Hydrographical changes of the southern Indian Ocean over the last 230 kyr, is reconstructed using a 17‐m‐long sediment core (MD 88 770; 46°01′S 96°28′E, 3290m). The oxygen and carbon isotopic composition of planktonic (N. pachyderma sinistra and G. bulloides) and benthic (Cibicidoides wuellerstorfi, Epistominella exigua, and Melonis barleeanum) foraminifera have been analysed. Changes in sea surface temperatures (SST) are calculated using diatom and foraminiferal transfer functions. A new core top calibration for the Southern Ocean allows an extension of the method developed in the North Atlantic to estimate paleosalinities (Duplessy et al., 1991). The age scale is built using accelerator mass spectrometry (AMS) 14C dating of N. pachyderma s. for the last 35 kyr, and an astronomical age scale beyond. Changes in surface temperature and salinity clearly lead (by 3 to 7 kyr) deep water variations. Thus changes in deep water circulation are not the cause of the early response of the surface Southern Ocean to climatic changes. We suggest that the early warming and cooling of the Southern Ocean result from at least two processes acting in different orbital bands and latitudes: (1) seasonality modulated by obliquity affects the high‐latitude ocean surface albedo (sea ice coverage) and heat transfer to and from the atmosphere; (2) low‐latitude insolation modulated by precession influences directly the atmosphere dynamic and related precipitation/ evaporation changes, which may significantly change heat transfer to the high southern latitudes, through their control on latitudinal distribution of the major frontal zones and on the conditions of intermediate and deep water formation.

Simple box model calculations are used to simulate the oceanic circulation during the last glacial maximum (LGM). These experiments show that the main features of the δ13C and Δ14C distributions and of the lysocline depth may be explained by a circulation pattern very different from the modem one. Intermediate and upper deep waters were produced in the North Atlantic Ocean, whereas deep waters of Subantarctic Mode type, forming at the northern edge of the Subantarctic convergence, invaded the main oceanic basins. The Southern Ocean, mainly self ventilated, had a reduced deep component that flew southward along the East Pacific Ridge and the Australian west cost. The thermodynamic fractionation that occurs during air‐sea exchange has only contributed slightly to the glacial deep δ13C distribution through surface water temperature variations.

(Paleoceanography, 11(1), 57-76, 1996), the rightmost colunto (phases, in kiloyears) in Table 5 is wrong. The results of the phase calculation, in radian and degree, were divided by a wrong constant for the equivalence in kiloyears. Transformation from radian to kiloyear should be done by the following formula: phase (kyr) = phase (rad) x period (kyr) / (2x3.1416). All calculations were done with the obliquity period (--43 kyr), instead 9f using the values corresponding to the three spectral bands (106, 43, and 21 kyr). The corrected Table 5 In the third paragraph on page 69, the third sentence should read "In the precession band, the foraminiferal SSST record lags the June 15 summer insolation by 1 kyr (17ø)."In the fourth paragraph on page 69, the first sentence should read "The diatom transfer function has a significantly different response, considering the + 1-to 2-kyr phase uncertainty calculated by the SPECMAP method (Table 5)

No abstract

No abstract