems, R., . . . A. (2011). Timing
and duration of the Central Atlantic Magmatic Province in the Newark and Culpeper Basins,
eastern USA. Lithos, 175-188.

Marzoli, A., Renne, P., Piccirillo, E., Ernesto, M., Bellieni, G., & De Min, A. (1999). Extensive
200-million-year-old continental flood basalts of the Central Atlantic Magmatic Province.
Science, 616-618.

Matthews, K.J., Maloney, K.T., Zahirovic, S., Williams, S.E., Seton,W. & Muller, R.D. (2016)
Global plate boundary evolution and kinematics since the late Paleozoic. Global and Planetary
Change, 146, 226-250
McElwain, J., Beerling, D., & Woodward, F. (1999). Fossil plants and global warming at the
Triassic-Jurassic boundary. Science, 285, 1386-1390.

McGlade, C., & Ekins, P. (2015). The geogrpahical distribution of fossil fuels unsued when
limiting global warming to 2 degress C. Nature, 517.

McHone, J. (2003). Volatile Emissions from Central Atlantic Magmatic Province Basalts: Mass
Assumptions and Environmental Consequences. In W. Hames, J. McHone, P. Renne, & C.
Ruppel, The Central Atlantic Magmatic Province: Insights from Fragments of Pangea (p. ).
Washongton D.C. : American Geophysical Union.

McNeil, B. I., & Sasse, T. P. (2016). Future ocean hypercapnia driven by anthropogenic
amplification of the natural CO2 cycle. Nature, 529(7586), 383-386.

65
Morse, J., & He, S. (1993). Influences of T,S and PCO2 on the pseudo-homogenous precipitation
of CaCO3 from seawater: implications for whiting formation. Marine Chemistry, 291-297.

Morse, J., Gledhill, D., & Millero, F. (2003). CaCO3 precipitation kinetics in waters from the
great Bahama bank: Implications for the relationship between bank hydrology and whitings.
Geochemica et Cosmochemica Acta, 2819-2826.

Mussard, M., Le Hir, G. Fluteau, F., Lefebvre, V. & Godderis, Y. (2014) Modeling the carbon-
sulfate interplays in climate changes related to the emplacement of continental flood basalts. The
Geological Society of America Special Paper 505, 339-352
Neal, C., Mahone, J., Kroenke, L., Duncan, R., & Petterson, M. (1997). The Ontong Java
Plateau. Geophysical Monograph-American Geophysical Union 100, 183-216.

Olsen, P., Kent, D., Et-Touhami, M., & Puffer, J. (2003). Cyclo-, magneto- and bio-stratigraphic
constraints on the duration of the CAMP event and its relationship to the Triassic-Jurassic
boundary. In The Central Atlantic Magmatic Province: Insights from fragments of Pangea (pp.
7-32). American Geophysical Union Monograph 136.

Opdyke, B. N., Wilkinson, B.H. (1993) Carbonate mineral saturation state and cratonic
limestone accumulation. Am. J. Sci. 293, 217–234.

Palfy, J., Demeny, A., Haas, J., Hetenyi, M., Orchard, M., & Veto, I. (2001). Carbon isotope
anamoly and other geochemical changes at the Triassic-Jurassic boundary. Geology, 29, 1047-
1050.

Palmer, M. R., & Ernst, G. G. J. (1998). Generation of hydrothermal megaplumes by cooling of
pillow basalts at mid-ocean ridges. Nature, 393(6686), 643-647.

Panchuk, K., Ridgwell, A., & Kump, L. (2008). Sedimentary response to Paleocene-Eocene
Thermal Maximum carbon release: A model-data comparison. Geology, 36, 315-318.

Paris, G., Donnadieu, Y., Beaumont, V., Fluteau, F., & Godderis, Y. (2016). Geochemical
consequences of intense pulse-like degassing during the onset of the Central Atlantic Magmatic
Province. Palaeogeography, Palaeoclimatology, Palaeoecology, 441, 74-82.

Polteau, S., Mazzini, A., Galland, O., Planke, S., & Malthe-Sorenssen, A. (2008). Saucer-shaped
intrusions: Occurences, emplacement and implications. Earth and Planetary Science Letters,
414-427.

Rampino, M.R. & Stothers, R.B. (1988). Flood basalt volcanism during the past 250 million
years. Science, 241, p. 663-668

Reidel, S. P., Camp, V. E., Ross, M. E., Wolff, J. A., Martin, B. S., Tolan, T. L., & Wells, R. E.
(2013). The Columbia River Flood Basalt Province (Vol. 497). Geological Society of America.

66
Ridgwell, A. (2005). A Mid Mesozoic Revolution in the regulation of ocean chemistry. Marine
Geology, 217, 339-357.

Ridgwell, A., & Hargreaves, J. (2007). Regulation of atmospheric CO2 by deep-sea sediments in
an Earth System Model. Global Biogeochemical Cycles, 21(2).

Ridgwell, A., & Zeebe, R. (2005). The role of the global carbonate cycle in the regulation and
evolution of the Earth System. Earth and Planetary Science Letters, 299-315.

