Methodology:

The data was generated using samples from International Ocean Discovery Program Site U1343 (Leg 323) in the eastern Bering Sea. The samples were freeze dried at -45°C and 0.2 mbar for 48 hours using a Thermo Savant Modulyo D freeze drier and an Edwards K4 Modulyo freeze drier at Plymouth and Cardiff University, respectively. The highly branched isoprenoid (HBI) lipids were extracted from 3g freeze dried sediment as described in Belt et al. (2012). Additionally sulphur removal using tetrabutylammonium sulphite reagent (TBA) (Cabedo-Sanz and Belt 2015) was performed for all samples prior to silica column chromatographic purification of the total organic extracts. The samples were measured on the gas chromatography mass spectrometer (Agilent 7890A GC coupled to a 5975 series mass selective detector fitted with an Agilent HP-5ms column) at Plymouth University using the operating conditions specified in Belt et al. 2012. The identification of individual lipids was based on their characteristic retention time and mass spectra and quantification was achieved by integrating the peak area of selected ions (m/z 350 (IP25); 346 (HBI III); 470 (brassicasterol)) in comparison to the peak area of the internal standards added to each sample (Belt et al. 2012). Quantification of individual lipids also considers an instrumental response factor obtained from known concentrations of biomarker lipids in the standard sediments (Belt et al. 2012). The data is provided in ng g-1 dry sediment (sed).

Data quality:

A procedural blank was added to each extraction batch to screen for contamination. The 4-year instrumental uncertainty is < 10%.

The average sample resolution for all three time intervals is 5 ka, however, the sample resolution is highly variable.