Late Miocene through early Pleistocene nutrient utilization and export production in the Antarctic Zone of the Southern Ocean

Global and Planetary Change 100 (2013) 353–361

Contents lists available at SciVerse ScienceDirect

Global and Planetary Change journal homepage: www.elsevier.com/locate/gloplacha

Late Miocene through early Pleistocene nutrient utilization and export production in the Antarctic Zone of the Southern Ocean Katharina Billups a,⁎, Anthony Aufdenkampe b, Rebecca Hays a, 1 a b

University of Delaware, School of Marine Science and Policy, 700 Pilottown Road, Lewes, DE 19958, United States Stroud Water Research Center, 970 Spencer Road, Avondale, PA 193113, United States

a r t i c l e

i n f o

Article history: Received 15 June 2012 Accepted 27 November 2012 Available online 4 December 2012 Keywords: Pliocene sedimentary nitrogen isotopes Southern Ocean paleoproductivity Ocean Drilling Program Leg 119, Site 745

a b s t r a c t We use bulk sediment δ15N values and opal and carbon mass accumulation rates (MAR) to reconstruct nutrient utilization and export productivity at Ocean Drilling Program Site 745 (Antarctic Zone of the Southern Ocean) spanning the late Miocene through early Pleistocene (~6.5–1.4 Ma). We investigate whether early Pliocene climatic warmth and subsequent cooling can be related to changes in high latitude productivity. Results indicate that δ15N values increase to above late Holocene levels from the late Miocene through the late Pliocene (6.5 to 2 Ma). Opal and carbon MARs are low during the early Pliocene. Relatively high δ15N together with low export production is consistent with a more southerly position of the Polar Frontal Zone (PFZ) allowing the expansion of nitrate depleted, low nutrient upper waters south toward Site 745. The interpretation is supported by a relatively small δ15N gradient between Site 745 and a site in the Subantarctic Zone of the Southern Ocean (Site 1090). There are no unique changes in the Site 745 δ15N values or export productivity at 2.7 Ma. During the late Pliocene to early Pleistocene climate transition (between ~2.1 and at 1.7 Ma), δ15N values display large variations approaching those observed during the last glacial to interglacial transition in this latitude band. Opal and carbon MARs also show large fluctuations, but in the opposite sense with maxima corresponding to minima in the δ15N record and vice versa. The pattern of high δ15N values associated with low export production may reflect changes in nutrient utilization in response to changes in water column stratification once the PFZ has moved north of the location of Site 745. Our results provide a mechanism for enhancing early Pliocene CO2 concentrations via reduced uptake of CO2 due to low productivity in the Southern Ocean. Once the PFZ has moved north, the region may have become sensitive to changes in water column stratification, potentially contributing to fluctuations in CO2. © 2012 Elsevier B.V. All rights reserved.

1. Introduction Because of the direct link between export productivity and atmospheric CO2 levels, much research effort has been directed at reconstructing aspects of the carbon cycle during the past. The Antarctic Zone of the Southern Ocean may be a significant variable in regulating atmospheric CO2 levels as it is a vast region where CO2 (and nutrient)-rich deep waters upwell to the surface. The interplay between the rate of CO2 removal during phyotoplankton photosynthesis and the rate of CO2 and nutrient delivery from the deep affects the amount of CO2 escaping to the atmosphere thereby contributing to global climate change (Toggweiler and Sarmiento, 1985; Watson and Garabato, 2006). In the Antarctic Zone of the Southern Ocean, the region south of the Polar Frontal Zone, an important factor in determining the ⁎ Corresponding author. E-mail address: [email protected] (K. Billups). 1 Now at: Eastern University, 1300 Eagle Road, St. Davids, PA 19087, United States. 0921-8181/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.gloplacha.2012.11.014

direction of the CO2 flux with respect to the air–sea interface is the degree of upper water column stratification (Francois et al., 1997; Sigman et al., 2004; Watson and Garabato, 2006). Weak polar ocean stratification favors upwelling and hence venting of CO2 into the atmosphere while a well-stratified water column limits upwelling and thus the amount of CO2 brought to the sea surface. Changes in water column stratification arise essentially from two factors: changes in the Ekman pull of the westerly winds and buoyancy fluxes related to both mixing with dense bottom water currents and air–sea heat exchange (Watson and Garabato, 2006). Water column stratification has been inferred from the degree to which nutrients such as nitrate are consumed by phytoplankton (Francois et al., 1992; Altabet and Francois, 1994; Francois et al., 1997; Sigman et al., 1999; Horn et al., 2011). For example, enhanced water column stratification limits the amount of upwelling of deeper and nutrient-laden waters to the photic zone. Thus, by reconstructing the amount of nutrients in the upper water column, or the degree of nutrient utilization, it is possible to track temporal changes in water column stratification on geologic time scales.

354

K. Billups et al. / Global and Planetary Change 100 (2013) 353–361

The nitrogen isotopic composition (δ15N values) of bulk sediment reflects nitrate utilization, and it has been used to assess changes in water column stratification through time. Similar to the carbon isotopic composition of organic matter, δ 15N values are low in the organic phase with respect to dissolved inorganic source (i.e. NO3−) because biological assimilation discriminates against the heavier isotope ( 15N). However, the degree of discrimination decreases as dissolved inorganic nitrogen concentrations decrease and become increasingly limiting. As a result, with increasing NO3− depletion, or nitrate utilization, δ15N values will increase in the biomass (e.g., Mariotti et al., 1981; Montoya and McCarthy, 1995; Wu et al., 1997). Therefore, there is an inverse relationship between nitrate concentrations in surface water and the δ15N of nitrate and algal nitrogen, which is reflected in the δ15N of sedimentary nitrogen (Francois and Altabet, 1992; Wu et al., 1997; Sigman et al., 1999). For example, water column stratification in the Antarctic Zone has been invoked to explain higher bulk sediment δ15N values coupled with lower opal accumulation rates during the Last Glacial Maximum with respect to the late Holocene (Francois and Altabet, 1992; Altabet and Francois, 1994; Francois et al., 1997; Sigman et al., 1999; Horn et al., 2011). Here we use a bulk sediment δ 15N record together with opal and organic carbon mass accumulation rates (MAR) from the Antarctic Zone of the Southern Ocean (Ocean Drilling Program, ODP, Site 745) to assess potential changes in nitrate utilization and export production from the late Miocene through early Pleistocene. The early Pliocene (~5–3.5 Ma) is one of the most recent intervals of time characterized by prolonged relative global warmth (e.g., Ravelo et al., 2007) and higher CO2 levels with respect to pre-anthropogenic times (Pagani et al., 2010). Global cooling followed with the expansion of Northern Hemisphere glaciers between 3.5 and 2.5 Ma (Mudelsee and Raymo, 2005) and decreasing atmospheric pCO2 (Pagani et al., 2010). We hypothesize that the Southern Ocean nutrient utilization and export production may have played a key role in the long-term decrease in atmospheric CO2 levels and climatic cooling during this interval of time.

