Atmospheric CO2 decline during the Pliocene intensification of Northern Hemisphere glaciations



[1] Several hypotheses have been put forward to explain the onset of intensive glaciations on Greenland, Scandinavia, and North America during the Pliocene epoch between 3.6 and 2.7 million years ago (Ma). A decrease in atmospheric CO2 may have played a role during the onset of glaciations, but other tectonic and oceanic events occurring at the same time may have played a part as well. Here we present detailed atmospheric CO2 estimates from boron isotopes in planktic foraminifer shells spanning 4.6–2.0 Ma. Maximal Pliocene atmospheric CO2 estimates gradually declined from values around 410 μatm to early Pleistocene values of 300 μatm at 2.0 Ma. After the onset of large-scale ice sheets in the Northern Hemisphere, maximal pCO2 estimates were still at 2.5 Ma +90 μatm higher than values characteristic of the early Pleistocene interglacials. By contrast, Pliocene minimal atmospheric CO2 gradually decreased from 310 to 245 μatm at 3.2 Ma, coinciding with the start of transient glaciations on Greenland. Values characteristic of early Pleistocene glacial atmospheric CO2 of 200 μatm were abruptly reached after 2.7 Ma during the late Pliocene transition. This trend is consistent with the suggestion that ocean stratification and iron fertilization increased after 2.7 Ma in the North Pacific and Southern Ocean and may have led to increased glacial CO2 storage in the oceanic abyss after 2.7 Ma onward.

1. Introduction

[2] To better understand and constrain future climate change in a high-CO2 world, it is essential to study potential CO2 analogs in Earth's history and the climate at that time. The Pliocene epoch (5.33–2.6 Ma) prior to the intensification of Northern Hemisphere glaciation at about 2.75 Ma is a likely candidate for such a high-CO2 analog. With a similar land-ocean configuration as today, Pliocene climate prior to 3.0 Ma was on average 3°C warmer than today with smaller terrestrial and sea ice extent [Haywood and Valdes, 2004, and references therein] and a partially deglaciated West Antarctic Ice Sheet [Naish et al., 2009]. Greenland and North Canada were forested, with possible small glacier outlets and terrestrial summer temperatures between 5 and 4 Ma were 14°C warmer than today [Ballantyne et al., 2006]. Accumulating evidence suggests that Pliocene CO2 levels prior to 3.0 Ma were higher than the preindustrial level of 280 μatm [Seki et al., 2010; Pagani et al., 2010] and possibly close to today's level of 392 μatm (P. Tans and R. Keeling, NOAA ESRL,

[3] To explain the onset of Northern Hemisphere glaciations between 3.6 and 3.0 Ma [Mudelsee and Raymo, 2005] and its intensification about 2.75 Ma onward [Haug et al., 2005], several mechanisms have been put forward. In a previous modeling study, a linear decrease in CO2 from high Pliocene concentrations to Pleistocene concentrations triggers large-scale ice sheets on the Northern Hemisphere after 2.9–2.7 Ma when obliquity was lowest (i.e., low summer insolation) [Li et al., 1998]. A recent modeling study compared the respective effects of the closure of Panamanian seaways, the end of a permanent El Niño state in the equatorial Pacific, and the tectonic uplift of the Rocky Mountains on the size of the Greenland ice sheet [Lunt et al., 2008, and references therein]. This comparison suggested that a decrease in atmospheric CO2 from a Pliocene level of 400 μatm to the preindustrial maximum of 280 μatm is most likely the culprit causing the onset of permanent Northern Hemisphere glaciations [Lunt et al., 2008]. A Northern Hemisphere glaciation threshold of 280 μatm CO2 has also been suggested by DeConto et al. [2008], and Seki et al. [2010] presented evidence for CO2 dropping from 330 to 400 μatm to ∼280 μatm between 3.2 and 2.8 Ma, thus confirming the coincidence between CO2 decrease and Northern Hemisphere glaciations.

[4] However, the reasons for the decrease in atmospheric CO2 have not yet been determined. Moreover, a decrease in atmospheric CO2 may not be the only prerequisite to the persistence of ice sheets on the Northern Hemisphere. Two studies recently proposed that the permanent El Niño state characterizing the early Pliocene equatorial Pacific would have prevented the growth of ice sheets at various locations on the Northern Hemisphere prior to 2.75–3.0 Ma [Huybers and Molnar, 2007; Vizcaíno et al., 2010]. Therefore it is necessary to determine the exact timing of CO2 changes, so that the tectonic and biogeochemical mechanisms of climate change and their causal relations can be better understood.

[5] In this study, we attempt to answer these questions by reconstructing changes in the partial pressure of carbon dioxide (pCO2) in surface seawater over the time period 4.6 to 2.0 Ma from planktic foraminiferal boron isotope ratios. Building on earlier work [Hemming and Hanson, 1992], boron isotope ratios recorded in planktic foraminifer shells have been applied successfully to reconstruct changes in surface seawater pH and aqueous pCO2 over the Pleistocene [Foster, 2008; Hönisch and Hemming, 2005; Hönisch et al., 2009]. Boron isotopes in marine carbonates are a function of seawater pH, since the relative abundance and isotopic composition of the two dominant dissolved boron species in seawater, i.e., boric acid and borate, change with pH. The charged borate is preferentially incorporated into marine carbonates so that their recorded boron isotopic composition (δ11B) increases with seawater pH [Hemming and Hanson, 1992].

[6] To understand how the uncertainties related to the reconstructions of paleoenvironmental proxies for the Pliocene epoch influence the reconstruction of pCO2, we also performed a sensitivity study of pCO2 estimates to the uncertainties on sea surface temperature (SST) related to uncertainties on the evolution of the seawater Mg/Ca ratio (Mg/Cas.w.), and the seawater boron isotope ratio (δ11Bs.w.) used to calculate pCO2 from boron isotope estimates.

2. Material and Methods

2.1. Site Location

[7] Estimating atmospheric CO2 from marine proxy records requires a location where aqueous and atmospheric CO2 are in close equilibrium. We have selected ODP site 999, located in the Columbian Basin (12°N, 78°W, 2839 m water depth) in the Caribbean Sea, which has previously been studied by Seki et al. [2010], albeit in lower resolution. Modern surface water is slightly oversaturated with respect to CO2 (±17 μatm) and Caribbean waters therefore act as a minor annual source of CO2 to the atmosphere [Takahashi et al., 2007]. A boron isotope study covering the last 140 ky suggests that glacial/interglacial atmospheric CO2 amplitudes can be captured from this site [Foster, 2008], suggesting that the strength of the CO2 source did not change significantly during Pleistocene glacial/interglacial cycles.

