Cycles in fossil diversity

TL;DR: Rohde et al. as discussed by the authors used the International Commission on Stratigraphy's 2004 time scale to translate the stratigraphic records into a record of diversity vs. time, with details given in the supplement.

Abstract: Cycles in Fossil Diversity Robert A. Rohde & Richard A. Muller Department of Physics and Lawrence Berkeley Laboratory University of California, Berkeley California 94720 USA Report LBNL-56544 20 October 2004 submitted to Nature It is well-known that the diversity of life appears to fluctuate during the course the Phanerozoic, the eon during which hard shells and skeletons left abundant fossils (0-542 Ma). Using Sepkoski's compendium 1 of the first and last stratigraphic appearances of 36380 marine genera, we report a strong 62 ± 3 Myr cycle, which is particularly strong in the shorter-lived genera. The five great extinctions enumerated by Raup and Sepkoski 2 may be an aspect of this cycle. Because of the high stastical significance, we also consider contributing environmental factors and possible causes. Sepkoski's posthumously published Compendium of Fossil Marine Animal Genera , and its earlier versions, has frequently been used in the study of biodiversity and extinction 3-4 . For our purposes, diversity is defined as the number of distinct genera alive at any given time, i.e. those whose first occurrence predates and last occurrence postdates that time. Because Sepkoski references only 295 stratigraphic intervals, the International Commission on Stratigraphy's 2004 time scale 5 is used to translate the stratigraphic records into a record of diversity vs. time, with details given in the supplement. Though Sepkoski's is the most extensive compilation available, it is known to be subject to certain systematic limitations due primarily to the varying availability and quality of geologic sections 6-7 , the implications of this will be discussed where appropriate. Fig. 1A shows diversity vs. time for all 36380 genera in Sepkoski’s compendium. In Fig. 1B we show the 17797 genera that remain when we remove those with uncertain ages (given only at epoch or period level), and those with only a single occurrence. The smooth trend curve through the data is the third-order polynomial that minimizes the variance of the difference between it and the data. The overall shape of 1A and 1B is similar to those previously published for fossil families 2 and for genera 3 . It rises rapidly at the beginning of the Phanerozoic (right side), drops to a nadir near the Permian- Triassic boundary (251 Ma), and then rises steeply until the present. These variations may result from evolutionary and environmental drivers 8 , observational biases 6 , or changes in the number of available geologic sections 7 ; for example, the sharp rise towards the present may be driven by the greater availability and study of recent sections. Our focus is not on the trend but on the short-term variations shown in 1C, obtained by subtracting the trend from 1B. The Fourier spectrum of 1C is shown in 1E. It is dominated by a strong peak with period 62 ± 3 Myr (frequency 0.016 cycles/Myr). The sine wave corresponding to this cycle is also shown in Fig. 1C, where it accounts for 35% of the variance. Note that because steep drops in diversity are often followed by gradual recoveries, the peaks and valleys in the data don’t precisely align with those of the sine curve. Also, some abrupt features may appear more gradual because of incomplete records. 9 We indicate the 5 major extinction events of Raup and Sepkoski 2