MODERN SOIL CHEMISTRY DISTRIBUTIONS WITHIN THE SERENGETI ECOSYSTEM AND THEIR IMPLICATIONS FOR PALEOSOL-BASED PROXIES

Bulk geochemical oxide ratios are a common method of interpreting changes in weathering, parent material, climate, etc. in both modern and fossil soils (paleosols). We typically relate ratios of oxides to a soil attribute, like mean annual precipitation (MAP), using a database of modern soils. We then identify a relationship and then use the geochemistry as a proxy to reconstruct these attributes in the past. However, these bulk geochemical databases center on North America, which does not represent the full range of climates and soil geochemistry on Earth. For example, eastern Africa has two distinct wet/dry seasons, high soil temperatures (often >20°C), and unique chemistries due the volcanism associated with the rift system. To effectively use geochemical proxies as indicators of past climate in eastern Africa, we need to develop proxies specific to the region, and therefore, modern soil samples from 11 sites were collected across a transect of the Serengeti Ecosystem in Tanzania. Geochemistry (10 major oxides), pH, electrical conductivity (EC), and Munsell soil color was measured at 10 cm intervals through all soil profiles, and a subset of these samples were analyzed for bulk and clay mineralogy from each identified soil horizon. From east to west, pH, EC, CaO, and Na₂O all decrease likely due to the increase in MAP across the transect. In contrast, SiO₂ decreases and MgO, MnO, SrO, BaO, and P₂O₅ all increase from east to west likely due to changes in mineralogy and parent material. Mineralogy changes very little with depth; however, there are significant changes across the landscape in the relative mineral abundances where quartz and illite/smectite increase and calcite decrease from east to west. This is likely due to a combination of MAP increases and parent material changes from alkaline volcanics in the east to granitoids and mafic rocks in the west. Overall, the geochemistry indicates that there are statistically significant differences between these North American soil databases and African soils. The bulk geochemical proxies generally underpredict soil attributes like MAP and pH, and suggests a global dataset is needed to improve to applicability of these proxies beyond climates found in North America.