The Geological Record of Ocean Acidification
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
BÄRBEL HÖNISCH1*, ANDY RIDGWELL2, DANIELA N. SCHMIDT2, ELLEN
THOMAS3,4, SAMANTHA J. GIBBS5, APPY SLUIJS6, RICHARD ZEEBE7, LEE
KUMP8, ROWAN C. MARTINDALE9, SARAH E. GREENE9, WOLFGANG
KIESSLING10, JUSTIN RIES11, JAMES C. ZACHOS12, DANA L. ROYER4, STEPHEN
BARKER13, THOMAS M. MARCHITTO JR.14, RYAN MOYER15, CARLES
PELEJERO16, PATRIZIA ZIVERI17, GAVIN L. FOSTER5, BRANWEN WILLIAMS18
1
Lamont-Doherty Earth Observatory of Columbia University, Palisades, NY, USA, 2University of
Bristol, Bristol BS8 1SS, UK, 3Yale University, New Haven, CT 06520, USA, 4Wesleyan University,
Middletown, CT 06459, USA, 5National Oceanography Centre, Southampton SO14 3ZH, UK,
6
Universiteit Utrecht, 3584 CD Utrecht, Netherlands, 7University of Hawaii at Manoa, Honolulu, HI
96822, USA, 8Pennsylvania State University, University Park, PA 16802, 9University of Southern
California, Los Angeles, CA 90089, USA, 10Museum für Naturkunde at Humboldt University, 10115
Berlin, Germany, 11University of North Carolina - Chapel Hill, NC 27599, USA, 12University of
California Santa Cruz, CA 95064, USA, 13Cardiff University, CF10 3XQ, UK, 14University of
Colorado, Boulder, CO 80309, USA, 15US Geological Survey, St. Petersburg, FL 33701, USA, 16ICREA
and Institut de Ciències del Mar, CSIC, 08003 Barcelona, Catalonia, Spain, 17Universitat Autònoma de
Barcelona, 01893 Barcelona, Spain, 18The Claremont Colleges, Claremont CA 91711, USA,
*Corresponding author email: hoenisch@ldeo.columbia.edu
21
22
The future consequences of ocean acidification for marine ecosystems are currently difficult to
23
assess, in part because laboratory experiments are hindered by their necessary short time-scale
24
and reduced ecologic complexity. In contrast, the geological record is replete not only with a
25
variety of global environmental perturbations that may include ocean acidification, but also
26
associated biotic responses including adaptation and evolution. Our review of the past ~250 My
27
of Earth history highlights a number of events exhibiting evidence for elevated pCO2, global
28
warming and ocean acidification, some with contemporaneous extinction amongst marine
29
calcifiers. Although valuable insight can be gained, no event perfectly parallels the future in
30
terms of the balance of carbonate chemistry changes – a consequence of the unprecedented
31
rapidity of CO2 release currently taking place.
32
What constitutes ‘ocean acidification’?
1
33
The geological record is imprinted with the biotic responses to natural perturbations in global carbon
34
cycling and climate change (Figure 1). By reconstructing past changes in marine environmental
35
conditions we can test hypotheses for cause and effect of future-relevant pressures such as ocean
36
acidification on ecosystems (1). However, the current rate of anthropogenic CO2 release leads to a
37
surface ocean environment characterized not only by elevated dissolved CO2 and decreased pH, but
38
critically also with decreased saturation with respect to a common shell mineral – calcium carbonate
39
(CaCO3). In comparing past and future carbon cycle perturbations, it therefore has to be realized that
40
slower rates of CO2 release lead to a different balance of carbonate system changes and may induce
41
differential biotic response or even no response at all, which hence invalidates a direct analogue being
42
drawn. For instance, the Cretaceous, a time characterized by massive chalk deposits (mainly in the
43
form of nannofossil calcite) but under generally high atmospheric pCO2 (Fig. 2B), is commonly
44
misconceived as evidence that marine calcification will not be impaired under high future pCO2 [e.g.,
45
(2)]. Such arguments are flawed, as extended conditions of high pCO2 and low pH (Fig. 2B, D) do not
46
necessarily result in suppressed seawater saturation state (1, 3), which exerts an important control on
47
organisms’ calcification (REF). The reason for a smaller saturation response to slow CO2 release is that
48
the alkalinity (cations) released by rock weathering on land must ultimately be balanced by the
49
preservation and burial of CaCO3 in marine sediments, which itself is controlled by the carbonate
50
saturation state of the ocean [Fig. 3 and (4)]. Although weathering (and hence saturation) is also related
51
to atmospheric pCO2 (5), it is related much more weakly than ocean pH, which allows these two
52
carbonate variables to be almost completely decoupled for slowly increasing atmospheric pCO2.
