ORNL/CDIAC-160 NDP-094 Climatological Distributions of pH, pCO , Total CO , Alkalinity, 2 2 and CaCO Saturation in the Global Surface Ocean 3 Contributed by Taro Takahashi, Stewart C. Sutherland, David W. Chipman, John G. Goddard Lamont-Doherty Earth Observatory of Columbia University, Palisades, NY 10964 and Timothy Newberger and Colm Sweeney Cooperative Institute in Environmental Sciences, University of Colorado, Boulder, CO 80309 Prepared by Alex Kozyr Carbon Dioxide Information Analysis Center Oak Ridge National Laboratory Oak Ridge, Tennessee, USA Date published: September 2014 Prepared for the Climate Change Research Division Office of Biological and Environmental Research U.S. Department of Energy Budget Activity Numbers KP 12 04 01 0 and KP 12 02 03 0 Prepared by the Carbon Dioxide Information Analysis Center OAK RIDGE NATIONAL LABORATORY Oak Ridge, Tennessee 37831-6335 managed by UT-BATTELLE, LLC for the U.S. DEPARTMENT OF ENERGY under contract DE-AC05-00OR2272 2 ABSTRACT Takahashi, T., S. C. Sutherland, D. W. Chipman, J. G. Goddard, T. Newberger and C. Sweeney. 2014. Climatological Distributions of pH, pCO , Total CO , Alkalinity, 2 2 and CaCO Saturation in the Global Surface Ocean. ORNL/CDIAC-160, NDP-094. 3 Carbon Dioxide Information Analysis Center, Oak Ridge National Laboratory, U.S. Department of Energy, Oak Ridge, Tennessee. doi: 10.3334/CDIAC/OTG.NDP094 Climatological mean monthly distributions of pH in the total H+ scale, total CO 2 concentration (TCO ), and the degree of CaCO saturation for the global surface ocean 2 3 waters (excluding coastal areas) are calculated using a data set for pCO , alkalinity and 2 nutrient concentrations in surface waters (depths <50 m), which is built upon the GLODAP, CARINA and LDEO database. The mutual consistency among these measured parameters is demonstrated using the inorganic carbon chemistry model with the dissociation constants for carbonic acid by Lueker et al. (2000) and for boric acid by Dickson (1990). The global ocean is divided into 24 regions, and the linear potential alkalinity-salinity relationships are established for each region. The mean monthly distributions of pH and carbon chemistry parameters for the reference year 2005 are computed using the climatological mean monthly pCO data adjusted to a reference year 2 2005 and the alkalinity estimated from the potential alkalinity-salinity relationships. The equatorial zone (4°N-4°S) of the Pacific is excluded from the analysis because of the large interannual changes. The pH thus calculated ranges from 7.9 to 8.2. Lower values are located in the upwelling regions in the tropical Pacific and in the Arabian and Bering Seas; and higher values are found in the subpolar and polar waters during the spring- summer months of intense photosynthetic production. The vast areas of subtropical oceans have seasonally varying pH values ranging from 8.05 during warmer months to 8.15 during colder months. The warm tropical and subtropical waters are supersaturated by a factor of as much as 4.2 with respect to aragonite and 6.3 for calcite, whereas the cold subpolar and polar waters are less supersaturated only by 1.2 for aragonite and 2 for calcite because of the lower pH values resulting from greater TCO concentrations. In 2 the western Arctic Ocean, aragonite undersaturation is observed. This report describes the data assembled for the study and the methods used for the computation of the pH, TCO , and degree of saturation of CaCO in global ocean surface waters. 2 3 3 1. INTRODUCTION Most active biological production takes place in the sun-lit top layer of the global ocean. Yet, this is the layer that is being acidified as a result of the absorption of rapidly increasing CO from the atmosphere. Many biogeochemical reactions in general will be 2 influenced by changes in ocean water pH: especially, calcifying organisms such as corals, foraminifera, pteropods and coccolithophores will be impacted directly due to the reduced degree of saturation of CaCO in seawater (Feely et al., 2012). At several time 3 series stations such as BATS (Bermuda Atlantic Time series Study) (Bates et al., 1996; Bates, 2007; Bates et al., 2012), HOT (Hawaii Ocean Time-series) (Dore et al., 2003; Dore et al., 2009), ESTOC (European Station for Time series in the Ocean Canary Islands) (Gonzalez-Dávila et al., 2003; Gonzalez-Dávila et al., 2007; Santana-Casiano et al., 2007; Gonzalez-Dávila et al., 2010) and Iceland Sea (Olafsson et al., 2009), the acidification of seawater has been