SYSTEMATIC VARIABILITY IN CLUMPED ISOTOPE TEMPERATURES OF LOESS CARBONATES DURING LATE PLEISTOCENE CLIMATE CHANGE, PALOUSE LOESS, USA

The rich stratigraphic, geochemical, and palynological records preserved in Palouse loess deposits that blanket eastern Washington state, USA, inform investigations into the Quaternary history of Cordilleran Ice Sheet advance and retreat and associated climate change. Climate models suggest Last Glacial Maximum (LGM) mean annual temperature was on the order of ~5-7 °C colder than modern in this region. However, most Palouse loess paleoclimate proxy records offer only indirect or qualitative measures of Late Pleistocene temperature change. Here we use clumped isotope thermometry (T(D₄₇)) to determine direct measures of formation temperatures of carbonates in loess-paleosol sequences that span known climate variability during the Late Pleistocene. We measured the D₄₇ values of bulk loess carbonates and, where possible, accompanying pedogenic root pore cements for comparison. A loess-paleosol sequence (CLY-3) deposited proximal to the loess source at Eureka Flat records relatively large (> 10-15 °C) changes in T(D₄₇) that correlate with the degree of pedogenesis. Warmer temperatures (T(D₄₇) > 35 °C) are recorded by carbonates in weak-pedogenesis loess intervals in comparison to cooler temperatures (T(D₄₇) < 30 °C) recorded by carbonates in paleosols (e.g., Washtucna paleosol). This difference in temperature is consistent with widely-accepted interpretations that loess accumulated during glacial retreat when strong and sustained southwesterly winds prevailed, whereas the soils formed during glacial advances characterized by slower loess accumulation and cold and dry, anticyclone conditions. Measured carbonate d¹³C values (-8 to -4 ‰) and calculated equilibrium water d¹⁸O values (-11 to -5 ‰) are consistent with a pedogenic origin for all carbonate samples analyzed. It is therefore possible that the measured T(D₄₇) variations record actual changes in soil temperature with changing climate. It is also possible that the measured T(D₄₇) variations result from changes in the relative proportions of detrital and in-situ, pedogenic carbonate in Palouse loess as a function of climate state. The warm T(D₄₇)-loess intervals observed at CLY-3 are not as well preserved at a more distal, lower deposition-rate sequence (WA-5) leaving the WA-5 record biased towards cold, soil-forming climate states.