COMBINED CLIMATE AND ECOLOGICAL NICHE MODELING PREDICTS THE RELATIVE SEVERITY OF MARINE EXTINCTIONS DURING THREE PHANEROZOIC GREENHOUSE-ICEHOUSE TRANSITIONS

The Late Ordovician Mass Extinction (LOME) was one of the largest extinction events of the past 500 million years. It has long been recognized that this event is closely linked to climate change, with the first major extinction pulse coinciding with cooling, expansion of south polar ice sheets, and sea level fall near the Katian-Hirnantian boundary. However, it remains puzzling that the magnitude of extinction during the first pulse of the LOME far exceeds that associated with other Phanerozoic glaciations. This disparity has led to suggestions that cooling alone cannot account for the LOME, and that other factors, such as eutrophication, anoxia, and heavy metal toxicity, may be the primary cause of the event. Here we integrate global circulation models and ecological niche simulations to compare expected impacts of cooling on shallow marine fauna during the Late Ordovician, Late Eocene-Oligocene, and Plio-Pleistocene climate transitions. We simulate nine different types of species, varying in both thermal tolerance range and dispersal ability. We randomly assign a range center for each species and allow it to expand to fill the adjacent coastal cells that have pre-transition sea surface temperatures within its thermal tolerance range. We then simulate a transition to glacial conditions and determine which species would be able to disperse along coastlines to stay within their tolerance range and which would be driven to extinction. Our simulations predict much higher extinction magnitude during the Late Ordovician transition than during the Cenozoic transitions, particularly among species with narrow thermal tolerance ranges. These predictions are consistent with observed extinction patterns, and bolster evidence that cooling could have been one of the major drivers of Late Ordovician extinctions. Experiments with artificial thermal gradients only weakly support the hypothesis that the paleogeographic configuration of the Late Ordovician placed Late Ordovician species at greater risk of climatically-driven extinction than their Cenozoic counterparts. Instead, the critical difference between Late Ordovician and Cenozoic simulations lies in the magnitude of cooling predicted by circulation models and implied by available proxy data.