Organic, inorganic, and isotopic proxy temperature estimates from a ten-year sediment trap record in the eastern Mediterranean Sea

Estimates of sea surface temperature (SST) in the geologic past often rely on geochemical proxies, including alkenone unsaturation ratios (U^K’₃₇), the proportion of cyclopentane rings in glycerol dibiphytanyl glycerol tetraethers (GDGTs; TEX₈₆), δ¹⁸O and Mg/Ca in planktonic foraminifera, and clumped isotopes in biogenic carbonates (Δ₄₇). Global calibrations show robust agreements between proxies and the temperature of the environment. However, proxy values in the eastern Mediterranean Sea are often offset from expected values based on these global relationships.

Here we aim to explain the Meditrerranean Sea proxy bias using sediment traps moored in the water column at 500, 1500, and 2500 m depth in the Ionian Sea between 1999 and 2018. We compare proxy values in trap material to those in surface sediment from the same site. U^K’₃₇-based SSTs (~12-22 °C) and TEX₈₆-based SSTs (~28-30 °C) vary seasonally, although they differ from satellite-based SSTs (~16-27 °C). No significant variations in mean Mg/Ca or δ¹⁸O values in G. ruber or in coccolith Δ₄₇ values were observed, indicating that these proxies do not vary on seasonal timescales in sinking particles at this site. Ultimately, however, sediments should record a flux-averaged proxy value. To this end, we explore export patterns of biomarkers and foraminifera.

We find that export of alkenones, GDGTs, and foraminifera exhibit different seasonality. Specifically, alkenone export typically peaks twice per year in late winter and late summer, maximum GDGT export often occurs in late winter, and export of planktonic foraminifera generally peaks in summer. Furthermore, alkenone export exhibits very little inter-annual variability, whereas the timing of GDGT export is more variable, and the peak flux of foraminifera is highly variable between years.

Flux-weighted U^K’₃₇- and TEX₈₆-, and mean coccolith Δ₄₇-based SSTs of 15, 29, and 15 ± 2 °C, respectively are offset from mean annual average satellite-based SSTs of 21 °C, whereas Mg/Ca and δ¹⁸O of G. ruber both yield average temperatures that are similar to the satellite-derived SSTs. Notably, whereas most proxy values measured in the surface sediment closely reflect those in the sediment trap material, the U^K’₃₇-based SST (13 °C) differs from the flux-weighted average in the sediment trap.

These results will be discussed in the context of seasonality, nutrient availability, dwelling depth, ecology, and physiological processes which may influence proxy values measured in the sediment trap and in underlying sediments.