JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 103, NO. C2, PAGES 2883-2896, FEBRUARY 15, 1998

Distribution of COz species, estimates of net community production, and air-sea COz exchange in the Ross Sea polynya

Nicholas R. Bates, Dennis A. Hansell, and Craig A. Carlson Bermuda Biological Station for Research, St. George's, Bermuda

Louis I. Gordon

College of Oceanic and Atmospheric Sciences, Oregon State University, Corvallis

Abstract. Measurements of surface total carbon dioxide (TCO2) , alkalinity, and calculated pCO2, along with water column nutrients and hydrography, were made on two cruises to the Ross Sea polynya (NBP 94-6, November-December 1994 and NBP 95-8, December 1995 to January 1996). The polynya experiences an intense phytoplankton bloom during a short period of open water conditions from mid-December to mid- February each year. Our biogeochemical observations were used to determine the temporal variability of CO2, fluxes of carbon within the ocean, and rates of air-sea exchange of CO2. Depletions of TCO2, pCO2, and nitrate+nitrite were considerable (-70-150/•mol kg -•, 80-150/•atm, and 10-20/•mol kg -•, respectively) a. nd associated primarily with biological uptake during Phaeocystis and diatom blooms. Alkalinity was a conservative tracer of salinity and nitrate+nitrite. Surface pCO 2 was undersaturated by -50-150/•atm, and air-sea gas exchange of CO2 during open water conditions was directed from atmosphere to ocean. Observed surface stoichiometric C:N ratios were 6.66: 1 and 6.77:1 for the 2 years, consistent with global "Redfield" ratios, while C:P and N:P ratios were variable (75-141'1, 12-!8:1). Estimates of net community production (NCP) rates were made using in situ changes in TCO 2 and nitrate+nitrite across repeated transects along 76ø30'S. Mean NCP rates across the polynya ranged from 0.86 to 0.98 g C __-2 d-1 m . These values may be underestimated by 5-25% because of the contribution of atmospheric CO2 to the surface layer through gas exchange. Export of carbon from the surface to depth was at least 55-60% of NCP rates.

1. Introduction

The Southern Ocean has an important role in the global carbon cycle since much of the global oceanic uptake of atmo- spheric CO2 occurs in this area [e.g., Tans et al., 1990; Taka- hashi et at., 1993; Poisson et at., 1993; Robertson and Watson, 1995; Sarmiento and Le Qu•r•, 1996]. Model analyses indicate that solubility effects rather than biological production appear to be the primary driving force for sequestering atmospheric CO2 [Siegenthaler and Sarmiento, 1992; Sarmiento and Le Qu&•, 1996]. There is no clear consensus about the importance of biology for oceanic CO 2 uptake in the Southern Ocean, and an improved understanding of the biogeochemical controls of the carbon cycle is clearly needed.

In the Southern Ocean, geographically limited spring phy- toplankton blooms occur in the Antarctic Polar Front or within ice marginal continental seas such as the Weddell, Belling- hausen, and Ross Seas [Sakshaug, 1984; El-Sayed, 1988; Arrigo and McClain, 1994; Smith, 1995; Boyd et al., 1995]. The most spatially extensive and predictable phytoplankton bloom in the Southern Ocean occurs on the Ross Sea continental shelf

[Holm-Hansen et al., 1982; Zwally et al., 1983; Sullivan et al., 1983; El-Sayed et al., 1983; Smith and Nelson, 1985, 1990;

Copyright 1998 by the American Geophysical Union.

Paper number 97JC02473. 0148-0227/98/97JC-02473509.00

Comiso et al., 1993;Arrigo and McClain, 1994]. Katabatic Winds off the Ross Ice Shelf form a coastal polynya of thin ice and exposed surface waters surrounded by dense multiyear ice [Comiso et al., 1993; Arrigo and McClain, 1994]. Ice-free con- ditions are typically present between mid-December and mid- February. A phytoplankton bloom of Phaeocystis a•tarctica develops in the central and eastern parts of the polynya in late November to early December, reaching maximum biomass in January and February [El-Sayed et al., 1983; Smith and Gordon, 1997]. A diatomaceous bloom also develops in the western- most part of the polynya near the coast and within Terra Nova Bay [Smith and Nelson, 1985; Arrigo and McClain, 1994].

Biological production in high-latitude areas causes signifi- cant depletion of surface nutrients, total carbon dioxide (TCO2) , and pCO 2 [e.g., Codispoti et al., 1982, 1986; Anderson and Jones, 1991; Anderson et al., 1991; Hansell et al., 1993; Takahashi et al., 1993; Yager et al., 1995]. However, there are few studies of inorganic carbon in continental seas of the Ant- arctic or ice marginal zones of the Southern Ocean [e.g., Ta- kahashi and Chipman, 1982; Karl et al., 1991; Hoppema et al., 1995]. Consequently, there is a need to better understand the processes and controls of the biogeochemical cycle of carbon in these regions. In early November 1994 and December 1995 the Research Vessel Ice Breaker (RVIB) Nathaniel B. Palmer was the first ship to visit the Ross Sea polynya in spring to inves- tigate biological dynamics and carbon cycling [Smith and Gor-

2883

2884 BATES ET AL.: CARBON CYCLING IN THE ROSS SEA POLYNYA

170øE

74øS .......

75øS • •"••- 76os

77øs I ••y• • •ø•l•

175øE 180øW 175øW 170øW

Ross Sea

ß .

Figure 1. Location map of the western Ross Sea, with gen- eral position of transects along 76ø30'S. The approximate ex- tent of open water conditions (<15% sea-ice) is shown for December 1994 (long-dashed line) and January 1996 (short- dashed line). This is modified from an image reported by Smith and Gordon [1997].

don, 1997]. in this paper, we describe the first measurements of CO2 species in the Ross Sea polynya. During the two visits to the polynya, sampling of TCO2, alkalinity, and biogeochemical and hydrographic properties was repeated five times on a transect along 76ø30'S (November-December 1994 and De- cember 1995 to January 1996). Our field program sampled different conditions in the polynya: late wintertime with ice cover, initial opening of the polynya and development of phy- toplankton bloom, and time of maximum biomass of the bloom [Smith and Gordon, 1997]. We investigated the stoichiometric relationships between CO2 species and nutrients as a tool for estimating rates of net community production (NCP) in the polynya. Difficulties with extrapolating in vitro primary pro- duction estimates to regional and seasonal scales can be par- tially overcome by quantifying mesoscale in situ changes in nitrate+nitrite [e.g.,Jennings et al., 1984; Sambrotto et al., 1986; Minas and Codispoti, 1993] TCO 2 [e.g., Codispoti et al., 1982, 1986; Chipman et al., 1993; Yager et al., 1995], or oxygen [e.g., Shulenberger and Reid, 1981; Emerson et al., 1993]. Here we estimate NCP from in situ changes in TCO 2 and nitrate+nitrite along repeated transects (76ø30'S). Finally, we discuss the importance of biological production for air-sea gas exchange of CO2 in the Ross Sea polynya.

and January 8-12, 1996, hereinafter referred to as transect 3, transect 4, and transect 5, respectively.

TCO 2 and alkalinity were collected using standard sampling protocols [Dickson and Goyet, 1994], poisoned with HgC12 to prevent biological alteration, and shipped to the Bermuda Bi- ological Station for Research (BBSR) for later analysis. TCO 2 was determined by coulometry using a Single Operator Mul- tiparameter Metabolic Analyzer (SOMMA) system designed by K. M. Johnson and coworkers [Johnson et al., 1985, 1987, 1993] and standard analytical protocols [Johnson, 1992; Dick- son and Goyet, 1994]. A highly precise (0.02% or _+0.4/xmol kg -•) and accurate coulometric technique was used for TCO 2 analysis at BBSR. Since 1992, TCO 2 samples have been ana- lyzed from the U.S. Joint Global Ocean Flux Study (JGOFS) Bermuda Atlantic Time Series Study (BATS) site in the Sar- gasso Sea [Bates, 1993, 1995; Bates et al., 1996a, b]. The mean difference between replicate samples for over 1000 samples (including Ross Sea samples) analyzed at BBSR between 1992 and 1996 was 0.41/xmol kg -•, while the standard deviation of over 100 triplicate or greater samples was 0.37/xmol kg -•. The internal consistency and comparability of both Ross Sea data sets were ensured by routine analyses of seawater Certified Reference Material (CRM) (CRMs were provided by A. G. Dickson, Scripps Institution of Oceanography, 1995, 1996) [Dickson, 1990]. The accuracy of the TCO 2 measurement was 0.025% (_+0.5/xmol kg-•). Alkalinity was determined using a semiclosed potentiometric titration system modified from ear- lier techniques [Edmond, 1970; Bradshaw et al., 1981; Brewer et al., 1986] and described by Bates et al. [1996a, b]. The precision of the alkalinity measurement was 0.06% (-1.2/xmol kg-•). CRMs were also routinely analyzed for alkalinity. The accuracy of alkalinity was 0.15% (_+3 /xmol kg -•) when compared to alkalinity values determined for CRMs (A. G. Dickson has certified alkalinity values for most batches of CRM seawater on the following web page: http://www-mpl.ucsd.edu/people/ adickson/). TCO2, alkalinity, temperature, salinity, nutrient data, and dissociation constants of Goyet and Poisson [1989] and Roy et al. [1993] were used to calculate carbonate alkalinity and pCO 2 [Peng et al., 1987; Bates et al., 1996b]. Surface and atmospheric pCO 2 data were collected on transects 1 and 2 (1994) using an infrared detector to analyze equilibrated sea- water from a continuous stream of seawater (T. Takahashi, unpublished data, 1994). Comparison between direct surface pCO 2 and calculated measurements indicated that the uncer- tainty associated with calculating pCO 2 was _+5-10/xatm. All TCO 2 and alkalinity data are reported in Tables 1 and 2. Nutrient, temperature, and salinity data used here were re- ported previously [Gordon et al., 1996a, b].

