Inversion of diabatic and adiabatic controls of ocean stratification using Argo profiles.

The idea that turbulent transport rates can be inferred from observed characteristics of the broad-scale hydrography goes back to Iselin and likely even further. The main focus of this talk is on the upper 2000m of the global ocean where extended observation of stratification by Argo thus provides a path to improved estimates of regional turbulent transport rates, as demonstrated by results from the ECCO v4 (Estimating the Circulation & Climate of the Ocean, version 4) ocean state estimate that covers 1992-2011. The estimated parameters are time-invariant three-dimensional maps of bolus velocity coefficient, isopycnal diffusivity, and diapycnal diffusivity. The parameter adjustments are consistent with the observed ocean stratification by construction, which is well reproduced by ECCO v4. It is shown that ocean stratification is highly sensitive to the estimated parameter adjustments, and that the geography of estimated parameter adjustments is aligned with contours of observed ocean stratification. Thus, the constraint of fitting Argo profiles is identified as an effective observational basis for the inversion of turbulent transport parameters. As part of the inversion method evaluation it is shown that the estimated parameter adjustments reduce spurious model drifts in multi-century simulations. Notably, they yield major improvements in simulating biogeochemistry variables that were not involved in the parameters optimization. The estimated parameter adjustments themselves and the resulting adjustments in ocean stratification are physically plausible. The results provide observational evidence to support the notion of a ventilated thermocline, responding primarily to adiabatic circulation controls, embedded in an internal thermocline where diffusion plays an increased role.