the chaNgiNg carboN cycle iN the SoutherN oceaN

. Various human activities, including fossil fuel combustion and forest clearing, emit about eight petagrams (or billion tons) of carbon in the form of CO 2 into the atmosphere annually. The global ocean absorbs about two petagrams of CO 2 , and about a half of that amount is absorbed by the Southern Ocean south of 30°S, thus slowing the rapid accumulation of CO 2 in the atmosphere. Partial pressure of CO 2 ( p CO 2 ) is a measure of the chemical driving force for the CO 2 exchange between the ocean and the atmosphere. This paper discusses its space and time distribution over the Southern Ocean. The major sink zone for atmospheric CO 2 is located in a latitude belt between 30°S and 50°S, where the biological utilization of CO 2 and cooling of warm subtropical waters flowing southward produce low seawater p CO 2 . Strong winds in this zone also enhance the ocean’s uptake. Although the source-sink conditions vary over a wide range through the seasons in the areas south of 50°S, this zone is a small sink on an annual average. Winter observations show that surface water p CO 2 values in the source region for Antarctic Intermediate Water have increased at a rate faster than the atmospheric increase rate, suggesting that the ocean CO 2 sink intensity has been weakening for several decades and has changed from a net sink to a net source since 2005. The results of ocean general circulation-biogeochemistry model studies are found to be consistent with the observations.

), 13 C/ 12 C mass balance (Quay et al., 2003), atmospheric oxygen and CO 2 changes (Bender et al., 2005;Manning and Keeling, 2006), and CO 2 change in the ocean (Sabine et al., 2004). The model studies include coupled Ocean General Circulation-Biogeochemistry models (OGCM) (Mikaloff-Fletcher et al., 2006;Sarmiento and Gruber, 2006;Jacobson et al., 2007;Lenton and Matear, 2007;Gruber et al., 2009;Le Quérér et al., 2010), and inversion of atmospheric CO 2 data using Atmospheric General Circulation models (AGCM) (Gurney et al., 2008). Gruber et al. (2009) reviewed the estimates for CO 2 uptake flux over the contemporary global ocean obtained by four groups of independent methods: inversion of the ocean data using 10 OGCMs, 13 ocean forward models (OCMIP-2), inversion of atmospheric CO 2 data (Gurney et al., 2008), and sea-air pCO 2 difference (Takahashi et al., 2009). Although the mean air-to-sea flux estimates for the contemporary global ocean obtained by these methods are in general agreement at 1.5 ± 0.5 Pg C yr -1 , notable discrepancies are found in the Southern Ocean.
The ocean inversion methods suggest a relatively uniform weak sink in the areas south of 58°S, whereas the sea-air pCO 2 difference (ΔpCO 2 ) data in these areas suggest a CO 2 source. Processes governing atmosphere-ocean interactions in the Southern Ocean region are complex because of the large seasonal variability in temperature, wind regimes, ice/water conditions, and biological activities.
Although significant progress has been made in recent years due to improved research facilities, observations are still limited because of operational difficulties related to hostile weather conditions, and observation-based estimates are subject to considerable uncertainty.
Model results are also subject to uncertainties because of limited time-space resolutions and imperfections in the parameterizations for various processes, including eddy mixing, ice formation, and biological processes.
In this article, we review recent progress in biogeochemical studies on the carbon cycle with emphasis on the temporal and spatial variability of pCO 2 in Southern Ocean surface water. Here, the Southern Ocean is defined as the oceanic areas south of 30°S that include a major sink zone for atmospheric CO 2 centered at 40°S. First, we review climatological mean distribution of surface water pCO 2 and net sea-air CO 2 flux. Second, we discuss the change in surface water pCO 2 and the intensity of the ocean CO 2 sink in circumpolar waters.

Sea-air co 2 tr aNSFer oVer the SoutherN oceaN
Before we discuss CO 2 exchange over the Southern Ocean, we briefly review relevant oceanographic information.
Because pCO 2 is the primary quantity measured by our group, we next discuss the time-space distribution of surface water pCO 2 , and then present the net sea-air CO 2 flux.  (Moore and Abbott, 2000), surface water pCO 2 is reduced, creating a strong CO 2 sink zone centered around 40°S in the Atlantic, Indian, and most of the Pacific Oceans. These waters (Mode Water) sink to mid-depth and transport atmospheric CO 2 to the subsurface regime.  in seawater and air (ΔpCO 2 ) determines the direction and magnitude of the net CO 2 flux across the interface. When the pCO 2 in seawater is greater than that in the overlying air (ΔpCO 2 > 0), the net flux is from sea to air; when ΔpCO 2 < 0, the net flux is from air to sea. The net sea-air flux may be estimated by multiplying the sea-air pCO 2 difference by the gas transfer coefficient across the sea surface.

