On the Δ17O budget of atmospheric O2

We modeled the Δ17O of atmospheric O2 using 27 ordinary differential equations comprising a box model composed of the stratosphere, troposphere, geosphere, hydrosphere and biosphere. Results show that 57% of the deficit in 17O in O2 relative to a reference water fractionation line is the result of kinetic isotope fractionation attending the Dole effect, 33% balances the positive Δ17O of O(1D) in the stratosphere, and 10% is from evapotranspiration. The predicted Δ′17O O2 relative to waters is −0.410‰ as measured at the δ18O of air. The value for Δ′17O O2 varies at fixed δ18O with the concentration of atmospheric CO2, gross primary production, and net primary production as well as with reaction rates in the stratosphere. Our model prediction is consistent with our measurements of the oxygen isotopic composition of air O2 compared with rocks if rocks define a fractionation line with an intercept in δ′17O = 103ln(δ17O/103 + 1) vs. δ′18O = 103ln(δ18O/103 + 1) space less than SMOW but more positive than some recent measurements imply. The predicted Δ17O is less negative than that obtained from recent measurements of O2 directly against SMOW. Underestimation of Δ′17O O2 can only be ameliorated if the integrated (bulk) Δ′17O for stratospheric CO2 is significantly greater than measurements currently allow. Our results underscore the need for high-precision comparisons of the 17O/16O and 18O/16O ratios of atmospheric O2, VSMOW, and rocks.