Detecting hydrologic distinctions among Andean lakes using clumped and triple oxygen isotopes

Abstract

Oxygen isotope distributions from lacustrine carbonates provide insights into climate and hydrological change, but it is difficult to isolate the influences of catchment precipitation δ18O, water temperature, and evaporation on lacustrine carbonate δ18O values. Recent work shows the potential for using a combination of clumped (Δ47) and triple oxygen isotope (Δ′17O) measurements to identify the roles of temperature and evaporation on carbonate δ18O values in lakes, allowing precipitation δ18O values to be inferred and facilitating paleoclimate reconstructions. However, modern calibration of this approach has been mostly limited to arid regions with a high ratio of evaporative losses over inputs (XE) and low relative humidity (h<0.7). Developing this tool for paleoclimate and paleoelevation reconstructions requires expanding the modern calibrations to a greater range of climatic and hydrologic conditions. We sampled four lakes in different hydrologic states under a single, high humidity climate regime (h=0.7–0.9) in the Lake Junín region of central Peru. Clumped isotope temperatures from lake carbonates reflect water temperatures during carbonate formation. Lake hydrology is the main control on the Δ′17O values of carbonates and waters: Δ′17O values are lowest in the larger lakes with higher XE when compared to smaller, headwater lakes where evaporation is minimal and Δ′17O is indistinguishable from that of precipitation. Reconstructed unevaporated catchment precipitation δ′18O (δ′18Orucp) values from lake waters rely on accurate characterization of λlake, the triple oxygen isotope evaporation slope. We explore the influence of humidity on λlake using both new observations and modeled data. Accounting for local humidity improves λlake estimates, which allows for more accurate reconstructions of δ′18Orucp. We generate a δ′18Orucp value of −15.2±2.1‰ from modern carbonates and waters (n=15) in the Lake Junín region, which is similar to amount weighted mean annual precipitation −14.1 (±2.2‰). This study illustrates that (1) Δ′17O can be used to differentiate between lakes with differing XE in humid climates, (2) lake carbonate Δ′17O and δ18O values can be used to evaluate the influence of evaporation on lake water δ18O values in a range of climates, and (3) modeling λlake under appropriate humidity conditions improves δ′18Orucp estimates from lake carbonate Δ′17O.