Comment on "Enhanced open ocean storage of CO2 from shelf sea pumping".

In a well-designed North Sea study, Thomas et al. (1) found that atmospheric carbon dioxide (CO 2 ) was absorbed by continental shelf water and was eventually exported into the North Atlantic Ocean. The work confirmed preliminary observations in the same area (2, 3) and provided support for the continental shelf pump hypothesis (4, 5). Thomas et al. then extrapolated Bthe CO 2 uptake by the North Sea to the global scale[ and inferred a net oceanic uptake of atmospheric CO 2 by coastal oceans of 0.4 Pg C year. A previous global extrapolation based on limited observations in the East China Sea (ECS) suggested an even greater air-sea CO 2 flux of 1.0 Pg C year in the world_s continental shelf (4). We are concerned with such extrapolations of regional studies to the global scale without cautioning readers that no current consensus exists on this issue. Although most shelf CO 2 measurements have thus far revealed that shelves are sinks of atmospheric CO 2 , these shelves are located in mid-latitude zones that experience strong spring blooms and substantial seasonal changes Ei.e., the North Sea (50-N to 61-N) (1, 2), the Gulf of Biscay (42-N to 52-N) (2), the ECS (25-N to 38-N) (4, 6), and the MidAtlantic Bight (35.5-N to 41-N) (7)^. They absorb atmospheric CO 2 , as evidenced by very low sea surface partial pressure of CO 2 (pCO 2 ) during planktonic bloom seasons. To sustain this uptake, absorbed CO 2 must be exported to the open ocean as organic and inorganic carbon under favorable shelf export conditions, especially in winter. The Bcontinental shelf CO 2 pump hypothesis[ was proposed to describe such circumstances (4). However, the shelves listed above represent only a small fraction of global shelf area (8) and may not be representative of global continental shelves. The North Sea, for example, is characterized by massive input from the land. A recent report from the U.S. South Atlantic Bight (SAB) (27-N to 35-N) provided the first example of a major source of annual CO 2 to the atmosphere (9). The pCO 2 signal in the SAB is high during spring and summer and low during winter, which is the opposite of the trend observed in the North Sea and Gulf of Biscay (1, 2). Elsewhere, the shelf and upper slope area of the northern South China Sea (SCS) (20-N to 22-N) also act as an annual CO 2 source to the atmosphere (10). Thus, it is clear that not all margins are a sink for atmospheric CO 2 . Margins are the most heterogeneous areas of the world_s oceans, with potentially very different magnitudes of physical and biogeochemical mechanisms. Sea surface pCO 2 may differ because of latitudinal differences as well as differences related to oceanographic settings. The Arctic and subarctic shelves may be CO 2 sinks (11–13). The shelves vary from strong to weak CO 2 sinks in the temperate areas (1–7). Farther south in the SAB and in the SCS, the shelves are sources of CO 2 to the atmosphere (9, 10). The tropical and subtropical shelves and marginal seas are most likely sources of CO 2 to the atmosphere, driven by either the high annual surface temperature, the lack of a strong spring bloom, inputs from marshes (9) and mangroves (14), or reef formation. Margins dominated by coastal upwelling are complex in that they receive deep water with high levels of both inorganic nutrient and CO 2 . Although precise annual fluxes are difficult to define for these shelves, it again appears that the low-latitude shelves act as CO 2 sources (15, 16), whereas those at mid to high latitudes act as CO 2 sinks (17–19). However, these systems have a rather small total area (8). Large river plumes may be a strong sink of atmospheric CO 2 but, aside from the Amazon system, they represent a limited surface area compared with the surrounding waters that often appear as CO 2 sources (20, 21). We are thus still at a stage of uncertainty about the magnitude of air-sea CO 2 exchange because of both the heterogeneous nature of ocean margins and the lack of spatial and temporal coverage of pCO 2 data. Highand low-latitude continental shelves have clearly not been sufficiently studied, and they deserve more attention in future research. The mechanisms that govern the net sink/source term and the magnitude of CO 2 exchange also require a more accurate understanding. Although the continental shelf pump hypothesis needs the scrutiny of further multidisciplinary field research, it is fair to suggest that low-latitude margins are not favorable for CO 2 absorption, in contrast to the case in midand high-latitude margins. We laud the effort to explore the global significance of continental shelves in the ocean carbon cycle (1, 4, 5) but are less confident in the global extrapolation of these studies. Atmospheric CO 2 uptake by continental shelves may have been overestimated given the latitudinal difference of air-sea exchange in marginal seas.