Earth and Planetary Sciences Faculty and Staff Publications

Publication Date

Spring 4-23-2021

Document Type



Multiple linear regression was used to determine the relationships between diversity-independent factors (i.e., abiotic, climatic) 2, 5, and 10 Myrs-prior to the most elevated Phanerozoic extinctions. We constructed five abiotic variables from Phanerozoic proxy records1–5 to compare to extinction rates: mean temperature, temperature instability, carbon cycle instability, continental weathering rates, and habitat instability. All three models were statistically significant (P < 0.05) and explained > 70% of the variation in Alroy’s6 three-timer generic extinction rates. However, the 2 Myr-prior model explained the most variance in extinction rates and had the most predictive power, based on adjusted and predictive R2 (~ 72% and 41%, respectively). Carbon cycle and habitat instabilities significantly contributed to this model (P < 0.05), thus suggesting that these variables positively contribute to the most severe extinctions during the Phanerozoic. However, carbon cycle and habitat instabilities seem to behave as extinction intensifiers, requiring an additional trigger to set off a major extinction event. Using the equation of the best fit line of the 2 Myr-prior model and the significant variables carbon cycle and habitat instabilities, we predicted a modern three-timer generic extinction rate of 0.85 (PI: 0.29, 1.40), falling between the end-Ordovician and end-Triassic mass extinctions in taxonomic severity. These results provide important information regarding the role diversity-independent factors play in intensifying the most elevated extinctions during the Phanerozoic and will continue to play in our present and future. Furthermore, these results support the importance the fossil record for contextualizing the potential severity of the modern extinction crisis.


mass extinction, environment, diversity-independence, anthropogenic, climate change, linear regression