2. Regional setting Site 745 (59°37′S, 85°52′E) is located in the Australian–Antarctic Basin in the Indian Ocean sector of the Southern Ocean in a water depth of 4082 m (Fig. 1). This region lies within the Antarctic Zone where nutrient and CO2 rich deeper waters rise to the surface (Pickard and Emery, 1990). The region is just 1.5° south of the Permanently Open Ocean Zone (Pondaven et al., 2000) with partial sea ice cover only during the winter (Cavalieri and Parkinson, 2008). As this is the time when primary productivity, because of low light levels, is already low (Bianchi et al., 1997) nutrient utilization and export production should be minimally affected by regional changes in sea ice cover. Thus, we surmise that this region reflects primarily open ocean processes such as the position of the Polar Frontal Zone and the strength of the Antarctic Divergence. Primary productivity in the Antarctic Zone of the modern Southern Ocean is low despite relatively high nitrate concentrations (Boyd et al., 2000) (e.g., Fig. 1B). This is because Fe limits productivity in this region today. During glacial intervals, however, increases in aeolian input of Fe resulted in enhanced nitrate consumption as recorded by higher bulk sediment δ15N values (Brzezinski et al., 2002). However, in this region, opal export fluxes decreased during the LGM suggesting that nitrate limited productivity at that time (Kumar et al., 1995; Francois et al., 1997; Brzezinski et al., 2002). Francois et al. (1997) argue that both reduced export production and high sedimentary bulk δ15N values can be explained by enhanced water column stratification limiting the amount of nutrients upwelled into the photic zone and hence primary and export productivity. Site 745 is unique because it is located south of the Polar Front, yet north of the region significantly affected by seasonal sea-ice cover, and also contains a complete late Miocene through Pleistocene sedimentary sequence. Although not multiply cored, recovery rate was near 100%, and shipboard biostratigraphy as well as magnetostratigraphy do not show the presence of significant gaps in the sedimentary record (Barron and Larsen et al., 1989). Late Miocene through early

Fig. 1. Location of Ocean Drilling Program Leg 119 Site 745 (59°35.71′S, 85°51.60′E, 4082 m water depth) in the Antarctic Zone of the Southern Ocean with respect to mean annual sea surface temperatures (A) and nitrate (B) (Levitus and Boyer, 1994; Conkright et al., 1994, respectively). Also shown are Sites 1090 (43°S, 20°E, Etourneau et al., 2009) and 1096 (66.5°S, 77°W, Sigman et al., 1999) for which published δ15N records exists that at least partially span the study interval. Site 745 lies to the south of the modern day Polar Front in this region and to the north of dense winter sea ice extent. The graph was generated using the interactive climate data library of the Lamont Doherty Earth Observatory.

K. Billups et al. / Global and Planetary Change 100 (2013) 353–361

Pleistocene sediments consist primarily of biogenic silica (primarily diatoms on average 40%), terrigenous material (clays, ~60% and ice rafted debris) with very little carbonate (b1%) (Ehrmann et al., 1991). Long-term variations in the relative amount of biogenic silica versus clay are attributed to large-scale climatic changes with generally higher silica content during warm intervals and higher terrestrial sediments during cold intervals (Ehrmann et al., 1991). Data from Ehrmann et al. (1991) also show that the percent opal in the sediments increase to above 50% at about 2 Ma, consistent with increases in opal at other Southern Ocean sites reflective of the establishment of the moderntype silica belt in this region (Cortese et al., 2004). 3. Material and methods 3.1. Sampling and analytical methods Samples were taken at 150 cm intervals from Site 745B. Approximately 1 g of sediment was freeze dried, ground with a mortar and pestle into a fine powder, and homogenized. About 90–100 mg of the thus prepared bulk sediment was weighed out into tin cups. Approximately 10 mg of powdered tin was added to each sample to aid with combustion given the low organic matter concentrations. Samples were then analyzed at the Stroud Water Research Center on a Thermo-Finnigan DeltaPlus XP Isotope Ratio Mass Spectrometer (IRMS) with Conflo III interfaced in continuous flow mode to a Costech 4010 CHNS-O Elemental Analyzer (EA). The instrument was modified in two ways to reduce background and improve sensitivity of nitrogen isotope analysis. First, an automated six-port valve fitted with an ascarite trap was installed in between the EA and IRMS in order to trap CO2 during elution of the N2 peak and therefore reduce formation of CO in the source and variation in mass 28, 29, and 30 backgrounds, as done by Brooks et al. (2003). Second, the IRMS was set to high-gain amplification to increase signal as done by Osburn and St-Jean (2007). Measured values were normalized to the δ 15N scale using the 2-point calibration approach advocated by Coplen et al. (2006) and with USGS40 (− 4.52 ± 0.12 per mil) and USGS41 (47.57 ± 0.22 per mil) as international references (Qi et al., 2003). The analytical precision of δ 15N was assessed for each sample based on propagating regression errors from calibration curves to final normalized values. This analytical uncertainty ranged from ±0.2 to ±0.5 per mil. The standard error of duplicate measurements of 18 of the samples from this study had a mean value of 0.29 per mil and a standard deviation of 0.22 per mil. All data are archived at the NOAA World Data Center-A for Paleoclimate. 3.2. Age model Site 745 contains a high quality paleomagnetic record measured at high resolution (30–50 cm for discrete samples and 10 cm whole round core) (Sakai and Keating, 1991). However, only few precise age–depth control points are provided by Sakai and Kealing: the depth of the Brunhes/Matayma boundary (42.6 mbsf), and the top and bottom of the Gauss (112.8–132.8 mbsf) and the Gilbert (133.3–186.2 mbsf). There appear to be uncertainties surrounding the depth of the oldest normal polarity chron within the Gilbert (C3An.4n) (Sakai and Keating, 1991), and we are not using it as an age control point. To obtain a higher resolution age model, we hand digitized the mbsf depths for other Mio/Pliocene chron boundaries based on Ehrmann et al. (1991) who show the high resolution paleomagnetic reversal pattern. We have then updated the ages for late Miocene/early Pleistocene chron boundaries to those of Berggren et al. (1995) (Fig. 2, Table 1). The thus derived depth–age control points vary smoothly downcore in a linear fashion between 42 and 92.3 mbsf (r= 0.993) and describe a cubic spline below 92.3 mbsf (r= 0.998) (Fig. 2A). We use the two curve fits to derive ages for all data points rather than linearly

