Exploring Continental Margin Carbon Fluxes on a Global Scale

K.-K. Liu1, L. Atkinson2, C.T.A. Chen3, S. Gao4, J. Hall5, R.W. Macdonald6, L. Talaue McManus7, R. Quiñones8

  1. National Center for Ocean Research, Taipei, Taiwan (Email: )
  2. Center for Coastal Physical Oceanography, Old Dominion University
  3. Institute of Marine Geology and Chemistry, National Sun Yat-sen University, Kaohsiung, Taiwan
  4. Institute of Oceanology, Chinese Academy of Sciences, Qingdao, China
  5. National Institute of Water and Atmospheric Research, Hamilton, New Zealand
  6. Institute of Ocean Sciences, Sidney, Canada
  7. Marine Science Institute, University of the Philippines,
  8. Departamento de Oceanografia, Universidad de Concepcion, Chile

Eos, Transactions, American Geophysical Union (Volume 81 Number 52)

December 26, 2000

How fast anthropogenic CO2 is being absorbed by the ocean remains one of the most critical and yet elusive fluxes sought by climate researchers. Since the ocean is a major sink of anthropogenic CO2, an accurate estimate of the present oceanic uptake rate of anthropogenic CO2 is essential for reliable prediction of the future CO2 level in the atmosphere. This number is difficult to pin down because the net air-to-sea CO2 flux is a small difference (about 2 Pg Cyr-1, 1 Pg = 1015 g) between two huge fluxes - namely, the uptake and the release of CO2 at sea surface (each about 90 Pg Cyr-1). A global ocean carbon budget helps to constrain the estimate of anthropogenic CO2 flux. To construct a complete carbon budget, we need to know not only the air-sea exchange but also the carbon fluxes from the continental margins [Liu et al., 2000]. For brevity, we use CO2 source to indicate release of CO2 from the ocean to the atmosphere and sink to indicate absorption of atmospheric CO2 by the ocean.

The riverine carbon fluxes from land sum up to about 0.8 Pg Cyr-1. Although dwarfed by the air-sea CO2 exchange, this total flux is on the same order of magnitude as the net air-to-sea CO2 transfer. Continental margins that receive riverine carbon fluxes are not merely reservoirs, which store all the discharged carbon, nor conduits, which pass everything to the interior ocean. They are sites of active physical and biogeochemical processes that transform, transport or bury carbon. Many on-going and recently completed regional studies (Fig. 1) indicate that active cross-shelf transport and biogeochemical processes in the margins influence the carbon cycle of the ocean as a whole. The findings strongly suggest that continental margins constitute important, but often neglected, components in the global carbon cycle. The intriguing role of margins in global biogeochemical cycles has been the topic of exciting new results presented and discussed at recent meetings, including the AGU/ASLO 2000 Ocean Sciences Meeting in San Antonio, Texas and the JGOFS Open Science Conference in Bergen, Norway.

Here, we highlight some of the exciting new findings and issues, and report on planned activities for data collection and synthesis. An attempt to synthesize the available information on continental margin carbon fluxes, co-sponsored by the Joint Global Ocean Flux Study (JGOFS) and the Land-Ocean Interaction in the Coastal Zone (LOICZ), two core projects of the International Geosphere-Biosphere Program (IGBP), is being conducted by the Continental Margins Task Team (CMTT). The goal is to establish a global estimate of continental margin carbon fluxes. Considering the diversity of continental margins and the different ways they may respond to global change, we seek contributions from all continental margin studies in this global synthesis effort. (Background information, reference materials, rationale and strategy for the CMTT synthesis plan, and means to participate are posted at the web site or its mirror sites which may be found at or