observations of Surface
Water pco 2 Equilibrators (bubbler, membrane, or shower types; e.g., Chipman et al., 1993;Newberger, 2004) were operated either in a seawater flowthrough mode for continuous underway measurement of water samples pumped from an intake located a few meters below the sea surface, or in a discrete water mode for water samples col-  Its formation is attributed primarily to high productivity in the high-chlorophyll zone observed by satellites over the same latitudes (see Plates 3 and 6 in Moore and Abbott, 2000). However, the colocation of the CO 2 sink zone and the high-productivity zone is only qualitative because seawater pCO 2 is governed by net community production, which includes primary production as well as the respiration, recycling, and export of organic carbon from the mixed layer.
Peaking of the sink intensity in August and September suggests that winter cooling of surface water plays an important role in the formation of the sink.
As the season progresses, the CO 2 sink zone moves southward due mainly to  2000, observing that the summertime seawater pCO 2 was lower than the atmospheric pCO 2 (ΔpCO 2 ~ -15 µatm) due to photosynthesis, and the winter pCO 2 was higher than the atmospheric (ΔpCO 2 ~ +10 µatm) due to the upwelling of high-CO 2 deep waters. Because the alkalinity was found to be similar, the biological effect on CO 2 is due mostly to the production of organic carbon.
Net Sea-air co 2 Flux from Δpco 2 The net CO 2 flux across the sea surface (F sea-air ) may be estimated by Equation 1, in which the main drivers are wind speed and sea-air pCO 2 difference (ΔpCO 2 ): F sea-air (g C m -2 month -1 ) = (1) 0.585 · Ko · (Sc) -1/2 · (U 10 ) 2 · ΔpCO 2 , where Ko is the solubility of CO 2 in seawater (mol CO 2 liter -1 atm -1 ; Weiss, 1974), Sc is the Schmidt number (see Wanninkhof, 1992), U 10 (m sec -1 ) is the wind speed at 10 m above the sea surface, and ΔpCO 2 is in µatm. The number 0.585 includes a unit conversion factor (changes from second to month, from liter to m 3 ), the gas Figure 3. Seawater pco 2 observed in ice field waters with temperatures less than -1.75°c in areas south of 60°S during June to September since 1998. The measurements were made possible by improvements in the intake port for the scientific water sampling line aboard rVib Nathaniel B. Palmer that prevent ice clogging. The data obtained in different years are averaged for each month, and one standard deviation is shown. The pco 2 in under-ice water increases as the season progresses. From Takahashi et al. (2009) transfer scaling factor of 0.26, and the reference Schmidt number of (660) 1/2 at 20°C for seawater (Takahashi et al., 2009). Although Ko and Sc vary with temperature, the temperature effects cancel in the ratio, and Ko/(Sc) 1/2 is nearly constant in the ocean temperature range. The 0.26 (± 30%) scaling factor for the gas transfer rate is deter-     Takahashi et al. (2009) which the ΔpCO 2 is represented by ice field measurements shown in Figure 3.
In June, the ΔpCO 2 in under-ice waters is negative, reflecting the low pCO 2 conditions produced by the biological pump during the preceding months. Hence, when water is exposed to the air, it acts as a sink for atmospheric CO 2 , and as the season progresses, it becomes a source by July. For the ice field surrounding the continent, the NCEP/DOE 2 Reanalysis (2005) ice cover data are regridded to our 4° x 5° grid and averaged for each month. When the ice cover is less than 10% in a 4° x 5° box area, it is assumed to be all water. Between 10% and 90%, the flux is computed proportional to the water area. Because ice fields have leads and polynyas due to dynamic motion of sea ice, we assume the fields to be 10% open water even though the satellite data report 100% ice cover (Worby et al., 2008). The strong CO 2 source zone centered around 60°S (yellow-orange) reflects the ice field edge zone in late winter months, when a large area of seawater with positive ΔpCO 2 values is exposed to the air, allowing gas exchange. Although the seasonal ice zone exhibits large seasonal changes in physical, biological, and chemical conditions, this zone appears to make a small contribution in terms of annual sea-air CO 2 flux to the global sea-air CO 2 budget.
On the annual mean, the zone cen- We have chosen wintertime SST as the indicator, and divided the data into five temperature zones between 0.8°C and 5.5°C. Figure 6 shows the time plots and data locations, and Table 1  indicates a weakening of the sink intensity due to much faster rates of oceanic pCO 2 increase than the atmospheric rate of 16 µatm decade -1 . On the other hand, a much lower rate of 6 µatm yr -1 was table 1. The mean decadal rate of change for wintertime surface water pco 2 and sea surface temperature (SSt) in five temperature zones. The second column shows the mean rate of pco 2 change as observed in Figure 6, and the SSt change in the third column is estimated using the temperature data obtained concurrently with the pco 2 data. The fourth column shows the pco 2 change corrected for the SSt change using 4.23% pco 2 change per °c.    1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 Rate of Change = 2.52 +/− 0.28 (N=56)

coNcluSioN
The global ocean is currently absorbing annually about 2 Pg C yr -1 of CO 2 from the air, and it plays a significant role in the uptake and long-term storage of anthropogenic CO 2 that is emitted to the atmosphere, affecting Earth's climate. The climatological mean sea-air flux is estimated by the observed sea-air pCO 2 difference and the gas transfer rate parameterized as a function of (wind speed) 2 . A Southern Ocean zone between 30°S and 50°S is found to be a major sink for atmospheric CO 2 , taking up 1.0 Pg C yr -1 . Thus, the Southern Ocean is a major ocean sink for atmospheric CO 2 . This paper discusses how this CO 2 sink is changing in response to recent climate change.
We investigated the multidecadal mean trends for surface water pCO 2 and temperature (0.8°-5.5°C) during winter in the formation region for AAIW. The wintertime waters were chosen because of minimal winter biological activity in order to avoid large biologically induced variability in seawater pCO 2 , and because of the maximal vertical mixing. The pCO 2 in the waters between 1.5°C and 4.5°C has increased at a rate of 23.9 ± 3.8 µatm decade -1 , which is