OceThe Officiala Maganzineog of the Oceanographyra Spocietyhy

CITATION Loose, B., L.A. Miller, S. Elliott, and T. Papakyriakou. 2011. Sea ice biogeochemistry and material transport across the frozen interface. 24(3):202–218, http://dx.doi.org/10.5670/ oceanog.2011.72.

COPYRIGHT This article has been published inOceanography , Volume 24, Number 3, a quarterly journal of The Oceanography Society. Copyright 2011 by The Oceanography Society. All rights reserved.

USAGE Permission is granted to copy this article for use in teaching and research. Republication, systematic reproduction, or collective redistribution of any portion of this article by photocopy machine, reposting, or other means is permitted only with the approval of The Oceanography Society. Send all correspondence to: [email protected] or The Oceanography Society, PO Box 1931, Rockville, MD 20849-1931, USA.

downloaded from www.tos.org/oceanography The Changing Arctic Ocean | Special Issue on the International Polar Year (2007–2009)

Sea Ice Biogeochemistry and Material Transport Across the Frozen Interface

By Brice Loose, Lisa A. Miller, Scott Elliott, and Tim Papakyriakou

Deployment of a surface flux chamber on the sea ice (Canadian Archipelago, spring 2010). The yellow box to the right contains the infrared analyzer. Credit: K. Swystsun

202 Oceanography | Vol.24, No.3 Abstract. The porous nature of sea ice not only provides a habitat for ice algae atmosphere, especially during winter but also opens a pathway for exchanges of organic matter, nutrients, and with (Untersteiner, 1964). The exploration the seawater below and the atmosphere above. These constituents permeate the ice of sea ice algal communities in brine cover through air-ice , brine drainage, seawater entrainment into the began in the 1960s; these communi- ice, and air-sea gas exchange within leads and polynyas. The central goal in sea ice ties held interest for polar biologists as biogeochemistry since the 1980s has been to discover the physical, biological, and both extremophiles and as potentially chemical rates and pathways by which sea ice affects the distribution and storage of important elements in the polar ocean biogenic gases (namely CO2, O2, and dimethyl sulfide) between the ocean and the food web (Arrigo et al., 2010; Caron atmosphere. Historically, sea ice held the fascination of scientists for its role in the and Gast, 2010). By 1965, it had been ocean heat budget, and the resulting view of sea ice as a barrier to heat and mass established that at least two micro- transport became its canonical representation. However, the recognition that sea algal communities existed in sea ice: ice contains a vibrant community of ice-tolerant organisms and strategic reserves (1) the epontic (under ice) community of carbon has brought forward a more nuanced view of the “barrier” as an active found principally in the skeletal layer participant in polar biogeochemical cycles. In this context, the organisms and their and bottom 20 cm of sea ice, and habitat of brine and salt crystals drive material fluxes into and out of the ice, regulated (2) the brown algal bands found below by and gas permeability. Today, scientists who study sea ice are acutely the ice freeboard level during the spring focused on determining the flux pathways of inorganic carbon, particulate organics, and summer (Burkholder and Mandelli, climate-active gases, excess carbonate alkalinity, and ultimately, the role of all of 1965). These organisms are able to thrive these constituents in the climate system. Thomas and Dieckmann (2010) recently in the ambient sea ice conditions in part reviewed sea ice biogeochemistry, and so we do not attempt a comprehensive review because of the stable light regime they here. Instead, our goal is to provide a historical perspective, along with some recent experience as compared to the mixed discoveries and observations to highlight the most outstanding questions and possibly layer. The two communities (epontic useful avenues for future research. and banded algae) have proven to be photo-adapted to very different light THE IMPORTANCE OF interest. Laboratory experiments have conditions, with the epontic variety SEA ICE IN GLOBAL been an essential substitute—but it is has exhibiting photoinhibition at 1 kilolux BIOGEOCHEMICAL CYCLES not been uncommon that phenomena (~ 14 µE m–2 s–1) and the algae in the Sea ice biogeochemistry is still a young revealed in the laboratory had to wait surface brown bands tolerating up discipline, with relatively few studies 30 years or more to be substantiated to 18 kilolux before photoinhibition dispersed through the last half century, by field data. (Burkholder and Mandelli, 1965). increasing in numbers during the last Since the 1960s, sea ice researchers The global abundance of the sea ice decade. We are only now progressing have been fascinated with brine in ice biome has been difficult to determine from the phase of exploratory discovery and its capacity to both alter the physical because of sampling difficulties coupled into that of defining and refining properties of ice and provide a vital with sea ice patchiness. Estimates of the details that are ultimately neces- ecological province. Early research was primary production in sea ice have sary to understand the controlling concerned with determining the effect been derived by estimating the standing processes. Biogeochemical cycles in of brine and sea ice algae on both the stock of ice algae (Legendre et al., sea ice occur on microscopic scales in mechanical and radiative properties of 1992), by 14C incubation (Burkholder a harsh, unstable environment that is sea ice (Weller, 1968). Principal among and Mandelli, 1965; Booth, 1984), and difficult to study without disturbing and these efforts was the determination of by immersion of microprobes fundamentally altering the processes of heat flux from the polar ocean to the for the measurement of net oxygen

