Refined Isotope Stratigraphy across the Continental Paleocene-eocene Boundary on Polecat Bench in the Northern

One of the most continuous and best studied continental stratigraphic sections spanning the Paleocene-Eocene boundary is preserved on Polecat Bench in the northern Bighorn Basin. The mammalian biostratigraphy of Polecat Bench sediments has been well documented, and includes a major reorganization of faunas at or near the P-E boundary. To complement the existing biostratigraphy, we measured the isotopic composition of paleosol carbonate nodules at soil-by-soil temporal resolution through the P-E boundary interval. These measurements provide a detailed record of the abrupt, transient, carbon isotope excursion that affected atmospheric and oceanic carbon reservoirs at ca. 55 million years before present [Ma]. Tests of soil thickness and diagenesis indicate that trends in the record are primary, and reflect syndepositional changes in the 613C value of atmospheric C02. The carbon isotope record suggests that the 613C value of atmospheric C02 dropped by ca. 8%0 during this interval, and then rebounded. The pattern of change is very similar to that of an independent high-resolution record of Bains et al. (submitted). Changes in the 6180 of paleosol carbonates are consistent with a significant increase in local mean annual temperature during the P-E boundary event. Comparison of biotic and isotopic stratigraphies on Polecat Bench shows that faunal changes at the P-E boundary lag behind major events in the carbon and oxygen isotope records by about 10 thousand years [k.y.]. INTRODUCTION time series can be derived from the oxygen and carbon isotope The isotopic chemistry of authigenic soil minerals can be a composition of paleosol carbonates, with 6180 recording sensitive recorder of climatic and environmental conditions. primarily changes in paleoclimatic variables, and 613C Authigenic carbonate from fossil soils (paleosols) is a valuable responding to paleoenvirOmental shifts. Here and widely used source of paleoclimatic and paleoenvirondelta values express deviations in per mil (%o) units. For mental proxy information, and the isotope systematics of example, 6180 is the deviation of the oxygen isotope ratio R of soil carbonate formation have been documented and a sample from that for a standard (V-PDB). This is calculated modeled (Cerling, 1984). Fossil soils are often developed many times in a stratigraphic section and thus very useful 6180 = lo00 (Rsample Rstandard)lRstandard In: Paleocene-Eocene Stratigraphy and Biotic Change in the Bighorn where R = 1801160. Similarly, 613C is the deviation (in %o) of and Clarks Fork Basins, Wyoming (P. D. Gingerich, ed.), University of the carbon isotope ratio of a sample from that of a standard (VMichigan Papers on Paleontology, 33: 73-88 (2001). PDB), where R = 13C112C. Extreme transient climate and carbon cycle events at the Paleocene-Eocene boundary have been documented in isotopic proxy records from a number of sources, including marine carbonates (Kennett and Stott, 1991; Zachos et al., 1993; Bains et al., 1999; Katz et al., 1999; Norris and Rohl, 1999), fossil pollen (Beerling and Jolley, 1998), and paleosol carbonates (Koch et al., 1992, 1995; Bowen et al., 2000; Bains et al., submitted). During the Paleocene-Eocene transition, marine bottom waters and high-latitude surface waters warmed dramatically (Kennett and Stott, 1991; Zachos et al., 1993), and substantial warming is indicated in terrestrial settings (Koch et al., 1995; Fricke et al., 1998). This transient climate event is associated with a large change in the carbon isotope composition of earth-surface carbon reservoirs. In the oceans, the 613C values of planktonic foraminifera dropped by ca. 2.5 to 4.5%0, while those of benthic species decreased by ca. 2.5%0 (Kennett and Stott, 1991; Zachos et al., 1993). Proxy records for atmospheric C02 show an abrupt decrease in 613C values of as much as 7%0 (Koch et al., 1992, 1995; Bowen et al., 2000; Cojan et al., 2000). This short-lived change in the isotopic composition of carbon at the earth's surface is best explained by the release and oxidation of at least 1500 gigatons of carbon from seafloor methane hydrate reservoirs, followed by sequestration of this excess carbon during carbon cycle re-equilibration (Dickens et al., 1997; Bains et al., 1999; Beerling, 2000). The stratigraphic sequence on Polecat Bench in the northern Bighorn Basin of Wyoming preserves a nearly continuous series of carbonate-bearing paleosols (e.g., Fig. 1) that represent late Paleocene and early Eocene time (Gingerich, 1983). These same soils preserve fossil mammals that have been collected and studied by paleomammalogists since the late nineteenth century. Previous study of paleosol carbonates from Polecat Bench and adjacent regions of the Clarks Fork Basin indicated that the Paleocene-Eocene boundary isotope excursion is recorded as a 6 to 7%0 drop in the 613C of soil carbonate nodules (Koch et al., 1992, 1995). Here we exploit recent advances in understanding of the stratigraphy on Polecat Bench (Gingerich, this volume) to present a