Boron in Foraminiferal Calcite as an Indicator of Seawater

Comments

Transcription

Boron in Foraminiferal Calcite as an Indicator of Seawater
Boron in Foraminiferal Calcite as an Indicator of
Seawater Carbonate Chemistry
Katherine A. Allen
Submitted in partial fulfillment
of the requirements for the degree of
Doctor of Philosophy
in the Graduate School of Arts and Sciences
COLUMBIA UNIVERSITY
2013
© 2013
Katherine A. Allen
All rights reserved
ABSTRACT
Boron in Foraminiferal Calcite
as an Indicator of Seawater Carbonate Chemistry
Katherine A. Allen
Foraminifera are unicellular organisms with a wide marine distribution. Many species
secrete carbonate tests whose physical and chemical nature reflect the seawater conditions in
which they grow. Thus, fossil tests preserved on the sea floor represent an archive that may
be used to investigate the composition of ancient seawater. With the aim of improving our
understanding of past ocean-climate links, I have tested proxies for seawater composition in
the modern ocean and applied them to a key period in Earth history.
The ratio of boron to calcium (B/Ca) in the calcite tests of planktic foraminifers has
previously been suggested as a proxy for past seawater carbonate chemistry, but controls on
B incorporation are not yet clear. The theoretical basis for this proxy is rooted in the pHdependent concentration of dissolved borate (B(OH)−
4 ) and its subsequent incorporation into
foraminiferal calcite. In this thesis, I present: 1) new insights into the environmental controls
on B/Ca revealed by culture experiments with living foraminifers, and 2) new reconstructions
of past seawater chemistry during the last deglaciation based on B/Ca of fossil calcite from
deep sea sediments.
To test environmental controls on B incorporation, I performed several culture experiments that quantified the effects of pH, temperature, salinity, dissolved boron and inorganic
carbon concentrations on the calcite tests of the planktic foraminifer species O. universa,
G. sacculifer, and G. ruber (pink). In these experiments, B/Ca increases with pH (lower
2−
−
[HCO−
3 ], higher [CO3 ] and [B(OH)4 ]) and salinity, but not with temperature. Thus, normal-
izing B/Ca data to a constant salinity (e.g., S=35) should improve our ability to isolate the
carbonate chemistry signal in B/Ca paleo-records and samples from different ocean sites.
In addition, B/Ca decreases with total dissolved inorganic carbon (DIC) at constant pH
2−
−
(higher [HCO−
3 ] and [CO3 ], constant [B(OH)4 ]), which suggests competition between aque-
ous boron and carbon species for inclusion into the calcite lattice. While different cultured
species exhibit similar B/Ca behavior in response to salinity, temperature, and pH changes,
their absolute B/Ca values are offset under identical seawater conditions. Thus, B/Ca is
both a function of environmental parameters that exert strong influence on test composition
as well as biological processes that result in species offsets.
To determine whether these culture calibrations are applicable in the open ocean, I used
equations relating the B/Ca of cultured foraminifers with experimental seawater properties
to predict B/Ca of wild specimens derived from sediment core-tops. Most measured core-top
values for O. universa and G. sacculifer are similar to values predicted by culture calibrations (average offsets are 4 and 15 µmol mol−1 , respectively) but values predicted for coretop
G. ruber deviate by up to 60 µmol mol−1 from predicted values. The greater discrepancy
observed for core-tops may suggest that our experiments still fall short of identifying all
environmental controls on B/Ca and/or that we need to revisit the growth conditions assumed for planktic foraminifers, in particular the depth habitat of G. ruber. Further, an
evaluation of planktic foraminiferal downcore data shows that B/Ca in planktic foraminifers
is not sufficiently sensitive to surface ocean carbonate chemistry to permit reconstruction
of Pleistocene atmospheric CO2 changes. However, B/Ca may serve as a useful proxy in
environments that experienced large carbonate system changes, such as upwelling zones, or
large events such as those during the Paleocene-Eocene.
In contrast to planktic foraminifers, B/Ca of benthic foraminifer tests appears to respond
to deep-water carbonate saturation state (∆CO2−
3 ). B/Ca of the benthic species Cibicidoides
wuellerstorfi increases linearly with ∆CO2−
3 in all major ocean basins, as demonstrated in the
modern core-top calibration of Yu and Elderfield (2007). To gain insight into carbon storage
in the deep ocean across the last glacial termination, I investigated the B/Ca composition of
C. wuellerstorfi in a sediment core from New Zealand’s Bay of Plenty, located at a depth of
−1
1,627 meters. The resulting reconstruction indicates that ∆CO2−
3 changed up to 30 µmol kg
across the deglaciation. Combined with benthic δ 13 C and independent paleo-O2 estimates,
the [CO2−
3 ] record indicates increased storage of CO2 in the deep ocean during the LGM,
with major roles for the biologic pump and carbonate compensation.
Contents
1. Introduction to Paleoclimate and Seawater Carbonate Chemistry
1
1.1. Climate History . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1
1.2. Seawater Carbonate Chemistry . . . . . . . . . . . . . . . . . . . . . . . . .
6
1.3. Unlocking Ocean History . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
15
1.3.1
Proxies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
15
1.3.2
Boron systematics . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
16
2. Controls on boron incorporation in cultured tests of the planktic foraminifer
Orbulina universa
21
2.1. Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
21
2.2. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
22
2.3. Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
27
2.3.1
Culturing procedures . . . . . . . . . . . . . . . . . . . . . . . . . . .
27
2.3.2
Elemental Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . .
29
2.3.3
Laser Ablation ICP-MS . . . . . . . . . . . . . . . . . . . . . . . . .
30
2.4. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
31
2.4.1
Carbonate system . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
31
2.4.2
Boron concentration . . . . . . . . . . . . . . . . . . . . . . . . . . .
32
2.4.3
Salinity and Temperature . . . . . . . . . . . . . . . . . . . . . . . .
33
2.4.4
Laser ablation profiles . . . . . . . . . . . . . . . . . . . . . . . . . .
33
2.5. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
35
i
2.5.1
Sensitivity to the Carbonate System in culture experiments . . . . . .
37
2.5.2
The role of boric acid . . . . . . . . . . . . . . . . . . . . . . . . . . .
37
2.5.3
The role of carbonate ion . . . . . . . . . . . . . . . . . . . . . . . . .
42
2.5.4
B/Ca variability within individual shells . . . . . . . . . . . . . . . .
43
2.5.5
Influence of [B]SW on partition coefficients . . . . . . . . . . . . . . .
44
2.5.6
Salinity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
46
2.5.7
Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
47
2.5.8
Proxy implications . . . . . . . . . . . . . . . . . . . . . . . . . . . .
47
2.6. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
49
2.7. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
50
3. The planktic foraminiferal B/Ca proxy for seawater carbonate chemistry: A critical evaluation
55
3.1. Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
55
3.2. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
56
3.3. Calibrations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
57
3.4. Down-core records . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
68
3.5. Recommendations for proxy development . . . . . . . . . . . . . . . . . . . .
71
3.5.1
Distinguish environmental controls . . . . . . . . . . . . . . . . . . .
73
3.5.2
Investigate the role of test size . . . . . . . . . . . . . . . . . . . . . .
74
3.5.3
Establish species-specific calibrations . . . . . . . . . . . . . . . . . .
75
3.5.4
Biological processes may influence B uptake . . . . . . . . . . . . . .
75
3.5.5
Constrain past seawater composition . . . . . . . . . . . . . . . . . .
77
3.6. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
78
3.7. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
79
4. Environmental controls on B/Ca in calcite tests of the tropical planktic
foraminifer species Globigerinoides ruber and Globigerinoides sacculifer 81
ii
4.1. Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
81
4.2. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
82
4.3. Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
84
4.3.1
Culture experiments . . . . . . . . . . . . . . . . . . . . . . . . . . .
84
4.3.2
Gulf of Mexico samples . . . . . . . . . . . . . . . . . . . . . . . . . .
87
4.3.3
Sample preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . .
87
4.3.4
Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
88
4.4. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
89
4.4.1
Culture experiments . . . . . . . . . . . . . . . . . . . . . . . . . . .
89
4.4.2
Gulf of Mexico core-top samples . . . . . . . . . . . . . . . . . . . . .
91
4.5. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
93
4.5.1
Carbonate system . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
93
4.5.2
Salinity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
95
4.5.3
Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
98
4.5.4
Species differences . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
4.5.5
Geochemical peculiarity of G. sacculifer ’s sac-like chamber . . . . . . 101
4.5.6
Comparison with core-top data . . . . . . . . . . . . . . . . . . . . . 103
4.5.7
Future proxy application and development . . . . . . . . . . . . . . . 106
4.6. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
4.7. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
5. Deep water CO2 loss during the Last Glacial Termination: Evidence
from New Zealand’s Bay of Plenty
113
5.1. Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
5.2. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
5.3. Core Site and Regional Oceanography . . . . . . . . . . . . . . . . . . . . . . 117
5.4. Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
5.5. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
iii
5.6. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
5.6.1
Implications for the Last Ice Age . . . . . . . . . . . . . . . . . . . . 125
5.6.2
Implications for the Termination . . . . . . . . . . . . . . . . . . . . 131
5.7. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
5.8. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
Appendix A: Artificial Seawater Recipe
137
Appendix B: Data and core photos from Bay of Plenty, New Zealand
139
iv
List of Figures
1.1
Solar Insolation and Global Ice Volume . . . . . . . . . . . . . . . . . . . . .
4
1.2
Bjerrum Plot: Dissolved Inorganic Carbon Species . . . . . . . . . . . . . . .
7
1.3
Alkalinity, Total Dissolved Inorganic Carbon, and pH . . . . . . . . . . . . .
12
1.4
Atlantic and Pacific Carbonate Ion Profiles . . . . . . . . . . . . . . . . . . .
14
1.5
Boron Isotope Systematics . . . . . . . . . . . . . . . . . . . . . . . . . . . .
18
2.1
Carbonate System Calibration . . . . . . . . . . . . . . . . . . . . . . . . . .
32
2.2
Boron concentration experiments . . . . . . . . . . . . . . . . . . . . . . . .
34
2.3
Salinity and Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . .
35
2.4
Laser Ablation Profiles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
36
2.5
The Empirical Partition Coefficient, KD
. . . . . . . . . . . . . . . . . . . .
38
2.6
Boron concentrations and isotopes vs. pH . . . . . . . . . . . . . . . . . . .
40
2.7
Controls on B/Ca . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
41
3.1
Coretop Calibrations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
59
3.2
Temperature and Carbonate Ion . . . . . . . . . . . . . . . . . . . . . . . . .
60
3.3
Comparison with Inorganic Calcite . . . . . . . . . . . . . . . . . . . . . . .
62
3.4
Denominator-driven KD Behavior . . . . . . . . . . . . . . . . . . . . . . . .
64
3.5
Examining a Downcore Calibration . . . . . . . . . . . . . . . . . . . . . . .
66
3.6
Primary B/Ca Relationships . . . . . . . . . . . . . . . . . . . . . . . . . . .
67
3.7
Glacial-Interglacial B/Ca Records . . . . . . . . . . . . . . . . . . . . . . . .
69
3.8
Comparison of pCO2 Estimates . . . . . . . . . . . . . . . . . . . . . . . . .
72
v
4.1
Carbonate System Experiments: pH
. . . . . . . . . . . . . . . . . . . . . .
89
4.2
Carbonate System Experiments: DIC . . . . . . . . . . . . . . . . . . . . . .
90
4.3
Salinity and Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . .
92
4.4
Testing Controls on B Incorporation . . . . . . . . . . . . . . . . . . . . . .
96
4.5
Activity and Concentration . . . . . . . . . . . . . . . . . . . . . . . . . . .
99
4.6
Core-top Comparison . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
5.1
Core Site in the Southwest Pacific Ocean: Map . . . . . . . . . . . . . . . . 119
5.2
Core Location: Bathymetric and Oxygen Cross-Sections
5.3
Age Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
5.4
B/Ca in C. wuellerstorfi as a proxy for ∆CO2−
. . . . . . . . . . . . . . . . 124
3
5.5
B/Ca Trace Element Record . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
5.6
Reconstruction of Seawater Carbonate Ion . . . . . . . . . . . . . . . . . . . 127
5.7
Estimating Paleo-Alkalinity . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
5.8
B/Ca vs. δ 13 C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
5.9
Photos of Core 83: Trigger Core (TC). . . . . . . . . . . . . . . . . . . . . . 144
5.10 Photos of Core 83: Jumbo Piston Core (JPC).
vi
. . . . . . . . . . . 120
. . . . . . . . . . . . . . . . 145
List of Tables
2.1
B/Ca results from Orbulina universa culture experiments . . . . . . . . . . .
52
2.2
Shell weights and diameters . . . . . . . . . . . . . . . . . . . . . . . . . . .
53
3.1
Glacial and Interglacial B/Ca Values . . . . . . . . . . . . . . . . . . . . . .
70
4.1
Seawater Properties and B/Ca Results . . . . . . . . . . . . . . . . . . . . . 110
4.2
Gulf of Mexico Core-top Results . . . . . . . . . . . . . . . . . . . . . . . . . 111
5.1
Age Constraints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
5.2
Trace element data from Core 83 TC and JPC. . . . . . . . . . . . . . . . . 143
vii
ACKNOWLEDGMENTS. I am fortunate to have a strong, loving, and inspiring
network of family, friends, and mentors. Throughout the trials and occasional triumphs
that graduate school brings, many people were there to provide support, guidance, a reality
check, a hug, a shoulder, an ear, a toast, a push ... whatever was needed to keep me moving
forward. I am deeply grateful to Bärbel Hönisch for her tireless support of all aspects of
my graduate work. Under her mentorship I have grown as a scientist and a teacher; I have
traveled the world, shared my ideas with the community both in person and in print, and
developed many field and laboratory skills. My thesis committee members – Bob Anderson
and Peter deMenocal – have been a continual source of insight and inspiration. I am also
grateful to Taro Takahashi for many discussions full of wisdom and encouragement. Serving
as a fellow of the LEEFS program enriched my graduate experience in more ways than I can
count; many thanks to Bob Newton and to Susan Vincent, my unstoppable partner in the
classroom and in many muddy adventures. For support in the laboratory and the field, I give
many, many thanks to Tali Babila, Kat Esswein, Nina Ruprecht, Steve Eggins, Howie Spero,
Ann Russell, Yair Rosenthal, Liz Sikes, Jimin Yu, Harry Elderfield, Aurora Elmore, Rebecca
Norman, Steve Doo, Spider Vetter, Jason Day, Julie Richey, Ben Flower, and Milton Carlo.
For infectious enthusiasm, unfailing style, and the battle cry “For The Love of Science!” I
thank the one and only Kelly Sztrepeck. Looking back, I can’t imagine these years without
my friends Carolyn and Jason Hudak, Elizabeth Pierce, Alison Hartman, Anna Foster, Meg
Reitz, Emmi Yonekura, Merry Yue Cai, Chris Hayes, Margaret Elvekrog, Sarah Hale, Aurora
Elmore, Jordan Landers, Stephanie Carr, Meg and Gordon Bromley, Laura Gonzalez, and
Catherine Delain. I thank my amazing sister, Emily, for rock-solid support, empathy, and
humor in the toughest of times. I thank my parents, Jon and Joanne, for a lifetime of love,
encouragement, and inspiration, and for instilling in me a deep curiosity and a passion for
nature. And finally, I thank my wonderful husband, Aaron, for supporting me through many
trials, for nurturing my creative spirit, and for always believing in me.
viii
To my parents,
who taught me to find truth and joy in nature.
ix
x
1
Chapter 1
Introduction to Paleoclimate
and Seawater Carbonate Chemistry
Conditions on the surface of the Earth have changed dramatically during our planet’s history.
The ocean, which today covers about 70% of Earth’s surface, has played an important role
in the joint evolution of rocks, atmosphere, ice sheets, and life since the ocean’s formation at
least 3.8 billion years ago (Nutman et al., 2012; Polat and Frei, 2005). Understanding how
the nature of seawater has changed in the past can shed light on many important processes
that are still shaping the world we live in today.
1.1
Climate History
Ice age cycles exert powerful influence on Earth’s landscape and are the defining feature
of late Cenozoic climate, but the driving mechanisms behind these multi-thousand-year
oscillations remain elusive. In the mid-1800s, geologists inferred from glacial landforms that
ice had covered large areas of Europe, Scandinavia, and North America (e.g., Agassiz, 1840).
About 100 years later, geochemists found additional evidence from deep marine sediments
that temperatures had fluctuated widely, and large ice sheets had grown and collapsed several
times during the past million years (Emiliani, 1955, 1966; Shackleton, 1987). Each of these
2
glacial cycles was characterized by slow cooling and ice sheet expansion followed by rapid
warming and ice sheet melting approximately every 100,000 years (100 ky). Before ∼ 900,000
years ago (900 ka), large ice sheets expanded every 40-41 ky instead of 100 ky, and before
2.7 million years ago (Ma), evidence for such large ice sheets in the Northern Hemisphere is
sparse during the Cenozoic (<65.5 Ma) (Raymo, 1994).
What caused these major climate transitions? Many scientists have considered the role
of the Sun, our climate system’s primary heat source. If the Earth were flat, and fixed at a
constant angle and distance far from the Sun, the amount of solar energy (insolation) received
at any given location would also be roughly constant through time (assuming constant solar
luminosity). However, the Earth is round, and revolves around a wobbly axis, and is not
located at a fixed distance from the Sun. How do these factors influence the radiation Earth
receives?
Equatorial regions receive the most concentrated solar energy because solar rays strike
the land surface at a direct angle, almost perpendicular to the land surface. In contrast,
polar regions’ land surface is oblique to incoming solar rays and so the energy is spread over
a much larger area, effectively diluted. Changing a region’s angle of tilt relative to solar
rays during a year gives us seasons (New York City is cold during the winter because rays
strike it at a more oblique angle and for a shorter daily period than during the summer). In
addition, the slow migration of equinoxes along Earth’s elliptical orbit (a phenomenon known
as precession) controls whether a particular season occurs when the Earth is at its closest
or farthest point from the Sun (91.5 and 94.5 million miles, respectively), or somewhere in
between, further modulating the intensity of a seasonal insolation.
When considering the average amount received across all latitudes and all days of the
year (global annual insolation), the only parameter that matters is Earth’s distance from the
Sun, which varies slightly due to the gravitational tug of other planets in our solar system.
Moving a planet closer to the Sun increases the radiation received (like stepping closer to a
bonfire), and vice versa. Figure 1.1 shows that during the past 1 million years, the global
3
annual insolation varied from 341.3 to 341.8 Watts m−2 , giving a total range of 0.5 W m−2 .
On the other hand, the insolation variability calculated for specific locations and seasons
during this same period can be larger than the global annual mean. For example, at 65
degrees North, the amount of summer (June 21) insolation has changed as much as 129 W
m−2 . For comparison, the modern summer-winter insolation range at 45 degrees North (e.g.,
the state of Maine) is 364 W m−2 . This means that the tilt-driven insolation change at
this latitude within a single year is more than 700x larger than the total global insolation
change over a million years, but only about 3x larger than the summer insolation change at
65°N. In summary, the total amount of solar energy striking the planet has been relatively
steady for the past few million years, but location-specific and seasonal insolation changes
have changed dramatically (Figure 1.1).
Scientists have argued about insolation’s role in climate change since the 1800s. One
landmark study revealed that for the past 500 ky, variations in sediment properties in the
Indian Ocean closely followed the rhythms of high-latitude summer insolation (Hays et al.,
1976). The behavior of these three different sediment properties, together reflecting global
ice volume and temperature, suggests a link between orbitally-controlled insolation changes
and climate. However, although variations in solar insolation are clearly important, they can
not fully explain the observed inception and pattern of ice ages. The magnitude of climate
response to solar forcing has not been consistent – as noted by Hays et al. (1976) and by
many others, ice ages have occurred approximately every 80 to 120 ky during the past million
years, but the 100 ky insolation forcing (eccentricity) is much weaker than the other, more
rapid frequencies of obliquity and precession. This mis-match between the strength of orbital
forcing and climate response has been dubbed ‘the 100 ky problem,’ and it has not yet been
resolved. Even before the beginning of widespread glaciation in the Northern Hemisphere,
climate changes occurred that can not be explained by insolation alone, such as dramatic
spikes in global temperature (e.g., the Paleocene-Eocene Thermal Maximum) and growth
of ice sheets on the Antarctic continent (Zachos et al., 2001, 2008). Such discrepancies
4
Solar Forcing
-2
Insolation (W m )
550
500
450
o
At 65 North on June 21
400
Annual Global Mean
-2
(range: 341.3 - 341.8 W m )
350
O (permil)
4.0
4.5
Colder,
More Ice
Benthic
18
3.5
d
Climate indicator
Warmer,
Less Ice
3.0
5.0
0
200
400
600
800
1000
Time (ka)
Figure 1.1: During the past 1 million years, changes in summer insolation at 65°N (black
solid line) are much larger than changes in annual global mean insolation (black dashed
line). The oxygen isotopic composition of benthic foraminiferal calcite (δ 18 O, colored line)
is a measure of global ice volume and deep water temperature (Lisiecki and Raymo, 2005).
This ‘saw-tooth’ pattern of slow cooling and rapid warming does not match orbitally-driven
changes in summer solar insolation at high latitudes (black solid line), indicating either that
this insolation parameter is not the main control on Northern Hemisphere ice volume or that
strong lags and nonlinear feedbacks exist within the climate system (e.g., Hays et al. (1976);
Imbrie et al. (1993)).
5
between climate records and insolation forcing indicate the presence of lags, feedbacks, and
additional forcing mechanisms within the climate system (Imbrie et al., 1993; Rind, 2002;
Wunsch, 2003).
The ocean has an immense capacity for heat transport (Ganachaud and Wunsch, 2000)
and gas storage (Millero, 2006). Changes in both of these processes likely play a role in
climate cycles (e.g., Denton et al. (2010)). For example, Earth’s surface temperature can
be increased by the heat-trapping action of greenhouse gases (GHGs) in the atmosphere
(Tyndall, 1861; Arrhenius, 1896; Hansen et al., 1984), and GHG abundances are influenced
by gas exchange between the atmosphere and ocean (Takahashi et al., 2009). Bubbles of
ancient air trapped in ice cores indicate that atmospheric concentrations of the greenhouse
gas carbon dioxide (CO2 ) were low during glacial periods (∼ 180 parts per million, ppm)
and high during interglacial periods (∼ 280 ppm), likely amplifying changes in global surface
temperature (Petit et al., 1999). Due to the larger extent of land-based ice sheets (Balco and
Rovey, 2010), the glacial terrestrial biosphere was smaller and could not have sequestered
large amounts of CO2 from the glacial atmosphere (Adams et al., 1990). The ocean, on the
other hand, contains ∼ 60 times more carbon than the atmosphere, and is the only surface
reservoir capable of absorbing enough carbon to account for the observed CO2 range on this
timescale (Sigman and Boyle, 2000).
Records of past natural fluctuations in seawater carbonate chemistry may thus help evaluate the patterns and drivers of carbon cycle and climate change. Specifically, marine records
can provide insight into ocean circulation and carbon storage, and even allow estimates of
past atmospheric CO2 beyond the temporal reach of ice cores. In this thesis, I develop methods for estimating past ocean carbonate chemistry (Chapters 1, 2, and 3) and present a new
sedimentary record of the last deglaciation (Chapter 4).
6
1.2
Seawater Carbonate Chemistry
To understand the ocean’s role in carbon cycle and climate change, we must consider the
behavior of carbon in seawater. Today, when carbon dioxide (CO2 ) dissolves into surface
seawater, most of it reacts with water (H2 O) to form carbonic acid (H2 CO3 ), which quickly
2−
dissociates into bicarbonate (HCO−
3 ) and carbonate ions (CO3 ) (Equation 1.1). The sum
of these dissolved carbon species is referred to as total dissolved inorganic carbon, or DIC.
If two of the measurable parameters CO2 (aqueous), DIC, pH, and alkalinity are known
in addition to temperature, salinity, and pressure, as well as the thermodynamic stability
constants for the relevant acid-base reactions, then a system of equations may be solved to
2−
−
+
quantify all chemical species of the carbonate system (CO2 , HCO−
3 , CO3 , OH , and H ).
These chemical species are involved in many interesting processes such as photosynthesis,
respiration, carbonate growth and dissolution, and continental weathering.
CO2 (aq) + H2 O H2 CO3 HCO3− + H + CO32− + 2H +
(1.1)
DIC = CO2 (aq) + H2 CO3 + HCO3− + CO32−
(1.2)
The rapid and almost complete conversion of dissolved gaseous carbon into ionic compounds dramatically slows the saturation of seawater with respect to CO2 , giving the ocean
its massive and climatically-important capacity for carbon uptake (Figure 1.2).
As CO2 dissolves into seawater, it also forces an increase in positive hydrogen ions (H+ ,
H3 O+ , H9 O+
4 ), resulting in lower seawater pH. This phenomenon, often referred to as ocean
acidification (Orr et al., 2005; Feely et al., 2004; Fabry et al., 2008; Doney et al., 2009), is
mainly the result of the following reactions:
CO2 + CO32− + H2 O 2HCO3−
(1.3)
7
-
CO2*
-1
Concentration ( mol kg )
2000
2-
HCO3
CO3
1500
m
1000
B(OH)3
500
H) 4
B(O
-
OH
0
5
6
7
8
9
10
pH (total scale)
Figure 1.2: Concentrations of major ions in the carbonate system. CO∗2 = [CO2 (aq) +
H2 CO3 ]. Curves were calculated in CO2SYS.M (van Heuven et al., 2011) assuming DIC
= 2000 µmol kg−1 , BT = 416 µmol kg−1 at S = 35, total pH scale, carbonic acid and
bicarbonate dissociation constants from Lueker et al. (2000), and KSO4 dissociation constant
from Dickson (1981).
8
HCO3− CO32− + H +
(1.4)
−
In this case, CO2−
3 ions are converting most of the invading CO2 to HCO3 (Equation
1.3), dampening the pH effect that CO2 would have had in pure water. This process is called
2−
+
buffering. If a small portion of that HCO−
3 then dissociates into H and CO3 (Equation
1.4), then the balance between total positive and negative charge in solution is maintained,
but the total amount of H+ increases slightly, corresponding to lower pH values 1 . This
chemical link between CO2 (gas) and surface ocean pH is the basis for atmospheric CO2
reconstructions from surface ocean pH proxies like foraminiferal boron isotopes, discussed
later (Sanyal et al., 1995; Pearson and Palmer, 2000; Hönisch and Hemming, 2005; Foster,
2008; Hönisch et al., 2009).
On long timescales (>1,000 years), ocean pH can be buffered not only by ions in solution,
but also by ions supplied from vast carbonate deposits on the sea floor (Broecker et al., 1979;
Archer, 1996). These carbonates are derived from the discarded tests of tiny pelagic calcifiers
like foraminifers and coccolithophorids, and they have played a key role in maintaining steady
carbonate saturation for over 100 million years (Zeebe and Westbroek, 2003). When deepsea carbonates dissolve, the liberated Ca2+ and CO2−
3 raise total seawater alkalinity, one of
the key parameters of the carbonate system.
In all solutions, the sum of positive and negative charges must be zero. This thermodynamic requirement is at the heart of the alkalinity concept. In natural seawater, the total
un-reactive or ‘conservative’ positive ion charge (e.g., Na+ , Mg2+ , Ca2+ ) is greater than the
−
−
total conservative negative ion charge (e.g., Cl− , SO2−
4 , Br , F ), and the difference is made
1
Many texts define pH as -log[H+ ]F , where [H+ ]F represents the free hydrogen ion concentration, including
hydrated forms (e.g., H3 O+ , H9 O+
4 ). Throughout this thesis, for greater accuracy and also for the sake of
consistency with previous studies, I use the total pH scale (Hansson, 1973), defined as:
pHT = −log([H + ]F + [HSO4− ])
(1.5)
Hydrogen sulfate ion [HSO−
4 ] is included in this equation because it is a major constituent of seawater, and
has the ability to release H+, making it functionally similar to a free hydrogen ion.
9
2−
up by the dissociation of weak acids like carbonic acid (H2 CO3 → HCO−
3 and CO3 ) to
give a net charge of zero. As described above (Equations 1.3 and 1.4), ions like CO2−
3 have
the ability to resist solution pH change by accepting a proton. Alkalinity can be thought of
as the total charge difference between such proton acceptors and proton donors in solution.
The higher a water sample’s alkalinity, the more acid it can neutralize. When alkalinity =
0, buffering no longer occurs, and pH will rise rapidly in response to acid addition.
For example, in a simple system composed of only water, sodium chloride salt (NaCl),
and dissolved carbon, the charge balance can be expressed as:
[N a+ ] + [H + ] − [Cl− ] − [HCO3− ] − 2 · [CO32− ] − [OH − ] = 0
(1.6)
and the alkalinity can be expressed as:
ALK = [HCO3− ] + 2 · [CO32− ] + [OH − ] − [H + ]
(1.7)
Carbonate ion is multiplied by two here because alkalinity keeps track of charges in
solution, and this ion has a charge of minus two. Sodium and chloride behave as conservative
ions in this situation, so they do not contribute to alkalinity. In this case, excluding sodium
and chloride, the sum of positive charges equals the sum of negative charges and the value
of alkalinity is zero.
Natural seawater is not so simple, and many additional chemical species contribute to
alkalinity, including borate, phosphate, and hydrogen silicate ions as well as organic acids
(Dickson, 1981; Wolf-Gladrow et al., 2007; Zeebe and Wolfe-Gladrow, 2001). In practical
terms, seawater alkalinity is simply the number of charge equivalents of hydrogen ion that can
be neutralized by a given volume of water, and it can be determined through acid titration
even if the identities of participating buffer ions are not entirely known (Wolf-Gladrow et al.,
2007; Zeebe and Wolfe-Gladrow, 2001).
Today, the main components of the ocean’s alkalinity budget are inputs from continen-
10
tal weathering and removal by carbonate burial (Archer, 1996). Chemical weathering of
continental silicate and carbonate rocks produces a mixture of clay, silicic acid, cations (primarily Ca2+ , Mg2+ , Na+ , K+ ), and anions (primarily HCO−
3 ). A generalized equation for
the break-down of carbonate minerals can be written as follows:
CaCO3 + CO2 + H2 O Ca2+ + 2HCO3−
(1.8)
This process produces two bicarbonate ions. One is from the combination of carbon
+
dioxide with water (CO2 + H2 O HCO−
3 + H ), and the other is from the dissolution of
2+
calcium carbonate (CaCO3 + H+ HCO−
3 + Ca ). On long timescales, carbonate weath2+
ering does not affect atmospheric pCO2 because the products, HCO−
3 + Ca , recombine
in the ocean, releasing the CO2 that was initially consumed back into surface waters and
ultimately back to the atmosphere.
Silicate weathering, on the other hand, results in net atmospheric CO2 consumption. The
reaction of potassium feldspar (a major rock-forming mineral) with CO2 and water produces
kaolinite clay, potassium cations, two bicarbonate anions, and silicic acid:
2KAlSi3 O8 (s) + 2CO2 + 11H2 O → Al2 Si2 O5 (OH)4 (s) + 2K + + 4H4 SiO4 + 2HCO3− (1.9)
Although this process produces two bicarbonate ions, just like carbonate weathering,
there is no major or favorable reaction by which the products are rapidly recombined in continental or marine waters. It has been suggested that silicate weathering acts as a homeostatic
control on Earth’s climate: when pCO2 increases, temperatures rise, chemical weathering
intensifies and CO2 is removed from the atmosphere until steady state is reached. Conversely, when pCO2 decreases, temperatures drop, chemical weathering slows and less CO2
is removed from the atmosphere.
Together, the processes of carbonate and silicate weathering contribute both cations and
11
anions (ultimately, alkalinity) to seawater via continental runoff. Some of this alkalinity
flux is removed quickly through carbonate deposition (discussed below), and some alkalinity
remains in seawater much longer, only removed through reactions with ocean crust or by
adsorption onto sedimentary clay particles.2 Chemical analyses of river water suggest that of
the total bicarbonate load washed into the ocean, about 1/3 is from dissolution of carbonate
minerals, 1/3 is from consumption of atmospheric CO2 by carbonate weathering, and 1/3 is
from consumption of atmospheric CO2 by silicate weathering (Emerson and Hedges, 2008).
In the deep sea, carbonate deposits only form in places where conditions are favorable
for preservation – e.g., in places where waters are supersaturated with respect to calcium
carbonate, and/or where burial outpaces dissolution. The presence or absence of carbonates
in sediment cores can provide insight into past ocean chemistry if we apply some basic
carbonate solubility relationships.
Dissolution of carbonate is described by:
CaCO3 Ca2+ + CO32−
Ksp =
aCa2+ · aCO32−
aCaCO3
(1.10)
(1.11)
Where Ksp , the solubility product of CaCO3 , is a function of system temperature, pressure, and composition (Ingle, 1975; Mucci, 1983), and ‘a’ indicates the activity, or effective
concentration, of a chemical species. For practical use, the measurable apparent solubility
product is often applied:
∗
= [Ca2+ ] · [CO32− ]
Ksp
(1.12)
Where the total concentrations (indicated by square brackets) include both free and
2
The fates of many seawater components are still unclear (e.g., Mg2+ , K+ ), and chemical oceanographers
have been working for decades to resolve apparent imbalances between observed riverine input and estimated
removal by mineral formation (after MacKenzie and Garrels, 1966).
12
2500
8.4
8
2450
Carb onate
dissolution
8.2
8.3
Alkalinity (µmol kg − 1)
8.1
8.2
7.9
2400
8
8
2350
8.1
8.2
7.9
CO 2
release
2300
8
7.8
CO 2 7.7
invasion 7.6
7.8
pH
7.6
8.1
2250
7.9
2200
2150
8
7.9
7.8
Carb onate
growth
7.5
7.7
7.3
7.5
7.7
7.2
7.2
7.4
1950
7.1
7.6
7.8
2100
1900
7.4
7.4
7.6
2000
2050
2100
2150
Dissolved Inorganic Carb on (µmol kg
2200
−1
)
Figure 1.3: The influence of CaCO3 growth/dissolution and CO2 invasion/release on alkalinity, DIC, and pH (colored contours). The slope of the growth/dissolution line is 2 because
CaCO3 growth consumes alkalinity and DIC in a 2:1 ratio. CO2 invasion and release can
raise and lower DIC, but do not affect alkalinity. Calculations were made in CO2SYS.M
assuming S = 35, T = 25°C, total pH scale, equilibrium constants from Lueker et al. 2000,
and Dickson’s KSO4 dissociation constant.
13
complexed ions.
3
The ratio of in situ (measured) ion concentrations to Ksp is a useful
predictor of carbonate dissolution or stability on the sea floor:
Ω=
[Ca2+ ]in situ · [CO32− ]in situ
Ksp
(1.13)
When Ω = 1, the solution is saturated with respect to CaCO3 . When Ω >1, the solution is
supersaturated, tipping the balance in favor of carbonate growth and preservation, and when
Ω <1, the solution is undersaturated, favoring dissolution. In seawater, Ca2+ behaves as a
conservative ion, varying in direct proportion to total dissolved salt. Salinity and Ca2+ vary
spatially throughout the surface ocean by roughly 10%, but are almost uniformly distributed
in the deep ocean. Because [Ca2+ ] can be assumed to be roughly constant in the deep ocean,
a parameter simply comparing in situ and saturation CO2−
3 concentrations is often used to
describe deep water carbonate saturation:
∆CO32− = [CO32− ]in situ − [CO32− ]sat
(1.14)
Because dissolved Ca2+ and CO32− ions take up less space than crystalline CaCO3 , they
are thermodynamically favored at higher pressures, making these dissolved ions the more
stable form at greater ocean depths (Hawley and Pytkowicz, 1969). As a result, carbonates
tend to accumulate on shallower ocean ridges and dissolve in the abyss, creating a landscape
often compared to snow-capped mountains (e.g. Archer 1998). The depth boundary at
which Ω = 1 is known as the saturation horizon, and the deeper boundary where carbonate
accumulation equals dissolution (where carbonates are essentially absent) is known as the
carbonate compensation depth (CCD). Today, the saturation horizon is located much deeper
in the North Atlantic Ocean (>4,000 m) than in the North Pacific Ocean (<1,000 m) largely
because deep water masses in the North Atlantic have higher [CO2−
3 ] (Figure 1.4) (Takahashi
and Broecker, 1981; Li et al., 1969).
3
Concentration and activity of any species ‘M’ are related by the equation: aM = γ · [M ], in which the
activity coefficient, γ, represents the fraction of M that is free to react.
14
0
Pacific saturation horizon (Ω = 1)
1000
TION
URA
SAT
Depth (m)
2000
3000
North Pacific
North Atlantic
4000
Atlantic saturation horizon (Ω = 1)
5000
0
50
100
150
200
250
-1
Carbonate ion (mmol kg )
Figure 1.4: Carbonate ion concentrations in the North Atlantic (blue dots) and North Pacific
Oceans (green dots) compared with values calculated for calcite saturation (dashed line). The
depth at which the in situ concentrations are equal to the saturation values is called the
saturation horizon. The saturation horizon is located at greater depth in the Atlantic than
the Pacific largely because Atlantic deep waters have higher carbonate ion concentrations.
