1 Introduction

The causes for Pleistocene ice ages are usually attributed to the astronomical forcing (Hays et al. 1976), which has been debated extensively. Antarctic ice cores have revealed low and high concentration of greenhouse gases during glacial and interglacial periods, respectively (Petit et al. 1999), which suggests that they too may be part of the cause. However, there is no general agreement among various records on the phase relationship between δ18O of planktic foraminifera and CO2 from the northern and southern hemispheres, especially during the deglaciation (Shakun et al. 2012). Hence, the CO2 role in driving the glacial and interglacial cycles has not been understood clearly. For example, based on proxy data, it is suggested that CO2 is the main driver of the ice ages (Luthi et al. 2008) on the one hand; on the other hand, interpretations such as (i) CO2 being a feedback from warming (Alley and Clark 1999), and (ii) CO2 is a consequence and may not be the cause of past climate change (Weaver et al. 1998), have also been made.

More recently, the new concept has been added to address the triggering force of deglaciation that the CO2 initiates the rise in temperatures (Shakun et al. 2012). The period from 11 to 19 kyr include the climate transition from the last glacial to the Holocene. The documentation from ice cores of polar region and other climate archives suggest that the climate variation patterns differ between Antarctica and the surrounding Southern Ocean and the northern hemisphere during this transition period (e.g., Pedro et al. 2011). The steady state of Antarctic deglacial warming reaches a maximum at ~14.5 ka, followed by cooler conditions during the Antarctic Cold Reversal (ACR). Conversely, Greenland records show two rapid warming phases, one at the onset of Dansgaard Oeschger Event1 (Bolling–Allerod interstadial) and the other at Holocene, and these two warm phases are separated by the Younger Dryas (YD) cold event. The sequence of these events was related to the thermal bipolar see-saw in the two hemispheres (Stenni et al. 2010).

Recent review by Shakun et al. (2012) reveals that the lead and lag of atmospheric CO2 with the rise of SST in different regions exhibit different patterns. Most of these records are from the north Atlantic and a few are from Pacific Ocean. No records are available from the Southern Ocean of the Indian sector; in this context, the present study is aimed at analyzing the CO2 and deglaciation records from the Southern Indian Ocean in order to understand whether this region exhibits lead or lag as compared to the CO2 variations in the Antarctica Ice Core record. The Southern Indian Ocean is primarily interesting to address this issue because it is highly influenced by the seasonal productivity variations driven by the monsoons besides the Antarctic Intermediate Water (AAIW) and Antarctic circum-polar current.

2 Material and methods

Core S2 was collected (42°0.258′S; 48°0600′E; water depth: 3305 m with a gravity corer) during the Austral Summer Cruise of 2006 onboard the Akademik Boris Petrov from the Indian Sector Southern Ocean (figure 1). Core location was south of southwest Indian Ridge, i.e., on the flanks of the ridge system. Core was sub-sampled at every 1 cm interval up to 100 cm depth and further down the core was sub-sampled at every 2 cm interval. Overall the time resolution between samples varies from 31 to 450 years. Five grams of dry sediment was soaked in distilled water overnight and washed through a sieve size of 125 µm. The residue was oven dried at a constant temperature of 60°C. Planktonic foraminifera Orbulina universa in the size range of 300–350 µm was analyzed for oxygen and carbon isotopes. O. universa is the most abundant surface dwelling planktonic foraminifera occurred in this core, hence it was chosen for oxygen and carbon isotope analyses to reconstruct the surface hydrographic changes. All foraminifera tests were cleaned ultrasonically in methanol for removal of adhering fine clay particles. The δ18O and δ13C measurements were carried out on a DeltaPlus Advantage Isotope Ratio Mass Spectrometer (IRMS) coupled with Kiel-IV automatic carbonate device at CSIR-National Geophysical Research Institute, Hyderabad, India. The foraminifera tests were treated with 100% H3PO4 at ~70°C in a vacuum system and the extracted CO2 was analyzed by IRMS. Isotopic compositions are expressed in δ notation as per mil deviation from VPDB standard. Analytical precision was better than 0.10‰ for δ18O and 0.05‰ for δ13C. Repeated measurements of international reference standards NBS-19 and NBS-18 were carried to achieve the calibration to the VPDB standard (Ahmad et al. 2012). The chronology of the core was established by AMS 14C dates from the AMS facility at the University of Arizona, Tucson, USA. 14C dates were calibrated using CALIB Program 6.0 (Stuvier et al. 1993) with reservoir age correction of 550 years (table 1). Bayesian analysis of the age-model is performed with the program ‘Bacon’ and final age-depth relationship is obtained (Blaauw and Christen 2011), which including all of the radiocarbon dates, are shown in table 1. Bacon methodology has an advantage over the linear interpolation between the age tying points as this approach is based on controlling sediment accumulation rates using a gamma autoregressive semi-parametric model with an arbitrary numbers of subdivisions along the sub-sampled intervals of the core.