documented for the past decades. The present pH conditions and future acidification of the global ocean waters by the ocean uptake of CO 2 have been reviewed by Feely et al. (2009) using a biogeochemical ocean GCM. More locally, the rapid acidification in the California Current system has been extensively investigated by Gruber et al. (2012) and Hauri et al. (2013) using a dynamic ocean circulation model combined with the multi-year observations. All of these studies indicate a significant degree of acidification in the surface ocean waters. However, although pH and its change have been measured at many study sites and during cruises using various pH sensor systems, the data are not globally comparable because of calibration problems and methodological differences associated with the pH measurements (Marion et al., 2011). Presently, the global distribution of ocean water pH is based on Biogeochemical Ocean General Circulation Models (BOGCM) (e.g. Orr et al., 2005; Doney et al., 2009; Feely et al., 2009) which include limited descriptions of marine ecosystems and community productions, but without land-ocean interactions. An observation-based global ocean pH distribution is desired for establishing a baseline for the future changes in oceanic carbonate chemistry, thus placing the ocean acidification study on a firmer ground. The objective of this study is to obtain the climatological mean distribution of pH and associated carbon chemistry properties in global surface ocean waters with a single unified scale for the calculated pH. These properties will be computed on the basis of the observations made for CO partial pressure (pCO ), total alkalinity (TALK) and total CO 2 2 2 concentration (TCO ) in surface waters (Z<50 m). It is imperative that these observations 2 are based on common standards which are stable over many decades to insure the compatibility of the data in space and time. In Section 2, the principles for the method employed in this study will be briefly explained. In Section 3, the potential alkalinity (PALK = TALK + nitrate) - salinity relationships observed in 24 oceanic regions will be 4 presented. The mean monthly distributions of TALK and pCO in surface waters of the 2 global ocean in a 4° x 5° spatial resolution, which will be used for computing pH and other carbon chemistry properties, will be shown. In Section 4, the computed results are tested against the observations made at various locations including the BATS, HOT and ESTOC sites, and the limitations for the method used for this study are discussed. In Section 5, the climatological distribution of pH, degree of CaCO saturation and TCO in 3 2 the global ocean will be presented. The database and results described in this report are used in the published paper by Takahashi et al. (2014). 2. PRINCIPLES The concentrations of H+ (or pH) and CO = ions and the degree of saturation for 3 calcite and aragonite in seawater may be computed using an inorganic equilibrium model for carbonate chemistry in seawater when temperature, salinity, pCO , TCO or the total 2 2 alkalinity are measured. The most desirable way for computing pH and carbonate chemistry parameters is to use pCO and TCO for two reasons: a) both pCO and TCO 2 2 2 2 measurements share the well-established air-CO gas mixture standards (e. g. Keeling, 2 1960 and the NOAA Earth System Research Laboratory, Boulder, CO); and b) the solubility of CO and dissociation constants for carbonic acid in seawater are the only 2 information needed. However, seasonal variability for pCO and TCO is large due to 2 2 seasonal changes in SST, net community production and upwelling of deep waters. While seasonal variability data for pCO are available for many locations in the global 2 oceans (e.g. Takahashi et al., 2009; Takahashi et al, 2013), the TCO observations are too 2 few to define seasonal changes other than those obtained at a few time-series stations. Hence, the TCO data are not sufficient for establishing the global distribution of pH and 2 other carbonate chemistry parameters, although the available data may be used for testing and validating the internal consistency of our analysis method. The carbonate chemistry in seawater may be also defined using a combination of pCO and the total alkalinity (TALK). However, this scheme requires additional 2 measurements for the concentrations of boric, phosphoric and silicic acids as well as the dissociation constants for each acid species, in order to correct for the contribution of these weak acids