Carbonate system in the water masses of the Southeast Atlantic sector of the Southern Ocean during February and March 2008

Biogeosciences, 8, 1401–1413, 2011 www.biogeosciences.net/8/1401/2011/ doi:10.5194/bg-8-1401-2011 © Author(s) 2011. CC Attribution 3.0 License. Biogeosciences Carbonate system in the water masses of the Southeast Atlantic sector of the Southern Ocean during February and March 2008 M. Gonz´ alez-D´ avila1, J. M. Santana-Casiano1, R. A. Fine2, J. Happell2, B. Delille3, and S. Speich4 1Departamento de Qu´ ımica, Facultad de Ciencias del Mar, Universidad de Las Palmas de Gran Canaria, 35017, Spain 2Rosenstiel School, University of Miami, 4600 Rickenbacker Causeway, Miami, FL 33149-1098, USA 3Unit´ e d’Oceanographie Chimique, Astrophysics, Geophysics and Oceanography department, University of Li` ege, All´ ee du 6 Aoˆ ut, 17 (Bˆ at B5), 4000 Li` ege, Belgium 4Laboratoire de Physique des Oceans (LPO), CNRS/IFREMER/UBO, Brest, France Received: 14 December 2010 – Published in Biogeosciences Discuss.: 17 January 2011 Revised: 20 May 2011 – Accepted: 23 May 2011 – Published: 31 May 2011 Abstract. Carbonate system variables were measured in the South Atlantic sector of the Southern Ocean along a transect from South Africa to the southern limit of the Antarctic Circumpolar Current (ACC) from February to March 2008. Eddies detached from the retroflection of the Agulhas Current increased the gradients observed along the fronts. Minima in the fugacity of CO2,fCO2, and maxima in pH on either side of the frontal zone were observed, noting that within the frontal zone fCO2reached maximum values and pH was at a minimum. Vertical distributions of water masses were described by their carbonate system properties and their relationship to CFC concentrations. Upper Circumpolar Deep Water (UCDW) and Lower Circumpolar Deep Water (LCDW) offered pHT,25 values of 7.56 and 7.61, respectively. The UCDW also had higher concentrations of CFC-12 (>0.2pmolkg−1) as compared to deeper waters, revealing that UCDW was mixed with recently ventilated waters. Calcite and aragonite saturation states () were also affected by the presence of these two water masses with high carbonate concentrations. The aragonite saturation horizon was observed at 1000m in the subtropical area and north of the Subantarctic Front. At the position of the Polar Front, and under the influence of UCDW and LCDW, the aragonite saturation horizon deepened from 800m to 1500m at 50.37◦S, and reached 700m south of 57.5◦S. High latitudes proved to be the most sensitive areas to predicted anthropogenic carbon increase. Buffer coefficients related to changes in [CO2], [H+] and with changes in dissolved inorganic carbon (CT) Correspondence to: M. Gonz´ alez-D´ avila ([email protected]) and total alkalinity (AT)offered minima values in the Antarctic Intermediate Water and UCDW layers. These coefficients suggest that a small increase in CTwill sharply decrease the status of pH and carbonate saturation. Here we present data that suggest that south of 55◦S, surface water will be under-saturated with respect to aragonite within the next few decades. 1 Introduction The Southern Ocean plays an important role in modulating the global climatic system by transporting and storing heat, fresh water, nutrients, and anthropogenic carbon dioxide (CO2)(e.g., Lovenduski and Gruber, 2005). This region is predicted to be greatly influenced by global change, given that polar marine ecosystems are particularly sensitive to carbonate change (Sarmiento et al., 1998; Orr et al., 2005). Since pre-industrial times, the uptake of CO2has modified the chemistry of the ocean, lowered the pH and the concentration of carbonate ions (CO2− 3)with high latitudes among the most affected areas (Caldeira and Wickett, 2003; Orr et al., 2005). Surface ocean pH levels have already decreased by 0.1 units in the Southern Ocean (McNeil and Matear, 2007; Key et al., 2004) and are projected to decline by around 0.3 until the year 2100 (McNeil and Matear, 2008). Orr et al. (2005) predicted that the Southern Ocean will begin to experience aragonite under-saturation by the year 2050. On the other hand, there is a study by McNeil and Matear (2008) which is based on a large-scale Southern Ocean observational analysis that also considers the seasonal magnitude and variability of CO2− 3and pH. This study suggests that the Southern Ocean aragonite under-saturation in Published by Copernicus Publications on behalf of the European Geosciences Union. 1402 M. Gonz´ alez-D´ avila et al.: Carbonate system in the Southern Ocean in 2008 winter will already occur by the year 2030. As the dissolution of anthropogenic carbon increases, the total inorganic carbon concentration of the surface waters, and the capability to buffer these inputs decreases. This results in a much greater sensitivity to local variations in total inorganic carbon and total alkalinity. The lowest buffer values have been observed in the Southern Ocean (Egleston et al., 2010), as this area is particularly sensitive to increased CO2. The Southern Ocean is particularly efficient when ventilating deep and bottom waters (e.g. Toggweiler et al., 2006). Deep ventilation takes place south of the Polar Front (PF). There are clear links between the seasonal carbon dynamics and the Antarctic Bottom Water (AABW) formation regions (McNeil et al., 2007). The entrainment and upwell of Circumpolar Deep waters, rich in