EarthRef.org Reference Database (ERR) -- Bard 2001

To a certain degree, all sea surface temperature (SST) proxies suffer from uncertainty due to signal preservation problems and to perturbations by other environmental variables. However, a multiproxy reconstruction is justified because most biases should, in principle, cancel out as they are specific to the biological groups and/or to the analyzed chemical species. Such a multiproxy strategy has proven crucial for calibrating these proxies and is also useful in assessing downcore profiles. Until now there have been only a dozen deep-sea cores where it is possible to compare directly alkenone-based SSTs with other paleotemperature records. From this still limited data set the broad picture is of general agreement for the amplitude of changes at low and middle latitudes. The few observed discrepancies suggest that alkenone-based temperatures are more comparable to those based on Mg/Ca ratios than those obtained through statistical analysis of the foraminiferal distribution. More systematic differences are noted at higher latitudes, where alkenones indicate warmer temperatures during the glacial period than transfer functions based on foraminifera or diatoms. Several reasons have been invoked to resolve the mismatch, but more ground truthing work is needed before deciding which temperature changes are the most accurate.