The Ocean-Atmosphere pCO Gradient

Broecker and Peng's conjecture about the ocean's role in the natural CO cycle rests on one critical point. The North Atlantic must currently be taking up an unusually large amount of CO in order to satisfy the CO uptake demanded by both the natural cycle and anthropogenic uptake. Since the North Atlantic is not especially large in terms of area, the partial pressures of CO in North Atlantic surface waters must be especially low in order to take up the required amount of carbon. This should be easily measurable by current techniques.

A global compilation of the partial pressure of CO (pCO) in ocean surface waters by Taro Takahashi, one of the Tans et al. (1990) authors, limits the CO sink in the North Atlantic to 0.53 GtC/yr. Because the North Pacific has no deep water formation its capacity to take up carbon is small. Takahashi's estimate for the North Pacific is only 0.06 GtC/yr. This leaves a total northern hemisphere ocean sink of about 0.6 GtC/yr. Maintaining an interhemispheric CO transport consistent with 2 GtC/yr of ocean uptake requires a northern hemisphere ocean sink more like 1.5 GtC/yr (Keeling et al., 1989). This is clearly not supported by the observations on hand. The limited ability of the North Atlantic to take up CO is a cornerstone of the Tans et al. argument.

The North Atlantic is particularly famous for its large spring phytoplankton blooms which remove CO from the water and draw down the pCO. The CO uptake during spring blooms is capable of pulling the oceanic pCO well below atmospheric levels and is strong enough to overcome the thermodynamic relationship between pCO and temperature (Watson et al., 1991; Takahashi et al., 1993). Watson et al. suggest that undersampling of North Atlantic spring blooms might significantly change the North Atlantic carbon balance, but it is doubtful that the effect of undersampling could be as large as 1 GtC/yr.

At the risk of confusing the reader, it is necessary to point out some of the uncertainty surrounding the estimation of ocean- atmosphere CO fluxes by pCO. Converting an ocean-atmosphere pCO difference into a CO flux requires knowledge of the rate at which gasses are exchanged across the air- sea interface. Ocean-atmosphere gas exchange rates are currently uncertain by a factor of two (Wanninkhof, 1992). Tans et al. (1990) combined gas exchange estimates with average ocean- atmosphere pCO differences over 6 latitude bands to determine the ocean-atmosphere CO flux as a function of latitude. The Tans et al. compilation yields a total ocean uptake of 1.6 GtC/yr which is skewed toward the 15-50 latitude band in the southern hemisphere.

The Tans et al. pCO compilation puts the main CO sink in the 15-50 latitude band during southern winter, a time of year for which very few observations were available before 1990. Tans et al. ultimately chose to downplay the poorly constrained pCOs from the southern hemisphere and to focus on the better constrained pCOs from the North Atlantic. Subsequent work has shown that this was a good strategy: poorly sampled areas of the South Pacific and South Indian Oceans do not yield the large CO sink implied by the original pCO compilation. Murphy et al. (1991) showed that the western South Pacific is a CO sink, but the eastern South Pacific is a CO source. Similarly, work in the southern Indian Ocean (Metzl et al., 1991) has shown that the areas of CO uptake are also negated by other areas of CO source.