355

interpolating between control points. The advantage of the curve fit is that it smoothes over rapid fluctuations and abrupt changes in sedimentation rates. The largest of a potential bias occurs during chron C3n.1n at 142 mbsf (~4.18 Ma) resulting in ages that are about 180 kyr too old (roughly twice the sampling interval). As we are focusing on longer-term changes over the past 6 million years, this uncertainty will not affect our data interpretations. The resulting sedimentation rates show two changes, one between 214 and 185 mbsf when sedimentation rates decrease from ~6 cm/kyr to 2 cm/kyr, and at ~92 mbsf when sedimentation rates increase rapidly to ~5 cm/kyr (Fig. 2B). These sedimentation rates yield a temporal resolution of ~30–75 kyr for our 150 cm sampling interval. Potential sources of uncertainties in the age model are introduced by having to estimate the mbsf depth corresponding to the chron boundaries from a figure published by Ehrmann et al. (1991). We estimate that our depth picks are precise to within +/− 10 cm. A relatively minor source of uncertainty is related to the resolution of the magnetic inclination measurement, which is based on discrete sampling every 30–50 cm and whole core pass-through cryogenics every 10 cm (Sakai and Keating, 1991). Depending on sedimentation rates, each of the two potential errors contributes a ~ 2–5 kyr age uncertainty (~4–20 kyr total). This is less than the temporal resolution of this study (~ 30–75 kyr) and does not affect our data interpretations at Site 745. Uncertainties in the age model and the choice of the Berggren et al. (1995) time scale limits comparisons of the Site 745 data to high resolution proxy records from other sites, but comparison with long-term trends is robust.

3.3. Mass accumulation rates Opal MARs in general are used as a proxy for export productivity (e.g., Charles et al., 1991; Froelich et al., 1991). At Site 747, the biogenic silica sediments of the late Miocene through Pliocene contain abundant diatoms throughout (15–80% diatoms, Barron et al., 1989), with diatoms dominating during the Pleistocene. Thus, at Site 745 opal MARs reflect export of organic matter from primary productivity. However, dissolution in the water column and in the sediments limits the degree to which opal MARs reflect primary productivity in surface waters (e.g., Charles et al., 1991). Over the time interval of interest here, which includes a fundamental change in the silica budget of the Southern Ocean at around 2 Ma (Cortese et al., 2004), it cannot be assumed that silica production and/or preservation remained constant. Specifically, we suspect that prior to the establishment of the modern silica belt in the Southern Ocean at ~ 2 Ma, the water column may have been greatly undersaturated with respect to silica. This means that the preservation potential of diatoms during the late Miocene/Pliocene may have been lower than during the Pleistocene and may bias interpretation of opal MARs with respect to export production. Organic carbon MARs are also commonly used to assess changes in surface water productivity. As opposed to opal MARs, which primarily reflect diatom productivity, organic carbon is more directly linked to overall productivity independent of the particular producers, although, organic matter is highly susceptible to respiration in the water column and sediments. In conjunction with the opal MARs, it should be possible, however, to determine robust trends in export production through time. MARs (in g/cm 2/kyr) are calculated by multiplying the fraction of opal (and organic carbon) in each core interval (data from Ehrmann et al., 1991) with the corresponding sedimentation rate (cm/kyr) and the dry bulk density (g/cm 3) (Barron et al., 1989). This means that MARs are dependent on the derivation of sedimentation rates and hence age control. Here we compare the MARs with the percent sediment component to assess the degree to which down-core changes in MARs are potentially driven by sedimentation rates.

356

K. Billups et al. / Global and Planetary Change 100 (2013) 353–361

Fig. 2. Derivation of the age model for Site 745. Left hand column shows the paleomagnetic reversal record based on discrete samples and whole core measurements (Ehrmann et al., 1991; Sakai and Keating, 1991) with chron boundaries labeled after Berggren et al. (1995). The middle panel shows depth–age control points used for the age model (summarized in Table 1), and the right hand panel shows the corresponding sedimentation rates.

4. Overview of late Miocene through early Pleistocene climate change Benthic foraminiferal δ 18O values outline the evolution of climate between the late Miocene and early Pleistocene (Fig. 3A). Relatively high δ 18O values during the late Miocene (6.4–5.3 Ma) reflect the relatively cold conditions associated with expansion of Antarctic ice and

Table 1 Summary of age control points. Chron

MBSF (m)

B85a (myr)

B95b (myr)

C1n C1r.1n C2n C2An.1n C2An.2n C2An.3n C3n.1n C3n.2n C3n.3n C3n.4n C3An.1n C3An.2n

0–42.3 50.64–54.2 91.6–92.3 112.3–123.5 125.5–126.8 128.3–133.2 142.3–145.8 150.0–155.5 161.3–163.5 ? 186.2–196.1 206.1–214.2

0–0.73 0.91–0.98 1.66–1.88 2.47–2.92 2.99–3.08 3.18–3.4 3.88–3.97 4.10–4.24 4.40–4.47 4.57–4.77 5.35–5.53 5.68–5.98

0–0.78 0.99–1.07 1.77–1.95 2.58–3.04 3.11–3.22 3.33–3.58 4.18–4.29 4.48–4.62 4.80–4.89 4.98–5.22 5.89–6.15 6.30–6.58

a b

Berggren et al. (1985). Berggren et al. (1995).

the establishment of the West Antarctic ice sheet (Barker et al., 1998). The early Pliocene warm interval corresponds to an interval of relatively low δ 18O values with respect to the late Miocene (and the late Holocene) beginning at about 5 Ma and ending at ~ 3 Ma. Global sea level estimates vary and are controversial, but may have been 10– 35 m lower than present day implying at the very least the loss of the West Antarctic ice sheet (Kennett and Hodell, 1993; Pollard and DeConto, 2010). The increase in δ 18O values between 3.5 Ma and 2.5 Ma (e.g., Fig. 3A) outlines the ice volume increase and global cooling associated with the beginning of significant expansion of Northern Hemisphere Glaciation (e.g., Mudelsee and Raymo, 2005). Although not a time of large changes in ice volume (in comparison to the last glacial to interglacial cycle, for example), the late Pliocene/ Pleistocene (~ 2 Ma and 1 Ma) is a time of transition toward cooler mean climate (Ravelo et al., 2007) evidenced by continued increase in benthic foraminiferal δ 18O values (e.g., Fig. 3A). At ~ 2 Ma, cooling becomes evident in the Benguela upwelling region (Etourneau et al., 2009) and later, between 1.8 Ma and 1.2 Ma in the Subantarctic Southern Ocean (Martínez-Garcia et al., 2010). In addition, during this interval of time, El Niño-like conditions in the equatorial Pacific give way to a stronger Walker circulation (Ravelo et al., 2007). Ice rafted debris (IRD, all visible grains >2 mm per section, Ehrmann et al., 1991) at Site 745 underscores large-scale climatic trends (Fig. 3B). Counts are highest during the early Pliocene warm

