ALBERT

Description: Despite a clear correlation of alkenone unsaturation and sea surface temperatures (SST) throughout most parts of the ocean, scatter of the regression for various calibration equations has been shown to increase significantly at low SSTs. In this study, we combine previously published (n = 101) and new (n = 51) surface sediment data from the northern North Atlantic to constrain uncertainties of alkenone paleothermometry at low SSTs and to discuss possible sources of the increased scatter in the regression. The correlation between alkenone unsaturation and SSTs is strongest, in particular at the cold end (SSTs 〈 10°C), when the tetra-unsaturated alkenones (C37:4) are included in the unsaturation index (expressed as inline image) and regressed against spring-summer temperature. Surface ocean salinity and sea ice cover are not correlated with inline image per se. However, samples located in regions of permanent winter sea ice cover exhibit a significant warm bias. Deviation from the linear regression is posited to be related to a number of additional non-exclusive factors, such as advection of allochthonous material, local temperature stratification, and uncertainty in the absolute age of surface sediment samples assumed to be equivalent to modern conditions. We conclude that alkenone unsaturation allows accurate reconstruction of SST records from many regions of the North Atlantic if the factors confounding alkenone paleothermometry detailed here can be excluded.