The fate of riverine carbon fluxes

Prior to the Industrial Revolution, the ocean carbon cycle was presumably close to steady state [Siegenthaler and Sarmiento, 1993]. If there were no carbon transfer between land and ocean, the exchange fluxes of CO2 between the atmosphere and the ocean would have matched each other. However, as pointed out by Smith and Mackenzie [1987], the global ocean must consume more organic carbon than it produces due to the input of organic carbon from land. The current estimate of the total riverine flux is about 0.8 PgCyr-1 (the dashed arrow under “Rivers” in Fig. 2), of which 45% is organic and the rest inorganic. For a balanced carbon budget before the Industrial Revolution, the CO2 release to the atmosphere must have exceeded the absorption flux by an amount equal to the riverine carbon influx minus the small burial flux of carbon to sediments. Following the Industrial Revolution, atmospheric CO2 has increased from a total amount of 600 Tg C to more than 750 Tg C at present due to fossil fuel burning and deforestation [Siegenthaler and Sarmiento, 1993]. The increased CO2 air concentration has resulted in surface seawater undersaturation on average, which renders the ocean a net CO2 sink, with the consequence that the carbon content of the ocean is presently increasing with time. Despite the increase in the atmospheric CO2, human perturbation to land-ocean carbon transfer is probably limited and, therefore, the estimate of current total riverine carbon flux is assumed applicable also to the pre-Industrial era.

The oceanic uptake rate of anthropogenic CO2, which is defined as the net carbon influx to the ocean, equals the total increase rate of the oceanic carbon pool (Fig. 2), which is 0.4 PgCyr-1 for the surface ocean plus 1.6 PgCyr-1 for the intermediate and deep waters. Because the burial rate of carbon in marine sediments is estimated at 0.2 PgCyr-1 (the dashed arrow marked as “Sedimentation” in Fig. 2) the remainder of the river input (0.8 PgC yr-1), about 0.6 PgCyr-1 (the dashed arrow from ocean to atmosphere in Fig. 2), must have been outgassed by the ocean before the Industrial Revolution [Siegenthaler and Sarmiento, 1993]. This outgassing flux, which is presumed still operating as before, offsets the downward flux of CO2 and makes the net air-to-sea transfer smaller than the anthropogenic CO2 invasion rate. Therefore, the oceanic uptake rate of anthropogenic CO2 equals the net air-to-sea flux of CO2 plus the 0.6 Pg Cyr1 from continental margins. This explains the rather modest value of the net air-to-sea CO2 flux of 1.4 Pg Cyr-1 based on earlier observations compared to the model-predicted oceanic uptake of anthropogenic CO2 at 2 Pg Cyr-1 [Siegenthaler and Sarmiento, 1993]. The latter estimate was accepted by the Intergovernmental Panel on Climate Change (IPCC) 1995 assessment of the global carbon budget (Fig. 2). Recently the revised air-to-sea CO2 fluxes based on newly obtained data of differences in CO2 partial pressure PCO2) across surface seawater from WOCE-JGOFS joint effort has been increased to as high as 2.2 Pg Cyr-1 [Takahashi et al., 1999]. If amended by the terrestrial input, the uptake rate of anthropogenic CO2 would be as high as 2.8 Pg Cyr-1, which is significantly higher than the model prediction.

One explanation offered for this discrepancy between models and observations is that the coarsely gridded PCO2 data set misses most, if not all, of the continental margins, and, therefore, it represents the air-sea CO2 exchange appropriate only to the interior ocean. If the river-discharged carbon fluxes were mostly outgassed in continental margins and little carbon got transported across the shelf, the margins would have little influence on the carbon budget in the interior ocean. Furthermore, continental margins would be a net CO2 source (0.6 Pg Cyr-1) which, for a total area of 36x1012 m2 [Liu et al., 2000], implies an average outgassing flux per unit area of 17 gCm-2 yr-1. Recent observations clearly contradict both of these implications.