K. Billups et al. / Global and Planetary Change 100 (2013) 353–361

357

Fig. 3. Summary of proxy records from Site 745 (A) Benthic foraminiferal δ18O record (Site 849, Mix et al., 1995), (B) ice rafted debris (Ehrmann et al., 1991), (C) bulk sediment δ15N values, (D) opal mass accumulation rates (black symbols) and percent opal content of the sediments (gray symbols, data from Ehrmann et al., 1991), (E) organic carbon mass accumulation rates (black symbols) and percent sediment content (gray symbols, data from Ehrmann et al., 1991), and (F) linear sedimentation rates derived by linear interpolation between age control points. Crosses in (C) are duplicate δ15N measurements through which the average is plotted. The heavy lines in A–E represent smoothed fit to the data using a polynomial function. The vertical gray bars highlight the δ15N maxima associated with opal MAR minima. The horizontal dashed line in panel C indicates average δ15N values for the Last Glacial Maximum in this latitude band (Francois et al., 1997).

interval reflecting ice sheet instabilities (Ehrmann et al., 1991), a pattern that has been observed at many other Southern Ocean sites (Hodell and Warnke, 1991). IRD counts decrease during the late Pliocene and are essentially zero between 3 and 2.4 Ma. Small increases occur thereafter, but overall, the low abundance of IRD at this latitude after the late Pliocene is consistent with the general stability of ice sheets since then (Ehrmann et al., 1991). In the Southern Ocean, the modern-day silica belt becomes established during the late Pliocene at ~ 2 Ma (Cortese et al., 2004). According to Cortese et al. (2004), this major hydrographic change is ultimately a consequence of the closure of the Central American Seaway, which established the large-scale ocean circulation patterns that characterize the modern ocean. Regardless of the causes, a major shift occurred in the silica budget of the Southern Ocean across the Plio/Pleistocene boundary that potentially affected primary productivity and export production, and thereby atmospheric CO2 levels.

5. Results 5.1. Site 745 bulk sediment proxies Bulk sediment δ 15N values show an overall increase over the study interval (6.5 to 1.6 Ma, Fig. 3C). The increase is not monotonic, it occurs in three steps. Between 6.4 Ma and 4.5 Ma, δ 15N values are generally well within the range of late Holocene values in the 60– 50°S latitude band south of the Polar Front (Francois and Altabet, 1992; Francois et al., 1997; Sigman et al., 1999). At 4.6 Ma values begin to increase and remain relatively high, generally above late Holocene levels between 4.5 and 2.5 Ma. Averaged values begin to increase further at 2.5 Ma and reach distinct maxima centered at 2.1 Ma and again at 1.7 Ma. These δ 15N maxima (5.6 and 5.7 per mil, respectively) are clearly higher than late Holocene values and approach those recorded during the Last Glacial Maximum (~6 per mil, Francois and Altabet, 1992; Francois et al., 1997). The maxima are

358

K. Billups et al. / Global and Planetary Change 100 (2013) 353–361

separated by a pronounced minimum with values as low as 1 per mil at 1.9 Ma. A second such minimum occurs at the top of the section at 1.5 Ma. Both minima are recorded by multiple data points and/or duplicate measurements and are thus robust features in the record (as opposed to the singularly high δ 15N value at 4.5 Ma). Such low bulk sediment δ 15N values are not typical of average Holocene bulk sediment δ 15N data south of the Polar Front (e.g., Francois and Altabet, 1992; Francois et al., 1997). However, a core 5° to the north of Site 747(MD 84-552) displays similarly low values during the Holocene (Sigman et al., 1999). Opal MARs decrease during the late Miocene, remain relatively low (0.5–1 g/cm 2/kyr) during the Pliocene until 3 Ma and then increase again: slowly at first until ~ 2 Ma and then rapidly toward maximum values of about 1.8 g/cm 2/kyr at 1.9 Ma and again at 1.5 Ma (Fig. 3D, black curve). Beginning at 3 Ma distinct variations with a range of 0.7 g/cm 2/kyr punctuate the long-term trend, while variability before 3 Ma is smaller and rather noisy (e.g., defined only by single data points). The long-term trend in the opal MARs essentially follows the sedimentation rate pattern (Fig. 3F). At the finer scale, however, the rapid, step-like increase between 2.0 and 1.9 Ma predates by 1 million years the step-up in sedimentation rates between 1.9 and 1.8 Ma, and is thus not a function of age control. Percent opal in the sediments, a parameter not affected by the instantaneous sedimentation rates (albeit by changes in the relatively proportions of other sediment components), shows very similar pattern both over the long-term and in the superimposed variability (Fig. 3D, gray curve). Percent opal remains relatively constant at 40% before 4 Ma (highlighting the sedimentation rate effect on the opal MARs during the late Miocene), drops to about half of this amount between 4 Ma and 3 Ma, and then increases, slowly at first until 2 Ma, and rapidly thereafter. As in the MAR record, individual fluctuations gain strength after 3 Ma with a distinct maximum at 1.9 Ma, coincident with the δ 15N minimum. Organic carbon MARs and the percent organic carbon in the sediments essentially mimic each other (Fig. 3E). Both record generally high values during the late Miocene decreasing toward a minimum by 3.5 Ma. Organic carbon MARs and percent content are particularly low between 4 and 3.5 Ma, the interval characterized by minimum δ 18O values (e.g., Fig. 3A). Organic carbon MARs and content both increase again beginning at 3.5 Ma and reach a broad maximum centered on 2.4 Ma. The proxies display a pronounced minimum at 2.1 Ma. This minimum is coincident with minima in the opal MAR (and percent, Fig. 3D), as well as the first of the two δ 15N maxima (Fig. 3C). Disregarding single-point peaks, a second organic carbon maximum occurs at ~ 1.9 Ma, coincident with the first of the two δ 15N minima, albeit it is not apparent in the corresponding MAR. 5.2. Southern Ocean δ 15N gradients For the late Plio/Pleistocene interval (3.5–1.7 Ma) the Site 745 δ 15N record adds a spatial perspective to previously published records from the Polar Antarctic Front (Site 1096, Sigman et al., 1999) and the Subantarctic Zone (Site 1090, Etourneau et al., 2009) (Fig. 1). In the modern ocean, or during the late Holocene, δ 15N values increase toward the north across the Polar Front reflecting enhanced nitrate utilization (Sigman et al., 1999). Within the Antarctic Zone, on the other hand, δ 15N values increase toward the south reflecting a decrease in the fractionation factor between nitrate and the assimilated biomass (DiFiore et al., 2010). During the Pliocene a relatively constant offset exists over the long-term in the δ 15N values between these three regions characterized by the lowest values at the southernmost Site 1096 and highest values at the northern most Site 1090 (Fig. 4). With respect to late Holocene, Site 1090 and Site 1096 δ 15N values are both lower while Site 745 δ 15N values are higher. Thus the gradient between the Antarctic and Subantarctic Zone is smaller than during the late Holocene