[8] Potential changes in sea-air equilibrium during the Pliocene may include upwelling of nutrient-rich and CO2-rich intermediate water at site 999, and enhanced surface water productivity. Some authors have suggested that Pacific waters rich in upwelled CO2 would have entered the Caribbean Sea prior to the final closure of the Panama seaway, elevating surface pCO2 at site 999 as compared to the atmosphere [Seki et al., 2010]. The inflow of Pacific water into the Caribbean can be monitored by comparison with site 1241 in the equatorial Pacific [Bartoli et al., 2005; Steph et al., 2006, 2010]. Seki et al. [2010] calculated pCO2 values at sites 1241 and 999 from alkenone ɛp estimates and found higher pCO2 on the Pacific side of the Isthmus as compared to the Caribbean side for the last 4.5 My. The authors implied that prior to the closure of Panama at 3.5 Ma upwelled CO2-rich surface waters entered the Caribbean Sea and rendered site 999 a minor source of CO2 to the atmosphere. However, the raw ɛp values and trends at sites 1241 and 999 are actually similar, and do not indicate any coincident change with the final closure of Panama at 3.5 Ma or later. There is therefore no direct proxy evidence that the Pacific surface inflow into the Caribbean increased Caribbean pCO2 prior to 3.5 Ma.

[9] In addition, the Pacific surface waters entering the Caribbean Sea through the Panamanian seaways, and monitored at site 1241, most likely derived from the North Equatorial Countercurrent, which originates in the oligotrophic West Pacific Warm Pool [Steph et al., 2010]. A study of changes in the thermocline depth at site 1241 based on Mg/Ca measured on the thermocline dwelling Globorotalia tumida shows no upwelling conditions between 5.5 and 2.0 Ma [Steph et al., 2010]. Upwelling did occur along the Peruvian coast, monitored at site 1243, and intensified after 3.9 Ma but never reached site 1241 [Steph et al., 2010]. The evolution of surface water stratification at site 999 is monitored by Mg/Ca-based SST from the surface dwelling Globigerinoides sacculifer and the thermocline dwelling Neogloboquadrina dutertrei [Steph et al., 2006, 2010]. They show a deepening of the thermocline following the closure of Panama between 4.8 and 4.0 Ma but constant surface water stratification conditions thereafter [Steph et al., 2006, 2010]. Similarly, the δ13C record of G. sacculifer at site 999 [Haug et al., 2001, unpublished data, 1999] does not show any significant changes between 5.0 and 2.0 Ma, leading us to believe that there were no major changes in surface productivity or nutrient supply that would characterize upwelling conditions. Finally, a study of Caribbean mean annual temperature range off the Panamanian coasts based on the size of bryozoan zooids suggests that coastal upwelling in the southern Caribbean Sea stopped after the closure of Panama between 4.25 and 3.45 Ma [O'Dea et al., 2007], although it is not sure whether these upwelled waters would have reached offshore site 999 prior to the closure. There is no absolute certainty that surface waters at site 999 were always in CO2 equilibrium with the atmosphere during the Pliocene. However, based on the evidence described above, we assume that surface waters were close to equilibrium with the atmosphere at least from 3.45 Ma onward, and that there is little evidence for upwelling between 3.45 and 4.6 Ma.

2.2. Boron Isotope Analysis

[10] Our aqueous pCO2 estimates are based on boron isotope ratios (11B/10B) in shells of the planktic foraminifer G. sacculifer. Hönisch and Hemming [2004] observed that shells >425 μm in diameter record surface seawater pH. Although several studies have observed a dissolution effect on smaller individuals [Hönisch and Hemming, 2004; Ni et al., 2007], the largest shell size classes of G. sacculifer are most resistant to dissolution, and their boron isotopic composition records surface seawater pH variations [Hönisch and Hemming, 2004; Ni et al., 2007]. Sediment samples were selected according to glacial and interglacial extremes in the benthic oxygen isotope stratigraphy of ODP 999A. Prior to washing the sediment, the first set of samples prepared at ETH Zürich was soaked for a brief period (no more than 2–3 h) in 3% H2O2 to help disaggregate the sediment (Data Set S1 in the auxiliary material). The second set of samples prepared at LDEO was freeze-dried (Data Set S1). Samples were then washed through a 63 μm sieve and then oven dried. Between 30 and 80 specimens with a total sample weight of 1–4 mg were analyzed.

[11] Shells were cracked between two glass slides, bleached overnight with 5% NaOCl to remove organic matter, and then rinsed 10 times with quartz distilled water under repeated ultrasonication and centrifugation steps. Mechanical cleaning by ultrasonication efficiently removes adherent clay particles [Martin and Lea, 2002; Deyhle and Kopf, 2004], which could be a source of B contamination. Samples were dissolved with 2N quartz-distilled HCl and aliquots containing ∼1 ng B were loaded onto outgassed zone refined Re filaments. 1 μl of boron-free seawater was added to each aliquot to facilitate ionization and discourage fractionation. Boron isotopes were then measured as BO2 ions on a Thermo Scientific TRITON thermal ionization mass spectrometer (TIMS) at the Lamont-Doherty Earth Observatory (LDEO). Analyses were done at a filament temperature of 980°C and a minimum of 3 acceptable analyses were collected of each sample solution to eliminate analytical artifacts such as excessive fractionation (>1‰ over the 20–30 min of data acquisition) and/or isobaric interference on mass 42 (>5,000 counts on mass 26) by organic matter contamination. δ11B was normalized to NIST 951 boric acid standard, similarly loaded with boron-free seawater and measured at 980°C. Three out of a total 195 analyses were excluded as analytical outliers because their isotope ratio deviated by >2 sd from the population mean of all acceptable analyses (Chauvenet's criterion) collected on the respective samples (Data Set S1 and Figure S1). Because each sample solution is homogenous, such a deviating analysis was rejected as an outlier. The analytical uncertainty reported on our analyses (Figure 1) is the 2σ standard error (2 standard errors = 2 sd/√N), which takes into account the number of acceptable analyses (N = 3–7) for each sample and the reproducibility of the NBS 951 boric acid standard (2 standard errors = 0.09‰) or, if higher, the 2 standard error external error of an in-house carbonate standard routinely analyzed at LDEO. The uncertainty of repeat data analysis was generally better than 0.34‰ (Figure 1).

Figure 1.

Planktic foraminiferal δ11B and benthic foraminiferal δ18O records. Repeat boron isotope analyses of individual sample solutions are indicated by dark blue dots, average values are shown in light blue, and uncertainties reflect internal or external 2 standard error reproducibility, whichever is larger. Grey crosses depict benthic foraminiferal δ18O at site 999 measured on C. wuellerstorfi [Haug and Tiedemann, 1998]. The black line shows the benthic foraminiferal δ18O stack LR04 [Lisiecki and Raymo, 2005], which reflects changes in deep ocean temperature and global ice volume. Glacial (blue) and interglacial (orange) MIS used for the descriptions of amplitudes are labeled according to the nomenclature by Lisiecki and Raymo [2005]. Note the right y axis change of scale from top to bottom.