2. Methods

Seawater samples were collected for TCO 2 and alkalinity analyses during two cruises to the Ross Sea polynya in 1994 and 1995-1996 aboard the RVIB Nathaniel B. Palmer. On the

first cruise (NBP 94-6) from November to December 1994, 28 surface stations were sampled in the Ross Sea. The dates of two transects along 76ø30'S from 175øW to 168øE were No- vember 14-16 and December 2-6 (Figure 1), which we refer to as transect 1 and transect 2, respectively. On the second cruise (NBP 95-8) from December 1995 to January 1996, 36 surface stations were sampled during three repeat transects along 76ø30'S from 171øW to 163øE. The dates of the transects were

December 21-29, 1995, December 30, 1995, to January 8, 1996,

3. Results and Discussion

3.1. Patterns of Hydrography and Biogeochemistry

The pattern, magnitude, and timing of production in the Ross Sea polynya appear to be similar each year [Arrigo and McClain, 1994; Smith and Gordon, 1997]. Our five repeat hy- drographic transects along 76030 ' S, although separated in time by a year, provide "windows of observation" on key hydro- graphic and biogeochemical changes in the short-lived polynya. Such changes encompass conditions ranging from late winter ice cover to opening of the polynya and exposure of surface waters and to subsequent development, acme, and senescence of phytoplankton blooms.

On the first visit to the Ross Sea polynya in November 1994

BATES ET AL.: CARBON CYCLING IN THE ROSS SEA POLYNYA 2885

Table 1. Summary Data From Cruise NBP-6 (November-December 1994) to the Ross Sea

Temperature, Phosphate, Nitrate + Nitrite, TCO2, Alkalinity, Station Cast øC Salinity /xmol kg -• /xmol kg -• /xmol kg-• /xmol kg-•

15 30 - 1.799 34.740 2.16 30.488 2242.3 2354.0 21 44 - 1.797 34.698 2.03 27.844 2218.1 2352.4 23 49 - 1.795 34.440 2.14 30.773 2221.3 2334.1 27 58 - 1.811 34.435 1.84 27.718 2218.1 2338.9 31 64 - 1.775 34.497 1.99 27.392 2197.3 2339.8 33 67 - 1.754 34.493 1.89 25.731 2203.2 2342.7

36 72 - 1.873 34.531 2.04 28.129 2190.3 2345.3 39 83 - 1.609 34.212 2.11 29.94 2213.6 2321.0 43 91 - 1.791 34.271 2.03 28.772 2201.9 2323.3 45 93 - 1.782 34.362 1.93 26.866 2200.9 2353.4 47 97 - 1.753 34.365 1.78 26.938 2182.9 NM 51 105 - 1.835 34.643 2.12 30.543 2234.5 2352.0 53 109 - 1.723 34.365 1.76 23.619 2168.7 2333.8 54 112 - 1.733 34.283 1.86 25.65 2181.6 2329.4 56 116 - 1.746 34.303 1.77 24.343 2176.5 2324.7 58 119 - 1.733 34.313 1.80 22.263 NM NM

58 120 - 1.761 34.305 1.68 21.131 2158.2 2332.8 60 123 - 1.788 34.308 1.98 27.816 2198.4 2328.3 62 126 - 1.718 34.308 1.82 24.136 2172.8 2337.6

64 130 - 1.764 34.441 1.60 19.725 2148.0 2350.5 66 133 - 1.672 34.411 1.71 22.331 2162.9 2332.5 68 136 - 1.663 34.325 1.78 23.951 2166.1 NM 70 139 - 1.642 34.444 1.58 18.896 2145.6 NM 72 145 - 1.767 34.609 1.70 22.089 2174.4 NM 74 148 - 1.771 34.663 1.99 28.947 2223.1 NM

75 150 - 1.84 34.647 1.95 27.978 2214.8 NM

77 153 - 1.798 34.480 1.76 23.494 2182.8 2336.9

Nitrate+nitrite and phosphate data have been previously reported by Gordon et al. [1996a]. NM, no measurement.

we encountered thin ice coverage and strong sustained winds (up to 18 m s -•) along transect 1. Surface waters exhibited temperature ranges between - 1.65 ø and - 1.85øC, salinity vari- ations from 34.2 to 34.7, and an average mixed layer depth of 29 m (Figures 2a, 2b, and 2d). Surface hydrographic properties were similar to intermediate- and deep-water values and were representative of wintertime conditions. Although the polynya was ice covered, a small bloom of Phaeocystis antarctica had begun with average primary production rates of 0.52 g C m -2 d -• [Smith and Gordon, 1997]. At most stations, surface and mixed layer nitrate + nitrite and phosphate concentrations were similar to intermediate- and deep-water concentrations (e.g., 30.1 and 2.17 •mol kg -•, see Figures 3a and 5a). At several stations in the central part of the polynya, surface nitrate de- pletion was observed, coinciding with elevated chlorophyll a concentration [Smith and Gordon, 1997]. On transect 1, sta- tions with wintertime concentrations of nitrate+nitrite and

phosphate (i.e., 30.1 and 2.17 •mol kg -•, respectively) had TCO 2 values between -2200 and 2250 •mol kg -• and surface seawater pCO 2 values of 300-360 •atm, close to atmospheric values (Figures 3b and 3d).

Several weeks later (transect 2), surface thermohaline con- ditions were similar to transect 1, but open water conditions were present. P. antarctica dominated the phytoplankton as- semblage, and chlorophyll a concentrations and primary pro- duction rates were higher [Smith and Gordon, 1997]. In the central region of the polynya, nitrate+nitrite and phosphate depletion was observed (Figure 3) compared to the earlier transect and wintertime values (e.g., 30.1 and 2.17 •mol kg-•). TCO 2 and p CO 2 concentrations along 76ø30'S on transect 2 were considerably lower compared to wintertime values, rang- ing from -2150 to 2190 •mol kg-• and from 240 to 270 •atm, respectively (Figure 3). The loss of TCO 2 was of the order of

50-90 •mol kg- • and located in the central part of the polynya (-170øE to 178øW) where production and nutrient depletion were greatest. Similar depletion of TCO 2 and pCO 2 has been observed in the Weddell Sea [Takahashi and Chipman, 1982; Hoppema et al., 1995] and Bransfield Strait [Karl et al., 1991].

One year later, we repeated three transects along 76ø30'S between mid-December 1995 and mid-January 1996 (transects 3-5). We encountered open water conditions and moderate sustained winds (5-10 m s -•) on all three transects. Surface temperatures were significantly warmer than late wintertime conditions the previous year (Figure 2a). On Transects 3 and 4 (Figure 2a) surface temperatures ranged between -0.5 and 1.0øC in the central and eastern parts of the polynya. By mid- January (transect 5) surface waters had decreased by 0.1-0.3øC because of cooling of air temperatures and deepening of the mixed layer (Figure 2d). Surface salinity in the central and eastern parts of the polynya showed little variability, ranging between 34.0 and 34.4 (Figure 2b). In the westernmost part of the polynya between 165øE and 170øE (Figures 2a and 2b), slightly lower temperatures (-1.0 ø to -1.5øC) and salinities (33.5-34.0) were observed. Differences in phytoplankton as- semblage were also observed [Carlson et al., 1997]. P. antarctica dominated in the central and eastern parts of the polynya, while diatoms dominated in the westernmost part of the area. Similar spatial distributions in the phytoplankton assemblage have been previously observed in the Ross Sea [Nelson and Smith, 1986; Nelson et al., 1991; Smith et al., 1996; Nelson et al., 1996]. Considerable nutrient depletion (when compared to wintertime and deep-water values) was observed on transect 3. Further nutrient depletion was observed by transects 4 and 5, and nitrate + nitrite concentrations reached a minimum of- 11

•mol/kg in the western part of the polynya (Figure 3). TCO 2 and pCO 2 concentrations in the central and eastern parts of

2886 BATES ET AL.: CARBON CYCLING IN THE ROSS SEA POLYNYA

Table 2. Summary Data From Cruise NBP 95-8 (November-December 1994) to the Ross Sea

Temperature, Phosphate, Nitrate + Nitrite, TCO2, Alkalinity, Station Cast øC Salinity /xmol kg- • /xmol kg- • /xmol kg -• /xmol kg- •

4 7 - 1.289 34.026 1.81 26.3 2169.4 2317.0 5 10 - 1.260 33.819 1.47 22.6 2137.6 2302.0 7 13 - 1.297 34.097 1.70 24.4 2160.8 2319.5 9 18 -0.969 34.252 1.56 22.7 2158.3 2329.6

10 19 - 1.051 34.240 1.56 23.1 2162.0 2331.5 12 23 -0.797 34.321 1.66 22.8 2163.2 2332.2 14 26 -0.945 34.479 1.73 22.4 2157.8 2347.7 16 29 -0.757 34.429 1.70 22.3 2161.8 NM 17 33 -0.623 34.497 1.62 19.9 2140.0 2349.9 18 35 -0.423 34.353 1.71 23.5 2170.1 2337.2 20 38 -0.490 34.381 1.54 21.0 2151.4 2314.2 22 43 0.745 34.268 1.21 12.0 2082.0 2350.2 23 45 - 1.411 34.009 1.39 17.0 2096.7 2321.1 24 46 - 1.147 33.938 1.22 19.0 NM NM 26 49 0.663 34.519 1.01 15.8 2131.4 2358.7 27 51 0.899 34.058 1.10 16.4 2111.9 2327.3 28 53 0.975 34.451 1.26 13.6 2095.9 2354.0 32 58 1.397 34.427 1.49 18.9 2139.8 2343.1 35 61 1.222 34.450 1.44 18.3 2137.6 2353.7 36 62 1.029 34.239 0.89 15.4 2116.5 2340.2 40 69 0.477 34.277 0.90 14.5 2103.8 2344.1 47 80 -0.724 33.596 0.63 12.4 2056.0 2299.6 49 83 0.200 34.090 1.21 14.0 2093.5 2334.7 54 88 0.155 34.382 1.64 22.3 2160.6 2346.8 61 96 0.092 34.265 1.42 19.8 2132.1 2339.7 65 100 -0.074 34.110 1.63 23.9 2152.6 2324.2 68 104 -0.477 34.073 1.42 21.4 2139.3 2326.9 70 107 0.606 34.127 1.62 22.4 2162.9 2324.2 73 112 -0.085 34.242 1.47 20.1 2136.5 2331.2 74 114 -0.125 34.194 1.47 20.1 2135.5 2335.7 75 115 -0.639 34.436 1.77 23.8 2175.1 2345.3 77 118 -0.266 34.377 1.72 23.2 2167.8 2340.3

79 121 -0.294 34.278 1.84 25.4 2169.7 2333.1 81 124 -0.065 34.348 1.61 22.6 2161.8 2342.8 83 127 -0.120 34.389 1.61 21.3 2161.7 2347.3 86 132 -0.971 33.409 1.11 15.0 2056.9 2291.5

Nitrate+nitrite and phosphate data have been previously reported by Gordon et al. [1996b]. NM, no measurement.

the polynya (transects 3-5) were also significantly lower than wintertime concentrations (Figure 3). The lowest TCO 2 and pCO 2 concentrations (---2050/xmol kg -• and ---160/xatm, re- spectively) were observed in the eastern part of the polynya (transects 3-5), associated with the diatom bloom. This feature is partly due to ice melt but is still apparent when TCO 2 is normalized to a constant salinity (nTCO2) to remove the ef- fects of physical processes such as ice melt, evaporation, and precipitation (Figure 3).

