TRIPLE OXYGEN ISOTOPES IN SOILS AND LAKES
We present Δ17O data on soil carbonates from North America, East Africa and East Asia and build a dynamic model to simulate triple oxygen isotopic compositions in soil water. Our model shows that (1) persistent evaporation will result in lowered Δ17O values of soil water and (2) these 17O-depleted signals in soil water diminish sharply as depth increases. The model results are supported by our new data. For example, the triple oxygen isotopic compositions of deep soil carbonates (>50 cm) are similar to or yield slightly lower Δ17O values than meteoric water.
We also develop a model to predict triple isotope compositions of lake waters for different hydrological scenarios. Our model predicts that lake waters with the same δ18O value may have different Δ17O values due to different relative influences of isotopic compositions in precipitation, lake water mixing and evaporation. Through-flowing lakes with minimal evaporation should have Δ17O values similar to local precipitation, whereas highly evaporated lakes should have distinctly negative Δ17O values.
In summary, our models and preliminary data suggest that Δ17O will be a valuable addition to the terrestrial paleoclimate toolkit, most notably for its ability to record the effects of aridity.