Figure 1
figure 1

Bathymetric map of Southern Ocean of Indian Sector showing major tectonics of Southwest Indian Ridge and its flanks, current pattern and core location.

Full size image
Table 1 Details of AMS 14C dates obtained for the Core S2 to establish its chronology. Delta R= 150 years (overall reservoir correction: 550 yrs). 14C ages are calibrated to calendar years by CALIB Program 6.0. (Stuiver and Reimer 1993).
Full size table

3 Oceanography

In the present core location, sea surface temperature (SST) varies from 9.5 to 13.5°C and salinity varies from 34.3 to 34.6 PSU (Levitus and Boyer 1994) (figure 2). This region is highly influenced by summer monsoon with lowest temperature during Austral winters and warmest temperatures during Austral summers (figure 2). Salinity changes in this region are mainly related to ocean precipitation, whereas evaporation more or less remains the same during both winter and summer (Levitus and Boyer 1994). During Austral summers, more productivity is caused due to the deepening of the mixed layer driven by the summer monsoon winds (Wyrtki 1973).

Figure 2
figure 2

Sea surface temperature and salinity variations at the core location (Levitus and Boyer 1994). 3–4°C in SST difference between the Austral Summer and winter with a minor seasonal salinity changes at the core location; therefore, seasonal SST dominates the hydrographic changes in the Indian sector of the Southern Ocean.

Full size image

This region is characterized by two surface water masses of southern tropical and subtropical. The former water mass lays northwards parallel 30–32°S latitude to a depth of 500–600 m. The temperature of this water mass varies 16–27°C and salinity from 34.8 to 35.3‰. Whereas subtropical water mass lies between 30°S and 45–50°S (Subantarctic Front), and varies in temperature from 17 to 24°C and in salinity from 35.4 to 35.8‰ (Levitus and Boyer 1994). Water circulation over the ridges rotates extensively east and southwards from Madagascar Island, and shift of warm water from north is caused by the southern branch South Equatorial Current, Mozambique Current and its continuation as Cape Agulhas current. Larger part of the Cape Agulhas current turns to south and east, and forms the Agulhas Retroflection. One part of this current returns to the Cape Agulhas current system, and the other moves to east. This chain of the subtropical rotation called Southern Indian Ocean Current carries the water mass to the east with a speed of 0.6–1.0 knots. The Agulhas Retroflection and South Indian Ocean Current confluence with the Antarctic Circumpolar Current in the southern part of the area by moving with a speed of 0.4–0.8 knots in November–March, while during May–September with a speed of up to 0.5–1.3 knots. As a result of confluence of these three currents, the sub-Antarctic zone extending from the surface down to depths 80–1200 m is formed. In general, this circulation extends down to 200 m (Shcherbachev et al. 1989). Local gyres are formed over the seamounts of the ridges, and subject to the synoptic changes in water movement. The intense upwelling area is higher during May–September than in during November–March, and as a result, the active layer gets enriched with nutrients and oxygen (Gebruk et al. 1997).

4 Oxygen and carbon isotope ratios in planktonic foraminifera

Oxygen isotope ratios in planktonic foraminifera are controlled by the water temperature and isotopic values of water in which they calcified (Erez and Luz 1983). Therefore, oxygen isotope ratios in planktonic foraminifera have been used to reconstruct the temperatures over the years (Emiliani 1955). Also the evaporation and precipitation influences the oxygen isotope values of planktonic foraminifera hence δ18O planktonic foraminifera along the magnesium/calcium derived sea surface temperatures have been used extensively to reconstruct monsoon variability in the Indian Ocean (e.g., Govil and Naidu 2010).

Carbon isotopic ratios (13C/12C) of planktonic foraminifera represent δ13C of total dissolved inorganic carbon (Williams et al. 1977). Surface waters are generally enriched in 13C compared to the deeper subsurface water due to photosynthesis uses 12C preferentially in the formation of organic matter. Thus the export flux is enriched in 12C and 13C tends to be left behind. This process of enrichment depends on carbon fixation into organic matter and is limited by the supply of nitrate and phosphate, which also get incorporated into the organic matter. Hence δ13C of planktonic foraminifera has been used to reconstruct the atmosphere/sea water CO2 exchange, biological productivity and nutrient cycling in surface waters (e.g., Broecker and Peng 1982).