[12] The δ11B data measured on the Thermo Scientific TRITON are lighter in comparison to data measured on the NBS design TIMS at SUNY Stony Brook [Hönisch and Hemming, 2004, 2005]. An interlaboratory comparison of modern and fossil G. sacculifer samples measured at LDEO on the TRITON and at SUNY Stony Brook on the NBS instrument, revealed a consistent offset of −1.1‰ (±0.19‰) on the TRITON that is not accounted for by standardization against NIST 951 [Hönisch et al., 2009]. The only difference other than the instrument is the use of boron free seawater at LDEO, which was added to both, samples and the boric acid standard. We do not know what causes this difference but we found that poor ionization at 980°C and excessive fractionation at temperatures >1020°C prohibit the analysis of foraminifer samples without boron free seawater on the TRITON. This instrument offset affects the absolute value but not the relative difference between glacial and interglacial samples, which allows for straightforward translation of boron isotope values into pH estimates [Hönisch et al., 2009].

2.3. Evaluation of Dissolution Bias and Reproducibility

[13] Although site 999 is characterized by excellent carbonate preservation after 4.6 Ma [Haug and Tiedemann, 1998], boron isotopes in planktic foraminifer shells are believed to be prone to partial shell dissolution [Hönisch and Hemming, 2004] and it is imperative to exclude potential dissolution bias on the measured δ11B. We therefore took Scanning Electron Microscopy images of individual shells randomly selected from the samples (Figure 2a and Data Set S1). SEM images do not show any dissolution artifact and show a pristine gametogenic calcite layer, which is typical for well-preserved specimens [Hönisch and Hemming, 2004]. We also compared measured δ11B with the shell weight of analyzed foraminifers (Figure 2b and Data Set S1). Dissolution would lower the foraminifers' shell weight and decrease the measured δ11B, although other factors may also modify the initial foraminifer shell weight such as the carbonate ion concentration at the time of shell formation. No systematic correlation between low shell weight and low measured δ11B was observed (Figure 2b). Two samples at 4.2 and 2.2 Ma yielded out-of-range δ11B values that could not be replicated by additionally picked foraminifer shells from the same samples (Data Set S1). The samples were depleted in large shells after these analyses and without additional confirmation data from these two samples were not interpreted further. Such poor reproducibility is unusual, as can be inferred from comparison of three shell-size specific sample pairs that reproduce within ±0.13‰ (Data Set S1).

Figure 2.

(a) SEM photos of the shell surface of two uncleaned G. sacculifer specimens taken from samples dated at (top) 2.16 Ma and (bottom) 4.03 Ma. Shells show no obvious sign of dissolution such as widened pore funnels or fissures. See also Data Set S1 for additional images. (b) Planktic foraminiferal δ11B with analytical error bars versus foraminiferal shell weight measured after cleaning. No correlation between shell weight and δ11B values confirms that the picked foraminifer shells used for δ11B analysis are unlikely to be biased by dissolution.

2.4. Seawater pH and Aqueous pCO2 Calculations

2.4.1. Calculations of pH

[14] To reconstruct seawater pH and aqueous pCO2 from δ11B (Figure 3a), a number of parameters are necessary that are summarized in Figures 3b, 3c, and 3e. To calculate pH (Figure 3d), species-specific empirical calibration curves are applied, which consistently show a shallower δ11B/pH slope than suggested for dissolved borate in seawater [Klochko et al., 2006]. Therefore instead of using an aqueous fractionation factor that does not describe the empirical pH dependence of our measured δ11B data, we apply the species-specific empirical calibration curves established by NTIMS analysis. To reflect the difference between the aqueous fractionation ɛ = 27.2‰ and the value that best describes the empirical data, we distinguish the use of the empirical value as ɛ* = 19.4‰ at 25°C and atmospheric pressure. The corresponding fractionation factor α* can then be calculated as α* = ɛ*/1000 + 1. To translate our measured boron isotope data into pH values, we used equation (1) after Hönisch et al. [2007]:

display math

where pKB is the equilibrium constant for the boric acid/borate system [Dickson, 1990], δ11Bs.w. is the δ11B value of seawater, δ11Bc is the isotopic composition of the measured carbonate, and ‘a’ = −4.2‰ is the constant offset between the apparent borate and empirical δ11B/pH calibration curve for G. sacculifer, adjusted for the additional instrumental difference described above [Sanyal et al., 2001; Hönisch et al., 2009].

Figure 3.

Parameters used to reconstruct seawater pH and pCO2 at ODP site 999. (a) ratios of δ11B measured on G. sacculifer (425–500 μm) with analytical error bars. Grey shading indicates Pleistocene minimal and maximal values obtained via the same method on G. sacculifer at site 668B [Hönisch et al., 2009], corrected for the 1.1‰ offset between the two laboratories (see text). (b) SST obtained from Mg/Ca ratios measured on G. sacculifer (315–400 μm) at site 999 [Groeneveld, 2005] (open dots), Mg/Cas.w. (open squares), and SST adjusted to secular changes in Mg/Cas.w. (solid dots). [Fantle and DePaolo, 2005, 2006] using the procedure outlined by Medina-Elizalde et al. [2008]. (c) SSS reconstructed from δ18Os.w. (solid triangles; see text) and from the record of ice volume change in open triangles [Mudelsee and Raymo, 2005]. (d) Surface water pH reconstructions according to the four scenarios defined in the text: 1, unadjusted SST and modern δ11Bs.w.; 2, adjusted SST and modern δ11Bs.w.; 3, unadjusted SST and Pliocene δ11Bs.w.; and 4, adjusted SST and Pliocene δ11Bs.w.. Grey shading indicates Pleistocene maximal and minimal pH values [Hönisch et al., 2009]. Average error bar is indicated. (e) Carbonate ion concentration ([CO32−]) estimated from past variations in the global CCD and in [Ca2+], using SSS in Figure 3c and adjusted SST (solid dots) and unadjusted SST (open dots) in Figure 3b. (f) Surface seawater pCO2 reconstructions according to the four scenarios defined in the text and indicated by numbers (see Figure 3d). Grey shading indicates preindustrial CO2 level of 280 μatm and minimal Pleistocene CO2 level of 180 μatm. Average error bar is indicated. Records of SST, SSS, ice volume, and Mg/Cas.w. have been interpolated to fit our time resolution.