Data shown here were derived from the Global Ocean Data Analysis Project (cite Key et
al.), bounded within the following regions: 42.8 - 48.2 °N, 25.0 - 33.2 °W (Atlantic); 26.0 47.0 °N, 148.1 - 160.4 °W (Pacific). Saturation values were calculated in Ocean Data View,
assuming constants as in the DOE 2007 handbook (cite Dickson).
15
If the alkalinity supplied by terrestrial weathering exceeds alkalinity consumption through
shallow carbonate deposition, then seawater [CO2−
3 ] will increase and the saturation horizon
will deepen, allowing carbonate accumulation at greater depths and across a larger area.
Conversely, if deep sea [CO2−
3 ] decreases, the saturation horizon will shoal, less CaCO3
2+
will be preserved, and deep carbonates will dissolve, releasing CO2−
3 (and Ca ) back into
seawater. In this way, deep pelagic burial compensates for imbalances between weathering
and shallow burial fluxes (Broecker and Peng, 1993). Carbonate burial and dissolution are
intimately tied to seawater [CO2−
3 ], which is a key parameter in climate studies because it
can exert a strong control on atmospheric pCO2 .
In summary, any two carbonate system parameters may be used in conjunction with
temperature, pressure, and salinity to constrain the entire carbonate system and quantify
the abundances of key ions in seawater. In Chapter 4, I apply proxies for these parameters
to geological archives in order to estimate past carbonate system states and test mechanisms
for carbon cycle and climate change.
1.3
1.3.1
Unlocking Ocean History
Proxies
The early paleoceanographic records of Césare Emiliani, Nicolas Shackleton, and others
were based on chemical signatures locked within microfossils – specifically, the calcite tests
of tiny, single-celled marine organisms from the phylum Foraminifera. As discussed above,
the discarded tests of individual foraminifers have been accumulating over much of the sea
floor for millions of years, constituting an enormous archive of ocean history. In the midto late-1900s, scientists pioneered the study of the oxygen isotopic composition (δ 18 O) of
foraminiferal calcite, which reflects the organism’s habitat temperature as well as δ 18 O of
seawater, an indicator of continental ice volume. Thus δ 18 O of carbonate acts as a ‘proxy’,
a measurable property preserved within a physical archive that reflects its contemporaneous
16
environmental conditions. In other words, a proxy is a key to the past. Many proxies have
been developed in recent decades, including the boron isotopic composition of boron atoms
locked within foraminiferal calcite (δ 11 B), which is an indicator for seawater pH. Studies of
δ 11 B also provide the foundation for a new proxy: the ratio of boron to calcium (B/Ca), which
can provide further insight into the seawater carbonate system. In this thesis, I examine
B/Ca of foraminiferal calcite derived from 1) culture experiments with living specimens and
2) marine sediment cores composed of fossil material.
1.3.2
Boron systematics
Approximately 0.1% of Earth’s total B inventory is stored in the ocean (Leeman and Sisson,
2002), the rest residing in crustal and mantle rocks. Major sources of B to the ocean reservoir
include riverine discharge (38 × 1010 g B yr−1 ) and fluids expelled from hydrothermal vents
and at convergent margins (6 × 1010 g B yr−1 ); major sinks include low-temperature crust
alteration (27 × 1010 g B yr−1 ) and clay adsorption (13 × 1010 g B yr−1 ) (Lemarchand et al.,
2000, 2002; Spivack et al., 1987). Hydrothermal processes at mid-ocean ridges, atmospheric
particulate and gas phases, and deposition of marine carbonate and silicate also play minor
roles (Leeman and Sisson, 2002; Vengosh et al., 1991). The variable δ 11 B and relative fluxes
of these sources and sinks likely resulted in significant shifts in δ 11 BSW through time. Recent
estimates for the residence time of B in the ocean range from 14 to 20 million years (Simon
et al., 2006; Lemarchand et al., 2000), but B fluxes between major reservoirs have not yet
been well-constrained, and models describing the evolution of δ 11 BSW through time do not
agree (Simon et al., 2006; Lemarchand et al., 2000). However, on short timescales (<1 My),
it is reasonable to assume that concentrations of B in seawater are constant.
In seawater, B behaves as a conservative element, maintaining a consistent boron to
salinity ratio of 0.1336 ±0.0005 mg kg−1 · S −1 , which equals 433 µmol B kg−1 seawater at S
= 35 (Lee et al., 2010). The isotope ratio (11 B/10 B) of boron in seawater does not vary with
depth, salinity, or temperature, giving a mean ocean value of δ 11 B 39.61 ± 0.04 h (Foster
17
et al., 2010).
4
Most B in modern seawater exists in the form of B(OH)3 or B(OH)−
4 (Byrne and Kester,
1974). In average surface seawater of pH 8 (total scale), ∼ 80% of B exists in the form of
−
B(OH)3 and ∼ 20% exists as B(OH)−
4 , with approximately 44% of B(OH)4 being complexed
by Na+ , Mg2+ , and Ca2+ (Byrne and Kester, 1974). Some evidence suggests the presence
of small amounts of polyborates and other species in solution as well (Ingri, 1963; Baes and
Mesmer, 1976), though equilibrium measurement data indicate that these are only detectable
in solutions whose [B] is more than 50 × that of natural seawater (Byrne et al., 2006; Liu
and Tossell, 2005). Since the present study focuses on natural seawater, the behavior of
these minor polyborate species are not discussed further.
Equilibrium exchange between B(OH)3 and B(OH)−
4 in seawater is described by the
equation:
+
B(OH)3 + H2 O B(OH)−
4 +H
(1.16)
whose balance is described by pKB (at 25°C and 35 S, pKB = 8.597). pKB depends on temperature, pressure, salinity (Dickson, 1990) and solution composition (Dyrssen and Hansson,
1973). All else being equal, [B(OH)−
4 ] increases with pH at the expense of [B(OH)3 ]. Due
to the constant fractionation between δ 11 B of B(OH)3 and B(OH)−
4 (27.2 h, Klochko et al.
(2006)), the δ 11 B of B(OH)−
4 increases with pH as well (Figure 1.5).
In natural marine carbonates and also in cultured foraminiferal calcite, both δ 11 B and
B/Ca increase with seawater pH (Sanyal et al., 1996; Hemming and Hanson, 1992), support−
ing a B(OH)−
4 -only (or B(OH)4 -dominant) model of B incorporation during foraminiferal
4
The delta notation used here indicates a normalized ratio of 11B to 10B:
 11

B
11
δ B=
10 B
sample
11 B
10 B
− 1] × 1000
(1.15)
SRM
where SRM indicates the standard boron reference material provided by the National Institute of Standards
and Technology (SRM 951).
-1
Concentration ( mol kg )
18
400
B(O
)
4
H
B(O
H)
3
300
m
200
100
0
d
11
B (permil)
60
εB
50
B(OH)3
40
total seawater = 39.61
30
B(OH)4-
20
7.0
8.0
9.0
10.0
pH (total scale)
Figure 1.5: Concentrations and boron isotopic compositions of boric acid (B(OH)3 ) and
borate (B(OH)−
4 ) in seawater. As pH increases, borate’s abundance and boron isotopic
composition both increase (green curves). Calculation assumes alkalinity = 2300 µmol kg−1 ,
S = 35, T = 25°C, E= 27.2h (cite Klochko), δ 11 BSW = 39.61 (cite Foster), B/Chl = 0.2414
(cite Lee), and carbonate system constants as in Figure 1.2.
19
calcite growth.
Based on these observations, the following incorporation equation was proposed:
−
CaCO3 + B(OH)−
4 CaHBO3 + HCO3 + H2 O
(1.17)
Applying the observed relationships between pH and δ 11 B to fossil foraminifer tests from
Pleistocene marine sediments reveals a surface ocean pH record with values ∼ 0.15 units
lower during interglacial than glacial periods (Sanyal et al., 1995; Palmer, 2003; Hönisch and
Hemming, 2005; Foster, 2008). These records fulfill the expectation that seawater pH should
decrease as pCO2 increases (Equations 1.3 and 1.4, Figure 1.2). In this way, δ 11 B records
from marine sediments may be used to estimate atmospheric CO2 beyond the temporal reach
of ice cores (Hönisch et al., 2009).
As with all proxies, δ 11 B has its limitations. Reconstructing past pH from δ 11 B requires
knowledge of past δ 11 B of seawater (δ 11 BSW ), which changes through time and can be difficult
to constrain (Simon et al., 2006; Lemarchand et al., 2000). An additional complication
is that the δ 11 B recorded by foraminiferal tests may be altered by dissolution on the sea
floor (Hönisch and Hemming, 2004), which limits this technique to well-preserved sediments.
Developing additional proxies to test or complement δ 11 B – such as B/Ca – allows more
robust reconstruction of past seawater chemistry.
20
21
Chapter 2
Controls on boron incorporation in
cultured tests of the planktic
foraminifer Orbulina universa
Note: This chapter has been published in Earth and Planetary Science Letters (2011) Vol.
309, No. 3-4, pp. 291-301
2.1
1
Abstract
Culture experiments with living planktic foraminifers reveal that the ratio of boron to calcium (B/Ca) in Orbulina universa increases from 56 to 92 µmol mol−1 when pH is raised from
7.6 ± 0.02 to 8.7 ± 0.03 (total scale). Across this pH range, the abundances of carbonate,
bicarbonate, and borate ions also change (+530, -500, and +170 µmol kg−1 , respectively).
Thus specific carbonate system control(s) on B/Ca remain unclear, complicating interpre1
AUTHORS: Katherine A. Allena *, Bärbel Hönisch a , Stephen M. Eggins b , Jimin Yu a , Howard J. Spero
, Henry Elderfield d
a
Lamont-Doherty Earth Observatory of Columbia University, 61 Route 9W, Palisades, NY 10964, USA
b
Research School of Earth Sciences, The Australian National University, Canberra ACT, Australia
c
Department of Geology, University of California, Davis, CA 95616, USA
d
Department of Earth Sciences, University of Cambridge, Downing Street, Cambridge, CB2 3EQ, UK
corresponding author: [email protected], tel: (845) 365-8668, fax: (845) 359-2931
c
22
tation of paleorecords. B/Ca in cultured O. universa also increases with salinity (55 – 72
µmol mol−1 from 29.9 – 35.4 h) and seawater boron concentration (62 – 899 µmol mol−1
from 4 – 40 ppm B), suggesting that these parameters may need to be taken into account
for paleorecords spanning large salinity changes (∼ 2 h) and for samples grown in seawater
whose boron concentration ([B]SW ) differs from modern by more than 0.25 ppm. While
our results are consistent with the predominant incorporation of the charged borate species
(B(OH)−
4 ) into foraminiferal calcite, behavior of the partition coefficient KD (defined as
−
[B/Ca]calcite /[B(OH)−
4 /HCO3 ]seawater ) can not be explained by borate incorporation alone,
and suggests the involvement of other pH-sensitive ions such as CO2−
3 . For a given increase
−
in seawater B(OH)−
4 , the corresponding increase in B/Ca is stronger when B(OH)4 is raised
by increasing [B]SW than when it is raised by increasing pH. These results suggest that B incorporation controls should be reconsidered. Additional insight is gained from laser-ablation
ICP-MS profiles, which reveal variable B/Ca distributions within individual shells.
2.2
Introduction
The ocean exerts a strong influence on atmospheric CO2 levels and the short-term carbon
cycle, and plays a key role in regulating Earth’s climate. For these reasons, it is important
to constrain past variations in seawater carbonate chemistry and saturation state. Insight
into this system has already been gained through analysis of marine sediment records whose
properties are sensitive to changes in ocean chemistry. Further development of such proxies
will better enable us to constrain past conditions and establish records against which models
can be tested. In ocean regions where the marine carbonate system is closely linked to the
atmosphere via gas exchange, quantifying changes in the surface ocean’s carbonate system
(including variables such as pH, alkalinity, and [CO2−
3 ]) would enable researchers to recon-
23
struct paleo-pCO2 levels and, in places of upwelling or downwelling, may shed light on past
ocean-atmosphere carbon exchange.
Fully constraining the carbonate system requires knowledge of temperature, salinity,
2−
pressure, and two of the following six parameters: pH, [CO2 + H2 CO3 ], [HCO−
3 ], [CO3 ],
total alkalinity (TA), and total dissolved inorganic carbon (DIC). Proxies for some of these
parameters have been developed, but their associated secondary effects and calibration uncertainties (e.g. Katz et al., 2010) make it desirable to develop additional tools which can
verify or complement these paleo-estimates.
The theory behind pH-dependent boron incorporation into the shells of foraminifera is
rooted in the observed behavior of boron isotopes in seawater and carbonates. In seawater,
the abundance and boron isotope composition (δ 11 B) of borate (B(OH)−
4 ) both increase
with pH (Hemming and Hanson, 1992). If borate is then incorporated during calcification
of marine carbonates, the boron concentration and δ 11 B of carbonates should be sensitive to
seawater pH. An incorporation equation was proposed by Hemming and Hanson (1992):
−
CaCO3 + B(OH)−
4 CaHBO3 + HCO3 + H2 O
Keq =
aCaHBO3 · aHCO3− · aH2 O
aCaCO3 · aB(OH)−4
(2.1)
(2.2)
where Keq describes the phases of Eq. 1 at equilibrium, and is constant for a given temperature, pressure, and seawater composition. Yu et al. (2007) then related this incorporation
model to a property measurable by mass spectrometry: the ratio of boron to calcium in
calcite (B/Ca). Assuming that in calcite, Ca2+ ≈ CO2−
3 and CaHBO3 /CaCO3 ≈ B/Ca, and
−
that B(OH)−
4 and HCO3 are the species relevant for calcification, gives:
KD ≈
[B/Ca]calcite
−
[B(OH)−
4 /HCO3 ]seawater
(2.3)
24
where KD is an apparent partition coefficient describing incorporation of boron into calcite
relative to seawater composition. The exact chemical species and mechanisms involved,
however, remain uncertain.
Although some B/Ca data are consistent with the current model’s expectations, direct
measurements on solid-phase coordination are controversial (Sen et al., 1994; Klochko et al.,
2009). In solution, the borate ion is tetrahedrally-coordinated, while boric acid is trigonallycoordinated. Different interpretations of nuclear magnetic resonance (NMR) analyses on
marine carbonates have recently sparked debate on which of these aqueous species is in−
corporated into calcite: only [B(OH)−
4 ] (Sen et al., 1994), or a mixture of [B(OH)4 ] and
[B(OH)3 ] (Klochko et al., 2009). The presence of trigonally-coordinated B in calcite led
Klochko et al. (2009) to suggest that boric acid may be incorporated during calcification.
However, NMR can not definitively resolve this question because aqueous species coordination could change during adsorption and crystal formation (Tossell, 2006; Sen et al., 1994).
It is possible that some tetrahedrally-coordinated borate is adsorbed, and then transformed
to trigonal coordination in the solid. The presence of trigonal B in the lattice thus does not
necessarily prove aqueous B(OH)3 incorporation, and the issue remains open. The resolution
of this debate is important to interpretations of boron data from fossils because an uncertain
mixture of boric acid and borate in shell calcite would severely complicate the relationship
of B/Ca to past carbonate system conditions. One of the main goals of our study is to
re-evaluate the incorporation equation (Eq. 1) and partition coefficient (Eq. 3) using tight
constraints from laboratory experiments with living foraminifera.
Empirical calibrations are necessary because foraminiferal calcite does not form in complete chemical equilibrium with seawater. Evidence of disequilibrium includes large B/Ca
differences between species (Yu et al., 2007; Foster, 2008), covariance of oxygen and carbon
isotopes in foraminiferal shells (Bijma et al., 1999; Spero et al., 1997), and biological mediation of calcification (Bentov et al., 2009; Zeebe and Sanyal, 2002; de Nooijer et al., 2008;
25
Lea, 2003). These organisms grow within a seawater microenvironment (diffusive boundary
layer, ∼ 500 µm) whose chemistry is influenced by respiration, calcification, and by photosynthesis of symbionts, if present (Wolf-Gladrow et al., 1999a; Rink et al., 1998; Jørgensen
et al., 1985), which can lead to offsets from bulk seawater chemistry. The kinetics of the
system (including diffusion and hydration/dehydration of ions) may also play a role in B
incorporation, though the experimentally indistinguishable diffusion coefficients of B(OH)−
4
and B(OH)3 in seawater (Mackin, 1986) indicate that relative transport of these two ions
in solution does not likely influence B incorporation. Finally, the presence of organic membranes and compounds can also influence the growth rate and composition of carbonates
(e.g. Zuddas et al. (2003); Langer (1992); Stumm and Morgan (1996)), and may exert some
control over calcite structure and element composition. Due to these considerations, we can
not predict the partitioning of B from pure theory. To increase proxy robustness, we must
instead isolate and identify the environmental control(s) and thoroughly test any possible
interfering effects.
Some empirical relationships between solution and solid B chemistry have already been established for marine carbonates. The relationship between δ 11 B of surface-dwelling foraminifers
with pH has been calibrated in laboratory experiments (Sanyal et al., 1996; Hönisch and Hemming, 2004; Hönisch et al., 2003), and downcore records have also revealed tight co-variation
of δ 11 B with ice core CO2 (Hönisch et al., 2009; Hönisch and Hemming, 2005; Sanyal et al.,
1995; Foster, 2008). Similarly, B/Ca of synthetic calcites (Hobbs and Reardon, 1999) and
the concentration of boron in calcite ([B]calcite ) of another species of cultured foraminifera,
Globigerinoides sacculifer, both increase with pH (Sanyal et al., 1996). Studies of planktic
foraminifera shells from sediment coretops have suggested that KD can be used to determine
pH (once temperature is taken into account, Yu et al. (2007)) and/or [CO2−
3 ] (Foster, 2008;
Hendry et al., 2009), but the relative influences remain elusive. If the primary control were
pH, then B/Ca could serve as a valuable companion proxy for δ 11 B, increasing confidence in
paleo-pH reconstructions. If B/Ca instead reflects [CO2−
3 ], then it could theoretically serve
26
as the second parameter needed for carbonate system calculations, although the close covariation of pH and CO2−
3 in seawater leads to large propagated uncertainties when predicting
other carbonate system parameters (Rae et al., 2011; Yu et al., 2010b).
The influences of non-carbonate system variables on B/Ca are uncertain as well. Some
planktic foraminifera species from pelagic sediments exhibit a positive relationship between
KD and temperature in Globorotalia inflata, Globigerina bulloides (coretop samples), and
Globigerinoides ruber (downcore) (Yu et al., 2007; Tripati et al., 2009). Alternatively, Foster (2008) observed a negative KD -temperature relationship for G. sacculifer, G. ruber, and
Neogloboquadrina dutertrei (coretop). Inconsistent KD -T relationships from field data suggest that these pelagic samples are influenced by multiple seawater variables, that individual
species are influenced by different mechanisms, and/or that the controlling factors have not
yet been adequately identified.
Because salinity, temperature, and pH often covary in the modern surface ocean, it can
be difficult to discern their separate influences on shell composition solely from sediment-trap
and core-top samples. Another method to evaluate the physical controls on B incorporation is through elemental analysis of calcite precipitated by foraminifera from experimental
solutions for which physical and chemical properties are well-constrained and can be varied
independently. Here, we present calibration results from living planktic foraminifers grown
in controlled laboratory experiments. The symbiont-bearing, subtropical-temperate species
Orbulina universa was grown under a wide range of pH, temperature, salinity, and boron
concentration ([B]SW ). We observe higher B/Ca of foraminiferal calcite grown in solutions
with higher [B(OH)−
4 ], which is consistent with (but does not yet prove) a predominant incorporation of B(OH)−
4 . Our data also indicate a small salinity and negligible temperature
effect on B/Ca.
27
2.3
2.3.1
Methods
Culturing procedures
Culturing experiments were performed at the Wrigley Marine Science Center on Santa
Catalina Island (33◦ 26’40”N, 118◦ 28’55”W), during July and August 2008. Juvenile foraminifers
were individually collected by scuba divers approximately 2 km NNE of Big Fishermans Cove,
at a water depth of 2 – 8 m. Average surface water temperature and salinity during the
study period were 21.1◦ C (19.5 – 22.0 °C) and 33.0 (32.5 – 33.4 on the Practical Salinity
Scale PSS-78, quoted hereafter as “h”), respectively. Immediately after collection, samples
were brought to the laboratory where the foraminifers were identified and maximum shell
diameter was determined using light microscopes. If foraminifers did not already have food
caught in their spines, they were fed a one-day old brine shrimp nauplius (Artemia salina)
and then transferred to the experimental seawater. All specimens were subsequently fed an
Artemia nauplius every 3 days.
All seawater used in experiments was collected offshore during collection dives, and then
filtered through 0.8 µm filters. Each experiment was designed to shift either the carbonate
system, temperature, salinity, or [B]SW . In each experiment, 15 to 20 juvenile (pre-sphere)
individuals were incubated in separate 125 mL borosilicate glass jars. The large volume
of culture water per organism ensured that the carbonate system of experimental seawater
was not significantly affected by symbiont photosynthesis or respiration (Spero and Parker,
1985). Over the course of 3 to 14 days these foraminifers grew a spherical chamber typical for
O. universa. Experiments ended after specimens underwent gametogenesis, at which point
the spherical chamber was devoid of cytoplasm (Hamilton et al., 2008; Bé and Anderson,
1976).
We varied the pH of 4L batches of filtered seawater by titrating drop-wise with NaOH
28
or HCl (7.6 – 8.6, total scale). Salinity was modified between 29.9 – 35.4 ± 0.1 hby diluting
seawater with deionized water or by partial evaporation under a heat lamp at 60°C. Boron
concentrations were increased to approximately 5x and 10x the natural concentration by
adding 404 mg and 908 mg of boric acid to 4L seawater, followed by titration with NaOH
to reestablish an ambient pH of 8.0 (total pH scale). Culture jars were placed in separate
temperature-controlled water baths so specimens could be grown between 17.7 ± 0.5 and
26.5 ± 0.2℃. Temperature uncertainty represents the standard deviation of measurements
made by in-tank sensors every 5 minutes for the duration of each experiment.
To minimize gas-exchange with the atmosphere, jars were completely filled with experimental seawater to eliminate gas headspace and sealed with parafilm and tight-fitting lids.
To monitor potential atmospheric CO2 exchange during feeding (when caps must be removed), alkalinity and pH were measured at the beginning (∼2 water samples) and end of
each experiment (∼4 water samples). Alkalinity was measured using a Metrohm open cell
auto-titrator, calibrated against Dickson-certified alkalinity standards. For all experiments,
the initial-final alkalinity difference was <1%. The average initial-final pH change for all experiments was 0.02, with a maximum of 0.06 pH units for the high pH seawater, suggesting
invasion of a small amount of CO2 during feeding for that experiment.
Carbonate system parameters were calculated using Mehrbach constants for the dissociation of carbonic acid and bicarbonate, K1 and K2, refit by (Lueker et al., 2000) on the
total pH scale. The Matlab program csys3.m (Zeebe and Wolf-Gladrow, 2004) was modified to allow full specification of seawater variables relevant to our experiments, including
initial boron concentrations that deviate from the natural boron-to-salinity ratio. For each
experiment, temperature, alkalinity and pH were measured several times as described above;
averages were used to calculate the values presented in Table 1. To determine the upper
(and lower) influence bounds of each measured seawater variable (T, S, alkalinity, and pH) on
calculated carbonate system parameters, standard deviations were added to (and subtracted
29
from) average values and used to fully re-calculate the carbonate system. Each seawater variable’s corresponding parameter ranges were then root-mean-square combined, and reported
in Table 1 (±σ).
Light levels were held at 406 ± 108 µmol photons m−2 s−1 by overhead fluorescent lamps
(cool white, high output), monitored bi-weekly using a light meter. These light levels exceed the saturation intensity for the symbiotic dinoflagellate Gymnodinium béii associated
with O. universa (386 µmol photons m−2 s−1 , Spero and Parker (1985)) but falls well below
photoinhibition, which has not been observed below 700 µmol photons m−2 s−1 (Rink et al.,
1998). All experiments were maintained on 12:12h light:dark cycles. Every day, foraminifers
were observed using 10x hand-lenses. Spine and symbiont presence, degree of chamber-fill,
and sphere formation were recorded until each experiment was completed.
Experiments ended after foraminifers had undergone gametogenesis (dropped spines and
released gametes). Approximately 90% of all individuals reached gametogenesis, and only
these gametogenic foraminifers were archived for later analysis. Empty shells were rinsed in
deionized water, dried, weighed and measured to determine growth during the experimental
period.
2.3.2
Elemental Analysis
Shells were crushed and cleaned following the methods of Russell et al. (2004). Previous laser
ablation and bulk shell chemical analyses indicate that ∼ 15 – 20 shells are needed to get a
reliable average, so we combined at least 15 specimens from an experiment for each sample.
Samples were rinsed with ultrapure (QD) water to remove fine particles, and then oxidized
twice for 30 minutes with hot (70◦ C), buffered H2 O2 (0.1N NaOH, 15% v/v H2 O2 Seastar)
to remove organic matter. Following oxidation, samples were rinsed 5 times with ultrapure
water and leached 3 times with 0.001N HNO3 to remove adsorbed ions. The low weight
(5 – 10%, Spero and Parker (1985)) and extreme fragility of juvenile calcite mean that its
30
contribution to bulk shell chemistry (especially after sonication, rinsing and cleaning) was
likely minimal.
Elemental analyses were carried out in the Godwin Laboratory at Cambridge University
following the methods of Yu et al. [2005]. Calcium concentrations were first measured
by Inductively-Coupled Plasma Atomic Emission Spectrometry (ICP-AES). Aliquots of the
same solution were diluted to 100 ppm [Ca] to minimize matrix effects during subsequent
analysis by Inductively-Coupled Plasma Mass Spectrometry (ICP-MS, PerkinElmer SCIEX
Elan DRC II). Standards prepared with Milli-Q+ allowed a B/Ca detection range of 0 –
260 µmol mol−1 , after Yu et al. (2007). Standard solutions were measured every 3 – 5
samples, and the long-term (2 year) relative standard deviation (RSD, 1σ) for B/Ca is
2.6% (Cambridge in-house elemental standard). Boron memory effect was reduced by use
of a quartz spray chamber and long wash-out times, and drift was corrected using external
standards.
Boron concentrations in experimental seawater were also measured by ICP-MS. Seawater
samples (15 µl) were taken from each experiment, filtered at 0.2 µm, acidified with 0.5 µl 15N
HCl, and stored in nalgene bottles sealed with parafilm. Samples were analyzed 6 months
later at Cambridge on the same instrument as the foraminiferal calcite. Prior to analysis,
10 µl of seawater was diluted 500x in a solution of 1% quartz-distilled HNO3 and internal
standard (1 ppb Rhodium, Indium), which was used to correct for drift. During the run,
RSD of the internal standard was 2.3 % (Rh) and 1.4 % (In).
2.3.3
Laser Ablation ICP-MS
Element profiles were measured at The Australian National University in Canberra with
a pulsed ArF excimer laser (ń= 193 nm) coupled to an Agilent 7500s inductively coupled plasma mass spectrometer (ICP-MS), following previously-described procedures (Eggins
et al., 2002, 2003). Two foraminifers per experiment were split in half and treated with an
oxidative solution of 50% analytical-grade H2 O2 buffered with 0.001N NaOH to remove resid-
31
ual organic matter. Pre-analysis SEM photos of the inner shell surface indicated that the
juvenile portion of the shell was not present in most specimens selected for laser ablation.
Any juvenile calcite remaining on the few shells where it was observed was removed by the
cleaning procedure prior to ablation. Cleaned and Milli-Q rinsed shell halves were then
mounted on carbon tape and profiled 2 – 4 times each, perpendicular to the shell surface.
Multiple element profiles from each specimen were then matched via the Linage function in
Analyseries 2.0 (Paillard and Yiou, 1996) and then combined into a single average profile.
These procedures yielded B/Ca and Mg/Ca profiles for 25 cultured O. universa shells.
2.4
2.4.1
Results
Carbonate system
Ratios of boron to calcium (B/Ca) measured via ICP-MS are presented in Figure 2.1 and
Table 1. Between pH 7.6 and 8.6 (total scale), B/Ca in cultured O. universa shells increases
from 56 ±1 to 92 ±2 µmol mol−1 . This experimental pH range corresponds to ranges of 70
−
– 577 µmol [CO2−
3 ] and 33 – 195 µmol [B(OH)4 ].
Because several seawater parameters are physically and chemically linked, some of our
experiments varied more than one parameter. For example, temperature influences the
carbonate system equilibrium constants K1 and K2 , leading to slightly lower pH at higher
temperature. At 17.7 and 26.5 °C, pH was 8.1 and 8.0 (total scale), respectively, creating a
pH gradient of 0.1 ± 0.03 units. Impact of this difference on the B/Ca-temperature results
is discussed later. Other carbonate system parameters gave small differences across the
temperature range: 4 ± 23 µmol kg−1 alkalinity, 1 ± 10 µmol kg−1 carbonate ion, and 10
± 23 µmol kg−1 dissolved inorganic carbon, or DIC (± root mean square error). It is also
important to note that adding acid or base to seawater simultaneously changes pH, carbonate
ion, and alkalinity. Thus we will refer to these as “carbonate system” experiments rather
than “pH” experiments. Alkalinity and CO2−
3 differences between the lowest and highest pH
32
seawater (7.7 – 8.7) were 745 ± 7 and 527 ± 24 µmol kg−1 , respectively, due to changes in
aqueous speciation. However, no carbon was added to seawater by adding HCl or NaOH,
so the DIC difference across our pH range was negligible (0.2 ± 17.7 µmol kg−1 ) and driven
by gas exchange during water transfer and feeding. Finally, between our lowest and highest
salinity experiments (29.9 – 35.4 h), carbonate system differences were: 0.05 ± 0.04 pH
units, 345 ± 9 µmol kg−1 alkalinity, 310 ± 18 µmol kg−1 DIC, and 31 ± 13 µmol kg−1 CO2−
3 .
100
c
b
90
80
m
B/Ca ( mol mol-1)
a
70
60
50
7.6
8.0
8.4
8.8
0
pH (total)
200
400
600
CO32- (mmol kg )
-1
0
50
100
150
200
B(OH)4- (mmol kg-1)
Figure 2.1: Ratios of boron to calcium (B/Ca) in cultured O. universa calcite vs. carbonate
system parameters. In this series of experiments, pH was varied while T, S, and [B]SW
were held constant. The same B/Ca data are plotted against different parameters: pH (a),
carbonate ion (b), and borate (c). Error bars are ±2σ.
2.4.2
Boron concentration
B/Ca increases linearly with boron in seawater (Figure 2.2), and can be described by the
following equations:
B/Ca = 23.4 ∗ [B]SW − 43.4
B/Ca = 1.4 ∗ [B(OH)−
4 ] − 39.5
(R2 = 0.99, n = 4)
(R2 = 0.99, n = 4)
(2.4)
(2.5)
33
where [B]SW is the measured concentration of boron in seawater (ppm) and [B(OH)−
4 ] is
calculated from pH and KB (Dickson, 1990).
While higher salinity should concentrate all major dissolved elements equally (i.e., both
B and Ca), our high-[B] experiments increased [B] alone. These experimental conditions
differed from ambient conditions most dramatically with respect to [B]SW (raised 400 and
900%, to 2036 and 4070 µmol B kg−1 seawater, respectively) but total alkalinity also increased
11 and 30% due to the added boron alkalinity. Other parameters (pH, temperature, salinity)
were held constant. The response of B/Ca to [B(OH)−
4 ] in these [B]SW experiments (Equation
5) is ∼5x stronger than in carbonate system experiments, where [B(OH)−
4 ] increased due to
a shift in speciation rather than a rise in [B]SW (Figure 2.2).
2.4.3
Salinity and Temperature
B/Ca increases from 55 to 72 µmol mol−1 between salinity 29.9 and 35.4 h(Figure 2.3). A
linear least-squares fit gives: B/Ca = 2.6 ∗ S − 20.6
(R2 = 0.6). This relationship is
significant at a 95% confidence level as determined by the Student’s t-test (n = 7, t >1.90).
The temperature-B/Ca relationship between 17.7 and 26.5◦ C, however, is not significant at
a 95% confidence level (n = 5, t <2.57).
2.4.4
Laser ablation profiles
Individual shell profiles reveal variable element patterns. Even specimens grown under identical experimental conditions exhibit different cross-sectional trends and micro-scale B/Ca
to Mg/Ca relationships (Figure 2.4, a and b). Given the small number of individuals analyzed per experiment (n = 1 to 3), it is unlikely that we have captured the full range of
variability in this species. Despite the observed intra-shell variation, we find that in 21 out of
the 25 analyzed specimens, B/Ca increases from the inner to the outer surface of the shell.
Furthermore, although the absolute maximum and minimum ratios vary, the total B/Ca
34
variable [B]SW
variable pH
-1
B/Ca ( mol mol )
800
600
m
400
200
B/Ca = 1.4 * B(OH)4- - 39.5
0
0
200
400
600
B(OH)4 (mmol kg )
-
-1
Figure 2.2: In carbonate system experiments (circles), total [B]SW is held constant and the
concentration of borate increases with pH due to a pH-driven shift in aqueous speciation.
By contrast, in [B]SW experiments (triangles), pH is held constant (pH = 7.99 ±0.03 and
CO2−
3 = 167 ± 11) and the activity of borate simply increases with total [B]SW . For a given
borate increase, the corresponding B/Ca increase in O. universa calcite is ∼ 5x higher in
[B]SW than in carbonate system experiments. This might be due to aqueous complexation
or competition with other ions. Error bars are ±2σ.
35
100
A
B/Ca = 2.6 * S - 20.6
2
R = 0.6
a
< 0.05
B/Ca = 0.4 * T + 53.4
2
R = 0.7
a
> 0.05
80
m
B/Ca ( mol/mol)
90
B
70
60
50
30
32
34
Salinity (permil)
36
18
20
22
24
26
28
o
Temperature ( C)
Figure 2.3: A) B/Ca of O. universa vs. changes in salinity (29.5 – 35.5 h), and B) temperature (17.7 – 26.5°C). Error bars are ±2σ; dotted lines are 95% confidence limits. Temperature
fit is not shown because although the R2 seems high, its slope is not statistically different
from zero (α >0.05).
range within each shell is consistently between 40 and 60 µmol mol−1 even across different
experiments (pH, salinity, etc). Some profiles also exhibit an inverse relationship between
B/Ca and Mg/Ca (Figure 2.4, c and d).
2.5
Discussion
In our culture experiments, the carbonate system exerts a measurable control on B/Ca (Figure 2.1) of O. universa. Our results are broadly consistent with the existing theoretical
framework proposed for the proxy, because B/Ca does increase with aqueous borate con2−
−
−
centration. However, because [B(OH)−
4 ], [B(OH)4 /HCO3 ], pH and [CO3 ] covary in our
carbonate system experiments, it remains unclear which is truly the controlling parameter. In addition, we see that salinity and seawater boron concentration ([B]SW ) exert an
influence on B/Ca. A deeper understanding of the relative influences of these and other possible controls is needed to accurately interpret B/Ca from paleo-records. Here, we discuss
new insights gained into B incorporation with the aim of clarifying the proxy’s theoretical
36
7
6
60
5
50
4
0
20
40
60
80
80
2
60
40
60
80
100
outside
Distance (%)
6
60
5
50
4
20
40
60
80
100
10
180
B/Ca ( mol/mol)
4
20
70
D
160
8
140
6
120
m
B/Ca ( mol/mol)
m
100
inside
7
100
4
80
Mg/Ca (mmol/mol)
6
120
Mg/Ca (mmol/mol)
8
140
0
8
80
0
10
C
9
90
100
180
160
100
m
70
B/Ca ( mol/mol)
8
80
10
B
Mg/Ca (mmol/mol)
9
90
Mg/Ca (mmol/mol)
100
m
B/Ca ( mol/mol)
10
A
2
60
0
20
40
60
inside
80
100
outside
Distance (%)
Figure 2.4: B/Ca and Mg/Ca profiles of four individual O. universa: ‘a’ and ‘b’ were grown
under the same ambient experimental conditions (experiment 3); ‘c’ and ‘d’ were grown
under high pH conditions (experiment 7). Thick line is Mg/Ca; dotted line is B/Ca. Profiles
are plotted against % distance from the inner shell surface (estimated from ablation time).
37
framework and eventually resolving discrepancies between existing core-top datasets.
2.5.1
Sensitivity to the Carbonate System in culture experiments
Equation 1 suggests that calcite grown from solutions with greater B(OH)−
4 should have
higher B/Ca, all else being equal, and this prediction holds true in our carbonate system experiments (Figures 2.1 and 2.2). Interestingly, we see the partition coefficient (KD , equation
3) decreasing at higher pH, suggesting less efficient incorporation of B into calcite (Figure 2.5). In other words, although the absolute B/Ca values are increasing, for a given
−
B(OH)−
4 /HCO3 ratio less B is incorporated into calcite at high pH than at low pH. To
understand this behavior, we compare different culture experiments and consider several
possible controls below.
