Alan Cohn

AOSC 658a

Carbon Cycle & Climate



The Anthropogenic Ocean Carbon Sink


Due to oceanic and land sinks, less than half of the CO2 emissions from the industrial period remain in the atmosphere.  Difficulty arises when trying to attribute how much of the sink is due to land and how much is due to the ocean.  The O2/N2 technique used in the 2001 IPCC report has come under scrutiny in recent studies, particularly due to inconsistencies in observed oceanic oxygen concentrations with the CO2 inventory.  A more recent study by McNeil et al. based their oceanic CO2 inventory on the distribution of chlorofluorocarbons (CFCs) in the ocean.  Their method used CFC concentration to estimate the “age” of the water, or time since it was last in contact with the surface.  CFC and anthropogenic CO2 concentrations are different due to differences in carbonate chemistry, solubility, and air-sea gas exchange.  Thus, they combined their estimates of water mass ages with the atmospheric CO2 history and carbonate chemistry equilibrium equations to calculate changes in dissolved organic carbon (DIC).  Their estimates of anthropogenic CO2 accumulations from 1980 to 1999 are seen in Figure 1.  However, since CFCs are a relatively new addition to the ocean, this method is insufficient to assess CO2 concentrations in older waters.  For another inventory of anthropogenic CO2 in the ocean, we look at a paper by Sabine et al.

 The most direct way of assessing anthropogenic CO2 concentrations in the ocean is to compare changes in observed DIC.  In the 1990s, the World Ocean Circulation Experiment (WOCE) and the Joint Global Ocean Flux Study (JGOFS) surveyed DIC distributions in the oceans.  To separate the anthropogenic CO2 component from the measured DIC concentrations, the carbon tracer ∆C* method was used.  Sabine et al. estimated a global anthropogenic CO2 sink of 118 ± 19 Pg C.  The North Atlantic had the highest vertically integrated concentrations of CO2 with 23% of the global oceanic anthropogenic CO2.  The Southern Ocean south of 50°S, on the other hand, only contains 9% of the global inventory.  However, the Southern Hemisphere oceans contain about 60% of the total oceanic anthropogenic CO2 inventory, mostly owing to its immense area.  Figure 2 depicts the global distribution of vertically integrated anthropogenic CO2 concentration.

            The highest concentrations of anthropogenic CO2 are found in near-surface waters, since CO2 enters the ocean by air-sea gas exchange.  Variations in the concentrations of CO2 at the surface are related to how long the water has been exposed to the atmosphere and the Revelle factor, which describes how the partial pressure of CO2 in seawater changes for a given change in DIC.  The capacity for ocean waters to take up anthropogenic CO2 is inversely proportional to the Revelle factor; the highest anthropogenic CO2 concentrations are found in the subtropical Atlantic surface waters due to the low Revelle factors there.  North Pacific surface waters have a high Revelle factor and therefore lower anthropogenic CO2 concentrations.  The Revelle factor limits uptake so that the CO2 inventory in the ocean is significantly lower than what it would be if one was to neglect its influence.

            Most anthropogenic CO2 is found in the thermocline.  Concentrations at depth are determined by how rapidly the near-surface anthropogenic CO2 is transported into the ocean interior along surfaces of constant density, or isopycnals.  The deepest penetration occurs at convergence zones in middle latitudes.  Low vertical penetration is found in upwelling regions, such as the Equatorial Pacific.  Another factor for CO2 concentration in the ocean interior is the atmospheric CO2 concentration when the water was last in contact with the surface; lower atmospheric CO2 concentrations in the past resulted in lower oceanic CO2 concentration than today’s surface water.

            Other factors contributing to the large CO2 concentrations in the Southern Hemisphere include the formation of Antarctic Intermediate Water (AAIW) and sub-Antarctic Mode Water (SAMW), along with high winds and low initial anthropogenic CO2 content of the water.  Thus, high uptake and rapid sinking contribute to the large CO2 inventory.  The attribution of the anthropogenic signal in the North Pacific, however, is not as easy to distinguish due to the complex interplay of different intermediate-type waters.  In the Indian Ocean, anthropogenic CO2 penetrates deep due to young, dense water, high in anthropogenic CO2, from the Red Sea and Persian Gulf.  In the North Atlantic, rapid sinking of water allows anthropogenic CO2 to penetrate to mid and abyssal depths.  As these waters move southward, the anthropogenic CO2 concentration decreases as mixing with older water occurs.  Although deepwater formation also occurs near Antarctica, there is little or no anthropogenic CO2 in Antarctic Bottom Water.  The likely reasons for this are a very high Revelle factor, limited contact with the surface due to rapid sinking and the presence of sea-ice, and dilution with older waters.