2.4.2. Uncertainties of Seawater δ11B

[15] The present-day value of δ11Bs.w. has recently been estimated to ∼39.61‰ [Foster et al., 2010] but modeling studies [Lemarchand et al., 2000; Simon et al., 2006] suggest it was different during the Pliocene, despite the residence time of B in the ocean being about 14 My [Spivack and Edmond, 1987]. Indeed, δ11Bs.w. can be modified by a number of sources and sinks and Lemarchand et al. [2000] estimated a secular decrease of ∼0.1‰/My from 0 to 20 Ma, i.e., δ11Bs.w. ≅ 39‰ at 5 Ma. Similarly, Simon et al. [2006] estimated the δ11B ratio of the oceanic crust and performed a sensitivity study on δ11Bs.w., considering various exchange rates between the oceanic crust and the ocean. Their data, excluding variations in riverine input as studied by Lemarchand et al. [2000, 2002], show a decrease in δ11Bs.w. from the present-day value to 34–38‰ at 5 Ma. Similarly, Pearson and Palmer [1999] developed a δ11Bs.w. model based on the water column pH profile estimated from various deep-dwelling planktic foraminifer species, which has then been used by Pearson and Palmer [2000] to estimate Cenozoic pCO2. These estimates are uncertain because Pearson and Palmer [1999] have not considered the potential vital effects in different foraminifer species. However, despite any reservation to the accuracy of their estimate, their values yield intermediate δ11Bs.w. estimates of 38.6–39‰ for the period 6–3.9 Ma. A recent estimation of δ11Bs.w. from Messinian marine halite yield a much lower value of 34.65‰ (±0.3‰) at 5.5 Ma [Paris et al., 2010]. Such low δ11Bs.w. would imply Pliocene surface seawater pH > 8.5, and in turn would decrease δ11B-based Pliocene atmospheric pCO2 (in this study and for data by Seki et al. [2010]) to levels lower than Pleistocene level. Such an estimate seems inconsistent with warmer climate at that time. Because uncertainties in all these estimates are large, we chose to adopt Lemarchand's estimate of δ11Bs.w., which suggests a decrease of 0.1‰/My from today's value. For our pH reconstruction we thus consider two cases with (1) constant δ11Bs.w. of 39.61‰ and (2) δ11Bs.w. decreasing from today's value of 39.61‰ to 39.15‰ at 4.6 Ma. Due to the slow rate of change in the processes that govern the oceanic boron budget, short-term changes in δ11Bs.w. are not to be expected [Lemarchand et al., 2002].

2.4.3. Uncertainties of Sea Surface Temperature Reconstructions

[16] The calculation of pKB requires knowledge of sea surface temperature (SST), sea surface salinity (SSS) and water pressure. Pliocene SST was reconstructed from Mg/Ca measured on G. sacculifer (315–400 μm) at site 999 by Groeneveld [2005]. The incorporation of Mg in foraminifer shells during calcification mainly depends on the temperature of the seawater in which they grow, but also on the Mg/Ca ratio of the growth medium [Delaney et al., 1985]. Pliocene variations in seawater Mg/Ca (Mg/Cas.w.) may not be negligible, as Mg/Cas.w. may have increased by 1 mol/mol between 4 and 2.5 Ma [Fantle and DePaolo, 2005, 2006]. Adjusting Mg/Ca-based SST for these secular changes in seawater Mg/Ca contributes to increase Pliocene SST by 3–6°C (4°C on average) at site 999 (Figure 3b). However, it should be noted that these estimates of Mg/Cas.w. [Fantle and DePaolo, 2005, 2006] have a low temporal resolution and are still controversial, although Coggon et al. [2010] also found lower-than-modern Mg/Cas.w. during the Neogene. Moreover, it is uncertain how the partition coefficient of foraminiferal Mg/Ca is influenced by changes in Mg/Cas.w. [Medina-Elizalde et al., 2008]. Similar to δ11Bs.w., we calculated seawater pH according to two SST scenarios (1) with unadjusted and (2) with adjusted SST to Mg/Cas.w., following Medina-Elizalde et al. [2008]. SST were not reconstructed for samples younger than 2.3 Ma by Groeneveld [2005] and an extrapolated temperature of 25.7°C was used to calculate pH and pCO2 for these four samples.

2.4.4. Uncertainties of Sea Surface Salinity Reconstructions

[17] Local changes in SSS at site 999 (Figure 3c) were estimated using the equation by Steph et al. [2006] from seawater δ18O corrected for changes in global ice volume, where ice volume estimates were taken from Mudelsee and Raymo [2005] (Figure 3c). Seawater δ18O was calculated following Shackleton [1974] from unadjusted SST and δ18O measured on G. sacculifer [Steph et al., 2009]. For comparison, SSS was also calculated following the same method but using the estimations of sea level changes to correct for changes in global ice volume (Figure S2). Any estimation of Pliocene SSS changes is greatly hindered by the lack of reliable global ice volume or sea level changes for this epoch. However, SSS exerts only a small influence on the uncertainty of pH (±0.01 pH units per 1‰ SSS) and of pCO2. For instance, pCO2 calculated with a constant salinity of 33.5‰ and of 37.5‰ differ by only 10–20 μatm in comparison.

2.4.5. Uncertainty of Carbonate Ion Concentration Reconstructions

[18] In addition to the pH values derived from δ11B (Figure 3d) a second parameter of the ocean carbonate system is needed to estimate aqueous pCO2 in seawater. Here we reconstructed the carbonate ion concentration ([CO32−]) following the method described by Tyrrell and Zeebe [2004], which is based on the observation that the calcite saturation state of seawater (Ω) was quasi-constant over the past 100 My. The calcite saturation state is given by

display math

where Ksp* is the stoichiometric solubility product of calcite. Given the calcium concentration and Ksp*, and assuming Ω approximately constant over the time interval of interest, global ocean [CO32−] can be calculated back in time [see Tyrrell and Zeebe, 2004]:

display math

[19] In our calculation of carbonate chemistry parameters, we included changes in the ocean's carbonate compensation depth over time [Tyrrell and Zeebe, 2004] and effects of whole ocean [Ca2+] and [Mg2+] (and the Mg/Ca ratio, [Tyrrell and Zeebe, 2004]) on stoichiometric equilibrium constants K1*, K2*, and Ksp* [Zeebe and Wolf-Gladrow, 2001]. Effects of local changes in SST and SSS at site 999 (Figures 3b and 3c) on these equilibrium constants were also taken into account for the calculation of [CO32−] and pCO2 (Figures 3e and 3f).

[20] In addition, because all parameters are estimated relative to modern conditions, a modern glacial/interglacial average [CO32−] value is required for site 999. GLODAP data from stations near the core site [Key et al., 2004] indicate a modern (interglacial) CO32− concentration of about 290 μmol kg−1. For the last glacial period, [CO32−] can be estimated as ∼340 μmol kg−1, given an atmospheric CO2 concentration of 200 ppmv. Thus a modern glacial/interglacial average value of [CO32−] = 315 μmol kg−1 was chosen for site 999 [(290 + 340)/2 = 315], and the global estimates of Tyrrell and Zeebe [2004] were scaled relative to this value. An uncertainty of 1 μmol kg−1 in [CO32−] translates into an uncertainty of 2 μatm in pCO2. Finally, given the reconstructed [CO32−] and pH values, aqueous pCO2 can be calculated using carbonate chemistry routines as given by Zeebe and Wolf-Gladrow [2001] and Tyrrell and Zeebe [2004]. For comparison, pCO2 was also estimated using various constant values of [CO32−] and total alkalinity (Figure S3). However, Pliocene [CO32−] and total alkalinity were unlikely constant from 1.8 to 4.7 Ma but likely differed from modern values as discussed by Tyrrell and Zeebe [2004].