53
We illustrate the progressive coupling between saturation and pH as the rate of atmospheric pCO2
54
change increases and sources (weathering) and sinks (CaCO3 burial) of alkalinity can no longer be
55
balanced, with the aid of a global carbon cycle model (6). For rapid (century-scale) and thus future-
56
relevant increases in atmospheric pCO2, both surface ocean pH and saturation state decline in tandem
2
57
(Fig. 4). The projected decrease in ocean surface saturation state – here, with respect to aragonite
58
aragonite) – is an order of magnitude larger for a rapid CO2 increase than for a slow (100-kyr) CO2
59
increase. Ultimately, saturation recovers while the pH remains suppressed, reflecting how changes in
60
the oceanic concentrations of dissolved inorganic carbon (DIC) and alkalinity make it possible to have
61
simultaneously both higher CO2 and carbonate ion concentration saturation ([ CO32- ], which controls
62
saturation), but with the relatively greater increase in [CO2] causing lower pH. The key to unlocking
63
the geological record of ocean acidification is hence to distinguish between long-term steady states and
64
transient changes. We use the term ‘ocean acidification event’ for time intervals in Earth’s history that
65
involve both a reduction in ocean pH and a substantial lowering of CaCO3 saturation, implying a time-
66
scale on the order of 10,000 years and shorter (Fig. 4).
67
For the fossil record to be of direct utility in assessing future ecosystem impacts, the occurrence and
68
extent of past ocean acidification must be unambiguously identified. In recent years, a variety of trace-
69
element and isotopic tools have become available that can be applied to infer past seawater carbonate
70
chemistry. For instance, the boron isotopic composition (δ11B) of marine carbonates reflects changes in
71
seawater pH, the trace element (e.g., B, U, and Zn) to calcium ratio of benthic and planktic foraminifer
72
shells records ambient [ CO32- ], and the stable carbon isotopic composition (δ13C) of organic molecules
73
(alkenones) can be used to estimate surface ocean aqueous [ CO 2 ] (6). Because such geochemical data
74
are still relatively scarce, possible past ocean acidification is often inferred from a decrease in the
75
accumulation and preservation of CaCO3 in marine sediments, potentially indicated by an increased
76
degree of fragmentation of foraminiferal shells (7). However, a difficulty here is in distinguishing
77
between the original calcification responses to chemical changes in the surface ocean and post-mortem
78
conditions at the seafloor. For instance, planktic calcifiers may secrete heavier or lighter shells (8) but
79
that signal may be further modified at the seafloor through dissolution or overgrowth after deposition
80
(9, 10). This duality can introduce controversy over cause-and-effect, the drivers of biological change,
3
81
and whether past intervals of ocean acidification are characterized by environmental conditions
82
relevant for the near future.
83
Are there indications of ocean acidification in the geological record?
84
With these criteria in mind, we review (in reverse chronological order) the intervals in Earth history for
85
which ocean acidification has been hypothesized, along with the evidence for independent geochemical
86
and biotic changes (summarized in Table S1). We confine this review to the past ~250 My because the
87
earlier Phanerozoic (and beyond) lacks the pelagic calcifiers that not only provide key proxy
88
information but also create the strong deep-sea carbonate (and hence atmospheric pCO2) buffer that
89
characterizes the modern Earth system (4). Our criteria for identifying potentially future-relevant past
90
ocean acidification are: (1) massive CO2-release; (2) a time-scale of CO2 release of less than 10-20 ky;
91
(3) ocean acidification (pH and/or saturation decline); and (4) evidence for global warming. Events are
92
given a similarity index that is based on available geochemical data (Table S1, Fig. 2).
93
Late Pleistocene deglacial transitions. The last deglaciation is the best documented past event
94
associated with a large (76 µatm) CO2 rise occurring between 17.8 to 11.6 ky B.P. (11). Boron isotope
95
estimates from planktic foraminifers show a 0.15 0.05 unit decrease in sea surface pH (12) across the
96
deglacial transition – an average rate of decline of ~0.002 units per 100 years compared to the current
97
rate of more than 0.1 units per 100 years (Table S1). Planktic foraminiferal shell weights decreased by
98
40-50% (8), and coccolith mass decreased by ~25% (13). In the deep ocean, changes in carbonate
99
preservation [e.g., (14)], pH [from foraminiferal 11B, (15)] and [CO32-] [from foraminiferal B/Ca and
100
Zn/Ca, (16, 17)] differed between ocean basins, reflecting co-varying changes in deep-water circulation
101
and an internal carbon shift within the ocean, and between the ocean and atmosphere (and terrestrial
102
biosphere) rather than an input from external, geological reservoirs. The regional nature of these
4
103
variations highlights the general need for careful evaluation of regional versus global effects in paleo-
104
studies.