detailed paleosol-bypaleosol analysis of isotopic change across the PaleoceneEocene boundary interval on Polecat Bench. ISOTOPIC SIGNATURE OF SOIL CARBONATE Processes that affect the isotopic composition of soil carbonate have been well documented in modern settings (Cerling, 1984; Quade et al., 1989; Cerling et al., 1991). Carbon in soil carbonate is derived from soil C02, which is a mixture of carbon dioxide from the atmosphere, and from organic decomposition and root respiration within the soil. The mixing of C02 from these sources has been modeled (Cerling, 1984; Cerling et al., 1991), and in soils with moderate to high productivity, minimal influence of atmospheric C02 is observed below ca. 30 cm depth. Below this depth, soil C02 is enriched in 13C by ca. 4.4%0 relative to plant tissues and respired C02, due to the effects of diffusion (Cerling et al., 1991). This offset, along with consistent fractionations associated with carbonate precipitation (ca. 10.5%0) add to produce soil carbonate with a F13C value ca. 15%0 greater than that of overlying vegetation (Cerling, 1984). The isotopic composition of vegetation should track the Fl3C of atmospheric C02 with a consistent fractionation of approximately -19%0 for Paleocene-Eocene plants using a C3 photosynthetic pathway (Bocherens et al., 1993; Arens et al., 2000). Therefore, changes in the carbon isotope composition of the atmosphere are propagated and recorded in the F13C of soil carbonate C02. Oxygen in soil carbonate is derived from soil water, which is derived from rain and snow (Cerling, 1984). Temporal variation in the 6180 value of precipitation has commonly been interpreted to reflect variation in mean annual temperature, following the relationship developed by Dansgaard (1964). Recent work, however, has shown that this relationship is not quantitatively applicable on geological time scales (e.g., Boyle, 1997; Fricke and O'Neil, 1999). Study of modern soil carbonates has demonstrated that their oxygen isotope composition is not typically affected by rock-water interactions in the soil zone, but may reflect significant evaporative enrichment of soil waters (Quade et al., 1989). MATERIALS AND METHODS Fossil soil development is nearly ubiquitous within finegrained sediments on Polecat Bench, and paleosols in this study were recognized as brown, green, orange, red, or purple mudstone horizons. Colored mudstones typically represent the Bhorizons of ancient soils, and commonly contain pedogenic carbonate nodules as well as other pedogenic minerals (Bown and Kraus, 1981,1987; Kraus, 1987). Paleosol carbonates were sampled near the base of paleosol B-horizons in freshly-exposed outcrops where there is no evidence of modem soil development (e.g., modern roots). Samples were collected at multiple stratigraphic levels within some of the thicker paleosols. Paleosols were sampled in situ in four local sections, which were mapped concurrently (Gingerich, this volume). Stratigraphic thicknesses were measured by hand leveling using a 1.5-meter Jacob's staff and by differential GPS, and detailed stratigraphic sections can be found elsewhere in this volume. Local sections were correlated by tracing marker beds (Gingerich, this volume), and results are presented on a composite stratigraphic scale. A similar record, generated independently, is presented in Bains et al. (submitted). In the laboratory, carbonate nodules were polished flat on a lapidary wheel using 600-grit silica carbide powder, washed, then dried in a low temperature oven. Samples (ca. 100 pg) were drilled from the polished surfaces under a binocular microscope using a mounted dental drill. Primary rnicritic carbonate and secondary diagenetic spar were sampled independently, and, where possible, two samples were drilled from each of two nodules from every soil in each local section. Samples were roasted in vacuo at 400" C for 1 hour to remove organic contaminants. These were analyzed using a Micromass Optima or Prism gas-source mass spectrometer, following reaction with 100% phosphoric acid at 90" C, with the aid of an FIGURE 1 Orange, red, and purple paleosol horizons (numbered darker bands) at the base of the carbon isotope excursion on Polecat Bench (Bains and Norris numbering system is shown here). Principal marker beds of Gingerich (Gingerich, this volume: figs. 7 and 8) are equivalent to beds numbered here, as follows: Thick Orange = beds -23 through -2 1 ; Purplish red mudstone = bed -1 9; Red mudstones = bed -1 1 ; Purple-0 = bed -5; Top Brown = bed -1 ; Lower Double-Red A and B = beds 0 and 1 ; and Purple-2 = bed 4. The carbon isotope excursion begins at the thin red mudstone represented by bed -8 here. University of Michigan vertebrate locality SC-404 with Meniscotherium priscum is at the level of beds -3 and -4. The Wa-0 faunal zone begins at bed 0. In this section the top of Wa-0 and the top of the carbon isotope excursion are missing due to truncated by the scour-and-fill sequence starting above bed 7. automated Isocarb device. Analytical precision, based on repeated analysis of an in-house standard, was better than 0.1%0 for both carbon and oxygen.