2.4.6. Propagated Uncertainties

[21] To estimate the contribution of each parameter to the final pH uncertainty, we used an uncertainty of ±1°C (uncertainty on Mg/Ca-derived SST), ±1‰ salinity (because of the uncertainty on ice volume changes), and the analytical uncertainty on δ11B (±0.34‰ on average). The main uncertainty on the pH reconstruction (Table 1) stems from the analytical uncertainty of boron isotope measurements and uncertainties on the SST estimate (±0.04 and ±0.02 pH units, respectively), giving a propagated 2σ uncertainty of ±0.044 pH units on average for all parameters. For estimating uncertainty on our pCO2 estimates, we estimated a pH uncertainty increasing from −37/+44 μatm at 200 μatm pCO2 increasing to −73/+88 μatm at 400 μatm pCO2, and an uncertainty of ±25 μmol kg−1 on [CO32−]. The average propagated uncertainty on pCO2 estimates (absolute values) is −70/+80 μatm (Table 1).

Table 1. Averaged Uncertainties on pH and pCO2 Determination Calculated With ±1°C, ±1‰ salinity, ±0.34‰ Measured δ11B, and a ±25 μmol kg−1 Error on the Carbonate Ion Concentrationa
 SST (±1°C)SSS (±1‰)δ11B (±0.34‰ 2 Standard Error on Average)Propagated Error (Average)
  • a

    Error propagations are the square root of the sum of all squared deviations for each parameter.

pH with δ11Bs.w. = 39.61‰ pH units
 pH (±0.044)[CO32−] (±25 μmol kg−1)  
ΔpCO2 (μatm)−53/+6628 −60/+70 μatm

3. Age Model

[22] The age model for site 999 was originally established by Haug and Tiedemann [1998] based on benthic foraminiferal δ18O stratigraphy. Here the benthic δ18O record from site 999 was correlated to the global stack of 57 benthic foraminiferal δ18O records (“LR04”) [Lisiecki and Raymo, 2005] without changing the definition of the marine isotope stages (MIS) by Haug and Tiedemann [1998]. The result is displayed in Figure 1. Where it was necessary, SST and SSS, ice volume, Mg/Cas.w. and δ11Bs.w. were linearly interpolated to fit our time resolution using the AnalySeries 2.0.3 package [Paillard et al., 1996].

4. Results

4.1. Boron Isotope Ratios and Sensitivity Study for pH and pCO2

[23] Pliocene δ11B values ranged between 20.11‰ and 21.70‰ over the time period 4.6–2.0 Ma (Figure 1) and sample preparation did not appear to influence the measured δ11B (Data Set S1). For comparison, Pleistocene (2.1–0 Ma) δ11B ratios obtained using the same analytical technique but originating from another core site [Hönisch et al., 2009] range between 22.29‰ and 20.55‰ (−1.1‰ for the offset between laboratories, Figure 4a). Compared to the Pleistocene, the lowest δ11B values of the early Pliocene were on average 0.59‰ lower and the highest values were on average 0.65‰ lower, confirming a shift from the early Pliocene toward the Pleistocene following the increase in global ice volume (Figure 4b). The δ11B minima-maxima amplitude increased from 0.58‰ during the early Pliocene to 0.94‰ during the late Pliocene, and to 1.1‰ during the early Pleistocene. Between 3.6 and 3.2 Ma, the δ11B minima-maxima amplitude was reduced to 0.39‰, and in particular between 3.4 to 3.32 Ma. After 2.7 Ma, the δ11B minima-maxima amplitude increased gradually (Figure 4a) toward the Pleistocene amplitudes.

Figure 4.

Planktic foraminiferal δ11B ratios, benthic δ18O, and estimated pCO2 during the Plio-Pleistocene. (a) Planktic foraminiferal δ11B ratios measured on G. sacculifer (black) with open symbols for the Pleistocene epoch [Hönisch et al., 2009] and with solid symbols for the Pliocene epoch, corrected by −1.1‰ for an offset between the two laboratories (see text) with analytical error bars. (b) Benthic foraminiferal δ18O stack LR04 [Lisiecki and Raymo, 2005] reflects changes in deep-sea temperature and ice volume. (c) Atmospheric pCO2 measured in ice cores [Lüthi et al., 2008] in dark gray, estimated from G. sacculifer δ11B in light blue [Hönisch et al., 2009] and dark blue (this study) and from G. ruber in orange [Seki et al., 2010], estimated from alkenone ɛp in light green squares [Seki et al., 2010] and from stomatal indices in dark green squares [Kürschner et al., 1996]. Upper and lower CO2 estimates from Seki et al. [2010] were obtained by using two different estimations of alkalinity. Red vertical dashed lines denote major climatic transitions: the mid-Pleistocene transition, the late Pliocene transition, and the start of Northern Hemisphere glaciations at 3.3 Ma. The turquoise dashed lines embrace the dominant CO2 changes as obtained from our boron isotope estimates.

[24] To decipher the uncertainties in pCO2 linked to uncertainties in δ11Bs.w. and Mg/Cas.w., four scenarios were considered for our pH and pCO2 estimates (Figures 3d and 3f): Scenario 1 considers unadjusted SST and constant δ11Bs.w., scenario 2 considers adjusted SST and constant δ11Bs.w., scenario 3 considers unadjusted SST and varying δ11Bs.w., and scenario 4 considers adjusted SST and varying δ11Bs.w.. Adjusting the SST estimate for variations in Mg/Cas.w. (scenario 2) translates into lowest pH and highest pCO2 estimates, whereas adjusting for δ11Bs.w. (scenario 3) results in the highest pH and lowest pCO2 estimates. Scenarios 1 (no adjustment) and 4 (adjusting both Mg/Cas.w. and δ11Bs.w.) yield very similar results, which are intermediate between scenarios 2 and 3. This may not be surprising as the secular variations in oceanic δ11Bs.w. and Mg/Cas.w. may be coupled via global runoff and weathering as suggested by Paris et al. [2010].

[25] The difference in pCO2 between the four scenarios decreases from 200 μatm at 4.7 Ma to 50 μatm at 2 Ma. Adjusting Mg/Ca-based SST for changes in Mg/Cas.w. (scenarios 2 and 4) or using a Pliocene value for seawater δ11Bs.w. (scenarios 3 and 4) both increase the pCO2 by 20 to 100 μatm (on average 40–50 μatm). In all scenarios, except scenario 3, early Pliocene pCO2 estimates are higher than the preindustrial maximum of 280 μatm. Because both models of secular evolution in Mg/Cas.w. and δ11Bs.w. are fraught with uncertainty and because it is unlikely that secular changes occurred only in Mg/Cas.w. or only in δ11Bs.w., we consider the intermediate estimates the best guess at this time. Our further discussions focus on the results of scenario 1.