dissolved inorganic carbon and carbonate poor, into the surface layer lowers the carbonate concentration considerably (McNeil and Matear, 2008). A recent study showed that the upwelling of deep waters, rich in CO2, is the most dominant driver of winter carbon cycling, as compared to temperature-driven differences in solubility or biological processes (McNeil et al., 2007). Physical processes such as deep-water formation in the Weddell Sea, the upwelling of deep water in the divergence zone and the formation of intermediate water in the Antarctic Polar Zone (APZ) affect the carbonate system parameters, with consequences for the CO2air-sea fluxes. In this region where several frontal systems are observed, sharp gradients in temperature and salinity (Lutjeharms and Valentine, 1984; Belkin and Gordon, 1996) and important changes in the CO2air-sea exchange have been described (Bakker et al., 1997; Hoppema et al., 1995; Chierici et al., 2004; McNeil et al., 2007). In the framework of the BONUS-GoodHope project, the parameters of the carbonate system, pH, total alkalinity (AT) and total dissolved inorganic carbon concentration (CT)were measured in the southeast Atlantic sector of the Southern Ocean (Fig. 1). The main objective of this work was to characterise the carbonate system of the water masses present in the area, and determining their sensitivity to an increase of CO2in this sector of the Southern Ocean. 2 Hydrography of the area of study The region studied (Fig. 1) is described in detail in Chever et al. (2010), based on Gladyshev et al. (2008). It is divided into three main regimes, namely, the sub-tropical domain north of 40◦S–42◦S, the Antarctic Circumpolar Current (ACC) between 40◦S–42◦S and 55◦S–57◦S, and the boundary region between the ACC and Weddell Gyre/the northern edge of Weddell Gyre (Park et al., 2001; Gladyshev et al., 2008). In this region, several frontal systems were described in a review by Orsi and Whitworth (2005), using potential temperature (θ), salinity and oxygen as indicators. These frontal zones were defined both by sharp changes in temperature and Figure 1. 812 813 814 Fig. 1. Map showing the cruise track and sea surface height (SSH) for the southwest Atlantic sector of the Southern Ocean during the BONUS GoodHope 2008 cruise. The track is plotted over an altimetry image for 27 February and the fronts are identified: STF (Subtropical Front), SAF (Subantarctic Front), PF (Polar Front), SACCF (Southern ACC Front) and SBdy (Southern Boundary). Cyclonic (Ci)and anticyclonic (Aiand M) Agulhas rings are also marked. salinity and by enhanced chlorophyll (Chl)-aconcentrations and reduced values of partial pressure of CO2(expressed as fugacity of CO2,fCO2)(Smith and Nelson, 1986, 1990; Chierici et al., 2004; Laika et al., 2009). In the sub-tropical domain, the Sub-tropical Front (STF) divides the warmer tropical waters and the colder sub-Antarctic waters. The area is then divided by the North Sub-tropical Front (N-STF) and the South Sub-tropical Front (S-STF). The strong interaction between the Agulhas Current and slope and shelf waters in the Agulhas Bank (Boebel et al., 2003; Richardson 2007) produces the formation of cyclonic and anti-cyclonic Agulhas rings that strongly interact with fronts affecting their boundaries. These features are common in the Cape Basin and have been proved recently to define the position of the N-STF and S-STF (Dencausse et al., 2010). In the ACC domain, four main fronts are identifiable. The Subantarctic Front (SAF) is located at around 44◦S, with a sharp decrease in both salinity and temperature. The Polar Front, PF, is found at around 50◦S, characterized by weak surface temperature and salinity gradients. A deep-reaching front observed to the south is indicative of the presence of the southern ACC front (SACCF), located at 52◦–53◦S. In Biogeosciences, 8, 1401–1413, 2011 www.biogeosciences.net/8/1401/2011/ M. Gonz´ alez-D´ avila et al.: Carbonate system in the Southern Ocean in 2008 1403 this region, the southern boundary of the ACC (SBdy) is located at around 56◦S. During late summer, the position of this front cannot be detected by surface temperature gradients, but can be defined by an increase in salinity. South of the SBdy is the region of the Weddell Sea. The change from low salinity surface water in the ACC band to higher salinity waters is associated with the presence of the Weddell Gyre. The narrow band of more saline surface waters close to the frontal area comes from the upwelling of deep, salty water during the course of the water from the western part of the Weddell Gyre through to the Prime Meridian. Along the transept, several water masses can be distinguished (see, e.g., Withworth and Nowlin, 1987; Arhan et al., 2011; Speich et al., 2011) as are briefly summarized here (Fig. 3). South of the SACCF at the SBdy, the Winter Water (WW) of the Antarctic Zone is observed as a sub-surface tongue centred at 150m that extends northwards to the position of the Antarctic PF. In this region, two circumpolar deep waters (CDW) are distinguished: Upper Circumpolar Deep Water (UCDW) and Lower Circumpolar Deep Water (LCDW) (Whitworth and Nowlin, 1987). Circumpolar Deep Waters are a composite of deep waters flowing from the Indian, Atlantic and Pacific basins. When approaching the Antarctic Continent, these eventually mix with younger waters such as the Weddell Sea Deep Water (WSDW), WW and the Ice Shelf Water, forming