Impact of terrestrial organic matter input on distributions of hydroxylated isoprenoidal GDGTs in marine sediments: Implications for OH-isoGDGT-based temperature proxies

Isoprenoidal glycerol dialkyl glycerol tetraethers (isoGDGTs) and hydroxylated isoGDGTs (OH-isoGDGTs) are widespread in marine, lacustrine, and terrestrial environments and serve as paleoenvironmental proxies. Several indices based on their distributions, such as ${TEX}_{86}$ (based on isoGDGTs), $\text{RI-OH}$, $\text{RI-OH}\prime$ (both based on OH-isoGDGTs), $\%OH$ and ${TEX}_{86}^{OH}$ (both based on isoGDGTs and OH-isoGDGTs), have been utilized as sea water temperature proxies. Although terrestrial organic matter (OM) input of isoGDGTs may affect the ${TEX}_{86}$ in coastal marine settings, relatively little is known on its consequences for OH-isoGDGT distributions and related proxies. Here, we studied the distributions of OH-isoGDGTs in soils, rivers, and coastal marine settings at three locations: the Kara Sea, the Iberian margin, and the northern Gulf of Mexico, receiving terrestrial OM through the Yenisei, the Tagus and the Mississippi Rivers, respectively. In general, we observe higher relative abundances of OH-isoGDGTs ($\%OH$) in coastal marine environments compared to soils and rivers from the same area. Comparison of OH-isoGDGT distributions shows that, in particular, the abundance of OH-isoGDGT with one cyclopentane moiety relative to total OH-isoGDGTs was lower in terrestrial settings. In general, the $\text{RI-OH}$ was higher in terrestrial settings, while such a consistent offset was not observed for $\text{RI-OH}\prime$. The ${TEX}_{86}^{OH}$ index, exhibits a distinct difference between terrestrial and marine settings, similar to the pattern observed for the ${TEX}_{86}$. This similarity is primarily attributed to the relatively minor influence of terrestrial OH-isoGDGTs compared to regular isoGDGTs in the ${TEX}_{86}^{OH}$ index. Despite these differences, only the coastal sediments of the Kara Sea showed indications of a potential bias of OH-isoGDGTs-based proxies caused by terrestrial OM input. However, these distributional variations may also be caused by the large salinity gradient in the Kara Sea, since it has been established that salinity has an effect on the distributions of OH-isoGDGTs. Our results show that caution should be exercised when interpreting temperature estimates based on OH-isoGDGT proxies in marine settings affected by large river outflows resulting in a substantial terrestrial OM input and/or have a strong salinity gradient.