(reduced from about 3 per mil to 0.8 per mil), while the gradient within the Antarctic Zone is reversed (from ~1 per mil higher values at the more polar site to ~0.6 per mil lower values at the polar site). Lower Pliocene δ 15N values with respect to late Holocene at Sites 1090 and 1096 are consistent with observations at sites from other ocean basins (Etourneau et al., 2009; Filippelli and Flores, 2009). Etourneau et al. (2009) attribute low Pliocene δ 15N values at Site 1090 to differences in the nutrient source regions within the Southern Ocean. Filippelli and Flores (2009) illustrate that an increase in δ 15N values occurred during the Pliocene/Pleistocene transition in the Pacific Ocean. These authors propose that a change in the nitrogen cycle, specifically an overall increase in denitrification, may explain the pattern. Importantly, with its high values relative to late Holocene the δ 15N record at Site 745 opposes the more global pattern highlighting the importance of regional processes. 6. Discussion 6.1. Diatom species A consideration in any geochemical study involves potential species or trophic-related effects. Published records focusing on Southern Ocean diatom species during the last glacial cycle indicate that a diatom species effect is relatively small for δ15N (Des Combes et al., 2008). Furthermore, the one species that tends to have higher δ15N values according to the study by Des Combes et al. (2008) (Eucampia antarctica) is not abundant in the Southern Ocean during the Miocene (Censarek and Gersonde, 2003) and Pliocene, and it is rare during the early Pleistocene (Shipboard Scientific Party, 1999). Although, there are no studies that address potential species-specific isotopic effects beyond the last glacial cycle, we cannot rule out entirely that species effects contribute to the magnitude of the observed δ15N range. Additionally, it has been shown that δ15N values of sinking particles are higher if algae use ammonia rather than nitrate to build biomass (Lourey et al., 2003). These effects would limit quantification of nitrate utilization and nutrient fluxes, but permits interpretation of the records in a more qualitative sense. 6.2. Diagenesis A primary assumption in paleoceanographic studies applying δ 15N values of bulk sediment is that effects of diagenesis are negligible, or rather, constant down-core. Sediment δ 15N values are higher than the δ 15N values of organic matter assimilated at the surface, which reflects to a large degree early diagenesis in the sediments (Robinson et al., 2012). On the regional scale, the offset is consistent and down-core changes in sedimentary δ 15N can be used to assess relative changes in nitrogen isotope fractionation at the sea surface (Altabet and Francois, 1994; Holmes et al., 1999). Although diagenetic effects cannot be ruled out, differential alteration does not control down-core variation in bulk sediment δ 15N records (Robinson et al., 2012). Thus, it seems justified to use bulk sediment δ 15N records to reconstruct aspects of nitrogen cycling (Francois and Altabet, 1992; Francois et al., 1997; Holmes et al., 1999; Sigman et al., 1999; Sigman et al., 2004; Liu et al., 2008; Etourneau et al., 2009). We are certain that the overall trend in the δ 15N record from Site 745 is not a diagenetic one. Diagenesis would increase the δ 15N values down-core reflecting enhanced organic mater remineralization with time of deposition. The δ 15N record, however, decreases down-core in opposition of the direction expected from diagenesis. 6.3. Pliocene warmth Bulk sediment δ 15N values are relatively high, at least as high as during the Holocene and higher at times. As this observation is unique to Site 745 (Sites 1090 and 1096 both having lower values

K. Billups et al. / Global and Planetary Change 100 (2013) 353–361

359

Fig. 4. Comparison of the Site 745 bulk sediment δ15N record with published records from the Subantarctic Zone of the Southern Ocean (Site 1090, Etourneau et al., 2009) and the Polar Antarctic Zone (Site 1096, Sigman et al., 1999). The horizontal lines indicate respective late Holocene δ15N values (Francois et al., 1997).

in agreement with global patterns, Fig. 4), the observation is consistent with a regional process, such as a more southerly position of the Polar Frontal Zone. In the modern ocean, δ 15N values of bulk sediment increase to the north across this hydrographic boundary due to high 15N/ 14N ratios in advected surface waters and in situ nitrate depletion (Francois et al., 1997). Thus, southward retreat of the Polar Frontal Zone should cause a southward expansion of more 15N enriched, nitrate depleted surface waters. Numerous studies have invoked a southerly position, or southward expansion, of the Polar Frontal Zone during the early Pliocene warm interval to explain the relatively warm conditions in the Southern Ocean (e.g., Abelmann et al., 1990; Froelich et al., 1991; Hodell and Warnke, 1991; Hodell and Venz-Curtis, 1992; Whitehead and Bohaty, 2003). It is thus conceivable that a southward retreat of the Polar Front from the late Miocene to the early Pliocene as climate warmed caused a poleward expansion of more nitrate depleted, 15N enriched surface waters more typically associated with the Subantarctic Zone. The early Pliocene warm interval precedes the establishment of the modern-day silica belt in the Southern Ocean indicating that the silica budget was fundamentally different than today (Cortese et al., 2004). Consequently, phytoplankton communities may not have been as diatom dominated as during the Pleistocene (Giuliani et al., 2011). In this case opal MARs would underestimate export of organic matter into sediments. The generally low opal MARs, however, are accompanied by a decrease in organic carbon content and MARs (and lower nitrogen content). Taken together, low opal and carbon MARs during the early Pliocene support that Southern Ocean export production was low during this interval of relative climatic warmth. We suggest that reduced primary productivity and export production in the Southern Ocean, either from opal producers or others, may have contributed to on average higher CO2 levels during the early Pliocene in comparison to preindustrial levels (Pagani et al., 2010). 6.4. Late Pliocene/Pleistocene cooling At Site 745, late Pliocene/Pleistocene cooling beginning at ~ 3.5 Ma (e.g., Fig. 3A) is accompanied by increases in opal content and opal MARs both also beginning to fluctuate with enhanced variability with respect to the earlier time interval. At 1.9 Ma, opal MARs (and content) reach a first pronounced maximum with opal content > 80% (Fig. 3D). The Site 745 opal record is in line with trends recorded at sites located nearer to and to the north of the Polar Frontal Zone (e.g., Cortese et al., 2004). Enhanced export production deduced from the opal MAR records is supported by concurrent increases in organic matter burial (Fig. 3E). Global cooling may thus be related to CO2 draw-down due, at least in part, to enhanced Southern Ocean diatom productivity.

