AllIGATR: A Multi‐Dimensional Approach for Marine Archaeal Lipid‐Based Paleothermometry

Abstract

Archaeal glycerol dibiphytanyl glycerol tetraether membrane lipids (GDGTs) are ubiquitous in marine sediments. The number of ring structures in six specific GDGTs varies with seawater temperature, leading to their use in paleothermometry. GDGT temperature dependence is typically quantified in TEX ₈₆ . However, calculating a single value from a six‐parameter archive reduces information prior to temperature determination. Here we use the full suite of GDGTs in surface and ancient sediments, mesocosms, and cultures to show that fractional abundances of GDGTs each have a distinct temperature optimum. Subsequently, we devise the ALL Isoprenoid GDGT Adaptation to Temperature in R (AllIGATR) calibration, which represents GDGT‐specific temperature responses as Gaussian functions. AllIGATR uses surface sediment data to constrain how this curve relates to sea surface temperature (SST) and depth‐integrated temperatures of 0–200 m (TO0TH). Compared to TEX ₈₆ , AllIGATR provides a slightly improved fit to ocean temperatures in surface sediments (AllIGATR root mean square error (RMSE) 4.5°C; TEX ^H ₈₆ RMSE 4.6°C), but shows large improvements in temperature prediction based on mesocosm and culture data, especially over 30°C (AllIGATR RMSE 6.3°C; TEX ₈₆ RMSE 8.4°C). In downcore records, AllIGATR performs well within the modern temperature range and yields reasonable paleotemperatures for most downcore GDGT distributions we test, but extreme temperature prediction in warm climates of the Eocene and Cretaceous highlight underconstrained GDGT response to temperature in these GDGT distributions. We suggest that inclusion of hydroxylated GDGTs and GDGT‐4 could possibly further improve the temperature fit, as would exploring mixing models to investigate GDGT contributions from different ecological groups.