Michael Everett Wickett

The Gosac Project to Predict the Efficiency of Ocean CO2 Sequestration Using 3-D Ocean Models

Publisher Summary This chapter presents a set of standard injection simulations that were made by eight ocean modeling groups to evaluate the efficiency of the ocean in retaining purposefully sequestered CO 2 . Injection was made simultaneously at seven separate sites; separate 7-site simulations were made for injection at 800 m, 1500 m, and 3000 m. For injection at 3000 m, all models showed 85% or greater global efficiency in year 2200—that is, 100 years after the end of the specified 100-year injection period; at the same time, the 1500-m injection is 60-80% efficient and 800-m injection is only 42-61% efficient. Most of the CO 2 injected at 3000 m was lost from the Southern Ocean (the principal region by which the deep ocean is ventilated); at shallower depths, relatively more was lost sooner, from the northern hemisphere and the tropics. The simulated global injection efficiency at 3000 m is correlated with both the simulated global mean CFC-11 inventory and deep-ocean natural inc. Based on these correlations, the global observational constraints for these two tracers, and model diversity, it appears likely that the range of model-predicted efficiencies would bracket real ocean behavior under the same 3000-m injection scenario.

Narrowing the uncertainty for deep-ocean injection efficiency

Publisher Summary The chapter proposes a basic ground rule for future studies of ocean injection efficiency: to be credible they must also demonstrate the associated models skill in simulating the global inventory of GFG-11 and the global mean for radiocarbon in the deep ocean. A model that performs well in regards to both those constraints will be more likely to simulate reasonable global injection efficiencies. Nonetheless, efficiencies for a given injection site in coarse resolution models could be biased. For instance, the majority of injection sites will be located on eastern or western boundaries, which have known problems in coarse resolution models. Furthermore, coarse-resolution grids are unable to resolve important subgrid-scale processes (e.g., eddies, boundary currents, convection). Properly accounting for these processes may affect large-scale transport and could alter model predictions of CO2 sequestration efficiency. Although, global-scale ocean general circulation models are now becoming available which do resolve these processes, their high resolution means that they can only be integrated for relatively short periods, a few decades at most.