3.2. Relationships Between CO2 Species, Nutrients, and Hydrography

Few alkalinity measurements have been previously reported for the Southern Ocean or continental seas. Our observations

in the Ross Sea polynya indicate that surface alkalinity was well correlated with salinity (Figure 4a) and those processes that influence salinity (e.g., ice melt or ice formation, evaporation or precipitation). Similar observations have been reported for other oceanic regions [e.g., Brewer et al., 1986; Bates et al., 1996b]. However, differences exist between the salinity-alkalinity regres- sions observed for 1994 and 1995-1996 (Figure 4a). These differ- ences are removed if alkalinity+nitrate +nitrite is plotted against salinity (Figure 4b). The inclusion of nitrate+nitrite accounts for the effect of the H+ flux associated with the biological uptake of nitrate +nitrite [Brewer and Goldman, 1976]. As a comparison, we have plotted alkalinity+nitrate +nitrite versus salinity from the

Ross Sea with stations from the Geochemical Ocean Sections

Study (GEOSECS) program in the Atlantic, Pacific, and Indian Oceans (Figures 4e and 4f). Differences between both data sets probably relate to differences in the calibration of acids used in the titration of seawater to determine alkalinity.

Alkalinity-TCO2 relationships can be used to determine whether the causes of variability of alkalinity and TCO 2 are due to processes such as photosynthesis/respiration or calcium carbonate production/dissolution. First, alkalinity and TCO 2 are normalized to a constant salinity to remove variability associated with salinity effects. Photosynthesis and respiration change TCO 2 but not alkalinity (see Figure 4d). Calcium car- bonate production/dissolution changes both TCO 2 and alka- linity in a ratio between 1:1 and 1:2 depending on the ratio of organic carbon production to calcium carbonate production [Robertson et al., 1994]. In Figure 4c, alkalinity variability is small (---10/xmol kg -•) compared to TCO 2 variability (---150 /xmol kg-•). This pattern indicates that photosynthesis/ respiration, rather than calcium carbonate production/ dissolution, controls TCO 2 variability. This agrees with previ- ous observations of a general lack of calcifying organisms in the Southern Ocean waters [Honjo, 1991] and biogenic silica gen- erally dominating the material collected in sediment traps or at the seafloor [e.g., DeMaster et al., 1991, 1996].

Stoichiometry of carbon, nitrogen, and phosphate was inves-

BATES ET AL.' CARBON CYCLING IN THE ROSS SEA POLYNYA 2887

ß Transect 1 and 2, 1994 a ,': ........ O ........ Transect 3, 1995/96

0 ......... :/d'"\ .•, ........ q) ........ Transect 4, 1995/96 •i•:::Q .'/ /:'"}'. •'"' ':': ........ ß ....... Transect 5, 1995/96

..' ..... .\ ,"?"/ ....... :::::::::::::::::::::::::: ........ ...'"' ...... ................... * ..... _ •." '%,/'/'/ 'o" "'"o"" .......... :• ........ l----e ..... e-----

'•i' / .... o o ..... .... o .... ""'o .... o ..... o., 100 km

165øE 170øE 1 øE 180 ø 175øW 170øW

.o ,...,•..•.,:....•....,,, .•....•.:::.• • * ::::...::•:,.• ...•. _ .q ,." .... •'"•"::•'::• ..... :..':• ..... • ...... • • ..... 0 ..• i • .•" " .'" G '• ':•----'• .......... "

0-'7? X ......

?' • Transact 1 and 2, 1994 •" I • ..... •.:/ ........ O ........ Transact 3, 1995/96 ........ • ........ Transact 4, 1995/96

'" .• ..... • ........ • ........ Transact 5, 1995/96 ',.

, , , , •'"' , , , , , , , • , , ] , , , , •0 ' ' I I .... 165øE 1 øE 175øE 180 ø 175øW 170øW

o ............. o @ ...... •....• .... 8'----.• .... .--. ........ .•' . . ', •-•':' '•' .... •":'"'•::::•:'::-•""•" 'O'::-,-o ..... ,-, ß "., .:•.:.:d "•¾: ' •.. ,. ,, .. ~.. ' "::::"t•:-::'-t ".. b'-:;? * Trans•t 1 •nd 2 •4 ...... ':::'"•

', .',' ' O ..... O' : ,."." ........ O ........ Transect 3, 1995/96

........ • ........ Transact 4, 1995/96

........ ß ........ Transect 5, 1995/96

0.5-

0-

-0.5 -

-1-

-1.5 -

-2

160øE 34.8

ß

34.6 ß

34.4 .

34.2 .

34 -

33.8 .

33.6 .

33.4

33.2

160øE

28.2

27.8 -

27.6

27.4 -

27.2 _

27 _

26.8

26.6

160øE

90 .

80 _

70

60- -

50 .

40 30

20 10

0

160OE

165OE

.......... Transect 1, 1994

.................... Transect 2, 1994

Transact 3, 1995/96

7 I I I 1 0øE 175øE 180 ø 175øW

.......... Transect 4, 1995/96 ß \

Transect 5, 1995/96 ',, ..;r '"/ \'/' ' ,,'..-\' ,, -, , ,:' "../',, '-- ,•.- 'x : . x, / ..... ..,.;.' [ ':'

•".,.,...•

170øW

165øE 170øE 175øE 180 ø 175øW 170øW

Figure 2. Surface properties along 76ø30'S in the Ross Sea observed during transects 1 (November 14-16) and 2 (December 2-6) conducted in 1994 and transects 3 (December 21-29, 1995), 4 (December 30, 1995, to January 8, 1996), and 5 (January 8-12, 1996) conducted in 1995-1996. (a) Temperature (degrees Celsius), (b) salinity, (c) sigma theta, and (d) mixed layer (meters). Mixed layer depths were based on a 0.1 density criterion.

tigated in the Ross Sea polynya (Figures 5b-5e, Table 3). Surface C:P and N:P ratios (-75-141:1 and -12-18:1) were highly variable but comparable with C:P and N:P ratios ob- served globally [Redfield et al., 1963; Takahashi et al., 1985; Minster and Boulahdid, 1987; Boulahdid and Minster, 1989; Anderson and Sarmiento, 1994] or in the Southern Ocean dur- ing GEOSECS sampling. Mean surface C:N ratios of 6.62:1 (and 6.67:1 for normalized TCO2) agree with C:N ratios pre- dicted by Redfield stoichiometry [Redfield et al., 1963]. Similar observations of C:N ratios in the Southern Ocean and the

Bellinghausen and Weddell Seas (data of T. Takahashi, re-

ported by U.S. JGOFS [1992]) suggest that the Redfield cycling of carbon and nitrogen is likely widespread across the South- ern Ocean and Antarctic ice marginal seas. Thus marked dif- ferences in carbon and nitrogen stoichiometry exist between environments like the Ross Sea polynya and other temperate and subpolar seas where C:N ratios are higher [e.g., Codispoti et al., 1986; Karl et al., 1991; Sambrotto et al., 1993]. It is not clear why this difference exists, but it may be associated with differential partitioning of organic carbon into particulate or- ganic carbon (POC) or dissolved organic carbon (DOC), which in turn may be controlled by phytoplankton community struc-

2888 BATES ET AL.: CARBON CYCLING IN THE ROSS SEA POLYNYA

• 26- • 24-

• 22-

• 20-

Z 18-

+ 16- 14-

0a 12-

ß -, 10 z

160øE 2250

2200-

2150

2100

2050

160øE

2300 _

2250

-

2200

2150

2100 160OE

380 _

360

340 ' -

320 300-

-

280 260 -

_

240 .

220 200

-

180 -

160

140

160øE

9,-4>-8

q• O • - Wransect 1, 1994 , O_•> I "O--0-O" -_<•:>- /•__• Transect 2, 1994 • ,-•x •x / O ', ,. • ,.'rx ,, ,,, ' '•' .B::--o.., ......... • •-e'- • O ø"• "' -•. '•".•.• ....... • • -'--

: • "'"':/' ............... ": • 'u'""gY .... O":7 ß "::•?'c•':':'-• ...y?,.<:e? o .... .{. . ß ........ ' ..... ::.. ........... .O.. ,:' :/ ............. .. .. : ,,? ,.

0 ............. <: ..... e. "0.... d' ,d ./" - ....... o ........ T:•e•t 3, 1995/96 -- • :'1•".. ', • .'•:• ,' (•. .;:: ..... "-.?¾ ..... :/-•... / ........ 'e ........ Tmn•½½t 4,, 1995196

"•1" ' I ' I '

165øE 170øE 175øE 180 ø 175øW 170øW

O

ß % o b x O

Transect 1, 1994 x • .' Transect 2, 1994 • - , r-, .....

