A 3-D Model of the Ocean Carbon Cycle: Sensitivity to Mixing Based on Bathymetry Roughness

Earlier as well as more recent model circulation assessments using natural C-14 have revealed the possibility for large differences between ocean models in simulating deep ocean C-14 penetration, for example in the North Atlantic where many models appear to have too slow deep circulation resulting in an overestimation of the ventilation timescales for deep water renewal. Significant differences between models in simulating the vertical modes of circulation, many at variance with observations, have also been shown in recent studies using CFC-11 as a chemical tracer. At the Canadian Climate Centre, an ocean carbon model is being developed and tested as part of a coupled carbon general circulation model for use in projections of future climate change. Experiments with the abiotic carbon model that included C-14 and CFC-11 exhibited similar tendencies: the mid to deep ocean circulation, especially in the North Atlantic, appeared to be too slow. As a remedy, a novel mixing scheme was incorporated in the model. It involves spatially variable eddy diffusivity as a function of sub-grid scale bathymetry roughness. Model experiments using this scheme instead of constant background eddy diffusivity resulted in noticeable improvement in the North Atlantic and the Pacific ocean circulations as diagnosed by the Delta-C14 ratio and CFC-11. Model results were compared to data.