the Modified CDW (M-CDW) or Antarctic Bottom Water (AABW). This mixing produces ventilated waters whose injection into mid-depth and bottom layers contributes to the ventilation of the deep Southern Ocean (Orsi et al., 2002; Lo Monaco et al., 2005; Bakker et al., 2008). Below 2000m, influence of the WSDW can be discerned; the water mass originates from the Weddell Sea below about 1500m. Below 4000m, in the Agulhas basin, cores of CFC-rich waters identify the Antarctic Bottom Water (AABW), formed at several places around Antarctica (Orsi et al., 2002). The near-surface water at the Antarctic Polar zone is the least saline near-surface water in the ACC band. It is observed continuously at all longitudes (Orsi and Whitworth, 2005). It subducts northwards at the SAF to feed the Antarctic Intermediate Water (AAIW), located in a 600–1000m band further north. Moreover, the salinity maximum of the deep water northwards to the SAF is associated with the diluted North Atlantic Deep Water (NADW) (Arhan et al., 2003). 3 Data and methods The BONUS-GoodHope cruise took place on board the French R/V Marion Dufresne in the southeast Atlantic sector of the Southern Ocean in the region 33◦580S-57◦330S, 17◦130E–0◦E (Fig. 1). It started on 13 February 2008, off Cape Town, and was completed on 17 March 2008. During the cruise, full depth CTD data were registered at 79 stations and samples were taken at 22 depths for the measurements of salinity, oxygen, pH, ATand CT. Oxygen concentrations were measured on board by Winkler titration. Chlorophyll-a (Chl-a) was also measured on board by fluorometric analysis of acetone (90%) extracts (Speich and Dehairs, 2008). Samples were collected for the posterior laboratory analysis of two chlorofluorocarbons, CFC-11 and CFC-12. The three variables of the carbonate system were measured on board the Marion Dufresne in order to achieve the highest level of data quality and resolution. The hydrocast stations (78 stations plus station zero) were sampled for pH on total scale at 25◦C (pHT,25, with [H+] in µmolkg−1), total alkalinity (AT, in µmolkg−1)and total dissolved inorganic carbon concentration (CT, in µmolkg−1). There were a total of 1639 bottles in hydrocast CTD stations at non-repetitive depths, and in some cases samples were flagged. As a result, high quality data are available for pH from 1609 samples, from 1559 samples for ATand from 1504 samples for CT. 3.1 Sampling procedure 500 ml glass bottles were used for the analytical determination of both pH and AT. 100ml glass bottles were used to analyse CT. The bottles were rinsed twice with seawater and were overfilled with seawater. Samples were shielded from the light and analysed between stations. At shallow stations and when samples could not be analysed for CTin less than 6h after sampling, they were poisoned with HgCl2(60µl, saturated solution). 3.2 pH measurements We measured pH on the total scale (pHT)at a constant temperature of 25◦C (pHT,25). An automated system based on the spectrophotometric technique of Clayton and Byrne (1993) with m-cresol purple as an indicator (the dye effect was removed for each pH reading) and with an uncertainty of 0.002 units was used (Gonz´ alez-D´ avila et al., 2003). 3.3 Total dissolved inorganic carbon measurements A VINDTA 3C system (Mintrop et al., 2000) (www. MARIANDA.com), with coulometer determination was used for the titration of total dissolved inorganic carbon concentrationafterphosphoricacidaddition. Thetitrationofcertified reference material for oceanic CO2, CRMs (#85), supplied by Andrew Dickson at Scripps Institution of Oceanography, was used to test the performance of the equipment. A CRM was analysed every time a new titration cell for CTdetermination was prepared (once a day); in total 31 CRMs were analysed. We measured 1996.0±1.6 µmolkg−1for CT, while the certified value is 2000.4±0.4 µmolkg−1. A study done on board indicates that this difference relates to the temperature at which CTis determined which was 25◦C in our case. Raw data were corrected for this offset by www.biogeosciences.net/8/1401/2011/ Biogeosciences, 8, 1401–1413, 2011 1404 M. Gonz´ alez-D´ avila et al.: Carbonate system in the Southern Ocean in 2008 multiplying by the factor 1.0022. Each CRM sample was also analysed for total alkalinity. 3.4 Total alkalinity measurements Samples for ATwere potentiometrically titrated with standardized 0.25 M HCl (0.45M in NaCl) to the carbonic acid end point using a titration system described in detail in Mintrop et al. (2000). The titration of CRMs (#85) was used to test the performance of the titration system. Measurements of CRMs were within ±1.1µmolkg−1of the certified value. The agreement between on board experimental data of AT(2184.0±1.1) and the certified value (2184.0±0.8) indicates accurate HCl concentration and pipette volume for the titration system. No correction was carried out on the experimental data. 3.5 CFC sampling and measurement 1191 samples (a mean of 18 samples per hydrocast) were collected from Niskin bottles. The samples (about 125ml) were taken via Viton tubes connected to glass bottles with connectors. The bottles and caps were thoroughly rinsed with the water to be sampled. The bottles were filled and capped underwater in a 1l beaker. At the University of Miami laboratory, water samples were analysed for CFC-11 and CFC-12 using an extraction system and gas chromatograph following established procedures (Bullister and Weiss, 1988). Analytical uncertainties for CFC-11 (CCl3F) and CFC-12 (CCl2F2) are each ±8%. 