In our culture experiments, we increase [B(OH)−
4 ] in two separate ways: (1) by raising
pH, which shifts B speciation in favor of B(OH)−
4 while the total amount of B dissolved in
seawater ([B]SW ) is held constant, and (2) by raising [B]SW , which increases both B(OH)−
4
and B(OH)3 , while pH is held constant. Interestingly, the B/Ca response of O. universa to
increased B(OH)−
4 is ∼5x stronger when [B]SW is changed than when pH is changed (Figure
2.2). The weaker response of B/Ca to borate raised by acid/base addition (rather than by
[B]SW ) may be caused by: A) an increasingly negatively charged calcite surface at higher
pH, making it more difficult for borate to adsorb (Morse, 1986; van Cappellen et al., 1993),
and/or B) the involvement of other pH-sensitive ions in calcification. Our data can not be
used to address possibility A, but we will discuss B in the sections below. Specifically, we
will consider the potential roles of boric acid and carbonate ions.
2.5.2
The role of boric acid
The boron isotopic composition of O. universa calcite falls close to the isotopic composition of
borate in seawater, suggesting preferential incorporation of borate (Hemming and Hanson,
1992; Foster, 2008; Foster et al., 2010; Hönisch et al., 2007). However, empirical boron
KD (x 1000)
38
measured
"borate only"
3
2
1
0
7.6
8.0
8.4
8.8
400
600
CO32- (mmol kg )
-1
pH (total)
KD (x 1000)
200
3
2
1
0
18
20
22
24
26
o
Temperature ( C)
30
32
34
36
Salinity (permil)
Figure 2.5: Apparent behavior of the empirical boron partition coefficient KD in carbonate
system, temperature, and salinity experiments. Black dots represent KD values calculated
from measured B/Ca; white dots represent KD values for which the δ 11 B-inferred boric acid
contribution to B/Ca has been removed (“borate only” KD , see text for details). Vertical
error bars represent the total propagated 2σ error for uncertainties associated with both
B/Ca analysis (±σ 2.6%) and calculation of borate and bicarbonate from experimental
temperature, salinity, pH and alkalinity measurements.
39
isotope calibration curves involving biogenic and inorganic carbonates grown over a wide
pH range in the laboratory (Hönisch and Hemming, 2004; Sanyal et al., 1996, 2000, 2001;
Krief et al., 2010) have shallower slopes than expected from the measured boron isotope
fractionation between dissolved boric acid and borate in seawater (Klochko et al., 2006).
Because carbonates do not strictly follow the expected isotopic composition of B(OH)−
4 in
solution, Klochko et al. (2009) have suggested that some B(OH)3 may also be incorporated
at low pH. Since B(OH)3 is isotopically heavier (i.e. has higher
11
B/10 B) than B(OH)−
4 , its
incorporation at lower pH might explain the unexpectedly high δ 11 B of carbonates (Figure
2.6). Similarly, B(OH)3 incorporation might also explain the elevation of our lowest-pH
B/Ca ratio above the value predicted by the [B] experiments (Equation 5, Figure 2.7).
Assuming end-member values for boric acid and borate as predicted by the measured fractionation factor ε = 27.2 h (Klochko et al., 2006) and using the empirical δ 11 B calibration
for O. universa (Sanyal et al., 1996), we can apply an isotope mass-balance calculation to
estimate how much isotopically heavy boric acid would have to be incorporated to explain
the unexpectedly high δ 11 B of carbonates at lower pH (Figure 2.6). The calculated boric acid
contribution to total [B] in cultured O. universa calcite is only 3 – 8%, exerting a small effect
on B/Ca (4.4 – 2.1 µmol mol−1 ) and a negligible effect on KD (Figure 2.5). These results
support dominant incorporation of B(OH)−
4 (Hemming and Hanson, 1992; Hemming et al.,
1995; Sen et al., 1994) rather than a significant contribution of B(OH)3 (Klochko et al., 2009)
and are also consistent with evidence for exclusive borate incorporation from benthic species
(Rae et al., 2011). One puzzling observation is that the empirical δ 11 B-pH calibration for
O. universa can not be completely explained by mixtures of borate and boric acid because
the highest measured isotope value at pH 8.9 (total scale) falls below the δ 11 B of borate
(outside the two endmembers), and suggests either an analytical offset or an unidentified
fractionation process during calcification such as kinetic or biologic effects. We conclude
that despite remaining uncertainties, the isotope data of Sanyal et al. (1996) suggest that
incorporation of boric acid must be small and we must instead turn to other explanations
40
300
B(OH)4-
B(OH)3
200
100
m
mol / kgSW
400
70
B (‰)
60
ε = 27.2 ‰
B(OH)3
50
40
11
d
-
B(OH)4
30
20
10
7
8
9
10
pH (total)
Figure 2.6: Speciation (top) and isotopic composition (bottom) of dissolved borate in seawater both change with pH. Because bonds in borate and boric acid are characterized by different vibrational frequencies, boric acid is isotopically heavier than borate by 27.2 h(Klochko
et al., 2006). Klochko et al. (2009) have suggested that at low pH, incorporation of boric
acid into calcite could explain higher-than-expected boron isotope values. A simple mixing
line drawn between boric acid and borate shows that the δ 11 B of cultured O. universa (black
dots, Sanyal et al. (1996)) could be explained by incorporation of 3 – 8% boric acid (below
pH 8), but the value of the highest pH calibration point falls below the predicted borate
curve. See text for further discussion.
41
pH experiment data
B/Ca expected for pH experiment
based on [B]SW results, assuming
that B(OH)4- and HCO3are the only controls on B/Ca
200
m
B/Ca ( mol mol-1)
300
100
0
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
-1
B(OH)4- / HCO3- (mol mol )
Figure 2.7: B/Ca values observed in our carbonate system experiments do not agree with
values predicted in accordance with the current B incorporation model. Predicted B/Ca
−
values (white circles) were derived by applying the B/Ca-B(OH)−
4 /HCO3 relationship ob−
−
served in [B]SW experiments to B(OH)4 /HCO3 values calculated for each carbonate system
−
experiment. Thus, the predicted values reflect what the B/Ca would be if B(OH)−
4 /HCO3
were the only control on B/Ca. The mismatch between predicted and observed B/Ca in
our carbonate system experiment series thus suggests that other pH-sensitive ions may be
involved in calcification.
42
for observed B/Ca behavior.
2.5.3
The role of carbonate ion
Another possible explanation for observed B/Ca behavior lies in the role of [CO2−
3 ] during
2−
calcification. If aqueous B(OH)−
4 and CO3 compete for the same lattice site (Sen et al.,
1994), and carbonate ions are increasingly used in foraminiferal calcification at higher pH,
then B(OH)−
4 incorporation may correspondingly become less favorable. This idea is supported by a comparison of the predicted and observed B/Ca curves shown in Figure 2.7.
−
The predicted values are derived from the relationship between B/Ca and B(OH)−
4 /HCO3
observed in [B]SW experiments, where pH = 7.99 ±0.03 and CO2−
3 = 167 ± 11. Because pH
and CO2−
3 were held constant in those experiments, we may assume the dominant influence
−
on B/Ca was a purely [B]SW -driven parameter, in this case B(OH)−
4 /HCO3 . By contrast, in
carbonate system experiments, pH and CO2−
3 both increase, and we see lower B/Ca values
−
than predicted from increasing B(OH)−
4 /HCO3 alone. This may indicate competition with
another ion whose abundance also increases with pH, such as CO2−
3 .
Greater incorporation of CO2−
3 at higher pH has already been proposed to explain carbon
and oxygen isotope values measured in O. universa (Spero et al., 1997; Zeebe, 1999). Carbon
and oxygen isotope composition of foraminiferal calcite is influenced by the proportion of
2−
CO2−
and HCO−
is more enriched in the lighter
3
3 used during calcification because CO3
isotopes of carbon and oxygen than HCO−
3 . Precipitation experiments that change solution
[CO2−
3 ] while holding pH constant could be conducted to test the hypothesis of competition
2−
between B(OH)−
4 and CO3 . B/Ca of the polar foraminiferan Neogloboquadrina pachyderma
(sinistral) may also respond to CO2−
3 (Hendry et al., 2009), suggesting that this phenomenon
may not be limited to O. universa.
43
2.5.4
B/Ca variability within individual shells
Laser ablation profiles of individual shells have revealed that foraminiferal calcite is compositionally heterogeneous, and that the magnitude of observed Mg/Ca variability exceeds
what could be explained by temperature changes due to vertical water column migration
(Eggins et al., 2003, 2004). Similarly, the B/Ca heterogeneity observed in our culture experiments (Figure 2.4) can not be explained by changes in seawater composition or temperature
because these were held constant within each experimental group. Light was the only experimental parameter that varied on a diurnal basis (12 hours of constant light alternating
with 12 hours of constant darkness throughout each experiment), yet in some specimens we
see many alternating bands of high and low B/Ca. Understanding possible controls on such
intra-shell variation is important to proxy development because changes in processes that
control the banding could also influence the overall relationship between shell and seawater
chemistry. To apply modern calibrations to past environments, we need to identify and
evaluate the influence of such additional controls.
As an individual O. universa ages, its symbiotic dinoflagellates divide and increase in
number (Spero and Parker, 1985). Assuming that most calcite is added progressively on
the sphere’s outer surface (Lee and Anderson, 1991; Spero, 1988), observed increases in
B/Ca from the inner to the outer shell could be caused by increasing numbers of photosynthetic symbionts progressively removing more CO2 from the microenvironment during a
foraminifer’s lifetime. This would raise local pH and carbonate ion (Hönisch et al., 2003;
Rink et al., 1998), either or both of which increase B/Ca (Figure 2.1). In some profiles,
B/Ca decreases slightly at the outer shell edge, perhaps reflecting symbiont rejection immediately before gametogenesis. However, low symbiont abundance or absence the day before
gametogenesis was only observed for half of these specimens, so the control on this pattern
remains unclear.
Photosynthesis, respiration, and calcification all occur during the day, but at night photosynthesis ceases while the other two processes continue. Transition between these regimes
44
can create a day-night shift of ∼ 1 pH unit in the microenvironment (Rink et al., 1998),
though this diurnal pH range may vary with specimen size and number of symbionts. Applying our calibration to Rink’s measured day-night pH ranges predicts a B/Ca range of ∼
40 µmol mol−1 . Assuming that our bulk B/Ca-pH trend is applicable to individual calcification bands, the average ∼60 µmol B/Ca range observed in banded profiles might be partially
explained by diurnal microenvironment changes. However, even though all cultured specimens possessed symbionts and should have endured large diurnal pH shifts, some individuals
did not exhibit any diurnal B/Ca variability, which does not support a symbiont-driven B
incorporation pattern.
Although wide variability within our dataset prevents a clear explanation of shell heterogeneity (Figure 2.4), our results raise two important points. One is that individual specimens grown under identical conditions may have different intrashell element profiles. This
highlights the need to combine multiple specimens when applying trace element proxies to
paleoceanographic applications. The second point is that any distribution coefficient derived from bulk shell measurements averages these different compositional bands. Bulk shell
measurements may still largely reflect external seawater conditions, but it is important to acknowledge that any seawater influence on B/Ca must first pass through a variable biological
and microenvironment filter.
2.5.5
Influence of [B]SW on partition coefficients
Another tool for examining calcite precipitation chemistry is the empirical partition coefficient, or D value (DB = [B/Ca]solid / [B/Ca]solution , cf. Morse and Bender, 1990). In our
study, DB ≈ 1.5×10−3 in ambient seawater. If this DB were applicable at higher [B]SW , then
the measured B/Ca of calcite grown in higher-[B] solutions should equal D×[B/Ca]SW . To
test this, we calculate [B/Ca]SW by dividing experimental seawater B concentrations (measured by ICP-MS, Table 1) by Ca (assumed to scale with salinity according to Dickson et al.
(2007); Riley and Tongudai (1967)), and then multiply this ratio by the D value determined
45
from our ambient seawater experiment. For calcite grown in seawater with 20 and 40 ppm
B, this predicts B/Ca values of 297 µmol mol−1 and 578 µmol mol−1 , respectively. Instead,
calcite from our 20 and 40 ppm B experiments has higher B/Ca values of 404 and 899 µmol
mol−1 , resulting in D values of 2.1×10−3 and 2.4×10−3 , respectively. The reason for excess
B/Ca enrichment at elevated boron seawater concentrations is not clear.
Ni et al. (2007) have suggested a calcification-rate control on B/Ca of Globigerinoides
sacculifer and G. ruber based on shell size trends, with higher B/Ca in larger specimens.
Because the observations of Ni et al. were not made on Orbulina, it is not clear whether
the suggested controls are applicable to our results. Because all shells from each experiment
needed to be combined to make a sample large enough for ICP-MS analysis, we could not
evaluate the shell size-B/Ca relationship. Even if separation by shell size had been possible, it would have been very difficult to quantify actual calcification rates, which may vary
throughout the life of a foraminifer. Calcite saturation (Ω) in our different [B] experiments
are the same within uncertainty (2σ), so it seems that elevated calcification rates and preferential incorporation of elements closer to seawater ratios (Morse and Bender, 1990) are not
a likely explanation. Alternatively, the B/Ca-size relationship observed by Ni et al. (2007)
may be related to the depth(s) at which individual foraminifers calcify. Laboratory and field
studies have linked higher light levels with increased symbiont photosynthesis and enhanced
calcification (Bé, 1972; Spero et al., 2003; Hönisch and Hemming, 2004), suggesting that
larger individuals grow at shallower depths. If symbiont activity in larger G. sacculifer and
G. ruber from the Ni et al. study were higher than in smaller individuals, the local pH and
B(OH)−
4 might also have been higher, making conditions favorable for greater B incorporation. This possible control could be tested both with plankton tow and laboratory-based
light experiments.
Another possible explanation for the unexpected B partitioning observed in our culture
experiments may be a modified crystal habit occurring in experimental solutions with greater
than 4 ppm B (Hemming et al., 1995). The appearance of a heavily-striated calcite form
46
with more rounded growth steps such as that observed by Hemming et al. (1995) at high
[B]SW may present a surface more conducive to B adsorption and incorporation. Finally,
we note that although D has provided insight into trace element behavior in other studies
(Elderfield et al., 1996), it may not be as applicable in this case if B and Ca are built into
different lattice sites with different size/charge requirements, and are not directly competing
with each other or with the same types of ions in solution.
2.5.6
Salinity
Our salinity (S) experiments span most of the natural growth range for O. universa, except
very high-salinity habitats like the Red Sea and the Mediterranean (up to ∼ 40 h). Exper-
imental seawater boron concentrations measured by ICP-MS indicate that the B/Salinity
ratio of our experimental seawater differs from the average modern ocean value (0.1336, Lee
et al, 2010) by a maximum of 5.1% and an average of 2.6% (Table 1). We do not consider
this to indicate significant B contamination from glass culture jars either through leaching
or adsorption. In addition, the ratio of boron to salinity measured for the high salinity
experiment was slightly higher than would be predicted from Lee et al. (2010), suggesting
that boron was not lost through boric acid volatilization.
These results show that when salinity is raised by evaporation, [B]SW increases proportionally to salinity (S). The slope defined by [B] experiments predicts a B/Ca increase of 15
µmol mol−1 between 29.9 and 35.4 S (3.84 and 4.55 ppm B, respectively), which agrees with
the actual salinity experimental data within measurement error. Thus although the B/Ca
response to S is somewhat scattered, it might be explained by higher [B]SW . Overall, the
results of our experiments indicate that the concentration of B in solution has a strong and
linear influence on foraminiferal B/Ca, which needs to be accounted for in both modern and
paleo records. If this effect is removed, it decreases the magnitude of glacial-interglacial B/Ca
records by ∼2.6 µmol mol−1 , further exacerbating the low-carbonate-system-sensitivity issue
discussed above in Section 5.2.
47
2.5.7
Temperature
Theoretically, trace element partitioning (described by Keq ) should vary with temperature
(T) (Stumm and Morgan, 1996). However, because the mechanism for B incorporation is
not known, and furthermore since the seawater and calcite phases are not in equilibrium, we
have to rely on empirical relationships. B/Ca values measured on shells grown between 17.7
– 26.5 ℃ agree within error (Figure 2.1), and the slope is statistically indistinguishable
from zero (Student’s t-test, 95% confidence level). Accounting for the slight pH difference
between our lowest and highest temperature experiments (-0.1 unit) raises the B/Ca of
the 26 °C sample from 65 to 68 µmol mol−1 , but does not change the result of our slope
t-test. The lack of a significant influence of temperature on B/Ca (as opposed to KD ,
which can yield artificial relationships), suggests that for paleo-applications, temperature
corrections are not necessary for O. universa shells. This further suggests that greater local
CO2 sequestration by photosynthetic symbionts at higher temperature (as demonstrated by
Bemis et al. (2000)) does not significantly affect B/Ca, at least across the temperature range
covered by our experiments.
The impact of temperature on shell chemistry appears to vary by species: B/Ca increases
with temperature in G. inflata (Yu et al., 2007), but no effect was observed on the highlatitude species Neogloboquadrina pachyderma (sinistral) (Hendry et al., 2009). We recommend that until the mechanism(s) of temperature influence on B/Ca are better understood,
empirical, species-specific temperature influences should be applied to paleo-reconstructions.
2.5.8
Proxy implications
To gain insight into past ocean conditions from fossil calcite records, we must first consider
how and where the calcite formed. Orbulina universa grows several thin, trochospiral chambers as a juvenile, then secretes a large spherical chamber as an adult, and finally adds a
layer of calcite immediately prior to gamete release (“gametogenic calcite”) (Spero, 1988).
The juvenile portion of its shell only comprises 5 - 10 % of the total weight (Spero and
48
Parker, 1985), and is generally too fragile to be preserved in marine sediment. This means
that geochemical signatures within fossil O. universa shells are less likely to have been influenced by ontogenetic (growth) effects than species in which the early-growth or juvenile
calcite is preserved. The adult, spherical chamber initially calcifies upon an organic template, and thickens over the course of ∼ 1 week (Spero, 1988). While most adult calcification
is confined to the photic zone to accommodate the light requirements of its symbionts (Bé
et al., 1982), some gametogenic calcite (∼ 4 – 20 % of total shell mass) may form at greater
depths as individuals sink to the thermocline to reproduce (see Hamilton et al. (2008) for
a full discussion). All specimens for the present study were collected within ∼ 5 meters
of the ocean surface, confirming the presence of this species (both juveniles and adults) in
the upper water column. It has been suggested that a few symbiotic foraminifers may even
calcify below the thermocline (Lohmann, 1995), but this would put these individuals at a
severe disadvantage for survival, calcification and reproduction (Hamilton et al., 2008; Bé,
1972) and such extreme deep-dwellers are rare (Fairbanks et al., 1982). Collectively, these
observations suggest that fossil O. universa calcite is likely to be dominantly composed of
adult-sphere calcite which grew mostly in the photic zone.
−1
If surface ocean pH and CO2−
higher during
3 were ∼ 0.15 pH units and ∼ 80 µmol mol
glacial periods than interglacial periods (Hönisch and Hemming, 2005; Barker and Elderfield,
2002; Sanyal et al., 1995), then based on our calibrations we would expect glacial O. universa
B/Ca to be higher by approximately 5 to 6 µmol mol−1 . Although it is not yet clear whether
one or both (or neither) of these two parameters are truly controlling B incorporation, the
response predicted above is greater than the predicted B/Ca increase of 2.6 µmol mol−1
in response to a glacial salinity increase of ∼ 1h, suggesting that on glacial-interglacial
timescales the carbonate system probably has a stronger influence than salinity on B/Ca.
Given the typical analytical uncertainty of 2 – 5% for B/Ca measurements (Yu et al.,
2007; Foster, 2008), and requiring a B/Ca separation of > 4 –10% to determine a significant (>2σ) difference between measured values, then at or near modern ocean conditions
49
([B/Ca]universa ∼ 62µmol mol−1 ) the smallest pH change resolvable with B/Ca in O. universa
is ∼ 0.1 – 0.2 pH units. This is as large or larger than the pH difference between the Holocene
and Last Glacial Maximum estimated from boron isotopes (Sanyal et al., 1995; Hönisch and
Hemming, 2005; Foster, 2008). Published paleoreconstructions using the foraminifer species
Globigerina inflata, Globigerinoides sacculifer and G. ruber have found maximum differences between glacial and interglacial B/Ca to be between ∼10 and 25 µmol mol−1 (Foster,
2008; Tripati et al., 2009; Yu et al., 2007). This comparison suggests that the B/Ca sensitivity of G. inflata, G. sacculifer, and G. ruber may be higher than that of O. universa,
but the species’ relative sensitivities need to be further tested, and we can not yet rule out
enhancement or dampening of the observed glacial-interglacial B/Ca ranges by competing
factors.
The observed influence of [B]SW on B/Ca in our culture experiments also suggests that
paleo-reconstructions using B/Ca may need to account for secular changes in the average
boron concentration in the ocean. If we apply the relationship observed in culture to the
variable-river-input [B]SW curve of Lemarchand et al. (2000), assuming constant pH of 8
(total scale), we find that the B/Ca adjustments are <1 µmol mol−1 back to 1.2 million
years ago (Ma), increasing to 8 µmol mol−1 at 15 Ma. This suggests that [B]SW changes are
not of great concern until the mid-Miocene. Considering the wide range of B/Ca observed
in different contemporaneous species and the possibility of different responses to [B]SW , this
recommendation does not necessarily apply to all planktic foraminifera.
2.6
Conclusions
1. The carbonate system in seawater exerts a measurable control on B/Ca of cultured O.
universa calcite. Results are broadly consistent with the proxy’s existing theoretical
−
−
framework; however, distinguishing between pH, CO2−
3 , or B(OH)4 / HCO3 control
requires further study.
50
2. Foraminiferal B/Ca increases with seawater B concentration and salinity.
3. Temperature does not significantly influence B/Ca in O. universa across a range of
17.7 – 26.5℃.
4. Laser-ablation profiles through cultured specimens reveal heterogeneous B/Ca distributions. B/Ca tends to increase from the inner to the outer shell surface, and an
anti-correlation with Mg/Ca is observed in some individuals.
5. Existing coretop and downcore B/Ca data present some puzzling discrepancies, including both positive (Yu et al., 2007) and negative (Foster, 2008) KD -temperature
relationships for G. ruber, as well as variable CO2−
3 influences on KD . Correction for
additional controls such as salinity revealed by culturing may remove or reduce these
differences and help establish a firmer and more consistent basis for the B/Ca proxy.
Though foraminifer shells from coretops grew under natural conditions, their seawater
parameters are difficult to constrain. In contrast, culture conditions can not exactly
mimic the foraminifers natural environment, but their growth conditions are very wellknown. Combination of these methods can thus support and enhance the use of B/Ca
as a carbonate system proxy.
2.7
Acknowledgments
We are grateful to all the staff at the USC Wrigley Institute for providing outstanding
technical and moral support during our field work on Catalina Island. These experiments
would not have been possible without the help of Steve Doo, Kelly James, and Lael Vetter,
who were dedicated members of our scuba-diving and laboratory team. We also thank Ann
Russell and Emily Allen for their assistance during the field season, and Jason Day for
analysis of seawater at Cambridge University. The manuscript benefitted from the remarks
of three anonymous reviewers. This research was funded by NSF grants OCE 07-51764 (BH)
51
and 05-50703 (HJS), ARC grant DP0880010 (SME), and ERC grant 2010-Adg-267931 (HE).
3a
3b
1
2
4
5
6
7
8a
8b
9a
9b
10
11
12
Sample
8.12
8.06
7.97
7.61
8.30
8.67
8.06
33.2
33.0
33.0
33.3
33.3
33.3
29.9
31.5
35.4
32.7
33.5
17.7
19.6
26.5
22.1
22.1
22.1
22.1
22.3
22.3
22.3
22.3
8.00
7.96
8.00
8.02
total
8.03
h
33.3
°C
22.3
pH
S
T
0.03
0.04
0.03
0.05
0.03
0.01
0.01
0.02
0.02
0.03
0.03
σ
0.02
±
2385
2485
2915
2163
2233
2229
2237
2058
2425
2803
2040
1
5
8
13
21
5
10
2
4
7
9
σ
5
µmol
kg
2239
±
Alkalinity
191
157
180
165
184
175
183
72
312
600
161
182
µmol
kg
CO2−
3 *
19
19
13
27
17
5
6
7
24
44
16
σ
11
±
4.80
20.06
39.92
4.43
4.45
4.32
4.45
4.64
4.48
4.40
4.21
ppm
4.25
[B]SW
82
303
665
74
82
75
76
35
124
200
74
74
µmol
kg
B(OH)−
4*
8
40
63
12
8
4
2
3
10
15
7
σ
5
±
0.043
0.162
0.353
0.042
0.046
0.042
0.042
0.019
0.074
0.145
0.045
mol mol
0.041
B(OH)−
4
*
HCO3−
−1
61.5
62.3
61.1
62.9
64.9
55.7
69.3
92.0
55.2
52.9
66.5
65.0
71.7
403.8
899.2
µmol
mol
B/Ca
c
σ
1.6
1.6
1.6
1.6
1.7
1.4
1.8
2.4
1.4
1.4
1.7
1.7
1.9
10.5
23.4
±
x 1000
1.48
1.5
1.32
1.51
1.53
2.97
0.93
0.63
1.23
1.17
1.57
1.54
1.68
2.49
2.55
KD *
Table 2.1: Table 1. Results of culture experiments with O. universa. Samples labeled ‘a’ and ‘b’ are analytical splits of material
from a single experiment. B/Ca uncertainties were determined from long-term standard solutions Yu et al. (2007). Uncertainties
for calculated carbonate system parameters (±σ) represent the combined influences of measured pH, alkalinity, and temperature
variance during each experiment.
∗Calculated values. All others are measured.
B
S
pH
T
ambient
Experiment
52
53
Experiment
3a
1
2
4
5
6
7
8
9
10
11
12
ambient
T (°C)
17.7
19.6
26.5
pH (total)
7.6
8.3
8.7
S (h)
29.9
31.5
35.4
B (ppm)
20.1
39.9
Shell weight
µg
54
±
σ
35
Shell diameter
µm
571
±
σ
97
16
25
24
18
24
7
10
411
460
430
100
58
69
19
15
15
25
43
33
22
31
7
478
523
488
93
116
42
15
15
15
39
54
25
29
44
10
536
577
443
97
151
50
20
17
15
36
33
9
33
528
480
46
480
15
16
n
Table 2.2: Average shell weights and diameters for each experiment.
54
55
Chapter 3
The planktic foraminiferal B/Ca
proxy for seawater carbonate
chemistry: A critical evaluation
Note: This chapter has been published in Earth and Planetary Science Letters (2012) Vol.
345-348, pp. 203-211
3.1
1
Abstract
The ratio of boron to calcium (B/Ca) in the carbonate tests of planktic foraminifers has
been proposed as a proxy for surface ocean pH or carbonate ion concentration ([CO2−
3 ]),
with a possible secondary influence of temperature. In both cultured and wild specimens,
−
B/Ca generally increases with [B(OH)−
4 /HCO3 ], consistent with the proxy’s theoretical ba-
sis. However, close examination of the available data reveals that calibrations using the
−
empirical boron partition coefficient, KD = [B/Ca]calcite /[B(OH)−
4 /HCO3 ]seawater , can be
−
driven by independent relationships between [B(OH)−
4 /HCO3 ] and other environmental pa1
AUTHORS: Katherine A. Allen*, Bärbel Hönisch
Lamont-Doherty Earth Observatory of Columbia University, 61 Route 9W, Palisades, NY 10964, USA
corresponding author: [email protected], tel: (845) 365-8668, fax: (845) 359-2931
56
rameters, not by B/Ca itself. If the influence of B/Ca on a calibration is negligible, it follows
that combining that calibration with down-core B/Ca can not yield new information about
past ocean conditions. In this study, we evaluate existing calibrations and down-core records
with the aim of establishing a framework that allows consistent calibration methods to yield
accurate proxy reconstructions from the fossil record. While many issues still need to be
addressed, B/Ca does respond to seawater chemistry, and does change across some major
climate transitions. Because it may provide insight into major carbon cycle perturbations,
we make specific recommendations on how to tackle current uncertainties and develop B/Ca
into a more robust proxy.
3.2
Introduction
The theoretical basis for pH-dependency of B incorporation into marine carbonates is rooted
in the aqueous speciation of dissolved boron. Boron dissolved in seawater exists primarily
as boric acid (B(OH)3 ) and borate (B(OH)−
4 ). As pH increases, the proportion of boron
in the form of borate increases, while boric acid decreases (Dickson, 1990). Because the
boron isotopic composition of marine carbonates falls close to the isotopic composition of
borate in seawater, Hemming and Hanson (1992) suggested that B(OH)−
4 is the dominant
form adsorbed and incorporated into marine carbonates. On this basis, B/Ca of carbonates
is expected to increase with pH.
However, B/Ca behavior differs among marine calcifying taxa, defying a simple explanation. For example, B/Ca varies widely within tropical, aragonitic coral skeletons and is not
directly related to seawater pH (Allison et al., 2010; Sinclair, 2005), instead varying inversely
with temperature in some specimens (Trotter et al., 2011). In some coccolithophorids, B/Ca
even decreases with pH (Stoll et al., 2012). In the calcite tests of the benthic foraminifer
species Cibicidoides wuellerstorfi and Cibicidoides mundulus, B/Ca is more closely related
2−
2−
to ∆CO2−
3 (defined as [CO3 ]insitu - [CO3 ]saturation ) than any other oceanographic param-
57
eter (Rae et al., 2011; Yu et al., 2007, 2010b), an observation that is not predicted by the
theoretical framework of the proxy. B/Ca in several planktic foraminifer species has recently been examined in marine core-top sediments (Foster, 2008; Ni et al., 2007; Yu et al.,
2007), sediment traps (Hendry et al., 2009), down-core sediment records (Foster, 2008; Seki
et al., 2010; Tripati et al., 2009; Yu et al., 2007), and in culture experiments with live
specimens (Allen et al., 2011; Sanyal et al., 1996). These studies reveal that B incorporation is species-specific, where the average modern B/Ca increases from Neogloboquadrina
dutertrei <Globigerina bulloides <Globorotalia inflata <Orbulina universa <Globigerinoides
sacculifer <Globigerinoides ruber, thus hinting at a potential influence of depth habitat (N.
dutertrei, G. inflata, and G. bulloides can live in deeper waters where pH and CO2−
3 tend
to be lower), and/or symbiosis (O. universa, G. sacculifer, and G. ruber have autotrophic
symbionts that consume CO2 and raise local pH) on B incorporation. In addition, questions
have been raised about the influence of temperature (Tripati et al., 2011; Yu et al., 2007),
the possible incorporation of boric acid (Klochko et al., 2009; Pagani et al., 2005; Tossell,
2005), and the influence of other parameters such as carbonate ion or salinity (Allen et al.,
2011; Foster, 2008; Hendry et al., 2009; Tripati et al., 2011; Yu et al., 2007).
Here, we examine the existing evidence for planktic foraminiferal B/Ca as a proxy for
the carbonate system in the surface ocean. While recent studies show promise for the use of
this proxy, we also discuss some problems with the way the empirical B partition coefficient
−
KD has been calibrated and applied (KD = B/Cacalcite / ([B(OH)−
4 ] / [HCO3 ]seawater )). We
synthesize the available data from calibrations and paleo-reconstructions, and conclude with
implications and recommendations for further proxy development.
3.3
Calibrations
Several B/Ca calibrations have been published in recent years using different approaches
and a variety of foraminifer species. Core-top calibrations include Globorotalia inflata and
58
G. bulloides (Yu et al., 2007), G. ruber (white), G. sacculifer, and N. dutertrei (Foster,
2008). B/Ca has also been measured in N. pachyderma (sinistral) collected from Antarctic
sediment traps (Hendry et al., 2009). Calibrations using down-core material include G.
ruber (white) (Tripati et al., 2009; Yu et al., 2007) and G. sacculifer (Tripati et al., 2009).
In core-tops, B/Ca of G. inflata and G. bulloides appears to increase with temperature, but
in G. ruber and G. sacculifer it does not (Figure 13.1). Though B/Ca in many foraminifers
appears to increase with pH, the magnitude of B/Ca increase in response to carbonate
system parameters also varies by species (Figure 3.1, A-C) (Tripati et al., 2011). Isolating
environmental parameters is challenging because several physical and chemical seawater
parameters covary in the surface ocean, making it difficult to discern the parameters’ separate
influences on B/Ca. For example, temperature and carbonate ion (CO2−
3 ) exhibit a broad
positive correlation in the top 100 meters of the water column ocean-wide (Figure 3.2).
Temperature and CO2−
3 concentrations from previous core-top calibration sites follow this
trend (Foster, 2008; Yu et al., 2007), creating uncertainty regarding specific controls on
B/Ca.
Culture experiments with live planktic foraminifers allow isolation of individual chemical
and physical parameters, and provide a different way to test environmental controls. Laboratory cultures reveal that raising seawater pH by addition of base (which raises [B(OH)−
4 ] and
−
[CO2−
3 ], but decreases [HCO3 ]) does increase the boron concentration in O. universa calcite
(Figure 3.1C, Allen et al. (2011); Sanyal et al. (1996)). This is consistent with inorganic
calcite experiments, most core-top calibrations, and with the theoretical basis proposed for
B incorporation (Allen et al., 2011; Sanyal et al., 1996, 2000). The mechanism of this increase with pH is not yet certain, but likely involves the behavior of pH-sensitive ions such
−
2−
as B(OH)−
4 , HCO3 and CO3 , discussed in more detail below. In addition, cultured O.
universa specimens do not record a strong dependence of B/Ca on temperature at constant
pH (Figure 3.1, D-E), suggesting that the inconsistent temperature relationships exhibited
by different foraminifer species from core-top sediments may not reflect species-specific tem-
-1
B/Ca ( mol mol )
59
B
A
120
C
100
80
m
60
40
0
200
400
-1
B/Ca ( mol mol )
120
600
CO3 (mmol kg )
-1
2-
D
0
50
100
150
200
B(OH)4 (mmol kg )
-1
-
E
m
80
60
40
10
20
30
o
Temperature ( C)
30
8.0
8.4
pH (total)
Globigerinoides ruber
Globigerinoides sacculifer
Orbulina universa
Globorotalia inflata (N. Atlantic)
Globorotalia inflata (S. Ocean)
Globigerina bulloides
Neogloboquadrina
pachyderma (sinistral)
100
0
7.6
32
34
36
38
Salinity ( )
−
Figure 3.1: B/Ca from modern samples vs. [CO2−
3 ], [B(OH)4 ], pH, temperature, and salinity,
updated from Tripati et al. (2011). Data shown are from Foster (2008: G. ruber, G.
sacculifer, core-top), Yu et al. (2007: G. inflata, G. bulloides, core-top), Hendry et al.
(2009: N. pachyderma, sediment trap), and Allen et al. (2011: O. universa, cultured). The
wide range of responses indicates species-specific behavior and/or unidentified controls on
B/Ca.
60
perature sensitivities, but rather the influence of other parameters. Before these laboratory
calibrations can be applied, however, they must be ground-truthed in the natural ocean.
Specific recommendations for testing these calibrations are given in Section 5.
-1
Carbonate ion ( mol kg )
350
Pre-industrial surface ocean (0 - 100m)
Foster (2008) calibration sites
Yu et al (2007) calibration sites
300
m
250
200
150
100
0
5
10
15
20
25
30
o
Temperature ( C)
Figure 3.2: Temperature and carbonate ion values for core-top sites studied by Foster (2008)
and Yu et al. (2007), plotted with modern oceanographic data from the upper 100 meters
of the water column (Key et al., 2004), which is the typical depth range of the planktic
foraminifer species.
Here, we discuss the use of KD , an empirical boron partition coefficient. KD was first
derived by Hemming and Hanson (1992) from a substitution equation:
−
B(OH)−
4 + CaCO3 ↔ CaHBO3 + HCO3 + H2 O
(3.1)
In this scenario, borate adsorbed at the CaCO3 surface substitutes for carbonate ion
during CaCO3 growth. Hemming and Hanson (1992) then proposed the following partition
coefficient:
61
KD =
[HBO32− /CO32− ]solid
−
[B(OH)−
4 /HCO3 ]f luid
(3.2)
Later, this equation was simplified to:
KD =
[B/Ca]solid
−
[B(OH)−
4 /HCO3 ]f luid
(3.3)
2+
assuming that in calcite, [CO2−
3 ] ≈ [Ca ] (Yu et al., 2007; Zeebe and Wolfe-Gladrow,
2001). Because the boron content of inorganic calcites (Sanyal et al., 2000) and marine
carbonates (Hemming and Hanson, 1992) appeared to be well-described by constant KD
values, Yu et al. (2007) expected that planktic foraminiferal calcite should be as well.
Instead, measured foraminiferal KD values were not constant, ranging across an order
of magnitude in cultured (Sanyal et al., 1996) and in core-top specimens (Yu et al., 2007).