105
Oligocene-Pliocene. The climate of the Oligocene to Pliocene (34-2.6 Ma) contains intervals of
106
elevated temperature and modest deviations of atmospheric pCO2 from modern (Fig. 2). Of particular
107
interest has been the Pliocene warm period [3.29 to 2.97 Ma, (18, 19)], which is characterized by global
108
surface temperatures estimated to be ~2.5°C higher than today (19), atmospheric pCO2 between 330-
109
400 µatm [Fig. 2C, (18, 20)], and sea surface pH(T) ~0.06-0.11 units lower (18) than the preindustrial.
110
Ecological responses to the warming include migration of tropical foraminifer species towards the
111
poles (21) but there are no documented calcification responses or increased nannoplankton extinction
112
rates (22). The early to middle Miocene (23-11 Ma) and Oligocene (34-23 Ma) were also characterized
113
by elevated temperatures and slightly higher pCO2 compared to preindustrial values (Fig. 2C), but, as
114
we discussed previously, such long intervals of quasi steady-state conditions cannot serve as future-
115
relevant ocean chemical analogues.
116
Paleocene/Eocene. Evidence for rapid carbon injection associated with the Paleocene-Eocene Thermal
117
Maximum (PETM, 56 Ma) as well as a number of smaller transient global warming events
118
(hyperthermals) during the late Paleocene and early Eocene (58 to 51 Ma) comes primarily from
119
observations of large (up to -4‰) negative 13C excursions (23) associated with pronounced decreases
120
in carbonate preservation (24). Depending on the assumed source, rate and magnitude of CO2 release
121
[e.g., (25)], a 0.25 to 0.45 unit decline in surface seawater pH is possible (1) with a reduction in mean
122
surface ocean aragonite saturation from Ω=3 to 1.5-2. The carbonate compensation depth (CCD, see:
123
Box) was shallower than 1.5 km in places (24) (compared to >4km today), although as with the last
124
glacial transition, deep sea geochemistry was likely regionally modulated by ocean circulation changes
125
(26). Although a pH decrease or pCO2 increase remains to be confirmed by geochemical proxies for
126
any of the hyperthermal events, the amount of carbon injected can be modeled based on consistent
5
127
carbonate 13C and CCD changes, yielding between about 2000 and 6000 PgC for the onset of the
128
PETM (27, 28).
129
PETM sediments record the largest extinction amongst deep-sea benthic foraminifers of the last 75
130
My (29). A major change in trace fossils also indicates a disruption of the macrobenthic community
131
(30). However, the co-variation of ocean acidification, warming and corresponding oxygen depletion
132
[Fig. S1 and e.g., (23)] makes the attribution of a single cause of this extinction difficult (1, 29). In
133
shallow water environments, a gradual shift from calcareous red algae and corals to larger benthic
134
foraminifers as dominant calcifiers started in the Paleocene and was completed at the PETM with the
135
collapse of coralgal reefs and larger benthic foraminiferal turnover (31). This event is recognized as
136
one of the four major metazoan reef crises of the past 250 My (32). In marginal marine settings,
137
coccolithophore (33) and dinoflagellate cyst (34) assemblages display changes in species composition
138
but these are interpreted to reflect sensitivity to temperature, salinity stratification and/or nutrient
139
availability (34, 35), not acidification (Fig. S1). In the open ocean, the occurrence of deformities in
140
some species of calcareous nannoplankton has been described (36), but despite a strong change in
141
assemblage, there is no bias in extinction or diversification in favor or against less or more calcified
142
planktic species (37).
143
End Cretaceous. The well-known mass extinction at 65 Ma was very likely caused by a large asteroid
144
impact (38). In addition to potential terrestrial biomass or fossil carbon burning, the impact may have
145
caused the emission of SO2 from vaporized gypsum deposits at the impact site and/or nitric acid
146
aerosols produced by shock heating of the atmosphere, which could have led to acid rain and hence
147
potentially to acidification of the surface ocean (38). Planktic calcifiers exhibit elevated rates of
148
extinction and reduced production (22, 39), however, the event was not associated with major
149
extinction of reef corals (32) (Fig. 2B), nor were benthic foraminifers affected in either shallow or deep
150
waters (29). Because multiple environmental changes co-varied and proxy data for marine carbonate
6
151
chemistry are not yet available, unambiguous attribution of the extinctions to ocean acidification is
152
currently not possible.
