Sampling biases in the rise and fall of fossil marine animal genera

An important question in evolutionary biology and macroecology is whether taxa show systematic trajectories in occupancy, the proportion of geographic area occupied, over macroevolutionary timescales. Past studies have used fossils to document these trajectories, showing a symmetric rise and fall. In this study, I focus on several biases in the analyses of fossil occupancy trajectories that have been unaccounted for. First, better sampling of boundary bins in a taxon’s stratigraphic range, paradoxically, results in lower mean occupancy of taxa in those bins. This is because better sampling allowed more taxa with low occupancies to be included in the mean occupancies of those bins compared with intermediate bins. Second, the possibility that taxa may have incomplete durations within boundary bins could also lower occupancies in those bins. Finally, a bias can also exist when the number of sampled sites is not constant throughout a taxon’s stratigraphic range. I use simulations to show that the first bias can be corrected by conditioning these boundary bins to be sampled in the same way as intermediate bins. To mitigate the second bias, I use higher-resolution time bins to constrain the intervals over which taxa’s occupancies are measured so that they are comparable between boundary and intermediate time bins. I also present an approach that can correct for the last bias by subsampling geographic sites, testing its impact in a simulation. Considering these factors, the occupancy trajectories of marine animal genera look to be a relatively gradual rise post-origination with a sudden decline before extinction.