PP046-0009
Lateral heterogeneity in paleosol geochemistry and resampling to improve proxy uncertainty

Wednesday, 16 December 2020
Poster
Rebecca Dzombak, Rebekah Stein and Nathan D Sheldon, University of Michigan Ann Arbor, Ann Arbor, MI, United States
Abstract:
Paleosols (fossil soils) serve as key terrestrial records of climate as they form at the surface, in constant contact with precipitation, temperature, and the atmosphere. Many proxies exist to reconstruct these variables, but typically, proxy uncertainty is limited to calibration error and does not take other types of uncertainty (e.g., spatial heterogeneity) into account. Specifically, we seek to improve the uncertainty associated with using a small number of paleosol profiles to represent landscape-scale climate and weathering processes. A previous study (Hyland et al., 2016) analyzed a 1.5-km laterally-continuous, consistently oxidized paleosol of the Eocene Wasatch Fm. at Honeycomb Buttes in the Green River Basin (GRB), and found it to be geochemically homogenous at that length-scale except for δ18O of pedogenic carbonates. That work focused on the reliability of pedogenic carbonate δ18O for paleo-altimetry studies but did not address the uncertainty of other paleoclimate proxies. To expand on this work, we sampled a 3-km laterally-continuous floodplain paleosol of the Wasatch formation at Oregon Buttes in the GRB. These 11 paleosol profiles had varying redox conditions but were characterized by geochemical homogeneity with respect to weathering indices and paleoclimate proxies. By resampling a random subset (n=1 thru 11) of the profiles and recalculating paleoclimate variables based only on their geochemistry, we improve statistical uncertainties on some paleosol-based paleoclimate proxies, which is critical because rarely is a single paleosol traceable for hundreds of meters. Additionally, inter-basin comparison of penecontemporaneous paleosols from the Wind River Basin (Hyland et al., 2013) reinforces that paleoclimate interpretations should be regionally limited. Paleosol-based paleoclimate studies should sample multiple profiles whenever available to increase certainty and should incorporate this type of uncertainty into their reconstructions.