Marine nitrogen cycling dynamics under altering redox conditions: insights from deposition of sapropels S1 and the ambiguous S2 in the Eastern Mediterranean Sea Version 17

The eastern Mediterranean Sea (EMS) sedimentary record is periodically interspersed with organic- rich ‘sapropel’ layers. Sapropels are characteristic of basin-wide anoxic events, triggered by precession-forced insolation maxima. Relatively subdued insolation maxima, however, are not always expressed as distinct sapropel events. The EMS sedimentary record hereby allows anoxia and nitrogen (N) cycling tipping points to be investigated, which may act as analogues for modern deoxygenation. To this end, we investigated a ~68 kyr EMS sedimentary record, containing the well-established sapropel S1 (deposited in two phases: S1a [~10.5─8.5 ka BP] and S1b [~7.8─6.1 ka BP]) and sediments timed to the ambiguous S2 sapropel (~53 ka BP). We focus on lipid biomarkers of microorganisms to reconstruct key components of the N cycle: (1) anaerobic ammonium oxidation (anammox) using ladderanes and a stereoisomer of bacteriohopanetetrol (BHT-x), (2) dinitrogen gas (N2) fixation using heterocyte glycolipids (HGs), and (3) nitrification by Thaumarchaeota using crenarchaeol. During S1, export-productivity (indicated by the barium to aluminum ratio) and anoxia (indicated by redox-sensitive trace elements and benthic foraminifer assemblages) are enhanced. Thaumarchaeota are most abundant in S1a, while anammox is enhanced throughout S1. N2-fixation (indicated by bulk sedimentary δ15N) occurs throughout S1, while the highest heterocyte cyanobacteria abundance is at the S1a termination and S1 interval. Ladderane presence suggests additional episodes of bioavailable N removal between ~69 to 39 cal ka BP. These episodes correspond to brief periods of water column deoxygenation, with anoxia occurring at the sediment-water interface in sediments timed to S2 (53─51 cal ka BP). During these episodes, ladderanes co-occur with the BHT-34R stereoisomer, but not BHT-x. Compound-specific δ13CBHT-34R indicates an anammox source. Our results highlight various modes of operation of the N cycle during the different deoxygenation events. During S1a, a combination of N-loss and P-supply may have reinforced anoxia, by favoring diatom-diazotroph symbiotic consortia. Conversely, a coupling between N2-fixation and anammox was not observed during S1b and the S2-timed interval, either because loss of bioavailable N was insufficient or diazotrophs were nutrient limited. During these periods, anammox may have provided negative feedback on anoxia by quenching primary production.