3.6 Calculations 3.6.1 Calcite and aragonite saturation state The saturation state of seawater with respect to calcite and aragonite () was calculated as the product of the calcium (Ca2+)and carbonate ion (CO2− 3)concentrations at in situ temperature, and the salinity and pressure, divided by the stoichiometric solubility product (K∗ sp)for those compounds cal = [Ca2+][CO2− 3]/K∗ sp,cal (1) ara = [Ca2+][CO2− 3]/K∗ sp,ara (2) where the calcium concentration is estimated from salinity (Lewis and Wallace, 1998), and the carbonate ion concentration is calculated from ATand CT, and computed by using CO2sys.xls v12 (Lewis and Wallace, 1998), using the carbonic acid dissociation constants of Mehrbach et al. (1973) as in Dickson and Millero (1987), the sulphate dissociation constant by Dickson (1990a) and the K∗ sp from Mucci (1983). Other values are described in detail in Lewis and Wallace (1998). The nutrient data were considered in all the computations (Speich and Dehairs, 2008; Branellec et al., 2010). 3.6.2 Normalisation procedure The pHTat in situ conditions (pHTis)was computed by applying the CO2sys.xls v12 programme and the set of constants mentioned above, to the experimental pHT,25 and AT pairs of data. The nutrient data were also considered. The regional normalisation to a constant salinity (Sref) proposed by Friis et al. (2003) was applied to the alkalinity and inorganic carbon concentration (XT), instead of the traditional normalisation NXT=XT/Smea ·Sref, where Smea is the salinity measured. This alternative procedure removes the effects of evaporation and precipitation as well as the salinity-proportional parts of mixing/upwelling and non-zero end-member, and improves the interpretation of their variations in the water column. In this normalisation, the salinity adjustment is based on a constant and region-specific term for S=0, which expresses river run-off, upwelling from below the lysocline, calcification, and lateral sea surface-water exchange. NXT=(Xmea T−XS=0 T)/Smea ·Sref +XS=0 T(3) Our surface data for ATand CTwere fitted to the equation XT=mTT+mSS+b0, AT= −3.18(±0.17)T+59.95(±1.65)S +266.6(±55.4)(4) CT= −10.49(±0.14)T +41.97(±1.40)S +757.4(±46.9)(5) with standard error of estimate of ±5.14 and ±4.88 for AT and CT, respectively. According to Eqs. (4) and (5), AS=0 T= 266.6±55.4 and CS=0 T=663.1±122.6. Sref was fixed to the traditional salinity of 35. 3.6.3 Buffer coefficients The explicit expressions for seven buffer factors that quantify the ability of ocean chemistry to resist changes in dissolved inorganic carbon and alkalinity as a function of proton, carbonate and borate ion concentrations were computed following Egleston et al. (2010), using their Table 1. Calculations of the several parameters used in these expressions were obtained after applying CO2sys program to pairs of data of CT and AT, assuming the set of constants indicated above. Six of those factors quantify the sensitivity of [CO2] (γi), [H+] (βi) and (ωi)to changes in dissolved inorganic carbon concentration (CT)and alkalinity (AT), while βHrefers to the traditional buffer capacity of the system, quantifying the resistance to changes of the pH of a chemical system to additions of a strong acid or base and it equals to −2.3 βAT. We have defined Q=[HCO− 3] + 4[CO2− 3] + [H+][B(OH)− 4]/(Kb+[H+]) + [H+]-[OH−], where Kb is the acidity constant for boric acid (Dickson, 1990b); P = 2[CO2]+[HCO− 3]; Ac=[HCO− 3] + 2[CO2− 3]. In order to avoid confusion, noted that the letter Qwas used instead of Sin the Egleston et al. (2010) paper. Biogeosciences, 8, 1401–1413, 2011 www.biogeosciences.net/8/1401/2011/ M. Gonz´ alez-D´ avila et al.: Carbonate system in the Southern Ocean in 2008 1405 γCT=∂ln[CO2] ∂CT−1 =CT−A2 c Q γAT=∂ln[CO2] ∂AT−1 =A2 c−CT·Q Ac βCT=∂ln[H+] ∂CT−1 =CT·Q−A2 c Ac βAT=∂ln[H+] ∂AT−1 =A2 c CT−Q ωCT=∂ln ∂CT−1 =CT−Ac·P [HCO− 3] ωAT=∂ln ∂AT−1 =Ac−CT·[HCO− 3] P(6) 4 Results and discussion 4.1 Surface distribution During the BONUS-GoodHope cruise, the expected trend of decreasing surface temperature towards the south was observed indeed (Fig. 2). This temperature gradient accompanied a decrease in pHT,25 and an increase in the surface dissolved inorganic carbon concentrations CT(Fig. 2). By using Sea Surface Temperature (SST) and Sea Surface Salinity (SSS) data from this research and the definitions of the characteristics of the major fronts south of Africa, the five major oceanic frontal structures have been identified and are marked in Fig. 2. The south Subtropical Front (S-STF) was located at 42◦20S. The north Sub-tropical Front (N-STF) was located west of the cruise line at around 38◦S (Chever et al., 2010). SST drops from 20.95◦C at 37.7◦S to 15.31◦C at 38.83◦S, while SSS decreases from 35.52 to 34.6. At those latitudes, the cruise crossed a warm, saltwater layer of two Agulhas rings at their boundaries. These were located at 36◦S–38◦S (A2) and 35◦S (A1), respectively (see Fig. 1). The influence of a cyclonic ring close to 36◦S (C1) increased the gradient observed in this region. The cyclonic structure C1 was injected into the region from the African slope (from the Agulhas Bank) as can be seen both from tracking using satellite altimetry (Fig. 1) and its hydrologic characteristics (e.g., salinity, oxygen, Fig. 3). From 41.60◦S to 42.03◦S, at the S-STF, the SST values dropped from 15.64 to 12.06◦C and the SSS fell from 34.75 to 34.22. From 39.2◦S to 40.2◦S, SSTs as high as 17◦C and SSS of 35 were found to be related to the influence of another Agulhas ring (A3), centred at 40◦S, 14◦E (Fig. 1). The changes in both temperature and salinity also affected the carbonate system variables which can also be used to distinguish the presence of the fronts. At around 38◦S, the presence of the Agulhas rings in the 36◦S– 38◦S (A1, C1 and A2) increased the pHT,25 from 7.97 at 815 816 817 Figure 2. 