Based on observed co-variation of B/Ca with the temperature-proxy Mg/Ca, Yu et al. (2007)
attributed core-top KD variability to the influence of temperature on B partitioning. However, this mechanism can not explain the wide KD range observed in foraminifera cultured
at constant temperature (Allen et al., 2011; Sanyal et al., 1996). Even in inorganic calcite
(Sanyal et al., 2000), KD ranges from 1.2 to 2.0 (Figure 3.3). Given these observations, it
seems that the assumption of constant KD should be revisited.
First, empirical constants (K*) apply to systems at equilibrium or steady-state. Because
foraminiferal life processes have been successfully modeled assuming steady-state conditions
(Wolf-Gladrow et al., 1999a; Zeebe et al., 1999), and the diffusional time scale of B ions
(1.5 minutes, assuming a 300 µm boundary layer (Wolf-Gladrow and Riebesell, 1997; WolfGladrow et al., 1999b)) is much faster than the timescale of calcification, it seems reasonable
to expect some constant K* to apply. Second, empirical constants are valid only for the
conditions in which they are measured – e.g., a single temperature, pressure, solution composition, and ionic strength (Stumm and Morgan, 1996). Consequently, changes in KD of
planktic foraminifers with temperature and salinity are to be expected. In contrast, although
62
=0
.00
2
140
0.
00
1
K
D
=
100
K
B/Ca ( mol/mol)
D
120
80
60
KD
05
.00
0
=
40
20
0
0.00
Inorganic calcite
Cultured O. universa calcite
0.04
0.08
B(OH)4 /
HCO3
0.12
-
Figure 3.3: Comparison of inorganically-precipitated calcite (Sanyal et al., 2000) with calcite
from culture experiments with the planktic foraminifer O. universa (Allen et al., 2011).
−
Measured B/Ca values are plotted against solution [B(OH)−
4 /HCO3 ]; thus, lines through
−
the origin represent constant KD values (KD = [B/Ca]calcite/[B(OH)−
4 /HCO3 ]solution). If
samples grow in a steady-state system, and the theoretical incorporation equation is accurate,
KD is expected to remain constant for a given temperature, pressure, solution composition,
and ionic strength. These parameters are held constant in both the inorganic precipitation
and culture experiments, but contrary to expectations, KD values vary with pH.
63
varying pH can affect ion-pairing behavior (Skirrow, 1975), in seawater with a total salt concentration ∼ 10x larger than the reacting ions discussed here (i.e. borate and bicarbonate),
ion-pairing shifts should not affect K* values (Stumm and Morgan, 1996). Although KD
decreases from 0.003 to 0.0006 in laboratory culture (Figure 3.3), which is not consistent
with the expectation above, the almost-linear behavior of KD with seawater-pH still suggests
potential as an empirical tool for determining past seawater carbonate chemistry.
However, an important practical issue with KD is that the division of B/Ca by ([B(OH)−
4]
/ [HCO−
3 ]) can create artificial correlations and obfuscate true ones. This is indicated by the
observation that in some core-top suites, B/Ca in planktic foraminifers does not covary with
temperature or carbonate ion, but KD does (Foster, 2008). To demonstrate the artificial
nature of this correlation, we examine the G. ruber data from Foster (2008) as an example.
Figure 3.4 illustrates that in this dataset, B/Ca does not depend on seawater carbonate
ion, but KD does. The apparent [CO2−
3 ] control on KD stems from the chemical links
−
between CO2−
and the KD denominator, [B(OH)−
3
4 ]/[HCO3 ]. The relative abundances of
−
these dissolved species are linked by equilibrium constants, where [CO2−
3 ] and [B(OH)4 ]
increase, and [HCO−
3 ] decreases with pH. In other words, this calibration curve is primarily
−
2−
−
−
2−
a function of 1/([B(OH)−
4 ]/[HCO3 ]) vs. [CO3 ], where [B(OH)4 ]/[HCO3 ] and [CO3 ] are
closely related. As an additional complication, carbonate ion also happens to be roughly
correlated with temperature in this dataset, making it difficult to discern their separate
influences (Foster, 2008), see Figure 3.2.
The artificial nature of the KD relationships is further illustrated by using the average
B/Ca value of the sample population instead of the measured B/Ca values to determine individual KD values. Using this population-average approach, B/Ca equals 115 µmol mol−1 for
−
each sample, which is then divided by the sample location-specific [B(OH)−
4 ]/[HCO3 ] ratio.
The trend-line resulting from this population-average approach (black line) is essentially the
same as that derived from the measured B/Ca (dashed line), suggesting that this KD calibration is effectively independent of the B/Ca data. It is important to be alert to denominator-
64
2.2
-1
B/Ca ( mol mol )
130
2.0
1.8
m
average
110
1.6
KD =
B/Ca
B(OH) - / HCO -
100
4
KD (x 1000)
120
1.4
3
1.2
220
240
260
280
-1
CO32- (mmol kg )
300
220
240
260
280
-1
300
CO32- (mmol kg )
Figure 3.4: Example of an artificial KD relationship. In this dataset (red circles, panel
A, Foster, 2008), B/Ca does not exhibit a relationship with carbonate ion. A relationship
−
only appears when B/Ca is divided by [B(OH)−
4 ]/[HCO3 ] (panel B). Solid black line in A
represents the population B/Ca average; solid black line in B represents a fit through KD
−
values calculated by combining core-site specific [B(OH)−
4 ]/[HCO3 ] with this average B/Ca
value for each sample. The resulting calibration line is essentially the same as that using
the individual, measured B/Ca data (dashed line), illustrating that this KD relationship is
driven by the denominator, and B/Ca only plays a negligible role in this calibration.
65
driven KD relationships such as these because they do not depend on the primary measured
data (B/Ca) and thus can not provide new paleo-information. This phenomenon is not
unique to (Foster, 2008); other core-top calibrations exhibit denominator-driven behavior as
well (e.g., G. bulloides of Yu et al., 2007).
Similar to these core-top calibrations, B/Ca of down-core calibration datasets (Tripati
et al., 2009; Yu et al., 2007) does not exhibit a relationship with Mg/Ca-derived sea surface temperature (SST) in either G. ruber or G. sacculifer (Figure 3.5A), but KD does
(Figure 3.5B). The positive KD -SST relationship is created solely by dividing B/Ca by
−
[B(OH)−
4 ]/[HCO3 ], which shows a broad negative correlation with SST over the 200-ky
study period. This is the same fundamental problem described above for the KD variation
with [CO2−
3 ] (Figure 3.4), except that in this case it is a down-core rather than a core-top
calibration (as in the G. ruber calibration of Yu et al., 2007).
By contrast, the G. inflata data of Yu et al. (2007) show a direct correlation between
B/Ca and temperature and/or [CO32− ], and in this case the calibration is not solely driven
by the denominator (Figure 3.6). Similar to Figure 3.4, the average B/Ca value of the sample population (thick black horizontal line in 5A) is used to determine KD values. In this
dataset, the KD calibration using the original B/Ca data (circles, dashed line fit) is very different from the calibration using the population average (solid black line), illustrating that
the B/Ca data do have a significant effect on the calibration. A temperature effect on B
incorporation might stem from its influence on the system’s free energy. Unfortunately, this
thermodynamic effect can not be quantified because formation-energy data are not available
2
for CaHBO3 . The temperature and [CO2−
3 ] correlations with raw B/Ca data (R = 0.77 and
0.80, respectively) are stronger than with calculated KD values (R2 = 0.65 for both, Figure
−
3.6), indicating that the division by ([B(OH)−
4 ]/[HCO3 ]) mutes the B/Ca signal, rather than
amplifying it. In light of this correlation decrease, we recommend that the partition coefficient be avoided. Instead, calibrations should be based on primary relationships between
B/Ca and environmental parameters, at least until the relevance of individual ion species on
66
130
A
B
1.4
-1
B/Ca ( mol mol )
120
G. ruber
100
m
90
G. sacculifer
KD (x 1000)
110
1.2
1.0
80
0.8
70
60
26
27
28
29
30
o
Temperature ( C)
26
27
28
29
30
o
Temperature ( C)
Figure 3.5: Down-core calibration. In the Pleistocene calibration dataset of Tripati et
al. (2009), B/Ca does not exhibit a relationship with Mg-derived SST (A), but KD does
(B). Similar to Fig. 3, the positive KD -SST relationship is created by dividing B/Ca by
−
[B(OH)−
4 ]/[HCO3 ]. Black horizontal lines in (A) represent the average of all G. ruber and
G. sacculifer data in the calibration dataset (114 and 74 µmol mol−1 , respectively). If these
−
average values are divided by core-site specific [B(OH)−
4 ]/[HCO3 ] to determine KD , the fits
to the resulting values (solid lines) are similar to fits reported by Tripati et al. (2009), which
use the original B/Ca data (red and blue dotted lines). This implies that in this case, the
B/Ca values have little influence on the calibration.
67
-1
80
1.9
B
A
1.8
75
1.7
70
1.6
65
average
m
1.5
60
1.4
55
50
1.3
2
2
R = 0.80
150
R = 0.65
160
170
180
150
CO3 (mmol kg )
-1
160
170
180
-1
2-
1.9
C
D
1.8
75
1.7
70
1.6
65
m
average
1.5
60
1.4
55
50
KD (x 1000)
B/Ca ( mol mol )
1.2
CO3 (mmol kg )
-1
2-
80
KD (x 1000)
B/Ca ( mol mol )
B incorporation is better understood.
1.3
2
2
R = 0.77
8
R = 0.65
10
12
o
Temperature ( C)
8
1.2
10
12
o
Temperature ( C)
Figure 3.6: Example of primary B/Ca relationships in G. inflata—. In this dataset (Yu et
al., 2007), B/Ca increases with [CO2−
3 ] (A) and temperature (C). In contrast to Figure 3.4,
the KD calibration resulting from using the B/Ca population average (solid black lines, B
and D) is different from that obtained by using the individual, measured B/Ca data (dashed
lines, B and D). This indicates that in this case, B/Ca does have a significant influence on
the calibration. The correlation (quantified by R2 values) of the KD relationships (B and D)
is weaker than that of the B/Ca relationships.
The second issue is the use of a “down-core calibration”, i.e. correlations between B/Ca
and environmental parameters through time. Because temperature and pCO2 covary on some
geologic and climatic timescales (e.g., Petit et al. (1999)), down-core records do not allow
separation of these two variables’ influences on proxies such as B/Ca and δ 11 B (Hemming and
Honisch, 2007). Such ambiguous calibrations can not be used to test past climate sensitivity
68
to CO2 because this approach automatically and inappropriately projects the Pleistocene
covariation of SST and pCO2 into the past, when their relationship may have been different.
For example, when using a Pleistocene KD -T calibration as in Yu et al. (2007) and Tripati
et al. (2009), a high SST corresponds to a high KD , which automatically translates into
low pH and high pCO2 . This forces temperature into a driving role in the reconstruction,
and prevents independent evaluation of SST and pCO2 shifts. Consequently, by applying
down-core calibrations to previous time periods, we can not learn anything new about how
the links between the ocean carbon cycle and climate may have changed through time.
3.4
Down-core records
Down-core records from planktic foraminifer species show no consistent B/Ca pattern for
the past 400,000 years, and G. ruber specimens from different locations yield different B/Ca
values despite similar seawater pH (Figure 3.7, Table 1). As a simple test of B/Ca variation
in Pleistocene records, we calculated the single-species means for glacial and interglacial
periods (as defined by benthic oxygen isotopes, illustrated in Figure 3.7), and found that
glacial and interglacial B/Ca at each site is the same at a 95% confidence level except
for the G. ruber (white) record from Foster (2008), whose means were statistically distinct
(Table 1). Data that fall within Terminations (rapid transitional periods between glacial and
interglacial conditions, as defined by Cheng et al. (2009)) were excluded from these averages.
The lack of consistent glacial and interglacial B/Ca behavior in Figure 3.7 is consistent with
the small glacial-interglacial B/Ca range of 10 µmol/mol predicted from O. universa culture
experiments. This B/Ca range was predicted assuming interglacial and glacial conditions
of 280 and 180 µatm pCO2 , 34.8 and 35.8 salinity, respectively, with alkalinity scaling with
salinity as in Hönisch and Hemming (2005). During the last glacial maximum, B/Ca values in
the G. ruber record of Foster (2008) were lower than Holocene values, which contradicts the
expectation of higher B/Ca at higher glacial pH. An additional consideration is if B/Ca of the
69
foraminifer species used in these reconstructions increases with salinity, as it does in cultured
O. universa (Allen et al., 2011), a salinity effect should be removed before determining pH.
Taking salinity into account for glacial samples – e.g., by subtracting 2.6 µmol/mol B/Ca
per salinity unit increase (Figure 3.1E) would slightly exacerbate the issue of already-low
glacial-interglacial B/Ca sensitivity.
-1
B/Ca ( mol mol )
140
120
G. ruber
100
m
80
G/I range predicted from
cultured O. universa
G. sacculifer
60
G. inflata
d
3.0
interglacial
4.5
5.0
0
100
200
300
O (‰)
4.0
glacial
18
3.5
400
Time (ka)
Figure 3.7: Glacial-interglacial B/Ca records. Data are from Tripati et al. (2009, red circles
and green diamonds), Foster (2008, black circles), and Yu et al. (2007, dark blue circles
and light blue triangles). Benthic foraminiferal oxygen isotope data are from Lisiecki and
Raymo (2005) (black line). B/Ca theory and culture calibrations both predict high B/Ca
during glacial intervals (light blue line) and low B/Ca during interglacial intervals (red line).
This B/Ca range was predicted assuming interglacial and glacial conditions of 280 and 180
µatm pCO2 , 34.8 and 35.8 salinity, respectively, with alkalinity scaling with salinity as in
Hnisch and Hemming (2005), and the salinity effect on B/Ca observed in culture (Figure
3.1E, Allen et al., 2011) taken into account. This predicted behavior is not observed in
down-core records, and there is no consistent B/Ca pattern between species and core sites.
Given the large differences between existing B/Ca records (Figure 3.7) and calibrations,
it is remarkable that these studies all yield pCO2 estimates that closely follow ice-core pCO2
trends (Foster, 2008; Tripati et al., 2009; Yu et al., 2007). This serves as an indirect illustration of how such KD -based reconstructions can be and often are driven by other inputs
70
Study
Species
Yu et al. (2007)
G. ruber
Yu et al. (2007)
G. inflata
Tripati et al. (2009) G. ruber
Tripati et al. (2009) G. sacculifer
Foster (2008)
G. ruber
Glacial
98.5
58.9
111.8
73.0
110.3
n
11
12
22
15
8
Interglacial n Z value
98.8
12 0.101
58.5
12 0.267
115.3
11 1.269
74.3
9
0.931
115.3
19 2.215
Critical value
2.086
2.042
2.042
2.060
2.042
Table 3.1: Comparison of B/Ca averages from glacial and interglacial intervals. The Z value is
a standard test value calculated from the means and standard deviations of the populations
being compared. The two-tailed critical values against which Z is tested are taken from
Borradaile (2003). If the Z value is less than the critical value, the two population means
are indistinguishable at a 95% confidence level. Boundaries between glacial and interglacial
intervals are based on oxygen isotope data, and data that fall within terminations (as defined
by Cheng et al., 2009) are excluded from the glacial and interglacial averages.
to the pCO2 calculation, not by B/Ca itself. In one down-core record from the Caribbean,
the pCO2 estimates derived from combining pH (δ 11 B) with [CO2−
3 ] (B/Ca) are almost indistinguishable from those derived from pH (δ 11 B) and alkalinity (salinity) (Foster, 2008).
This close similarity is due to the dominant influence of pH on pCO2 calculations in the
Pleistocene carbonate system (e.g. Foster, 2008; Hönisch and Hemming, 2005).
In contrast to the Caribbean record (Foster, 2008), where the pCO2 estimate agrees with
ice core records primarily because of a strong relationship between atmospheric pCO2 and
surface ocean pH (recorded by δ 11 B), reconstructions using a temperature-dependent KD
(Tripati et al., 2009; Yu et al., 2007) match ice core records primarily because sea surface
temperature and atmospheric pCO2 covary. In these cases, Mg/Ca-derived temperature
is the main control on pCO2 estimates, with B/Ca itself having very little impact. Tem−
perature does exert a real, natural influence on the abundances of [B(OH)−
4 ] and [HCO3 ]
in seawater through its impact on equilibrium constants K1 , K2 , and KB (Dickson, 1990;
Lueker et al., 2000), but the influence here is very small. The driving mechanism behind the
−
SST-[B(OH)−
4 ]/[HCO3 ] relationship in the case of these glacial-interglacial paleo-records is
the systemic climate correlation between surface temperature and pCO2 (which influences
−
the seawater carbonate system, and thus [B(OH)−
4 ]/[HCO3 ], through air-sea gas exchange).
71
It is the application of this down-core relationship, not the B/Ca record, which artificially
creates large changes in reconstructed pCO2 (as discussed in Section 3). An example of
this phenomenon is illustrated in Figure 3.8. pCO2 estimates in this figure were calculated
according to the temperature-dependent KD method applied by Yu et al. (2007) and Tripati
et al. (2009). To evaluate the impact of B/Ca values on pCO2 estimates, we compared
the original estimates (symbols connected by solid lines) with estimates calculated using the
same B/Ca value for all samples across the whole record (symbols connected by dashed lines,
using average B/Ca of G. ruber = 101 µmol mol−1 for Yu et al. (2007), and G. ruber = 108
µmol mol−1 , and of G. sacculifer = 75 µmol mol−1 for Tripati et al. (2009)). The similarity
of these results illustrates that, in these two cases, the influence of B/Ca on pCO2 estimates
is minimal, and it is the assumptions built into the calibration that dominate the final result.
3.5
Recommendations for proxy development
We agree with the recommendation of Tripati et al. (2011) that higher-resolution records
with greater spatial distribution are needed to assess the link between pCO2 and climate, but
we also echo the caution of Seki et al. (2010), who suggested that applying KD is premature
in the light of uncertainties controlling B incorporation. We propose that until better constraints are placed on the specific incorporation stoichiometry (and by extension, KD ), direct
relationships between raw B/Ca data and well-known modern or recent seawater parameters
should instead be used to interpret B/Ca records. B/Ca does change with pH and carbonate
ion in culture experiments, and does exhibit some variability across major climate transitions (Seki et al., 2010; Tripati et al., 2011), so while many issues still need to be addressed,
the B/Ca proxy may eventually provide insight into past carbon cycle perturbations such
as the Paleocene-Eocene Thermal Maximum or Triassic-Jurassic boundary. Below, we make
specific recommendations on how to tackle current uncertainties and develop B/Ca into a
more robust proxy.
72
pCO2 (ppmv)
400
Yu et al. (2007)
350
300
Measured B/Ca
G. ruber, Yu et al. (2007)
250
G. ruber, Tripati et al. (2009)
200
G. sacculifer, Tripati et al. (2009)
Tripati et al. (2009)
pCO2 (ppmv)
350
Constant B/Ca
B/Ca = 101 mmol mol
300
B/Ca = 108 mmol mol
250
-1
-1
-1
B/Ca = 75 mmol mol
200
150
0
100
200
300
400
Age (ka)
Figure 3.8: Comparison of three paleo-pCO2 estimates using measured B/Ca data (red and
blue symbols, data from Yu et al. 2007 and Tripati et al. 2009) and constant B/Ca values
(black circles and diamonds), corresponding to the average B/Ca values of each species. All
calculations use the temperature-dependent KD calibrations applied in the original studies.
The constant-B/Ca and variable B/Ca methods give similar results, demonstrating the minor
role of B/Ca in the reconstruction and the dominance of other input parameters during this
time interval. Note also that different species yield different pCO2 values, which reflects
species-specific differences in the calibrations.
73
Even if B/Ca is not able to resolve glacial-interglacial shifts in atmospheric pCO2 (at least
not with current analytical precision and with the species investigated to date), it might still
be useful in settings where larger carbonate system shifts occurred. A potential target could
be upwelling regions, where carbonate system shifts can be several times larger than those
driven by atmospheric CO2 . However, the only study available to date measured B/Ca of
G. sacculifer from the northern Arabian Sea (Palmer et al., 2010), a region characterized by
strong seasonal upwelling. In that study, δ 11 B of calcite and δ13C of alkenones both indicate
elevated surface seawater CO2 between 11-17 ka, but B/Ca remained relatively constant,
suggesting that carbonate system changes were not large enough to be detected by this
proxy. Separate and/or competing controls might also explain the lack of B/Ca variability.
More data from other upwelling regions (preferably with tight constraints on species depth
habitats and seasonal production) are needed to more deeply evaluate these possibilities.
3.5.1
Distinguish environmental controls
To establish clearer evidence for or against temperature and/or carbonate ion control, we recommend that these relationships should be tested further using more extensive live-plankton
and core-top comparisons in areas where temperature and [CO2−
3 ] are decoupled either spatially or temporally. In the G. inflata calibration dataset of Yu et al. (2007), temperature
2
and [CO2−
3 ] covary (R = 0.78, Figure 3.2, squares) and their respective relationships with
B/Ca are similar (Figure 3.6A, 3.6C). Yu et al. (2007) discuss this issue, and point to the
inconsistency between KD and [CO2−
3 ] relationships observed in their down-core record at
668B in G. ruber (negative) and in the North Atlantic in G. inflata (positive) as evidence
against CO2−
3 control. They also compare core-top specimens grown at similar temperatures
2−
but different CO2−
3 concentrations (Southern Ocean sites had slightly lower CO3 than the
North Atlantic, Figure 3.1). However, it is difficult to draw a firm conclusion from this comparison because of the two Southern Ocean core top samples, one agrees with B/Ca grown
at the same temperature in the North Atlantic within error while the other does not. Ex-
74
panding upon this work would allow firmer conclusions to be drawn regarding environmental
controls on B/Ca.
For example, surface waters in the Mediterranean and Black Seas span a wide salinity range (corresponding to a large [CO2−
3 ] range) but relatively small temperature range
(MEDAR, 2001), and there are many areas in which these parameters are decoupled. At the
Hawaii Ocean Time-Series (http://hahana.soest.hawaii.edu/), [CO2−
3 ] and temperature appear largely independent (R2 =0.28 for 1989-2010 data, assuming [CO2−
3 ] = alkalinity - DIC).
Such areas might serve as useful testing-grounds for temperature vs. [CO2−
3 ] controls. In
addition, using only modern samples will place these investigations on firmer ground because
modern seawater properties have been measured directly, and thus are better constrained
than past conditions, which must be reconstructed from other proxies.
3.5.2
Investigate the role of test size
Field studies are also needed to investigate the relationship between test size and B/Ca.
In the core-top study of Ni et al. (2007), B/Ca increases with test diameter of G. ruber
(white and pink) and G. sacculifer. The authors suggest that variable calification rates may
explain the observed B/Ca behavior. Trace element partitioning often varies with crystal
growth rates, possibly due to changes in mass transport or surface reaction kinetics (Morse
and Bender, 1990), but the specific influence of calcification rate on B partitioning has not
yet been quantified. Ni et al. (2007) also suggest that habitat depth and proportion of
gametogenic to ontogenetic calcite dissolution may influence B/Ca. However, independent
estimates of shell corrosion, such as shell weights, were not provided, thus preventing unambiguous identification of the controlling factors. It is important to investigate these possible
influences so researchers can make informed decisions about which size fraction(s) to analyze, and to be aware of possible biases in down-core records. For example, previous work
suggests that larger individuals are less affected by dissolution and selecting large shells
therefore minimizes potential dissolution bias (Hönisch and Hemming, 2004; Ni et al., 2007).
75
Yu et al. (2007) observed no B/Ca difference between G. inflata samples deeper than 3.9
km (n=4) and samples shallower than 3.4 km (n=36), but a comparison involving more
species, more samples, and broader spatial coverage would help assess the general applicability of this result. Comparing individuals from sediment traps (or deep vertical plankton
tows) with core-top material at the same sites might further improve our understanding of
dissolution effects. Laser ablation ICP-MS could enable separate analysis of ontogenetic and
gametogenic calcite layers.
3.5.3
Establish species-specific calibrations
It is also important to investigate whether the B/Ca of different species may respond differently to changes in carbonate chemistry. In Figure 3.8, the pCO2 estimates given by
individual species G. ruber and G. sacculifer are offset. While Tripati et al. (2009) com−
bined [B(OH)−
4 ]/[HCO3 ] ratios from both species to construct a final pCO2 record, separate
calculations for each species yield different results. For samples between 0 – 1 Ma, the average difference between G. ruber - and G. sacculifer -derived pCO2 is 70 ppm when using a
temperature-dependent KD . Such species offsets can be avoided by applying species-specific
calibrations. More culture calibrations, sediment trap studies, and core-top studies should
be performed on species used for paleo-reconstructions to determine their individual B/Ca
sensitivity to carbonate chemistry, temperature, and salinity.
3.5.4
Biological processes may influence B uptake
Inorganic calcite growth is primarily controlled by the availability and type of nucleation
surfaces, solution trace element composition, ionic strength, and saturation state (Stumm
and Morgan, 1996). Biogenic calcite growth is generally subject to the same parameters as
inorganic precipitation. However, because organisms have developed many ways to strongly
modify the microenvironments in which they grow, the calcite they produce is often characterized by crystal forms, growth rates, and chemical compositions that are very different
76
from inorganic calcite (Erez, 2003; Weiner and Dove, 2003).
The carbonate chemistry in a foraminifer’s immediate surroundings (∼ 2 mm radius),
especially in spinose species, is influenced by a number of physiological processes. These
include calcification, respiration and, for those that bear symbionts, photosynthesis (Zeebe,
1999). Calcification consumes DIC and alkalinity in a 1:2 ratio, which raises [CO2 ] and
lowers local pH. Respiration consumes oxygen and produces CO2 (lowering pH), while photosynthesis consumes CO2 (raising pH) and produces oxygen. Photosynthesis exerts only
a minor influence on alkalinity through its consumption of nitrate (Wolf-Gladrow et al.,
2007). As a result of these processes, foraminifer species hosting photosynthetic symbionts
(such as the dinoflagellate Gymnodinium beii ) can experience shifts from pH 8.7 to 7.7 (total scale) under light and dark conditions, respectively (Jørgensen et al., 1985; Rink et al.,
1998). Such shifts in microenvironment pH alter local aqueous boron speciation, and could
ultimately influence the abundance of B incorporated into calcite, as suggested by boron
isotope analyses of symbiont-bearing foraminifer shells grown in the light and near-darkness
(Hönisch et al., 2003). Experiments to test the effects of these physiological processes on
the B/Ca proxy have not yet been performed, but, in theory, first insight into the potential
for such effects can be gained from laser-ablation ICP-MS profiles through individual tests.
These profiles sample across individual calcite layers precipitated alternately during day and
night throughout a foraminifer’s lifetime. Such analyses of natural and laboratory-grown
foraminifer tests display wide variability, with B/Ca as high as 110 µmol mol−1 and as low
as 50 µmol mol−1 within a single shell (Allen et al., 2011). This suggests a strong biological
control on B incorporation, but the observed patterns can not yet be linked to any specific
process (Allen et al., 2011; Hathorne et al., 2009). Based on B/Ca variations with test size
in core-tops, Ni et al. (2007) have suggested that B incorporation may be influenced by
calcification rate. Varying rates of calcification could lead to intra-test B/Ca variation, but
this has not yet been tested in culture.
77
3.5.5
Constrain past seawater composition
In the modern ocean, boron behaves as a conservative element, with a consistent boron to
salinity ratio of 0.1336 ± 0.0005 mg kg−1 * S−1 , which equals 433 µmol B kg−1 seawater
at S = 35 (Lee et al., 2010). Boron fluxes between major reservoirs have not yet been
well-constrained (for review see Tripati et al., 2011). The relatively poor constraints on the
secular evolution of the boron cycle limit our ability to apply boron-based proxies to earlier
times in Earth history. While relative changes within short intervals (<1 million years) are
possible, absolute estimates comparing events that are several million years apart require a
much improved understanding of the boron sources and sinks through time. Some paleoδ 11 Bseawater estimates have been made (e.g., Pearson (1999)), but higher-resolution studies
and paleo-[B] estimates are needed.
It may be important to consider not only the concentration of B in ancient seawater, but
of other elements as well. The balance of B(OH)3 and B(OH)−
4 in solution (described by the
equilibrium constant pKB ) depends on temperature, pressure, salinity (Dickson, 1990)), and
solution composition (Dyrssen and Hansson, 1973). Potentiometric measurements suggest
that borate ion activity (i.e., the tendency to interact with other dissolved chemical species)
varies with solution composition (Bassett, 1980; Hershey et al., 1986; Simonson et al., 1987).
For example, pKB values determined in pure Na-Cl, Na-Mg-Cl, and Na-Ca-Cl solutions at
ionic strength 0.65 (Hershey et al., 1986) give borate concentrations of 53, 69, and 76 µmol
mol−1 , assuming a total pH of 8 and total boron (BT ) of 433 µmol kg−1 . The results from this
study imply that dissolved B may have behaved differently in ancient oceans with higher or
lower Mg, Ca, Na, and/or K concentrations. However, the impact of such variable aqueous
B behavior on the incorporation of B into fossil biogenic calcite is uncertain. All studies to
date have assumed that the activity of dissolved B species has been constant through time,
or at least that any changes did not impact B incorporation.
78
3.6
Conclusions
In planktic foraminifer culture experiments, inorganic precipitation experiments, and most
sediment core-tops, B/Ca of calcite increases with seawater pH. This increase is consistent
with the theoretical basis proposed for B incorporation. The specific mechanism for observed
B/Ca behavior is uncertain, but may involve competition between the pH-sensitive anions
−
2−
B(OH)−
4 , HCO3 , and/or CO3 for inclusion in the calcite lattice.
Relationships between KD and seawater parameters can sometimes be driven by the
−
denominator ([B(OH)−
4 ]/[HCO3 ]), and not by B/Ca. Reconstructions based on such B/Ca-
independent relationships are susceptible to being driven by other environmental parameters.
Application of the empirical boron partition coefficient, KD , should be avoided until more
is known about the relative influences of different chemical species on B incorporation. We
recommend that where no direct relationships between B/Ca and environmental parameters
can be observed, B/Ca-based reconstructions should not be attempted.
Temperature-KD calibrations that are based on down-core records of both temperature
and ocean carbon chemistry inherently contain the relationship between SST and pCO2 of
that particular time period. Such calibrations thus can not be used to test how the SST-pCO2
relationship may have changed in earlier periods of Earth history.
In Pleistocene down-core records, planktic foraminiferal B/Ca exhibits species- and sitespecific offsets but no consistent temporal pattern. B/Ca response to recent shifts in seawater
carbonate chemistry (e.g., glacial-interglacial cycles) might be too small to be detected with
current methods. However, more dramatic carbonate system shifts should cause larger,
more pronounced B/Ca variation in the planktic fossil record. As the controls on this proxy
become clearer, planktic foraminiferal B/Ca may yield new insight into past surface ocean
chemistry.
79
3.7
Acknowledgments
We are grateful to Bob Anderson, Peter deMenocal, Jesse Farmer, Gary Hemming, James
Rae, Taro Takahashi, and Jimin Yu for helpful discussions. We thank Harry Elderfield, Gavin
Foster, Aradhna Tripati, and an anonymous reviewer for thoughtful, insightful reviews that
greatly improved this manuscript. This work was supported by NSF grant OCE 07-51764.
80
81
Chapter 4
Environmental controls on B/Ca in
calcite tests of the tropical planktic
foraminifer species Globigerinoides
ruber and Globigerinoides sacculifer
Note: This chapter has been published in Earth and Planetary Science Letters (2012) Vol.
351-352, pp. 270-280
4.1
1
Abstract
The ratio of boron to calcium (B/Ca) in the calcite tests of planktic foraminifers may serve
as a proxy for past seawater chemistry, but controls on B incorporation are not yet certain.
Here we present the results of laboratory culture experiments with live specimens of Globigerinoides ruber (pink) and Globigerinoides sacculifer, which provide new insight into B
1
AUTHORS: Katherine A. Allena *, Bärbel Hönischa , Stephen M. Egginsb , Yair Rosenthalc
Lamont-Doherty Earth Observatory of Columbia University, 61 Route 9W, Palisades, NY 10964, USA
b
Research School of Earth Sciences, The Australian National University, Canberra ACT, Australia
c
Institute of Marine and Coastal Science, Rutgers University, New Brunswick, NJ 08901, USA
corresponding author: [email protected], tel: (845) 365-8668, fax: (845) 359-2931
a
82
incorporation controls. We find that in G. sacculifer, B/Ca increases with increasing pH
2−
−
(lower [HCO−
3 ], higher [CO3 ] and [B(OH)4 ]), but decreases with increasing total dissolved
2−
−
inorganic carbon (DIC) (higher [HCO−
3 ] and [CO3 ], constant [B(OH)4 ]). This suggests
competition between aqueous boron and carbon species for inclusion into the calcite lattice.
Similar to previous experiments with the subtropical-temperate Orbulina universa, B/Ca
increases with salinity, but not with temperature. We evaluate possible carbonate system
control parameters, and compare our tropical culture calibrations with new and published
core-top data.
4.2
Introduction
The ratio of boron to calcium (B/Ca) in fossil calcite tests of planktic foraminifers has
recently been used to investigate the carbonate chemistry of ancient oceans (Foster, 2008;
Palmer et al., 2010; Seki et al., 2010; Tripati et al., 2009; Yu et al., 2007). The theoretical
basis for this proxy stems from the pH-dependent concentration of dissolved borate (B(OH)−
4)
and its subsequent incorporation into foraminiferal calcite (Hemming and Hanson, 1992).
Controls on boron incorporation in several planktic species have been investigated through
core-top studies (Foster, 2008; Ni et al., 2007; Yu et al., 2007), sediment-trap samples (Hendry
et al., 2009), and culture experiments (Allen et al., 2011; Sanyal et al., 2001). Results vary by
region and by species, and no consensus has yet been reached on the influence of temperature
or individual carbonate system parameters (Foster, 2008; Yu et al., 2007).
B/Ca has been applied as a proxy for either the concentration of carbonate ion ([CO2−
3 ])
−
in seawater (Foster, 2008), or the ratio of aqueous borate to bicarbonate ([B(OH)−
4 ]/[HCO3 ]),
which can be used to estimate seawater pH if total dissolved boron and carbon concentrations are also known (Palmer et al., 2010; Tripati et al., 2009; Yu et al., 2007). Combination
of [CO2−
3 ] or pH with another parameter (such as alkalinity or total dissolved inorganic carbon) then allows the entire carbonate system to be constrained. However, because carbonate
83
system parameters often covary in modern seawater, core-top and culture experiment studies to date have not been able to tease apart their separate influences. Yu et al. (2007)
suggested that B incorporation increases with temperature in Globorotalia inflata and Globigerina bulloides derived from core-top sediments. By contrast, in Globigerinoides ruber
and Globigerinoides sacculifer, B/Ca appears to decrease with temperature (Foster, 2008).
In cultured Orbulina universa, B/Ca does not change significantly across an 8°C temperature
range (95% confidence level, Allen et al., 2011). Culture experiments also reveal that B/Ca
in O. universa increases with salinity (Allen et al., 2011)), but evaluation of a salinity effect
on existing core-top specimens is difficult due to the correlation of salinity with other variables in the modern surface ocean. In summary, it is not clear whether planktic foraminiferal
−
−
B/Ca is controlled by temperature, salinity, [CO2−
3 ], [B(OH)4 ]/[HCO3 ], or some combina-
tion of these or other parameters. To improve confidence in the interpretation of B/Ca, we
need to understand which carbonate system parameter(s) controls B incorporation, and to
quantify the influence of other environmental variables such as temperature and salinity.
Some basic principles provide a starting place when considering the process of trace element incorporation. Typically, relative rates of ion adsorption and desorption during crystal
growth determine the amount of a chemical species (e.g., B(OH)−
4 ) that is ultimately incorporated into a solid (e.g., calcite). For inorganic calcites, adsorption scales with ion flux to
the surface and the availability of kink sites, while detachment depends on the strength of
an ion’s bond to its neighbors, which is a function of temperature (De Yoreo and Vekilov,
2003). This suggests that dissolved ion concentrations and temperature both have the potential to influence B incorporation. However, to interpret geochemical signatures from the
fossil record, we need to understand the trace element composition of calcite that has been
secreted by a living organism. This requires that we consider an additional set of controls,
because marine organisms exert considerable energy to nucleate and build carbonate tests
from seawater (for a review, see Weiner and Dove (2003)). Several strategies are used by
different calcifying species to create conditions favorable for mineral growth, which include:
84
raising the saturation state (e.g., by increasing ionic strength or pH), regulating the concentrations of inhibitors or impurities, forming a preliminary amorphous carbonate phase, or
providing organic templates for carbonate formation (Erez, 2003; Feng, 2011; Weiner and
Addadi, 1997, 2011). These processes, especially those that influence the seawater carbonate
system of a foraminifer’s microenvironment, might influence B incorporation.