to the TALK. While the errors are negligibly small for the low nutrient subtropical gyre waters, the errors may be large for high nutrient waters in high latitude and upwelling areas. Surface water TALK is governed primarily by a) water balance (i.e. evaporation - precipitation), b) upwelling of high alkalinity deep waters (due to the dissolution of CaCO ), c) production of CaCO shells within the mixed layer and d) 3 3 mixing between waters with different characteristics. Since the vertical gradient for TALK in the upper five hundred meters of water column is much smaller (1/2 to 1/10) than that for TCO , and since the biological production of CaCO shells is commonly less 2 3 than 1/10 the net community production of organic carbon (with the exception of the coccolithophore blooms (Balch et al., 2005)), TALK in surface water is primarily related to salinity, and the TALK values normalized to a constant salinity are found to be seasonally invariant in the low nutrient oligotrophic waters (e. g. Bates et al., 1996; Bates, 2001). On the other hand, TALK depends also on the concentration of NO - (and 3 to small extent on phosphate), which varies seasonally between 0 and 30 µmol kg-1 (e. g. Takahashi et al., 1993; Hales and Takahashi, 2004) in high latitude waters due to deep water upwelling and biological utilization. Accordingly, TALK may be altered 5 seasonally by as much as 30 µeq kg-1. We, therefore, look for the salinity dependence of the potential alkalinity (PALK = TALK + NO -, Brewer and Goldman, 1976), which 3 compensates for the effect of change in NO -. An advantage of this scheme is that the 3 seasonal variability of PALK is generally small, and the seasonal change in TALK caused by nutrient change may be explicitly estimated from the salinity and nitrate data in a specific region. The observed regional differences in the PALK - salinity relationships reflect primarily changes in CaCO production in surface waters and mixing with high 3 TALK deep waters produced by the dissolution of CaCO . 3 Our approach for estimating TALK is in contrast to the previous work by Millero et al. (1998) and Lee et al. (2006), who parameterized TALK as a function of salinity and SST without including parameter(s) representing explicitly the net community production. In their formulations, SST was used as a proxy for the combined effect of complex seasonal processes including net community production, CaCO formation and 3 lateral/vertical mixing. In our study, the TALK change caused by seasonal net community production is corrected using seasonal NO - data. The results of these studies 3 will be compared with this study in Section 3.3.1. In this report, the climatological distributions for pH (in the total H+ scale), CO = 3 and other related values in a 4° x 5° grid are computed with an inorganic carbonate chemistry model using the climatological mean values for surface water pCO for the 2 reference year 2005 (an up-dated version of Takahashi et al., 2009), total alkalinity, salinity and temperature. In order to demonstrate the global integrity of the computed pH and other values, TCO values which are computed from the pCO -TALK data with the 2 2 chemical model are compared with the available TCO data. As shown in Section 4, the 2 computed values are found to be consistent with the observed values, with a few exceptions attributable to the relatively coarse 4° x 5° spatial resolution used in this study. The equilibrium carbon chemistry model consists of the following dissociation constants and CO solubility in seawater: Lueker et al. (2000) for carbonic acid, Dickson 2 (1990) for boric acid, Dickson and Riley (1979-a) for water, Dickson and Riley (1979-b) for phosphoric acid, Sjorberg et al. (1981) for silicic acid and Weiss (1974) for the CO 2 solubility. The dissociation constants for carbonic acid by Lueker et al. (2000) are selected because (a) the experiments are conducted by measuring all four variables (pCO , TCO , alkalinity and pH) in natural seawater samples, and (b) the pCO and 2 2 2 TCO are determined using the manometric standards established by Keeling (1960) and 2 the NOAA Earth System Research Laboratory, that are also used for the field observations used in this study. In this model, the total H+ ion scale (concentrations of free H+ plus HSO - ions (Dickson, 1993)) is used, and the effects of organic acids are not 4 included. The degree of saturation for calcite and aragonite are computed using the apparent solubility products (Ksp’) of Mucci (1983). 