, ..... •-----., ..... •:•-::::?'g5':•2-• -----•.. o.--o .............. o. ....... .v-",•..- - ............................... • ',., .......... :• ........ :,•.

,.,. :/// /" •, •,:: ..... ;x ...... • O Transect 3, 1995/96 ,.

,,xx .......... •77 'O ........ • ........ Transect 4, 1995/96 •" • ........ ß ........ Transect 5, 1995/96

I I I I I

165øE 170øE 175øE 180 ø 175øW 170øW

Transect 1, 1994 xx • • /• a•'' .-C•. ............ Transect 2, 1994 x_•..-• ..... QC"•'Z"Z' ..... • .......... ::•- ....... • .... O ..... .::5 .... •::::-•'5"•

....... O .... • .; '• ............. •::::2-- • / .......... :: .......... ......... . .... •? ...... e ..... •:::::-• ......

O ............. •. //--'" .,," ,,,," --. "-?::::• ........... . .............. O ........ Transect 3, 1995/96

........ , ..... , ,,

• • ...•s ........ •.. ........ • ........ Transect 4, 1995/96 ........ ß ........ Transect 5, 1995/96

I I I I I

165øE 170øE 175øE 180 ø 175øW 170øW

• Transect 1, 1994 - •- - Transect 2, 1994

165øE

d • o o

" 'ø'">?'::::'ø'::'"e ............. e•-7'"' ..... • ........... e' o.•"':::""'"ø"::::•-.-cs ...... .•:; ........ 9' •-•--.•:.-.:.:-4 ....... -"' ';•. - '• .... ...' ................. :.- ........ ..." L./; ............ :'o- ......... -'-,---.7--::•:----;•;: ...........

..-"'" /" ...'"" - ....... O ........ Transect 3, 1995/96 ........ ß ........ Transect 4, 1995/96 ........ ß ........ Transect 5, 1995/96

170øE 175øE 180 ø 175øW 170øW

Figure 3. Surface properties along 76ø30'S in the Ross Sea observed during transects 1 (November 14-16) and 2 (December 2-6) conducted in 1994 and transects 1 (December 21-29, 1995), 2 (December 30, 1995, to January 8, 1996), and 3 (January 8-12, 1996) conducted in 1995-1996. (a) Nitrate+nitrite (/xmol kg-•), (b) TCO 2 (/xmol kg-l), (c) nTCO 2 (/xmol kg-•; TCO 2 was normalized to a constant salinity of 35), and (d) seawater calculated pCO 2 (•atm).

ture and foodweb dynamics. In the Ross Sea polynya, Redfield C:N stoichiometry occurs in the Phaeocystis-dominated region. Carlson et al. [1997] reported that 89% of the total organic carbon was partitioned primarily into particulate organic car- bon (POC) with the remaining partitioned into dissolved or- ganic carbon (DOC). It may be that the partitioning of organic matter into DOM is greater in regions where elevated C:N stoichiometry is observed and differential recycling rates of C, N, and P from DOM may occur. It is interesting to note that C:P and N:P ratios were particularly low in the diatom- dominated westernmost part of the Ross Sea polynya where

partitioning of organic carbon into DOM may be greater when compared to Phaeocystis-dominated areas (C. A. Carlson and D. A. Hansell, unpublished data, 1996.)

3.3. Estimates of Net Community Production

Estimates of net community production (in the sense of Williams [1993]) have been made by observation of in situ water column changes in oxygen [e.g., Shulenberger and Reid, 1981; Emerson e! al., 1993], nitrate+nitrite [e.g.,Jennings e! al., 1984; Sambrotto e! al., 1986; Minas e! al., 1986; Codispoti e! al., 1991; Hansell e! al., 1993; Minas and Codispoti, 1993], TCO 2

2400

,•, 2380-

• 2360-

• 2340

• 2320

.• 2300

2280

33

2500

2450

2400

2350

2300

2375

2350

2325

2300

2275

2250

O

33.5

O

34 34.5

Salinity

:;so :;o :oo= normalized TCO2 (!.tmoles/kg)

o

o Q

o (

_+

•I 2400

• 2380

•, 2360

Z+ 2340

Z,• 2320

• 2300

• 2280 35 .•

¸ 2500 Z +

Z: 2450

4-

.• • 2400

,'• '• 2350

2300

2400

2375

2350

2325

• 2300

• 2275 35 • 33 33.5 34 34.5 33

Salinity

33 33.5 34 34.5 35

Salinity

Photo•ynthesis/respirati•on

on

normalized TCO2 (gmoles/kg)

•x

33.5 34 34.5 35

Salinity

Figure 4. Scatterplots of surface properties measured on transects along 76ø30'S and selected stations across the Ross Sea. Regression lines include an estimate of rms error for calculation of alkalinity (TA) from salinity. Properties are normalized to a constant salinity of 35 where appropriate. Diamonds denote 1994 data, and open circles denote 1995-1996 data. (a) Salinity is plotted against alkalinity (in •mol kg-•). Regression lines are as follows: November-December 1994 data, TA = 63.799 x S + 139.59 (_+4.41 •mol kg-•), r 2 -- 0.938; December 1995 to January 1996 data, TA = 60.057 x S + 279.15 (_+5.17 •mol kg-•), r 2 = 0.866. (b) Salinity is plotted against alkalinity plus nitrate+nitrite (in gmol kg-•). Regression lines are as follows: November- December 1994 data, TA + NO 3 + NO 2 -- 70.557 x S - 66.324 (_+3.79 •mol kg-•), r 2 = 0.914; December 1995 to January 1996 data, TA + NO 3 q- NO 2 -- 63.646 x S + 176.282 (_+2.93 •mol kg-•), r 2 = 0.966. (c) Normalized TCO 2 (nTCO2) is plotted against normalized alkalinity (nTA) (in •mol kg-•). (d) Normalized TCO 2 (nTCO2) is plotted against normalized alkalinity plus nitrate+nitrite (nTA + NO 3 q- NO2) (in •mol kg-•). Arrows denote the influence of photosynthesis/respiration and calcification on TCO 2 and alkalinity. (e) Salinity is plotted against alkalinity (in •mol kg -•) for Ross Sea and Southern Ocean (GEOSECS data). Regression lines for the Ross Sea data are the same as in Figures 4a and 4b. Regression lines for GEOSECS stations in the Southern Ocean are as follows: Atlantic Ocean (stations 78-89), TA = 68.238 x S - 24.466, r 2 : 0.912; Pacific Ocean (stations 278-292), TA = 73.653 x S - 211.056, r 2 = 0.817; Indian Ocean (stations 429-432), TA = 62.010 x S + 185.279, r 2 = 0.929. (f) Salinity is plotted against alkalinity plus nitrate+nitrite (•mol kg -•) for Ross Sea and Southern Ocean (GEOSECS data). Regression lines for the Ross Sea data are the same as in Figures 4a and 4b. Regression lines for GEOSECS stations in the Southern Ocean are aõ follows: Atlantic Ocean (stations 78-89), TA + NO 3 q- NO 2 -- 78.698 x S - 354.372, r 2 -- 0.961; Pacific Ocean (stations 278-292), TA + NO 3 q- NO 2 -- 83.087 x S - 505.421, r 2 = 0.835; Indian Ocean (stations 429-432), TA + NO 3 q- NO 2 -- 73.278 x S - 169.804, r 2 = 0.967.

2890 BATES ET AL.: CARBON CYCLING IN THE ROSS SEA POLYNYA

1.5

0.5

0

-0.5

-

-1

-1.5

2250

'•' 2200

• 2150

2100

2050

O O

Nov. - Dec. 1994

Dec. 1995- Jan. 1996¸ Winter source water ¸

O o o 0

O O O

a

33.5 3 34.5

Salinity

O Nov.- Dec. 1994 • •

¸ Dec. 1995- Jan. 1•96••

OO

O O

0 15 20 25 30 35

Nitrate + Nitrite (lamoles/kg)

35

• 30

25

z 20

4-

O Nov. - Dec. 1994 / O Dec. 1995-Jan. 1996 • / ß Wint s

0.5 1 1.5 2 2.5

Phosphate ([tmoles/kg)

2300

2250

2200

2150

2100

Nov. - Dec. 1994

Dec. 1995- Jan. 19960 O

Nitrate + Nitrite ([tmoles/kg)

2300 ] O Nov. - Dec. 1994 2250 O Nov.-Dec. 1994 <•/• '• t O Dec. 1995-Jan. 1996 ø I ' 1 2200 • 2250 O

o • 2200

o

• 2100 • 2150 2050 2100 •5 0.5 1 1.5 2 2.5 0.5 1 1. 2 2.5

Phosphate (gmoles/kg) Phosphate (gmoles/kg)

Figure 5. Scatterplots of surface properties measured on transects along 76ø30'S and selected stations across the Ross Sea. Regression lines include an estimate of rms error for calculation of TCO 2 from nitrate + nitrite. (a) Salinity is plotted against temperature (degrees Celcius). Solid square denotes winter source water chosen as the mean and standard deviation of salinity and temperature at 150 m from every station across the Ross Sea. (b) Phosphate (/xmol kg -l) is plotted against nitrate+nitrite (/xmol kg-•). Solid square denotes winter source water chosen as the mean and standard deviation of phosphate and nitrate+nitrite at 150 m from every station across the Ross Sea. Regression lines: November-December 1994 data, NO3 + NO2 = 18.427 x PO 4 - 9.176, /.2 = 0.971; December 1995 to January 1996 data, NO3 + NO2 = 12.264 x PO 4 + 2.048,/.2 = 0.803. (c) Nitrate+nitrite (/xmol kg -•) is plotted against TCO 2 (/xmol kg-•). Regression lines: November- December 1994 data, TCO 2 - 7.289 x NO3 + NO2 + 2002.938 (_+9.59/xmol kg-•), r 2 = 0.971; December 1995 to January 1996 data, TCO2 = 7.518 x NO3 + NO2 + 1986.950 (_+12.78/xmol kg-•). (d) Nitrate+nitrite (/xmol kg -•) is plotted against normalized TCO 2 (/xmol kg-•). TCO 2 is normalized to a constant salinity of 35. Regression lines: November-December 1994 data, nTCO2 = 6.662 x NO3 + NO2 + 2054.491 (_+7.24 /xmol kg-•), r 2 = 0.971; December 1995 to January 1996 data, nTCO 2 = 6.769 x NO3 + NO2 + 2050.250 (_+6.33 /xmoles kg-•). (e) Phosphate (/xmol kg -•) is plotted against TCO 2 (/xmol kg-•). Regression lines: November-December 1994 data, TCO 2 = 140.952 x PO 4 + 1926.694, /.2 ___ 0.805; December 1995 to January 1996 data, TCO 2 = 95.117 x PO 4 + 1999.047, r 2 = 0.732. (f) Phosphate (/xmol kg -•) is plotted against normalized TCO 2 (/xmol kg-•). TCO 2 is normalized to a constant salinity of 35. Regression lines: November-December 1994 data, nTCO 2 = 126.502 x PO 4 + 1985.463, r 2 = 0.816; December 1995 to January 1996 data, nTCO2- 75.304 x PO 4 + 2073.823,/.2: 0.653.