153
Cretaceous and Jurassic Oceanic Anoxic Events (OAEs). The Mesozoic OAEs (in particular OAE 2
154
~93 Ma, OAE1a ~120 Ma, and Toarcian OAE ~183 Ma) were intervals during which the ocean’s
155
oxygen minimum and deep anoxic zones expanded markedly (40). The onsets of these OAEs have been
156
linked to the emplacement of large igneous provinces, degassing large amounts of CO2, and associated
157
environmental consequences of warming, lower oxygen solubility, and possibly ocean acidification
158
(40). Some of the Cretaceous OAEs are associated with turnover in the plankton communities (41).
159
Deformities and some minor size reduction in coccoliths as well as a massive increase in the abundance
160
of the heavily calcified nannoconids have been observed (42, 43). However, similar to more recent
161
events, there is difficulty in unequivocally attributing observations to surface water acidification given
162
the co-variation of environmental changes (44).
163
Because most old seafloor (~180 Ma or older) is subducted, the sedimentary record of the Toarcian
164
OAE is now restricted to former continental margins. Sedimentary organic and inorganic carbon
165
deposits display initially negative followed by positive 13C excursions, consistent with an influx of
166
CO2 into the atmosphere followed by organic carbon burial (40). The negative isotopic transition
167
occurs in distinct negative 13C shifts, each estimated to occur in less than 20 ky (45) and possibly in as
168
little as 650 years (46). The Toarcian OAE is associated with a reef crisis that was particularly selective
169
against corals and hypercalcifying sponges (i.e. animals with a large skeletal to organic biomass ratio)
170
(32) (Fig. 2B) and a decrease in nannoplankton flux (47). Again, these observations could have been a
171
response to any one or combination of a number of different contemporaneous environmental changes
172
Triassic-Jurassic (T/J). The Triassic-Jurassic mass extinction has been linked to the coeval
173
emplacement of the Central Atlantic Magmatic Province (48). Across the T/J boundary (~200 Ma),
174
proxy records suggest a doubling of atmospheric pCO2 over as little as 20 ky (49, 50), although the
7
175
absolute pCO2 estimates differ greatly between proxies, with leaf stomata giving 700-1500 µatm, while
176
pedogenic carbonates suggest 2000-4400 µatm (6) (Fig. 2C). Decreased carbonate saturation has been
177
inferred from reduced pelagic carbonate accumulation in shelf sediments (51), although it has to be
178
appreciated that shallow water carbonate deposition can vary in response to many parameters, not only
179
acidification. A calcification crisis amongst hypercalcifying taxa has been inferred for this period (Fig.
180
2B), with reefs and scleractinian corals experiencing a near-total collapse (32). However, the
181
observation that tropical species were more affected than extra-tropical species suggests that global
182
warming may have been an important contributor or even dominant cause of the extinction (32).
183
Permian/Triassic (P/T). The Permo-Triassic mass extinction (251.4 Ma) was the most severe of the
184
Phanerozoic Era and coincided, at least in part, with one of the largest known continental eruptions, the
185
Siberian trap basalts. Recent estimates for the total CO2 release put it at ~13,000-43,000 PgC/400 ky
186
(52, 53), which puts the annual carbon release to ~0.03-0.11PgC [compared to 9.9 PgC yr-1 in 2008,
187
(54)]. There is some observational evidence for carbonate dissolution in shelf settings (52) but its
188
interpretation is again debated (55). There is abundant evidence for ocean anoxia, photic zone euxinia
189
(i.e. enrichment in hydrogen sulfide) [e.g., (56)] and strong warming (52) but no direct proxy evidence
190
for pH or carbonate ion changes. Knoll et al. (56) inferred the preferential survival of taxa with
191
anatomical and physiological features that should confer resilience to reduced carbonate saturation state
192
and hypercapnia (i.e. high CO2 in blood) and preferential extinction of taxa that lacked these traits, e.g.
193
reef builders (32).
194
Is there a geologic analogue for future ocean acidification?