3. OBSERVATIONS AND ANALYSIS The availability of the observations needed for this study varies widely in each parameter. While about 6 million pCO data are available for the various seasons since 2 1950’s over the global oceans (Takahashi et al., 2013), there are only about 16,000 TCO 2 and alkalinity measurements which are mostly limited to summer time since 1980’s. In only 2,200 of them, pCO , TCO and TALK have been measured in the same samples. 2 2 6 We will first test the mutual consistency of these three properties using this set of data and the chemical model described in Section 2. Once this is confirmed, then the global distribution of pH, TCO and other carbon chemistry parameters will be computed using 2 the climatological mean monthly distribution of pCO and the alkalinity estimated using 2 the characteristic PALK-salinity relationships found in various ocean regions. The observed TCO values will be compared with the calculated values for the validation of 2 the method. 3.1. The TALK-TCO -pCO Database: 2 2 We assembled a carbon-nutrient database (named LDEO_SurCarbChem) for the upper 50 meters of ocean water measured during the TTO, SAVE, WOCE and CARINA programs. This database, which is described in this report, is built upon the data synthesis of the GLODAP database (Key et al., 2004), and also includes the TCO -TALK 2 pairs from the CARINA program (Tanhua et al., 2009; Key et al., 2010) and pCO and 2 TCO data from the LDEO database (also available to the public at the Biological and 2 Chemical Oceanography Data Management Office (BCO-DMO), Woods Hole Oceanographic Institution). The pCO data are based on the air-CO gas mixture 2 2 standards certified by Pieter Tans of NOAA Earth System Research Laboratory (ESRL), and the TCO data are consistent with the CO reference solutions prepared by Dickson 2 2 (2001), which are based on the Keeling CO standards (which are consistent with the 2 NOAA/ESRL values). Of about 16,000 records listed in this database, about 4,800 records are from the CARINA database, and 2,600 pCO data are from LDEO database. 2 Several thousand TCO measurements made by the LDEO group are already included in 2 the GLODAP database. Of these, 2,205 sets have the pCO , TCO and TALK in the 2 2 same samples. The mutual consistency between these three parameters in this subset is tested by comparing computed values with measured values (Figure 1). The calculated values are found to be consistent with the measured values within the measurement uncertainties respectively: the TALK values computed using pCO and TCO as inputs 2 2 are in agreement with the measured values with a root mean square deviation (RMSD) of ± 3.9 µeq kg-1; the computed TCO values from the pCO and alkalinity data are in 2 2 agreement with the measured values with ± 3.3 µmol kg-1; and the computed pCO values 2 from the TCO and TALK data are in agreement with the measured values with ±6.8 2 µatm. This demonstrates that these three properties are mutually consistent within the context of the equilibrium model, and that one property may be computed reliably when two others are given. --------------------------------------------------------- Figure 1 – Comparison of the observed values with the calculated values from the pCO , 2 TCO and alkalinity data for 2,205 sets of observations in surface waters. The black line 2 shows the linear regression line, and the red crosses and the red right-hand axis indicate the deviations around the regression line. (Top) The calculated TALK values using pCO 2 and TCO are in agreement with the measured values within RMSD of ± 3.9 µeq kg-1; 2 (Middle) the calculated TCO values using pCO and alkalinity are in agreement with the 2 2 measured values within RMSD of ± 3.3 µmol kg-1; and (Bottom) the calculated pCO 2 values using TCO and TALK are in agreement with the measured values within ± 6.8 2 µatm. 7 --------------------------------------------------------- 3.2 Potential Alkalinity-Salinity Relationships: We use the “LDEO SurfCarbChem” database described above, and demonstrate that the PALK varies linearly with salinity within respective oceanic regions. The value of TALK, which is needed for the carbonate chemistry calculation, is estimated using the PALK-salinity relationship and the NO - data. The PALK-salinity relationship depends 3 primarily on the evaporation-precipitation (E-P) of water, the formation and dissolution of CaCO as well as mixing with waters of different characteristics (including river and 3 deep waters). Five different cases showing how the slope of the linear PALK-salinity relationships varies with different oceanographic situations are illustrated in Figure 2. 