Here, we present the results of culture experiments with live specimens of G. ruber and
G. sacculifer. We discuss both biological and inorganic processes that may control B incorporation into foraminiferal calcite. Because foraminiferal calcite does not typically form in
equilibrium with seawater with respect to trace elements (Elderfield et al., 1996), we test
and discuss empirical relationships rather than equilibrium partition coefficients. Finally, we
compare B/Ca results from culture experiments with new sediment core-tops from the Gulf
of Mexico and with published core-top results from the Atlantic Ocean.
4.3
4.3.1
Methods
Culture experiments
Culture experiments were performed at the University of Puerto Rico’s Marine Sciences
Center on Isla Magueyez during March and April 2010. Scuba divers hand-collected juvenile
foraminifers 8 nautical miles offshore (17°52’ N, 66°58’ W) between 2 – 6 meters water depth.
During the collection period, average surface water temperature was 27.8°C (range 27.5 –
28.5°C), and salinity was 35.4 (range 35.0 – 35.8) on the practical scale (Lewis and Perkin,
1978; Poisson and Curie, 1981). Immediately after collection, foraminifers were brought to
the laboratory and identified. Specimens were measured under a light microscope (maximum diameter) and then transferred to experimental seawater. If no food was observed
in foraminifers’ spines, they were fed a one-day-old brine shrimp (Artemia salina). Subsequently, G. ruber (pink) and G. sacculifer juveniles were each fed a live Artemia nauplius
every 2 days.
85
Most seawater used in these experiments was collected offshore and filtered (0.8 µm) to
remove large particles. The one exception was our low-DIC experiment, which consisted of
half natural and half artificial seawater. Experiments were designed to vary either temperature, salinity, or carbonate system parameters. In each experiment, ∼20 G. sacculifer and
∼50 G. ruber specimens were individually cultured in 120 mL glass jars. All jars were placed
in temperature-controlled water baths topped with cool-white fluorescent lamps whose output was measured biweekly with a light meter. A strict diurnal light cycle was maintained
throughout all experiments: 12 hours light, 12 hours dark. Over the course of 1 – 3 weeks,
foraminifers added new chambers to their calcite tests to accommodate cytoplasm growth.
Specimens were observed daily through a 10x hand lens to determine spine and symbiont
presence, chamber formation, and degree of chamber-fill by cytoplasm. After an individual
underwent gametogenesis, the remaining empty calcite test was rinsed in de-ionized water,
dried, and archived for later analysis. The maximum diameter after incubation was measured
to determine growth.
We lowered the salinity of ambient seawater to 33 by adding de-ionized water, and raised
salinity to 40 by partial evaporation under a heat lamp at ∼60°C. Temperatures of 24, 26, and
30°C were established in constantly-circulating water baths whose conditions were monitored
by HOBO TidbiT® temperature loggers every 5 minutes. To raise or lower seawater pH,
we added NaOH or HCl, respectively. In these experiments, raising or lowering pH also
raised or lowered the dissolved [CO2−
3 ]. To test the effect of varying carbonate ion separately
from pH, we also conducted experiments that maintained constant pH but changed [CO2−
3 ]
by varying the concentration of DIC. To raise DIC, we dissolved 0.75 grams NaHCO3 in
4L ambient seawater, and then immediately added NaOH to return the solution to ambient
pH of 8.0 (total scale). To lower DIC, we mixed two liters of ambient seawater (DIC=2033
µmol kg−1 ) with two liters of synthetic seawater (DIC=0 µmol kg−1 ) and titrated pH back
to ambient seawater conditions (for synthetic seawater recipe, see Appendix A). Final DIC
was not estimated assuming linear mixing of these two solutions, but instead was calculated
86
from pH and alkalinity measurements made on the final natural-synthetic seawater mixture
(details below).
To minimize gas-exchange with the atmosphere, jars were filled without a gas headspace,
and then sealed with Parafilm and tight-fitting (snap-cap) plastic lids. To monitor potential
atmospheric CO2 exchange during feeding (when caps must be removed), alkalinity and pH
were measured at the beginning (2-3 water samples) and end of each experiment (3-8 water
samples). Alkalinity and pH were measured using a Metrohm 809 open cell auto-titrator
and pH meter, calibrated against NIST buffers and Dickson-certified alkalinity standards.
Calculation of the seawater carbonate system was performed with the MATLAB program csys3.m (Zeebe and Wolf-Gladrow, 2004), modified by K. Allen to allow input of total
dissolved boron (Uppström, 1974). Standard deviations (1σ) for experimental temperature
were calculated from continuous TidbiT meter data measured every 5 minutes throughout
each experiment (n ≈ 10,000). Standard deviations for alkalinity and pH were calculated
from suites of measurements made on experimental seawater throughout each experiment
(average n = 13). The 1σ measurement error given by Orion Star Thermo Scientific for
the conductivity meter was 0.1. Confidence bounds (±1σ) for each parameter (temperature,
salinity, pH and alkalinity) were propagated through the full MATLAB calculation, yielding upper and lower estimates for all calculated carbonate system parameters (e.g., [CO2 ],
[CO2−
3 ]). Finally, these upper-lower estimate ranges were combined to yield a composite
uncertainty value according to the equation below:
q
σA = (AU − AL )2T emp + (AU − AL )2Salinity + (AU − AL )2pH + (AU − AL )2Alkalinity
(4.1)
where A is the calculated parameter (e.g., [CO2−
3 ]), and subscripts U and L indicate
upper and lower estimates.
87
4.3.2
Gulf of Mexico samples
Sediment box cores were collected in 2007 from the Garrison (PE07-2) and Fisk Basins
(PE07-5, PE07-6), and in 2003 from the Pigmy Basin (PBBC-1) by the R/V Longhorn
(Richey et al., 2011). A radiocarbon-based age model was established by Richey et al. (2007),
indicating a core-top age of ∼1950 A.D. Orbulina universa and G. sacculifer specimens
between 500-710 µm diameter were picked from surface sediment samples. Surface seawater
properties for these core sites were interpolated from modern transects published by the Gulf
of Mexico and East Coast Carbon Cruise (Peng, 2007). Alkalinity and DIC measurements
were used to estimate the rest of the carbonate system using a modified version of csys3.m
(same as above). Surface ocean DIC values could not be corrected for industrial CO2 invasion
because CFC data were not available for these sites.
4.3.3
Sample preparation
The final calcite tests produced by foraminifers in our culture experiments were composed
of chambers grown in the open ocean before collection (under unconstrained conditions),
and chambers added after insertion into experimental seawater (known conditions). To
analyze only the calcite that grew in experiments, we used a medical scalpel blade and a
light microscope to amputate the calcite chambers that were grown after collection under
controlled conditions. Amputated chambers from each experimental condition were pooled
to make a single sample. On average, each chamber weighs 20 µg for G. sacculifer and 5
µg for G. ruber, requiring approximately 20 to 80 chambers per experiment, respectively, to
yield a total of 400 µg of calcite.
Cultured samples were cleaned according to the methods of Russell et al. (2004), using
two 30-minute treatments of a hot, buffered hydrogen peroxide solution (equal parts 0.1N
NaOH + 30% H2 O2 , 70-80°C) to remove organic matter.
Core-top samples were crushed, then rinsed and sonicated with Milli-Q+ water 5x (or
until no fine particles were visible) to remove clays. Samples were treated with a reductive
88
solution (anhydrous hydrazine, ammonium hydroxide, and citric acid) to remove oxides, and
an oxidative solution (hydrogen peroxide and sodium hydroxide) to remove organic matter
following the modified protocol of Boyle and Keigwin (1985) as outlined in Rosenthal et al.
(1997).
Finally, all cleaned samples were rinsed 3 times with ultrapure, B-filtered Milli-Q+ water,
leached with 0.001N HNO3 , and rinsed again 3 times with Milli-Q+ . Immediately prior to
analysis, samples were dissolved in 0.65N HNO3 (trace element grade).
4.3.4
Analysis
Elemental ratios of calcite samples were determined on a sector-field inductively coupled
plasma mass spectrometer (Thermo Scientific Element XR) at Rutgers University. Precise
and accurate determinations of B/Ca ratios may be compromised by both high B blanks and
a memory effect. To reduce blank levels during the analysis, we used a cyclonic quartz spray
chamber coupled with a sapphire injector, and we minimized memory effects by injecting
ammonia gas directly into the spray chamber at a flow rate of 0.07 ml min−1 (Al-Ammar
et al., 2000). All reagents were made with B-free water. Typical B blank was ∼0.02 ppb
or less than 3% of the foraminiferal content throughout the run. Standard solutions with
differing [Ca2+ ] (1.5 - 8 mM) were run to quantify and correct for any matrix effects on trace
element ratios. The influence of [Ca2+ ] on B/Ca was minor. Within-run precision of the
matrix-corrected B/Ca data was 2.7, 1.2, and 3.0% as determined by duplicate analysis of
three consistency standards at the beginning and end of the run.
89
4.4
Results
4.4.1
Culture experiments
Carbonate system
Between the low pH (7.49 ± 0.05) and high pH (8.45 ± 0.03) experiments, B/Ca increased
from 134 to 212 µmol mol−1 in G. ruber, and from 87 to 137 µmol mol−1 in G. sacculifer
(Figure 4.1a). In these experiments, adding NaOH or HCl shifted not only pH but also the
−1
aqueous speciation of boron and carbon: [CO2−
(Figure
3 ] increased from 66 to 498 µmol kg
−1
4.1b); [B(OH)−
(Figure 4.1c); [HCO−
4 ] increased from 31 to 177 µmol kg
3 ] decreased from
1896 to 1564 µmol kg−1 , and alkalinity increased from 2061 to 2755 µmol kg−1 (Table 1).
Globigerinoides ruber
Globigerinoides sacculifer
Orbulina universa
-1
B/Ca ( mol mol )
a
c
b
200
150
m
100
50
7.6
8.0
8.4
pH (total)
8.8
0
200
400
600
-1
CO3 (mmol kg )
2-
0
50
100 150 200
-1
B(OH)4 (mmol kg )
-
Figure 4.1: B/Ca of G. ruber, G. sacculifer and O. universa grown in experiments where pH
2−
was shifted by adding acid or base. In these experiments, [B(OH)−
4 ], [CO3 ], and alkalinity
all increase with pH. The B/Ca behavior of different species is similar, but offset.
The high-pH G. sacculifer experiment yielded enough calcite that we were able to combine
several final sac chambers to make a separate sample. This final-chamber calcite had lower
90
B/Ca (94 ± 5 µmol mol−1 ) than non-sac chambers from the same experiment (145 ± 8
µmol mol−1 ) (Table 1). A weighted average of sac and earlier chambers (based on initial
calcite sample weights) yields a with-sac value of 133 µmol mol−1 . This value is provided to
facilitate direct comparison with other G. sacculifer samples whose sac chambers were not
amputated.
Separate experiments that varied DIC while keeping pH constant at 8.0 (total scale)
revealed a negative correlation between DIC and B/Ca, which decreased from 144 to 72
µmol mol−1 between 1046 and 4196 µmol DIC kg−1 seawater (Figure 4.2).
c
0
2
-1
6
8
10
12
14
constant pH
increasing pH
160
B/Ca ( mol mol )
4
140
s
e
as
re
120
-
)4
c
] in
H
(O
[B
100
80
[B(OH
)4 - ] is
60
100
200
300
2-
cons
tant
400
500
600
-1
CO3 ( mol kg )
Figure 4.2: When DIC is raised by addition of NaHCO3 and seawater pH is held constant,
2−
[B(OH)−
4 ] remains constant and [CO3 ] increases. In these experiments, B/Ca of G. sacculifer decreases (black triangles). By contrast, when seawater pH is raised and DIC is held
2−
constant, both [B(OH)−
4 ] and [CO3 ] are increased, and B/Ca increases (blue triangles).
2−
These results suggest that B(OH)−
4 and CO3 ions may compete for the same site in the
calcite lattice.
91
Temperature and Salinity
B/Ca does not exhibit a significant relationship with temperature in G. ruber (R2 = 0.38,
p = 0.57) or G. sacculifer (R2 = 0.02, p = 0.87) (Figure 4.3). In both species, B/Ca (µmol
mol−1 ) increases with salinity:
(B/Ca)G.ruber = 4.51 × S − 23.6
R2 = 0.91, p = 0.20
(4.2)
(B/Ca)G.sacculif er = 5.10 × S − 85.9
R2 = 0.96, p = 0.02
(4.3)
Standard deviations of water bath temperatures, measured every 5 minutes for the duration of each experiment (lasting 3 - 5 weeks) were less than ± 0.5 °C. Between the beginning and end of each experiment, pH drifted no more than 0.05 units. Exceptions are the
low-salinity experiment, in which pH decreased by 0.15 units and the low-pH experiment,
in which pH increased by 0.07 units. These changes were likely caused by CO2 exchange
during brief intervals when lids were opened for feeding, and may be strongest in these experiments because they were farthest from equilibrium with atmospheric pCO2 . For display and
interpretation of all results, we used the average of initial and final experimental conditions.
4.4.2
Gulf of Mexico core-top samples
Estimated average surface seawater conditions (0 - 50 meters below sea surface) for both coretop locations are 27.9°C, 35.9 salinity, 2040 µmol kg−1 DIC and 2386 µmol kg−1 alkalinity.
G. sacculifer and O. universa B/Ca values fall between 111 - 122 µmol mol−1 and 71 - 76
µmol mol−1 , respectively (Table 2).
92
180
Globigerinoides ruber
Globigerinoides sacculifer
Orbulina universa
-1
B/Ca ( mol mol )
160
140
slope = 4.5
2
R = 0.91, p = 0.20
2
R = 0.38, p = 0.58
120
m
100
slope = 5.1
2
R = 0.96, p = 0.02
2
R = 0.02, p = 0.87
80
slope = 2.6
60
2
2
R = 0.63, p = 0.03
R = 0.62, p = 0.11
40
30
32
34
36
Salinity
38
40
16
18
20
22
24
26
28
30
o
Temperature ( C)
Figure 4.3: B/Ca increases with salinity in all three cultured species: G. ruber (R2 = 0.91,
p = 0.20), G. sacculifer (R2 =0.96, p = 0.02), and O. universa (R2 =0.63, p=0.03). B/Ca
does not exhibit a strong relationship with temperature in G. ruber (R2 = 0.38, p = 0.58)
or G. sacculifer (R2 = 0.02, p = 0.87), and increases slightly in O. universa (R2 = 0.62, p
= 0.11).
93
4.5
4.5.1
Discussion
Carbonate system
Adding acid (base) to experimental seawater resulted in lower (higher) pH, [B(OH)−
4 ] and
[CO2−
3 ] of the solution, and lower (higher) B/Ca of both G. sacculifer and G. ruber calcite
(Figure 4.1). These observations are consistent with B/Ca incorporation behavior of O.
universa (Allen et al., 2011), and suggest a common control across these species. Higher
2−
seawater [B(OH)−
4 ] could have led to greater B inclusion, and/or higher [CO3 ] could have
caused more rapid calcification rates, which might also lead to greater incorporation of B. To
investigate these possibilities, we performed another set of experiments with G. sacculifer, in
−
which we raised [CO2−
3 ] by increasing DIC while holding pH and [B(OH)4 ] constant. These
experiments reveal that, all else being equal, B/Ca decreases as [CO2−
3 ] increases (Figure
4.2).
These results imply that calcite saturation state (Ωc = ([Ca2+ ][CO2−
3 ])seawater /
([Ca2+ ][CO2−
3 ])saturation ) is not a major control on B/Ca. Between the low and high DIC
experiments, Ωc increased from 2.2 to 11.6. The saturation state increase between the low
and high pH experiments was similar: from 1.6 to 12.0. Although we could not measure
foraminiferal calcification rates directly, calcification rate is expected to increase with saturation (DePaolo, 2011; Zuddas and Mucci, 1998). If so, calcification in both the “high
[CO2−
3 ]” seawater (Ωc = 11.6) and the “high pH” seawater (Ωc = 12.0) should have been
more rapid. However, the high DIC experiment yields very low B/Ca, and high pH yields
very high B/Ca. Instead of a saturation state control, the B/Ca of calcite appears to be primarily influenced by the relative abundances of B(OH)−
4 and carbonate species in seawater.
2−
−
In the following discussion, we test different combinations of [B(OH)−
4 ], [HCO3 ] and [CO3 ]
as control parameters. If one or some combination of these parameters consistently controls
B/Ca, such a relationship may be useful for paleoclimate reconstructions.
2−
First, we consider a scenario in which B(OH)−
4 and CO3 are competing with each other
94
for inclusion in the calcite lattice. If B/Ca increases with [B(OH)−
4 ] and decreases with
−
2−
[CO2−
3 ], then B/Ca should exhibit a positive relationship with [B(OH)4 ]/[CO3 ]. How2−
ever, B/Ca increases with [B(OH)−
4 ]/[CO3 ] in the pH experiments, and decreases with
2−
[B(OH)−
4 ]/[CO3 ] in the DIC experiments (Figure 4.4G). This inconsistent behavior indi2−
cates [B(OH)−
4 ]/[CO3 ] is an incorrect or incomplete control parameter.
−
−
−
If we replace [CO2−
3 ] with [HCO3 ] (giving [B(OH)4 ]/[HCO3 ]), then the B/Ca relation-
ships in the pH and DIC experiments become consistent (Figure 4.4H). This unified B/Ca
−
behavior suggests that HCO−
3 is the dominant carbon species in competition with B(OH)4
for inclusion in the calcite lattice, which is consistent with the original incorporation stoichiometry hypothesized by Hemming and Hanson (1992).
−
−
In a third scenario, CO2−
3 and HCO3 both compete with B(OH)4 during calcite growth.
Near ambient seawater pH (∼ 8.0 total scale), [CO2 ] is very small, so DIC ≈ [CO2−
3 ]
+ [HCO−
3 ]. In Figure 4.4I, B/Ca of both pH and DIC experiments is plotted against
−
−
−
[B(OH)−
4 ]/DIC. The [B(OH)4 ]/DIC and [B(OH)4 ]/[HCO3 ] calibrations are very similar,
which is not surprising because [HCO−
3 ] is the dominant component of DIC in seawater
between pH 6 and 9. Given the available data, it is difficult to discern which calibration
parameter ([HCO−
3 ] or DIC) most accurately quantifies the control on B incorporation. A
−
2−
key unknown is the amount of CO2−
3 incorporated relative to HCO3 . If CO3 incorporation
−
is negligible, then [B(OH)−
4 ]/[HCO3 ] is the more appropriate control parameter. On the
−
other hand, if a mixture of CO2−
3 and HCO3 is incorporated into calcite, then the reality
−
−
may lie somewhere between [B(OH)−
4 ]/[HCO3 ] and [B(OH)4 ]/DIC.
2−
For now, let us assume that both HCO−
are involved in calcification, and
3 and CO3
explore the possibility of [B(OH)−
4 ]/DIC control on B/Ca. In three species - G. ruber
and G. sacculifer (this study), and O. universa (Allen et al., 2011) - B/Ca increases with
[B(OH)−
4 ]/DIC. This relationship may be expressed as a simple linear regression of the form
B/Ca = m*([B(OH)−
4 ]/DIC) + b. This allows a slightly better fit than a multivariate regression of the form B/Ca = m1 *[B(OH)−
4 ] + m2 *DIC + b. Here, we report linear fits to
95
culture data that have been normalized to S = 35:
B
B(OH)−
4
=
× 1152 + 104
Ca G.ruber
DIC
B
B(OH)−
4
=
× 853 + 63
Ca G.sacculif er
DIC
B
B(OH)−
4
=
× 473 + 53
Ca O.universa
DIC
R2 = 0.89
(4.4)
R2 = 0.79
(4.5)
R2 = 0.97
(4.6)
where B/Ca is in µmol mol−1 and [B(OH)−
4 ]/DIC is unitless. Salinity normalization
was done for G. ruber and G. sacculifer using equations 1 and 2, respectively, and for O.
universa using the equation given in Figure 3a from Allen et al. (2011). Equation 4 includes
the high- and low-DIC experiments, and all G. sacculifer data include sac chambers (to be
consistent with the other with-sac samples, the weighted-average of amputated sac-chambers
and ontogenetic calcite B/Ca is used as the high-pH experiment value). These equations will
be tested against core-top B/Ca data, and their down-core applicability will be discussed in
Section 5.7.
4.5.2
Salinity
B/Ca increases with salinity at rates of 4.5 and 5.1 µmol mol−1 per salinity unit in G. ruber
and G. sacculifer, respectively. A positive relationship was also observed in O. universa (2.6
µmol mol−1 per salinity unit, Allen et al., 2011). This raises the question: if all dissolved
elements are concentrated equally as salinity is raised, why does B incorporation increase?
Below, we explore a few possibilities.
Salinity influences the equilibrium constants of dissolved boron and carbon species, described by KB , K1 , and K2 (Dickson, 1990; Lueker et al., 2000; Mehrbach et al., 1973). In our
culture experiments, salinity ranged from 33 to 40, corresponding to [B(OH)−
4 ]/DIC of 0.041
and 0.048, respectively. Given this increase in [B(OH)−
4 ]/DIC of ∼0.007, the calibrations
from carbonate system experiments predict B/Ca increases of 9 and 7 µmol mol−1 for G.
110
100
90
80
70
60
A
B
C
D
E
F
G
H
I
220
200
180
160
140
m
G. sacculifer
-1
B/Ca ( mol mol )
m
G. ruber
-1
B/Ca ( mol mol )
m
O. universa
-1
B/Ca ( mol mol )
96
140
120
100
80
0.2
0.4
0.6
B(OH)4
2CO3
-
0.8
0.05
0.10
B(OH)4
HCO3
0.15
-
0.05
B(OH)4
DIC
0.10
-
Figure 4.4: B/Ca of O. universa, G. ruber, and G. sacculifer plotted against three different
dissolved ion ratios. G. sacculifer was the only species with which we performed increasing
pH/constant DIC experiments (blue triangles) and constant pH/increasing DIC experiments
2−
(black triangles). B/Ca patterns in the two experiments differ when [B(OH)−
4 ]/[CO3 ] is
−
plotted as the controlling parameter (A), but become consistent when [B(OH)−
4 ]/[HCO3 ] is
−
used (B). If DIC replaces [HCO3 ], the relationship is similar (C).
97
ruber and G. sacculifer, respectively. However, the increases observed in salinity experiments
are larger: 33 and 35 µmol mol−1 . This implies that the influence of salinity on equilibrium
constants can not fully explain our observations, and that there is an additional salinity
effect at work.
The B content of mollusk shells also increases with salinity (Furst et al., 1976)) by
approximately 1.7 µmol mol−1 B/Ca per salinity unit, though the mechanism for this effect
is uncertain. Because the calcification strategies of mollusks and foraminifera differ (Feng,
2011), their common B increase may be most easily explained by a common, inorganic
mechanism rather than a biogenic one. Indeed, Kitano et al. (1978) observed that the boron
content of inorganically-precipitated calcite crystals increases with salinity. Because Kitano
et al. (1978) raised salinity simply by modifying the concentration of NaCl (not dissolved
B), the resulting increase in [B] of calcite points to some mechanism involving ionic strength
and/or ion-ion or ion-surface interactions. It is possible that the mechanism for a salinity
effect on the B content of inorganic calcite crystals is also responsible for the response seen in
marine biominerals. Although it is difficult to compare the results of Kitano et al.’s inorganic
experiments with our cultures quantitatively, because the pH of their solutions was not held
constant during calcite precipitation (ranging from 7.5 - 8.2, uncertain scale), their results
do point towards an inorganic origin of the salinity effect on B/Ca.
A salinity effect on B incorporation might stem from the influence of seawater ionic
strength on individual ion activities. The ionic strength of seawater scales with the total
amount and charge of ions dissolved in solution. It influences the activity (i.e. effective
concentration, or tendency to react) of different ions in different ways because each ion has
unique properties that dictate its behavior in solution – its attraction to other ions, its
2−
−
hydration sphere, etc. If B(OH)−
4 , CO3 , and HCO3 compete for inclusion into calcite’s
carbonate ion site, then changes in the relative activities of these three ions with salinity
might explain the observed change in B incorporation. Using an ion pairing model, Millero
−
and Schreiber (1982) estimated activity coefficients of free, unpaired B(OH)−
4 , HCO3 and
98
CO2−
3 in seawater between salinities of 1 - 40. In a solution of ionic strength 0.68 (average
seawater ≈ 0.7) and 25°C, approximately 56% of B(OH)−
4 is free, the remaining 44% being
complexed with Na, Mg, and Ca ions (Byrne and Kester, 1974). We can combine Millero
and Schreiber’s coefficients with typical ion concentrations in seawater to estimate free ion
activities according to the equation
α = γi × mi
(4.7)
where γi is the activity coefficient of chemical species i, m is the concentration, and α is
the activity (effective concentration). This calculation yields different ion activities at higher
−
−
salinity: CO2−
3 activity remains roughly constant, while B(OH)4 and HCO3 activities both
increase (Figure 4.5). Higher ion activity leads to greater incidence of bonding with the
growing calcite surface (De Yoreo and Vekilov, 2003), so higher borate activity should lead
to higher B incorporation into the solid. However, the effective concentration increase in
−
−
HCO−
3 is much larger than that of B(OH)4 (Fig. 5). If HCO3 is the dominant species
competing with B(OH)−
4 for inclusion into the calcite lattice, then it seems unlikely that
changes in ion activity can explain our observed increase in B/Ca with salinity. On the
other hand, the activity of carbonate ion increases less with salinity than that of borate ion.
This may be consistent with a role of carbonate ion in foraminiferal calcification.
In sum, the B/Ca increase observed in salinity experiments appears too large to be
explained by salinity’s effect on equilibrium constants or ion activities. While we can not
fully explain its origin, we recommend that the influence of salinity on B/Ca be considered
when interpreting sediment records.
4.5.3
Temperature
Temperature influences the equilibrium constants K1 , K2 , and KB , which in turn affect the
[B(OH)−
4 /DIC] ratio of seawater. In prior culture experiments performed with O. universa on
99
-1
Concentration and Activity [ mol kg ]
-
HCO 3
1000
-
HCO3
m
2-
CO 3
100
-
B(OH) 4
-
B(OH) 4
10
2-
CO3
15
20
25
30
35
40
Salinity
Figure 4.5: The ionic strength of seawater increases with salinity, and influences the activity
(i.e. effective concentration, or tendency to react) of dissolved ions. In this figure, ambient
2−
−
B(OH)−
4 , CO3 , and HCO3 concentrations (gray lines) have been combined with activity
coefficients in seawater to estimate activities of the free, unpaired ions (black lines). The
activity of all three ions increases at higher salinity, but the activity of HCO−
3 increases to
−
2−
a much greater degree than that of B(OH)4 and CO3 . These results can not explain the
observed salinity effect on B/Ca. See text for details.
100
Catalina Island (Allen et al., 2011), calculated [B(OH)−
4 /DIC] of culture seawater increased
by 0.01 between 18 and 26°C. This difference, according to our carbonate system calibrations,
should translate into a B/Ca increase of 4 µmol mol−1 for O. universa. Between 24 and 30°C
in Puerto Rico, [B(OH)−
4 /DIC] increases by 0.008. This difference should translate into B/Ca
increases of 9 and 7 µmol mol−1 for G. ruber and G. sacculifer, respectively. These estimated
differences are small and consistent with the insignificant temperature relationships of B/Ca
in G. ruber (R2 = 0.38, p = 0.57), G. sacculifer (R2 = 0.02, p = 0.87) and O. universa (R2 =
0.62, p = 0.11). Thus, although B/Ca is expected to increase slightly due to temperature’s
influence on equilibrium constants, changes are too small to be resolved by current analytical
techniques across the temperature ranges provided in our experiments. Our observations
confirm the earlier inference of Foster (2008), who argued that temperature can not be the
primary control on B/Ca because core-top and down-core samples from the Atlantic are
characterized by markedly different B/Ca-temperature relationships.
4.5.4
Species differences
When grown at ambient seawater pH (∼8.0, total scale), G. ruber, G. sacculifer, and O.
universa tests yield B/Ca values of 142, 93, and 62 µmol mol−1 , respectively (corresponding
to 140, 91, and 66 µmol mol−1 when normalized to S=35 using the calibrated salinity relationships). These large and consistent B/Ca differences between foraminifer species grown
in the same seawater composition indicate that these organisms actively influence B incorporation. A better understanding of the biological controls would improve our ability to
predict whether primary seawater variables (like salinity) will be unaffected by the organism’s biological controls and be preserved in fossil calcite.
The boron isotopic composition of foraminiferal calcite increases with seawater pH, and
individual species’ calibrations are also offset from each other (Hönisch et al., 2007). At any
given seawater pH value, cultured O. universa have both lower B/Ca and lower δ 11 B than
G. sacculifer (Sanyal et al., 2001, 1996). It is possible that these species offsets are caused
101
by internal biological controls such as ion pumping, but such processes are currently not well
understood in planktic foraminifers. It is interesting to note that across a wide seawater pH
range (∼ 7-9), δ 11 B offsets between species are roughly constant, but B/Ca offsets are not.
It is also interesting that regressions of B/Ca against these environmental control parameters (Fig. 1, 3 and 4, equations 1-6) do not intersect the origin. This is not an artifact of
−
−
using [B(OH)−
4 ]/DIC as the control parameter, as calibrations versus [B(OH)4 ]/[HCO3 ] yield
similar results. In a conceptual model where B in calcite comes exclusively from B(OH)−
4 in
solution, B/Ca should approach zero when [B(OH)−
4 ] approaches zero. It is possible that the
relationships are not linear, or that the relationship between B/Ca and seawater composition
is complicated by active biological inclusion or exclusion of ions. It is also possible that a
small amount of boric acid (B(OH)3 ) is incorporated at low pH (Allen et al., 2011; Klochko
et al., 2009). More work is needed to rigorously assess this possibility, and to determine the
best form of the curve fit.
4.5.5
Geochemical peculiarity of G. sacculifer ’s sac-like chamber
Globigerinoides sacculifer often produces a final, “sac-like” chamber a day or two before gametogenesis (Bé, 1980; Caron et al., 1990). A few hours before releasing gametes, individuals
(both with and without this chamber) drop their spines and secrete a final biomineral layer
over the entire outer test surface. This layer, known as gametogenic calcite, is estimated
to contribute up to ∼ 30% of the total test weight (Bé, 1980; Caron et al., 1990). Final
sac chambers amputated from individuals in our high-pH experiment yield a B/Ca value of
94 µmol mol−1 , ∼43 µmol mol−1 lower than non-sac chambers from the same experiment.
While it should be noted that we did not analyze final non-sac chambers, and therefore
cannot determine whether this observation is exclusive to sac chambers or universal to all
final chambers, potential explanations for B depletion in these chambers involve changes in
symbiont activity and/or calcification mechanism prior to gametogenesis.
Symbiont photosynthesis removes CO2 from a foraminifer’s microenvironment, raising
102
11
local pH and [CO2−
3 ]. This enhances calcification (Bé et al., 1982), increases δ B (Hönisch
et al., 2003), and decreases test δ 18 O (Bemis et al., 1998). In G. sacculifer, the sac chamber
has higher δ 18 O than the rest of the test, and the enrichment is greater when the chamber
is secreted closer in time to gamete release (Spero and Lea, 1993). If a foraminifer possesses
fewer symbionts when the sac chamber is formed (or if it has sunk into deeper water with
lower light levels), photosynthetic CO2 sequestration would be lower, leading to lower [CO2−
3 ]
and pH, and perhaps higher δ 18 O and lower B/Ca of foraminiferal calcite. Changes in
symbiont numbers through a foraminifer’s life cycle are not well known, but in order to
survive, a foraminifer’s symbionts must abandon their host before it sinks too far from the
photic zone. If abandonment begins before or during final-chamber formation, [B(OH)−
4]
and [B(OH)−
4 ]/DIC within a foraminifer’s microenvironment would likely decrease. If the
pH change due to loss of symbionts were of the same magnitude as that observed between
darkness and full light conditions (Jørgensen et al., 1985), [B(OH)−
4 ] would decrease by ∼90
−1
µmol mol−1 and [B(OH)−
4 ]/DIC by ∼0.04 mol mol . According to Equation 4, this would
lead to a lowering of B/Ca by approximately 40 µmol mol−1 . This is similar to the difference
observed between sac and ontogenetic chambers. To test this mechanism, microelectrodes
could be used to monitor pH within the G. sacculifer microenvironment throughout the
life cycle. Previously-applied techniques for counting symbionts require destruction of both
symbionts and host (Spero and Parker, 1985), but if non-destructive methods were applied,
symbiont counts throughout the lives of individuals could be accomplished and help resolve
this issue.
Differences in the formation of gametogenic and ontogenetic calcite might also explain
the lower B/Ca of sac-chamber calcite. After spines are discarded, the rhizopodial network
collapses and gametogenic calcite forms (Bé, 1980). This calcite is lumpy and uneven (Bé,
1980; Bé and Hemleben, 1970), suggesting a different biomineralization process from the
neatly-layered and structured ontogenetic calcite. If gametogenic calcite is deposited evenly
across the entire test, the final chamber would likely receive a greater proportion than the
103
other chambers, because it is larger and thinner, and does not have as many ontogenetic
layers as the older chambers (Bé and Hemleben, 1970; Erez, 2003). However, the B/Ca
composition of G. sacculifer ’s gametogenic calcite is uncertain. In O. universa, B/Ca tends
to decrease towards the outermost part of the test, but not always (Allen et al., 2011), and
the boundary between gametogenic and ontogenetic calcite in specimens analyzed to date is
not clear.
4.5.6
Comparison with core-top data
To evaluate whether our culture calibrations are applicable to core-top specimens, we use
Equations 4, 5 and 6 as well as oceanographic data to predict B/Ca of the coretop G. ruber
and G. sacculifer studied by Foster (2008), Ni et al. (2007), Tripati et al. (2009), and our
samples from the Gulf of Mexico. To compare directly the culture and core-top B/Ca data,
we must account for a systematic analytical B/Ca offset between data measured at Rutgers
University, where most cultured samples were analyzed, and Bristol University, where most
core-top data were measured. The conversion equation is derived from an inter-laboratory
comparison between Rutgers and Cambridge (Yair Rosenthal and Jimin Yu, personal communication), including calcite from both G. ruber and G. sacculifer :
B
Ca
= 1.085 ×
Rutgers
B
Ca
− 1.09
(4.8)
Cambridge
We assume that this equation (R2 = 0.94, n = 7) is also applicable to data derived
from Bristol, because no B/Ca offset has been observed between Bristol and Cambridge
(Rae et al., 2011). On average, these adjustments raise Cambridge and Bristol G. ruber, G.
sacculifer, and O. universa B/Ca values by ∼7%.
To estimate the pre-industrial conditions in which foraminifers likely grew, Foster (2008)
subtracted anthropogenic CO2 reported by the Global Ocean Data Analysis Project (Key
et al., 2004) from DIC, and Ni et al. (2007) assumed surface ocean CO2 equilibrium with
104
the pre-industrial atmosphere. Invasion of anthropogenic CO2 in the Gulf of Mexico has not
been quantified, but the modern age of the sediment at our core-top sites (Richey et al.,
2007) suggests that adjustments likely to be minor or unnecessary.
The resulting comparison between predicted and measured B/Ca values is illustrated in
Figure 4.6. All measured B/Ca values, both cultured and core-top, have been normalized
to S=35 and, if necessary, converted to the Rutgers B/Ca scale using equation 8. After
accounting for salinity differences, there is good agreement between predicted and measured
B/Ca (R2 = 0.7, p ¡ 0.01). Most measured core-top values for O. universa and G. sacculifer
are similar to values predicted by culture calibrations (average offsets are 4 and 15 µmol
mol−1 , respectively) but values predicted for coretop G. ruber deviate by up to 60 µmol
mol−1 from predicted values. The greater discrepancy observed for core-tops may suggest
that our experiments still fall short of identifying all environmental controls on B/Ca and/or
that we need to revisit the growth conditions assumed for planktic foraminifers, in particular
the depth habitat of G. ruber.
We consider the following possible explanations for discrepancies between predicted and
measured B/Ca of core-top samples. First, some of the core-top sediments may have accumulated as early as 4,000 years ago (Foster, 2008). Surface seawater parameters ascribed
to core-top sites were measured during the 20th and 21st centuries (Key et al., 2004), and
although the surface carbonate system values in these datasets have been corrected for
industrial CO2 invasion (Sabine et al., 2004), it is unclear how well these pre-industrial
oceanographic data approximate mid-Holocene conditions.