195
Figure 2A and Table S1 summarize the degree to which past ocean carbon cycle perturbation events are
196
judged qualitatively similar to anthropogenic ocean acidification. Difficulties exist for pre-Pliocene
197
times where proxy evidence for lower pH and reduced saturation is currently lacking. Global carbon
8
198
cycle models can help to infer the magnitude of carbon release by fitting observed changes in the 13C
199
of calcium carbonates and organic remnants [e.g., (57)]. However, as well as needing information on
200
the source and isotopic composition of the added carbon, the time-scale of 13C change is critically
201
important to the estimation of CO2 fluxes (25). Due to the lack of open-ocean sediments and
202
increasingly poor temporal and spatial resolution of the geological record further back in time, placing
203
the necessary constraints on the duration and rate of CO2 release is difficult. Radiometric dating
204
techniques are not accurate enough to identify Mesozoic intervals of 10-kyr duration, although orbital
205
spectral analysis of highly resolved isotope records can help to partly overcome this, e.g., if a 13C
206
excursion is shorter or longer than one precession cycle [i.e. 21 ky, (49)]. Even for the well-studied
207
PETM, the duration of the main phase of this carbon injection is still debated (35, 58) and model
208
inferred peak rates of 1 PgC per year (27, 58) could potentially be an underestimate.
209
Additional complications arise because carbon may not have been released at a uniform rate, and in
210
the extreme, may have occurred solely in the form of rapid pulses of CO2-release. In such cases,
211
assuming an average emissions rate throughout the entire duration of the pulsed release will fail to
212
capture the potential for episodes of intense acidification. For instance, although the total duration of
213
the CO2 release from the Triassic/Jurassic-age Central Atlantic Magmatic Province was estimated to be
214
~600 ky, pulses as short as ~20 ky have been suggested (49, 59). Similarly, the main phase of OAE1a
215
(excluding the recovery interval) was ~150 ky (43) and hence too slow for carbonate saturation to be
216
significantly affected (Fig. 4), but major volcanic eruptions and thus rapid CO2 release could
217
potentially have produced future-relevant perturbations in the carbon cycle. Substantially improved
218
chronologies and high-resolution records are needed to refine estimates of rate.
219
Given current knowledge of the last 250 My of Earth history (Table S1, Fig. 2), the PETM and
220
associated hyperthermal events, the T/J, and potentially the P/T, all stand out as having excellent
221
potential as analogue events, although the T/J and P/T will always be much more poorly constrained
9
222
due to the absence of deep-sea carbonate deposits. OAEs may also hold some relevance but suffer from
223
being less severe in continental volcanism (CO2 release) than the older events (P/T) and less well
224
characterized than the younger (e.g., PETM) events. The last deglacial transition, although
225
characterized by temperature and CO2-increase, is two orders of magnitude slower than current
226
anthropogenic change. It is also thought to largely represent a reorganization of carbon within the
227
ocean and to the atmosphere and terrestrial biosphere, and hence will not have the same globally-
228
uniform impact that an input from geological reserves has.
229
Perspectives for utilizing the geological record to help project future global change
230
Only rapid or pulsed CO2 release events can provide direct future relevant information. Assessment of
231
such events critically depends on independent geochemical quantification of the associated changes in
232
the carbonate system, specifically seawater-pH and carbonate saturation. Geochemical proxy estimates
233
are currently absent for the Cretaceous and beyond, and these measurements need to be filled in to
234
verify whether ocean acidification did indeed happen at those times. This is challenging, because in
235
addition to the potential for increasing diagenetic alteration and reduced stratigraphic exposure,
236
uncertainty over the chemical and isotopic composition of seawater increases and limits our
237
interpretation of these proxies [e.g., (60, 61)]. Future studies will have to improve and expand
238
geochemical estimates and their uncertainties of surface and deep-ocean carbonate chemistry changes
239
associated with carbonate dissolution and ecological changes. This includes finding new archives to
240
study the secular evolution of seawater chemistry, but also the study of living proxy carriers under
241
mimicked conditions of past seawater chemistry in laboratory culture. An unfortunate aspect of the
242
geological record, however, is the lack of deep-sea carbonates in the early Jurassic and beyond, which
243
further reduces our ability to characterize the carbonate chemistry of those older events.