5 Results

Chronology of the core was based on AMS 14C ages calibrated to calendar years by using reservoir age of 550 years, the sedimentation rate varies 8–32 cm/kyr (figure 3). Time resolution between the sample intervals varies from 31 years till 400 years BP, decadal resolution is documented up to 3 ka and centennial resolution up to 8 ka and multi-centennial up to 26 ka. The δ18O values of planktonic foraminifera (Orbulina universa, denoted by δ18Oc) vary from 1.5 to 2.5‰, documenting a glacial to Holocene shift of 1.0‰ (figure 4). A pronounced high δ18Oc between 10 and 8.5 ka and greater δ18Oc fluctuating during Late Holocene was documented (figure 4). Strikingly, the δ18Oc pattern during deglaciation more or less mimics the δ18O of the Antarctic EPICA ice core (figure 4). The initiation of warming as evidenced by the depleting δ18O in the core S2 also coincides very well with Antarctica warming around 18 ka. Comparison of δ18Oc of core S2 with data from the GISP2 ice core clearly indicates asynchrony between the deglaciation and the initiation of warming (figure 4). However, δ18O record of the Southern Indian Ocean (Core S2) clearly shows an asynchrony with both Antarctic and GISP2 cores pointing the hydrography of this region during the Holocene was entirely controlled by the local changes, suggesting absence of any link with either of the high latitude records.

Figure 3
figure 3

Depth vs. ages of Core S2 derived from AMS 14C dates.

Full size image
Figure 4
figure 4

(a) δ18Oc variations in Core S2 (blue) and δ18O of Ice, EPICA Dome C (red) (from Stenni et al. 2004) representing that a simultaneous depletion trend of δ18Oc in Core S2 and δ18O of Ice in EPICA dome indicates a synchrony between these two records. (b) δ18O variations in Core S2 (blue), and δ18O of Ice, GISP2 (red) (from Grootes et al. 1993). Comparison of δ18Oc of S2 core with GISP2 δ18O clearly indicates that asynchrony between these two data sets during deglaciation, which is shown in yellow and green shades.

Full size image

Comparison of δ18Oc with atmospheric CO2 record shows that the rise of the atmospheric CO2 started from 18 kyr, which coincides with the start of depleting δ18O in the core S2 (figure 5). Further, the patterns of δ18O and CO2 during deglaciation show a striking similarity during the last glacial period and during deglaciation. During the Holocene, δ18Oc exhibits rapid fluctuations, on the contrary atmospheric CO2 concentrations document a gradual rise.

Figure 5
figure 5

δ18Oc (blue line) and δ13C (green line) variations in Core S2, and atmospheric CO2 variations (red line) documented in an Antarctic ice core by Monin et al. (2001). Note a distinct synchrony during deglaciation between the atmospheric CO2 and δ18Oc of Core S2 suggests that the degassing in Southern Ocean plays an important role in rising atmospheric CO2 particularly within deglaciation period shown with green shade.

Full size image

6 Discussion

6.1 Sequence of deglaciation in the Southern Ocean

The climate records from Ice cores of the Greenland and Antarctica document asynchronous temperature variations on millennial time scales during the last glacial period (EPICA Community Members 2006). The transition from glacial to interglacial period exhibited markedly different warm conditions in both the hemispheres, and the pattern is attributed to the thermal bipolar see-saw (Stocker et al. 2003). Distinct variations in the Antarctic records point to the differences in the climate evolution of the Indo-Pacific and Atlantic sectors of Antarctica has been reported recently (Stenni et al. 2010). δ18Oc record of our core S2 from the Indian Sector Southern Ocean shows the initiation of gradual deglacial warming from 18 ka, which is in phase with the Antarctic δ18O record (figure 4). Similar observation was made on the SST records of several cores from the northern Indian Ocean tropics (e.g., Naidu et al. 2010), indicating that the entire Indian Ocean deglaciation occurs around 18 ka.

The warming during the transition period, i.e., from glacial to interglacial conditions, was distinctly different in the two hemispheres. In Greenland, an increase of 6‰ δ18O is noticed from 12 to 10 kyr, whereas in Antarctica ~1‰ increase is noticed in δ18Oc during the same period, which suggests that warming is quite rapid in the northern hemisphere as compared to southern hemisphere (figure 4). This difference in the rate of warming between two hemispheres was attributed to the difference in the ratio of land to ocean between the two hemispheres. Two step deglaciation warming is documented in the Core S2 (figure 4), and ~0.5‰ in δ18Oc shifts between 18–14 and 12–10 ka are noticed. The present core location is highly influenced by seasonal SST change, while the seasonal salinity variations are small. Therefore, we attribute the shift in δ18Oc during the above-mentioned events primarily to SST change. The ~0.5‰ shift corresponds to ~2°C rise (Epstein et al. 1953) within the deglaciation period. The spatial variability in the global pattern of deglacial warming reflected in the leads of the Antarctic temperatures over the global temperature. Compilation of temperature stacks for the Northern and Southern Hemispheres suggests that the magnitudes of deglacial warming in the two hemispheric stacks are nearly identical (Pedro et al. 2011). The stack of each hemisphere also shows two-step warming as observed in global stack and the CO2 record.

