Mixed models elucidate local- and regional-scale drivers of paleoenvironmental change in eastern Africa during the emergence of Paranthropus and Homo

Abstract

Climate and environmental change are considered important drivers of hominin speciation and diversification, including the emergence of Paranthropus and Homo. Therefore, making connections between potential drivers of hominin evolution on multiple spatial scales is a critical issue in hominin paleoecology. This is especially true in light of evidence that paleoenvironmental proxies operating at different spatial scales lead to divergent interpretations about whether environmental change in eastern Africa was primarily driven by regional-scale or local-scale factors. This study takes a novel approach to addressing issues of scale in hominin paleoecology by using linear mixed models to test the relative impact of local factors versus regional factors in driving variation in aridity and vegetation at important hominin paleontological sites across eastern Africa between 3.7 Ma and 2.3 Ma. Results indicate there are no significant trends in aridity as evidenced by stable oxygen isotopes of paleosol carbonates. However, the results for stable carbon isotopes of paleosol carbonates suggest regional-scale factors drove a uniform increase in grassy vegetation at nearly all sites. Additionally, we find that while regional-scale factors drive vegetation change, local-scale factors were important in determining the unique mix of trees/shrubs and grass that was present at each site and upon which vegetation change acted. Thus, while sites across eastern Africa experienced the same degree of vegetation change, at any given time the vegetation at each site would have differed from other sites. This result offers a possible resolution between regional-scale proxies which show an overall increase in C4 photosynthesizing grasses and local-scale proxies which indicate that paleoenvironments at different sites were varied. It also connects regional-scale drivers to local-scale environmental change which would have had an important impact on hominin evolution and adaptation. This study introduces linear mixed models as a methodology which to date has not been widely applied to questions of hominin paleoecology and has the potential to address critical issues of scale.