Ridgwell, A., Hargreaves, J., Edwards, N., Annan, J., Lenton, T., Marsh, R., . . . Watson, A.
(2007). Marine geochemical data assimilation in an efficient Earth system model of global
biogeochemical cycling. Biogeosciences, 4, 87-104.

Ries, J. B., Ghazaleh, M. N., Connolly, B., Westfield, I., & Castillo, K. D. (2016). Impacts of
seawater saturation state (ΩA= 0.4–4.6) and temperature (10, 25 C) on the dissolution kinetics of
whole-shell biogenic carbonates. Geochimica et Cosmochimica Acta, 192, 318-337.

Schaller, M., Wright, J., & Kent, D. (2011). Atmospheric pCO2 perturbation associated with the
Central Atlantic Magamatic Province. Science, 331, 1404-1409.

Schaller, M., Wright, J., Kent, D., & Olsen, P. (2012). Rapid emplacement of the Central
Atlantic Magmatic Province as a net sink for CO2. Earth and Planetary Science Letters, 323-
324, 27-39.

Self, S., Schmidt, A., & Mather, T. (2014). Emplacement characteristics, time scales and
volcanic gas release rates of continental flood basalt eruptions on Earth. Geological Society of
America Special Paper 505.

Self, S., Thordarson, T., & Widdowson, M. (2005). Gas fluxes from flood basalt eruptions.
Elements, 283-287.

Self, S., Widdowson, M., Thordarson, T., & Jay, A. (2006). Volatile fluxes during flood basalt
eruptions and potential effects on the global environment: A Deccan perspective. Earth and
Planetary Science Letters, 518-532.

Seton, M., Muller, R.D., Zahirovic, S., Gaina, C., Torsvik, T., Shephard, G., Talsma, A.
…Chandler, M. (2012) Global continental and ocean basic reconstructions since 200 Ma. Earth-
Science Reviews, 113, 212-270
Simon, L., Godderis, Y., Buggisch, W., Strauss, H., & Joachimski, M. (2007). Modeling the
carbon and sulfur isotope compositions of marine sediments: Climate evolution during the
Devonian. Chemical Geology, 246, 19-38.

Sobolev, S., Sobolev, A., Kuzmin, D., Krivolutskaya, N., Petrunin, A., Arndt, N., . . . Vasiliev,
Y. (2011). Linking mantle plumes, large igneous provinces and environmental catastrophes.
Nature, 312-316.

67
Song, H., Wignall, P. B., Chu, D., Tong, J., Sun, Y., Song, H., ... & Tian, L. (2014). Anoxia/high
temperature double whammy during the Permian-Triassic marine crisis and its
aftermath. Scientific reports, 4(1), 1-7.

Storey, M., Duncan, R. A. & Swisher, C. C. Paleocene-Eocene thermal maximum and the
opening of the Northeast Atlantic. Science 316, 587–589 (2007)

Svensen, H., Planke, S., Malthe-Sorenssen, A., Jamtveit, B., Myklebust, R., Eidem, T., & Rey, S.
(2004). Release of methane from a volcanic basin as a mechanism for initial Eocene global
warming. Nature, 542-545.

Tanner, L., Hubert, J., & Coffey, B. (2001). Stability of atmospheric CO2 levels across the
Triassic-Jurassic boundary. Nature, 411, 675-677.
Thordarson, T., & Self, S. (2003). Atmospheric and environmental effects of the 1783-1784 Laki
eruption: A review and reassessment. Journal of Geophysical Research-Atmosphere, 1-29.

Tejada, Maria Luisa G., et al. "Ontong Java Plateau eruption as a trigger for the early Aptian
oceanic anoxic event." Geology 37.9 (2009): 855-858.Wignall, P. (2005). The end-Permian mass
extinction-How bad did it get? Geobiology, 293-297.
Urban, M. C. (2015). Accelerating extinction risk from climate change. Science, 348(6234), 571-
573.

Walker, J., Hays, P., & Kasting, J. (1981). A negative feedback mechanism for the long-term
stabilization of Earth's surface temperature. Journal of Geophysical Research, 9776-9782.

Walter, L., & Morse, J. (1985). The dissolution kinetics of shallow marine carbonates in
seawater: a laboratory study. Geochemica et Cosmochemica Acta, 1503-1513.

Ward, P., Haggart, J., Carter, S., Wilbur, D., Tipper, H., & Evans, T. (2001). Sudden
productivity collapse associated with the Triassic-Jurassic boundary mass-extinction. Science,
292, 1148-1151.

White, R.V. & Saunders A.D. (2005) Volcanism, impact and mass extinctions: Incredible or
credible coincidences? Lithos, 79, p. 299-316
Whiteside, J., Olsen, P., Kent, D., Fowell, S., & Et-Touhami, M. (2007). Synchrony between the
Central Atlantic magmatic province and the Triassic-Jurassic mass extinction event?
Palaeogeography, Palaeoclimatology, Palaeoecology, 345-367.

Wignall, P. (2001). Large igneous provinces and mass extinctions. Earth Science Reviews, 1-33.

