In this study, we update a modeling approach called TEXAS, which uses laboratory data from living microbes to improve how we relate TEX₈₆ to ocean temperature. Unlike earlier models, our new version uses a statistical method called Bayesian inference to better handle uncertainty and include prior scientific knowledge. When we apply this improved model to existing PETM data, we find that ocean warming was likely more modest than some previous studies suggested—about 8°C rather than 10°C or more.

Our work shows that TEX₈₆-based temperature estimates can be made more reliable by including information about microbial biology and environmental conditions. We are now expanding this model to also consider nutrient levels in the ocean, which can affect the TEX₈₆ signal and help us make even better reconstructions of ancient climate.