Pliocene and late Miocene soil temperatures in the Chinese Loess Plateau based on clumped-isotope thermometry of paleosol carbonates

We apply the carbonate clumped-isotope paleothermometer—a proxy based on the relative abundance of 13C-18O bonds in carbonate ions—to Pliocene and late Miocene soil carbonate from the Baode area of the Chinese Loess Plateau (CLP). Carbonate clumped-isotope temperatures of recent temperate-latitude soil carbonates are similar to warm-season temperatures, suggesting that soil carbonate forms primarily during the warmest months of the year. We therefore surmised that this proxy could help to address the following question: did putative elevated mean annual temperatures during the Pliocene result from a seasonally-uniform increase in temperature, or was the shift mainly related to cool-season temperatures and warmer winters than present. If the latter scenario is correct, then Pliocene summer temperatures would have been similar to present summer temperatures, and thus carbonate clumped-isotope temperatures of Pliocene paleosol carbonates should also be similar to present summer temperatures. Clumped-isotope Δ47 values of 19 time-horizons spanning 7.2 to 2.7 Ma at Baode (northern Loess Plateau) and 9 time-horizons spanning 6.2 to 2.7 Ma at Lantian (southern Loess Plateau) range between 0.652‰ and 0.681‰ at Baode, and 0.651‰ and 0.659‰ at Lantian. These values imply temperatures of soil carbonate mineralizastion ranging between 23.7°C and 17.4°C, averaging 20.0°C at Baode, and 23.9°C and 22.2°C, averaging 23.2°C at Lantian. These temperatures are similar to, or cooler than, modern warm-season temperatures in the Loess Plateau (JJA mean temperature = ca. 22°C at Boade and 26°C at Lantian). In a broader sense, our results can be explained by one or more scenarios, including: 1) the paleosol carbonates formed during warm seasons, and these were not warmer than today and 2) the paleosol carbonates did not form during warm-season peak temperatures, and, for example, formed later in the season when soils dewatered following wet summer monsoons. Scenario 1 is consistent with the existing calibration of the clumped-isotope system in temperate soil carbonates, but we point out that this calibration is biased towards arid climates where soil dewatering during the warm season is likely. Therefore, extension of the calibration to wetter monsoon climates may help to evaluate the plausibility of scenario 2.