We examined the response
of both alkenone concentrations, a proxy for haptophyte production and
preservation, and alkenone paleotemperature estimates across two glacial
terminations. There is considerably more context in which to interpret the
record of the last deglaciation (MIS 2-1 transition) because of greater
supporting information on sea-level changes (Fairbanks, 1989), high-frequency
global climate oscillations such as the Younger Dryas Event (Fairbanks, 1989),
and previous studies from the Cariaco Basin (Hughen et al., 1996; Lin et al.,
1997). The time scale for the results presented here comes from the isotope
stratigraphy of Peterson et al. (Chap.
4, this volume) based on the planktonic foraminifer G.
ruber. Because the G.
ruber record clearly differs in detail from the global SPECMAP
isotopic stack, we caution that the time scale may be somewhat less accurate
than a chronology based on benthic 
18O
data representative of global ice-volume changes. We also note that the
high-frequency variations seen in many intervals of both 18O
and Uk´37
records suggest that both time series are still undersampled and that it would
be premature to compare isotopic and alkenone data too closely.
Although both alkenone concentrations and paleotemperatures display substantial variations over the last full glacial cycle in the Cariaco Basin, the alkenone concentration measurements resemble the oxygen isotopic record more than they resemble the paleotemperature data (Fig. 4). Alkenone concentrations vary positively both with ice volume and with total organic carbon (TOC) (Haug et al., 1998). The alkenone concentration record differs from the TOC record in that the amounts of long-chained ketones are higher on a relative basis during MIS 5 than are the TOC contents. The median alkenone quantity (4.4 µg/g) over the entire record is about one order of magnitude higher than is typical of pelagic sediments (cf. Prahl et al., 1989; Lyle et al., 1992) and is slightly higher than the average concentration of long-chained ketones at Santa Barbara Basin, another silled and frequently anoxic basin, over the same time interval (Herbert et al., 1995).
Alkenone paleotemperatures range from a high of 27.3°C during marine isotope Substage 5E to lows of 21.0°-22.0°C during brief intervals across the MIS 2-1 transition and during MIS 3-5 (Fig. 5). The mean temperature estimated for the entire record is 24.8°C, similar to modern core-top temperatures. MIS 3 and 4 and Substages 5A-5C contain high-frequency swings in estimated temperatures, with some estimates exceeding modern values juxtaposed with intervals as cold as 21°C. Both the LGM and Stage 6 intervals show very little cooling from Holocene and Substage 5E conditions, respectively. The average late Holocene temperature at Hole 1002C, which is nearly identical to our box core composite modern temperature of 25.5°C, is 0.5°-1°C warmer than our reconstructed temperature at the LGM (Fig. 5). Similar patterns occurred from MIS 5 to 6. The small degree of glacial cooling inferred from alkenones is consistent with the dominance of G. ruber during the late glacial record observed by Peterson et al. (1991) but perhaps anomalous in light of oxygen isotope data, as discussed below.
Large, short-lived temperature anomalies seem to have occurred at glacial-interglacial transitions (Fig. 5). The ~3°C cooling coincides with the first signs of high alkenone abundance and the first laminated facies of the Holocene (Shipboard Scientific Party, 1997). It significantly precedes the interval correlated with Younger Dryas time, which does display alkenone temperatures ~1°C cooler than average Holocene values. The remainder of the Holocene shows modest temperature variations around the Holocene optimum of 25.8°C between 4 and 6 ka. A cooling of similar magnitude and duration identified at the MIS 5-6 transition occurs at a position stratigraphically identical to the well-defined temperature minimum reconstructed here at 13-14 ka (Fig. 4).
Very low alkenone concentrations consistently occur during MIS 2 and 4 (Fig. 4). The detailed record presented in Figure 5 shows that rather constant values were deposited from ~15-25 ka, despite the presence of laminated intervals (Shipboard Scientific Party, 1997). This interval also displays little variation in alkenone unsaturation index, an indication that the mixed oxic/anoxic conditions did not promote significant diagenetic alterations in the Uk´37 index.