8 ---------------------------------------------------- Figure 2 – Variations of the potential alkalinity-salinity relationships. The effect of evaporation and precipitation is indicated by the light solid line. The slope of the line is altered by the production/dissolution of CaCO , upwelling of deep waters, mixing with 3 river waters, and mixing with seawaters with different compositions. (A) (B) ------------------------------------------- (1) The change of PALK due to the evaporateon-precipitation (E-P) of water is depicted by a straight line passing through the origin and the source water (heavy open circle in Figures 2-A and-B). An example is found in the Central Tropical North Pacific region in Figure 5 (green line). (2) As shown in Figure 2-A, subtropical waters that have higher salinity and reduced PALK due to the growth of calcareous organisms (downward red arrow in the left panel) are represented by the small filled circle: in contrast, subpolar waters that are less salty and enriched in PALK due to deep water upwelling (small filled circle with a upward red arrow). The mixing line between these two water types (the heavy straight line) has a shallower slope than the E-P line. This type of relationships is commonly observed in the subtropical gyres in the Atlantic (Figure 4), Pacific (Figures 5 and 6) and Indian Ocean (Figures 7 and 8), and is represented by the positive intercept values in regression lines (summarized in Table 1). (3) Figure 2-B shows the case when PALK in lower salinity waters is reduced by the growth of calcareous organisms. The mixing line for this case has a slope steeper than the E-P line. Such a case is found in the Panama Basin (Figure 5, red line), which is represented by a negative intercept value in the regression line (Table 1). (4) The source water may be mixed with river water that has a higher PALK value as shown by the dashed line in Figures 2-A and -B. River waters are commonly fresh and contain PALK derived from the dissolution of limestones. This type of relationships is found in the Okhotsk Sea (black line in Figure 5) and the Bengal Basin (red line in Figure 7), which also have positive intercept values for regression lines (Table 1). 9 (5) The source water may be also mixed with waters which have high salinity and reduced PALK (marked Warm Marginal Sea in Figure 2-B). Such a trend may be found for the mixing with an outflow from a warm evaporative basin like the Red Sea (green line in Figure 7). The characteristic linear PALK-salinity relationships are observed in 27 ocean regions, and they are presented in Figures 3 through 9. However, the relationships in six regions can not be described satisfactorily using linear equations because of time-space variation of complex mixing of various water types: the subtropical-subarctic transition zones due to intense eddy mixing, and the Pacific equatorial belt affected by El Niño events due to reorganization of ocean dynamics (Feely et al., 2002). These zones are indicated with the tangerine color in the sample distribution maps of Figures 4 - 8, and are marked with special symbols in our maps. The linear regression equations and root mean square deviations (RMSD) are summarized in Table 1. Of the 33 regions listed in Table 1, five regions which are marked “Transition Zone” and one marked “El Niño Zone” are excluded from further analysis. In addition, the three Southern Ocean regions for the Atlantic, Indian and Pacific sectors are combined together as a single Circumpolar Southern Ocean region (see Section 3.2.7). The remaining 24 regions have the RMSD values ranging between ± 4 and ± 17 µeq kg-1 with a mean RMSD of ± 8.2 µeq kg-1 (excluding ± 60.5 µeq kg-1 for the Beaufort Sea, Table 1), and are used in this study. The level of deviations is considered acceptable in comparison with the precision of ± 6 µeq kg-1 estimated for the PALK measurements. A few regions with large RMSD will be commented in the respective sections below. ---------------------------------------------------------------------------------------------------------- Table 1 –The potential alkalinity-salinity relationships in 33 ocean areas. Five regions which are marked “Transition Zone” and one marked “El Niño Zone” are excluded from the analysis. The Atlantic, Indian and Pacific sectors of the Southern Ocean are combined together as a single Circumpolar Southern Ocean region (see Section 3.2.7). The remaining 24 regions are used in producing the climatological distributions of TALK. PALK (µeq kg-1) = (Slope) x (Salinity, PSS) + Intercept (µeq kg-1), and RMSD = Root mean square deviation (µeq kg-1). 