818 819 Fig. 2. (A) The sea surface temperature (SST), salinity (SSS), total dissolved inorganic carbon (CT, µmolkg−1)and pH in total scale at 25◦C (pHT,25)along the cruise track for samples analysed over the upper 10m. The figure shows the position of the major frontal zones during the BONUS GoodHope cruise. (B) Surface ocean Chlorophyll a, partial pressure of CO2in seawater expressed as fugacity, fCO2,sw (µatm) and pH in total scale at in situ conditions, pHTis. 36◦S to 8.040 all along 36.5◦S to 37.8◦S. Between A2 and A3, the pHT,25 shifted from 8.040 to 7.946, increasing to 7.97 at the position of the anticyclonic ring A3. At the S-STF, the pHT,25 decreased from 7.948 to 7.887, indicating a total change inside the STF area of 0.15 pH units. Again, from 39.2◦S to 40.2◦S, the pHT,25 increased from 7.95 to 8.00, in line with the observed temperature increase. Total alkalinity was strongly correlated with salinity. AT (data not shown) decreased from 2334 to 2296µmolkg−1 at 38◦S, between rings A2 and A3. The ATfrom 41.60◦S to 42.03◦S at the S-STF, dropped from 2294µmolkg−1to 2273µmolkg−1. There was also a noticeable increase in CT at both locations with changes of 33µmolkg−1at the 38◦N (from 2018.2 to 2051.6µmolkg−1)and around 20µmolkg−1 at the southern front, increasing from 2051.1µmolkg−1to 2072.8µmolkg−1. After normalisation to a constant salinity (data not shown), the NCTincreased by 70µmolkg−1at 38◦N and 31µmolkg−1at the S-STF. These variations indicate that the upwelling of deep CO2-rich waters takes place in this frontal area that, at least at the time of the cruise, overcompensates any reduction of CO2due to biological activity. An examination of Fig. 3 suggests a deep-reaching nature of the STF related to the presence of the Agulhas rings detached from the retroflection of the Agulhas Current. www.biogeosciences.net/8/1401/2011/ Biogeosciences, 8, 1401–1413, 2011 1406 M. Gonz´ alez-D´ avila et al.: Carbonate system in the Southern Ocean in 2008 820 Figure 3 821 Figure 3 cont 822 823 Fig. 3. The vertical distribution of temperature (T,◦C), salinity (S), pH in total scale at 25◦C (pHT,25), pH in total scale at in situ conditions (pHTis), dissolved oxygen (O2, µmolkg−1), total alkalinity (AT, µmolkg−1), total dissolved inorganic carbon (CT, µmolkg−1) and chlorofluorocarbon-12 (CFC-12, pmolkg−1)along the southeast Atlantic sector of the Southern Ocean during February to March 2008. The dots indicate the locations of discrete samples. In the ACC domain, four main fronts were identified. The Sub-Antarctic Front (SAF) was located at 44◦20S. The SSS dropped from 35.037 at 43◦190S to 33.93 at 44◦20S and the SST fell from 13.74 to 9.74◦C, located barely south of an old intense Agulhas Ring (M in Fig. 1). At these positions, the pHT,25 sharply decreased by 0.1 pH units from 7.938 to 7.839, the ATdropped from 2314 to 2265µmolkg−1while the CTincreased from 2070 to 2082µmolkg−1. The Polar Front, PF, was located at 50◦220S. A significant pHT,25 gradient was observed at the front changing from 7.768 to 7.740. The total alkalinity increased by 7µmolkg−1from 2280µmolkg−1while the CTincreased by 10µmolkg−1 from 2130µmolkg−1at 50◦220S to 2140µmolkg−1at 50◦380S. The southern ACC front (SACCF), was charted at 52◦390S. At this front, the SST slightly decreased from 2.44◦C at 52◦360S to 1.75◦C at 52◦550S while the SSS increased from 33.705 to 33.742. The position of the SACCF is seen more precisely on the pH gradient, with the pHT,25 decreasing from 7.716 to 7.696 as we moved southwards. At these positions, the CTincreased by 13µmolkg−1 from 2143µmolkg−1at 52◦360S to 2157µmolkg−1at 52◦550S. In the region studied, the southern boundary of the ACC (SBdy) was located at 55◦540S. It was detected by the increase in salinity from 33.836 at 55◦340S to 33.980 at 55◦540S. At that position, the ATincreased from 2293µmolkg−1to 2308µmolkg−1. After normalisation, the NATis still 4µmolkg−1higher related to the upwelling of deep water rich in alkalinity. From the pHTat in situ conditions (pHTis)computed fugacity of CO2,fCO2(AT,CT), and the Chl-a(Fig. 2b), we can see that the highest pH and lowest fCO2values were Biogeosciences, 8, 1401–1413, 2011 www.biogeosciences.net/8/1401/2011/ M. Gonz´ alez-D´ avila et al.: Carbonate system in the Southern Ocean in 2008 1407 observed in the same areas where there were Chl-amaxima. Along the section, there appeared to be two regions with different mean conditions for pHTis and fCO2. North of the SAF, a mean pHTis value of 8.102±0.014 with fCO2of 335±5µatm was observed, while to the south, the pHTis was 8.069±0.008 and the fCO2increased to 365±10µatm. The undersaturation observed in the sub-tropical zone with fCO2values below atmospheric (378µatm, see Speich and Dehairs, 2008 for details) is in concordance with the strong sink for atmospheric CO2previously described for the area (Siegenthaler and Sarmiento 1993, Metzl et al., 1995; Br´ evi` ere et al., 2006; Borges et al., 2008). The observed Chlaconcentration varied between 0.06 and 0.67mgm−3. In general, elevated Chl-alevels (reaching 0.4 and 0.6mgm−3) were associated with the shear area of the frontal zone, with very low values (below 0.2mgm−3)around the position of the rings. They were accompanied by minima of fCO2(e.g., 332µatm at 38◦N) and maxima of pHTis (e.g., 8.107 at 38.1◦S) on either side of the frontal zone, while at the position of the frontal zone, the fCO2was at a maximum (369µatm at the 38◦S area) and the pHTis was at a minimum (8.088, at 38◦S) (Fig. 2b). At the positions of the fronts, surface nitrate concentrations (data not shown) moved from below detection limit to north of 38.1◦S to 3.22µmolkg−1south and from 4.94µmolkg−1at 41.60◦S to 10.19µmolkg−1at 42.03◦S, under the influence of eddy C2. This implies that mixing takes place at the frontal zones, in particular where cyclonic rings are located, bringing up CO2rich (plus low pH and high nutrient) water that spreads out over the fronts where recent biological production favoured by the nutrient input increased the pHTis and decreased the fCO2levels. Eddies and rings provided the conditions to promote primary production, as has been indicated in the case of the Haida eddies (Crawford et al., 2005) and in other areas including oligotrophic areas (McGillicuddy et al., 2003; Pelegr´ ı et al., 2005; Gonz´ alez-D´ avila et al., 2006). A special event was observed south of 40◦S at the southern limit of the two detached Agulhas rings. Strong fCO2 under-saturation was detected together with high pH values and low Chl-avalues. These may be the result of a chemical memory effect indicating previous primary production, which was over at the time of sampling (and where Chl-a was thus, low again). South of 44◦S, the Southern Ocean surface waters were rich in nutrients (nitrate values >16µmolkg−1that increased to >25µmolkg−1south of the PF) but low in Chl-a(Fig. 2b), typically referred to as high-nutrient-lowchlorophyll (HNLC) waters. The primary production is not able to take up macronutrients in these areas due to iron limitation, unstable surface-water stratification, and light limitation, defined as being responsible for the maintenance of the HNLC condition in the ice-free Southern Ocean (e.g., Venables and Moore, 2010). Surface water alkalinity for the area between 30◦S and 70◦S and with SST<20◦C and 33<SSS<36 in the Southern Ocean can be estimated by the following relationship Eq. (7) (Lee et al., 2006): AT=2305+52.48(SSS−35)+2.85(SSS−35)2 −0.49(SST−20)+0.086(SST−20)2(7) The mean difference 1AT(measured - calculated) between values calculated using thes relationship and our measured values was −3.4±5.1µmolkg−1. The largest differences (−8.2) were observed at 40◦170S, between the two northern Agulhas Rings (St 17 and 34), where cyclonic rings C1 and C2 were located, at the Polar Front (St 80) and south of the SBdy area. If these areas are removed, 1ATis −2.5±4.3µmolkg−1. This result confirms both the validity of the relationship provided by Lee et al. (2006) for the full region except were mesoscale structures were observed, and that the surface alkalinity stays constant under CO2uptake by the ocean, at least in the short-to-mid-term future (e.g., Orr et al., 2005). 4.2 Carbonate system characteristics The oceanic region separating the African and Antarctic continents has been less studied than its two counterparts south of South America (the Drake Passage) and south of Australia. Recent observations from this region have significantly improved our knowledge of the properties and circulation of water masses exiting and entering the Atlantic Ocean south of Africa and processes influencing the Agulhas leakage (Hoppema et al., 2000; Richardson et al., 2003; Van Aken et al., 2003; Gladyshev et al., 2008). However, the carbonate properties in this region have not been fully described (Lo Monaco et al., 2005; Bakker et al., 2008). Due to water mass formation and transformation taking place in the Southern Ocean near the continent and water transport (Orsi and Whitworth, 2005), we will describe the carbonate system properties and their relationship to the CFC concentrations from south to north. South of the SACCF, the Winter Water (WW) of the Antarctic Zone is characterised by a sub-surface tongue centred at 150m with CFC-12 concentrations exceeding 2.5pmolkg−1(Fig. 3). The vertical transition between WW and the Circumpolar Deep Water south of the Southern ACC Front is marked by a pronounced halocline at ∼200m depth. The UCDW was characterised by a pHT,25 as low as 7.56, (low pHTis =7.87 and low oxygen, 175µmolkg−1)and the LCDW by high salinity and pHT,25 =7.61 (pHTis =7.93). Both the Circumpolar Deep Water masses are characterized by a maximum in CTconcentrations (2245–2260µmolkg−1, NCTof 2270–2280 µmolkg−1, from Eq. 3) attributed to the influence of old waters from the Indian and Pacific Oceans (Hoppema et al., 2000; Lo M´ onaco et al., 2005). The UCDW is located in the 300–600m range south of the PF with CT of 2255µmolkg−1(NCTof 2280µmolkg−1). This range of values compared well with those provided by Bakker et al. (2008) along 0◦and from 55◦S to 58◦S. They found www.biogeosciences.net/8/1401/2011/ Biogeosciences, 8, 1401–1413, 2011 1408 M. Gonz´ alez-D´ avila et al.: Carbonate system in the Southern Ocean in 2008 CTdata in the 2250–2260µmolkg−1range at the 300–400m range during December 