Sea surface temperature (SST) and salinity of the Western Pacific Warm Pool (WPWP) reflect global climate effects such as the El Niño-Southern Oscillation phenomenon. However, reconstructions of past changes in the WPWP from the geologic record vary depending on the specific proxy record used. Here we develop a multiproxy record of the last deglaciation from a radiocarbon-dated sediment core (MD97-2138) retrieved in the heart of the WPWP. SST reconstructions for the past 30,000 years based on planktonic foraminiferal Mg/Ca (Globigerinoides ruber and Globigerinoides sacculifer), alkenone unsaturation index, and foraminiferal transfer functions differ notably. Mg/Ca-based SST estimates from the surface dwelling species G. ruber in MD97-2138 indicate a larger surface cooling (3° ± 0.6°C) during the Last Glacial Maximum (LGM) than inferred from Mg/Ca ratios in G. sacculifer (2.3° ± 0.7°C), statistical transfer functions based on planktonic foraminiferal species assemblages, and UK37' (1°-2°C). These estimates are consistent with estimates from other WPWP cores, thereby suggesting that the discrepancy is due to proxy compatibility rather than differences in cores qualitity. Postdepositional dissolution above the lysocline might have altered the Mg/Ca-based temperature estimates in our site, but this effect is insufficient to resolve discrepancies between Mg/Ca in G. ruber and the other proxies. We suggest that the lower estimates obtained from Mg/Ca in G. sacculifer, faunal transfer functions, and Uk37' might reflect subsurface temperature changes rather than strict surface estimates. Accounting for potential artefacts, including dissolution and bioturbation, we suggest that the glacial WPWP SST was about 2.5° ± 0.7°C cooler than during the Holocene, whereas the subsurface/upper thermocline temperature change was only about 1.8° ± 0.7°C. Interpreting variations in d18OSW in terms of salinity changes suggests a possibly slight decrease in surface salinity at the site of MD97-2138 during the LGM. Though LGM freshening in MD97-2138 is not robust to postdeposition dissolution effects, this inferred freshening appears to be a general feature of the western equatorial Pacific.

The 14C ages were converted to calendar ages using the INTCAL98 calibration (1998) for the last 18,000 years and the polynomial fit of Bard [1998] for ages prior to the period covered by the INTCAL98 study.

Supplement to: de Garidel-Thoron, Thibault; Rosenthal, Yair; Beaufort, Luc; Bard, Edouard; Sonzogni, Corinne; Mix, Alan C (2007): A multiproxy assessment of the western equatorial Pacific hydrography during the last 30 kyr. Paleoceanography, 22(3), PA3204