RESULTS

The oxygen isotope stratigraphy at Hole 1087A is based on the continuous record of C. wuellerstorfi, allowing the identification of MISs 1 through 13, which spans the last 500 k.y. The timescale calibration of the isotopic stages corresponds to the orbitally derived scales of Imbrie et al., (1984), Martinson et al. (1987), and Bassinot et al. (1994). The age-control points are listed in Table T2. Linear interpolation was applied between adjacent age control-points, assuming a constant sedimentation rate within the time interval. The average sedimentation rate is estimated at 2.7 cm/k.y. for the entire 13-m interval. In fact, sedimentation rates are irregular, with higher rates (3-4 cm/k.y.) during MISs 1, 2, 4, 5, 8, 10, 11, 13 and lower rates (1-1.9 cm/k.y.) during MISs 3, 7, 9, and 12.

The ¹⁸O record of Cibicides is given both as a function of depth (Fig. F1A) and of age (Fig. F1B). The lowest ¹⁸O values are measured during isotopic Event 5.5 (2.19) and MIS 11 (2.37), which characterize the warmest periods. The highest ¹⁸O values (4.27) are measured during MIS 12, which marks the coldest stage. At Site 704, which is near the Polar Front Zone in the South Atlantic (i.e., in the source area of AAIW), the coldest stage is also MIS 12, but there is no significant difference between the interglacial stages for the last 450 k.y. except the long warm period during MIS 11 (Hodell, 1993; Hodell et al., 2000). If one assumes that the G-IG salinity changes were negligible in the source area of AAIW and subtracts 1.6 and 1.2% for the maximum global ice effect during Terminations V and II, respectively (McManus et al., 1999), the maximum difference of ¹⁸O values between MIS 12 and 11 and MIS 6 and 5 will equal 0.3 and 0.6, which correspond to a maximum temperature change of ~1.5° and 2.5°C in deep waters (using an ¹⁸O fractionation of -0.22/°C). Otherwise, this G-IG temperature variation would be less if part of the ¹⁸O variation is caused by increased freshwater dilution during interglacials. The difference of ¹⁸O values between the LGM and the Holocene is 1.6, indicating that during Termination I the temperature increase was <2°C in the deep waters at Site 1087.

The ¹⁸O records of the three species of planktonic foraminifers display similar variations, although the amplitudes of the difference between glacial and interglacial values (¹⁸O G-IG) differ for each species (Fig. F2). For the last four G-IG cycles, the ¹⁸O G-IG amplitude decreases from 2.1 in G. bulloides to 1.5 in G. inflata and to 1.2 in G. ruber. A comparison of the G. bulloides and G. inflata records shows in fact that the glacial values are very close for the two species, whereas the low interglacial values are much lower for G. bulloides than for G. inflata; this indicates that the temperature change in surface waters occurred mostly during interglacial stages.

As was the case for the benthic foraminifers, the highest ¹⁸O values of the planktonic foraminifers are found during MIS 12, which corresponds to the coldest stage of the last 500 k.y. (Fig. F2). The difference in ¹⁸O values between LGM and MIS 12 ranges from 1.2 (G. ruber) to 0.4 (G. bulloides), with G. inflata showing an intermediate value of 0.7. Subtracting the global ice effect difference of 0.4 between LGM and MIS 12 (McManus et al., 1999) translates into an MIS 12 cooling of surface waters compared to the LGM situation of ~4° and 1.5°C, as recorded by G. ruber and G. inflata, respectively.

The ¹³C record of C. wuellerstorfi shows a pattern different from the G-IG cycles but linked to large oscillations much higher than the 100-k.y. eccentricity period (Fig. F3). Two intervals of lower ¹³C values occur between 20-160 ka and 300-420 ka and are separated by an interval of higher ¹³C values during the period 160-300 ka. If the ¹³C values of benthic foraminifers are considered to be a proxy of ventilation and nutrient concentration at depth, the AAIW was well ventilated and nutrient poor prior to 420 ka and from 160 to 300 ka, whereas it was less ventilated and nutrient rich during both 20-160 ka and 300-420 ka intervals. Furthermore, ¹³C values tend to decrease at the IG-G transitions, a pattern particularly well expressed at the MIS 13/12, MIS 11/10, and MIS 7/6 transitions. This ¹³C decrease suggests that inputs of nutrients occurred during the beginning of cool conditions at the IG-G transitions and that the migration toward the equator of the Polar Front Zone was associated with a relatively lower rate of AAIW ventilation.

Similarly the planktonic foraminiferal ¹³C records do not display clear evidence of G-IG cyclicity (Fig. F4). Common to the three planktonic taxa is a large positive ¹³C oscillation spanning the time interval 260-425 ka; this increase is the highest (2.1) for G. bulloides, intermediate (1.2) for G. inflata, and the lowest (0.8) for G. ruber. In terms of productivity levels, this positive ¹³C excursion identifies a highly productive period that coincides with the global mid-Brunhes event (Jansen et al., 1986) known for its very high level of pelagic carbonate production during MISs 7, 9, and 11 (Hodell, 1993). A shorter period of ¹³C enrichment between 80 and 100 ka also suggests high productivity levels during the late MIS 5.