Duration and Intensity of End-Permian Marine Anoxia

Ocean anoxia is widely interpreted to have played an important role in the end-Permian mass extinction and uranium isotope data are among the most important tools for quantifying the global extent of ocean deoxygenation. Stratigraphic variations in carbonate δ238U and uranium concentrations are interpreted to reflect changes in anoxia across time. However, quantifying trends in redox conditions is challenging because of noise in the data and uncertainty regarding the values of key parameters in the geological uranium cycle, such as the length of a perturbation and the amount of anoxia experienced during the event. In this study, we use Monte Carlo simulations of a uranium cycle model and quantify uncertainty in interpretations of marine anoxia based on uranium data. Numerical simulations are compared against paired records of U concentrations and isotopic compositions from eight stratigraphic sections reported from previous studies. The 1% of model runs that are most similar to the data were determined using principle component analysis and the parameterization of these runs used to characterize posterior distributions of uncertainty for these parameter values. The results of this approach indicate a prolonged (million-year) perturbation in the seawater U cycle during which the levels of anoxia in the ocean increased from modern day values (0.2% of seafloor area) to approximately 4% of the seafloor. The results rule out scenarios involving a dramatic increase in ocean water anoxia that lasted for an extended period of time, favoring scenarios that include a relatively short (< 30 Kyr) yet pronounced increase in anoxia in the range of 15% to 20% followed by a more prolonged interval (1 Myr) of less extensive but still elevated anoxia in the range of 3% to 4% of the seafloor. This approach highlights a broader range of possible scenarios that previous attempts do not.