10
244
The sensitivity of ocean chemistry to CO2 release, and the relationship between induced pH and
245
pCO2 changes, vary through time and further complicate the picture. For instance, seawater calcium
246
and magnesium ion concentrations were different in the past (Fig. 2C), altering the ocean’s carbonate
247
ion buffering capacity and hence sensitivity of the Earth system to carbon perturbation (62). Varying
248
seawater Mg/Ca ratios also affect the mineralogy of marine calcifiers, where the more soluble high-Mg
249
calcite predominates in the Neogene ocean and during the Permian through mid Jurassic, and more
250
resistant low-Mg calcite predominated during the Late Jurassic through Paleogene (63). On this
251
mineralogical basis, the vulnerability of marine calcifiers to ocean acidification and seawater
252
geochemistry during the P/T and T/J was arguably closer to the modern ocean and greater than e.g., the
253
PETM.
254
Although we have concentrated on the prospects for extracting information from the geological
255
record concerning the impact of ocean acidification, it is really necessary to separate effects? (64). The
256
enhanced greenhouse effect associated with elevated atmospheric CO2 leads to warming in the surface
257
ocean (65) and a decrease in dissolved oxygen concentration (66). Because the consequences of
258
massive carbon release, whether future or past will share the same combination and sign of
259
environmental changes, the strength of the geological record lies in revealing past coupled warming
260
and ocean acidification (and deoxygenation) events as an ‘integrated’ analogue. In this context,
261
understanding and identifying past ocean acidification is important, as the coupled decline in pH and
262
saturation is an integral and characteristic part of the net environmental response to CO2 release. It is
263
also in this context that the current rate of (mainly fossil fuel) CO2 release stands out as potentially
264
unparalleled in at least the last 250 My of Earth history, suggesting that we are entering an unknown
265
territory of ecosystem change with the future magnitude of decline in carbonate saturation potentially
266
comparable to or exceeding that associated with some of the great mass extinctions in the ocean.
267
11
268
References
269
1.
A. Ridgwell, D. N. Schmidt, Nature Geosci 3, 196 (2010).
270
2.
J. B. Ries, in Science Now. (2009).
271
3.
R. E. Zeebe, P. Westbroek, Geochem. Geophys. Geosyst. 4, (Dec, 2003).
272
4.
A. Ridgwell, R. E. Zeebe, Earth and Planetary Science Letters 234, 299 (2005/6/15, 2005).
273
5.
D. Archer et al., Geophysical Research Letters 24, 405 (Feb, 1997).
274
6.
SOM, Methods are available on Science online.
275
7.
W. H. Berger, Deep-Sea Research 15, 31 (1968).
276
8.
S. Barker, H. Elderfield, Science 297, 833 (2002).
277
9.
S. Barker et al., Paleoceanography 19, doi:10.1029/2004PA001004 (2004).
278
10.
S. J. Gibbs et al., Earth and Planetary Science Letters 295, 583 (2010).
279
11.
E. Monnin et al., Science 291, 112 (January 5, 2001, 2001).
280
12.
B. Hönisch, N. G. Hemming, Earth and Planetary Science Letters 236, 305 (2005/7/30, 2005).
281
13.
L. Beaufort et al., Nature 476, 80 (2011).
282
14.
J. W. Farrell, W. Prell, Paleoceanography 4, 447 (1989).
283
15.
B. Hönisch et al., Earth and Planetary Science Letters 272, 309 (Jul, 2008).
284
16.
J. Yu et al., Science 330, 1084 (November 19, 2010, 2010).
285
17.
T. M. Marchitto et al., Earth and Planetary Science Letters 231, 317 (2005/3/15, 2005).
286
18.
O. Seki et al., Earth and Planetary Science Letters 292, 201 (2010).
287
19.
A. M. Haywood et al., Glob. Planet. Change 66, 208 (Apr, 2009).
288
20.
M. Pagani et al., Nature Geosci 3, 27 (2010).
289
21.
H. J. Dowsett, M. M. Robinson, Micropaleontology 53, 105 (2007).
290
291
22.
P. R. Bown et al., in Coccolithophores – From molecular processes to global impacts, H. R.
Thierstein, J. R. Young, Eds. (Springer, Berlin, 2004), pp. 481-508.
292
23.
J. P. Kennett, L. D. Stott, Nature 353, 225 (Sep, 1991).
12
293
24.
J. C. Zachos et al., Science 308, 1611 (June 10, 2005, 2005).
294
25.
J. C. Zachos et al., Earth and Planetary Science Letters 299, 242 (Oct, 2010).
295
26.
R. E. Zeebe, J. C. Zachos, Paleoceanography 22, (Jul, 2007).
296
27.
R. E. Zeebe et al., Nat. Geosci. 2, 576 (Aug, 2009).
297
28.
K. Panchuk et al., Geology 36, 315 (April 01, 2008, 2008).
298
299
29.
