Stable carbon isotopes from paleosol carbonate and herbivore enamel document differing paleovegetation signals in the eastern African Plio-Pleistocene

Abstract

Analyses of stable carbon isotopes (δ13C) from herbivore dental enamel and paleosol carbonates are important tools for Plio-Pleistocene paleovegetation reconstructions. A single herbivore tooth documents an isotopic record of vegetation on the order of 10–1–1 years and in proportion to that individual's foraging range. Paleosol carbonates, conversely, record environmental information on longer time scales (102–3 years) but smaller spatial scales (101 m2). Given that these two proxies document paleoenvironments at different spatiotemporal scales, it is worth comparing them to see if they offer redundant or complementary paleovegetation information. Here, we compare δ13C patterns from herbivore enamel and paleosol carbonates from geological (sub)members in the Awash Valley, Ethiopia, and Turkana Basin, Kenya, from ~ 4.4–1 million years ago. We find that median herbivore enamel δ13C is typically ~ 5–7 per mil (‰) higher than that from paleosol carbonates within a given (sub)member. The distributions of paleosol carbonate δ13C values usually have less spread (variation) than herbivore enamel. The bias in median and spread between these two data types likely reflects the different spatial and temporal scales at which these proxies record paleoenvironmental information. Most Plio-Pleistocene fossiliferous deposits are formed in fluvial settings in which paleosol carbonates sample the immediate habitat of floodplain woodlands and shrubs, resulting in a lower δ13C (i.e., more C3) signal. On the other hand, fossil teeth of wide-ranging herbivores could generate a higher (i.e., more C4) and more variable δ13C signal if some taxa fed in floodplain woodlands while others fed on open grasslands distal to the floodplain. We conclude that δ13C values from herbivore enamel and paleosol carbonates offer paleovegetation data at different spatiotemporal scales, both of which are informative for hominin habitat reconstructions. Careful consideration of the spatial and temporal signals inherent in these and other proxies should be applied in future studies.