Marine animal diversity across latitudinal and temperature gradients during the Phanerozoic

The latitudinal biodiversity gradient (LBG) is a fundamental biological pattern seen across taxa and ecosystems today, but its drivers remain uncertain despite intense study. Palaeontological data may add valuable evidence from diversity distributions during intervals with different Earth system configurations, including potential analogues of future climate regimes. However, accurately characterising these distributions is challenging because the geographic scope of fossil record coverage varies through time, introducing biases that have not been quantified by most previous studies. Here, we attempt a comprehensive documentation of latitudinal biodiversity distributions for marine invertebrates through the past 540 million years, explicitly accounting for regional variation in diversity and sampling. We demonstrate large uncertainties when using current fossil data at this scale. Nevertheless, some signals are detectable. We show that marine animal biodiversity declined with increasing palaeolatitude and with decreasing temperature during at least some intervals from the Permian onwards (298.9 Ma). Additionally, we find that the LBG was shallower on average when Earth’s climate was hotter, although this signal is weak. We also document a strong, systematic bias due to intense sampling of the fossil record in North America and especially Europe, which may have led previous studies to incorrectly infer a mid-latitude diversity peak during hot intervals of Earth history. Our results provide a baseline for what current fossil databases might tell us about Phanerozoic LBGs of marine animals, and suggests that quantitative evaluation of uncertainties and systematic bias will be central to advancing knowledge of geographic variation in diversity through Earth’s history.