2002. These values increased to over 2260µmolkg−1further south. The UCDW extends deeper north of the PF reaching 900–1500m at 49◦S with values of CTin the range 2240–2245µmolkg−1(NCTin the range of 2260 to 2265µmolkg−1)and low pHTis (7.87). The UCDW has higher concentrations of CFC-12 (>0.2pmolkg−1)as compared to deeper waters also revealing the mixing with recently ventilated waters. The influence of the Weddell Sea Deep Water (WSDW) was detected below 2000m and south of SBdy, with temperatures colder than 0◦C and salinities under 34.66. We found a pHT,25 of 7.62 (pHTis in the range 7.88–7.90), oxygen of 235µmolkg−1and CTaround 2250µmolkg−1in the WSDW. Close to the seafloor, the core of CFC-rich waters can be used to identify the presence of AABW (Mantisi et al., 1991; Orsi et al., 2002). The formation process of AABW also involves WW, Ice Shelf Water and deep waters of local origin including the Weddell Sea Deep Water (Wong et al., 1998), characterised by relatively high levels of CFC-12 >0.5pmolkg−1, and slightly higher oxygen (over 245µmolkg−1)and pHT,25 values (7.63) than in WSDW, but lower pHTis (below 7.85). A similar CFC distribution was also observed by Lo M´ onaco et al. (2005) along 30◦E in 1996. The Antarctic Intermediate Water (AAIW), located at the 600–1000m band is identified by the northward deepening of the salinity minimum at the near-surface waters of the Polar Frontal Zone, together with the deepening of CTand CFC-12 isolines. AAIW in this region is characterized by low pHT,25, ranging between 7.65 and 7.68 (pHTis =7.93) but slightly higher than those at UCDW. The AAIW core followed the 27.1 to 27.4 potential density line moving to 600–800m depth in the Cape Basin area, where it met the Indian AAIW injected with the Agulhas Rings. In the Cape Basin, salinity values were 0.2 units higher and the temperature was 2◦C warmer than at 45◦S. The AAIW also had higher dissolved inorganic carbon content, ranging from 2170µmolkg−1at 45◦S to values over 2185µmolkg−1in the Cape Basin, which also had lower CFCs than at 45◦S, in accordance with Fine et al. (1988). Along the northern part of the section, the deep salinity maximum is not associated with LCDW, but with diluted North Atlantic Deep Water. It has low CFC concentrations (<0.05pmolkg−1), as a sign of its age and its long passage across the Atlantic Ocean from its formation area. This water is one of the two NADW varieties found in the region. The present variety corresponds to the eastern NADW pathway that has crossed the South Atlantic at 20◦S–25◦S (Arhan et al., 2003) and then flows southeastward along the African slope as a slope current. It is usually found in the Cape Basin and north of the SAF. It is characterised by salinity maxima over 34.83. It is present as a uniform layer of pHT,25 >7.69 (pHTis =7.93), ATin the range 2340 to 2350µmolkg−1 (NATaround 2355µmolkg−1)and CTbetween 2210 and 2220µmolkg−1(NCTbetween 2215 and 2225µmolkg−1). The other variety of NADW is found south of the SAF in the APZ. It is characterised by salinity maxima between 34.74 and 34.79, and is associated with NADW injected in the ACC in the southwestern Argentine Basin (Whitworth and Nowlin, 1987). It should be regarded as a blend of LCDW and NADW injected into the ACC in the Argentine Basin. Concentrations of CFC-12 are in the range of 0.08 and 0.1pmolkg−1. The north to south uplifting of isolines showed the transition of water properties between mixed and pure LCDW water. South of 44◦S, this NADW variety presented pHT,25 between 7.66 and 7.67 (pHTis of 7.90–7.91), ATranging between 2345 and 2360µmolkg−1 (NATof 2365 and 2370µmolkg−1)and CTin the range of 2230–2240µmolkg−1, corresponding to a NCTof 2245 and 2255µmolkg−1. AABW can also be distinguished in the deepest part of the section in the form of a layer of slightly higher CFC12 values (>0.07pmolkg−1), low pHT,25 and pHTis (7.65 and 7.81) and high ATandNAT(2365–2370 and 2385– 2388µmolkg−1), spreading north reaching 36◦S. These values are lower for CFC-12 and slightly higher for CTand lower for pHTthan those shown above, south of the SBdy, indicating that this AABW is older and became diluted by the overlaying NADW from the south to the north, as indicated in Gladyshev et al. (2008). Calcite and aragonite saturation states, cal and ara, (Fig. 4) decrease from north to south in the first 600m. The isoline of cal =2 was located at 500–600m in the subtropical area, dropping to under 700m at the position of the old Agulhas ring M (the one located immediately north of the SAF), andreaching 100m inthe ACC zone, following what it has been shown by Hauck et al. (2010) on the same transect. The same vertical positions were followed for the isoline of ara =1.2. The influence of the UCDW and LCDW with higher carbonate ion concentration and older waters from the Indian and Pacific oceans maintained the cal over 1.5 at 1200–1500m south of the APF, and the isoline of cal =1.5 at 1000m north of the PF. Values below the calcite saturation horizon are located below 3800m in the subtropical area, 3300–3400m in the sub-Antarctic zone, and at around 3100–3200m south of the SBdy. The aragonite saturation horizon is at 1000m in the subtropical area and north of the SAF, as a result of the influence of eddy M, that presented decreased CTvalues (2101µmolkg−1at 500m) inside the eddy and higher values (2184µmolkg−1at 500m) outside the eddy field. The presence of the PF together with