Although phytolith analysis is a promising tool for reconstructing the palaeovegetation in grassland ecosystems, no phytolith study has yet attempted to quantitatively reconstruct the corresponding plant community characteristics. Questions remain regarding whether or not soil phytoliths can quantitatively reflect various aspects of grassland physiognomy, and how well modern phytoliths represent the quantitative characteristics of grassland ecosystems. These are important scientific issues in the field of phytolith analysis, which require further research on the modern phytoliths of grassland ecosystems. In order to explore the potential of phytolith analysis in the temperate grassland region, we have for the first time established a modern soil phytolith reference dataset by exploring the relationships between soil phytoliths and plant species inventory data from 77 sites in the Songnen grassland in Northeast China. The results show that phytolith assemblages can clearly distinguish Leymus chinensis grassland, Stipa grassland and mesophytic herb grassland. In addition, we found that the Ic phytolith index could discriminate samples of L. chinensis grassland from the other samples, while the Iph phytolith index proved to be less reliable in the region. In addition, in terms of quantitative plant community characteristics, the soil phytoliths showed a good correspondence with plant community cover and species richness, but not with biomass and diversity index. The transfer functions established in this study can provide reliable estimates of community cover and species richness in the Songnen grassland. In addition, phytoliths are shown to have the potential to reconstruct the taxonomic abundance of the aboveground plant community. This study is a major advance in the application of phytolith analysis to the quantitative estimation of plant community characteristics, and it provides a modern phytolith reference for the quantitative reconstruction of the paleovegetation of grassland ecosystems.
Read full abstract