BATES ET AL.: CARBON CYCLING IN THE ROSS SEA POLYNYA 2891

Table 3. Stoichiometric Ratios Observed in the Ross Sea, Derived From Regressions Plotted in Figures 5c-5f

Southern Ocean

Elemental Ratio 1994 1995-1996 (GEOSECS)

C:N ratios

TCO2:NO3 + NO2 7.29:1 7.52:1 10.53:1 nTCO2: NO 3 + NO 2 6.66:1 6.77:1 7.29:1 C:P ratios

TCO2: PO 4 140.95:1 95.12:1 142.77:1 nTCO2: PO 4 126.50:1 75.30:1 97.24:1 N:P ratios

NO 3 + NO2: PO 4 18.43:1 12.26:1 13.60:1

Data from the following GEOSECS stations in the Southern Ocean (below 55øS) were used to calculate mean C:N:P stoichiometry in the upper 300 m: Atlantic Ocean, stations 78-89; Pacific Ocean, stations 278-292; Indian Ocean, stations 429-432.

[e.g., Weiss et at., 1979; Codispoti et at., 1982, 1986; Karl et at., 1991; Chipman et al., 1993; Yager et al., 1995], andpCO2 [Rob- ertson and Watson, 1993]. Here we use a mass balance ap- proach to calculate mean daily and total net community pro- duction (NCP) rates from measurements of TCO 2 and nitrate+nitrite on repeat transects (76ø30'S) across the Ross Sea polynya (Tables 4a and 4b). A mean surface TCO 2 for each transect was calculated by linear interpolation between stations. Since sparse sampling was conducted along each transect, additional surface TCO 2 data were calculated from nitrate+nitrite concentrations at each transect station (Figure 6). The error in estimating TCO 2 from nitrate+nitrite was _+5.6 /•mol kg -• (from regression of normalized TCO 2 and nitrate+nitrite in Figure 5d), relatively small considering the 120-200/xmol kg -• changes in TCO 2. Since the calculation of TCO 2 from nitrate+nitrite compared favorably with measured TCO2, we estimated the depth-integrated changes in TCO 2 at each station by integrating the nitrate+ nitrite deficit observed at each depth from surface to 150 rn (Figure 6). The mean nitrate+nitrite concentration at 150 rn from every station was chosen as the representative source water for initial wintertime conditions. Nutrient concentrations remained unchanged at this depth during all the sampling periods (November- January), and we assumed that the water column was well mixed in winter (a recent JGOFS cruise to the Ross Sea in October 1996 confirms this). At each depth, the nitrate+ nitrite deficit was calculated from the mean source water

nitrate+nitrite concentrations (i.e., 30.1 /xmol kg -•) minus observed nitrate+nitrite. A stoichiometric ratio of 6.67 ob-

served in the Ross Sea polynya (Figure 5d) was then used to convert the nitrate+nitrite deficit to depth-integrated loss of TCO 2. The total net community production in the polynya was estimated from the difference between wintertime

nitrate+nitrite concentrations and deficits of nitrate+nitrite

observed on each transect.

Several important considerations and assumptions have been taken into account. We assumed that changes in the concentrations of TCO 2 and nitrate+nitrite result primarily from net community production (NCP). We also took other processes such as vertical and horizontal transport, salinity changes, calcium carbonate production, and air-sea gas ex- change into consideration. We assumed that vertical entrain- ment of TCO 2 and nitrate+nitrite into the surface layer was not quantitatively important during the period between sam- pling. The transect data chosen (transects 1 and 2 (1994) and 3 and 4 (1995-1996)) represent periods when the mixed layer was generally shoaling or stable. Furthermore, our integration of the nitrate+nitrite deficit to 150 m accounts for any poten- tial vertical redistribution of TCO 2 and nitrate+nitrite due to mixing, despite a fluctuation of mixed layers from -20 to 60 m. It was more difficult to account for the effect of horizontal

transport of waters. We calculated a mean NCP across the polynya which should integrate the contribution of mesoscale variability and horizontal transport in the polynya. Since the period between repeat transects was short (19 and 8.5 days), we assumed that horizontal transport had limited influence on geochemical stocks of TCO 2 and nitrate+nitrite. Jaeger et al. [1996] reported mean flow rates of 1.8-4.1 cm s -• (1.5-3.5 km

Table 4a. Estimates of Mean Daily Net Community Production in the Ross Sea Polynya

Mean Integrated NCP Mean Time Period, Mean Surface NCP, Integrated Over 0-150 m,

Date Transect days mg C m -3 d -• g C m -2 d -•

Dec. 14-16, 1994 1 19 20.9 0.98 Dec. 2-5, 1994 2 Dec. 21-29, 1995 3 8.5 44.7 0.86 Dec. 30, 1995, to Jan. 8, 1996 4

Mean surface NCP was calculated from changes in normalized TCO 2 (nTCO2) between repeat transects along 76ø30'S. For each transect, a mean nTCO2 was calculated along the line by linear interpolation between stations. The nTCO2 data at each station were calculated from nTCO2:nitrate+nitrite relationships shown in Figure 5d. Mean depth-integrated NCP was calculated from changes in nitrate+nitrite between repeat transects along 76ø30'S. Integrated nitrate+nitrite deficit in the upper 0-150 m was estimated by subtracting observed nitrate+nitrite from a mean winter source water (150 m) nitrate+nitrite value (31.0 mol/kg). Integrated nitrate+nitrite deficits for each station are converted to integrated nTCO2 loss using nTCO2:nitrate+nitrite regressions observed in the Ross Sea (Figures 5c and 5d). Mean transect integrated nTCO2 loss was estimated by linear interpolation between stations. Mean integrated NCP were then calculated on a daily rate using an average time difference (in days) between repeat transects. NCP, net community production.

2892 BATES ET AL.: CARBON CYCLING IN THE ROSS SEA POLYNYA

Table 4b. Estimates of Total Net Community Production in the Ross Sea Polynya

Total NCP, Transect g C m -2

1 7.7

2 26.4

3 41.0

4 48.3

5 47.9

d -•) directed eastward in the region of 76ø30'S, but it should be noted that a low-temperature feature observed at -t67øE in 1995-1996 did not change position. Our estimates of total NCP implicitly account for any potential horizontal advection, since all waters had initial wintertime nutrient concentrations.

Salinity effects were also considered by normalizing TCO 2 to a constant salinity of 35 to remove the potential effects of ice melt/formation or evaporation/precipitation. We found little difference between NCP estimates calculated from TCO 2 or salinity-normalized TCO 2. The phytoplankton community was dominated by noncalcifying organisms such as P. antarctica and diatoms. No alkalinity changes characteristic of calcification were observed in the polynya (Figures 5c and 5d), and we assumed that calcium carbonate production had an insignifi- cant effect on CO2 species and estimates of NCP. Finally, we accounted for the potential contribution of air-sea gas ex- change of CO2 on TCO 2 changes. On transect t, we assumed that gas exchange of CO 2 was negligible because the presence of thin ice provided a barrier to gas exchange. On transects 2-5 we assumed that gas exchange of CO 2 occurred during open water conditions. Mean winds during the sampling period were -5-10 m s -•, and the average difference between atmospheric and surface pCO 2 (i.e., ApCO2) was about -tOO _+ 30 tzatm (Figure 3d). Gas exchange of CO 2 was directed from atmo- sphere to ocean, and we calculate that the flux of CO2 was probably in the range of 4 to 26 mmol CO2 m -2 d- • (see Table At for further details on the calculation). It should be noted that the range in values reflects uncertainties in CO 2 gas ex- change; we have therefore used two different gas exchange formulations [Liss and Merlivat, 1986; Wanninkhof, 1992] to provide estimates of probable minimum and maximum CO2 fluxes. Between transect 3 (1995-96) and transect 4 (1995-96) we estimated that gas exchange contributed to a mean increase in oceanic TCO 2 of -0.024-0.17 tzmol kg- • over the upper 150 m (Table A1). Our calculated NCP rates may be under- estimated by 5-25%, since gas exchange contributed additional CO2 to the upper water column.

The mean daily NCP rates for the Ross Sea were 0.98 and 0.86 g C m -2 d -• for both years, respectively (Table 4a and Figure 6). Integrated net primary production rates observed for both cruises varied from 0.5 to 6 g C m -2 d -• along the 76ø30'S line [Carlson et al., 1997]. Our geochemical estimates of NCP compared well with other observations of integrated primary production in the Ross Sea. Arrigo and McClain [1994] reported mean production rates of 0.99-3.9 g C m -2 d- • from coastal zone color scanner imagery of the Ross Sea polynya. Rates of •4C primary production observed along 76ø30'S in the polynya ranged from 0.79 to 1.3 g C m -2 d -• during visits to the polynya in 1990 and 1992 [Nelson et al., 1996; Smith et al., 1996]. Our estimates of total NCP ranged from 7 to 48 g C m -2, with a total NCP of -48 g C m -2 by mid-January. This

compares with estimates of yearly primary production ranging from 45.6-50 g C m -2 yr -1 to 200 g C m -2 yr -1 [Smith and Nelson, 1985, 1990; Arrigo and McClain, 1994; Smith and Gor- don, 1997].

