Biotic impacts of temperature before, during, and after the end-Permian extinction: A multi-metric and multi-scale approach to modeling extinction and recovery dynamics

Publication Date

January 2019

Document Type


Publication Title

Palaeogeography, Palaeoclimatology, Palaeoecology





First Page


Last Page



Extinction and delayed recovery during the end-Permian extinction and Early Triassic has been linked to environmental instability brought on by volcanic outgassing and greenhouse conditions, but the relative importance of the myriad of environmental stressors at this time on recovery dynamics is not well understood. Previous workers have documented both overall delayed biotic recovery for the entirety of the Early Triassic, but also incipient recoveries that appear to occur relatively early after the initial extinction event. Here, we explore the patterns of extinction and recovery using several metrics of ecological complexity in marine benthic communities using a global dataset, and compare several multiple regression models to determine which set of abiotic factors best predicts extinction and recovery dynamics. We additionally test the importance of temporal scale of analysis in interpretations of recovery dynamics and modeling results, by including analyses at the epoch, stage, and substage scales bracketing the interval of extinction and recovery. We find differences in mode of recovery between the ecological metrics analyzed, with some metrics exhibiting an Early Triassic recovery lag, while others recover continuously or immediately following the initial extinction event. We also find evidence of a global ‘Dienerian minimum’, with overall levels of community complexity significantly lower than earlier Griesbachian communities, suggesting a synchronous disturbance to the progression of recovery at this time. The regression model with δ18Oapatite mean values as the response variable is most often found to be the best-fit model across all time scales analyzed, though proxies of rock record fidelity and paleontological sampling effort become more important in finer timescale analyses, likely due to dwindling sample numbers. Out of the models tested, these results suggest that global ocean temperatures best predict patterns of extinction and recovery across several ecological metrics, and that thermal episodes during the initial extinction event and subsequently in the Early Triassic recovery period significantly suppressed benthic marine community health.


Permian-Triassic, Paleoecology, Regression models


SJSU users: use the following link to login and access the article via OneSearch.