The response of thermocline ventilation to variability at the ocean surface from observations and offline tracer modeling.

摘要

Variability in oceanic ventilation can arise from either changes at the surface of the ocean or the ocean interior. Four studies are presented which advance our understanding on how these changes can be diagnosed in both observational and modeling contexts.;Chlorofluorocarbons (CFCs) and sulfur hexafluoride (SF6) have been used extensively to infer transit time distributions (TTDs) and ventilation ages in the ocean. An offline tracer model (Offtrac) is combined with a simple model of gas exchange to simulate the mixed layer saturations of CFC-11, CFC-12, and SF6. The large wintertime undersaturations of these tracers arise from the increase in solubility due to the cooling of the mixed layer and also from the entrainment of relatively tracer-poor waters as the mixed layer deepens. In the mode waters of the North Pacific, this can cause a bias TTD mean ages of up to 24%.;The Antarctic Circumpolar Current (ACC) is a strong dynamical feature in the Southern Ocean which transports water around the entirety of the Antarctic continent. Monte-Carlo simulations of a meandering Gaussian jet model in conjunction with distributions of sea level anomaly from 1992 to 2014 are used to determine the mean position and width of the fronts that form the boundaries of the ACC. The mean position of these fronts largely follow the underlying topography. Significant internannual variability in the location of the fronts was uncorrelated to changes in the Southern Annular Mode (SAM).;Offtrac is used to simulate CFCs, SF6, oxygen, ideal age, and transit time distributions using a boundary impulse response technique (TTD-BIR). The output from these simulations are used to evaluate how well tracers can constrain the timescales of oceanic ventilation. The inverse Gaussian solution to the 1d transport equation is shown to be a reasonable approximation to the TTD-BIR within the ventilated thermocline of the subtropical gyres, but a poor approximation in regions with strong gradients in age. 1d TTDs constrained by modeled CFC-12 and SF6 have a strong bias towards diffusively dominated transport. By comparing variability in oxygen and tracer-inferred TTD mean ages to changes in ideal age, guidelines are developed as to where observations of these tracers may robustly diagnose changes in ventilation.;The effect of how variability of a tracer at the surface maps to changes in the oceanic interior is examined using an analytic Fourier transform of the 1d TTD. The magnitude and phase of the Fourier coefficients of the 1d TTD demonstrate that interior variability is a result of ventilation behaving like a low-pass filter with nonlinear phase response applied to a time-varying surface boundary condition. Typical values of the parameters of the filter for the thermocline suggest that primarily the low-frequency (>5 years) part of the time-varying boundary condition are likely to observed within the interior.;Both climatological and hindcast adjoint TTDs are calculated in Offtrac to understand variability of the formation of North Pacific Mode Water (NPMW) and South Indian Ocean Subantarctic Mode Water (ISAMW). Using volume fraction (Vf) and mean ages calculated from these adjoint TTDs, NPMW is found to be formed within one region of the North Pacific, whereas ISAMW likely has at least two formation regions, one within the Indian Ocean and another southwest of Australia. Empirical orthogonal function analysis shows that about 54% of ISAMW Vf variability is captured by a mode, uncorrelated to SAM (p<0.1), with a 30-year oscillation between the two climatological formation regions. For NPMW, 69% of the variability is explained by a meridional mode centered around 35°N in the central North Pacific that is significantly correlated ( p<0.05) with the Pacific Decadal Oscillation.