Unraveling the mechanisms behind the triple isotopic composition of dissolved oxygen

Emeline Clermont , Ji-Woong Yang , Thomas Extier, Didier M. ROCHE

Estimating the evolution of biospheric oxygen production through time is a major challenge. The triple oxygen isotope composition of molecular oxygen (17Δ) has been proposed as a tracer of this production, as it integrates contributions from marine (17Δocean) and terrestrial (17Δterr) biospheric oxygen. The recent implementation of 17Δocean in the intermediate-complexity climate model iLOVECLIM now enables a mechanistic exploration of this proxy. This study aims to improve our understanding of the behavior and limitations of 17Δocean as an indicator of biological oxygen production across spatial scales. At large scales, the 17Δocean signal primarily reflects residual processes affecting dissolved oxygen, with photosynthesis exerting dominant control on its magnitude and variability. Our results further highlight the strong sensitivity of the simulated signal to biological isotopic fractionations. Sensitivity experiments based on a Latin hypercube approach constrain the 17O/18O fractionation during respiration to 0.517. While the model captures large-scale patterns and overall variability of δ18O-O2 and 17Δocean, it cannot reproduce their local variability. This mismatch reflects subgrid-scale variability that the model cannot resolve, as it represents grid-cell averages rather than individual observations. Consequently, no combination of biological fractionation coefficient fully reproduces the observed 17Δocean signal, pointing to both model limitations and observational uncertainties. Overall, these results confirm the first-order sensitivity of 17Δocean to photosynthetic processes, while indicating that its finer interpretation in global models remains uncertain. This calls for a reassessment of the isotopic and biological processes governing this proxy, and for improved observational constraints.