Decreasing Phanerozoic extinction intensity is a predictable consequence of Earth surface oxygenation and metazoan ecophysiology

Extinction events are often correlated with geochemical evidence for environmental change in the geologic record, and are increasingly being explored as deep-time analogues to modern climate change. The metabolic index, an ecophysiological technique that quantifies the extent of viable habitat for marine ectotherms, offers a mechanistic approach to evaluating the role of ocean warming and deoxygenation in driving extinctions observed in the fossil record. However, modern understanding of temperature-dependent hypoxia tolerances can only be applied to warming events in the geologic record in the context of long-term changes in surface oxygenation. Extinction rates have been shown to decrease through the Phanerozoic (~541 Ma, million years ago, to present), with particularly high extinction rates in the Cambrian- Ordovician (~541 to ~440 Ma). There is increasing support for limited Earth surface oxygenation (<40% present atmospheric levels) through this interval of high extinction rates, with an increase in atmospheric oxygen proposed sometime between 470 and 400 Ma. Here, we integrate ecophysiological methods with the cGENIE Earth system model to evaluate the role of atmospheric oxygen in the extinction vulnerability of marine animals through geologic time. Specifically, we model the impact of warming events on metabolic habitat viability for marine ectotherms at a range of atmospheric oxygen levels using an ensemble Earth system modeling approach and a probabilistic representation of the metabolic index. We demonstrate that extinction vulnerability to ocean warming (and associated solubility-driven deoxygenation) is significantly higher at the low atmospheric oxygen levels predicted for the early Paleozoic. We therefore argue that higher early Phanerozoic extinction rates are a predictable function of metazoan physiology, and that atmospheric oxygen is a key variable to constrain in the application of deep-time climate change analogues to past and future ocean warming.