10 REGRESSION   REGIONS   LOCATIONS   EQUATIONS   BOX  AREAS     Slope   Intercept   RMSD           ARCTIC  &  GIN  SEAS             1   West  GIN  Seas   66°N  -­‐  80°N,  30°W  -­‐  0°   14.12   1796.2   6.1   68N-­‐76N,  332.5  to  357.5   2   East  GIN  Seas   66°N  -­‐  80°N,  0°  -­‐  30°E   59.57   232.0   12.3   68N-­‐76N,  2.5  to  32.5   3   High  Arctic   N  of  80°N   27.30   1340.7   16.8   80N-­‐88N,  2.5  to  357.5   4   Beaufort  Sea   66°N  -­‐  80°N,  140°W    -­‐  180°   61.29   285.8   60.5   68N-­‐76N,  180  to  222.5   5   Labrador  Sea   55°N  -­‐  80°N,  85°W  -­‐  45°W   37.27   1016.2   17.2   56N-­‐76N,  277.5  to  312.5   6   Subarctic  Atlantic   55°N  -­‐  80°N,  40°W  -­‐  10°E   45.37   730.6   6.7   56N-­‐76N,  317.5  to  7.5   NORTH  &  SOUTH  ATLANTIC  OCEAN             7   N.  Atlantic  Drift   40°N  -­‐  55°N,  60°W  -­‐  10°E   45.30   733.0   6.5   44N-­‐52N,  302.5  to  7.5   8   Central  Atlantic   40°S  -­‐  40°N,  Coast  to  Coast   58.25   270.9   12.6   40S-­‐40N,  Ocean  width   9   S.  Atlantic  Transition  Zone   40°S  -­‐  48°S,  65°W  -­‐  30°E   30.27   1259.4   7.6   44S-­‐48S,  297.5  to  27.5   10   S.  Atlantic  Subpolar  Front  Zone   48°S  -­‐  60°S.  70°W  -­‐  30°E   Transition  Zone   48S-­‐56S,  292.5  to  27.5   11   Antarctic  (Atlantic)   60°S  -­‐  70°S,  70°W  -­‐  30°E   57.78   367.8   5.4   60S-­‐68S,  292.5  to  27.5   NORTH  &  SOUTH  PACIFIC  OCEAN             12   Kuroshio-­‐Alaska  Gyre   34°N  -­‐  40°N,  Ocean  width   36N  ,  Ocean  width          &  N  of  40°N,  E  of  150°W   44.88   724.8   9.0    40N-­‐60N,  107.5  to  207.5       13   N.  Central  Pacific   44°N  -­‐  60°N,  W  of  150°W   79.92   -­‐395.7   14.7   44N-­‐56N,  207.5  to  237.5   14   N.  Pacific  Subpolar  Front  Zone   40°N  -­‐  44°N,  140°E  -­‐  145°W   Transition  Zone   40N-­‐44N,  137.5  to  217.5   15   Okhotsk  Sea   47°N  -­‐  60°N,  W  of  152°E   59.37   301.4   8.9   48N-­‐60N,  122.5  to  152.5   16   N.  Pacific  Subtropical  Front  Zone   30°N  -­‐  34°N,  130°E  -­‐  120°W   Transition  Zone   32N,  142.5  to  242.5   17   Central  Tropical  N.  Pacific   5°N  -­‐  30°N,  W  of  120°W   65.55   9.4   8.9   8N-­‐28N,  107.5  to  122.5   18   Tropical  East  N.  Pacific   5°N  -­‐  30°N,  120°W  -­‐  95°W   82.20   -­‐553.2   9.7   8N-­‐28N,  127.5  to  262.5   19   Panama  Basin   0°N  -­‐  15°N,  95°W  -­‐  80°W   74.27   -­‐290.5   8.6   8N-­‐12N,  267.5  to  282.5   20   Equatorial  Pacific   5°S  -­‐  5°N,  130°E  -­‐  80°W   El  Nino  Zone   4S-­‐4N,  127.5  to  277.5   21   Central  South  Pacific   5°S  -­‐  35°S.  150°E  -­‐  84°W   66.64   -­‐28.4   9.4   8S-­‐36S,  142.5  to  277.5   22   E.  Central  South  Pacific   5°S  -­‐  60°S,  E  of  84°W   58.88   268.5   4.0   8S-­‐56S,  282.5  to  307.5   23   Subpolar  S.  Pacific   35°S  -­‐  50°S,  180°  -­‐  84°W   45.10   733.6   7.8   34S-­‐48S,  180  to  277.5   24   S.  Pacific  Subpolar  Front  Zone   50°S  -­‐  65°S,  150°E  -­‐  70°W   Transition  Zone   52S-­‐58S,  152.5  to  297.5   25   Antarctic  (Pacific)   65°S  -­‐  75°S,  140°E  -­‐  65°W   81.69   -­‐450.8   6.7   64S-­‐80S,  137.5  to  292.5   NORTH  &  SOUTH  INDIAN  OCEAN             26   Main  North  Indian   5°N  -­‐  25°N,  48°E  -­‐  100°E   57.07   302.3   6.7   4N-­‐24N,  47.5  to  102.5   27   Red  Sea   12°N  -­‐  27°N,  34°E  -­‐  48°E   26.27   1417.2   6.3   12N-­‐24N,  32.5  to  47.5   28   Bengal  Basin   10°N  -­‐  20°N,  80°E  -­‐  100°E   39.66   894.2   10.7   12N-­‐24N,  77.5  to  102.5   29   Main  South  Indian   5°N  -­‐  40°S.  30°E  -­‐  140°E   65.03   20.0   7.6   4N-­‐36S,  27.5  to  142.5   30   S.  Indian  Transition   40°S  -­‐  48°S,  30°E  -­‐  175°E   23.76   1486.1   5.5   40S-­‐48S,  27.5  to  172.5   32   S.  Indian  Subpolar  Front  Zone   48°S  -­‐  52°S,  30°E  -­‐  150°E   Transition  Zone   Overlap  above  &  below   32   Antarctic  (Indian)   52°S  -­‐  70°S,  25°E  -­‐  165°E   62.57   202.0   6.6   52S-­‐68S,  22.5  to  167.5   SOUTHERN  OCEAN             33   Circumpolar  Southern  Ocean   S  of  60°S,  All  Southern     74.13   -­‐192.3   9.1   60S-­‐84S,  2.5  to  357.5   -----------------------------------------------------------------------------------------------------------