Seeing through the evaporation effects on water δ18O
Sedimentary lake deposits from the Eocene of the western. U.S. can serve as archives of past tectonics because oxygen isotopes (δ18O) in authigenic minerals can reflect past rainfall δ18O, which tends to decrease with elevation. However, in many lakes the progressive loss of water via evaporation increases δ18O, perhaps masking the elevation signal. Fortunately, evaporation also impacts δ17O which, together with δ18O, makes up the “triple oxygen” isotope system (Δ17O). By measuring Δ17O in lacustrine sediments (here, chert) it is therefore possible to account for evaporation and estimate the δ18O of precipitation falling into the lake.
Estimating precipitation δ18O and Eocene elevation
We use the “back-projection” model of Passey & Ji (2019) to help estimate past precipitation δ18O. The model simulates data in δ18O-Δ17O space, intersecting the Meteoric Water Line at the “un-evaporated” δ18O value. New carbonate clumped isotope data helps us compare the lake-water δ18O reconstructed from carbonates and cherts. We then use the adiabatic uplift model of Rowley et al., (2001) to estimate paleo-evolution δ18O constraint.
-
![]()
Chert and carbonate isotope systematics
We identify a decrease in Δ’17O with increasing chert δ18O, consistent with evaporation controlling δ18O variability. We also find that oxygen isotopes of cherts and carbonates are hard to reconcile with “equilibrium” conditions—i.e. these minerals likely formed in waters with different isotopic compositions and/or different temperatures.
-
![]()
Two scenarios of precipitation δ18O
We don’t have direct temperature constraints on chert formation, so to account for temperature uncertainty we consider two scenarios for calculating precipitation δ18O. In the first, cherts form at just over 30°C, consistent with carbonate clumped-derived temperatures of overlying rocks. In the second, cherts form at 60°C during early burial diagenesis. The second scenario is consistent with evaporative lake waters and more similar to clumped-derived water δ18O from carbonates. We use the Passey & Ji (2019) equations to “back project” from evaporated lake waters to unevaporated meteoric waters, and estimate precipitation δ18O = ~-16‰.
-
![]()
Sensitivity to temperature, the MWL, and initial δ18O
We run a series of sensitivity tests to account for three sources of uncertainty. First, we find that the estimated precipitation δ18O is not very sensitive to the temperature of chert formation. The sensitivities to the Meteoric Water Line intercept and the initial (shoreline) precipitation δ18O, however, are higher. Fortunately, these sources of uncertainty can be improved with future work.