4.2. Pliocene CO2 Changes

[26] Calculated aqueous pCO2 values (scenario 1) at site 999 range between 195 and 425 μatm over the period 4.6–2.0 Ma (Figure 4c and Data Set S1). Small differences exist between the δ11B record (Figure 4a) and the calculated pCO2 record (Figure 4c), however the main trends are the same, confirming that the pCO2 estimates are not driven by secondary corrections such as SST or [CO32−]. The highest pCO2 values were reached during the early Pliocene prior to 3.6 Ma with a maximum of 425 μatm at 4.58 Ma and 410 μatm on average (Figure 4c). During this period, lowest pCO2 values ranged 290–330 μatm and 300 μatm on average. Between 3.4 and 3.32 Ma, maximal and minimal δ11B values were similar and translated into relatively low pCO2 values of 290 μatm on average, ranging 250–300 μatm (Figure 4c). During this period, minimal pCO2 values continued a gradual decrease from ∼300 μatm to ∼245 μatm at 3.1 Ma. After 3.1 Ma, maximal pCO2 values resumed to higher values of ∼400 μatm ranging 355–420 μatm until 2.7 Ma, when they decreased to ∼350 μatm, still higher by 50 μatm than Pleistocene values. Between 3.1 and 2.7 Ma, minimal (glacial) pCO2 values were relatively constant at ∼245 μatm. At 2.7 Ma minimal (glacial) pCO2 abruptly decreased to ∼200 μatm, a level similar to early Pleistocene glacial values (Figure 4c). The ∼45/50 μatm pCO2 decrease during glacial stages after 2.7 Ma coincides with a +0.4‰ increase in glacial benthic δ18O as reported by Lisiecki and Raymo [2005] (Figure 4b). At 2.2 Ma, maximal CO2 increased again to 380 μatm and finally decreased to the Pleistocene level of 300 μatm at 2.0 Ma (Figure 4c). During the Pliocene, minima/maxima amplitude of estimated pCO2 were as high as 100 μatm (Figure 4c), similar to the early Pleistocene [Hönisch et al., 2009].

[27] In summary, our estimates describe a gradual decrease in both minimal and maximal pCO2 estimates since 4.1 Ma. The total CO2 decline was ∼−100 μatm, with a noted step at 2.7 Ma, when Pleistocene glacial values were reached while interglacial pCO2 values remained +50 μatm higher than Pleistocene interglacial values until 2.0 Ma (Figure 4c).

5. Discussion

5.1. Comparison With Other CO2 Records

[28] Because CO2 is well mixed in the atmosphere and because surface ocean carbonate chemistry at site 999 is close to equilibrium with the atmosphere, we consider our aqueous pCO2 estimates approximate to atmospheric pCO2, hereafter referred to as δ11B-based pCO2. At 2.0 Ma our estimates overlap with pCO2 estimates from the eastern equatorial Atlantic obtained via the same method [Hönisch et al., 2009] (Figure 4), which raises confidence that boron isotope reconstructions from oligotrophic oceanic regions indeed yield reliable estimates of atmospheric pCO2.

[29] Few stomata index-based CO2 estimates of 280 and 370 μatm between 3.3 and 2.0 Ma [Kürschner et al., 1996] are also in general agreement with our data (Figure 4c). Pagani et al. [2010] reconstructed surface ocean pCO2 based on alkenone ɛp from 6 locations in the oligotrophic and mesotrophic Atlantic and Pacific Oceans (not shown). For all these locations, their estimates show consistently higher pCO2 for the Pliocene, although the absolute values differ between sites and only two of their records approach ice core CO2 values of 170–280 ppmv during the Pleistocene. Pagani et al. [2010] estimated average pCO2 values of 365–415 μatm during the warm early Pliocene, which is similar to our CO2 estimates (380–420 μatm), and described a decrease of atmospheric pCO2 by 45–144 μatm or 41–216 μatm, depending on their assumption of nutrient supply, over the last 4.5 Ma. Our boron isotope reconstruction is restricted to the period 4.2–2.0 Ma and registers a relatively similar decrease by 100 μatm.

[30] Pearson and Palmer's [2000] Cenozoic record is also based on planktic foraminiferal δ11B, using a different analytical technique than in this study. Although their Pliocene estimates show a similar decline in atmospheric CO2 from 280 to 210 ppm between 3.87 Ma to 3.0 Ma, their low temporal resolution and use of mixed planktic foraminifer species preclude any meaningful comparison. A more detailed comparison can be made with recently published pCO2 estimates from site 999 based on alkenone ɛp and δ11B [Seki et al., 2010]. In general both data sets are in good agreement and show similar CO2 trends and absolute values through time (Figure 4c). However, the sampling strategy applied by Seki et al. [2010] appears to have favored interglacial times, as they did not yield any pCO2 estimates lower than 250 μatm during the Pleistocene. Similarly, the pCO2 estimates from alkenone ɛp only reproduce intermediate values (∼240 μatm) if low nutrient concentrations are assumed. Their favored scenario of elevated nutrient concentrations agrees well with the boron isotope ratios measured on the same samples but suggests pCO2 > 250 μatm for the entire record (Figure 4c).

[31] Finally, reconstructions of Pliocene pCO2 include estimates from foraminiferal B/Ca ratios (not shown) spanning 3.4–2.4 Ma at western tropical Pacific site 806 [Tripati et al., 2009]. Their CO2 data present a maximum for the Pliocene epoch of only 300 μatm (±50 μatm) around 3.2 Ma and a decrease to 150 μatm (±50 μatm) around 2.8 Ma. This range seems biased toward too low values compared to other Pliocene records, which record atmospheric pCO2 higher than the preindustrial level for this warm epoch [Seki et al., 2010; Pagani et al., 2010; Kürschner et al., 1996; Raymo et al., 1996], including this present study. Because various calibration data sets for a pH effect on foraminiferal B/Ca ratios show conflicting results [Yu et al., 2007; Foster, 2008; Tripati et al., 2009], much more information is needed on this new proxy before past pCO2 reconstructions can be approached with confidence [Tripati et al., 2011].

[32] With the exception of the B/Ca estimates, all records agree that interglacial Pliocene pCO2 was higher by about +110 μatm compared to the Pleistocene, although small differences exist in the details of each record. In agreement with studies referenced above, our record shows that a gradual decline in CO2 by ∼100 μatm from 4.1 to 2.0 Ma coincides with the late Pliocene onset and intensification of Northern Hemisphere Glaciations, supporting the link between glaciation and decrease in atmospheric CO2 concentration, as previously postulated [Raymo et al., 1996; Seki et al., 2010; Pagani et al., 2010] and modeled [Lunt et al., 2008]. However, the higher temporal resolution of this record offers a more detailed perspective on the timing and the supposed causes of the atmospheric CO2 changes between 2.0 and 3.6 Ma, as described below.

