The Early Miocene rise of the grass-dominated ecosystem is a plausible trigger for a sharp Miocene increase in accumulation of nonmarine diatomaceous sediment as well as diversification of nonmarine diatoms. This grassland radiation introduced a biogeochemical mechanism for enhancing widespread and sustained mobilization of usable silica and other nutrients. Volcanism was probably responsible for episodic nonmarine diatomaceous sediments from the advent of the oldest known nonmarine diatoms in the Late Cretaceous through the Oligocene. Although prolific Miocene volcanism was undoubtedly still important in the development of many diatomites, feedback from grassland colonization of volcanic soils may explain why diatomaceous sedimentation surged in the Miocene following a more sparse pre-Miocene record.
The initial rise of the grass-dominated ecosystem, increased nonmarine diatomite accumulation, and Early Miocene evolutionary radiations of nonmarine diatom taxa are at least approximately coeval. Although the earliest known grass is Paleocene, multiple lines of evidence, including mollic-epipedon paleosols, fossil occurrences of hypsodontic ungulate grazers, and fossil phytoliths, suggest that the grass-dominated ecosystem did not expand significantly until Early Miocene. The grassland radiation apparently was delayed until Middle Miocene in parts of Eurasia, Africa, and Australia. If that delay is real, the diatom/diatomite record in those regions should coincide with it. The onset of increased Miocene diatomite accumulation is as yet imprecisely dated, but coincidence with the rise of the grass-dominated ecosystem is predicted herein. Early Miocene diversifications of Actinocyclus and Thalassiosira diatoms are consistent temporally with grassland expansion where it is Early Miocene.
Subsequent adjustments in the silica cycle also may be attributed to grasslands. Nonmarine diatom radiations in the Late Miocene and Pliocene coincide with sharp regressions that may have released nutrients and soluble phytolith opal stored in Miocene soils and paleosols as well as dissolved silica in soil pore waters. Regressional erosive pulses of phytoliths provide a new explanation for low Ge/ Si ratios in marine diatoms during Pleistocene glacial intervals. Nonmarine diatoms from regressive intervals should record lower Ge/Si ratios than before and after those regressions because of phytolith contributions with low Ge/Si ratios. Late Miocene radiations of C4 and moist tall-grass ecosystems may have mobilized even more silica than the short, dry-climate Early Miocene grasses. Abundance of diatomite may have fluctuated in concert with changes in degree of volcanism, even after grassland expansion, but at substantially higher levels than before this new terrestrial ecosystem arose.