Middle Miocene environments from herbivore stable isotopes and assessing C4 biomass seasonal variability

This study reports enamel stable carbon (δ13C) and oxygen (δ18O) isotope analyses of fossil tooth enamel from a broad suite of mammalian herbivores recovered from the late Early Miocene locality of Buluk, northern Kenya, with the aim of reconstructing local environments and assessing whether C4 plants were part of herbivore diets during the early–middle Miocene transition. Buluk is an important eastern African locality for documenting habitat heterogeneity across this interval because it preserves a diverse mammalian assemblage, including ruminant artiodactyls, suoids, anthracotheres, rhinocerotids, proboscideans, and hyraxes. The research had two explicit objectives: to establish dietary isotopic ranges for these fossil herbivore guilds and to compare Buluk δ13C enamel values with published enamel isotope records from three other East African localities (the Early Miocene Ugandan site Moroto and the Middle Miocene Kenyan sites Maboko Island and Fort Ternan) to examine inter-site variation and niche partitioning among Early and Middle Miocene herbivores.

Analyses of δ13C enamel and δ18O enamel indicate that most Buluk herbivores foraged in a C3-dominated mosaic of open-canopy woodland habitats without evidence for closed-canopy forest conditions. These enamel isotope results align with independent soil and biomarker analyses that point to predominantly C3 vegetation at the site, though paleosol and plant-wax geochemical indicators also document some presence of C4 vegetation on the Buluk landscape. Notably, despite geochemical evidence that C4 plants occurred in the regional flora, bulk δ13C enamel values from the sampled herbivores show no isotopic signature consistent with substantial consumption of C4 biomass. The authors discuss possible explanations for this discrepancy: C4 plants may have been present on the landscape but were not consumed in sufficient quantities by the locally sampled herbivores to be detectable in bulk enamel δ13C; the sampled taxa may not include the species that consumed C4 resources; or the different proxies (soil and plant-wax geochemistry versus enamel isotopes) integrate ecological information over different spatial and temporal scales, producing scale-dependent signals.

Comparative analyses place Buluk’s range of δ13C enamel values closer to the Early Miocene site of Moroto and significantly more enriched than the Middle Miocene localities Maboko Island and Fort Ternan. This pattern suggests that more water-stressed environments predominated at earlier Miocene sites in East Africa sampled here, and that the Buluk assemblage reflects a C3-dominated but spatially heterogeneous landscape with open woodland components. δ18O enamel data are used in the study to characterize drinking and water-use behaviors of herbivores and to provide additional context for local aridity and water-stress conditions. Together, the Buluk isotopic dataset and inter-site comparisons contribute to a regional picture of temporal and spatial habitat variation across the developing East African Rift System in the early to middle Miocene. By establishing isotopic dietary baselines for multiple herbivore groups and comparing these values across sites, the study refines understanding of niche partitioning before the widespread emergence of C4 grasslands around 10 Ma and highlights the complexity of paleoecological interpretations when different paleoenvironmental proxies record differing aspects of landscape composition and resource use.