Our NCP estimates approximated rates of new production in the polynya since CO2 was fixed with new nitrogen. How- ever, export production in the Ross Sea is a fraction of NCP rates since fixed CO2 is partitioned between suspended partic- ulate organic matter, dissolved organic matter, and sinking particles. Carlson et al. [1997] and D. A. Hansell and C. A. Carlson (unpublished data, 1996) estimated that during the November-December 1994 visit to the polynya, 35.4% of fixed CO2 was partitioned into suspended POC, 8.4% accumulated as DOC, and the remaining 56.2% was exported. During a 1992 visit to the polynya, Nelson et al. [1996] estimated that 42% of fixed carbon was available for export using lSN based measurements. Thus export production rates (i.e., flux of car- bon out of the euphotic zone) were probably close to 0.4-0.5 g Cm-2 d -1 '

Our method of estimating NCP revealed that there was no measured NCP between transects 4 and 5 (1995-1996) (Tables 4a and 4b). It seems likely that this observation results from a balance of production (net autotrophy) and respiration (net heterotrophy). The short sampling interval between transects suggests that mesoscale variability and horizontal transport have little impact on TCO 2 and nitrate + nitrite stocks, while any potential vertical redistribution is accounted for by depth integration to 150 m. Carlson et al. [1997] have suggested that the blooms of Phaeocystis in the Ross Sea were senescing by early January 1996. Arrigo and McClain [1994] noted the sea- sonal decline in pigment concentration began in mid-January to late January each austral summer. Carlson et al. [1997] reported a small DOC (and POC) accumulation in the upper water column between transect 3 and transect 4 but reported that this accumulation had disappeared by transect 5, with DOC (and POC) returning to levels observed on transect 3. Incubations of seawater sampled on the Ross Sea cruises sug- gest that accumulated DOC is rapidly mineralized by bacteria (C. A. Carlson et al., unpublished data, 1997). If suspended POC is not mineralized and sinks out at some point, it will contribute to export production already estimated for the polynya (0.5 g C m -2 d-•). However, the timing and magni- tude of this additional export production is unresolved at present.

3.4. Implications for Air-Sea Exchange of CO2 and Export Production

During our visits to the Ross Sea polynya, surface seawater pCO2 concentrations ranged from -140 to 250 gatm (Figure 3). In the Phaeocystis-dominated central and eastern parts of the polynya, pCO2 concentrations were -200-250 gatm. In the diatom-dominated westernmost part of the polynya, pCO2 concentrations were significantly lower at -140-160 gatm. Even lower pCO2 concentrations (<100 gatm) have been ob- served in the Ross Sea (T. Takahashi, personal communica- tion, 1996) and in the Weddell Sea (D. Karl, personal commu- nication, 1997). During the period of our observations in the Ross Sea polynya we estimate that air-sea flux of CO 2 was directed from atmosphere to ocean at probable rates in the range of 4-10 mmol CO2 m -2 d -• (Table A1). If we assume that open water conditions existed over the period of our observations (-60 days; mid-November to mid-January) and that the surface area of the polynya was -330,000 km 2 [Nelson

BATES ET AL.: CARBON CYCLING IN THE ROSS SEA POLYNYA 2893

2280

2260

2240

2220-

2200- 2180 -

2160 160øE

7000

........ ̧ ........ Transect 1, 1994

.... • .... Transect 2, 1994

165øE 170øE 175øE 180 ø 175øW

a

170øW

6O00

5000 4000

3000

2000

lOOO

o

160OE

2240

........ [] ........ Transect 1, 1994 - - ....... • ........ Transect 2, 1994

_

- ..... ß / ! .m"',/ "..., ,... /•.

- • • • ...... p ..... •. •8 • i• m • .• • //•, '• ..... R •, •. /'

- • •/ "• ..... • .... • • ............ •.•...• •' , ---• , • , ,

165øE 170øE 175øE 180 ø 175øW

b

170øW

2220

2200

2180 2160

2140

2120

7000

160øE

........ O ........ Transect 3, 1995/96

.... • .... Transect 4, 1995/96

ß Q

I .... I ' '

165øE 170øE 175øE 180 ø 175øW 170øW

6000

5000 4000

3000

:ooo 1000

[]

/ '...• ..... [].," EL.m, ,4,'" • " ..." ..... :•.

.......... '.,,,,, & /'.,/. .... "0'" :: '.. / '. ,' I'-1 : '., :.' •B / 0 .... [] '" "'

........ [] ........ Transect 3, 1995/96

.... • .... Transect 4, 1995/96

d

"'123 .... 0 ..'"' '", ..... 0 .... o" ':..,

0 ' I ' I ' 160øE 165øE 170øE 175øE 180 ø 175øW 170øW

Figure 6. Surface nTCO2 and depth-integrated NCP along 76ø30'S in the Ross Sea. Open symbols, nTCO2 calculated from stoichiometric relationships (Table 3); solid symbols, directly measured nTCO2. (a) Surface nTCO2 measured on transects 1 (November 14-16, 1994) and 2 (December 2-6, 1994). (b) Depth-integrated NCP on transects 1 (November 14-16, 1994) and 2 (December 2-6, 1994). (c) Surface nTCO2 measured on transects 1 (December 21-29, 1995-1996) and 2 (December 30 to January 8, 1995-1996). (d) Depth-integrated NCP on transects 1 (December 21-29, 1995-1996) and 2 (December 30 to January 8, 1995-1996).

et al., 1996], the flux of CO2 into the Ross Sea polynya ranged between 0.57 and 2.18 Tg C (10 TM g C). This is a relatively small quantity compared to global primary production and anthro- pogenic uptake of CO2 ("•10 •s gC). The small oceanic uptake of atmospheric CO2 in the Ross Sea may partially explain the limited anthropogenic CO2 contribution to deep Antarctic bot- tom waters formed in the Ross Sea and other Antarctic con-

tinental seas [Poisson and Chen, 1987]. Nelson et al. [1996] have estimated that export production

across the entire Ross Sea is --•20 Tg C at 50 m but that only 1.6 Tg C reaches sediment traps at 250 m deep. They suggested that sediment traps were accurate since no significant differ- ences were observed between net production values and opal fluxes at 250 m deep. The measured loss of carbon at 250 m (i.e., export production of 1.6 Tg C) was similar to our esti- mates of a gain of CO2 into the surface layer through gas exchange (--•0.95-3.8 Tg C). Carbon balance requires that the export production be balanced by input terms that include gas

2894 BATES ET AL.: CARBON CYCLING IN THE ROSS SEA POLYNYA

exchange, vertical redistribution of TCO2, and horizontal transport of water into the polynya. The relative importance of these processes remains to be resolved.

It is not clear what the net annual flux of CO2 in the polynya is. Our observations suggest that by mid-January, net hetero- trophic conditions exist in the polynya presumably because of the mineralization of suspended POM and accumulated DOM. Nutrients and TCO 2 return to wintertime concentrations by two likely processes: (1) deep mixing before sea ice covers the polynya (mid-March) and return of respired TCO 2 and nutri- ents to the surface during the winter months and (2) horizontal transport of surface waters. It is likely that after mid-January, surface seawater pCO 2 increased from --•140-250 /xatm to wintertime concentrations of --•340-360/xatm, close to atmo- spheric values. The presence of sea ice in the polynya between mid-March and mid-November does not automatically rule out gas exchange. Gas exchange can occur during ice coverage [e.g., Fanning and Totres, 1991], presumably through transient leads in the sea ice. However, our data indicate that ApCO2 values were probably low (_+0-20/xatm) during sea ice cover, and gas exchange rates were probably insignificant. If this is correct, then the Ross Sea polynya is a net sink for atmospheric CO2 primarily due because of intense biological production during open water conditions. This finding contrasts with the Southern Ocean, where solubility effects rather than biological production appear to be the primary driving force for seques- tering atmospheric CO2 [Siegenthaler and Sarmiento, 1992; Sar- miento and Le Qu•r•, 1996]. It is likely that solubility is the dominant control of oceanic uptake of CO2 across much of the Southern Ocean. However, in geographically limited areas where intense biological production occurs, such as the Ross Sea polynya and perhaps other ice marginal areas of the Ant- arctic, biological processes become important for transferring CO2 from the atmosphere to the ocean.

Table A1. Estimates of Air-Sea CO2 Gas Exchange in the Ross Sea Polynya

Parameter Value

Mean wind speed, m s -• Ap CO2,/xatm Daily CO2 flux, a mmol CO2 m -2 d-•

Liss and Merlivat [1986] Wanninkhof [1992]

ATCO 2 loss a Daily,/xmol kg d -• Total, mmol CO2 m -2

5

- 100

3.6

10.2

0.06-0.024

30.6-86.7

aTransects 1 and 2, 0-150 m.

k = 0.31u2(Sc/660) -ø's

where Sc is the Schmidt number for CO2. Sc was calculated using the equations of Wanninkhof [1992], while solubility of CO2, s, was calculated from the observed temperature and salinity using the equations of Weiss [1974].

A mean wind speed of 5 m s- • and an average ApCO2 of -100/xatm were used in the calculations (Table A1). A daily flux is estimated and converted to daily loss of TCO 2 (ATCO2) in the 0-150 m layer. The total loss of TCO 2 (ATCO2) in the 0-150 m layer is calculated over a mean period of 8.5 days between transects 1 and 2 (1995-1996).

Acknowledgments. We thank the crew of the RVIB Nathaniel B. Palmer. Three anonymous reviewers are thanked for their comments and suggestions for improvement of the paper. The authors acknowl- edge the support of NSF through grant OCE-9416565 (N.R.B.) and OPP-9317200 (D.A.H.). The U.S. Department of Energy CO2 Science Team is gratefully acknowledged for the use of a SOMMA CO2 ana- lyzer at BBSR. Andrew Dickson is thanked for the supply of certified reference material for seawater TCO 2 analysis. This is BBSR contri- bution 1473.

