The Use of Tracers in Modelling the Oceanic Uptake of Carbon Dioxide

Increasing levels of atmospheric carbon dioxide from the burning of fossil fuels and changes in land use pose a threat of significant global climatic change in the 21st century. Owing to uncertainty in the pre-industrial atmospheric CO₂ concentration and in CO₂ releases from land-use change, direct estimates of the airborne fraction of this man-made CO₂ are not well established. Effort has therefore been devoted to estimating the strengths of the oceanic and biospheric sinks as an alternative route to determining the airborne fraction. This paper reviews the development of oceanic CO₂-uptake models. One-dimensional box models, with first-order exchange kinetics or vertical diffusion to simulate CO₂ penetration into the deep ocean and calibrated against natural ¹⁴C distributions, appear inadequate. Their uptake is too small to be consistent with the recent atmospheric record and most estimated histories of CO₂ release from land use change. Models incorporating representations of specific oceanic mixing processes important in CO₂ uptake achieve somewhat larger CO₂ uptake, especially when calibrated against `short timescale' tracers, such as radiocarbon and tritium derived from bomb tests. Despite this general conclusion, substantial differences between such models remain. A comparison between my two-dimensional advection-diffusion model and Siegenthaler's outcrop-diffusion model illustrates how the relative importance of air-sea gas exchange rate and rate of ocean mixing in limiting CO₂ uptake depends critically upon modelling assumptions. The failure of most models calibrated with a single oceanic tracer to reproduce well the distribution of other tracers has encouraged the development of multi-box, geographically realistic, models whose circulation and mixing patterns are determined by simultaneous inverse solution of a set of conservation equations for a range of tracers. This technique, when augmented with additional dynamical constraints, probably offers the most promise for advancing CO₂-uptake modelling while suitable three-dimensional oceanic general circulation models are being developed. The latter, atmospherically driven, models will eventually play a key role in assessing how any future climatic change may feed back on atmospheric CO₂ levels. Feedback could arise either by alteration of the mixing processes responsible for man-made CO₂ uptake, or more fundamentally if changes in the surface-ocean productivity result from changes in circulation-mediated nutrient supply.