Second, even if surface seawater conditions at the time of core-top deposition are wellapproximated by pre-industrial surface conditions, it is still difficult to determine the seawater environment in which core-top-derived planktic foraminifers grew. Foraminifers migrate
vertically through the water column during their life cycle, which can expose them to different seawater conditions. At all core-top sites in Figure 4.6, B(OH)−
4 /DIC decreases with
depth (calculated using the GLODAP database). Thus, the offset between predicted and
Cultured
G. ruber pink (this study)
G. sacculifer (this study)
O. universa, Allen et al. (2011)
1
200
1:
-1
Measured B/Ca ( mol mol )
105
m
Core-top
G. ruber white, Foster (2008)
G. ruber white, Ni et al. (2007)
G. ruber white, Tripati et al. (2009)
G. ruber pink, Ni et al. (2007)
150
G. sacculifer, Foster (2008)
G. sacculifer, Ni et al. (2007)
G. sacculifer, Tripati et al. (2009)
G. sacculifer, Gulf of Mexico
(this study)
100
O. universa, Gulf of Mexico
(this study)
50
50
100
150
200
-1
Predicted B/Ca (mmol mol )
Figure 4.6: Culture calibrations are combined with surface seawater properties (GLODAP
database, Key et al., 2004) to predict foraminiferal B/Ca values at several core-top locations
using culture calibration equations 4, 5, and 6 (R2 =0.7). After normalizing to S=35, most
B/Ca predictions are similar to measured values (Foster, 2008; Ni et al., 2007; Tripati
et al., 2009), with G. ruber exhibiting the largest offsets. Differences between predicted
and measured B/Ca may be due to uncertain foraminifer depth habitats, differences in
age between seawater measurements and core-top sediment age, calibration limitations, or
an unidentified environmental control on B/Ca. Gray bar with arrows represents the B/Ca
range expected in G. ruber for a pH range of 8.0 - 8.15 (assuming T=25°C, S=35, DIC=1950
µmol kg−1 ). All B/Ca shown here have been converted to the Rutgers scale according to
equation 7 (see text for details).
106
measured B/Ca for G. ruber might be explained if these specimens lived at greater depths
than assumed. Applying our G. ruber calibration equation to depth profiles of B(OH)−
4 /DIC
at each core-top site allows us to predict B/Ca values of foraminifers growing at different
depths in the water column. Comparing the core-top B/Ca values measured by Foster (2008)
with these estimated profiles yields habitat depths of 61, 48, 41, and 38 meters for sites 925,
664, 668, and 999A, respectively. These depths differ from the widely adopted surface – 10
m growth habitat for G. ruber (e.g., Foster, 2008) but are within the habitat ranges observed
in plankton tow studies (Fairbanks et al., 1980; Williams et al., 1981), and may represent
an average of depths at which the foraminifers calcified.
Third, plankton tow and sediment trap studies have found that the depth habitats, seasonal abundances, and carbon and oxygen isotope signatures of the pink and white varieties
of G. ruber sometimes differ from each other (Deuser and Ross, 1989; Deuser et al., 1981;
Kuroyanagi et al., 2008; Tedesco and Thunell, 2003). However, B/Ca values determined
from pink and white G. ruber specimens in Ni et al. (2007) generally agree within measurement uncertainty. The white variety of G. ruber was not present in the waters off Puerto
Rico during our 2010 field season, therefore we could not compare B/Ca of pink and white
specimens in culture experiments.
Finally, B/Ca predictions based on culture calibrations may not match core-top B/Ca if
there are additional controls that have not yet been identified in culture, or if the parame−
−
ter [B(OH)−
4 ]/DIC (similar to [B(OH)4 ]/[HCO3 ]) is not an accurate approximation of the
carbonate system’s influence on B incorporation.
4.5.7
Future proxy application and development
Culture experiments have provided new insight into controls on B incorporation, but can this
study’s calibrations be useful to paleoceanographers? Here, we discuss some uncertainties
and recommend directions for future proxy development and application.
In our temperature and salinity experiments, foraminifers experienced conditions that
107
fall within the natural range of modern surface ocean variability. In contrast, our carbonate
system experiments subjected foraminifers to pH and DIC extremes that exceed natural
modern ranges (Table 1). As previously discussed, it was necessary to grow foraminifers
under these extreme conditions in order to distinguish the influence of separate carbonate
system controls on B/Ca. Now that B/Ca has been found to increase with borate and
decrease with DIC (or [HCO−
3 ]), more experiments are needed to better quantify B/Ca
relationships with these parameters within their natural range. In G. sacculifer and G.
ruber, the rate of B/Ca increase with [B(OH)−
4 ] is not constant. With only 3-5 data points,
it is difficult to determine whether the relationship is exponential, or subject to a specific
sensitivity threshold. A higher-resolution calibration with more points across the natural
seawater pH range is needed to answer these questions.
In the meantime, what can be done with the existing calibrations? As discussed above,
carbonate system controls need to be further tested and refined. B/Ca may respond to seawa−
−
ter [B(OH)−
4 ]/[HCO3 ], [B(OH)4 ]/DIC, or some variation of these. In theory, if total dissolved
−
boron (BT ) and DIC are known in addition to temperature and salinity, [B(OH)−
4 ]/[HCO3 ]
may be used to estimate seawater pH (e.g., Tripati et al., 2009; Yu et al., 2007). However,
lack of discernible B/Ca sensitivity to temperature in our culture experiments suggests that
such temperature-dependent KD calibrations may need to be revisited (see also Allen et al.,
in press).
For the sake of example, if we assume the [B(OH)−
4 ]/DIC equations are accurate, then
seawater B(OH)−
4 concentrations can be estimated from B/Ca and DIC:
[B(OH)−
4 ] = DIC ×
(B/Ca − b)
m
(4.9)
and seawater DIC can be estimated if both [B(OH)−
4 ] and B/Ca are known:
DIC = [B(OH)−
4]×
m
(B/Ca − b)
(4.10)
108
However, due to the limited number of data points used in each calibration, uncertainties
regarding the ability of these calibrations to predict DIC or [B(OH)−
4 ] are large. The 95%
prediction bands for equation 7 span ∼ 100 µmol kg−1 [B(OH)−
4 ]. If the uncertainty of an
individual sample ranges from 50 - 150 µmol [B(OH)−
4 ] per kg seawater, this corresponds to
a 0.6 unit range in estimated pH, roughly 3 times greater than the surface ocean pH change
between glacial and interglacial periods (Hönisch and Hemming, 2005). Although the cor−
−
relations between B/Ca and both [B(OH)−
4 ]/DIC and [B(OH)4 ]/[HCO3 ] appear promising
(R2 ¿ 0.75), the low number of calibration points translate into large predictive uncertainties.
For example, with our three G. ruber points we can not yet say at a 90% confidence level
whether the fit coefficients in Equation X are positive or negative. More data are needed to
improve the predictive power of these calibrations.
It is also important to note that the range of DIC in the G. sacculifer dataset (1046 4196 µmol kg−1 ) is much larger than in the G. ruber and O. universa pH-experiments (in
which DIC was constant at 2034 ± 30 and 1986 ± 3, respectively). This means that the
G. ruber and O. universa calibrations may not be applicable to systems or time periods
when DIC was significantly different from these mean experimental DIC values. High- and
low-DIC experiments with these species are needed in order to test calibration accuracy.
4.6
Conclusions
Culture experiments with live specimens of G. sacculifer indicate that B/Ca of foraminiferal
calcite depends on the relative abundance of dissolved [B(OH)−
4 ] and inorganic carbonate
species in seawater. B/Ca increases as seawater [B(OH)−
4 ] and pH are raised (while DIC is
held constant), but decreases as DIC is raised (while pH and [B(OH)−
4 ] are held constant).
This suggests competition between borate and carbonate species for inclusion into the crystal
lattice. Fossil foraminiferal calcite B/Ca thus may reflect the relative abundance of these
chemical species in ancient seawater. In cultured G. ruber and G. sacculifer, B/Ca does not
109
increase with temperature, but increases with salinity at rates of 4.5 and 5.1 µmol mol−1
per salinity unit, respectively. Normalizing B/Ca data to a constant salinity (e.g., S=35)
should improve our ability to isolate the carbonate chemistry signal in B/Ca paleo-records
and samples from different ocean sites. Final sac chambers amputated from G. sacculifer
specimens yield lower B/Ca than older chambers from the same specimen, suggesting that
the B/Ca composition of gametogenic calcite may be depleted in B compared to the rest of
the test.
4.7
Acknowledgments
This work would not have been possible without the dedicated efforts of Kelly Strzepek,
Rebecca Norman, Juan Pablo D’Olivo Cordero, Anand Patel, and Steve Doo, who were
members of our fearless scuba-diving and laboratory team in Puerto Rico. We are grateful
to the members of the University of Puerto Rico’s Isla Magueyez Laboratories for their
support, especially Milton Carlo and Godoberto Lopez, who ensured our safety on and under
water. Many thanks to Tali Babila and Katherine Esswein at Rutgers University for sharing
their time and analytical expertise, and to Julie Richey and Ben Flower at the University of
South Florida for generously providing the Gulf of Mexico core-top samples. This research
was supported by a Columbia Climate Center grant (KA and BH), NSF OCE 07-51764
(BH), and ARC DP8800010 (SE). The manuscript benefitted from thoughtful comments of
two anonymous reviewers.
1
2
4
5
7
12
13
G. ruber (p)
Ambient
Low Temp
High Temp
Low Sal
High Sal
Low pH
High pH
25.7
24.0
29.3
25.7
25.5
25.7
25.7
25.7
25.7
24.0
29.3
25.7
25.5
25.7
25.6
25.7
25.7
25.7
25.7
25.7
Temp
°C
0.4
0.2
0.2
0.2
0.3
0.5
0.5
0.4
0.4
0.2
0.2
0.2
0.3
0.2
0.3
0.5
0.5
0.5
0.5
0.5
±
35.4
35.6
35.2
33.0
40.0
35.5
35.4
35.4
35.4
35.6
35.2
33.0
40.0
35.4
35.4
35.5
35.5
35.4
35.4
35.4
Salinity
1800
1823
1809
1730
1963
1896
1564
1800
1800
1823
1809
1730
1963
947
3694
1896
1896
1564
1564
1564
µmol
kg
HCO−
3
61
36
83
195
77
44
95
61
61
36
83
195
77
104
158
44
44
95
95
95
±
221
212
220
184
265
66
498
221
221
212
220
184
265
91
480
66
66
498
498
498
µmol
kg
CO2−
3
22
11
34
77
19
14
38
22
22
11
34
77
19
32
36
14
14
38
38
38
±
92
90
89
78
108
31
177
92
92
90
89
78
108
76
96
31
31
177
177
177
µmol
kg
B(OH)−
4
9
5
14
33
7
6
14
9
9
5
14
33
7
27
6
6
6
14
14
14
±
DIC
2033
2047
2042
1927
2240
2005
2065
2033
2033
2047
2042
1927
2240
1046
4196
2005
2005
2065
2065
2065
µmol
kg
47
31
53
131
78
43
64
47
47
31
53
131
78
84
166
43
43
64
64
64
±
ALK
2342
2344
2347
2183
2609
2061
2755
2342
2342
2344
2347
2183
2609
1210
4757
2061
2061
2755
2755
2755
µmol
kg
36
28
9
26
41
15
21
18
18
14
4
26
41
29
89
15
15
21
21
21
±
5.3
5.1
5.4
4.5
6.1
1.6
12.0
5.3
5.3
5.1
5.4
4.5
6.1
2.2
11.6
1.6
1.6
12.0
12.0
12.0
Ωc
0.5
0.3
0.8
1.9
0.4
0.3
0.9
0.5
0.5
0.3
0.8
1.9
0.4
0.8
0.9
0.3
0.3
0.9
0.9
0.9
±
8.04
8.04
7.98
7.99
8.04
7.49
8.45
8.04
8.04
8.04
7.98
7.99
8.04
7.93
8.07
7.49
7.49
8.45
8.45
8.45
pH
total
0.01
0.03
0.04
0.12
0.02
0.05
0.03
0.06
0.06
0.03
0.04
0.12
0.02
0.10
0.02
0.05
0.05
0.03
0.03
0.03
±
141.9
127.5
139.2
121.4
154.8
134.2
212.4
95.9
90.0
91.9
92.9
84.3
119.0
144.0
72.3
86.2
88.4
145.0
129.5
94.2
µmol
mol
B/CaR
7.8
7.0
7.7
6.7
8.5
7.4
11.7
5.3
4.9
5.1
5.1
4.6
6.5
7.9
4.0
4.7
4.9
8.0
7.1
5.2
±
131.8
118.6
129.3
112.9
143.6
124.7
196.7
89.4
83.9
85.7
86.6
78.7
110.7
133.7
67.6
80.4
82.5
134.7
120.4
87.8
µmol
mol
B/CaCR
Table 4.1: Seawater properties and B/Ca results for culture experiments. Each experimental sample comprises approximately
40-50 G. ruber (pink) and 20-30 G. sacculifer specimens. Elemental analyses were performed on samples composed only of
calcite grown under experimental conditions. See text for details. Measured seawater parameters include temperature, salinity,
−
pH, and alkalinity, which were used to calculate [CO2−
3 ], [B(OH)4 ], DIC, and Ωc. Uncertainties reported for measured seawater
parameters represent 1-sigma standard deviations of all seawater measurements made throughout each experiment; uncertainties
for calculated values represent the root-mean-square combined uncertainty from measured parameters. B/CaR and B/CaCB
indicate values on the Rutgers University and Cambridge-Bristol University scales, respectively (see equation 8).
High pH
Low Temp
High Temp
Low Sal
High Sal
Low CO2−
3
High CO2−
3
Low pH
Final chamber
Sample
1a
1b
2
4
5
7
8
9
12a
12b
13a
13b
13c
G. sacculifer
Ambient
Experiment
110
111
Core
G. sacculifer
PE07-6I
PE07-5I
PE07-6II
O. universa
NPBC
PE07-6I
PE07-6II
Lat
°N
Long
°W
Temperature
°C
Salinity
DIC
ALK
(B/Ca)R
µmol
kg
µmol
kg
µmol
mol
27.417
27.550
27.417
92.117
92.167
92.117
27.89
27.88
27.89
27.194
27.417
27.417
91.409
92.117
92.117
27.96
27.89
27.89
±
(B/Ca)CB
35.83
35.82
35.83
2040
2040
2040
2386
2386
2386
121.7
113.2
110.9
6.1
5.7
5.5
113.2
105.4
103.2
35.99
35.83
35.83
2038
2040
2040
2387
2386
2386
75.7
72.5
70.8
3.8
3.6
3.5
70.7
67.8
66.2
µmol
mol
Table 4.2: Results of core-top samples from the Gulf of Mexico. Seawater parameters were
estimated from the 2007 GOMECC Report, assuming foraminiferal depth habitat of 0-50
meters. All foraminifers were derived from the 500-710 µm size fraction.
112
113
Chapter 5
Deep water CO2 loss during the Last
Glacial Termination: Evidence from
New Zealand’s Bay of Plenty
5.1
Abstract
Ocean carbonate chemistry likely played a major role in regulating atmospheric CO2 across
ice age cycles, but the mechanisms responsible for ocean sequestration and release of CO2
have not yet been fully explained. Here, we present new records of deep water carbonate
chemistry derived from core RR0503-83, located at 1,627 meters depth in New Zealand’s Bay
of Plenty in the Southwest Pacific Ocean. Trace element and stable isotopic composition
of foraminiferal calcite (the epibenthic species Cibicidoides wuellerstorfi ) reveal changes in
deep water properties during periods of atmospheric CO2 change. The boron to calcium
ratio (B/Ca) in these shells indicates that deep water carbonate ion ([CO2−
3 ]) changed up
to 30 µmol kg−1 across the deglaciation, settling to Holocene values ∼5 µmol kg−1 higher
than those during the glacial. Combined with deep water O2 and δ 13 C records (Jaccard
et al. 2009; Elmore and Sikes et al., in prep), our paleo-[CO2−
3 ] results provide evidence
114
for increased storage of respired CO2 in the deep ocean and a strong influence of CaCO3
dissolution on deep water alkalinity during the last glacial maximum. These records are
also consistent with the release of respired CO2 from deep waters during the termination,
and together suggest a series of events in which major oceanographic changes are intimately
linked with shifts in atmospheric circulation.
5.2
Introduction
During the last ice age, Northern Hemisphere ice sheets were larger than today, covering
much of North America, Scandinavia, and Europe (Denton and Hughes, 1981; Svendsen
et al., 2004). Global sea surface temperatures were a few degrees cooler (Waelbroeck et al.,
2009), eustatic sea level was about 120 meters lower (Peltier and Fairbanks, 2006), and the
concentration of carbon dioxide in the atmosphere (CO2 ) was about 100 parts per million
by volume (p.p.m.v.) lower than during warm, interglacial periods (Petit et al., 1999).
The transition from these glacial conditions to the present interglacial climate represents a
dramatic, global environmental shift, but its cause is not yet fully understood.
The termination of the last ice age involved a series of abrupt events including major
ice sheet melting in the north, release of fresh water and sea ice into the Atlantic, rapid
sea level rise, and shifts in atmospheric and oceanic circulation (reviewed in Denton et al.,
2010; Cheng et al., 2009). Climate records from geologic archives indicate that these events
are part of a larger pattern: Earth experienced several ice age cycles during the past ∼
850,000 years, all characterized by slow cooling and ice sheet growth followed by a rapid
return to warm conditions approximately every 100,000 years (Emiliani, 1955; Lisiecki and
Raymo, 2005). Identifying the cause of these rapid glacial terminations is a key step towards
understanding the mechanisms that drive periodic glaciation in the Pleistocene epoch.
The similar rhythms of ice volume and solar insolation changes at high latitude suggest
a driving role for insolation in ice age cycles (Hays et al., 1976). However, large magnitude
115
differences and timing offsets between past insolation change and observed climate response
indicate the presence of lags, strong feedbacks, and/or other driving mechanisms for ice
age inception and termination (Imbrie et al., 1993; Wunsch, 2003). Another issue is that
if glaciation were driven by high-latitude insolation changes (i.e., summer insolation intensity at 65°N), widespread glaciation should alternate between the Northern and Southern
Hemispheres (Milankovitch, 1941). Instead, observations from geologic archives show that
glacial conditions occur synchronously in both hemispheres, requiring a driver or feedback
with global impact (Hays et al., 1976).
Changes in ocean heat storage as well as changes in the concentration of greenhouse
gases in the atmosphere (e.g., CO2 and CH4 ) both have the potential to alter global surface
temperatures (Gordon, 1981; Lumpkin and Speer, 2007; Hansen et al., 1984). Indeed, ice core
records indicate that atmospheric pCO2 and temperature have been tightly linked during
the past 800,000 years (Petit et al., 1999), suggesting that CO2 may have played a role in
transitions between ice ages and warm interglacial periods (Barnola et al., 1987; Denton et al.,
2010; Shakun et al., 2012), an idea that has been discussed for over a century (Arrhenius,
1896). Because of its large capacity and rapid exchange of gases with the atmosphere, the
ocean is the reservoir most likely to have sequestered CO2 from the atmosphere during
glacial periods (Broecker and Peng, 1982); the terrestrial biosphere was smaller during ice
ages (Adams et al., 1990), and other reservoirs (such as crustal rocks) exchange too slowly
to account for glacial-interglacial CO2 fluctuations (Broecker and Denton, 1990). However,
the specific mechanisms that led to greater sequestration of CO2 in the glacial ocean and
the means by which this CO2 returned to the atmosphere during subsequent deglaciation
are still under intense investigation.
The effects of lower glacial sea surface temperature and higher glacial seawater salinity on
the solubility of CO2 are small and mutually canceling, together accounting for <10 p.p.m.v.
of the observed 80 – 100 p.p.m.v. glacial-interglacial change (Sigman and Boyle, 2000). This
points to a dominant role for seawater carbonate chemistry in driving atmospheric CO2
116
fluctuations. In seawater, the partial pressure of pCO2 is related to total alkalinity (TA) and
total dissolved inorganic carbon (DIC) by the following equation:
pCO2 ≈
K2 [2DIC − T A]2
K0 · K1 [T A − DIC]
(5.1)
where K0 , K1 , and K2 are the carbonate system equilibrium constants. Therefore, to
draw down CO2 to glacial levels, surface ocean DIC must have been lowered, or surface ocean
alkalinity must have been raised, or both. Processes capable of altering these parameters
include: export of organic matter and carbonate (CaCO3 ) from surface waters, variations in
ocean circulation and vertical mixing, continental weathering flux, and dissolution of CaCO3
(reviewed in Sarmiento and Gruber, 2006). Many of these processes are characterized by
very specific ratios of TA:DIC – for example, dissolution of CaCO3 adds TA:DIC to seawater
in a 2:1 ratio. If constraints can be placed on relative alkalinity and DIC changes across
glacial cycles, new insight may be gained into the processes controlling atmospheric CO2 .
Establishing the timing of seawater carbonate system changes can provide further insight
into ice age drivers. Evidence from both marine and terrestrial records indicate that a greater
amount of CO2 was stored in the deep sea during glacial periods than interglacial periods,
and that storage decreased during the glacial termination (e.g., Sikes et al., 2000; Jaccard
et al., 2009; Yu et al., 2010a; Marchitto et al., 2005; Skinner et al., 2010; Burke and Robinson,
2012). Between approximately 18 and 11 ka, ice core records indicate two periods of rapid
pCO2 rise and warming Antarctic surface temperatures separated by a short period, known
as the Antarctic Cold Reversal, in which pCO2 remained stable and Antarctic temperatures
briefly decreased (Monnin et al., 2001, see also Figure 5.6). During the two warming periods,
upwelling intensified in the Southern Ocean, likely accompanied by degassing of deep waters
(Anderson et al., 2009). However, direct evidence of oceanic CO2 release is sparse in this
region, and further constraints are needed to improve estimates of CO2 released during the
deglaciation. Such constraints may be gleaned from records of seawater carbonate chemistry
at sites where deep waters enter and exit the Southern Ocean.
117
In this study, a sediment core from the Southwest Pacific Ocean collected from a water
depth of 1,627 m is used to reconstruct changes in seawater properties from the last glacial
period through the termination (25 to 1 ka) using trace element ratios recorded within
fossil calcite. In some benthic foraminifer species, the ratio of boron to calcium (B/Ca)
is strongly correlated with the calcite saturation state of overlying deep waters as quantified by the parameter ∆CO2−
(Yu and Elderfield, 2007; Rae et al., 2011; Brown et al.,
3
2011). This parameter, representing the difference between the existing carbonate ion concentration and the concentration that would be required for calcite saturation (∆CO2−
3 =
2−
[CO2−
3 ]insitu - [CO3 ]sat ), is often used in lieu of the full calcite saturation parameter (Ω =
2−
2+
[Ca2+ ][CO2−
3 ]insitu / [Ca ][CO3 ]sat ) because the concentration of calcium ion in seawater
is roughly constant (Millero, 2006). B/Ca values vary among benthic foraminifer species,
with epifaunal species, particularly Cibicidoides wuellerstorfi, exhibiting the highest B/Ca
values and also the highest sensitivity to bottom water ∆CO2−
3 . No mechanism for this observed B/Ca-∆CO2−
3 relationship has yet been identified, but consistent results for benthic
foraminiferal samples derived from different ocean basins and measured in different laboratories provide strong evidence for a reproducible and global pattern (Rae et al., 2011; Yu
and Elderfield, 2007). If temperature, salinity, and pressure are known, the concentration of
2−
carbonate ion in seawater ([CO2−
3 ]) may be calculated from ∆CO3 . In combination with
respired CO2 levels estimated from paleo-O2 data, this [CO2−
3 ] may be used to estimate past
alkalinity, and ultimately, to constrain possible mechanisms of ocean CO2 sequestration and
release.
5.3
Core Site and Regional Oceanography
The sediment core RR0503-83 (36.7375°S, 176.6398°E, 1627 m) is a combination triggerpiston core collected by the RV Roger Revelle as part of a wide depth transect (663 to 3836
m) within New Zealand’s Bay of Plenty (Figures 5.1 and 5.2). The Bay of Plenty is located
118
north of New Zealand’s North Island and just west of the Tonga-Kermadec Ridge, which runs
roughly NNE-SSW at 180°W (Figure 5.1). The islands of New Zealand are flanked on the
south and east by the Campbell Plateau and Chatham Rise. Together, these features exert
a strong influence on South Pacific circulation. The tip of New Zealand’s South Island is
crossed by the Subtropical Front (STF), a sharp transition between warm, salty subtropical
water to the north and cooler, fresher Antarctic surface water to the south (Heath, 1985;
Orsi et al., 1995). South of the STF, the Antarctic Circumpolar Current (ACC) transports
waters eastward, and some of this deep circumpolar flow is diverted northwards and eastwards
around the Campbell Plateau and Chatham Rise as the Deep Western Boundary Current
(DWBC), while some intermediate and deep water is swept into the South Pacific anticyclonic
gyre (Reid, 1997; Orsi et al., 1999; Macdonald et al., 2009). The northward flow of Antarcticderived deep and bottom water in the Pacific is balanced by southward flow of Pacific Deep
Water (PDW) and Upper Circumpolar Deep Water (UCDW) (Sloyan and Rintoul, 2001).
Low-oxygen, high-nutrient PDW forms as deep and bottom water upwells slightly in the
North Pacific Basin (Reid, 1997). PDW then flows southwards, accumulating nutrients and
CO2 from the respiration of marine microbes, and flows into the Bay of Plenty from just
east of the Tonga-Kermadec Ridge. Today, the core 83 site is bathed by a mixture of PDW
and CDW (Figure 5.2B).
5.4
Methods
To establish an age model for RR0503-83, several tephra layers within the core were identified
by their unique geochemical compositions and then tied to a regional tephrostratigraphic
framework based on radiometric dating of volcanic deposits on land (Shane et al., 2006).
The trigger core (TC) and jumbo piston core (JPC) were aligned using stable isotope records
(Aurora Elmore, personal communication, see also Figure 5.5). Ages of sediment samples in
core 83 were calculated by simple linear interpolation between the ashes and radiocarbon date
119
0
−20
−40
SubTropical Front
D
LC
W
−60
Polar Front
cumpolar Current
ctic Cir
A ntar
−80
100
120
140
160
180
200
220
240
260
AAIW
280
300
Figure 5.1: Cartoon of deep water circulation in the Pacific Ocean. Northward flow of
Antarctic-derived deep and bottom waters (dark blue arrows) is balanced by southward flow
of Pacific Deep Water (PDW, red arrows). Upper Circumpolar Deep Water (UCDW, green
arrows) circulates at mid-depths (∼27.7 to 28.1 γ n ) between the Antarctic Circumpolar
Current (ACC, black arrows) and South Pacific gyre. Core site 83 (red circle) is influenced
by both PDW and UCDW.
listed in Table 5.3). Average sedimentation rate of TC is ∼9 cm/ky. If the isotope records
from TC and JPC are aligned, the resulting sedimentation rate for the upper portion of
JPC (<15 ka) decreases to ∼ 2 cm/ky, but the reason for this is not clear. It is possible
that the top of the piston core was compacted during coring, leading to lower apparent
sedimentation rates. To resolve this issue, radiocarbon ages will be measured on planktonic
foraminiferal calcite from the upper portion of JPC in the future. In the present chapter,
the ash constraints (both TC and JPC), radiocarbon date (TC), and isotope alignment (TC
and JPC) are used together to establish the age model shown in Figure 5.3 (see also Table
5.4). Close examination of core photos reveals a possible disturbance or double-penetration
in TC below 150 cm; no samples below that depth were analyzed in this study. Except for
this disturbance feature, the sediment archives of both TC and JPC appear continuous and
well-preserved (Appendix B).
This age model was used to select samples for trace element analysis that stretched from
120
-1000
North
Core Transect Bathymetry
WOCE Hydrographic P15
sea level
0
Bay of Plenty
Core 87
1000
Depth (m)
NORTH
Island
Core 79
ANTARCTICA
Core 83
2000
o
177 E
Campbell Plateau
Core 125
3000
Core 41
4000
o
195 E
5000
o
Chatham Rise
(165 W)
6000
-60
-50
-40
-30
-20
Latitude
0
26.25
Core 87
AAIW
Depth (m)
1000
Core 79
27.70
2000
28.01
-1
28.10
UCDW
γ = 28
.16
LCDW
3000
150
175
200
225
250
275
300
325
PDW
Core 125
n
4000
AABW
Oxygen
(μmol kg )
Core 83
Core 41
5000
n
6000
-3
γ = neutral density (kg m )
-60
-50
-40
-30
-20
Latitude
Figure 5.2: Core location. A. Bathymetric profile of World Ocean Circulation Experiment
(WOCE) transect P15 at 165°W superimposed on sediment core transect collected by the
Roger Revelle in 2009 (cruise RR0503) in New Zealand’s Bay of Plenty. Core 83, the focus of
the present study, is highlighted in red. B. Neutral density surfaces (gray lines) and oxygen
concentrations (colors) drawn for WOCE transect P15 with RR0503 Core locations (gray
and black circles) and approximate major water mass flow directions (dark arrows). Deep
Western Boundary Current is derived from Antarctic Bottom Water; CDW is Circumpolar Deep Water; PDW is Pacific Deep Water; AAIW is Antarctic Intermediate Water and
SAMW is SubAntarctic Mode Water. In the Bay of Plenty region, two AAIW types flow in
from the West (higher salinity) and Northeast (lower salinity) (Stanton et al., 2002).
121
JPC Depth
cm
–
–
(12)
(15)
(18)
(23)
68
107
164
TC Depth
cm
0
(9)
57.5
72.5
97
125
–
–
–
Age
cal yr
0
0.64
7.01
8.01
9.90
13.64
17.63
21.80
25.36
±
cal yr
0
0.01
0.16
0.05
0.17
0.43
0.50
0.16
Age constraint
References
assumed
Kaharoa ash (*)
Tuhua ash (TC)
Mamaku ash (TC)
14 C, G. inflata
Waiohau ash (TC)
Rerewhakaaitu ash (JPC)
Okareka ash (JPC)
Kawakawa ash (JPC)
Elmore et al., in prep.
Hogg et al. 2003
Hajdas et al. 2006
Hajdas et al. 2006
Elmore et al., in prep.
Hajdas et al. 2006
Hajdas et al. 2006
Reimer et al. 2004
Vandergoes et al., 2013
Table 5.1: Age constraints for site 83 trigger core (TC) and jumbo piston core (JPC). Bold
numbers indicate depths at which ashes have been identified by geochemical fingerprinting
(Shane et al., 2006); numbers in parentheses indicate JPC depths (column 1) that have been
inferred from oxygen isotope alignment with TC, and one TC depth (column 2) inferred
from oxygen isotope alignment with a nearby core that is not included in the present study
(Aurora Elmore, personal communication).
the Last Glacial Maximum (LGM, ∼23-19 ka) through the glacial termination, with highresolution (1 cm) sampling between 19 and 12 ka. Sediment samples were washed with
deionized water through a 63 µm sieve, and benthic foraminifers were hand-picked from
the >150 µm size fraction under a light microscope. If there were not enough Cibicidoides
wuellerstorfi specimens within a 1 cm depth interval to yield ∼250 µg calcite, adjacent
samples were combined to achieve an adequate sample weight, provided that the stable
isotopes in the two samples were similar. Of the final 70 samples analyzed for trace elements,
49 span 1 cm depth intervals and 21 represent combined depth intervals (average age span
for a combined sample = 312 years). Foraminiferal samples were partially crushed between
two clean glass slides, transferred to acid-cleaned vials, and subjected to a series of cleaning
steps to remove clay particles, oxides, organic matter, and adsorbed contaminants according
to the methods of Boyle and Keigwin (1985) as outlined in Rosenthal et al. (1997).
Cleaned samples were dissolved in 0.065N nitric acid (HNO3 ) and analyzed on a sectorfield inductively coupled mass spectrometer (Thermo Scientific Element XR) at Rutgers
University following methods detailed in Allen et al. (2012). Any samples with Al/Ca, Ti/Ca,
122
0
83
TC
2 cm
/ky
83
C
JP
Core depth (cm)
50
100
9
cm
150
y
/k
0
5
10
15
20
25
Age (ka)
Figure 5.3: Age model for sediment core RR0503-83 JPC based on tephras (filled triangles)
and one radiocarbon date from planktonic G. inflata (circle). Each ash’s age was derived
from radiocarbon measurements on contemporary or bounding organic material in terrestrial
deposits (Shane et al., 2006; Alloway et al., 2006). JPC = Jumbo Piston Core; TC = Trigger
Core. Open triangles in JPC represent depths that were aligned with the TC age model using
oxygen isotope records (Elmore et al., in prep,; see Figure 5.5). The reason for the difference
in apparent sediment accumulation rates between TC and JPC is not clear. See text for
details.
123
and/or Fe/Ca values 10x greater than the mean were assumed to be contaminated, and were
excluded from further interpretation (5 out of 70 samples). Analyses were performed in
December 2011 and August 2012, and between-run precision of matrix-corrected data was
better than 5% for all elements presented here.
2−
To calculate paleo-seawater carbonate ion concentrations ([CO2−
3 ]), we first derive ∆CO3
from B/Ca according to the modern core-top calibration of Yu and Elderfield (2007):
B/Ca (µmol mol−1 ) = 1.14 · ∆CO32− (µmol kg −1 ) + 177.1
(5.2)
2−
In the modern ocean, ∆CO2−
3 is a function of both pressure and [CO3 ]. In the Yu and
Elderfield (2007) calibration dataset, ∆CO2−
3 in the Pacific is controlled primarily by pressure
2−
([CO2−
3 ] is virtually constant throughout the water column), while ∆CO3 in the Atlantic
2−
is controlled by both pressure and [CO2−
3 ] (with [CO3 ] increasing at shallower depths).
Regardless of this environmental difference, when B/Ca values from both Atlantic and Pacific
ocean basins are plotted against ∆CO2−
3 , their trends are indistinguishable (Figure 5.4B).
With regard to paleoreconstructions – if pressure remains roughly constant at a given core
location, then any down-core changes in B/Ca may be inferred to simply reflect changes in
seawater [CO2−
3 ].
2−
Finally, ∆CO2−
3 values are converted to [CO3 ] according to the equation:
[CO32− ] = ∆CO32− + [CO32− ]sat
5.5
(5.3)
Results
Between 28 and 20 ka, B/Ca remains near 194 µmol mol−1 (Figure 5.5). From 19 ka to
16 ka, B/Ca values are variable, with average B/Ca decreasing slightly. After 16 ka, B/Ca
increases, reaching a maximum value at ∼14 ka, and between 12 and 11 ka, B/Ca exhibits a
124
Yu and Elderfield (2007)
240
220
200
m
B/Ca ( mol mol
-1
)
260
180
160
PACIFIC
ATLANTIC
Other
140
120
-20
0
20
40
60
-1
DCO3 (mmol kg )
2-
Figure 5.4: A compilation of C. wuellerstorfi B/Ca data from Yu and Elderfield (2007), Rae
et al. (2011), and Brown et al. (2011); linear fit (dashed black line) from Yu and Elderfield
(2007). Green circles and purple triangles are from the Atlantic and Pacific calibration sites
of Yu and Elderfield (2007); gray symbols include data from the Indian Ocean and Norwegian
Sea.
125
short, sharp decrease. After 10 ka, B/Ca returns to average values similar to those between
28 and 20 ka.
Pressure, temperature, and salinity (P, T, S) all have the potential to influence [CO2−
3 ]sat
through their influence on carbonate equilibrium constants. We evaluate the possible magnitude of glacial-interglacial changes in these parameters on [CO2−
3 ]sat at core site 83 by calculating Holocene and LGM [CO2−
3 ]sat from P, T, and S derived from independent records.
Assuming eustatic sea-level rise from the coral records of of Peltier and Fairbanks (2006)
and a corresponding salinity decrease in a well-mixed ocean, we estimate the influence of
deglacial sea level rise on bottom water P and S to be 135 dbar and 1.3 salinity units, respectively. In addition, bottom waters in the Pacific and Southern Oceans are estimated
to have warmed ∼ 2 °C between the LGM and the Holocene (Nelson et al., 1993; Adkins
et al., 2002; Malone et al., 2004; McCave et al., 2008). If all of these effects are combined,
−1
the resulting difference between Holocene and LGM [CO2−
3 ]sat is only 0.5 µmol kg , which
is smaller than the uncertainty in ∆CO2−
3 reconstructed from B/Ca. Therefore, we assume
−1
negligible influence of P, T, and S, and apply a constant [CO2−
3 ]sat value of 58.5 µmol kg
for the entire record (calculated assuming 2.74 °C, 34.57 S, and 1627 dbar, respectively, Key
et al., 2004). Combining this constant value with B/Ca-derived ∆CO2−
3 estimates gives a
−1
[CO2−
(Figure 3.2).
3 ] record that ranges from ∼ 50 – 80 µmol kg
5.6
5.6.1
Discussion
Implications for the Last Ice Age
Here, we examine the glacial [CO2−
3 ] record from Core 83 and discuss its implications for
processes that may have contributed to oceanic sequestration of atmospheric CO2 during
the glacial period. The surface-deep [CO2−
3 ] gradient is discussed in the context of vertical
stratification, which can facilitate greater storage of respired CO2 in deep waters. Then,
[CO2−
3 ] is combined with DIC inferred from O2 records in order to estimate deep water
126
220
B/Ca
-1
( mol mol )
210
200
190
m
180
170
3.5
1.0
4.0
d
13
C (permil)
0.8
O (permil)
3.0
18
d
2.5
0.6
0.4
0.2
0.0
-0.2
-0.4
0
5
10
15
20
25
Age (ka)
Figure 5.5: Stable isotope ratios and B/Ca from benthic foraminiferal calcite (C. wuellerstorfi ) in core RR0503-83 across the last glacial termination. Filled circles indicate samples
from JPC and open squares indicate samples from TC. Error bars on B/Ca represent the
long-term analytical precision on a consistency standard.