5.2. Implications for the Onset and Intensification of Northern Hemisphere Glaciations

[33] Prior to the start of Northern Hemisphere glaciations, early Pliocene δ11B-based pCO2 values between 4.6 and 3.6 Ma averaged maxima of 410 μtam and minima of 310 μatm, both above a suggested threshold value of 280 μatm below which Northern Hemisphere glaciations are possible [DeConto et al., 2008]. This may explain in part why large-scale glaciations on Greenland were not recorded during this time. Alternatively, Koenig et al. [2011] defined other CO2 threshold values between 200 and 400 μatm for the initiations of glaciations on Greenland by taking into account vegetation changes (i.e., forest versus tundra) over an initially ice-free Greenland. They concluded on a critical role of decreasing atmospheric CO2 in the glaciation of Greenland at 2.74 Ma, although other albedo-related feedbacks were also significant, including the vegetation cover on Greenland and the sea ice cover in Greenland and Labrador Seas [Koenig et al., 2011]. If the CO2 threshold defined by Koenig et al. [2011] is correct, Miocene and early Pliocene atmospheric CO2 estimates <400 μatm [Seki et al., 2010; Pagani et al., 2010; Kürschner et al., 1996; this study] would explain why ice-rafted detritus deposition in the North Atlantic could start as early as 5–10 Ma [Wolf and Thiede, 1991; Jansen and Sjøholm, 1991].

[34] After 4.2 Ma and until 2.0 Ma, minima/maxima amplitudes of estimated pCO2 were as high as 100 μatm (Figure 4c), similar to the Pleistocene glacial/interglacial CO2 amplitude [Hönisch et al., 2009]. While the Pleistocene amplitude of 100 μatm is still not fully understood [Kohfeld et al., 2005], it is clear that glacial/interglacial CO2 variations are mainly controlled by the Southern Ocean through ventilation of the deep ocean and biological pump strength [see Sigman et al., 2010]. Although the estimated Pliocene minima/maxima CO2 amplitudes of 100 μatm require validation by other atmospheric CO2 records, such amplitudes could be controlled by the state of the Southern Ocean. Indeed, glacial/interglacial oscillations in the size of the West Antarctic Ice Sheet during the early Pliocene [Naish et al., 2009] may have influenced the overturning of the Southern Ocean and the sequestration of atmospheric CO2 during the early Pliocene. Similarly, it was suggested by Sigman et al. [2004] that overturning of the polar oceans could be responsible for large amounts of CO2 degassing during the late Oligocene and middle Miocene.

[35] After the start of Northern Hemisphere glaciations around 3.6 Ma, maximal pCO2 estimates transiently decreased from 410 μatm to 260 μatm on average between 3.4 and 3.32 Ma, while minimal pCO2 estimates gradually decreased from 310 μatm to 245 μatm. A transient increase in glacial and interglacial ice volume is also recorded between 3.43 and 3.32 Ma by an increase of 0.20‰ in the benthic δ18O stack LR04 (Figure 4b) and/or by a transient increase of 12 m equivalent sea level (0.11‰ δ18O) around 3.3 Ma (Figure 3c) [Mudelsee and Raymo, 2005]. The transient glaciation on Greenland at this time also coincides with a transient decrease in North Atlantic SST between 3.3 and 3.5 Ma by 2°C during interglacials and 5°C during glacials at site U1313 [Naafs et al., 2010] and at site 982 [Lawrence et al., 2009]. Increased ice rafted-detritus deposition at 3.3 Ma in the North Atlantic [Jansen et al., 2000; Kleiven et al., 2002] and in the Labrador Sea [Sarnthein et al., 2009] also evidenced the transient increase in the size of the Greenland and Laurentide ice sheets. Although the transient decline in pCO2 coincides with transient glaciations on Greenland between 3.43 and 3.32 Ma, the maximal pCO2 estimates between 3.6 and 3.2 Ma are as low as Pleistocene interglacial values, an observation that may be unexpected from comparison with the benthic δ18O stack LR04 (Figure 4b). However, several lines of evidence suggest that the low pCO2 estimates could indicate temporary disequilibrium between surface seawater at site 999 and the atmosphere during this time. Increased nutrient supply to site 999 between 3.5 and 3.1 Ma [Kameo, 2002] and increased surface seawater primary productivity between 3.4 and 3.35 Ma at nearby site 502A [Bornmalm et al., 1999] have been documented for this time, suggesting a possible lowering of surface seawater pCO2 compared to the atmosphere via increased primary productivity [Wanninkhof et al., 2007]. Similarly, low surface water salinity at site 999 between 3.5 and 3.1 Ma could suggest an inflow of low-salinity and nutrient-rich Amazon and Orinoco river water, brought to the Caribbean Sea by the Guyana Current [Kameo et al., 2004]. In any case, the decrease in pCO2 estimates appears to have occurred during the start of Northern Hemisphere glaciations at 3.3–3.6 Ma.

[36] Between 3.2 and 2.7 Ma, maximal pCO2 estimates returned to values of 400 μatm on average until 2.77 Ma (Figure 4c), similar to present-day CO2 level. Interglacial ice volume also resumed to previous level during this time (Figure 4b). Although it was not the specific focus of this paper, we estimate high pCO2 of 410 μatm (MIS K1) and 350 μatm (MIS G19) during interglacials of the mid-Pliocene warm period (3.29–2.97 Ma). This confirms the earlier notion that the warmer climate characteristic of the mid-Pliocene warm period may have been partly driven by a stronger greenhouse effect due to higher atmospheric CO2 [Raymo et al., 1996]. Lunt et al. [2010] discussed in detail the implication of such a result for the estimation of Earth sensitivity through time. Between 3.2 and 2.7 Ma, minimal CO2 estimates stayed at a level of 245 μatm (Figure 4c), while maximal glacial ice volume also stayed constant between 3.1 and 2.8 Ma, as evidenced by the LR04 benthic δ18O stack (Figure 4b). Ice-rafted detritus deposition was also low in the North Atlantic after the transient increase around 3.3 Ma until 2.7 Ma [Kleiven et al., 2002]. Moreover, North Atlantic SST records off South Iceland increased by 2–3°C during interglacial and glacial stages at site 984 [Bartoli et al., 2005] and site 982 [Lawrence et al., 2009] between 3.05 and 2.80 Ma. This suggests that warming over the northern North Atlantic, associated with high atmospheric CO2 concentrations and low obliquity [Laskar et al., 2004] jointly prevented the further growth of the Greenland ice sheet between 3.2 and 2.7 Ma.

