Paleotemperature Estimation by Tandem δ18O Measurement of Biogenic Carbonate and Gypsum Hydration Water

One of the fundamental problems in oxygen isotope paleothermometry is the carbonate mineral-water temperature equation is often under constrained. The δ18O of calcite or aragonite can be measured on fossil shell material but a unique temperature solution is not possible without knowing the δ18Owater from which the carbonate precipitated. Gypsum (CaSO4*2H2O) is a hydrated mineral that contains 20.9% water by weight. The isotopic composition of the water from which the gypsum was precipitated can be calculated from the δ18O and δD of gypsum hydration water provided: (i) the fractionation factors between mother water and crystal water are known; and (ii) no further exchange has occurred between environmental and crystal water after deposition. Here we propose that tandem δ18O measurements of both carbonate and gypsum hydration water in the same samples offers a power tool for determining past temperature. We demonstrate the method using co-occurring gypsum and biogenic carbonate from sediment cores from Lake Peten Itza, Guatemala (17oN, 90oW). Oxygen and hydrogen isotopic results of interstitial and gypsum hydration water from glacial-aged deposits in Lake Peten Itza suggest the gypsum hydration water has not exchanged with sediment pore water. The δ18O and δD of lake water, calculated from gypsum hydration water and known fractionation factors, fall on a projection of the evaporative line (slope of ~5) defined using modern lakes in the region. Fifteen paired measurements of ostracods and gypsum hydration water in the same samples gave a mean temperature of ~16oC (range 13 to 19 oC) for the Late Glacial (18 to 10 ka), which is 9oC cooler than mean annual temperature in the region today (~25oC). The low glacial temperatures likely reflect winter climate that sets hypolimnetic temperature in warm, monomictic lakes. Our preliminary results support previous findings of much greater glacial tropical cooling during winter on land in Central America than indicated by marine proxies in the nearby Caribbean Sea. Accuracy of the method could be improved by verifying the oxygen and hydrogen isotope fractionation factors between mother water and gypsum hydration water determined in the 1960s and by culturing experiments of the ostracod species (Limnocythere opesta) used in this study to determine the offset from oxygen isotope equilibrium.