OXYGEN ISOTOPE RECORD

The lightest oxygen isotope values in the Site 1006 record were found in isotopic Stages 1, 5, 9, and 11. The Holocene values are very similar to the core-top G. ruber values as reported by Slowey and Curry (1987) from near Site 1006. The ¹⁸O values of the planktonic foraminiferal calcite from the Holocene section are in equilibrium with modern temperature and salinity values of the surface waters using the ¹⁸O-salinity relationship for the North Atlantic (Fig. 5) as given by Ganssen and Kroon (in press). Our observation that the most depleted values are seen in the Holocene section is unusual for deep-sea records; normally isotopic Stage 5e typically shows the lightest values, which makes the Holocene unique in the Bahamian area. It seems that this trend toward lighter values during the periods of maximum deglaciations began around isotopic Stage 11.

In general, the amplitude changes in the oxygen isotope record are large, especially in the top 17 m (Fig. 2, Fig. 3) because of the light values of ¹⁸O recorded during sea-level highstands since isotopic Stage 11 and maximum values of isotopic Stages 6 and 8 during sea-level lowstands. The Holocene/last glacial maximum transition shows a maximum isotopic shift of 2.15. The maximum interglacial-glacial amplitude difference is used here rather than an average (Slowey and Curry, 1987). We prefer the maximum amplitude ¹⁸O difference because bioturbation may have smoothed the record. About 1.1-1.3 can be ascribed to the ice volume/sea-level effect (Berger and Gardner, 1975; Chapell and Shackleton, 1986; Fairbanks, 1989), which leaves a residual ¹⁸O component of ~1 that can be attributed to salinity and temperature variations. Unfortunately, we cannot discriminate between these two parameters because independent proxies have not as yet been produced on this core. However, temperature reconstructions from other records in the western equatorial Atlantic have been a useful aid. The temperature proxies (Sr/Ca and ¹⁸O) from the Barbados coral record suggest a temperature shift of 5°C from the last glacial maximum to the Holocene (Guilderson et al., 1994), although it is not clear how unambiguous the interpretation of such a record is. Nonetheless it would fit with a residual component of ~1 in the Hole 1006A record (0.21°C, Epstein et al., 1953). This result means that salinity changes were small at the isotopic Stage 2/1 transition and had no significant effect on the ¹⁸O record. On the other hand, other proxies such as estimates from planktonic foraminiferal assemblages and alkenone analysis show relatively little cooling (1°-2°C) in the west equatorial Atlantic (CLIMAP, 1981; McIntyre et al., 1989; Sikes and Keigwin, 1994, Billups and Spero, 1996). For each 0.45 increase in ¹⁸O (Standard Mean Ocean Water [SMOW]), there is an increase of 1 practical salinity units (psu) in salinity (Fig. 5; Ganssen and Kroon, in press). This would imply a salinity effect of ~1 psu between the last glacial maximum and Holocene low-latitude Atlantic Ocean assuming a 2°C change, a result also found by Broecker (1989). Thus, the glacial Atlantic Ocean was saltier than today in this scenario. Unfortunately, it is not known which temperature proxy is more correct; this leaves us with the two different scenarios or a combination of the two, although in the discussion we argue that the temperature change based on coral data is probably more likely.

The isotopic shifts across the other transitions since isotopic Stage 12 are larger than the isotopic Stage 2/1 shift. The largest shifts occur from isotopic Stage 9 to 8 and from Stage 7 to 6. The maximum change of ~3 was measured from isotopic Stage 9 to 8 (Fig. 3), which could potentially imply a temperature drop of 9°C (assuming no salinity change) since the temperature optimum of isotopic Stage 9. However, some diagenetic effects may have altered the original isotope values of the foraminiferal shells in the samples from the glacial Stage 8 and 6 sediments. The glacial sediments are much coarser in this part of the record. Coarse-grained sediments facilitate diagenesis (Rendle et al., Chap. 6, this volume).

The trend toward lighter values over the course of the Quaternary can best be explained by a temperature shift (see "Discussion" section, this chapter). The trend toward lighter values was also seen in the oxygen isotope record of Deep Sea Drilling Project Hole 502B farther to the south in the Caribbean (Prell, 1982), although amplitude changes were slightly smaller because another species, Globigerinoides sacculifer, was used to generate the isotope record. The records from Holes 502B and 1006A suggest that the warming is potentially characteristic for the Caribbean and Bahamian area.

The ¹⁸O values of the Hole 1006A interglacial periods prior to isotopic Stage 11 are on average 0.75 lower than the values of the interglacial periods after Stage 12, or ~3.5°C cooler, assuming no salinity change. The interglacial/glacial variability prior to isotope Stage 11 is large compared to the changes at Site 806 (Fig. 6). Maximum amplitude changes at Site 1006 are ~1.5, implying large temperature shifts on a glacial/interglacial basis unless salinity changes play a major role.