[37] During the late Pliocene transition at 2.73 Ma, global ice volume increased by 22 m equivalent sea level [Sosdian and Rosenthal, 2009], glacial/interglacial cycles intensified [Lisiecki and Raymo, 2007], deepwater cooled by 2°C [Sosdian and Rosenthal, 2009], and ice-rafted detritus deposition increased over the North Atlantic, thus confirming the increase of the Greenland, Laurentide, and Scandinavian ice sheets [Kleiven et al., 2002]. After 2.7 Ma, our minimal pCO2 estimates decreased by 45 μatm (Figure 4c) and reached values similar to early Pleistocene glacial values of 200 μatm, while maximal pCO2 estimates decreased by 50 μatm but stayed +50 μatm higher than during early Pleistocene interglacials. Our results are consistent with the timing of increased stratification in the subarctic North Pacific [Sigman et al., 2004] and in the Southern Ocean [Hodell and Venz-Curtis, 2006; Waddell et al., 2009] at 2.73 Ma. It has been suggested that increased polar stratification led to enhanced sequestration of atmospheric CO2 in the oceanic abyss after 2.73 Ma, by preventing CO2-rich deep waters to reach the surface. Stratification thus could have played a major role in the onset of large-scale ice sheets in the Northern Hemisphere [Sigman et al., 2004]. North Pacific stratification alone could account for a decrease of 30–40 μatm [Haug et al., 1999], which is close to the 45 μatm glacial decrease registered in this study between 2.70 and 2.68 Ma. In addition, the strong North Pacific halocline may have persisted during both glacials and interglacial in the North Pacific [Swann, 2010] and thus may have played a role in the 50 μatm decrease during interglacials after 2.7 Ma. Recently, Martínez-Garcia et al. [2011] presented a record of Aeolian iron input to the Southern Ocean that suggests increased iron fertilization during glacials after 2.7 Ma. In addition to polar stratification, increased export production and nutrient utilization in the Southern Ocean thus may have contributed to the glacial decrease in atmospheric CO2 [Martínez-Garcia et al., 2011]. Increased dust supply to the North Pacific after 2.75 Ma [Bailey et al., 2011] may have had a similar fertilization effect on the North Pacific and adjacent seas, leading to higher CO2 drawdown in these regions as well. Our results suggest that a threshold was crossed after 2.7 Ma, when the onset of polar stratification and glacial iron fertilization in the North Pacific and Southern Ocean were responsible for the glacial atmospheric CO2 decrease to values characteristic of Pleistocene glacials. Interestingly, tropical SST changes in various ocean basins show a strong increase in the 41 ky orbital periodicity at 2.7 Ma [Herbert et al., 2010]. This suggests that atmospheric CO2 and glacial feedbacks after 2.7 Ma were strong enough to imprint a 41 ky periodicity on tropical SST records as compared to prior to 2.7 Ma [Herbert et al., 2010]. Our study agrees well with a decrease in glacial atmospheric CO2 to Pleistocene levels after 2.7 Ma.

[38] In comparison, maximal pCO2 estimates decreased by only 50 μatm after 2.7 Ma in our record (Figure 4c) but another high level of 380 μatm is observed at 2.2 Ma, suggesting that interstadial pCO2 stayed above early Pleistocene values until 2.0 Ma. A single estimate from stomata also records 360 μatm around 2.0 Ma (Figure 4c). Our findings may suggest that after the sequestration of atmospheric CO2 in the oceanic abyss after 2.7 Ma, another mechanism was responsible for the lowering interglacial atmospheric CO2 after 2.2 Ma, such as the intensification of the Benguela upwelling system between 2.1 and 1.9 Ma, which would have increased the efficiency of the biological pump [Marlow et al., 2000]. Alternatively, Etourneau et al. [2010] suggested that increased upwelling off Namibia between 2.4 and 2.0 Ma would have increased atmospheric CO2 by releasing deep water CO2 to the surface and thereby counteracting the effects of atmospheric CO2 sequestration elsewhere in the oceans until 2.0 Ma. This mechanism is also supported by our data showing increased atmospheric CO2 at 2.2 Ma compared to 2.5 Ma.

[39] In summary, the increase in the size of the Greenland ice sheet during the Pliocene epoch was short lived between 3.43 and 3.32 Ma [Lisiecki and Raymo, 2005] and permanent after 2.7 Ma [Haug et al., 2005], and is matched by decreased atmospheric CO2 concentrations as compared to the early Pliocene. Following the pattern given by the LR04 benthic δ18O stack (Figure 4b) and tropical SSTs [Herbert et al., 2010, Figure 6], minimal atmospheric CO2 concentrations decreased faster than maximal atmospheric CO2 concentrations over the course of the Plio-Pleistocene. It seems that an important climatic threshold was crossed at 2.7 Ma, when estimated pCO2 minima decreased by ∼45 μatm and approached early Pleistocene glacial values. This decrease and its timing are consistent with the onset of polar stratification [Haug et al., 1999; Sigman et al., 2004] associated with iron fertilization in the Southern Ocean [Martínez-Garcia et al., 2011] and in the North Pacific [Bailey et al., 2011]. It is possible that Southern Ocean stratification starting around 3.3 Ma [Hillenbrand and Cortese, 2006] has contributed to the early atmospheric CO2 decrease between 3.43 and 3.32 Ma, although the role of terrestrial weathering, in particular of the Tibetan-Himalayan Plateau between 4.2 and 2.7 Ma [Zhang et al., 2009], and other possible mechanisms also have to be considered. Based on the timing of the precursor and final closures of Panama dated at 3.15–3.3 Ma and at 2.82–2.95 Ma [Bartoli et al., 2005], the final closure of Panama did not influence the decrease in atmospheric CO2 recorded here. It can be postulated that a decrease in atmospheric CO2 concentrations was necessary in order to sustain large-scale glaciations in the Northern Hemisphere after 2.7 Ma, as opposed to the ephemeral glaciations of the Pliocene prior to 3.2 Ma.

6. Conclusions

[40] We reconstructed Pliocene atmospheric CO2 based on foraminiferal δ11B between 4.6 and 2.0 Ma. Our data contribute to previously published evidence that suggests atmospheric CO2 was higher during the Pliocene epoch by 130 μatm compared to preindustrial levels, and close to the modern level of 400 μatm. Estimated pCO2 decreased during both the start and the intensification of Northern Hemisphere glaciations around 3.3 Ma and at 2.7 Ma, respectively. After 2.7 Ma, minimal atmospheric CO2 dropped to values characteristic of early Pleistocene glacials, consistent with the suggested increased sequestration of atmospheric CO2 in the North Pacific and Southern Ocean via polar stratification and increased iron fertilization. In comparison, maximal CO2 estimates reached Pleistocene values only after 2.0 Ma and stayed below 300 μatm until the 1940s.


[41] We thank the Ocean Drilling Program for making the samples available for this study and M. Schmidt for sharing some of his samples. S. Bernasconi and T. Schmid are acknowledged for providing stable isotopes analyses, U. Brupbacher for laboratory assistance, and M. Medina-Elizalde for helping us with the adjustment of Mg/Ca-based temperatures. Discussions with G. Haug, S. Jaccard, and A. Martínez-Garcia were extremely useful. One anonymous reviewer, G. L. Foster, and the editor C. Charles are thanked for comments that helped improve this manuscript. This work was supported by Swiss National Foundation grant 200020-118045/1 and Individual short visit grant NF PI0I2-119420 (G.B.), by NSF OCE 06-23621 (B.H.), and by NSF OCE09-27089 (R.E.Z.). B.H. also thanks G. Lenfest for providing the funds for the TRITON at LDEO.