In contrast to the opal MAR records from sites located further south in the Polar Antarctic Zone near the Antarctic Peninsula (Sites 1096 and 1095, Hillenbrand and Fütterer, 2001; Sigman et al., 2004), Site 745 does not show a decrease at 2.7 Ma neither abrupt nor gradual. The difference may be related to the different hydrographic regimes at the two sites with Site 745 representing open ocean conditions less affected by sea ice than the other sites (Hillenbrand and Fütterer, 2001; Hillenbrand and Cortese, 2006). The shift in export productivity occurred in tandem with shifts at sites to the north at 2 Ma, rather than in concert with the sites to the south at 2.7 Ma. This suggests that Site 745 was located in a similar hydrographic regime as the sites to the north, which would be consistent with a more southerly location of the Polar Frontal Zone. The Polar Front had moved north by 2.4 Ma in response to intensified glaciation and Antarctic sea ice expansion (Hodell and Warnke, 1991). And the modern-like frontal system became established at around 2 Ma (Cortese et al., 2004). Against this backdrop, a pronounced inverse relationship between the bulk sediment δ 15N values and opal MARs (and content) begins at ~ 2.1 Ma. The range in δ 15N values of about 4 per mil exceeds the average range between the Last Glacial Maximum and the late Holocene recorded in sediments in the 50–60°S latitude band (Francois et al., 1997), but it is similar to the range observed in an Indian Ocean sector core just to the north of Site 745 (Sigman et al., 1999). Thus, the potential changes in nitrate availability occurring at 2.2–1.5 Ma are only a little bit smaller than the range recorded during one of the largest climatic changes of the Pleistocene. The δ15N maxima and opal/organic carbon MARs minima at 2.1 Ma could thus reflect the effects of a more stratified water column on nitrate utilization and export productivity. However, for this mechanism to be registered by the δ15N values, the Antarctic Zone must have become limited by nitrate. By extension, the apparent effect of nitrate utilization raising bulk sediment δ 15N values at 2.1 Ma would necessitate that other nutrients such as Fe were not limiting at this particular time. There is evidence from Site 1090 for enhanced Fe accumulation the Subantarctic Zone of the Southern Ocean beginning at 2.7 Ma (Martínez-Garcia et al., 2011). This Fe record also shows a more pronounced Fe increase at 2.1 Ma, and a smaller increase at around 1.7 Ma, the times of the two δ15N maxima at Site 745. Thus it is conceivable that the input of Fe at these times caused the upper water column to become nitrate limited (as export production is at a minimum at these times, e.g., Fig. 3D). Why no response was registered at 2.7 Ma may reflect that nitrate was not limiting at this time as discussed by Sigman et al. (2004). To account for nitrate limitation as inferred from high δ15N values, Fe input to the upper waters would need to be decoupled from nutrient (i.e., nitrate) input. Fe could potentially come from dust originating in

360

K. Billups et al. / Global and Planetary Change 100 (2013) 353–361

the Patagonian desert, having been carried to the study site by westerly winds, as has been argued by Kumar et al. (1995) to explain increased Fe content in Southern Ocean sediments from the Last Glacial Maximum. This hypothesis necessitates cooler and drier climatic conditions in the southern regions of South America at these times. Alternatively, the preformed Fe content in deep waters reaching the surface could have been higher at these times without changing upwelling rates and thus vertical nutrient fluxes (Latimer and Filippelli, 2001). In this hypothesis, Fe becomes enriched in deep waters due to enhanced weathering of continental margins during sea level low stands. In any case, the δ 15N proxy record is at insufficient resolution to establish concrete links with other climate proxy records to potentially differentiate between these mechanisms. 7. Summary and conclusions We have constructed bulk sediment δ 15N records in the Antarctic Zone of the Southern Ocean in order to investigate whether or not changes in nutrient utilization occurred that may provide evidence for hydrographic changes related to late Miocene through Pliocene climate change. The δ 15N records are accompanied by published records of opal and organic carbon MARs to constrain the potential role of export productivity on atmospheric CO2 concentrations. Our results indicate an overall increase in δ 15N values to above late Holocene levels between the late Miocene through late Pliocene (~ 6.5–2 Ma) while export production has a distinct minimum during the early Pliocene warm interval. Our results from the bulk sediment δ 15N records are consistent with a more southerly position of the Polar Frontal Zone during the early Pliocene interval of relative warmth. Export productivity was low at this time, which may have contributed to relatively high atmospheric CO2 levels. At 2.7 Ma, the time of the onset of significant Northern Hemisphere glaciation accompanied by abrupt decrease in opal MARs at Site 1096, Site 745 does not display any changes in δ 15N values or export productivity. The Site 745 record is thus more consistent with changes at sites nearer the modern-day Polar Frontal Zone than the more poleward located Site 1096. This observation may reflect that the Polar Frontal Zone was still to the South of Site 745 at this time. During the late Pliocene/early Pleistocene climate transition between 2.1 and 1.6 Ma, δ15N values and export production display large and antiphased variations. After the northward movement of the Polar Front between 2.5 and 2.1 Ma, and the development of the modern day silica belt by 2 Ma, the δ 15N data indicate that the region may have become sensitive to changes in water column stratification potentially contributing to fluctuations in atmospheric CO2 concentrations. Acknowledgments K. Billups and A. Aufdenkampe acknowledge the donors of the Petroleum Research Fund (PRF # 49945-ND2), administered by the American Chemical Society ACS. This research used samples provided by the Ocean Drilling Program (ODP). ODP is sponsored by the U.S. National Science Foundation (NSF) and participating countries under the management of the Joint Oceanographic Institutions (JOI) Inc. We particularly thank the reviewers, Gabriel Filippelli and anonymous, for their constructive suggestions that have helped us to improve the manuscript. References Abelmann, A., Gersonde, R., Speiss, V., 1990. Pliocene–Pleistocene paleoceanography of the Weddell Sea-siliceous microfossil evidence. In: Bleil, U., Thiede, J. (Eds.), Geologic History of the Polar Ocean: Arctic versus Antarctic. Kluwer, Amsterdam, pp. 729–759. Altabet, M., Francois, R., 1994. Sedimentary nitrogen isotopic ratio as a recorder for surface ocean nitrate utilization. Global Biogeochemical Cycles 8, 103–116.

