METHODS

Carbonate carbon contents were determined using a Coulometrics 5011 carbon dioxide coulometer as in shipboard measurements. Depending on carbonate contents of the samples, 20-200 mg of freeze-dried, ground sediment was reacted with 10% HCl. The liberated CO₂ was back titrated to a colorimetric endpoint, yielding the inorganic carbon content (IC [wt%]). By X-ray diffraction measurements of each sample, the bulk mineralogy was determined and calcite verified as the carbonate phase. Therefore, the percentage of carbonate is calculated as CaCO₃ (wt%) = IC (wt%) x 8.333. By measuring shipboard samples, replicates of samples, and calcite and calcite-quartz mixture standards, it was ensured that there is no offset compared to shipboard CaCO₃ data. The accuracy of the IC (wt%) determination is 0.05%, and the relative standard derivation of CaCO₃ contents of replicates is 0.6%.

Quantification of long-chain alkenone concentrations for the determination of the used the following method. Total extractable organic matter of 2-g freeze-dried sediment with added recovery standards was obtained by ultrasonic extraction for 15 min with 12.5 mL of methanol, then with 12.5 mL methanol:dichloromethane (1:1), and finally with 12.5 mL dichloromethane. After each extraction step, samples were centrifuged, and the solvent was collected by pipette. The three lipid extracts were combined, and salts were removed from the dichloromethane layer by addition of 10 mL H₂O_bidest. After concentration by rotary evaporation and removal of remaining solvent under a gentle stream of nitrogen, the total extract was redissolved in hexane and separated on a silica gel column in three compound classes by elution with 4-mL hexane (fraction 1), 4-mL ethylacetate:hexane (3:97) (fraction 2 containing alkenones), and 3-mL dichloromethane followed by 2-mL methanol (fraction 3). After saponification of the second fraction, lipids were redissolved in 100-µL hexane containing an injection standard. The sample (1-4 µL) was injected in solvent vent mode onto an HP 6890 gas chromatograph fitted with a 50-m glass capillary column (Ultra 2, ID 0.2 mm, film thickness = 0.11 µm) coupled to an HP 5973 mass selective detector. Helium was used as carrier gas (2 bar constant). The temperature program was 50° for 2.5 min, 250° at 20°/min, 290° at 3°/min, 310° at 0.5°/min, 320° at 10°/min, and isothermal at 320°C for 10 min. Quantification was performed on total ion current signals. Results of duplicate analyses (N = 20) indicate that the standard deviation of the measurements is 0.005 units. Replicate extractions (N = 4) of a pooled homogenized sample resulted in standard deviations up to 0.015 units. The is calculated according to the equation = (37:2)/(37:2 + 37:3) and converted into sea-surface temperatures according to Prahl et al. (1988) as SST (°C) = ( - 0.039)/0.034.