Recent observations of continental margin carbon fluxes

In the past ten years continental margin studies have been conducted all over the world (Fig. 1), some of them focussing on carbon fluxes, whereas others have emphasized biology, physics or sedimentology while still shedding light on the carbon cycle. PCO2 has been surveyed on continental margins in the North Sea, the Baltic Sea and the East China Sea (ECS), all of which have a significant terrestrial input of carbon. Contrary to conventional wisdom which suggests these margins to be releasing CO2 to the atmosphere, they have all been found to be net CO2 sinks. The ECS is an instructive example. It receives an enormous amount of carbon from some of the world’s largest rivers, such as Changjiang (the Yangtze River) and Huanghe (the Yellow River). The riverine carbon fluxes comprise 12 Tg Cyr-1 in organic form and 20 Tg Cyr-1 in dissolved inorganic form [Chen and Wang, 1999]. If half of this carbon flux were outgassed evenly over the ECS shelf (0.9x1012 m2), the average efflux per unit area would be 23 gCm-2 yr-1. Instead, the ECS absorbs atmospheric CO2 with an estimated mean flux per unit area as large as 35 gCm-2 yr-1 [Tsunogai et al., 1999]. Similarly, in the Baltic Sea, the mean absorption flux of CO2 per unit area was 11 gCm-2 yr-1 for an entire annual cycle [Thomas and Schneider, 1999].

Continental shelf waters have limited capacity to store absorbed CO2 and river-discharged carbon: all but a small fraction of the received carbon must either be exported to the open sea or be buried in sediments. The Shelf Edge Exchange Processes (SEEP) projects in the Mid Atlantic Bight were conducted precisely to look for the export flux of POC from the shelf. Although SEEP was prompted by the notion that a major fraction of the primary production on the shelf could be exported, the observed export fluxes were disappointingly small, and less than 5% of primary production in total [cited in Liu et al., 2000]. The observations during SEEP projects may, however, have missed important export processes and, like the ECS and Baltic Sea, the Mid Atlantic Bight appears also to be a net sink for CO2 as indicated by recent observations by M. DeGrandpre and colleagues.

The phenomenon of the shelf sea serving as a CO2 sink has been called the “continental shelf pump” [Tsunogai et al., 1999]. The ECS shelf pump’s estimated capacity of 20-30 Mt Cyr-1 is facilitated by the active biological carbon uptake in summer and high solubility of CO2 in winter. The estimated organic carbon burial rate in the ECS shelf sediments is no more than 10 Mt Cyr-1 [Chen and Wang, 1999; S. Gao, unpublished data). It is apparent that carbon burial in sediments cannot even store all of the riverine input let alone the absorbed CO2. A large fraction must be exported from the shelf as particulate organic carbon (POC), dissolved organic carbon (DOC) and/or dissolved inorganic carbon (DIC). Advective export of POC was observed in association with a cyclonic eddy at the shelf edge northeast of Taiwan [Liu et al., 2000] during the Kuroshio Edge Exchange Processes project. Similarly strong export was also found near Cape Hatteras on the eastern margin of the North Atlantic. On the western margin of the Atlantic, observations by the Ocean Margin Exchange project showed a 15% export from the shelf primary production in the northern Gulf of Biscay [R. Wollast and L. Chou, personal communication, 2000). The shelf pump is effected by physical and biogeochemical processes which are not yet fully understood, and warrant further study.

Very effective but entirely different export processes take place on polar margins and tropical coasts. In the Arctic, ice formation produces an algal habitat and alternately stablizes or destablizes the water column when it melts or forms. The coupling of the biological cycle with the ice cycle provides an effective carbon pathway to the deeper interior ocean, one that is particularly vulnerable to change. The Mackenzie River and shelf serves as one example for the polar margin. Its coastal zone is ice free and stratified in summer; productivity is high, and organic carbon is deposited on the shelf. In winter, sea ice formation produces brine which may sink from the surface down the continental slope. Dense, cold water not only carries anthropogenic CO2 but provides a mechanism to transport re-suspended organic matter and regeneration products to the deep basin. The strength of this process remains uninvestigated but the total flow, estimated at 1 Sv (32,000 km3 yr-1; Melling, 1993], suggests it ought not to be neglected. By virtue of brine export and sinking of algal mats from the ice, the Arctic contains potentially some of the most efficient shelf export regions in the world. In contrast, many tropical watersheds discharge a high volume of water and sediments to the ocean due to frequent floods. Indeed, more than half of the total runoff and land-derived sediments are discharged to the oceans from tropical coasts, especially the Indo-Pacific Archipelago [Nittrouer et al., 1995]. The narrow shelves and weak Coriolis force favor cross shelf transport of sediments, which carry a significant amount of carbon, to the deep ocean.