E. Thomas, in Geological Society of America Special Paper, S. Monechi, R. Coccioni, M. R.
Rampino, Eds. (Geological Society of America, 2007), pp. 1-23.
300
30.
F. J. Rodríguez-Tovar et al., Marine Geology 282, 178 (Apr, 2011).
301
31.
C. Scheibner, R. P. Speijer, Earth-Sci. Rev. 90, 71 (Nov, 2008).
302
32.
W. Kiessling, C. Simpson, Glob. Change Biol. 17, 56 (Jan, 2011).
303
33.
S. J. Gibbs et al., Geology 34, 233 (Apr, 2006).
304
34.
A. Sluijs, H. Brinkhuis, Biogeosciences 6, 1755 (2009).
305
35.
A. Sluijs et al., Nature Geosci 2, 777 (2009).
306
36.
I. Raffi, B. De Bernardi, Marine Micropaleontology 69, 119 (2008).
307
37.
S. J. Gibbs et al., Science 314, 1770 (2006).
308
38.
P. Schulte et al., Science 327, 1214 (Mar, 2010).
309
39.
S. D'Hondt et al., Geology 22, 983 (November, 1994, 1994).
310
40.
H. C. Jenkyns, Geochem. Geophys. Geosyst. 11, (Mar, 2010).
311
41.
R. M. Leckie et al., Paleoceanography 17, (Jul-Aug, 2002).
312
42.
E. Erba, F. Tremolada, Paleoceanography 19, (Feb, 2004).
313
43.
E. Erba et al., Science 329, 428 (Jul, 2010).
314
44.
S. J. Gibbs et al., Science 332, 175 (April 8, 2011, 2011).
315
45.
G. Suan et al., Earth and Planetary Science Letters 267, 666 (2008).
316
46.
A. S. Cohen et al., J. Geol. Soc. 164, 1093 (2007).
317
47.
E. Mattioli et al., Global and Planetary Change 65, 134 (2009).
318
48.
J. H. Whiteside et al., Paleogeogr. Paleoclimatol. Paleoecol. 244, 345 (Feb, 2007).
319
49.
D. B. Kemp et al., Nature 437, 396 (Sep, 2005).
13
320
50.
M. Ruhl et al., Science 333, 430 (July 22, 2011, 2011).
321
51.
A. E. Crne et al., Geological Society of America Bulletin 123, 40 (Jan, 2011).
322
52.
J. L. Payne et al., Proceedings of the National Academy of Sciences, (April 26, 2010, 2010).
323
53.
S. V. Sobolev et al., Nature 477, 312 (2011).
324
54.
C. Le Quéré et al., Nature Geosci 2, 831 (2009).
325
55.
P. B. Wignall et al., GSA Bulletin 121, 954 (2009).
326
56.
A. H. Knoll et al., Earth and Planetary Science Letters 256, 295 (Apr, 2007).
327
57.
J. Zachos et al., Science 292, 686 (2001).
328
58.
Y. Cui et al., Nature Geosci 4, 481 (2011).
329
59.
M. Ruhl et al., Earth and Planetary Science Letters 295, 262 (2010).
330
60.
D. Lemarchand et al., Nature 408, 951 (2000).
331
61.
R. M. Coggon et al., Science 327, 1114 (February 26, 2010, 2010).
332
62.
M. F. Stuecker, R. E. Zeebe, Geophysical Research Letters 37, (Feb, 2010).
333
63.
S. M. Stanley, L. A. Hardie, Paleogeogr. Paleoclimatol. Paleoecol. 144, 3 (Nov, 1998).
334
64.
C. Turley et al., Marine Pollution Bulletin 60, 787 (2010).
335
65.
N. A. Rayner et al., J. Clim. 19, 446 (Feb, 2006).
336
66.
R. F. Keeling et al., Annual Review of Marine Science 2, 199 (2010).
337
67.
J. A. D. Dickson, Science 298, 1222 (Nov, 2002).
338
68.
R. V. Demicco et al., Geology 33, 877 (November 01, 2005, 2005).
339
69.
A. Ridgwell, Marine Geology 217, 339 (2005).
340
341
70.
E. T. Sundquist, K. v. Visser, : Elsevier, Chapt. 9, in Treatise on Geochemistry:
Biogeochemistry, W. H. Schlesinger, Ed. (Elsevier, 2004).
