Evaluation of reconstructed sea surface temperatures based on [formula omitted] from sediment surface samples of the North Pacific

Abstract

The alkenone unsaturation index (U37k′) as proxy for sea surface temperature (SST) is an important tool in paleoclimatology for reconstructing past ocean temperature variability. Typically, U37k′ recorded in marine surface sediments shows a linear correlation with modern mean annual SST. However, in high-latitude oceanic regions, such as the subpolar Pacific, U37k′-based SSTs do overestimate the mean annual temperature by up to 6 °C, potentially leading to obscured paleoclimatic information drawn from stratigraphic U37k′-records. The reason for this “warm bias” is still not well understood. Here, we present a compilation of 97 sediment surface samples from Multicores collected in the Bering Sea, the Okhotsk Sea and the North Pacific to evaluate the alkenone-temperature proxy against observational data from the North Pacific. Sediment surface samples were analysed for alkenones and the derived U37k′-indices converted to water temperatures using different calibration equations established in the literature. U37k′-based SSTs were then compared to instrumental SST data, as well as modern alkenone flux data from sediment traps in the North Pacific. Our results confirm that most U37k′-based SSTs from the subpolar Pacific are 2–6 °C too warm compared to instrumental mean annual SSTs for calibrations applied. However, with an uncertainty at the level of ±1.5 °C or less reconstructed SSTs fit quite well to modern autumn temperatures north of the Subarctic Front (SAF), when maximum export flux of alkenones to the seafloor is indicated by sediment trap data. South of the SAF, reconstructed SSTs largely mimic the modern mean annual SST signal with an uncertainty of ±1.5 °C or less, which is likely due to the attenuation of seasonality and longer growth season of coccolithophorids according to sediment trap data. Our study further demonstrates that U37k′, when seasonality in alkenone production and export are known and considered, is able to provide reasonable estimates of SSTs in modern high-latitude ocean settings. We conduct a case study using available alkenone time-series derived from a sediment core collected from the south-western Okhotsk Sea to better understand the potential effect of seasonality in alkenone production on stratigraphic U37k′-record in the subpolar Pacific. The case study from the Okhotsk Sea indicates that even a small shift in seasonality may lead to strongly biased SSTs with broader regional implications for paleoclimate reconstructions in high-latitude ocean settings.