Oceanography | September 2011 203 production in artificial sea ice (Mock Thomas, 2005; Deming, 2010). In fact, BIOGENIC GASES IN SEA ICE: et al., 2002, 2003). In a review of the bacterial in sea ice may A BRIEF HISTORY AND AMERIEZ (Antarctic Marine Ecosystem at times exceed primary production THE STATE OF THE ART Research at the Ice Edge Zone) project, (Deming, 2010). The first hints that chemistry Legendre et al. (1992) estimated that ice The strategic importance of sea ice might be interesting in sea ice came algal primary production was between to polar ecology and carbon cycling from laboratory thermodynamic studies 63 and 70 Tg C yr–1 for the Southern may be based on the role that it plays of salt precipitation that focused on Ocean seasonal ice zone (SIZ), defined as nutrient supply and as feed stock understanding the structural properties by these authors as south of 65°S). They for production in the benthos, in the of sea ice (Gitterman, 1937; Nelson and also cited similar estimates for ice algal water column, and along its retreating Thompson, 1954; Assur, 1958). These production in the Arctic (for waters edge in spring. In the 1980s, Carey studies found that CaCO3 is likely the north of 65°N). Although these results (1987) documented the steady particle first salt to precipitate from cooling ice for primary production in ice were flux from sea ice providing a small but brines. However, gas originally thought to be too large, they significant carbon source for benthic (N2, O2, Ar, and CO2) in natural sea ice have withstood time and bracket subse- dwellers. The importance of the ice edge were not reported until the early 1960s quent measurements and model esti- to the food web was being established; (Miyake and Matsuo, 1963; Matsuo and mates (e.g., Lizotte, 2001; Arrigo et al., stable density stratification and water- Miyake, 1966). Although those CO2 2010). Based on these values, primary column seeding of algae and nutrients by results may have been compromised (see production in ice algae would appear to melting ice were invoked as explanations section on “Measuring Gas Dynamics represent at most 5% of water-column for the observation of ice edge blooms in Sea Ice”), they generated ideas about primary production in the Antarctic (Smith and Nelson, 1985, 1986). By the carbon cycling in polar waters that are and 17% of primary production in the 1990s, tests of the “seeding hypothesis” still being tested. In the 1970s, a group in Arctic on an annual basis. With the indicated that ice algae can provide a Fairbanks, Alaska, reported concentra- development of cold- micro- starter community for phytoplankton tions of CO2, CO, CH4, and N2O in sea biological methods over the last two growth in the water column at least ice (Kelley and Gosink, 1979; Gosink, decades, we have come to understand under some conditions (Kuosa et al., 1980) and confirmed that gases can that, in contrast to the photosynthetic 1992). Beyond algae, melting sea ice may move through sea ice, at least at temper- community, the heterotrophic microbial serve as a source of organic matter and atures above –15°C (Gosink et al., 1976). community in sea ice is distributed micronutrients, such as iron (Lannuzel In 2002 and 2003, three groups throughout the ice column (Mock and et al., 2007; Thomas and Dieckmann, independently attempted to measure

