Simulating the formation of herbivore tooth death assemblages to improve expectations for paleoenvironmental reconstruction from intra-tooth isotopic analysis

Isotopic analysis of serially-sampled dental enamel from fossil faunal assemblages is a popular paleoenvironmental proxy for its ability to inform on changes in seasonality and vegetation structure in the past. However, multiple factors contribute to patterning observed in isotope data from these enamel samples, including growth rate changes over the course of tooth development; variation in the amount of dental wear; seasonal variation in births that affect the timing of recorded environmental signals; and intra-annual variation in the isotopic composition of ingested water and plant foods. Here, we model tooth growth and wear in zebra ( Equus spp.) to generate simulated assemblages of teeth to characterize how the signal of an idealized oxygen isotope sine wave input is modulated by these sources of variation. In the model, a death assemblage of individual animals is generated using data from known populations. For each animal, a third molar (m3) is grown to a specified height, during which time it accumulates an isotopic signal, and then is worn based on empirically-observed rates of wear to its age at death. Simulated teeth are then sampled serially at designated intervals down the length of the tooth, generating an enamel isotope profile for each tooth. In a series of simulated experiments, we complexify the model by layering changing growth patterns and population demography and compare the resultant enamel isotope profiles. Our results demonstrate the dramatic impact these factors have on how isotopic signals are accumulated within teeth and within assemblages. To operationalize these results, we also compare our model output to measured oxygen isotope data from mid-Pleistocene zebra from the Elandsfontein dunefield in the southwestern Cape of South Africa. These comparisons indicate Elandsfontein may have experienced similar precipitation seasonality during the mid-Pleistocene to the southwestern Cape today. • Serial isotope profiles provide paleoecological insight on seasonal timescales. • Developmental and demographic factors complicate the interpretation of these data. • Simulations reveal how these factors drive variation in recovered isotope profiles. • Simulations also demonstrate how sampling strategy shapes isotope profiles. • We demonstrate how simulations can strengthen interpretation of fossil data.