Carbon and oxygen isotopic variability in Neogene paleosol carbonates: constraints on the evolution of the C4-grasslands of the Great Plains, USA

Understanding the development of grassland ecosystems is critical for understanding the evolution of their associated vertebrate faunas. We focus on the central and southern portions of the Great Plains, which is the largest continuous grassland in North America and has a well-documented mammalian fossil record. Modern biomass in this region is dominated by C4 grasses, but paleontological evidence indicates the region was forested prior to the middle Miocene, implying the region was formerly dominated by C3 plants. We examine longand short-term variations in the abundance of C4 grasses and climatic conditions in the central and southern Great Plains using the stable carbon and oxygen isotope composition of paleosol carbonates from 28 sections of 11 lithostratigraphic units in the region. The carbon isotope composition (dC) of paleosol carbonate reflects the proportion of C3 and C4 plants that grew in an ancient soil. The long-term pattern of dC values indicates that the percentage of C4 grasses in the Great Plains was moderate (ca. 20%) throughout most of the Miocene, increased from 6.4 to 4.0 Ma, and reached modern levels by 2.5 Ma. Based on the range in dC values within sections and the mean dC values for different carbonate morphologies that reflect different intervals of carbonate accumulation, the abundance of C4 biomass did not vary substantially on short time scales. The variability in C4 biomass on short time scales does not correlate with age of section, latitude, longitude, or the mean percentage of C4 biomass. The oxygen isotope composition (dO) of soil carbonate is controlled by soil temperature and the dO value of soil water. Great Plains paleosol dO values vary considerably in relation to time and latitude. Removing the variation due to latitude with least squares linear regression reveals a temporal pattern that is consistent with the global pattern of Neogene climate change. Quantitative interpretation of the dO record is difficult without independent control on temperature or the oxygen isotope composition of soil water, both of which have probably varied with long-term climate change. Comparisons of measured dO values for each section with predicted values based on modern temperature data and estimated modern meteoric water dO values suggest that (1) most Miocene sections have dO values consistent with warmer temperatures and more positively shifted soil water dO values than today, (2) three Miocene sections in western Nebraska probably had soil water compositions that were more negative than similar-aged localities in the region, and (3) the Plio-Pleistocene sections have dO values consistent with cooler temperatures and more negative soil water dO values than today. The paleosol dC values do not 0031-0182/$ see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.palaeo.2003.09.030 * Corresponding author. Current address: Department of Geology and Geophysics, University of Minnesota, Minneapolis, MN 55455, USA. Tel.: +1-612-624-6361; fax: +1-612-625-3819. E-mail addresses: [email protected] (D.L. Fox), [email protected] (P.L. Koch). 1 Tel.: +1-831-459-5861. D.L. Fox, P.L. Koch / Palaeogeography, Palaeoclimatology, Palaeoecology 207 (2004) 305–329 306 correlate strongly with the pattern of climate change recorded by the dO values, suggesting that long-term change in mean temperature is not the main control on the abundance of dC biomass in the Great Plains. We propose two testable hypotheses to account for our data and other published data that bear on the evolution of the Great Plains ecosystem. D 2004 Elsevier B.V. All rights reserved.