2010). Concrete field evidence for CO2 fluxes in sea ice in situ: one in Brice Loose ([email protected]) is these processes are still few in number, Alaska (Semiletov et al., 2004), one Postdoctoral Fellow, Woods Hole and it will be a challenge to establish a in the Weddell Sea (Delille, 2006), Oceanographic Institution, Woods Hole, connective relationship between ice algal and one in the Canadian Archipelago MA, USA. Lisa A. Miller is Research colonies and the phytoplankton blooms (Papakyriakou and Miller, 2011). These Scientist, Institute of Ocean Sciences, that succeed them. However, advances in three studies, using different methods in Fisheries and Oceans Canada, Sidney, BC, rapid DNA identification and quantifica- different places (but all on first-year ice Canada. Scott Elliott is a scientist at Los tion, such as quantitative polymerase in spring), observed surprisingly large

Alamos National Laboratory, Los Alamos, chain reaction (QPCR) techniques, may CO2 fluxes that contradicted expecta- NM, USA. Tim Papakyriakou is Acting revolutionize the field through their tions and challenged assumptions about Director, Centre for Earth Observation ability to readily establish species abun- the role of sea ice in biogeochemical Science, Department of Environment dance and distribution and thereby chart cycles. Since then, additional data sets and Geography, University of Manitoba, the temporal and spatial evolution of using increasingly refined eddy covari- Winnipeg, Manitoba, Canada. these communities. ance methods to directly measure gas

204 Oceanography | Vol.24, No.3 fluxes have been published, confirming can also consume atmospheric CO2 as the solution chemistry in sea ice brines gas fluxes over multiyear ice (Zemmelink a result of ice algal blooms in spring may also lead to gas transport through et al., 2006), showing that dimethyl (Delille et al., 2007). CaCO3 precipitation, leading to changes sulfide (DMS) is also released from sea Considering surface waters overlain in pCO2 and associated fluxes. Early ice (Zemmelink et al., 2008), and finding by sea ice, Kelley (1987) concluded observations that CaCO3 precipitates in that CO2 fluxes through the autumn that the ice-covered seas may serve sea ice at relatively high (up

(Else et al., in press) and into winter as a net source of atmospheric CO2, to –2.2°C; Gitterman, 1937; Nelson and (Miller et al., 2011). Notwithstanding simply because sea ice is porous and Thompson, 1954; Assur, 1958), coupled these advances, the magnitude and the underlying surface waters are often with knowledge of the marine carbonate prevalence of these fluxes have yet to be supersaturated in CO2. Similarly, Bakker system (Marion, 2001), implied that concretely established as a widespread et al. (1997) observed CO2 supersatura- CO2 partial could be quite sea ice process. In parallel to these tion in the Weddell Gyre even as ice high in sea ice brines, particularly at studies of gas fluxes over sea ice, the cover retreated, implying a net flux of low temperatures, a speculation that has last 10 years have also been a period of CO2 to the atmosphere. In contrast, now been supported by measurements intense research into the geochemistry the “seasonal rectification (of CO2)” (Figure 1; Delille et al., 2007; Miller et al., of the ice itself and the role it may play hypothesis for the Arctic proposed 2011; recent work of authors Miller, in gas exchange. that respired CO2 remains trapped by Papakyriakou, and colleagues). sea ice until spring, when sea ice melt Lyakhin (1970) was the first to Carbon and stratification cause high primary discuss the geochemical implications