Appendix: Estimates of Air-Sea COz Gas Exchange

The net flux of CO2 was calculated from the gas exchange equation

F -- ks (Ap CO2) ( 1 )

where k is the transfer velocity, s is the solubility of CO2, and ApCO2 is the difference in partial pressure between atmo- sphere and seawater surface pCO 2. The transfer velocity and gas exchange coefficient are both functions of wind speed and several different gas exchange-wind speed relationships are commonly used [Liss and Merlivat, 1986; Wanninkhof, 1992]. We have used two relationships to provide upper and lower error estimates for the flux calculation. The lower limit is based

on the equations of Liss and Merlivat [1986], which were de- rived mainly from wind tunnel and lake experimental data:

k = 0.17u(600/Sc) 2/3 u -< 3.6 m s -• (2)

k = (2.85u - 9.65)(600/Sc) •/2 3.6<u-<13ms -•

(3)

k = 5.9u - 49.3(600/Sc) 1/2 u > 13 m s -• (4)

where u is wind speed at 10 m above mean sea level. The upper limit is based on the equations of Wanninkhof [1992], which were based on a quadratic dependency between wind speed and transfer velocity:

References

Anderson, L. A., and J. L. Sarmiento, Redfield ratios of remineraliza- tion determined by nutrient data analysis, Global Biogeochem. Cy- cles, 8, 65- 80, 1994.

Anderson, L. G., and E. P. Jones, The transport of CO2 into Arctic and Antarctic Sea: Similarities and differences in the driving processes, J. Mar. Syst., 2, 81-95, 1991.

Anderson, L. G., O. Holby, R. Lindegren, and M. Ohlson, The trans- port of anthropogenic carbon dioxide into the Weddell Sea, J. Geo- phys. Res., 96, 16,679-16,687, 1991.

Arrigo, K. R., and C. R. McClain, Spring phytoplankton production in the western Ross Sea, Science, 266, 261-263, 1994.

Bates, N. R., Determination of total inorganic carbon (CT) by the coulometric procedure, in BATS Methods Manual, Version 3, March 1993, edited by A. H. Knap, et al., pp. 31-37, U.S. Joint Global Ocean Flux Study Plann. Off., Woods Hole, Mass., 1993.

Bates, N. R., Investigation of the physical and biological controls of the oceanic CO2 system in the Sargasso Sea, Ph.D. dissertation, 276 pp., Univ. of Southampton, Southampton, England, 1995.

Bates, N. R., A. F. Michaels, and A. H. Knap, Alkalinity changes in the Sargasso Sea: Geochemical evidence of calcification?, Mar. Chem., 51,347-358, 1996a.

Bates, N. R., A. F. Michaels, and A. H. Knap, Seasonal and interan- nual variability of the oceanic carbon dioxide system at the U.S. JGOFS Bermuda Atlantic Time-Series Site, Deep Sea Res., Part II, 43(2-3), 347-383, 1996b.

Boulahdid, M., and J.-F. Minster, Oxygen consumption and nutrient regeneration ratios along isopycnal horizons in the Pacific Ocean, Mar. Chem., 26, 133-153, 1989.

Boyd, P. W., C. Robinson, G. Savidge, and P. J. L. Williams, Water column and sea-ice primary production during austral spring in the

BATES ET AL.: CARBON CYCLING IN THE ROSS SEA POLYNYA 2895

Bellingshausen Sea, Deep Sea Res., Part H, 42(4-5), 1177-1200, 1995.

Bradshaw, A. L., P. G. Brewer, D. K. Shafer, and R. T. Williams, Measurements of total carbon dioxide and alkalinity by potentio- metric titration in the GEOSECS program, Earth Planet. Sci. Lett., 55, 99-115, 1981.

Brewer, P. G., and J. Goldman, Alkalinity changes generated by phy- toplankton growth, Limnol. Oceanogr., 21, 108-117, 1976.

Brewer, P. G., A. L. Bradshaw, and R. T. Williams, Measurements of total carbon dioxide and alkalinity in the North Atlantic Ocean in 1981, in The Changing Carbon Cycle: A Global Analysis, edited by J. R. Trabalka, and D. E. Reichle, pp. 348-370, Springer-Verlag, New York, 1986.

Carlson, C. A., H. W. Ducklow, D. A. Hansell, and W. O. Smith, Jr., Organic carbon partitioning during spring phytoplankton blooms in the Ross Sea Polynya and the Sargasso Sea, Limnol. Oceanogr., in press, 1997.

Chipman, D. W., J. Marra, and T. Takahashi, Primary production at 47øN and 20øW in the North Atlantic Ocean: A comparison between the •4C incubation method and the mixed layer carbon budget, Deep Sea Res., Part A, 40, 151-169, 1993.

Codispoti, L. A., G. E. Friederich, R. L. Iverson, and D. W. Hood, Temporal changes in the inorganic carbon system of the southeast- ern Bering Sea during spring 1980, Nature, 296, 242-245, 1982.

Codispoti, L. A., G. E. Friederich, and D. W. Hood, Variability in the inorganic carbon system over the southeastern Bering Sea shelf during spring 1980 and spring-summer 1981, Cont. Shelf Res., 5, 133-160, 1986.

Codispoti, L. A., G. E. Friederich, C. M. Sakamoto, and L. I. Gordon, Nutrient cycling and primary production in the marine system of the Arctic and Antarctic, J. Mar. Syst., 2, 359-384, 1991.

Comiso, J. C., C. R. McClain, C. W. Sullivan, J.P. Ryan, and C. L. Leonard, Coastal zone color scanner pigment concentrations in the Southern Ocean and relationships to geophysical surface features, J. Geophys. Res., 98, 2419-2451, 1993.

DeMaster, D. J., D. M. Nelson, S. L. Harden, and C. A. Nittrouer, The cycling and accumulation of biogenic silica and organic carbon in Antarctic deep-sea and continental margin environments, Mar. Chem., 35, 489-502, 1991.

DeMaster, D. J., O. Ragueneau, and C. A. Nittrouer, Preservation efficiencies and accumulation rates for biogenic silica and organic C, N, and P in high-latitude sediments: The Ross Sea, J. Geophys. Res., 101, 18,501-18,518, 1996.

Dickson, A. G., The oceanic carbon dioxide system: Planning for quality data, U.S. JGOFS News, 2(2), pp. 2-10, U.S. JGOFS Plann. Off., Woods Hole, Mass., 1990.

Dickson, A. G., and C. Goyet, Handbook of methods for the analysis of the various parameters of the carbon dioxide system in seawater, version 2.0, CO2 Science Team report, U.S. Dep. of Energy, Wash- ington, D.C., 1994.

Edmond, J. M., High precision determination of titration alkalinity and total carbon dioxide content of seawater by potentiometric titration, Deep Sea Res. Oceanogr. Abstr., 17, 737-750, 1970.

E1-Sayed, S. Z., Productivity of the Southern Ocean: A closer look, Comp. Biochem. Physiol., B, Comp. Biochem., 90, 489-498, 1988.

E1-Sayed, S. Z., D.C. Biggs, and O. Holm-Hansen, Phytoplankton standing crop, primary productivity, and near-surface nitrogenous nutrient fields in the Ross Sea, Antarctica, Deep Sea Res., Part A, 30, 871-886, 1983.

Emerson, S., P. Quay, C. Stump, D. Wilbur, and R. Schudlich, Deter- mining primary production from the mesoscale oxygen field, ICES Mar. Sci. Symp., 197, 196-206, 1993.

Fanning, K. A., and L. M. Torres, 222Rn and 226Rn: Indicators of sea-ice effects on air-sea gas exchange, Polar Res., 10(1), 51-58, 1991.

Gordon, L. I., A. A. Ross, and J. Wilson, Chemical data and contoured sections for Ross Sea Bloom Dynamics Experiment Nathaniel B. Palmer legs NBP 94-6 and NBP 95-8, report, Coll. of Oceanic and Atmos. Sci., Oregon State Univ., Corvallis, 1996a.

Gordon, L. I., A. A. Ross, and J. Wilson, Nutrient data listings for Nathaniel B. Palmer legs NBP 95-8, December/January 1995/1996, Ross Sea Bloom Dynamics Experiment, report, Coll. of Oceanic and Atmos. Sci., Oregon State Univ., Corvallis, 1996b.

Goyet, C., and A. Poisson, New determination of carbonic acid disso- ciation constants in seawater as a function of temperature and sa- linity, Deep Sea Res., Part A, 36, 1635-1654, 1989.

Hansell, D. A., T. E. Whitledge, and J. J. Goering, Patterns of nitrate utilization and new production over the Bering-Chukchi shelf, Cont. Shelf Res., 13, 601-628, 1993.

Holm-Hansen, O., A. Neori, and I. Koike, Phytoplankton distribution, biomass, and activity in the southwest Ross Sea, Antarct. J. U.S., 27, 150-152, 1982.

Honjo, S., Particle fluxes and modern sedimentation in the polar oceans, in Polar Oceanography, Part B, edited by W. O. Smith Jr., pp. 687-739, Academic, San Diego, Calif., 1991.

Hoppema, M., E. Fahrbach, M. Schr6der, A. Wisotzki, and H. J. W. de Baar, Winter-summer differences of carbon dioxide and oxygen in the Weddell Sea surface layer, Mar. Chem., 51,177-192, 1995.

Jaeger, J. M., C. A. Nittrouer, D. J. DeMaster, C. Kelchner, and R. B. Dunbar, Lateral transport of settling particles in the Ross Sea and implications for the fate of biogenic material, J. Geophys. Res., 101, 18,479-18,488, 1996.

Jennings, J. C., L. I. Gordon, and D. M. Nelson, Nutrient depletion indicates high primary productivity in the Weddell Sea, Nature, 309, 51-54, 1984.

Johnson, K. M., Manual for the single-operator multiparameter met- abolic analyser (SOMMA) for total carbon dioxide (CT) with cou- lometric detection, report, Brookhaven Natl. Lab., Upton, New York, 1992.

Johnson, K. M., A. E. King, and J. M. Sieburth, Coulometric TCO 2 analyses for marine studies: An introduction, Mar. Chem., 16, 61-82, 1985.