127
ACR
-1
90
LGM
30
80
m
20
m
70
2-
10
60
-1
D
0
50
200
C (permil)
1.0
CO2 (ppmv)
250
B
0.5
C
0.0
d
13
2-
CO3 ( mol kg )
HS1
CO3 ( mol kg )
A
0
5
10
15
20
25
Age (ka)
Figure 5.6: Past deep water reconstruction based on C. wuellerstorfi trace element composition from core RR0503-83 in New Zealand’s Bay of Plenty. Open symbols = TC; filled
symbols = JPC. (A) Calcite saturation (∆CO2−
3 ) calculated from B/Ca according to the calibration of Yu et al. (2007). Corresponding [CO2−
3 ] estimates (right axis) assume constant
2−
S, T, and P through time, giving [CO3 ]saturation = 58.5 µmol kg−1 . See text for details.
(B) Atmospheric CO2 from the EPICA Dome C ice core (Monnin et al., 2001), adjusted to
the GISP2 timescale. (C) δ 13 C of C. wuellerstorfi (Elmore, Sikes et al., in prep). ACR =
Antarctic Cold Reversal, HS1 = Heinrich Stadial 1, LGM = Last Glacial Maximum.
128
alkalinity, and thus gain insight into the relative contributions of the biologic pump and
carbonate compensation to glacial CO2 drawdown.
Given the minor impact of reconstructed salinity and temperature changes on glacial
pCO2 levels (Broecker and Peng, 1982; Sigman and Boyle, 2000) and the rapid exchange
of gases between the surface ocean and atmosphere (Takahashi et al., 2002), low glacial
atmospheric pCO2 values likely reflect a lower surface ocean pCO2 , corresponding to higher
carbonate ion concentrations. Models and records of surface [CO2−
3 ] infer glacial values at
least 40-50 µmol kg−1 greater than today (Lea et al., 1999; Barker and Elderfield, 2002). The
−1
B/Ca reconstruction from Core 83 yielded LGM ∆CO2−
lower than
3 values only 5 µmol kg
Holocene values (Figure 5.6). In light of the predicted surface ocean increase of 40-50 µmol
kg−1 , this new benthic record suggests that the surface-to-bottom carbonate ion gradient
in the Bay of Plenty was at least 45 µmol kg−1 greater than today during the LGM. This
enhanced vertical chemical gradient suggests greater ocean stratification in the Southwest
Pacific during the LGM, consistent with other sedimentary records of stratification and deep
water isolation in the Southern Hemisphere (Ninnemann and Charles, 2002; Anderson et al.,
2009; Burke and Robinson, 2012).
Greater stratification can lead to enhanced accumulation of CO2 in deep waters by increasing the efficiency of a process commonly known as the ‘biologic pump.’ In ocean surface
waters, photosynthetic organisms convert CO2 to organic matter, some of which sinks below
the photic zone. At greater depth, organic matter is transformed to CO2 via respiration,
a process that consumes dissolved oxygen (O2 ). These processes can be described by the
following generalized equation (Anderson, 1995):
106CO2 + 16HN O3 + H3 P O4 + 78H2 O
respiration
photosynthesis
C106 H175 O42 N16 P + 150O2
(5.4)
During glacial times, sedimentary recorders of dissolved oxygen indicate lower values in
129
the deep ocean, implying greater consumption of O2 by respiration, and correspondingly
increased CO2 storage at depth (Jaccard et al., 2009; Bradtmiller et al., 2010). Such an
increase in the CO2 concentration of seawater would have caused a transient carbonate
(CaCO3 ) dissolution event (‘carbonate compensation’), which also would have increased
seawater alkalinity and allowed sequestration of additional atmospheric CO2 (Boyle, 1988).
To constrain the relative roles of the biologic pump and carbonate compensation in altering
glacial seawater chemistry, we combine B/Ca-derived [CO2−
3 ] with DIC changes estimated
from O2 records (simply assuming 1 mole of CO2 (DIC) produced for every mole of O2
consumed during respiration). Starting from modern deep water composition at core site
83, applying a glacial salinity increase of 3% (Adkins et al., 2002) raises both TA and
DIC (Figure 5.7). Then, addition of 100 µmol kg−1 of respired CO2 (inferred from a ∼
100 µmol kg−1 O2 decrease, Jaccard et al., 2009) raises DIC to ∼2420 µmol kg−1 (Figure
5.7). This respired CO2 represents a contribution via the biologic pump to deep ocean DIC.
Finally, to achieve the ∼ 70 µmol kg−1 [CO2−
3 ] reconstructed from B/Ca, CaCO3 dissolution
must have raised DIC and TA to ∼ 2500 and 2590 µmol kg−1 , respectively (white arrow
labeled “CaCO3 dissolution,” Figure 5.7). Thus, the relative contributions of respired CO2
(biologic pump) and dissolved carbonates (carbonate compensation) to bottom-water glacialinterglacial difference in terms of DIC are 100 µmol kg−1 and 85 µmol kg−1 , respectively, a
little over a 1:1 influence.
Because these calculations involve the assumption of 1:1 CO2 :O2 stoichiometry during
respiration (which may actually range from 1:1 up to 1:1.5, Hedges et al., 2002), and do not
take into account any nutrient contribution to alkalinity (small amounts of NO−
3 released
during respiration would also contribute to TA), the resulting TA and DIC serve as preliminary estimates only. Tighter constraints on surface water export composition and the
influence of temperature and nutrients need to be established. However, taken at face value,
this result implies a major role for both the biologic pump and carbonate compensation in
controlling glacial-interglacial pCO2 .
130
2600
40
2550
TCO2 (µmol/kg)
2500
O
[C
23
]=
10
ion
O
CaC 3
2450
2400
2350
2300
2-
O3
[C
]=
20
2-
3
[CO
]=
Respired
CO2
30
2- ]
3
[CO 2- ] = 50
3
[CO
init
Sal
y
TE
NA /Ca
O
B
B
AR om
fr
2-
3
[CO
MODERN
DEEP
WATER
2200
]=
15
0
2-
3
[CO
]=
2300
2350
20
0
2-
3
[CO
Bay of Plenty
2,000 m
2150
GLACIAL
DEEP
WATER
ION
0
=4
2250
2100
2250
C
olut
diss
]=
25
0
30
2400
2450
2500
2550
0
2600
Alkalinity (µmol/kg)
Figure 5.7: When combined with estimates of respired CO2 from paleo-oxygen data, B/Ca2−
derived [CO2−
3 ] may be used to estimate past alkalinity. The glacial [CO3 ] value is from
preliminary work on Bay of Plenty cores; the respired CO2 value used here is drawn from Pacific sites (Jaccard et al., 2009), and is used for illustration only. The arrow labeled “CaCO3
dissolution” represents the net dissolution of both sedimentary and sinking carbonate. Temperature, nutrient changes, and variable O2 :CO2 respiration represent sources of uncertainty
that need to be addressed in the future.
131
5.6.2
Implications for the Termination
The early deglacial (19 to 16 ka).
During this time, the westerly wind belts were at their southernmost extent (De Deckker
et al., 2012; Broecker and Putnam, 2012) and atmospheric CO2 was rising rapidly (Monnin
et al., 2001; Petit et al., 1999) (Figure 5.6C). All else being equal, when CO2 is degassed
2−
from seawater, [CO2−
3 ] increases. Thus, the observed dip to lower [CO3 ] values at site 83
indicated by low B/Ca in the early deglacial period is unexpected (Figure 5.6). Rather than
reflecting changes in ocean-atmosphere gas exchange, this B/Ca decrease might be recording
a southward expansion of Pacific Deep Water (PDW) in response to changes in atmosphereocean circulation. If the observed [CO2−
3 ] decrease were the result of a simple increase in the
13
PDW “end member” as it is characterized today (low [CO2−
3 ] and low δ CDIC ), this change
should have been accompanied by a decrease in benthic foraminiferal δ 13 C. Instead, δ 13 C
remains roughly constant between 19 and 16 ka (Figure 5.5). What could have decoupled
these two parameters?
13
In deep-water formation regions and throughout much of the deep sea, [CO2−
3 ] and δ C
patterns are dominated by the degradation of organic matter (Broecker and Peng, 1982;
Yu et al., 2008). The release of CO2 with a ‘light’ isotope signature typical of marine
photosynthates has the two-fold effect of increasing DIC and decreasing [CO2−
3 ] as well
as driving the average δ 13 C composition of DIC to lower (more negative) values. Thus,
13
in modern deep waters, [CO2−
3 ] and δ C often decrease in constant proportion as organic
matter is respired (on average, 1/43 hper µmol kg−1 , see Figure 5.8). This process is largely
13
responsible for the very low [CO2−
3 ] and low δ CDIC in the modern deep Pacific. Changes in
deep water alkalinity can alter [CO32− ] without changing δ 13 C, however, a freshwater input
equivalent to over 1,000 meters of sea level rise would be required explain the decrease in
−1
[CO2−
3 ] from ∼70 to ∼ 55 µmol kg , assuming the ocean is a well-mixed reservoir on a ∼
1 ky timescale. Considering that sea level reconstructions indicate only ∼120 meters of sea
132
level rise during the deglaciation, freshwater dilution is not a reasonable explanation.
13
Alternatively, the observation of decreased [CO2−
3 ] at constant δ C could be the result
of mixing two water masses characterized by similar δ 13 C, but differing [CO2−
3 ]. During the
LGM and early deglacial period, benthic foraminiferal records indicate that North Pacific
Deep Water and Pacific Circumpolar Deep Water were characterized by similar δ 13 CDIC
(Ninnemann and Charles, 2002). If a greater proportion of NPDW reached site 83 at this
time, and NPDW [CO2−
3 ] was lower than that of CDW (like today), it could have depressed
13
[CO2−
3 ] without strongly altering δ C.
The mid-deglacial (16 to 11 ka).
−1
Between ∼16 and 14.3 ka, [CO2−
and δ 13 C increases ∼ 0.3 h. The
3 ] increases 30 µmol kg
13
concomitant change in [CO2−
3 ] and δ C suggests that deep waters bathing site 83 contained
less respired CO2 during this interval than during the glacial and early deglacial periods
(large gray arrow, Figure 5.8). Lower concentrations of deep water CO2 outcropping in the
Southern Ocean would lead to slower CO2 degassing, consistent with a decrease in slope of
the atmospheric CO2 curve at 16 ka observed in the ice core record (Monnin et al., 2001;
Barker et al., 2009). In general, increased [CO2−
3 ] at site 83 suggests the existence of more
‘degassed’ deep water masses in the mid-deglacial, consistent with scenarios in which abyssal
CO2 is released to the atmosphere primarily during the early deglaciation.
Deep water [CO2−
3 ] reached its peak near 14.3 ka and then began to decrease during
the Antarctic Cold Reversal (ACR), when the Antarctic warming trend briefly reversed
and cold conditions prevailed in much of the Southern Hemisphere (Putnam et al., 2010;
Monnin et al., 2001). During the ACR, % abundance of the tropical planktic foraminifer
species Globigerinoides ruber reached a low point at site MD03-2611 just south of Australia
(De Deckker et al., 2012). The % G. ruber decrease indicates a weakening of the warm
Leeuwin Current, likely due to a northward shift of the Subtropical Front (STF) (De Deckker
et al., 2012). At the same time, glaciers readvanced in New Zealand (Putnam et al., 2010),
133
Southern Ocean upwelling paused (Anderson et al., 2009), and atmospheric CO2 stabilized
(Monnin et al., 2001). Together, these observations suggest that westerly winds shifted
northwards during the ACR, causing a temporary return to stratified conditions across the
Southern Ocean. Greater ocean stratification could have led to increased amounts of respired
CO2 in CDW and PDW, which might explain the [CO2−
3 ] decrease observed during the ACR
at site 83. δ 13 C did not decrease during the ACR, as would be expected from an increase in
the concentration of respired CO2 . It is possible that this signal was influenced by mixing
with another water mass, as discussed above, or any decrease due to respired CO2 input
may have been counteracted by A) the whole-ocean increase in δ 13 C (∼ 0.3 h, Broecker
and McGee, accepted) and/or B) the decrease in Southern Ocean temperatures, which has
the potential to drive δ 13 CDIC to more positive values via a change in kinetic fractionation
during gas exchange (Lynch-Stieglitz et al., 1995).
5.7
Conclusions
At a ∼1600 m site in the Southwest Pacific, a [CO2−
3 ] record derived from benthic foraminiferal
B/Ca suggests that this region was more strongly stratified during the last ice age. Combined
with benthic δ 13 C and independent paleo-O2 estimates, the [CO2−
3 ] record also indicates increased storage of CO2 in the deep ocean during the LGM, with major roles for the biologic
pump and carbonate compensation.
During the early glacial termination (19 to 16 ka), B/Ca indicates a deep water [CO2−
3 ]
decrease, which may indicate expansion of low-[CO2−
3 ] Pacific Deep Water during this time.
−1
This decrease is followed by a rapid [CO2−
3 ] increase of ∼30 µmol kg , suggesting the
presence of better-ventilated deep waters. During the ACR, [CO2−
3 ] decreases again, perhaps
in response to restratification of the Southern Ocean.
Overall, this study’s B/Ca results provide new evidence for deep water carbonate chemistry changes during the LGM and key periods of CO2 rise. The [CO2−
3 ] reconstruction
134
1.0
'Re
s
eld
dfi
l op
e'
5
0.5
10
13
M
LG
0.0
-
ce
lo
Ho
ne
15
Age (ka)
C (permil)
Today
20
LGM
19 - 16 ka
~0.3 ‰
-0.5
25
whole-ocean shift
50
60
70
2CO3 (
80
90
-1
mol kg )
Figure 5.8: Relationships between δ 13 Cbenthic and B/Ca-derived [CO2−
3 ] at site 83 since 25
2−
13
ka. The general shift from low [CO3 ], low δ C values during glacial times (blue symbols)
13
to higher [CO2−
3 ], higher δ C values during the Holocene (red symbols) is consistent with a
transfer of ‘light’ respired CO2 from the deep ocean to the atmosphere. A shift in [CO2−
3 ]
13
independent of δ C during the early deglaciation (19 - 16 ka) may reflect a change in
alkalinity. Observed changes in δ 13 C likely reflect the combined effects of local and wholeocean δ 13 CDIC change (upward black arrow). Gray background symbols are modern ocean
data (Key et al., 2004).
135
is consistent with deglacial scenarios in which major oceanographic changes are intimately
linked with shifts in atmospheric circulation. Additional [CO2−
3 ] records from shallower and
deeper depths are needed to discern changes in water mass properties from migration of water mass boundaries, and to provide tighter constraints on global DIC and alkalinity changes
across ice age cycles.
5.8
Acknowledgments
I am grateful to Aurora Elmore for instruction in micropaleontological techniques and for
generously sharing and discussing her stable isotope records. Elisabeth Sikes provided key
guidance regarding the study’s conception and sampling scheme as well as expertise in the
oceanography and history of the Southwest Pacific region. I am also grateful to Tali Babila,
Julie Kalansky, and Yair Rosenthal for analytical assistance at Rutgers University, and to
Peter deMenocal, Alex Gagnon, Taro Takahashi, and Aaron Putnam for comments that
improved this chapter. Many thanks to Bob Anderson, who provided insightful discussion
and guidance , and to Bärbel Hönisch, who initiated and encouraged this project, and
provided support throughout this work.
136
137
Appendix A: Artificial Seawater
Recipe (Chapter 4)
Supplement for Chapter 4: “Environmental controls on B/Ca in calcite tests of the tropical
planktic foraminifer species Globigerinoides ruber and G. sacculifer ”
To create low-DIC seawater, we mixed 2L natural seawater (for properties, see ambient
experiment, Table 1 in Chapter 4) with 2L of the synthetic seawater mixture below.
To 4 L deionized water (preferably Milli-Q+ ), add:
Ingredient
NaCl
Na2 SO4
KI
B(OH)3
MgCl2 ·6H2 O
CaCl2 ·2H2 O
SrCl2 ·6H2 O
grams
98.950
16.548
3.141
0.108
44.280
6.230
0.100
Finally, after thoroughly mixing the natural and synthetic seawater in a large Erlenmeyer
flask, the solution was titrated to 8.19 pH (NBS) with NaOH. This resulted in seawater
with DIC=1030 and TA=1210 µmol kg−1 . The solution was immediately transferred into
tightly sealed jars to prevent CO2 exchange with the atmosphere.
138
Appendix B: B/Ca data and core
photos from Bay of Plenty, New
Zealand (Chapter 5)
139
22
47
49
82
104
108
120
123
132
17
18
20
22
24
27
28
TC
TC
TC
TC
TC
TC
TC
TC
TC
JPC
JPC
JPC
JPC
JPC
JPC
JPC
30
28
26
23
21
20
18
135
125
123
109
105
84
50
49
25
17
cm
cm
14
Depth max
Depth min
TC
Core
29.0
27.5
25.0
22.5
20.5
19.0
17.5
133.5
124.0
121.5
108.5
104.5
83.0
49.5
48.0
23.5
15.5
cm
Depth mean
14.2
14.0
13.8
13.3
11.8
10.6
9.6
13.7
13.5
13.2
11.4
10.9
8.8
6.0
5.8
2.5
1.5
ka
Age
208.2
203.4
199.9
205.4
197.2
200.2
203.8
186.7
198.0
195.9
175.6
193.0
199.1
196.8
195.9
197.6
27.3
23.0
20.0
24.8
17.6
20.2
23.4
8.5
18.3
16.5
-1.3
13.9
19.3
17.3
16.5
18.0
17.1
80.5
76.3
73.4
78.1
71.0
73.6
76.7
62.1
71.7
69.9
52.5
67.4
72.7
70.7
69.9
71.4
70.5
µmol
kg
µmol
kg
µmol
mol
196.6
CO2−
3
∆CO2−
3
B/Ca
140
30
31
32
34
35
35
36
37
39
40
41
43
44
45
45
46
48
JPC
JPC
JPC
JPC
JPC
JPC
JPC
JPC
JPC
JPC
JPC
JPC
JPC
JPC
JPC
JPC
JPC
49
47
47
46
45
44
42
41
40
39
37
36
36
35
33
32
31
31
cm
cm
28
Depth max
Depth min
JPC
Core
48.5
46.5
46.0
45.5
44.5
43.5
41.5
40.5
39.5
38.0
36.5
35.5
35.5
34.5
32.5
31.5
30.5
29.5
cm
Depth mean
15.9
15.7
15.7
15.6
15.5
15.5
15.3
15.2
15.1
15.0
14.8
14.7
14.7
14.7
14.5
14.4
14.3
14.2
ka
Age
181.3
188.6
177.5
195.8
190.0
186.2
186.2
186.4
188.0
192.6
181.4
184.3
188.0
192.7
194.7
188.4
211.6
3.7
10.1
0.4
16.4
11.3
8.0
8.0
8.2
9.6
13.6
3.8
6.3
9.5
13.6
15.4
9.9
30.3
26.9
57.4
63.6
54.1
69.9
64.8
61.6
61.6
61.8
63.2
67.1
57.5
59.9
63.1
67.1
68.9
63.5
83.4
80.1
µmol
kg
µmol
kg
µmol
mol
207.8
CO2−
3
∆CO2−
3
B/Ca
141
49
51
52
52
54
56
55
61
64
68
69
70
71
72
72
73
74
JPC
JPC
JPC
JPC
JPC
JPC
JPC
JPC
JPC
JPC
JPC
JPC
JPC
JPC
JPC
JPC
JPC
75
74
73
73
72
71
70
71
67
62
59
57
55
53
53
52
52
50
cm
cm
49
Depth max
Depth min
JPC
Core
74.5
73.5
72.5
72.5
71.5
70.5
69.5
69.5
65.5
61.5
57.0
56.5
54.5
52.5
52.5
51.5
50.5
49.5
cm
Depth mean
18.3
18.2
18.1
18.1
18.0
17.9
17.8
17.8
17.4
17.0
16.6
16.6
16.4
16.3
16.3
16.2
16.1
16.0
ka
Age
192.9
185.3
189.7
183.3
181.9
172.0
176.0
178.0
180.7
179.5
197.0
177.9
182.4
173.9
170.8
177.3
181.6
13.9
7.2
11.1
5.5
4.2
-4.5
-1.0
0.8
3.2
2.1
17.4
0.7
4.7
-2.8
-5.5
0.1
4.0
0.1
67.4
60.8
64.6
59.1
57.9
49.4
52.8
54.6
56.9
55.9
70.9
54.5
58.4
51.0
48.4
53.9
57.7
53.9
µmol
kg
µmol
kg
µmol
mol
177.2
CO2−
3
∆CO2−
3
B/Ca
142
76
77
78
80
90
111
133
134
179
JPC
JPC
JPC
JPC
JPC
JPC
JPC
JPC
JPC
180
136
134
112
94
82
80
78
77
77
179.5
135.0
133.5
111.5
92.0
81.0
79.0
77.5
76.5
76.5
75.5
cm
Depth mean
28.4
23.5
23.5
22.1
20.2
19.0
18.8
18.6
18.5
18.5
18.4
ka
Age
192.3
195.6
194.8
192.1
193.4
183.8
187.4
184.6
186.5
178.0
13.3
16.3
15.5
13.2
14.3
5.9
9.0
6.6
8.2
0.8
8.4
66.8
69.7
69.0
66.7
67.8
59.6
62.6
60.2
61.8
54.6
62.0
µmol
kg
µmol
kg
µmol
mol
186.7
CO2−
3
∆CO2−
3
B/Ca
Table 5.2: Trace element data from Core 83 TC and JPC.
76
JPC
76
cm
cm
75
Depth max
Depth min
JPC
Core
143
Figure 5.9: Photos of Core 83: Trigger Core (TC).
144
Figure 5.10: Photos of Core 83: Jumbo Piston Core (JPC).
145
146
Bibliography
Adams, J. M., Faure, H., Faure-Denard, L., McGlade, J. M., 1990. Increases in terrestrial
carbon storage from the Last Glacial Maximum to the present. Nature 348, 711–714.
Adkins, J. F., McIntyre, K., Schrag, D. P., 2002. The Salinity, Temperature, and δ 18 O of
the Glacial Deep Ocean. Science 298 (1769), 1769–1773.
Al-Ammar, A. S., Gupta, R. K., Barnes, R. M., 2000. Elimination of boron memory effect
in inductively coupled plasma-mass spectrometry by ammonia gas injection into the
spray chamber during analysis. Spectrochimica Acta Part B: Atomic Spectroscopy
55 (6), 629–635.
Allen, K. A., Hönisch, B., Eggins, S. M., Rosenthal, Y., 2012. Environmental controls on
B/Ca in calcite tests of the tropical planktic foraminifer species Globigerinoides ruber
and Globigerinoides sacculifer . Earth and Planetary Science Letters 351-352 (C),
270–280.
Allen, K. A., Hönisch, B., Eggins, S. M., Yu, J., Spero, H. J., Elderfield, H., 2011. Controls
on boron incorporation in cultured tests of the planktic foraminifer Orbulina universa.
Earth and Planetary Science Letters 309 (3-4), 291–301.
Allison, N., Finch, A. A., EIMF, 2010. δ 11 B, Sr, Mg and B in a modern Porites coral: the
relationship between calcification site pH and skeletal chemistry. Geochimica et
Cosmochimica Acta 74 (6), 1790–1800.
Alloway, B. V., Lowe, D. J., Barrell, D. J. A., Newnham, R. M., Almond, P. C.,
Augustinus, P. C., Bertler, N. A. N., Carter, L., Litchfield, N. J., McGlone, M. S.,
Shulmeister, J., Vandergoes, M. J., Williams, P. W., 2006. Towards a climate event
stratigraphy for New Zealand over the past 30000 years (NZ-INTIMATE project).
Journal of Quaternary Science 22 (1), 9–35.
Anderson, L. A., 1995. On the hydrogen and oxygen content of marine phytoplankton.
Deep-Sea Research Part I 42 (9), 1675–1680.
Anderson, R., Ali, S., Bradtmiller, L., Nielsen, S., Fleisher, M., Anderson, B., Burckle, L.,
2009. Wind-driven upwelling in the Southern Ocean and the deglacial rise in atmospheric
CO2 . Science 323 (5920), 1443–1448.
147
148
Archer, D., 1996. An atlas of the distribution of calcium carbonate in sediments of the deep
sea. Global Biogeochemical Cycles 10 (1), 159–174.
Arrhenius, S., 1896. On the Influence of Carbonic Acid in the Air upon the Temperature of
the Ground. The London, Edinburgh, and Dublin Philosophical Magazine and Journal of
Science Series 5 (41), 251.
Baes, C. F., Mesmer, R. E., 1976. The Hydrolysis of Cations. New York: Wiley.
Balco, G., Rovey, C. W., 2010. Absolute chronology for major Pleistocene advances of the
Laurentide Ice Sheet. Geology 38 (9), 795–798.
Barker, S., Diz, P., Vautravers, M. J., Pike, J., Knorr, G., Hall, I. R., Broecker, W. S.,
2009. Interhemispheric Atlantic seesaw response during the last deglaciation. Nature
457 (7233), 1097–1102.
Barker, S., Elderfield, H., 2002. Foraminiferal calcification response to glacial-interglacial
changes in atmospheric CO2 . Science 297 (5582), 833–836.
Barnola, J. M., Raynaud, D., Korotkevich, Y. S., Lorius, C., 1987. Vostok ice core provides
160,000-year record of atmospheric CO2. Nature 329, 408–414.
Bassett, R. L., 1980. A critical evaluation of the thermodynamic data for boron ions, ion
pairs, complexes, and polyanions in aqueous solution at 298.15 K and 1 bar. Geochimica
Et Cosmochimica Acta 44 (8), 1151–1160.
Bé, A. W. H., 1972. Zoogeographic Provinces of Holocene Planktonic Foraminiferida.
American Association of Petroleum Geologists Bulletin 56 (3), 603.
Bé, A. W. H., 1980. Gametogenic calcification in a spinose planktonic foraminifer,
Globigerinoides sacculifer (Brady). Marine Micropaleontology 5 (3), 283–310.
Bé, A. W. H., Anderson, O. R., 1976. Gametogenesis in planktonic foraminifera. Science
192 (4242), 890–892.
Bé, A. W. H., Hemleben, C., 1970. Calcification in a living planktonic foraminifer,
Globigerinoides sacculifer . N. Jb. Geol. Paläont. 134, 221–234.
Bé, A. W. H., Spero, H. J., Anderson, O. R., 1982. Effects of symbiont elimination and
reinfection on the life processes of the planktonic foraminifer Globigerinoides sacculifer .
Marine Biology 70 (1), 73–86.
Bemis, B. E., Spero, H. J., Lea, D. W., Bijma, J., 2000. Temperature influence on the
carbon isotopic composition of Globigerina bulloides and Orbulina universa (planktonic
foraminifera). Marine Micropaleontology 38 (3-4), 213–228.
Bemis, E. B., Spero, H. J., Bijma, J., Lea, D. W., 1998. Reevaluation of the oxygen
149
isotopic composition of planktonic foraminifera: Experimental results and revised
paleotemperature equations. Paleoceanography 13 (2), 150–160.
Bentov, S., Brownlee, C., Erez, J., 2009. The role of seawater endocytosis in the
biomineralization process in calcareous foraminifera. Proceedings of the National
Academy of Sciences of the United States of America 106 (51), 21500–21504.
Bijma, J., Spero, H., Lea, D., 1999. Use of Proxies in Paleoceanography: Examples from
the South Atlantic. Springer-Verlag, Ch. “Reassessing Foraminiferal Stable Isotope
Geochemistry”, pp. 489 – 512.
Borradaile, G., 2003. Statistics of Earth Science Data: Their Distribution in Time, Space
and Orientation. Springer-Verlag.
Boyle, E. A., 1988. Vertical Oceanic Nutrient Fractionation and Glacial/Interglacial CO2
cycles. Nature 331, 55–56.
Boyle, E. A., Keigwin, L. D., 1985. Comparison of Atlantic and Pacific paleochemical
records for the last 215,000 years: Changes in deep ocean circulation and chemical
inventories. Earth and Planetary Science Letters 76 (1), 135–150.
Bradtmiller, L. I., Anderson, R. F., Sachs, J. P., 2010. A deeper respired carbon pool in the
glacial equatorial Pacific Ocean. Earth and Planetary Science Letters 299, 417–425.
Broecker, W., Putnam, A. E., 2012. How did the hydrologic cycle respond to the two-phase
mystery interval? Quaternary Science Reviews 57 (C), 17–25.
Broecker, W. S., Denton, G. H., December 1990. What drives glacial cycles? Scientific
American, 49 – 56.
Broecker, W. S., Peng, T. H., 1982. Tracers in the Sea. Lamont-Doherty Geological
Observatory, 1–702.
Broecker, W. S., Peng, T. H., 1993. Greenhouse Puzzles. Eldigio Press.
Broecker, W. S., Takahashi, T., Simpson, H., Peng, T., 1979. Fate of Fossil-Fuel
Carbon-Dioxide and the Global Carbon Budget. Science 206 (4417), 409–418.
Brown, R. E., Anderson, L. D., Thomas, E., Zachos, J. C., 2011. A core-top calibration of
B/Ca in the benthic foraminifers Nuttallides umbonifera and Oridorsalis umbonatus: A
proxy for Cenozoic bottom water carbonate saturation. Earth and Planetary Science
Letters 310 (3-4), 360–368.
Burke, A., Robinson, L. F., 2012. The Southern Ocean’s Role in Carbon Exchange During
the Last Deglaciation. Science 335 (6068), 557–561.
Byrne, R. H., Kester, D. R., 1974. Inorganic Speciation of Boron in Seawater. Journal of
150
Marine Research 32 (2), 119–127.
Byrne, R. H., Yao, W. S., Klochko, K., Tossell, J. A., Kaufman, A. J., 2006. Experimental
11
evaluation of the isotopic exchange equilibrium 10 B(OH)3 + 11 B(OH)−
B(OH)3 +
4 =
−
10
B(OH)4 in aqueous solution. Deep-Sea Research Part I - Oceanographic Research
Papers 53 (4), 684–688.
Caron, D., Roger Anderson, O., Lindsey, J. L., Faber, W., Lin Lim, E. E., 1990. Effects of
gametogenesis on test structure and dissolution of some spinose planktonic foraminifera
and implications for test preservation. Marine Micropaleontology 16 (1), 93–116.
Cheng, H., Edwards, R. L., Broecker, W. S., Denton, G. H., Kong, X., Wang, Y., Zhang,
R., Wang, X., 2009. Ice Age Terminations. Science 326 (5950), 248–252.
De Deckker, P., Moros, M., Perner, K., Jansen, E., Mar. 2012. Influence of the tropics and
southern westerlies on glacial interhemispheric asymmetry. Nature Geoscience 5 (4), 1–4.
de Nooijer, L. J., Toyofuku, T., Oguri, K., Nomaki, H., Kitazato, H., 2008. Intracellular pH
distribution in foraminifera determined by the fluorescent probe HPTS. Limnology and
Oceanography-Methods 6, 610–618.
De Yoreo, J. J., Vekilov, P. G., 2003. Principles of crystal nucleation and growth. Reviews
in Mineralogy and Geochemistry 54 (1), 57–93.
Denton, G. H., Anderson, R. F., Toggweiler, J. R., Edwards, R. L., Schaefer, J. M.,
Putnam, A. E., Jun. 2010. The Last Glacial Termination. Science 328 (5986), 1652–1656.
Denton, G. H., Hughes, T. J. (Eds.), 1981. The Last Great Ice Sheets. New York: Wiley.
DePaolo, D. J., 2011. Surface kinetic model for isotopic and trace element fractionation
during precipitation of calcite from aqueous solutions. Geochimica et Cosmochimica
Acta 75 (4), 1039–1056.
Deuser, W., Ross, E., 1989. Seasonally abundant planktonic foraminifera of the Sargasso
Sea; succession, deep-water fluxes, isotopic compositions, and paleoceanographic
implications. The Journal of Foraminiferal Research 19 (4), 268–293.
Deuser, W. G., Ross, E., Hemleben, C., Spindler, M., 1981. Seasonal changes in species
composition, numbers, mass, size, and isotopic composition of planktonic foraminifera
settling into the deep Sargasso Sea. Palaeogeography, Palaeoclimatology, Palaeoecology
33 (1), 103–127.
Dickson, A. G., 1981. An exact definition of total alkalinity and a procedure for the
estimation of alkalinity and total inorganic carbon from titration data. Deep Sea
Research Part A Oceanographic Research Papers 28 (6), 609–623.
Dickson, A. G., 1990. Thermodynamics of the Dissociation of Boric Acid in Potassium
151
Chloride Solutions from 273.15-K to 318.15-K. Journal of Chemical and Engineering
Data 35 (3), 253–257.
Dickson, A. G., Sabine, C. L., Christian, J. R. (Eds.), 2007. Guide to best practices for
ocean CO2 measurements. Vol. 3. PICES Special Publication.
Doney, S. C., Fabry, V. J., Feely, R. A., Kleypas, J. A., 2009. Ocean Acidification: The
Other CO2 Problem. Annual Review of Marine Science 1, 169–192.
Dyrssen, D., Hansson, I., 1973. Ionic medium effects in sea water - A comparison of acidity
constants of carbonic acid and boric acid in sodium chloride and synthetic seawater.
Marine Chemistry 1, 137–149.
Eggins, S., De Deckker, P., Marshall, J., 2002. Laser ablation-ICP-MS compositional
profiling of chamber walls in planktonic Foraminifera: implications for Mg/Ca
thermometry. Geochimica et Cosmochimica Acta 66 (15A), A207–A207.
Eggins, S., De Deckker, P., Marshall, J., 2003. Mg/Ca variation in planktonic foraminifera
tests: implications for reconstructing palaeo-seawater temperature and habitat
migration. Earth and Planetary Science Letters 212 (3-4), 291–306.
Eggins, S. M., Sadekov, A., De Deckker, P., 2004. Modulation and daily banding of Mg/Ca
in Orbulina universa tests by symbiont photosynthesis and respiration: a complication
for seawater thermometry? Earth and Planetary Science Letters 225 (3-4), 411–419.
Elderfield, H., Bertram, C. J., Erez, J., 1996. Biomineralization model for the incorporation
of trace elements into foraminiferal calcium carbonate. Earth and Planetary Science
Letters 142 (3-4), 409–423.
Emerson, S., Hedges, J., 2008. Chemical Oceanography and the Marine Carbon Cycle.
Cambridge University Press.
Emiliani, C., 1955. Pleistocene Temperatures. Journal of Geology 63 (6), 538–578.
Emiliani, C., 1966. Isotopic Paleotemperatures. Science 154 (3751), 851–857.
Erez, J., 2003. The source of ions for biomineralization in foraminifera and their
implications for paleoceanographic proxies. Vol. 54. Mineralogical Society of America,
Washington, pp. 115–149.
Fabry, V. J., Seibel, B. A., Feely, R. A., Orr, J. C., 2008. Impacts of ocean acidification on
marine fauna and ecosystem processes. ICES Journal of Marine Science 65 (3), 414–432.
Fairbanks, R., Wiebe, P. H., Bé, A., 1980. Vertical distribution and isotopic composition of
living planktonic foraminifera in the western North Atlantic. Science 207 (4426), 61–63.
Fairbanks, R. G., Sverdlove, M., Free, R., Wiebe, P. H., Bé, A. W. H., 1982. Vertical
152
distribution and isotopic fractionation of living planktonic foraminifera from the Panama
Basin. Nature 298 (5877), 841–844.
Feely, R., Sabine, C., Lee, K., Berelson, W., Kleypas, J., Fabry, V., Millero, F., 2004.
Impact of anthropogenic CO2 on the CaCO3 system in the oceans. Science 305 (5682),
362–366.
Feng, Q., 2011. Molecular Biomineralization. Springer-Verlag, Ch. Principles of
Calcium-based Biomineralization, pp. 141–197.
Foster, G. L., 2008. Seawater pH, pCO2 and [CO23 -] variations in the Caribbean Sea over
the last 130 kyr: A boron isotope and B/Ca study of planktic forminifera. Earth and
Planetary Science Letters 271 (1-4), 254–266.
Foster, G. L., Pogge von Strandmann, P. A. E., Rae, J. W. B., 2010. Boron and magnesium
isotopic composition of seawater. Geochemistry Geophysics Geosystems 11, Q08015.
Furst, M., Lowenstam, H. A., Burnett, D. S., 1976. Radiographic study of the distribution
of boron in recent mollusc shells. Geochimica Et Cosmochimica Acta 40 (11), 1381–1386.
Ganachaud, A., Wunsch, C., 2000. Improved estimates of global ocean circulation, heat
transport and mixing from hydrographic data. Nature 408 (6811), 453–457.
Gordon, A. L., 1981. Seasonality of Southern Ocean Sea Ice. Journal of Geophysical
Research 86 (C5), 4193–4197.
Hamilton, C. P., Spero, H. J., Bijma, J., Lea, D. W., 2008. Geochemical investigation of
gametogenic calcite addition in the planktonic foraminifera Orbulina universa. Marine
Micropaleontology 68 (3-4), 256–267.
