Hot or not? Impact of seasonally variable soil carbonate formation on paleotemperature and O-isotope records from clumped isotope thermometry

Quantifying the timing and temperature of soil carbonate formation is important for interpreting isotopic records of Earth-surface temperature and soil water composition from paleosols. Pedogenic carbonates typically have been assumed to form at mean annual temperature, but recent work suggests warm-season bias in carbonate formation could impact the interpretation of delta O-18 values and carbonate formation temperatures for paleosol carbonates. We investigate the relationship between seasonally variable soil temperatures and carbonate formation temperatures estimated using clumped isotope thermometry (T(Delta(47))). Holocene soil carbonates were collected along the eastern flank of the Andes (33 degrees S) over 2 km of relief and a nearly 15 degrees C range of mean annual air temperature (MAAT). This area receives both westerly, winter season precipitation and easterly, summer season precipitation, leading to a reversal of soil moisture regimes that occurs at similar to 2 km elevation. In instrumented pits sampled to 1 m depth, we do not observe systematic variation in T(Delta(47)) with depth or with elevation-dependent MART. Average T(Delta(47)) values for soil carbonates collected above 2 km elevation reflect summer soil temperatures. In contrast, T(Delta(47)) values below similar to 2 km track mean annual soil temperature. These results may reflect the dominance of summer precipitation below similar to 2 km, which likely delays soil drying and carbonate growth until fall. This seasonal variability in carbonate formation greatly impacts carbonate formation temperature, with important implications for estimating soil water delta O-18 values. Calculating soil water delta O-18 values using T(Delta(47)) produces a soil water delta O-18 lapse rate of -4.6 parts per thousand/km, remarkably similar to that of modem surface waters (-4.8 parts per thousand/km), and significantly improving on previous soil water estimates assuming MAAT for carbonate formation temperatures. Although seasonal bias can prevent the straightforward translation of soil carbonate T(Delta(47)) into meaningful surface air temperatures, direct constraints on temperature from clumped isotopes can both provide a window into soil processes and aid in the interpretation of conventional stable isotopic data to reconstruct surface conditions in the past. (C) 2012 Elsevier B.V. All rights reserved.