Johnson, K. M., J. M. Sieburth, P. J. L. Williams, and L. Brandstrom, Coulometric total carbon dioxide analysis for marine studies: Auto- mation and calibration, Mar. Chem., 21, 117-133, 1987.

Johnson, K. M., K. D. Wills, D. B. Butler, W. K. Johnson, and C. S. Wong, Coulometric total carbon dioxide analysis for marine studies: Maximizing the performance of an automated gas extraction system and coulometric detector, Mar. Chem., 44, 167-188, 1993.

Karl, D. M., B. D. Tilbrook, and G. Tien, Seasonal coupling of organic matter production and particle flux in the western Bransfield Strait, Antarctica, Deep Sea Res., Part A, 38, 1097-1126, 1991.

Liss, P.S., and L. Merlivat, Air-sea gas exchange rates: Introduction and synthesis, in The Role of Air-Sea Exchange in Geochemical Cy- cling, edited by P. Buat-Menard, NA TO ASI Ser. C Math. Phys. Sci., 185, 113-128, 1986.

Minas, H. J., and L. A. Codispoti, Estimation of primary production by observation of changes in the mesoscale nitrate field, ICES Mar. Sci. Symp., 197, 215-235, 1993.

Minas, H. J., M. Minas, and T. T. Packard, Productivity in upwelling areas deduced from hydrographic and chemical fields, Limnol. Oceanogr., 31, 1182-1206, 1986.

Minster, J.-F., and M. Boulahdid, Redfield ratios along isopycnal sur- faces: A complementary study, Deep Sea Res., Part A, 34, 1981-2003, 1987.

Nelson, D. M., and W. O. Smith Jr., Phytoplankton bloom dynamics of the western Ross Sea ice edge, II, Mesoscale cycling of nitrogen and silicon, Deep Sea Res., Part A, 33, 1389-1412, 1986.

Nelson, D. M., J. A. Ahern, and L. J. Herlihy, Cycling of biogenic silica within the upper water column of the Ross Sea, Mar. Chem., 35, 461-476, 1991.

Nelson, D. M., D. J. DeMaster, R. B. Dunbar, and W. O. Smith Jr., Cycling of organic carbon and biogenic silica in the Southern Ocean: Estimates of water-column and sedimentary fluxes on the Ross Sea continental shelf, J. Geophys. Res., 101, 18,519-18,532, 1996.

Peng, T.-H., T. Takahashi, W. S. Broecker, and J. Olafsson, Seasonal variability of carbon dioxide, nutrients and oxygen in the northern North Atlantic surface water: Observations and a model, Tellus, Ser. B, 39, 439-458, 1987.

Poisson, A., and C. T.-A. Chen, Why is there little anthropogenic CO2 in Antarctic bottom water?, Deep Sea Res., Part A, 34, 1255-1275, 1987.

Poisson, A., N. Metzl, C. Brunet, B. Schauer, B. Brfis, D. Ruiz-Pifio, and F. Louanchi, Variability of sources and sinks of CO2 in the western Indian and Southern Oceans during the year 1991, J. Geo- phys. Res., 98, 22,759-22,778, 1993.

Redfield, A. C., B. H. Ketchum, and F. A. Richard, The influence of organisms on the composition of seawater, in The Sea, (edited by M. N. Hill), pp. 27-77, John Wiley, New York, 1963.

Robertson, J. E., and A. J. Watson, Estimation of primary production by observation of changes in the mesoscale carbon dioxide field, ICES Mar. Sci. Symp., 197, 207-214, 1993.

2896 BATES ET AL.: CARBON CYCLING IN THE ROSS SEA POLYNYA

Robertson, J. E., and A. J. Watson, A summer-time sink for atmo- spheric carbon dioxide in the Southern Ocean between 88øW and 88øE, Deep Sea Res., Part II, 42, 1081-1092, 1995.

Robertson, J. E., C. Robinson, D. R. Turner, P.M. Holligan, A. J. Watson, P. Boyd, E. Fernandez, and M. Finch, The impact of a coccolithophore bloom on oceanic carbon uptake in the northeast Atlantic during summer 1991, Deep Sea Res., Part I, 41, 297-314, 1994.

Roy, R. N., L. N. Roy, K. M. Vogel, C. Porter-Moore, T. Pearson, C. E. Good, F. J. Millero, and D. Campbell, Determination of the ioniza- tion constants of carbonic acid in seawater, Mar. Chem., 44, 249- 268, 1993.

Sakshaug, E., The physiological ecology of polar phytoplankton, in Proceedings of the Sixth Conference of the Comit• Artique Internatio- nale, edited by L. R. Rey and V. Alexander, pp. 61-89, E. J. Brill, Leiden, Netherlands, 1984.

Sambrotto, R. N., H. J. Niebauer, J. J. Goering, and R. L. Iverson, Relationship among vertical mixing, nitrate uptake, and phytoplank- ton growth during the spring bloom in the southeast Bering Sea middle shelf, Cont. ShelfRes., 5, 161-198, 1986.

Sambrotto, R. N., G. Savidge, C. Robinson, P. Boyd, T. Takahashi, D. M. Karl, C. Langdon, D. Chipman, J. Marra, and L. Codispoti, Elevated consumption of carbon relative to nitrogen in the surface ocean, Nature, 363, 248-251, 1993.

Sarmiento, J. L., and C. Le Qu6r6, Oceanic carbon dioxide uptake in a model century-scale global warming, Science, 274, 1346-1350, 1996.

Shulenberger, E., and J. L. Reid, The Pacific shallow oxygen maximum, deep chlorophyll maximum and primary productivity reconsidered, Deep Sea Res., 28, 901-919, 1981.

Siegenthaler, U., and J. L. Sarmiento, Atmospheric carbon dioxide and the ocean, Nature, 365, 119-125, 1992.

Smith, W. O., Jr., Phytoplankton biomass and productivity in the Ross Sea polynya, paper presented at International Symposium on Car- bon Fluxes and Dynamic Processes in the Southern Ocean: Present and Past, Sci. Comm. on Oceanic Res., Sci. Comm. on Antarct. Res., Brest, France, Aug. 28-31, 1995.

Smith, W. O., Jr., and L. I. Gordon, Hyperproductivity of the Ross Sea (Antarctica) polynya during austral spring, Geophys. Res. Lett., 24, 233-236, 1997.

Smith, W. O., Jr., and D. M. Nelson, Phytoplankton bloom produced by a receding ice-edge in the Ross Sea: Spatial coherence with the density field, Science, 227, 163-167, 1985.

Smith, W. O., Jr., and D. M. Nelson, Phytoplankton growth and new production in the Weddell Sea marginal ice zone in the austral spring and autumn, Limnol. Oceanogr., 35, 809-821, 1990.

Smith, W. O., Jr., D. M. Nelson, G. R. DiTullio, and A. R. Leventer, Temporal and spatial patterns in the Ross Sea: Phytoplankton bio- mass, elemental composition, productivity, and growth rates, J. Geo- phys. Res., 101, 18,455-18,465, 1996.

Sullivan, C. W., K. R. Arrigo, C. R. McClain, J. C. Comiso, and J. Firestone, Distributions of phytoplankton blooms in the Southern Ocean, Science, 262, 1832-1837, 1983.

Takahashi, T., and D. W. Chipman, Carbon dioxide partial pressure in surface waters of the Southern Ocean, Antarct. J. U.S., 17, 103-104, 1982.

Takahashi, T., W. S. Broecker, and S. Langer, Redfield ratio based on chemical data from isopycnal surfaces, J. Geophys. Res., 90, 6907- 6924, 1985.

Takahashi, T., J. Olafsson, J. G. Goddard, D. W. Chipman, and S.C. Sutherland, Seasonal variation of CO2 and nutrients in the high- latitude surface oceans: A comparative study, Global Biogeochem. Cycles, 7, 843-878, 1993.

Tans, P. P., I. Y. Fung, and T. Takahashi, Observational constraints on the global atmospheric CO2 budget, Science, 247, 1431-1438, 1990.

U.S. Joint Global Ocean Flux Study (JGOFS), U.S. Joint Global Ocean Flux Study: Southern Ocean process study, U.S. JGOFS Plann. Rep., 16, 114 pp., U.S. JGOFS Plann. Off., Woods Hole, Mass., 1992.

Wanninkhof, R., Relationship between wind speed and gas exchange over the ocean, J. Geophys. Res., 97, 7373-7382, 1992.

Weiss, R. F., Carbon dioxide in water and seawater: The solubility of a non-ideal gas, Mar. Chem., 2, 203-215, 1974.

Weiss, R. F., H. G. Ostlund, and H. Craig, Geochemical studies of the Weddell Sea, Deep Sea Res., 26, Part A, 1093-1120, 1979.

Williams, P. J. L., On the definition of plankton production terms, in Measurement of Primary Production From the Molecular to the Global Scale, edited by W. K. W. Li and S. Y. Maestrini, ICES Mar. Sci. Symp., 197, 9-19, Academic, San Diego, Calif., 1993.

Yager, P. L., D. W. R. Wallace, K. M. Johnson, W. O. Smith Jr., P. J. Minnett, and J. W. Deming, The Northeast Water Polynya as an atmospheric CO2 sink: A seasonal rectification hypothesis, J. Geo- phys. Res., 100, 4389-4398, 1995.

Zwally, H. J., J. C. Comiso, C. L. Parkinson, W. J. Campbell, F. D. Carsey, and P. Gloerson, Antarctic sea ice, 1973-1976: Satellite passive-microwave observations, NASA Spec. Publ., 459, 1-206, 1983.

N. R. Bates, C. A. Carlson, and D. A. Hansell, Bermuda Biological Station for Research, Ferry Reach, St. George's, Bermuda GE01. (e-mail: nick@bbsr.edu)

L. I. Gordon, College of Oceanic and Atmospheric Sciences, Ore- gon State University, 104 Ocean Administration Building, Corvallis, OR 97331-5503. (e-mail: lgordon@oce.orst.edu)

(Received December 22, 1996; revised July 1, 1997; accepted July 26, 1997.)