In contrast to the margins mentioned above, some others are indeed CO2 sources. Recent results from Biogeochemical Budgeting Modeling Project of LOICZ led by Stephen Smith and Fred Wulff indicate that many coastal embayments and lagoons export more phosphate (an essential nutrient for marine primary producers) than they receive, suggesting these systems to be net heterotrophic or respiring CO2. Apparently, a fraction of the riverine carbon does get remobilized and released to the atmosphere in nearshore environments. It seems that systems with sluggish water exchange tend to be CO2 sources. Eastern boundary current systems form another type of coastal CO2 source. For example, observations on the western coast of Iberian Peninsula indicate it to be a weak net CO2 source supported by the upwelling of the CO2 laden Eastern North Atlantic Central Water rather than the input of terrestrial carbon. Strong outgassing of CO2 was also observed in the coastal zones off Oregon and Chile during strong upwelling events. Although coastal upwelling stimulates phytoplankton growth, which consumes CO2, the high partial pressure of CO2 in the upwelled water often overwhelms the consumption of CO2 and results in a net release of CO2 to the atmosphere. A confounding phenomenon within these systems is the juxtaposition of strong CO2 sources and sinks which are respectively associated with poorly- and well-developed phytoplankton communities. Recent findings suggest that iron supply from shelf sediments may play a role in regulating the growth of phytoplankton.

To summarize, riverine carbon fluxes do not necessarily make an adjacent continental margin a CO2 source. Instead, the margin may serve as a net CO2 sink by exporting a significant fraction of locally produced or river discharged organic carbon to the interior ocean. The shelf biological pump is driven by both the riverine nutrients, and, more importantly, the upwelled nutrients. Shelf released iron is also an important, but poorly quantified, micro-nutrient controlling air-sea CO2 exchange. The export of shelf primary production accounts for 7-27% of the global biological pump [Liu et al., 2000], which is defined as the carbon transfer from the surface water to the intermediate and deep waters mediated by biological activities. To illustrate how continental margin carbon fluxes may affect the global ocean carbon budget, we have constructed one likely scenario based on the following assumptions: (See Fig. 3 for fluxes in discussion that are noted below.)

  • The continental margins as a whole is a weak CO2 sink (0.1 PgCyr-1 as indicated by the net downward flux between atmosphere and margins).
  • The storage of anthropogenic CO2 in margins (0.05 PgCyr-1 in parenthesis) is proportional to their seawater volume.
  • Three quarters of the carbon burial occurs in margins (dashed arrow of 0.15 PgCyr-1)
  • Half of the remobilizable carbon from the river runoff is released to the atmosphere from margins (dashed arrow of 0.3 PgCyr-1 to the atmosphere).
  • The export of shelf primary production accounts for 20% of the global biological pump (2 PgCyr-1 from margins to intermediate and deep waters), which is compensated by upwelling of dissolved inorganic carbon of equal strength.

Given these assumptions, the net air-to-sea CO2 flux in the interior ocean would be 1.3 Pg Cyr-1 (net downward flux between atmosphere and the surface water of interior ocean as shown in Fig. 3). For this we also assume the total oceanic uptake of anthropogenic CO2 and the total physical and biological pumps remain the same as those in Fig. 2. However, if the net air-to-sea CO2 flux in the interior ocean is as large as the current estimate (2.2 Pg Cyr-1), the actual uptake rate of anthropogenic CO2 would be as large as 2.9 Pg Cyr-1 due to the net input of terrestrial carbon (0.6 Pg Cyr-1) and net CO2 uptake in the margins (0.1 Pg Cyr-1).

Global synthesis efforts and ongoing regional studies

The new findings highlighted above clearly demonstrate that continental margins are complicated systems with sufficiently active physical and biogeochemical processes to affect the global carbon cycle. Within the world oceans, these margins are almost certainly the most vulnerable to perturbations from human activities. Continental margins are much more heterogeneous than the interior ocean so that continental margin carbon fluxes – and perturbations to them – cannot be adequately represented by the coarsely gridded maps of global ocean carbon fluxes we have today. The CMTT is keen to synthesize the currently available data on as many different margins as possible to obtain a global estimate.