CENOZOIC POLYCYSTINE RADIOLARIAN DIVERSITY AND THE EFFECT OF EOCENE-OLIGOCENE BOUNDARY GLOBAL CLIMATE CHANGE

Unlike for carbonate plankton, until now only subjective estimates of Cenozoic radiolarian diversity dynamics have been available. Textbooks usually suggest diversity decline vs the Mesozoic, and stasis or further decline from Paleogene to Neogene (e.g. De Wever et al., 2001). We provide a first global quantitative analysis of Cenozoic radiolarian diversity, using the NSB marine microfossil occurrences database, and a new comprehensive revision of its species-level taxonomy (Lazarus et al., in prep.). All data were filtered for reworked occurrences, plus pacman range trimming (Lazarus et al., 2012) to remove remaining outliers. Results are robust to levels of trimming and subsampling algorithm used (primarily SQS - Alroy, 2010, or classical rarefaction). Data were analysed globally and by latitudinal band. Data in NSB was insufficient in quantity below ca 45 Ma for analysis.

Radiolarian global diversity increased strongly over the last 45 Ma, tripling to peak values in the latest Miocene, before declining slightly in the Plio-Pleistocene. Rates of turnover (Foote metrics) were high and increasing in the Paleogene, peaking at the E-O boundary, driven by rapid diversity increase and establishment of a diverse, largely endemic Southern Ocean fauna. Global rates of turnover were lower throughout the Oligocene-Miocene, increasing again in the last ca 5 Ma due to elevated rates of extinction. Tropical radiolarian diversity declined by nearly half between the late mid Eocene and the mid Oligocene, and only gradually recovered to mid Eocene levels in the late Miocene, when Southern Ocean diversity also peaked, and coeval major diversification in north polar regions raised global diversity to a Cenozoic maximum. The Eocene-Oligocene climate shift thus fundamentally altered radiolarian global faunas, shifting the locus of diversification to polar regions and altering global rates of bidiversity dynamics. These results largely confirm earlier regional studies by other authors but contradict the current textbook view. It is also worth noting that little of this dynamic behavior would be visible using a (typical for many paleobiology studies) single global curve with ca 10 my time bins.