13 Some Remarks on a Control Volume Experiment for North Atlantic JGOFS

James F. Price

Woods Hole Oceanographic Institution

Department of Physical Oceanography

Woods Hole, MA 02543

The budget equation of carbon integrated over an upper ocean control volume will include source/sink terms due to chemical and biological processes, as well as horizontal and vertical advective terms. While the advective terms are extremely important for some aspects of the carbon budget (for example, the advective terms couple the carbon budget to the general circulation), it remains that for the purpose of deducing the source/sink terms the advective effects are mainly a noise process that obscures the signal of most interest.

Some idea of the advection that can be anticipated in the Bermuda region can be formed from the historical SOFAR float data set (Owens, 1991) some of which, especially at 700 m, reaches the Bermuda area. The float data suggest that Bermuda is near the southern edge of the Gulf Stream recirculation gyre, and that the mean current is toward the west-southwest at a speed of about 0.05 m/sec. The rms speed of the currents associated with mesoscale eddies is somewhat larger, about 0.15 m/sec, so that the instantaneous current could be in almost any direction. The dominant eddy period is about two months, so that mesoscale eddies produce particle displacements during a period of a month (roughly the duration of a complete bloom event) of very roughly 200 km. If there is substantial variation of the tracer on such a scale, then advection is likely to be an important term in the budget equation and a three-dimensional control volume experiment would be required to sort out advective effects.

A control volume experiment could seek to minimize the noise effects of advection by making measurements over a mesoscale-sized region (O(200 km) on a side) and in a way that allows the advection terms to be estimated directly. This would require measurements of both the advecting velocities and of the tracer field itself. The former could be greatly enhanced by the use of satellite altimetry, though the tracer itself is likely to be measurable at depth only by in situ methods.

The resources required to carry out a control volume experiment are sure to be substantial, and will have to be justified by careful planning involving, among other things, the use of high resolution numerical models to simulate field experiments. The key question is whether the effort required to observe a three dimensional budget will result in a substantially improved estimate of the source/sink terms. Said differently, can advective effects be estimated to within an error bound that is significantly smaller than the present uncertainty on the source/sink terms? If the answer is found to be yes and the required level of effort is affordable and manageable, then a control volume experiment offers a means to advance our knowledge of the carbon cycle.

The discussion above has been cast in a rather elementary form emphasizing that advection is a noise to be compared to the signal of source/sink terms. It could turn out that a three-dimensional experiment would reveal systematic three-dimensional circulation effects involving, say, the Gulf Stream recirculation, that could be of far greater interest than implied by a random noise process alone.