EVALUATING THE IMPACT OF TAPHONOMY ON THE FIDELITY OF GEOGRAPHIC RANGE-SELECTIVE EXTINCTION PATTERNS IN TERRESTRIAL MAMMALS

Geographic range size is a fundamental characteristic of species and frequently correlated with extinction risk, both in the modern and for various times in geologic history. While recent simulation studies have supported the fidelity of the spatial fossil record, questions remain regarding the impact of taphonomy on our ability to accurately interpret the geographic range size selectivity of extinction events. To that end, we simulated fossil records of North American terrestrial mammals using a series of preservation biases based on published relationships between mammal body mass, population density, and expected number of carcasses, and tested the sensitivity of these patterns to the inclusion of fossil microsites, those that overwhelmingly preserve small vertebrates. We simulated extinction by removing 25%, 50%, or 75% of small-ranged taxa, small-bodied taxa, or taxa at random. All scenarios were tested for extinction selectivity by performing a binary logistic regression on simulated geographic range size. This framework allowed us to determine (1) the ability to preserve range size selective extinction patterns using realistic taphonomic parameters for terrestrial mammals, (2) the potential to misinterpret body size driven extinction events as range size driven events, and (3) the chances of falsely generating a range size selective extinction pattern through taphonomic processes alone. Results suggest that microsites must be analyzed in conjunction with traditional macrofossil sites for geographic range size selective extinction patterns to be accurately reconstructed. When looking only at macrofossils, body mass driven taphonomic bias destroys our ability to reconstruct a true geographic range size selectivity pattern. Conversely, during a body size driven extinction, removal of small-bodied mammals frequently generates a geographic range size selectivity pattern (a false positive) in macrofossil assemblages, though the inclusion of microfossils reverses this trend. The likelihood of randomly generating a false positive extinction selectivity signal is high, though consistently less common than in range size selective simulations. This suggests a need for extreme caution when interpreting extinction selectivity results derived exclusively from macrofossil records.