Oxygenation as a driver of the Great Ordovician Biodiversification Event

The largest radiation of Phanerozoic marine animal life quadrupled genus-level diversity towards the end of the Ordovician Period about 450 million years ago. A leading hypothesis for this Great Ordovician Biodiversification Event is that cooling of the Ordovician climate lowered sea surface temperatures into the thermal tolerance window of many animal groups, such as corals. A complementary role for oxygenation of subsurface environments has been inferred based on the increasing abundance of skeletal carbonate, but direct constraints on atmospheric O2 levels remain elusive. Here, we use high-resolution paired bulk carbonate and organic carbon isotope records to determine the changes in isotopic fractionation between these phases throughout the Ordovician radiation. These results can be used to reconstruct atmospheric O2 levels based on the O2-dependent fractionation of carbon isotopes by photosynthesis. We find a strong temporal link between the Great Ordovician Biodiversification Event and rising O2 concentrations, a pattern that is corroborated by O2 models that use traditional carbon–sulfur mass balance. We conclude that that oxygen levels probably played an important role in regulating early Palaeozoic biodiversity levels, even after the Cambrian Explosion. During the Ordovician period one of the greatest biological radiations of the Phanerozoic took place, when genus-level diversity quadrupled and ecospace utilization increased1. During this Great Ordovician Biodiversification Event (GOBE), marine communities expanded into new niches such as epifaunal suspension feeding and deep burrowing2. This new niche space was largely exploited by members of the Palaeozoic Evolutionary Fauna (EF), which includes articulated brachiopods, crinoids, ostracodes, cephalopods, corals, and bryozoans1. Both the Cambrian EF (trilobites and inarticulate brachiopods) and Modern EF (bivalves, gastropods, fish, and so on) also diversified, but it was the expansion of the Paleozoic EF that drove the GOBE. The causes of the GOBE remain poorly understood and may include both intrinsic biological factors and external environmental drivers (such as global cooling, nutrient delivery from erosion, higher sea levels that expanded habitable platform area, and oxygenation)1,3,4,5. Oxygen isotopes from well-preserved conodont apatite provide proxy evidence for high sea surface temperatures (~40 °C) at the onset of the Ordovician that may have inhibited diversification, but global cooling throughout the Early–Middle Ordovician brought temperatures closer to modern conditions and possibly into the tolerance window (27–32 °C) for members of the Palaeozoic EF6. Cooling oceans could also store more dissolved oxygen and more effectively ventilate subsurface environments, which would in turn create a stronger vertical gradient in carbonate saturation that lowered the metabolic costs of skeletal carbonate biomineralization in surface waters 7c animal life 9 and expanded the range of habitable ecospace for infaunal burrowers deeper into the sediment 0. A more oxygenated ocean could also have supported more predators in the food chain (fish and cephalopods), setting into motion an evolutionary ‘arms race’11. Ordovician global cooling is generally thought to have been caused by decreasing atmospheric CO2 (the cause of this drop is itself not well understood, but hypotheses include increased silicate weathering 12 and the advent of land plants 13), but a role for increased atmospheric O2 is possible via an increase in total atmospheric pressure and the associated inhibition of solar optical depth, scattering incident solar radiation that would have otherwise contributed to the surface latent heat flux 14. These arguments for linking oxygenation to cooling and biodiversification, while compelling, are hindered by poorly constrained ocean–atmosphere oxygen records. Existing isotope mass balance models are hampered by coarse time resolution (typically 10 Myr bins) that are not capable of resolving changes in atmospheric O2 as a cause of the main pulses of biodiversification across the GOBE 15,16,17 (Supplementary Fig. 1). The absence of charcoal is interpreted to reflect atmospheric O2 levels below 13–15% until the Late Silurian 18, but primitive non-vascular land plants only expanded into terrestrial environments by the Middle–Late Ordovician 13, thus the charcoal record is not well suited to constrain Early Ordovician O2 levels. Land plant expansion is thought to have increased organic burial rates and oxygenated the atmosphere to near modern levels 13, but the timing and magnitude of this oxygenation is also poorly resolved. Similarly, although iron-based redox proxies suggest that O2 levels were between 2% and 21% throughout the Ordovician, their resolution is not yet sufficient to resolve finer temporal trends 19. Here we apply a new approach to reconstruct the changes in atmospheric oxygen with high age-resolution based on the O2-dependence of carbon isotope fractionation during photosynthesis 20.