The largest mass extinction in Earth’s history happened approximately 252 million years ago at the end of the Permian period. This end-Permian event, commonly termed the “Great Dying,” resulted in the extinction of 80% of marine life as well as of many terrestrial vertebrates. It was driven by eruptions of molten rock from volcanoes in Siberia over the course of hundreds of thousands of years; the heat and carbon dioxide released from these eruptions led to ocean acidification and anoxia.
Previous explanations linked the end-Permian extinction to these lower oxygen levels, but so far, there’s been no scientific consensus about the exact way atmospheric changes caused the die-off. Now, a team led by biogeochemical modeler Dominik Hülse, a postdoc at the University of California, Riverside, is hoping to change that. Using a 3D Earth system model, researchers are proposing a temperature-driven, mechanistic explanation for the end-Permian extinction that incorporates proxy data types in a new way, according to Kimberly Lau, an assistant professor of geosciences at Pennsylvania State University and coauthor of the findings, published in Nature Geoscience.
Complex System, Complex Modeling
Prior modeling studies used a static model of the marine carbon pump, a set of biological processes in which carbon from the atmosphere cycles through the ocean and oceanic sediment. Static modeling studies assumed that organic matter degraded at uniform rates throughout the ocean’s depths.
The new model provides a broader view of the end-Permian extinction, more realistically accounting for conditions that affected the die-off. “It’s a complex system, but [these] modeling results show that just including the temperature effect on the biological pump can have major impacts on the spatial distribution of carbon isotopes, which has really big implications for how we interpret them and link them to carbon perturbations in the past,” Lau said.
Hülse, Lau, and the team used a dynamic, temperature-based model that accounted for variable organic matter degradation rates—and therefore for areas of low oxygen—as the ocean rapidly warmed. They also lowered the sinking rate of particulate organic matter by 22%, a shift that reflects the new understanding that organisms were smaller and lacked carbonate shells during the Permian period, Hülse said. Finally, the model accounted for sulfurization, the reaction of organic matter compounds with the hydrogen sulfide that builds up when oxygen levels are low. (Microbes have difficulty consuming the resultant compounds.) Where older models had estimated nearly complete anoxia, or lack of oxygen, on the seafloor, Hülse’s model estimated less extreme anoxia, in one instance just 30%.
“Previous models, in order to create either larger areas of anoxia at the bottom of the ocean or hydrogen sulfide higher up in the ocean, if they could model it, usually had to increase the nutrient input into the ocean by a lot: 10 times more, or sometimes even higher,” Hülse said. With the new model, nutrient levels (in this case, phosphorus) were required only to double.
“With this temperature-dependent approach,” Hülse continued, “microbes remineralize—respire—this organic matter faster when it’s warmer. It’s like when you have food on a table outside. If it’s warmer, it also rots faster.…It’s similar in the ocean. If it gets warmer, the microbes work faster, and respire things faster, and therefore, they get respired further up in the water column.”
Ying Cui, a paleoclimatologist at Montclair State University in New Jersey, said the new model’s changes and assumptions are reasonable—if not the only plausible explanations for the end-Permian event. Cui, who was not involved in the study, pointed to the complexity of trying to reconstruct the mass extinction. “It raises a bunch of new questions,” Cui said. “How do we incorporate an Earth system model that is capable of simulating all of these processes that may have [occurred] together that triggered the end-Permian extinction?”
Hülse agreed that modeling the end-Permian extinction is something of a conundrum, as it involves Earth system processes relevant on varying timescales. Models that are able to provide higher spatial resolution typically cannot incorporate data from processes that occurred over longer time spans, while models run for longer time frames necessarily incorporate less spatial detail of the time span studied.
—Robin Donovan (@RobinKD), Science Writer