Hansen, J., Lacis, A., Rind, D., Russell, G., Stone, P., Fung, I., Ruedy, R., Lerner, J., 1984.
Climate Processes and Climate Sensitivity. Vol. Maurice Ewing Volume 5 of AGU
Geopysical Monograph 29. American Geophysical Union, Ch. Climate Sensitivity:
Analysis of Feedback Mechanisms, pp. 130–163.
Hansson, I., 1973. A new set of pH-scales and standard buffers for sea water. Deep Sea
Research and Oceanographic Abstracts 20, 479–491.
Hathorne, E. C., James, R., Lampitt, R. S., 2009. Environmental versus biomineralization
controls on the intratest variation in the trace element composition of the planktonic
foraminifera g. inflata and g. scitula. Paleoceanography 24.
Hawley, J., Pytkowicz, R. M., 1969. Solubility of calcium carbonate in seawater at high
pressures and 2 ◦ C. Geochimica et Cosmochimica Acta 33 (12), 1557–1561.
Hays, J. D., Imbrie, J., Shackleton, N. J., 1976. Variations in the Earth’s orbit: pacemaker
of the ice ages. Science 194, 1121–1132.
153
Heath, R. A., 1985. A review of the physical oceanography of the seas around New Zealand
. New Zealand Journal of Marine and Freshwater Research 19 (1), 79–124.
Hedges, J. I., Baldock, J. A., Gelinas, Y., Lee, C., Peterson, M. L., 2002. The biochemical
and elemental compositions of marine plankton: A NMR perspective. Marine Chemistry.
Hemming, N., Hanson, G., 1992. Boron Isotopic Composition and Concentration in
Modern Marine Carbonates. Geochimica et Cosmochimica Acta 56 (1), 537–543.
Hemming, N. G., Honisch, B., Jan. 2007. Boron Isotopes in Marine Carbonate Sediments
and the pH of the Ocean. In: Developments in Marine Geology. Elsevier, pp. 717–734.
Hemming, N. G., Reeder, R. J., Hanson, G. N., 1995. Mineral-fluid partitioning and
isotopic fractionation of boron in synthetic calcium carbonate. Geochimica et
Cosmochimica Acta 59 (2), 371–379.
Hendry, K. R., Rickaby, R. E. M., Meredith, M. P., Elderfield, H., 2009. Controls on stable
isotope and trace metal uptake in Neogloboquadrina pachyderma (sinistral) from an
Antarctic sea-ice environment. Earth and Planetary Science Letters 278, 67 – 77.
Hershey, J. P., Fernandez, M., Milne, P. J., Millero, F. J., 1986. The ionization of boric
acid in NaCl, Na-Ca-Cl and Na-Mg-Cl solutions at 25 degrees C. Geochimica et
Cosmochimica Acta 50 (1), 143–148.
Hobbs, M. Y., Reardon, E. J., 1999. Effect of pH on boron coprecipitation by calcite:
Further evidence for nonequilibrium partitioning of trace elements. Geochimica et
Cosmochimica Acta 63 (7-8), 1013–1021.
Hönisch, B., Bijma, J., Russell, A. D., Spero, H. J., Palmer, M. R., Zeebe, R. E.,
Eisenhauer, A., 2003. The influence of symbiont photosynthesis on the boron isotopic
composition of foraminifera shells. Marine Micropaleontology 49 (1-2), 87–96.
Hönisch, B., Hemming, N. G., 2004. Ground-truthing the boron isotope-paleo-pH proxy in
planktonic foraminifera shells: Partial dissolution and shell size effects.
Paleoceanography 19 (4), PA4010.
Hönisch, B., Hemming, N. G., 2005. Surface ocean pH response to variations in pCO2
through two full glacial cycles. Earth and Planetary Science Letters 236 (1-2), 305–314.
Hönisch, B., Hemming, N. G., Archer, D., Siddall, M., McManus, J. F., 2009. Atmospheric
Carbon Dioxide Concentration Across the Mid-Pleistocene Transition. Science
324 (5934), 1551–1554.
Hönisch, B., Hemming, N. G., Loose, B., 2007. Comment on “A critical evaluation of the
boron isotope-pH proxy: The accuracy of ancient ocean pH estimates” by M. Pagani, D.
Lemarchand, A. Spivack and J. Gaillardet. Geochimica et Cosmochimica Acta 71 (6),
1636–1641.
154
Imbrie, J., Berger, A., Boyle, E. A., Clemens, S. C., Duffy, A., Howard, W. R., Kukla, G.,
Kutzbach, J., Martinson, D. G., McIntyre, A., Mix, A. C., Molfino, B., Morley, J. J.,
Peterson, L. C., Pisias, N. G., Prell, W. L., Raymo, M. E., Shackleton, N. J., Toggweiler,
J. R., 1993. The 100,000-Year Cycle. Paleoceanography 8 (6), 699–735.
Ingle, S. E., 1975. Solubility of calcite in the ocean. Marine Chemistry 3 (4), 301–319.
Ingri, N., 1963. Polyborates in 3.0 M Na(Br), 3.0 M Li(Br) and 3.0 M K(Br). Acta
Chemica Scandinavica 17, 581–589.
Jaccard, S. L., Galbraith, E. D., Sigman, D. M., Haug, G. H., Francois, R., Pedersen,
T. F., Dulski, P., Thierstein, H. R., 2009. Subarctic Pacific evidence for a glacial
deepening of the oceanic respired carbon pool. Earth and Planetary Science Letters
277 (1-2), 156–165.
Jørgensen, B. B., Erez, J., Revsbech, N. P., Cohen, Y., 1985. Symbiotic photosynthesis in a
planktonic foraminiferan, Globigerinoides sacculifer (Brady), studied with
microelectrodes. Limnology and Oceanography 30 (6), 1253–1267.
Katz, M. E., Cramer, B. S., Franzese, A., Hönisch, B., Miller, K. G., Rosenthal, Y.,
Wright, J. D., 2010. Traditional and Emerging Geochemical Proxies in Foraminifera.
Journal of Foraminiferal Research 40 (2), 165–192.
Key, R. M., Kozyr, A., Sabine, C. L., Lee, K., Wanninkhof, R., Bullister, J. L., Feely,
R. A., Millero, F. J., Mordy, C., Peng, T. H., 2004. A global ocean carbon climatology:
Results from Global Data Analysis Project (GLODAP). Global Biogeochemical Cycles
18 (4), GB4031.
Kitano, Y., Okumura, M., Idogaki, M., 1978. Coprecipitation of borate-boron with calcium
carbonate. Geochemical Journal 12 (3), 183–189.
Klochko, K., Cody, G. D., Tossell, J. A., Dera, P., Kaufman, A. J., 2009. Re-evaluating
boron speciation in biogenic calcite and aragonite using 11 B MAS NMR. Geochimica et
Cosmochimica Acta 73 (7), 1890–1900.
Klochko, K., Kaufman, A. J., Yao, W., Byrne, R. H., Tossell, J. A., 2006. Experimental
measurement of boron isotope fractionation in seawater. Earth and Planetary Science
Letters 248 (1-2), 276–285.
Krief, S., Hendy, E. J., Fine, M., Yam, R., Meibom, A., Foster, G. L., Shemesh, A., 2010.
Physiological and isotopic responses of scleractinian corals to ocean acidification.
Geochimica et Cosmochimica Acta 74 (17), 4988–5001.
Kuroyanagi, A., Tsuchiya, M., Kawahata, H., Kitazato, H., 2008. The occurrence of two
genotypes of the planktonic foraminifer Globigerinoides ruber (white) and
paleo-environmental implications. Marine Micropaleontology 68 (3-4), 236–243.
155
Langer, M. R., 1992. Biosynthesis of glycosaminoglycans in Foraminifera - A Review.
Marine Micropaleontology 19 (3), 245–255.
Lea, D. W., 2003. Elemental and Isotopic Proxies of Past Ocean Temperatures. Vol. 6 of
Treatise on Geochemistry. Elsevier.
Lea, D. W., Bijma, J., Spero, H. J., Archer, D., 1999. Implications of a Carbonate Ion
Effect on Shell Carbon and Oxygen Isotopes for Glacial Ocean Conditions. Use of
Proxies in Paleoceanography: Examples from the South Atlantic, 513 – 522.
Lee, J. J., Anderson, O. R. (Eds.), 1991. Biology of Foraminifera. Academic Press.
Lee, K., Kim, T. W., Byrne, R. H., Millero, F. J., Feely, R. A., Liu, Y. M., 2010. The
universal ratio of boron to chlorinity for the North Pacific and North Atlantic oceans.
Geochimica et Cosmochimica Acta 74 (6), 1801–1811.
Leeman, W., Sisson, V., 2002. Geochemistry of Boron and its implications for crustal and
mantle processes. Vol. 33 of Reviews in Mineralogy. Mineralogical Society of America,
Ch. 12, pp. 645–695.
Lemarchand, D., Gaillardet, J., Lewin, E., Allègre, C. J., 2000. The influence of rivers on
marine boron isotopes and implications for reconstructing past ocean pH. Nature
408 (6815), 951–954.
Lemarchand, D., Gaillardet, J., Lewin, E., Allègre, C. J., 2002. Boron isotope systematics
in large rivers: implications for the marine boron budget and paleo-pH reconstruction
over the Cenozoic. Chemical Geology 190 (1-4), 123–140.
Lewis, E., Perkin, R. G., 1978. Salinity: Its Definition and Calculation. Journal of
Geophysical Research 83, 466–478.
Li, Y.-H., Takahashi, T., Broecker, W. S., 1969. Degree of saturation of CaCO3 in the
oceans. Journal of Geophysical Research 74 (23), 5507–5525.
Lisiecki, L. E., Raymo, M. E., 2005. A Pliocene-Pleistocene stack of 57 globally distributed
benthic delta O-18 records. Paleoceanography 20 (1), PA1003.
Liu, Y., Tossell, J. A., Aug 2005. Ab initio molecular orbital calculations for boron isotope
fractionations on boric acids and borates. Geochimica et Cosmochimica Acta 69 (16),
3995–4006.
Lohmann, G. P., 1995. A model for variation in the chemistry of planktonic foraminifera
due to secondary calcification and selective dissolution. Paleoceanography 10 (3),
445–457.
Lueker, T. J., Dickson, A. G., Keeling, C. D., 2000. Ocean pCO2 calculated from dissolved
inorganic carbon, alkalinity, and equations for K-1 and K-2: validation based on
156
laboratory measurements of CO2 in gas and seawater at equilibrium. Marine Chemistry
70 (1-3), 105–119.
Lumpkin, R., Speer, K., Oct. 2007. Global Ocean Meridional Overturning. Journal of
Physical Oceanography 37 (10), 2550–2562.
Lynch-Stieglitz, J., Stocker, T. F., Broecker, W. S., Fairbanks, R., 1995. The influence of
air-sea exchange on the isotopic composition of oceanic carbon: Observations and
modeling. Global Biogeochemical Cycles 9 (4), 653–665.
Macdonald, A. M., Mecking, S., Robbins, P. E., Toole, J. M., Johnson, G. C., Talley, L.,
Cook, M., Wijffels, S. E., Sep. 2009. The WOCE-era 3-D Pacific Ocean circulation and
heat budget. Progress in Oceanography 82 (4), 281–325.
Mackin, J. E., 1986. The free-solution diffusion coefficient of boron: Influence of dissolved
organic matter. Marine Chemistry 20, 131 – 140.
Malone, M. J., Martin, J. B., Schönfeld, J., Ninnemann, U. S., Nürnburg, D., White, T. S.,
2004. The oxygen isotopic composition and temperature of Southern Ocean bottom
waters during the last glacial maximum. Earth and Planetary Science Letters 222 (1),
275–283.
Marchitto, T. M., Lynch-Stieglitz, J., Hemming, S. R., 2005. Deep Pacific CaCO3
compensation and glacial-interglacial atmospheric CO2 . Earth and Planetary Science
Letters 231, 317–336.
McCave, I. N., Carter, L., Hall, I. R., 2008. Glacial–interglacial changes in water mass
structure and flow in the SW Pacific Ocean. Quaternary Science Reviews 27 (19-20),
1886–1908.
MEDAR, 2001. Mediterranean and black sea database of temperature, salinity, and
bio-chemical parameters. Tech. rep., Climatologial Atlas, IFREMER Edition ed.
Mehrbach, C., Culberson, C. H., Hawley, J. E., Pytkowicz, R. M., 1973. Measurement of
the apparent dissociation constants of carbonic acid in seawater at atmospheric pressure.
Limnology and oceanography 18 (6), 897–907.
Millero, F., 2006. Chemical Oceanography. Taylor and Francis.
Millero, F., Schreiber, D. R., 1982. Use of the ion pairing model to estimate activity
coefficients of the ionic components of natural waters. American Journal of Science
282 (9), 1508–1540.
Monnin, E., Indermüle, A., Dällenbach, A., Flückiger, J., Stauffer, B., Stocker, T. F.,
Raynaud, D., Barnola, J. M., 2001. Atmospheric CO2 Concentrations over the Last
Glacial Termination. Science 291 (5501), 112–114.
157
Morse, J. W., 1986. The Surface Chemistry of Calcium Carbonate Minerals in Natural
Waters - An Overview. Marine Chemistry 20 (1), 91–112.
Morse, J. W., Bender, M. L., 1990. Partition Coefficients in Calcite - Examination of
Factors Influencing the Validity of Experimental Results and their Application to
Natural Systems. Chemical Geology 82 (3-4), 265–277.
Mucci, A., 1983. The solubility of calcite and aragonite in seawater at various salinities,
temperatures, and one atmosphere total pressure. American Journal of Science 283,
780–799.
Nelson, C., Cooke, P., Hendy, C., Cuthbertson, A., 1993. Oceanographic and climatic
changes over the past 160,000 years at Deep Sea Drilling Project Site 594 off
southeastern New Zealand, southwest Pacific Ocean. Paleoceanography 8 (4), 435–458.
Ni, Y., Foster, G. L., Bailey, T., Elliott, T., Schmidt, D. N., Pearson, P., Haley, B., Coath,
C., 2007. A core top assessment of proxies for the ocean carbonate system in
surface-dwelling foraminifers. Paleoceanography 22 (3), PA3212.
Ninnemann, U. S., Charles, C. D., 2002. Changes in the mode of Southern Ocean
circulation over the last glacial cycle revealed by foraminiferal stable isotopic variability.
Earth and Planetary Science Letters 201 (2), 383–396.
Nutman, A. P., Bennett, V. C., Friend, C. R. L., 2012. Waves and weathering at 3.7 Ga:
Geological evidence for an equitable terrestrial climate under the faint early Sun.
Australian Journal of Earth Sciences 59 (2, SI), 167–176.
Orr, J., Fabry, V., Aumont, O., Bopp, L., Doney, S., Feely, R., Gnanadesikan, A., Gruber,
N., Ishida, A., Joos, F., Key, R., Lindsay, K., Maier-Reimer, E., Matear, R., Monfray, P.,
Mouchet, A., Najjar, R., Plattner, G., Rodgers, K., Sabine, C., Sarmiento, J., Schlitzer,
R., Slater, R., Totterdell, I., Weirig, M., Yamanaka, Y., Yool, A., SEP 29 2005.
Anthropogenic ocean acidification over the twenty-first century and its impact on
calcifying organisms. Nature 437 (7059), 681–686.
Orsi, A. H., Johnson, G. C., Bullister, J. L., 1999. Circulation, mixing, and production of
Antarctic Bottom Water. Progress in Oceanography 43 (1), 55–109.
Orsi, A. H., Whitworth, T., Nowlin, W. D., 1995. On the meridional extent and fronts of
the Antarctic Circumpolar Current. Deep-Sea Research Part I 42 (5), 641–673.
Pagani, M., Lemarchand, D., Spivack, A., Gaillardet, J., 2005. A critical evaluation of the
boron isotope-pH proxy: The accuracy of ancient ocean pH estimates. Geochimica et
Cosmochimica Acta 69 (4), 953–961.
Paillard, D., L. L., Yiou, P., 1996. Macintosh program performs time-series analysis. Eos
Trans. AGU 77, 379.
158
Palmer, M. R., 2003. A 23,000-Year Record of Surface Water pH and pCO2 in the Western
Equatorial Pacific Ocean. Science 300 (5618), 480–482.
Palmer, M. R., Brummer, G. J., Cooper, M. J., Elderfield, H., Greaves, M. J., Reichart,
G. J., Schouten, S., Yu, J. M., 2010. Multi-proxy reconstruction of surface water pCO2 in
the northern Arabian Sea since 29ka. Earth and Planetary Science Letters 295 (1-2),
49–57.
Pearson, P. N., 1999. Middle Eocene Seawater pH and Atmospheric Carbon Dioxide
Concentrations. Science 284 (5421), 1824–1826.
Pearson, P. N., Palmer, M. R., 2000. Atmospheric carbon dioxide concentrations over the
past 60 million years. Nature 406 (6797), 695–699.
Peltier, W. R., Fairbanks, R., 2006. Global glacial ice volume and Last Glacial Maximum
duration from an extended Barbados sea level record. Quaternary Science Reviews.
Peng, T., 2007. Gulf of mexico and east coast carbon cruise report. Tech. rep., NOAA.
Petit, J. R., Jouzel, J., Raynaud, D., Barkov, N. I., Barnola, J. M., Basile, I., Bender, M.,
Chappellaz, J., Davis, M., Delaygue, G., 1999. Climate and atmospheric history of the
past 420,000 years from the Vostok ice core, Antarctica. Nature 399 (6735), 429–436.
Poisson, A., Curie, M., 1981. Conductivity ratio measurement of diluted or concentrated
standard seawater and KCl solution used to establish the new practical salinity scale,
1978. UNESCO Technical Papers in Marine Science.
Polat, A., Frei, R., 2005. The origin of early Archean banded iron formations and of
continental crust, Isua, southern West Greenland. Precambrian Research 138 (1-2),
151–175.
Putnam, A. E., Denton, G. H., Schaefer, J. M., Barrell, D. J. A., Andersen, B. G., Finkel,
R. C., Schwartz, R., Doughty, A. M., Kaplan, M. R., Schlüchter, C., 2010. Glacier
advance in southern middle-latitudes during the Antarctic Cold Reversal. Nature
Geoscience 3 (10), 700–704.
Rae, J. W. B., Foster, G. L., Schmidt, D. N., Elliott, T., 2011. Boron isotopes and B/Ca in
benthic foraminifera: Proxies for the deep ocean carbonate system. Earth and Planetary
Science Letters 302, 403 – 413.
Raymo, M. E., 1994. The initiation of northern hemisphere glaciation. Annual Review of
Earth and Planetary Sciences 22, 353–383.
Reid, J. L., 1997. On the total geostrophic circulation of the Pacific Ocean: flow patterns,
tracers, and transports. Progress in Oceanography 39 (4), 263–352.
Richey, J. N., Hollander, D. J., Flower, B. P., Eglinton, T. I., 2011. Merging late Holocene
159
molecular organic and foraminiferal-based geochemical records of sea surface
temperature in the Gulf of Mexico. Paleoceanography 26 (1).
Richey, J. N., Poore, R. Z., Flower, B. P., Quinn, T. M., 2007. 1400 yr multiproxy record of
climate variability from the northern Gulf of Mexico. Geology 35 (5), 423.
Riley, J. P., Tongudai, M., 1967. The major cation/chlorinity ratios in seawater. Chemical
Geology 2, 263 – 269.
Rind, D., Apr. 2002. The Sun’s Role in Climate Variations. Science 296 (5568), 673–677.
Rink, S., Kuhl, M., Bijma, J., Spero, H. J., 1998. Microsensor studies of photosynthesis and
respiration in the symbiotic foraminifer Orbulina universa. Marine Biology 131 (4),
583–595.
Rosenthal, Y., Boyle, E. A., Labeyrie, L., 1997. Last Glacial Maximum paleochemistry and
deepwater circulation in the Southern Ocean: Evidence from foraminiferal cadmium.
Paleoceanography 12 (6), 787–796.
Russell, A. D., Hönisch, B., Spero, H. J., Lea, D. W., 2004. Effects of seawater carbonate
ion concentration and temperature on shell U, Mg, and Sr in cultured planktonic
foraminifera. Geochimica et Cosmochimica Acta 68 (21), 4347–4361.
Sabine, C. L., Feely, R. A., Gruber, N., Key, R. M., Lee, K., Bullister, J. L., Wanninkhof,
R., Wong, C. S., Wallace, D., Tilbrook, B., Millero, F. J., Peng, T. H. Kozyr, A., Ono,
T., Rios, A., 2004. The Oceanic Sink for Anthropogenic CO2 . Science 305 (5682),
367–371.
Sanyal, A., Bijma, J., Spero, H. J., Lea, D. W., 2001. Empirical relationship between pH
and the boron isotopic composition of Globigerinoides sacculifer : Implications for the
boron isotope paleo-pH proxy. Paleoceanography 16 (5), 515–519.
Sanyal, A., Hemming, N. G., Broecker, W. S., Lea, D. W., Spero, H. J., Hanson, G. N.,
1996. Oceanic pH control on the boron isotopic composition of foraminifera: Evidence
from culture experiments. Paleoceanography 11 (5), 513–517.
Sanyal, A., Hemming, N. G., Hanson, G. N., Broecker, W. S., 1995. Evidence for a higher
pH in the glacial ocean from boron isotopes in foraminifera. Nature 373 (6511), 234–236.
Sanyal, A., Nugent, M., Reeder, R. J., Bijma, J., 2000. Seawater pH control on the boron
isotopic composition of calcite: Evidence from inorganic calcite precipitation
experiments. Geochimica et Cosmochimica Acta 64 (9), 1551–1555.
Sarmiento, J. L., Gruber, N., 2006. Ocean Biogeochemical Dynamics. Princeton University
Press.
Seki, O., Foster, G. L., Schmidt, D. N., Mackensen, A., Kawamura, K., Pancost, R. D.,
160
2010. Alkenone and boron-based Pliocene pCO2 records. Earth and Planetary Science
Letters 292 (1-2), 201–211.
Sen, S., Stebbins, J. F., Hemming, N. G., Ghosh, B., 1994. Coordination Environments of
B-impurities in Calcite and Aragonite Polymorphs: A 11 B MAS NMR-Study. American
Mineralogist 79 (9-10), 819–825.
Shackleton, N., 1987. Oxygen Isotopes, Ice Volume, and Sea-Level. Quaternary Science
Reviews 6 (3-4), 183–190.
Shakun, J. D., Clark, P. U., He, F., Marcott, S. A., Mix, A. C., Liu, Z., 2012. Global
warming preceded by increasing carbon dioxide concentrations during the last
deglaciation. Nature 484, 49–54.
Shane, P., Sikes, E. L., Guilderson, T. P., Jun. 2006. Tephra beds in deep-sea cores off
northern New Zealand: implications for the history of Taupo Volcanic Zone, Mayor
Island and White Island volcanoes. Journal of Volcanology and Geothermal Research
154 (3-4), 276–290.
Sigman, D. M., Boyle, E. A., 2000. Glacial/interglacial variations in atmospheric carbon
dioxide. Nature 407 (6806), 859–869.
Sikes, E. L., Samson, C., Guilderson, T. P., Howard, W., 2000. Old radiocarbon ages in the
southwest Pacific Ocean during the last glacial period and deglaciation. Nature
405 (6786), 555–559.
Simon, L., Lécuyer, C., Maréchal, C., Coltice, N., 2006. Modelling the geochemical cycle of
boron: Implications for the long-term δ 11 B evolution of seawater and oceanic crust.
Chemical Geology 225 (1-2), 61–76.
Simonson, J. M., Roy, R. N., Roy, L. N., Johnson, D. A., 1987. The thermodynamics of
aqueous borate solutions, 1. Mixtures of boric acid with sodium or potassium borate and
chloride. Journal of Solution Chemistry 16 (10), 791–803.
Sinclair, D., 2005. Correlated trace element “vital effects” in tropical corals: A new
geochemical tool for probing biomineralization. Geochimica et Cosmochimica Acta
69 (13), 3265–3284.
Skinner, L. C., Fallon, S., Waelbroeck, C., Michel, E., Barker, S., 2010. Ventilation of the
Deep Southern Ocean and Deglacial CO2 Rise. Science 328 (5982), 1147–1151.
Skirrow, G., 1975. Chemical Oceanography. Academic Press, London.
Sloyan, B. M., Rintoul, S. R., 2001. The Southern Ocean Limb of the Global Deep
Overturning Circulation. Journal of Physical Oceanography 31 (1), 143–173.
Spero, H. J., 1988. Ultrastructural Examination of Chamber Morphogenesis and
161
Biomineralization in the Planktonic Foraminifer Orbulina universa. Marine Biology
99 (1), 9–20.
Spero, H. J., Bijma, J., Lea, D. W., Bemis, B. E., 1997. Effect of seawater carbonate
concentration on foraminiferal carbon and oxygen isotopes. Nature 390 (6659), 497–500.
Spero, H. J., Lea, D. W., 1993. Intraspecific stable isotope variability in the planktic
foraminiferaGlobigerinoides sacculifer: Results from laboratory experiments. Marine
Micropaleontology 22 (3), 221–234.
Spero, H. J., Mielke, K. M., Kalve, E. M., Lea, D. W., Pak, D. K., 2003. Multispecies
approach to reconstructing eastern equatorial Pacific thermocline hydrography during
the past 360 kyr. Paleoceanography 18 (1), 1022.
Spero, H. J., Parker, S. L., 1985. Photosynthesis in the Symbiotic Planktonic Foraminifer
Orbulina universa and its Potential Contribution to Oceanic Primary Productivity.
Journal of Foraminiferal Research 15 (4), 273–281.
Spivack, A., Palmer, M. R., Edmond, J. M., 1987. The sedimentary cycle of the boron
isotopes. Geochimica et Cosmochimica Acta 51 (7), 1939–1949.
Stoll, H., Langer, G., Shimizu, N., Kanamaru, K., 2012. B/Ca in coccoliths and
relationship to calcification vesicle pH and dissolved inorganic carbon concentrations.
Geochimica Et Cosmochimica Acta 80 (C), 143–157.
Stumm, W., Morgan, J. J., 1996. Aquatic Chemistry: Chemical Equilibria and Rates in
Natural Waters. John Wiley and Sons, Inc.
Svendsen, J. I., Alexanderson, H., Astakhov, V. I., Demidov, I., Dowdeswell, J. A., Funder,
S., Gataullin, V., Henriksen, M., Hjort, C., Houmark-Nielsen, M., 2004. Late Quaternary
ice sheet history of northern Eurasia. Quaternary Science Reviews 23 (11), 1229–1271.
Takahashi, T., Broecker, W. S., 1981. The alkalinity and total carbon dioxide concentration
in the world oceans. Carbon Cycle Modelling 16, 271–286.
Takahashi, T., Sutherland, S., Sweeney, C., Poisson, A., Metzl, N., Tilbrook, B., Bates, N.,
Wanninkhof, R., Feely, R. A., Sabine, C., 2002. Global sea-air CO2 flux based on
climatological surface ocean pCO2, and seasonal biological and temperature effects. Deep
Sea Research Part II: Topical Studies in Oceanography 49 (9-10), 1601–1622.
Takahashi, T., Sutherland, S. C., Wanninkhof, R., Sweeney, C., Feely, R. A., Chipman,
D. W., Hales, B., Friederich, G., Chavez, F., Sabine, C., Watson, A., Bakker, D. C. E.,
Schuster, U., Metzl, N., Yoshikawa-Inoue, H., Ishii, M., Midorikawa, T., Nojiri, Y.,
Körtzinger, A., Steinhoff, T., Hoppema, M., Olafsson, J., Arnarson, T. S., Tilbrook, B.,
Johannessen, T., Olsen, A., Bellerby, R., Wong, C. S., Delille, B., Bates, N. R., de Baar,
H. J. W., Apr. 2009. Climatological mean and decadal change in surface ocean pCO2,
162
and net sea–air CO2 flux over the global oceans. Deep Sea Research Part II: Topical
Studies in Oceanography 56 (8-10), 554–577.
Tedesco, K. A., Thunell, R. C., 2003. Seasonal and interannual variations in planktonic
foraminiferal flux and assemblage composition in the Cariaco Basin, Venezuela. Journal
of Foraminiferal Research 33 (3), 192–210.
Tossell, J. A., 2005. Boric acid, “carbonic” acid, and N-containing oxyacids in aqueous
solution: Ab initio studies of structure, pKa, NMR shifts, and isotopic fractionations.
Geochimica Et Cosmochimica Acta 69 (24), 5647–5658.
Tossell, J. A., 2006. Boric acid adsorption on humic acids: Ab initio calculation of
structures, stabilities, 11 B NMR and 11 B, 10 B isotopic fractionations of surface
complexes. Geochimica et Cosmochimica Acta 70 (20), 5089–5103.
Tripati, A. K., Roberts, C. D., Eagle, R. A., 2009. Coupling of CO2 and Ice Sheet Stability
Over Major Climate Transitions of the Last 20 Million Years. Science 326 (5958),
1394–1397.
Tripati, A. K., Roberts, C. D., Eagle, R. A., Li, G., 2011. A 20 million year record of
planktic foraminiferal B/Ca ratios: Systematics and uncertainties in pCO2
reconstructions. Geochimica Et Cosmochimica Acta 75 (10), 2582–2610.
Trotter, J., Montagna, P., McCulloch, M. T., Silenzi, S., Reynaud, S., Mortimer, G.,
Martin, S., Ferrier-Pagès, C., Gattuso, J.-P., Rodolfo-Metalpa, R., 2011. Quantifying the
pH ‘vital effect’ in the temperate zooxanthellae coral Cladocora caespitosa: Validation of
the boron seawater pH proxy. Earth and Planetary Science Letters 303, 163–173.
Tyndall, J., 1861. On the absorption and radiation of heat by gases and vapors, and on the
physical connexion of radiation, absorption, and conduction. Philosophical Magazine
Series 4 (22), 169–194, 273–285.
Uppström, L., 1974. The boron/chlorinity ratio of deep-sea water from the Pacific Ocean.
Deep Sea Research 21, 161–162.
van Cappellen, P., Charlet, L., Stumm, W., Wersin, P., 1993. A Surface Complexation
Model of the Carbonate Mineral-Aqueous Solution Interface. Geochimica et
Cosmochimica Acta 57 (15), 3505–3518.
van Heuven, S., Pierrot, D., Rae, J., Lewis, E., Wallace, D., 2011. MATLAB Program
Developed for CO2 System Calculations. Tech. rep., Carbon Dioxide Information
Analysis Center.
Vengosh, A., Kolodny, Y., Starinsky, A., Chivas, A. R., McCulloch, M. T., 1991.
Coprecipitation and isotopic fractionation of boron in modern biogenic carbonates.
Geochimica et Cosmochimica Acta 55 (10), 2901–2910.
163
Waelbroeck, C., Paul, A., Kucera, M., eacute, A. R.-M., Weinelt, M., Schneider, R., Mix,
A. C., Abelmann, A., Armand, L., Bard, E., Barker, S., Barrows, T. T., Benway, H.,
Cacho, I., Chen, M. T., Cortijo, E., Crosta, X., de Vernal, A., Dokken, T., Duprat, J.,
Elderfield, H., Eynaud, F., Gersonde, R., Hayes, A., Henry, M., Hillaire-Marcel, C.,
Huang, C. C., Jansen, E., Juggins, S., Kallel, N., Kiefer, T., Kienast, M., Labeyrie, L.,
Leclaire, H., Londeix, L., Mangin, S., Matthiessen, J., Marret, F., Meland, M., Morey,
A. E., Mulitza, S., Pflaumann, U., Pisias, N. G., Radi, T., Rochon, A., Rohling, E. J.,
Sbaffi, L., fer Neth, C. S. a., Solignac, S., Spero, H., Tachikawa, K., Turon, J. L., 2009.
Constraints on the magnitude and patterns of ocean cooling at the Last Glacial
Maximum. Nature Geoscience 2 (2), 127–132.
Weiner, S., Addadi, L., 1997. Design strategies in mineralized biological materials. Journal
of Materials Chemistry 7 (5), 689–702.
Weiner, S., Addadi, L., 2011. Crystallization Pathways in Biomineralization. Annual
Review of Materials Research 41 (1), 21–40.
Weiner, S., Dove, P. M., 2003. An overview of biomineralization processes and the problem
of the vital effect. Vol. 54. Mineralogical Society of America, Washington.
Williams, D. F., Bé, A. W. H., Fairbanks, R., 1981. Seasonal stable isotopic variations in
living planktonic foraminifera from Bermuda plankton tows. Palaeogeography,
Palaeoclimatology, Palaeoecology 33 (1), 71–102.
Wolf-Gladrow, D., Riebesell, U., 1997. Diffusion and reactions in the vicinity of plankton: a
refined model for inorganic carbon transport. Marine Chemistry 59 (1), 17–34.
Wolf-Gladrow, D. A., Bijma, J., Zeebe, R. E., 1999a. Model simulation of the carbonate
chemistry in the microenvironment of symbiont bearing foraminifera. Marine Chemistry
64 (3), 181–198.
Wolf-Gladrow, D. A., Riebesell, U., Burkhardt, S., Bijma, J., 1999b. Direct effects of CO2
concentration on growth and isotopic composition of marine plankton. Tellus B
51461-476.
Wolf-Gladrow, D. A., Zeebe, R. E., Klaas, C., Koertzinger, A., Dickson, A. G., JUL 2007.
Total alkalinity: The explicit conservative expression and its application to
biogeochemical processes. Marine Chemistry 106 (1-2), 287–300.
Wunsch, C., 2003. The spectral description of climate change including the 100 ky energy.
Climate Dynamics 20 (4), 353–363.
Yu, J., Broecker, W. S., Elderfield, H., Jin, Z., McManus, J., Zhang, F., Nov. 2010a. Loss
of Carbon from the Deep Sea Since the Last Glacial Maximum. Science 330 (6007),
1084–1087.
164
Yu, J. M., Elderfield, H., Jun 2007. Benthic foraminiferal B/Ca ratios reflect deep water
carbonate saturation state. Earth and Planetary Science Letters 258 (1-2), 73–86.
Yu, J. M., Elderfield, H., Hönisch, B., 2007. B/Ca in planktonic foraminifera as a proxy for
surface seawater pH. Paleoceanography 22 (2), PA2202.
Yu, J. M., Elderfield, H., Piotrowski, A. M., 2008. Seawater carbonate ion-δ 13 C systematics
and application to glacial-interglacial North Atlantic ocean circulation. Earth and
Planetary Science Letters 271 (1-4), 209–220.
Yu, J. M., Foster, G. L., Elderfield, H., Broecker, W. S., Clark, E., 2010b. An evaluation of
benthic foraminiferal B/Ca and δ 11 B for deep ocean carbonate ion and pH
reconstructions. Earth and Planetary Science Letters 293 (1-2), 114–120.
Zachos, J., Pagani, M., Sloan, L., Thomas, E., Billups, K., 2001. Trends, rhythms, and
aberrations in global climate 65 Ma to present. Science 292 (5517), 686–693.
Zachos, J. C., Dickens, G. R., Zeebe, R. E., 2008. An early Cenozoic perspective on
greenhouse warming and carbon-cycle dynamics. Nature 451 (7176), 279–283.
Zeebe, R., Bijma, J., Wolf-Gladrow, D., 1999. A diffusion-reaction model of carbon isotope
fractionation in foraminifera. Marine Chemistry 64 (3), 199–227.
Zeebe, R., Westbroek, P., 2003. A simple model for the CaCO3 saturation state of the
ocean: The “Strangelove”, the “Neritan”, and the “Cretan” Ocean. Geochemistry
Geophysics Geosystems 4 (12).
Zeebe, R., Wolfe-Gladrow, D., 2001. CO2 in Seawater: Equilibrium, Kinetics, Isotopes.
Elsevier Oceanography Series.
Zeebe, R. E., 1999. An explanation of the effect of seawater carbonate concentration on
foraminiferal oxygen isotopes. Geochimica et Cosmochimica Acta 63 (13-14), 2001–2007.
Zeebe, R. E., Sanyal, A., 2002. Comparison of two potential strategies of planktonic
foraminifera for house building: Mg2+ or H+ removal? Geochimica et Cosmochimica
Acta 66 (7), 1159–1169.
Zeebe, R. E., Wolf-Gladrow, D. A., 2004. A matlab program for calculating the carbonate
system (csys3.m).
URL http://www.soest.hawaii.edu/oceanography/faculty/zeebe
Zuddas, P., Mucci, A., 1998. Kinetics of calcite precipitation from seawater: II. The
influence of the ionic strength. Geochimica et Cosmochimica Acta 62 (5), 757–766.
Zuddas, P., Pachana, K., Faivre, D., 2003. The influence of dissolved humic acids on the
kinetics of calcite precipitation from seawater solutions. Chemical Geology 201 (1-2),
91–101.

Similar documents

Detecting anthropogenic carbon dioxide uptake and ocean

Detecting anthropogenic carbon dioxide uptake and ocean dioxide (CO2 ) to the atmosphere, with the ocean absorbing approximately 30 % (Sabine et al., 2004). Ocean uptake and chemical equilibration of anthropogenic CO2 with seawater results in a gradual ...

More information