MULTI-PROXY (PHYTOLITHS, STABLE ISOTOPES, TRACE FOSSILS) HIGH-RESOLUTION RECORD OF LATE MIOCENE ECOSYSTEM VARIABILITY

The tempo and mode of the spread of C₃/C₄ grassland ecosystems in North America have recently become clearer thanks to use of several new lines of evidence for vegetation, such as stable carbon isotopes and plant silica (phytolith) assemblages. However, elucidating the details of this profound ecological transition is complicated because different proxies (e.g., macrofossils vs. paleosols) give drastically dissimilar evolutionary histories of grasslands in terms of their gross spatial and temporal distribution and their development. Often, only one data source for reconstructing paleovegetation has been applied in one area or section, so little is understood about possible taphonomic biases in sampling.

We combine stable carbon isotopic paleovegetation data (% C₄), phytolith assemblage data, and paleopedological information about rooting and bioturbation for reconstruction of early late Miocene habitats in southwestern Montana. The studied 34-meter section is of the Sixmile Creek Formation at Timber Hills, Montana and contains 35 paleosols. Preliminary results show that vegetation structure varied through the section, from open grasslands with a minor woody component to relatively closed forest near water, as indicated by abundant wetland indicators (e.g., sedge phytoliths, diatoms). Percentage C₄ vegetation based on isotopic proxy data and phytolith analysis are roughly similar, although error margins are large due to current uncertainty in how well grass phytolith shape reflects C₄ photosynthesis. Root diameter also varies widely, and reaches its highest variance at intermediate levels of bioturbation. The current data show that the highest degrees of bioturbation are linked to more closed habitats near water, whereas the widest root diameters are associated with more open habitats, counter to expectation. These results suggest that although some paleovegetation data (phytoliths vs. isotopes) correspond fairly well, others (phytoliths vs. trace fossils) may suffer from different taphonomic biases, for example in time averaging (phytolith assemblages may form over longer time spans than preserved trace fossils); these relationships need to be taken into account in paleovegetation reconstruction.