The distribution of C₃ and C₄ graminoids in southern Africa responds to a complex set of climatic variables, floral evolution patterns, topography, and geology. This study analyzes the geographic distribution of diagnostic opal phytoliths in Restionaceae, Cyperaceae, and Poaceae in 140 soil surface samples across various biomes in South Africa and Botswana. Frequencies of phytolith morphotypes are correlated with qualitative variables (biome and vegetation type) and quantitative variables (aridity index, percent of winter/summer rain, precipitation variability, and January temperature). The classification of Poaceae phytoliths is based on a scheme of grass silica short cells (GSSC) previously published by Cordova and Scott (2010, Palaeoecology of Africa, vol. 30). The classification uses unique and shared morphotypes commonly found in C₃ (Pooideae, Danthonioideae, Ehrhartoideae, Arundinoideae) and C₄ (Panicoideae, Chloridoideae and Aristidoideae) grass subfamilies. Although topography and geology influence the distribution grasses, many of the morphotypes studied are strongly controlled by growing season temperatures and precipitation seasonality.
The main purpose of the classification presented here is to create a protocol for reconstructing past vegetation and climatic conditions from graminoid opal-phytolith assemblages in paleosols, lacustrine deposits, hyrax middens, and herbivore dental calculus/plaque. Because the winter and summer rainfall areas have each characteristic phytolith morphotypes, this study ultimately aims at reconstructing rainfall seasonality fluctuations during glacials and interglacials. Furthermore, strong relation between phytolith morphotypes and vegetation types and biomes is a resource employed in reconstruction of herbivore paleodiets and movement patterns.