342
343
344
345
346
347
348
349
Funding for the “Workshop on Paleocean Acidification and Carbon Cycle Perturbation Events”
was provided by NSF OCE 10-32374 and PAGES. We thank the workshop participants for
stimulating discussions and contributions to this manuscript, and the USC Wrigley Institute on
Catalina Island for hosting the workshop. Particular thanks are owed to Thorsten Kiefer of
PAGES, who initiated the workshop and supported it at all stages. This work is a contribution to
the "European Project on Ocean Acidification" (EPOCA).
14
Figure 1. Extinction and radiation amongst the dominant calcareous and organic fossil
lineages suggest events of major environmental change throughout the past 300 million
years. The line thickness indicates relative, period-averaged species diversity. Note that
Cretaceous dinocyst diversity is studied in greater detail than other periods, possibly
biasing evolutionary trends. Highlighted events (vertical red lines) have been associated
with potential ocean acidification events. However, note that calcareous organisms were
not uniformly affected at all times, suggesting the importance of synergistic
environmental factors to extinction, adaptation and evolution. Identification of a
paleocean acidification event therefore requires independent geochemical evidence for
ocean chemistry changes. The thickness of the lines indicates species richness, where
calcareous plankton is shown in black, calcareous benthos in blue, and organic fossils in
green. Images of organisms are exemplary, see (6) for references and further information
on the displayed organisms.
Figure 2. Compilation of data-based (panels B+C) and model-reconstructed (D+E)
indicators of global carbon cycle evolution over the past 300 My together with candidate
ocean acidification events (panel A). Panel A summarizes events (Table S1) that have
some degree of similarity to modern ocean acidification. The similarity index (see text
and Table S1 for details) is color-coded, where red == 3/most similar, orange == 2/partly
similar, yellow == 1/unlike. (B) Proxy-reconstructed atmospheric pCO2 (6) grouped by
proxy: yellow = paleosol 13C, cyan = marine phytoplankton 13C, red = stomatal
indices/ratios, blue = planktic foraminiferal 11B, magenta = liverwort 13C, black =
sodium carbonates, with 10 My averages shown by grey bars. For plotting convenience,
estimates exceeding 2500 µatm are not shown [primarily paleosol 13C from the
uppermost Triassic, (6)]. (C) Ocean Mg/Ca ratios (left axis), reconstructed from fluid
inclusions [black circles, (6)] and echinoderm fossil carbonate [red circles, (67)] together
with the Phanerozoic seawater model of Demicco et al. (68) (dashed line). Also shown
15
(right axis) is [Ca2+] from fluid inclusions [grey triangles, (6)] and models [solid line,
(68)]. (D) Model-reconstructed changes in mean ocean surface pH at 20 My intervals
[black line, (69)].
Figure 3. When CO2 dissolves in seawater, it reacts with water to form carbonic acid,
which then dissociates to bicarbonate, carbonate and hydrogen ions. The higher
concentration of hydrogen ions makes seawater acidic but this process is buffered on long
time scales by the interplay of seawater, seafloor carbonate sediments and weathering on
land. Shown are the major pathways of reduced carbon (black) and of alkalinity (e.g.,
Ca2+) (yellow). Processes leading to ocean acidification and/or reduction of [ CO32-] are
indicated in red and processes leading to ocean alkalinisation and/or [ CO32-]-increases are
indicated in blue. Anthropogenic perturbations are marked in italics. Approximate fluxes
are printed in parentheses (PgC yr-1), while reservoir inventory values are shown in
brackets [PgC]. Natural carbon cycle fluxes from Sundquist and Visser (70);
anthropogenic fluxes for 2008 from Le Quéré et al. (54), which for the land sink is
significantly above its 1990-2000 average of 2.6 PgC yr-1 due to the 2008 La Niña state.
See Box for a more detailed explanation of carbonate equilibria in seawater.
Figure 4. Mean ocean surface pH and aragonite saturation (aragonite) in an Earth system
model (6) are progressively decoupled as the rate of atmospheric pCO2 change increases.
The four panels show: (A) prescribed linear increases of atmospheric pCO2 (on a log10
scale) from ×1 to ×2 pre-industrial CO2, (B) evolution of mean surface pH, (C) evolution
of mean surface aragonite, and (D) a cross-plot illustrating how aragonite is progressively
decoupled from pH as the rate of pCO2 increase slows, color-coded from red (= 'fast') to
blue (= 'slow'). Carbonate dissolution and terrestrial weathering on long time scales add
alkalinity to the ocean, which attentuates the decrease in aragonite saturation. These
model results include climate feedback but not very long-term (silicate) weathering
feedback.
16
F
igure
1
17
Figure 2
18
Figure 3
19
Figure 4
20