Barker, P.F., Camerlenghi, A., Acton, G.D., 1998. Proc. ODP, Init. Repts., 178. Ocean Drilling Program, Texas A&M University, College Station, TX 77845-9547, U.S.A. Barron, J., Larsen, B., et al., 1989. Proc. ODP, Init. Repts., 119. Ocean Drilling Program, College Station, TX. Berggren, J.A., Kent, D.V., Flynn, J.J., 1985. Cenozoic geochronology. Geological Society of America Bulletin 96, 1407–1418. Berggren, J.A., Kent, D.V., Swisher, C.C., Aubry, M.-P., 1995. A revised Cenozoic geochronology and chronostratigraphy. Geochronology Time Scales and Global Stratigraphic Correlation. SEPM. Special Publication 54, 129–212. Bianchi, M., Feliatra, F., Tréguer, P., Vincendeau, M.A., Morvan, J., 1997. Nitrification rates, ammonium and nitrate distribution in upper layers of the water column and in sediments of the Indian sector of the Southern Ocean. Deep Sea Research Part II: Tropical Studies in Oceanography 44, 1017–1032. Boyd, P.W., Watson, A., Law, C.S., Abraham, E.R., Trull, T., Murdich, R., Bakker, C.E., Bowie, A.R., Buesseler, K., Chang, H., 2000. A mesoscale phytoplankton bloom in the polar Southern Ocean stimulated by iron fertilization. Nature 407, 695–702. Brooks, P.D., Geilmann, H., Werner, R.A., Brand, W.A., 2003. Improved precision of coupled δ13C and δ15N measurements from single samples using an elemental analyzer/isotope ratio mass spectrometer combination with a post-column sixport valve and selective CO2 trapping; improved halide robustness of the combustion reactor using CeO2. Rapid Communications in Mass Spectrometry 17, 1924–1926. Brzezinski, A., Pride, C.J., Franck, V.M., Sigman, D.M., Sarmiento, J.L., Matsumoto, K., Gruber, N., Rau, G.H., Coale, K.H., 2002. A switch from Si(OH)4 to NO3-depletion in the glacial Southern Ocean. Geophysical Research Letters 29, 1564. Cavalieri, D.J., Parkinson, C.L., 2008. Antarctic sea ice variability and trends. Journal of Geophysical Research 113, 1979–2006. Censarek, B., Gersonde, R., 2003. Data report: relative abundance and stratigraphic ranges of selected diatoms from Miocene sections at ODP Sites 689, 690, 1088, and 1092 (Atlantic Sector of the Southern Ocean). Proceedings of the Ocean Drilling Program — Scientific Results 177. Charles, C.D., Froelich, P.N., Zibello, M.A., Mortlock, R.A., Morley, J.J., 1991. Biogenic opal in Southern Ocean sediments over the last 450,000 years: implications for surface water chemistry and circulation. Paleoceanography 6, 697–728. Conkright, M.E., Levitus, S., Boyer, T.P., 1994. World Ocean Atlas 1994 Volume 1: Nutrients. NOAA Atlas NESDIS 1. U.S. Department of Commerce, Washington, D.C. 150 pp. Coplen, T.B., Brand, W.A., Gehre, M., Gröning, M., Meijer, H.A.J., Toman, B., Verkouteren, R.M., 2006. New guidelines for δ13C measurements. Analytical Chemistry 78, 2439–2441. Cortese, G., Gersonde, R., Hillenbrand, C., Kuhn, D.G., 2004. Opal sedimentation shift in the world ocean over the last 15 Myr. Earth and Planetary Science Letters 224, 509. Des Combes, H.J., Esper, O., De La Rocha, C.L., Abelmann, A., Gersonde, R., Yam, R., Shemesh, A., 2008. Diatom δ13C, δ15N, and C/N since the Last Glacial Maximum in the Southern Ocean: potential impact of species composition. Paleoceanography 23. DiFiore, P.J., Sigman, D.M., Karsh, K.L., Trull, T.W., Dunbar, R.B., Robinson, R.S., 2010. Poleward decrease in the isotope effect of nitrate assimilation across the Southern Ocean. Geophysical Research Letters 37. Ehrmann, W.U., Grobe, H., Fütterer, D.K., 1991. Late Miocene to Holocene glacial history of East Antarctica as revealed by sediments from sites 745 and 746. Proceedings of the Ocean Drilling Program — Scientific Results 119, 239–260. Etourneau, J., Martinez, P., Blanz, T., Schneider, R., 2009. Plio-Pleistocene variability of upwelling activity, productivity, and nutrient cycling in the Benguela region. Geology 37, 871–874. Filippelli, G., Flores, J.-A., 2009. From the warm Pliocene to the cold Pleistocene: a tale of two oceans. Geology 37, 959–960. Francois, R.M., Altabet, B.L., Burckle, L.H., 1992. Glacial to interglacial changes in surface nitrate utilization in the Indian Sector of the Southern Ocean as recorded by sediment δ15N. Paleoceanography 7, 589–606. Francois, R., Altabet, M.A., Yu, E.-F., Sigman, D.M., Bacon, M.P., Frank, M., Bohrmann, G., Bareille, G., Labeyrie, L.D., 1997. Contribution of Southern Ocean surface-water stratification to low atmospheric CO2 concentrations during the last glacial period. Nature 389, 929–935. Froelich, P.N., Malone, D.A., Hodell, P.F., Ciesielski, D.A., Warnke, F., Westall, E.A., Hailwood, D., Nobes, C.J., Fenner, J., Mienert, C., Mwenifumbo, J., Muller, D.W., 1991. Biogenic opal and carbonate accumulation rates in the Subantarctic South Atlantic: the late Neogene of the Meteor Rise Site 704. Proceedings of the Ocean Drilling Program — Scientific Results 114, 515–532. Giuliani, S., Capotondi, L., Maffioli, P., Langone, L., Giglio, F., Yam, R., Frignani, M., Ravaioli, M., 2011. Paleoenvironmental changes in the Pacific sector of the Southern Ocean (Antarctica) during the past 2.6 Ma. Global and Planetary Change 77, 34–48. Hillenbrand, C.-D., Cortese, G., 2006. Polar stratification: a critical view from the Southern Ocean. Palaeogeography, Palaeoclimatology, Palaeoecology 242, 240–252. Hillenbrand, C.-D., Fütterer, D.K., 2001. Neogene to Quaternary deposition of opal on the Continental Rise West of the Antarctic Peninsula, ODP Leg 178, Sites 1095, 1096, and 1101. Proceedings of the Ocean Drilling Program — Scientific Results 178 (215.2001). Hodell, D.A., Venz-Curtis, K.A., 1992. Toward a high resolution stable isotopic record of the Southern Ocean during the Pliocene–Pleistocene (4.8–1.2 Ma), in the Antarctic paleoenvironment: a perspective on global change. Antarctic Research Series 56, 265–310. Hodell, D., Warnke, D., 1991. Climate evolution of the Southern Ocean during the Pliocene Epoch from 4.8 to 2.6 million years ago. Quaternary Science Reviews 10, 205–214. Holmes, M.E., Eichner, C., Struck, U., Wefer, G., 1999. Reconstruction of surface ocean nitrate utilization using stable nitrogen isotopes in sinking particles and sediments. In: Fischer, I., Wefer, G. (Eds.), Use of Proxies in Paleoceanography: Examples from the South Atlantic. Springer-Verlag, Berlin, pp. 447–468.

