Intact polar and core glycerol dibiphytanyl glycerol tetraether lipids in the Arabian Sea oxygen minimum zone. Part II: Selective preservation and degradation in sediments and consequences for the TEX86

The TEX86 is a proxy based on a ratio of pelagic archaeal glycerol dibiphytanyl glycerol tetraether lipids (GDGTs), and used for estimating past sea water temperatures. Concerns exist that in situ production of GDGTs lipids by sedimentary Archaea may affect its validity. In this study, we investigated the influence of benthic GDGT production on the TEX86 by analyzing the concentrations and distributions of GDGTs present as intact polar lipids (IPLs) and as core lipids (CLs) in three sediment cores deposited under contrasting redox conditions across a depth range from 900 to 3000 m below sea level in and below the Arabian Sea oxygen minimum zone (OMZ). Direct analysis of IPLs with crenarchaeol as CL via HPLC/ESI-MS2 revealed that surface sediments in the OMZ were relatively depleted in the phospholipid hexose, phosphohexose (HPH)-crenarchaeol compared to suspended particulate matter from the water column, suggesting preferential and rapid degradation of this IPL. In sediment cores recovered from deeper, more oxic environments, concentrations of HPH-crenarchaeol peaked at the surface, probably due to in situ production by ammonia-oxidizing Archaea, followed by a rapid decrease with increasing depth. No surface maximum was observed in the sediment core from within the OMZ. In contrast, the glycolipids, monohexose-crenarchaeol and dihexose-crenarchaeol, did not change in concentration with depth in the sediment, indicating that they were relatively well preserved and likely mostly derived from fossil pelagic GDGTs.These results suggest that phospholipids are more sensitive to degradation, while glycolipids might be preserved over longer time scales, in line with previous incubation and modeling studies. Furthermore, in situ produced IPL-GDGTs did not accumulate as IPLs, and did not influence the CL-TEX86. This suggests that in-situ produced GDGT lipids were more susceptible to degradation than fossil CL and IPL and did not accumulate as CL. In agreement, no significant changes of TEX86 with sediment depth in the core lipids were observed. However, consistent differences between IPL-derived TEX86 and CL-TEX86 were found. These could be explained by a different composition of CL-GDGT of the glyco- and phospholipids, in combination with dissimilar degradation rates of phospholipids vs. glycolipids. We also observed consistent differences in both IPL-derived and CL-TEX86 between the different cores, equivalent to 3 °C when converted to temperature, despite the proximity of the core locations. These differences may potentially be due to a larger addition of GDGTs produced in deeper, colder waters to the (sub)surface-derived GDGTs for the deeper core sites