the influence of the UCDW and the LCDW made the aragonite saturation horizon deepen from 800m at 49.57◦S to 1500m at 50.37◦S. South of this latitude, the position of the saturation horizon shoaled at 700m at 57.5◦S. This distribution should be strongly seasonally affected (McNeil and Matear, 2008) due to the ventilation of deeper waters in the Southern Ocean south of the PF as a result of upwelling, due to winter cooling and strong persistent winds. These deep waters are CTrich and carbonate ion poor, lowering the carbonate Biogeosciences, 8, 1401–1413, 2011 www.biogeosciences.net/8/1401/2011/ M. Gonz´ alez-D´ avila et al.: Carbonate system in the Southern Ocean in 2008 1409 824 825 826 Figure 4 827 828 Fig. 4. The vertical distribution of the calcite saturation, cal, and aragonite saturation, ara, from February to March 2008 in the southeast Atlantic sector of the Southern Ocean. The calcite and aragonite saturation horizon is marked. ion concentration considerably. During summertime (McNeil and Matear, 2008), the shallow mixed layers evolve where the biological production depletes the CTand enriches carbonate ion concentrations. Subduction of these waters explains the vertical distribution of calcite and aragonite saturation states in the Polar sector observed during this cruise. 4.3 Sensitivity of carbonate system to increasing CO2 The decrease in the carbonate ion concentration and the pH as a direct result of the increased CO2in the atmosphere and in the surface ocean affects and will continue to affect the ocean chemistry. In order to account for the sensitivity to future changes, we used the experimental data to compute the fractional changes in [CO2] (γi), [H+] (βi)and  (ωi)induced by changes in CTand AT. Following Frankingnoulle (1994) and updated by Egleston et al. (2010) (Eq. 6) seven buffer factors have been used to quantify the ability of the ocean to slow down changes in carbonate chemistry. γCT=(∂ ln[CO2]/∂CT)−1 γAlk =(∂ ln[CO2]/∂Alk)−1 βCT=(∂ ln[H+]/∂CT)−1 βAlk =(∂ ln[H+]/∂Alk)−1= −βH/2.3 ωCT=(∂ ln/∂CT)−1 ωAlk =(∂ ln/∂Alk)−1(8) The buffer factors have dimensions of molkg−1of seawater (mmolkg−1throughout the text). Low values imply low buffering capacity and larger changes in [CO2], [H+] and for a given change in CTor AT. Rising atmospheric CO2concentrations increase the CTof the ocean, without changing its AT. However, biological feedback including biogenic calcification or enhanced calcite dissolution and increased mineral weathering may increase the alkalinity, and thus the capacity to store CO2(Egleston et al., 2010). Figure 5 depicts the vertical distribution of the seven buffer coefficients over the Bonus GoodHope section. Figure 5 also shows the ratio of carbonate alkalinity to inorganic carbon Ac/CT, which quantifies the number of protons released by CO2at the pH of the sample. Minimum absolute values for the buffer coefficients are found in waters with similar CT and ATvalues and pH of about 7.5. The presence of borate results in a minimum value not exactly located half-way between the two acidic constants for the carbonate system. In this region of pH (Santana-Casiano and Gonz´ alez-D´ avila, 2011), the CO2− 3, CO2and B(OH)− 4are present at very low concentrations, and small additions of acid or base reacting with HCO− 3strongly affect the concentration of CO2− 3, CO2, and pH. It should be noted that CO2additions leave the [HCO− 3] roughly constant while [H+] increases and [CO2− 3] decreases. Moreover, as [Ca2+] and the solubility products of calcium carbonate are invariant with changes in CTand AT(at constant salinity and temperature), ωCTand ωAT(see data and method section) describe how the [CO2− 3] changes with CTand AT, respectively. The values of γCTand βCT are positive while ωCTis negative because the addition of CO2to the seawater increases [CO2] and [H+] but decreases [CO2− 3]. Reversely, γATand βATare negative while ωATis positive because the addition of a strong acid (base) to seawater increases (decreases) [CO2] and [H+] but decreases (increases) [CO2− 3]. As indicated in Egleston et al. (2010), the Revelle factor, R, can be computed as R=CT/γCT. Over the whole of the cruise section, the ATwas higher than the CT, with an AT/CTratio ranging from 1.16 to 1.02. Ratios close to 1 and lower pH values correspond to minimum absolute values for the different buffer factors. These were observed in the 1000–1500m range north of the SAF, and approaching the 250 to 400m range south of the PF. These minimum values were found in the layer where the AAIW and UCDW were located. They clearly indicate that these water masses are particularly sensitive to increases in [CO2], and small increases of CTwill strongly decrease the pH and the carbonate saturation states. Mixing processes in the Southern Ocean bring up relatively low AT/CTwaters that mix with waters where biological production is not able to use up the macronutrients of these HNLC waters and to draw down the inorganic carbon. This results in low AT/CTratios. The pH and saturation states are also highly sensitive to changes in CTand ATin these waters, with values at the end of the austral summer of βH=0.36mmolkg−1and ωCT=0.12mmolkg−1in surface waters. These are minimum values for both parameters in the ocean, and the chemistry of these surface waters becomes much more sensitive to local variations in both CTand AT. www.biogeosciences.net/8/1401/2011/ Biogeosciences, 8, 1401–1413, 2011