Joint estimation of sampling and turnover rates from fossil databases: capture-mark-recapture methods revisited

Abstract

The estimation and interpretation of temporal patterns in origination and extinction rates is a major goal of paleobiology. However, the possibility of coincident variation in the quality and completeness of the fossil record makes the identification of such patterns particularly difficult. Previously, Nichols and Pollock (1983) proposed that capture-mark-recapture (CMR) models be adapted to address this problem. These models can be used to estimate both sampling and turnover rates, reducing the risk of confounding the two quantities. Since that time, theoretical advances have made possible the application of these tools to a much broader range of problems. This paper reviews those advances likely to be of greatest relevance in paleobiological studies. They include (1) joint estimation of per-taxon origination and extinction rates, (2) modeling sampling or turnover rates as explicit functions of causal variables, (3) ranking of alternative models according to their fit to the data, and (4) estimation of parameter values using multiple models. These are illustrated by application to an Ordovician database of benthic marine genera from key higher taxa. Robustness of these methods to violation of assumptions likely to be suspect in paleobiological studies further suggests that these models can make an important contribution to the quantitative study of macroevolutionary dynamics.