            About two-thirds of anthropogenic CO2 emissions have remained in the atmosphere.  The results of Sabine et al. suggest that land has been a net source since 1800 AD, whereas the ocean has been the only true net sink for anthropogenic CO2.  They estimate that atmospheric CO2 would be about 55 ppm higher today that presently observed if it weren’t for oceanic uptake.  The question now becomes, will the ocean continue to take up CO2 at its current rate?  There have already been signs of slowing—the uptake fraction has decreased from 28-34% to about 26%.  The slow mixing time of the ocean already limits its rate of uptake.  Given thousands of years, the ocean would likely absorb about 90% of anthropogenic CO2 emissions, but the current fraction is less than 30%.  In essence, the ocean cannot keep up with us.  However, both positive and negative feedbacks may take effect.  These can result from changes in the circulation, stratification, and biology of the ocean.  McNeil et al. suggest that as the ocean warms and becomes more stratified, transport into the ocean interior will slow, but this will not have much of an effect on oceanic uptake.  Chemical feedbacks, which are mostly positive, may be more significant.  If PCO2 concentrations of the surface ocean continue to increase in proportion with the atmospheric PCO2 increase, a 2XCO2 scenario would result in about a 30% decrease in carbonate ion concentration and thus an increase in the Revelle factor.  This, in turn, would decrease the ocean’s ability to absorb CO2.

            For a model simulation of anthropogenic CO2 uptake, we look at a study by Fung et al.  Their model includes, in simplified form, the main processes for the solubility carbon pump, organic and inorganic biological carbon pumps, and air-sea CO2 flux.  They use two emission scenarios (Figure 3) along with the historical emissions trajectory for the 19th and 20th centuries.  One scenario assumes balanced energy sources; this is A1B from the Special Report on Emission Scenarios (SRES).  In A1B, fossil fuel emission increases until 2050 AD and decreases thereafter.  The other scenario, called A2, assumes business-as-usual; emission increases exponentially.  Here, carbon sequestration in the ocean and on land cannot keep up with the emissions.  In addition, the capacities of both sinks diminish as CO2 increases.  In A1B, mixing of CO2 into the deep ocean maintains a slower surface ocean CO2 increase, so that the oceanic sink steadily increases.

            The authors experiment with models that did (ROL and RO) and did not (OL and O) include carbon-climate coupling.  Models runs looked at the oceanic sink without the land sink.  The coupled model that included land and ocean for “balanced energy sources” scenario A1B output an atmospheric CO2 concentration in 2100 of 661 ppmv and a temperature increase of 1.21 K.  The oceanic CO2 fraction, as compared to the airborne and land CO2, is about 24%.  For the “business-as-usual” A2 scenario, coupled and accounting for land and oceanic sinks, atmospheric CO2 concentration is 792 ppmv in 2100 with a temperature increase of 1.42 K.  In this case, oceanic CO2 fraction is about 21%.  See Table 1 for more detailed figures.

The results of comparing the A2 scenario with and without land suggest that the ocean circulation slows, thus reducing uptake.  However, the net biological effect is to lower surface pCO2 and increase ocean uptake, which partially compensates for the slower circulation.  The models did not, however, include ocean acidification and the increasing Revelle factor.  This could significantly decrease oceanic uptake, leading to higher atmospheric CO2 concentrations and higher temperatures in both scenarios.  Regardless of the role of the oceanic CO2 sink, it appears from these studies that atmospheric CO2 concentrations will be much higher than today, and global temperatures will most likely continue to increase well into the 21st century without significant measures to reduce or negate emissions.


Figure 2. 




Fung et al. (2005)  Evolution of carbon sinks in a changing climate.  PNAS, 102, 11201-11206.


Gruber, N., Sarmiento, J.L., and T. Stocker (1996)  An improved method for detecting anthropogenic CO2 in the oceans, Global Biogeochemical Cycles, 10, 809-837.


McNeil et al. (2003)  Anthropogenic CO2 Uptake by the Ocean Based on the Global Chlorofluorocarbon Data Set, Science, 299, 235-239.


Sabine et al. (2004)  The Oceanic Sink for Anthropogenic CO2, Science, 305, 367-371.