K. Billups et al. / Global and Planetary Change 100 (2013) 353–361 Horn, M.G., Beucher, C.P., Robinson, R.S., Brzezinski, M.A., 2011. Southern Ocean nitrogen and silicon dynamics during the last deglaciation. Earth and Planetary Science Letters 310, 334–339. Kennett, J.P., Hodell, D.A., 1993. Evidence for relative climate stability of Antarctica during the early Pliocene: a marine perspective. Geografiska Annaler 75, 205–220. Kumar, N., Anderson, R.F., Mortlock, R.A., Froelich, P.N., Kubik, P., Dittrich-Hannen, B., Suter, M., 1995. Increased biological productivity in the glacial Southern Ocean. Nature 378, 675–680. Latimer, J.C., Filippelli, G.M., 2001. Terrigenous input and paleoproductivity in the Southern Ocean. Paleoceanography 16, 6. http://dx.doi.org/10.1029/2000PA000586. Levitus, S., Boyer, T.P., 1994. Temperature, NOAA Atlas NESDIS 4. World Ocean Atlas 1994, vol. 4. Nat. Oceanic and Atmos. Admin., Silver Spring, MD . (129 pp.). Liu, Z., Altabet, M.A., Herbert, T.D., 2008. Polio-Pleistocene DE nitrification in the eastern Tropical North Pacific: intensification at 2.1 Ma. Geochemistry, Geophysics, Geosystems 9. Lourey, M.J., Trull, T.W., Sigman, D.M., 2003. Sensitivity of δ15N of nitrate, surface suspended and deep sinking particulate nitrogen to seasonal nitrate depletion in the Southern Ocean. Global Biogeochemical Cycles 17. Mariotti, A., Germon, J.C., Hubert, P., Kaiser, P., Letolle, R., Tardieux, A., Tardieux, P., 1981. Experimental determination of nitrogen kinetic isotope fractionation: some principles; illustration for the denitrification and nitrification processes. Plant and Soil 62, 413–430. Martínez-Garcia, A., Rosell-Melé, A., McClymont, E.L., Gersonde, R., Haug, G.H., 2010. Subpolar link to the emergence of the modern equatorial Pacific cold tongue. Science 328, 1550–1553. Martínez-Garcia, A., Rosell-Mele', A., Jaccard, S.L., Geibert, W., Sigman, D.M., Haug, G.H., 2011. Southern Ocean dust–climate coupling over the past four million years. Nature 476, 312. Mix, A.C., Pisias, N.G., Rugh, W., Wilson, J., Morey, A., Hagelberg, T.K., 1995. Benthic foraminifer stable isotope record from Site 849 (0–5 Ma): local and global climate changes. Proceedings of the Ocean Drilling Program — Scientific Results 138, 371–412. Montoya, J.P., McCarthy, J.J., 1995. Isotopic fractionation during nitrate uptake by marine phytoplankton grown in continuous culture. Journal of Plankton Research 17, 439–464. Mudelsee, M., Raymo, M.E., 2005. Slow dynamics of the Northern Hemisphere glaciations. Paleoceanography 20, PA4022. Osburn, C.L., St-Jean, G., 2007. The use of wet chemical oxidation with highamplification isotope ratio mass spectrometry (WCO-IRMS) to measure stable isotope values of dissolved organic carbon in seawater. Limnology and Oceanography: Methods 5, 296–308. Pagani, M., Liu, Z., LaRiviere, J., Ravelo, A.C., 2010. High Earth-system climate sensitivity determined from Pliocene carbon dioxide concentrations. Nature Geoscience 2, 27–30.

361

Pickard, G.L., Emery, W.J., 1990. Descriptive Physical Oceanography. Pergamon Press (320 pp.). Pollard, D., DeConto, R.M., 2010. Modelling West Antarctic ice sheet growth and collapse through the past five million years. Nature 458. Pondaven, P., Ragueneau, O., TreÂguer, P., Hauvespre, A., Dezileau, L., Reyss, J.L., 2000. Resolving the ‘opal paradox’ in the Southern Ocean. Nature 405, 168–172. Qi, H., Coplen, T.B., Geilmann, H., Brand, W.A., Bohlke, J.K., 2003. Two new organic reference materials for δ13C and δ15N measurements and a new value for the δ13C of NBS 22 oil. Rapid Communications in Mass Spectrometry 17, 2483–2487. http:// dx.doi.org/10.1002/rcm.1219. Ravelo, A.C., Billups, K., Dekens, P.S., Herbert, T.D., Lawrence, K.T., 2007. Onto the ice ages: proxy evidence for the onset of Northern Hemisphere glaciation. In: Williams, M., Haywood, A.M., Gregory, J., Schmidt, D. (Eds.), Deep-time Perspectives on Climate Change: Marrying the Signal from Computer Models and Biological Proxies: The Geological Society of London and The Micropalaeontological Society. Special Publications, pp. 563–573. Robinson, R.S., Kienast, M., Luiza Albuquerque, A., Altabet, M., Contreras, S., De Pol Holz, R., Dubois, N., Francois, R., Galbraith, E., Hsu, T.-C., Ivanochko, T., Jaccard, S., Kao, S.-J., Kiefer, T., Kienast, S., Lehmann, M.F., Martinez, P., McCarthy, M., Möbius, J., Pedersen, T., Quan, T.M., Ryabenko, E., Schmittner, A., Schneider, R., Schneider-Mor, A., Shigemitsu, M., Sinclair, D., Somes, C., Studer, A., Thunell, R., Yang, J.-Y., 2012. A review of nitrogen isotopic alteration in marine sediments. Paleoceanography 27, PA4203. http://dx.doi.org/10.1029/2012PA002321. Sakai, H., Keating, B., 1991. Paleomagnetism of Leg 119-Holes 737A, 738C, 742A, 745B, and 746A. Proceedings of the Ocean Drilling Program — Scientific Results 119, 751–770. Shipboard Scientific Party, 1999. Leg 177. Proc. of the ODP. Init. Repts. Sigman, D., Altabet, M., Francois, R., McCorkle, D., Gaillard, J.-F., 1999. The isotopic composition of diatom-bound nitrogen in Southern Ocean sediments. Paleoceanography 14, 118–134. Sigman, D.M., Jaccard, S.L., Haug, G.H., 2004. Polar ocean stratification in a cold climate. Nature 428, 59–63. Toggweiler, R.J., Sarmiento, J.L., 1985. Glacial to interglacial changes in atmospheric carbon dioxide: the role of ocean surface water in high latitudes. Geophysical Monograph 32, 163–184. Watson, A.J., Garabato, A.C.N., 2006. The role of Southern Ocean mixing and upwelling in glacial–interglacial atmospheric CO2 change. Tellus 58B, 73–87. Whitehead, J.M., Bohaty, S.M., 2003. Pliocene summer sea surface temperature reconstruction using silicoflagellates from Southern Ocean ODP Site 1165. Paleoceanography 18. http://dx.doi.org/10.1029/2002PA000829. Wu, J., Calvert, S.E., Wong, C.S., 1997. Nitrogen isotope variations in the northeast subarctic Pacific: Relationships to nitrate utilization and trophic structure. Deep Sea Research Part 1 (44), 287–314.