Vertebrate diversity across the end-Permian mass extinction — Separating biological and geological signals

Abstract

This study examines regional diversity patterns of Permian–Triassic terrestrial vertebrates of the South African Karoo Basin and the Russian fore-Ural region. Previous studies clearly established that the diversity patterns of synapsids, the dominant terrestrial tetrapods of their time, are significantly biased by heterogeneity in the Permian–Triassic terrestrial rock record. This conclusion was based on a positive correlation between taxonomic diversity estimates (TDEs) and varying sampling proxies: (1) number of formations and (2) outcrop area. Recently, these proxies have been criticized to be non-independent variables or problematic for other reasons and (3) number of localities or (4) number of specimens have been suggest as more reliable proxies. In the present study, I test the relevance of the latter proxies (2–4) with respect to vertebrate diversity in South Africa and Russia. On these regional scales, log-transformed and generalized differenced TDEs of vertebrates at varying taxonomic levels are (with few exceptions) not significantly correlated with the various sampling proxies (2–4) for the complete time series. After eliminating the impact of the end-Permian extinction event by excluding the earliest Triassic time intervals, selected vertebrate TDEs in South Africa and Russia show statistically significant strong positive correlations with outcrop area, number of localities, and number of specimens. Diversity residuals, resulting from modeled diversity estimates, exhibit clade-specific patterns with little support for a mid-Permian (end-Guadalupian) event but strong support for a severe end-Permian extinction. The results imply that, although vertebrate diversity patterns in South Africa and Russia are at least partially biased by the Permian–Triassic terrestrial rock record, they still preserve genuine biological signals. Finally, it is important to note that outcrop area represents a valid and useful sampling proxy at the regional scale of the South African Karoo Basin and that it should be used in concert with other sampling proxies, such as the number of localities and specimens, that accurately capture various aspects of sampling in a given study area, as it has been performed in the present study.