6.2 Synchrony between CO2 and δ18O records

The concentration of atmospheric CO2 has been steadily increasing since the beginning of industrialization. By investigating past natural CO2 variations, the information on feedbacks of the carbon cycle and climate and also the possible impact of atmospheric CO2 on the climate system can be obtained. The transition from the Last Glacial Maximum (LGM) to the Holocene, during which CO2 increased by ~40%, is a key period to evaluate three issues: (i) role of CO2 on deglaciation, (ii) the spatial and temporal variation of CO2 during this period, and (iii) how the CO2 variation corresponds to temperature changes during the deglaciation.

The Vostok results suggest the important role played by the Southern Ocean in regulating the glacial–interglacial CO2 changes (Petit et al. 1999). The role of Southern Ocean is confirmed by measurements obtained for shorter time intervals from the Taylor Dome in the last glaciation (Indermuhle et al. 2000). The starting point of CO2 increase around 18 ka, occurred before any substantial warming in the northern hemisphere, is consistent with the present view of the role of the Southern Hemisphere in causing the CO2 increase. Comparison of δ18Oc record of Core S2 with the atmospheric CO2 record (figure 5) during the deglaciation reveals that the rise of atmospheric CO2 and start of 18Oc depletion and enrichment of δ13C began at ~18 ka. Further, both CO2 and δ18O exhibit similar patterns of fluctuations from 18 to 11 ka. Rate of change in warming around 16.7 ka is conspicuously noticed in the Antarctic Ice Core and our Core S2 records (figure 5); similar change is also observed in an alkenone sea surface temperature record near the Chilean coast (Lamy et al. 2007). Conversely, the alkenone SST record in a marine core from the southwest Pacific shows an uninterrupted deglacial warming (Pahnke and Sachs 2006). Therefore, marine and ice core sequences depict a different rate of warming in the South Chilean coast/south Atlantic and the Indo/South Pacific from around 16 ka. The slowdown of warming at EPICA Dronning Maud Land-Vastok Dome F (EDML-DF) is synchronous with the deceleration of the CO2 rise (Monin et al. 2001), weaker AMOC (McManus et al. 2004) and the distinct change in the strength of Asian monsoon (Broecker et al. 2010). However, the pattern of δ18Oc fluctuations from 11 to 1 ka in Core S2 does not show any correspondence with CO2 record, and this could be due to evaporation and precipitation changes during the Holocene.

6.3 Synchrony in the global context

Recent studies of the deglaciation show strong correlation between times of minima in AMOC and maxima in CO2 release, consistent with temperature difference between northern and southern hemispheres (Shakun et al. 2012), suggesting that a change in AMOC may have also contributed to CO2 degassing from the deep Southern Ocean. This might have happened through its influence on the extent of Southern Ocean sea ice (Skinner et al. 2010), which has been recently confirmed by Fogwill et al. (2020), the position of southern westerlies (Toggweiter et al. 2006), and the efficiency of the biological pump (Schmitter et al. 2008). A near synchronous see-saw response is seen from the high northern latitudes to the mid-southern latitudes, whereas strong Antarctic warming and the increase in CO2 concentration lag the AMOC change (Barker et al. 2009). The present study shows that the Indian sector of the Southern Ocean exhibits synchronous deglacial warming and similar deglacial events as documented in the Antarctic Ice δ18O record, which corresponds to the rise of atmospheric CO2 concentrations. Recently, Parrenin et al. (2013) have reported a similar synchronous change of atmospheric CO2 and Antarctica temperatures during the last deglaciation. This evidence lends support to the hypothesis that the internal heat redistribution related to AMOC explains the lead of Antarctic temperature over CO2 in a few regions while global temperature was in phase or slightly lagged CO2.

7 Summary and conclusion

Detailed analyses of a sediment core from the Indian sector of the Southern Ocean revealed that the pattern of deglaciation in the Southern Ocean mimics the Antarctic pattern. Warming in the Southern Ocean and CO2 rise, both are strongly coupled throughout the deglacial period. Synchrony between the δ18Oc record of Southern Indian Ocean, Antarctic ice core δ18O record and the rise of atmospheric CO2 during the last deglaciation signifies that degassing in Southern Ocean plays an important role in rising atmospheric CO2 during the last deglaciation. Better resolved records with independent estimates of SST are needed from the Southern Oceans to address the issues raised in this study.