A SIMULATION FRAMEWORK FOR UNDERSTANDING THE INFLUENCE OF SAMPLING ON APPARENT DIVERSITY PARTITIONING PATTERNS WITHIN DEPOSITIONAL BASINS

Paleoecological analyses often focus on changes in diversity partitioning through time or across habitats. Many ecological dynamics can create similar spatiotemporal diversity patterns, making it difficult to test specific hypotheses that identify causal relationships between abiotic or biotic interactions and diversity trends. To account for confounding factors, we need to model how fossil abundances are generated by age, location, deposition, and sampling. We present a framework for simulating marine taxa that are under zero net diversification and distributed uniformly across a water-depth gradient, whose abundances are sampled in a sequence-stratigraphic context. We simulate sampling six stratigraphic columns from the shore to the middle of a basin filled during several cycles of sea-level. Each species-response curve follows a normal distribution centered on its preferred depth, but each species has a right-skewed probability of being sampled. Sampling frequency decreases with water depth. We compare alpha and beta-diversity among facies and within stratigraphic columns with diversity metrics of varying abundance sensitivity. The model deliberately removes biotic and abiotic factors such as time averaging, species interactions, temperature, and oxygen availability that may covary with depth to demonstrate that even in their absence some diversity metrics exhibit conspicuous depth gradients. For example, alpha diversity is lowest in the shallowest-water facies, where there are no taxa at the upper end of their depth range. Alpha diversity also increases towards more basinward stratigraphic columns, because these columns include more facies and deeper-water facies than nearshore columns. However, diversity metrics that incorporate relative abundance are less sensitive to these confounding trends, suggesting abundance data is better than occurrence data for distinguishing ecologically-driven diversity patterns. This model framework is a starting point for simulating changes in species interactions, thermohaline gradients, extinction events, and other topics of broad interest.