Wignall, P. (2005). The end-Permian mass extinction-How bad did it get? Geobiology, 293-297.

Zeebe, R., & Wolf-Gladrow, D. (2001). CO2 in seawater: equilibrium, linetics, isotopes. Gulf
Professional Publishing.

68
Zheng, Xin-Yuan, et al. "Changing ocean circulation and hydrothermal inputs during Ocean
Anoxic Event 2 (Cenomanian–Turonian): evidence from Nd-isotopes in the European shelf
sea." Earth and Planetary Science Letters 375 (2013): 338-348.
Zheng, Xin-Yuan, et al. "A climatic control on reorganization of ocean circulation during the
mid-Cenomanian event and Cenomanian-Turonian oceanic anoxic event (OAE 2): Nd isotope
evidence." Geology 44.2 (2016): 151-154.

Abstract (if available)

Abstract Four of the “Big Five” mass extinctions in Earth’s history have been associated with intense volcanism from Large Igneous Provinces (LIPs), suggesting close ties between carbon emissions and biotic crisis. Yet not all LIP events had the same environmental impacts, nor did all cause mass extinctions. Here we use an Earth system model to address the question of why some LIPs were associated with severe environmental collapse while others were not. We find that the production of carbonate by calcifying organisms in the open ocean can significantly buffer the Earth system, reducing the impact of CO2 release on the oceans and atmosphere when compared to carbonate production only at reefs. The evolution of open ocean calcifiers in the mid-Mesozoic (~ 200 Myrs ago) may thus at least in part explain less severe impacts from LIPs over the past 150 Myrs compared to prior eruptions, pointing to the possibility that life and the Earth system co-evolved to buffer carbon cycle perturbations.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Paleoenvironmental and paleoecological trends leading up to the end-Triassic mass extinction event

PDF

The geobiological role of bioturbating ecosystem engineers during key evolutionary intervals in Earth history

PDF

Sedimentary geochemistry associated with the end-Triassic mass extinction: changes to the marine environment from an age constrained sedimentary section

PDF

The early Triassic recovery period: exploring ecology and evolution in marine benthic communities following the Permian-Triassic mass extinction

PDF

Benthic and pelagic marine ecology following the Triassic/Jurassic mass extinction

PDF

New insights into glacial-interglacial carbon cycle: multi-proxy and numerical modeling

PDF

Ecological recovery dynamics of the benthic and pelagic fauna in response to extreme temperature events and low oxygen environments developed during the early Triassic

PDF

The impact of Lagrangian environmental variability on the growth of phytoplankton

PDF

Carbonate dissolution at the seafloor: fluxes and drivers from a novel in situ porewater sampler

PDF

Environmental controls on alkalinity generation from mineral dissolution: from the mineral surface to the global ocean

PDF

The impact of mesoscale and submesoscale physical processes on phytoplankton biomass, community composition, and carbon dynamics in the oligotrophic ocean

PDF

Paleoecology of Upper Triassic reef ecosystems and their demise at the Triassic-Jurassic extinction, a potential ocean acidification event

PDF

The geobiology of fluvial, lacustrine, and marginal marine carbonate microbialites (Pleistocene, Miocene, and Late Triassic) and their environmental significance

PDF

Flowstone ideograms: deciphering the climate messages of Asian speleothems

PDF

Benthic paleoecology and macroevolution during the Norian Stage of the Late Triassic

PDF

Calcite and aragonite dissolution in seawater: kinetics, mechanisms, and fluxes in the North Pacific

PDF

Investigations on marine metal cycling through a global expedition, a wildfire survey, and a viral infection

PDF

Discerning local and long-range causes of deoxygenation and their impact on the accumulation of trace, reduced compounds

PDF

Ages, origins and biogeochemical role of water across a tropical mountain to floodplain transition

PDF

The dynamic regulation of DMSP production in marine phytoplankton

Asset Metadata

Creator Stellmann, Jessica Lynn (author)

Core Title How open ocean calcifiers broke the link between large igneous provinces and mass extinctions

School College of Letters, Arts and Sciences

Degree Master of Science

Degree Program Geological Sciences

Degree Conferral Date 2022-12

Publication Date 10/03/2022

Defense Date 09/09/2022

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag carbon cycle,earth systems modeling,large igneous province,mass extinction,OAI-PMH Harvest

Format theses (aat)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor West, Joshua (committee chair), Corsetti, Frank (committee member), Levine, Naomi (committee member)

Creator Email jessica.stellmann@gmail.com,stellman@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC112114247

Unique identifier UC112114247

Identifier etd-StellmannJ-11260.pdf (filename)

Legacy Identifier etd-StellmannJ-11260

Document Type Thesis

Format theses (aat)

Rights Stellmann, Jessica Lynn

Internet Media Type application/pdf

Type texts

Source 20221017-usctheses-batch-986 (), University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright. The original signature page accompanying the original submission of the work to the USC Libraries is retained by the USC Libraries and a copy of it may be obtained by authorized requesters contacting the repository e-mail address given.

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA

Repository Email cisadmin@lib.usc.edu