A fossilised shell can reveal whether a species was doomed long before it vanished. This project will analyse 485 million years of fossil data from shelly marine invertebrates to identify which biological traits and environmental stressors—such as changes in temperature, pH, oxygen, or food supply—best predict extinction risk. The researchers will test whether external pressures are stronger predictors than internal factors like body size or geographic range, and whether combinations of stressors (for example, heat plus low oxygen) are more dangerous than any single one. This matters because current conservation efforts lack a long-term baseline. Most extinction risk assessments rely on recent observations, which may miss slow-acting or cumulative threats. If the models succeed, they could identify which modern marine species are most vulnerable under future climate scenarios, giving conservation planners a tool grounded in deep evolutionary history. The project will also produce new three-dimensional reconstructions of ancient ocean chemistry, which could improve how Earth system models simulate past and future environmental change. While the research is fundamentally curiosity-driven—aimed at understanding macroevolutionary dynamics—it may ultimately help prioritise species for protection before they reach crisis point.
View original technical description
Extinction is a natural process that has occurred continuously since life began on Earth. Patterns of extinction in the fossil record help to reveal those species at higher risk of extinction based on their traits. For example, species that occupy large regions (i.e., have large geographic ranges) are less at risk of extinction than those that are found in only restricted regions. Determining the factors that elevate extinction risk is important, because it helps us to better understand evolutionary history and potentially identify those species at risk of extinction today. This project aims to uncover the biological and environmental factors which have led to elevated extinction risk among marine organisms over 485 million years of evolutionary history. Our analyses will use shelly marine invertebrates, as these organisms preserve well as fossils. We will specifically test whether external, environmental stressors, such as changes in temperature, pH, oxygen, and food supply, are better predictors of extinction risk than biological traits, such as body size, geographic range size, or the thermal tolerance of a species. Our state-of-the-art, spatially explicit analyses will analyse both new and known determinants of extinction risk to robustly characterise how the relative strength of predictors may have varied over Earth history, providing a step-change in our understanding of macroevolutionary dynamics. We will additionally examine whether the synergistic interaction between environmental stressors (e.g., temperature and dissolved O2) better predict extinction risk than any variable alone. Using extinction models calibrated on fossil data, we will then ask which modern-day taxa might be most at risk of future extinction under different global change scenarios. These models will provide an estimate of intrinsic extinction risk for modern taxa using the most up-to-date understanding from palaeontological baselines. To produce estimates of the regional environmental changes experienced by taxa over the Phanerozoic for use in our extinction risk models, we will develop new, state-of-the-art biogeochemical ocean reconstructions, which we refer to here as ‘PhanerO-3D’. These new reconstructions will integrate two different Earth system models, SCION and cGENIE, to provide the most up-to-date understanding of spatial variations in marine oxygen, pH, and nutrients over Earth history. Overall, our novel modelling framework will allow us to disentangle the correlates of extinction over millions of years, providing new insight on what regulates large-scale evolutionary patterns and which taxa may be most prone to extinction today.
Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.
Is something wrong? Let us know