Inverse methods for extracting seasonality signals from ungulate tooth enamel

The variability in the isotopic composition of tooth enamel recorded in serially-sampled ungulate teeth provides a valuable archive of changes in the food and water consumed by an animal on the temporal scale of months to years. These records provide evidence of paleoseasonality and insight into animal behavior and ecology over seasonal temporal scales relevant to selection. However, interpreting these records proves difficult due to the complex nature of enamel mineralization, which variably compresses, lengthens, and dampens the input isotopic signal during growth. Despite these complications, “inverse modeling” can recover original input signals from serial chemical measurements in teeth, by modeling enamel mineralization and the incorporation of mineral content with specific isotopic compositions through time. Here we present a set of models to execute this process, which fit the enamel mineralization patterns of plains zebra ( Equus quagga ) but are parameterized to be adaptable to the growth patterns of other mammalian teeth. Through sensitivity testing and the application of the inverse model to real-world data from modern plains zebra from Ol Pejeta Conservancy in central Kenya, we show that these models recover an accurate seasonality signal from preserved enamel, particularly closer to the coronal end of the crown and over longer seasons (i.e., longer periodicities). These models are a valuable expansion of the existing toolkit for the analysis of enamel isotope data as a paleoecological proxy. • Isotopic analysis of serially-sampled tooth enamel can track paleoseasonality. • Interpretation of isotope values is complicated by how enamel forms. • Inverse modeling recovers input seasonal δ 18 O signals from Equus teeth. • More realistic enamel mineralization models recover a more accurate input signal. • Wavelet analysis of inverse model results reveals subannual periodicities in δ 18 O.