Uncovering a potential bias in the RI-OH paleothermometer in methane seepage environments

Hydroxylated isoprenoid glycerol dialkyl glycerol tetraethers (OH-GDGTs) are ubiquitous, degradation-resistant archaeal membrane lipids. In marine environments, they are primarily thought to be produced by ammonia-oxidizing archaea Nitrososphaeria, and previous studies have shown that the number of cyclopentane moieties is positively correlated with upper ocean temperatures. This relationship has led to the development of the RI-OH and RI-OH’ paleothermometers. However, the low abundance of OH-GDGTs in marine settings, particularly in low-latitude regions, makes OH-GDGT-based paleothermometers relatively susceptible to the contribution of other archaea groups, such as methane-cycling benthic archaea, to the sedimentary OH-GDGT pool. This can potentially bias the relationship between the number of cyclopentane moieties and upper ocean temperatures.

Until now, the potential influence of OH-GDGT production by anaerobic methanotrophic archaea has remained unexplored.  To address this gap, we studied four sediment cores collected using a remotely operated vehicle within a ~200 m radius at three sites in the Four Way Closure Ridge cold seep system, offshore southwestern Taiwan. Three cores were collected from black patches associated with active seepage, while one control core represented a "normal" marine setting unaffected by seepage. A suite of data, including gas, porewater, bulk sediment geochemistry, archaeal genomics, and GDGT distributions, was generated.

We tested various calibrations using the control core, and found that the latest global RI-OH calibration yields temperatures that agree with the annual mean sea surface temperatures (SST). In contrast, other RI-OH and RI-OH’ calibrations produce temperatures that skew toward the summer and winter months, respectively. In three other sediment cores affected by anaerobic methanotrophic archaea, we observed that the Methane Index, which estimates the contribution of anaerobic methanotrophic archaea to the isoprenoid GDGT pool, reached as high as 0.7. Furthermore, we observed that when Methane Index values increase, the relative abundance of OH-GDGT-1 tends to increase as well, while that of OH-GDGT-2 tends to decrease. These changes corresponds to a temperature decrease of up to ~4°C as reconstructed using the RI-OH paleothermometer, while only marginally impacting the RI-OH' paleothermometer and the isoprenoid GDGT-based TEX86 (< 1°C). These findings suggest that anaerobic methanotrophic archaea may contribute to the sedimentary OH-GDGT pool in addition to the isoGDGT pool, potentially introducing a previously unidentified cold bias in the RI-OH paleothermometer. This underscores the need for further investigations into the role of anaerobic methanotrophic archaea in OH-GDGT production to improve the accuracy of RI-OH-based temperature reconstructions.