The idea that sea ice might participate production (Yager et al., 1995). In this of CaCO3 precipitation in sea ice as a in the carbon cycle dates to the work scenario, sea ice-covered regions, like possible explanation for the high CaCO3 of Miyake and Matsuo (1963), where the Arctic, are a net sink for atmospheric saturation states that he observed in they speculated that melting sea ice CO2. Several studies, including Gibson Okhotsk Sea waters. Jones and Coote would release CO2 to the atmosphere. and Trull (1999) and Sweeney (2003), (1981) next proposed that high-pCO2 ice

Their back-of-the-envelope calcula- reported a net sink for the Southern brines (produced by CaCO3 precipita- tions concluded that while CO2 release Ocean. However, all of these studies used tion) would effectively export inorganic from sea ice would be about an order of the wind-speed parameterization for carbon (TIC) into the ocean, possibly magnitude less than that from seawater, air-sea gas exchange in the sea ice zone, even contributing to the global it was still significant. Recent laboratory and there is evidence that this scaling pump, and others developed this idea studies have supported the hypothesis is not applicable (Loose et al., 2009). further (Alekseev and Nagurny, 2007; that sea ice is not only gas permeable Furthermore, gas exchange and primary Delille et al., 2007; Rysgaard et al., 2007; (Nomura et al., 2006; Loose et al., 2010) production are leading-order processes see Figure 2 for a schematic descrip- but it can also act as a direct CO2 source in determining net flux from the surface tion of the TIC pump). However, the to the atmosphere (Nomura et al., 2006). ocean. Collectively, these studies demon- hypothesis of the “TIC pump” has not However, the rate of diffusive fluxes strate that it is difficult to establish been consistently supported by field reported by Nomura et al. (2006) and the net effect of sea ice on the carbon data, which sometimes show only insig-

Loose et al. (2010) indicate that CO2 budget because the presence of sea ice nificant alkalinity fractionation in the fluxes in winter would only be significant inhibits gas exchange and supports ice (Anderson and Jones, 1985; Miller if pCO2 (CO2 partial ) in the primary production to varying degrees et al., 2011). Furthermore, the TIC ice were several orders of magnitude at different times of the year and over a pump requires certain physical forcing higher than that in the water. On the range of spatial scales. conditions that may not take place other hand, Nagurnyi (2008) found that In addition to the autotrophic and everywhere in the ice pack. Specifically, only deformed or broken ice released heterotrophic communities found in sea ice formation and convection must

CO2, and evidence exists that sea ice sea ice, which may influence gas fluxes, be strong enough to produce brine

Oceanography | September 2011 205 1.2 depending on time and place. During Arctic, Autumn 2007 (Miller et al., in press) freeze-up, strong downward fluxes in Arctic, Winter 2004 (Miller et al., 2011) 1.0 the presence of at least some remaining Antarctic, Spring 1999 (Delille et al., 2007) open water punctuated by sharp, brief

)

4 upward fluxes (Else et al., in press) 0.8 support both the Anderson et al. (2004)

hypothesis of enhanced CO2 drawdown 0.6 into open waters in which sea ice is

(ppmv x 10 2

CO beginning to form and the suggestion p 0.4 that CaCO3 precipitation as ice cools

increases pCO2 within the ice, which 0.2 can then outgas to the atmosphere. To

date, only one investigation of CO2 in 0.0 winter sea ice has been published (Miller –25 –20 –15 –10 –5 et al., 2011), and that study found that Temperature (˚C) very small CO2 fluxes (both upward Figure 1. Carbon dioxide (pCO2) values from sea ice, as determined to and downward) above the ice persisted date. The trend indicates thatp CO2 inside sea ice generally increases with decreasing throughout the coldest part of the year. temperature, which may be the result of CaCO3 precipitation in ice brines and/or reduced gas solubility at high salinities. The scatter may indicate the influence of other Because the ice was extremely cold processes, such as biological consumption, remineralization, and increased solubility with and likely impermeable (Golden et al., decreasing temperature. 1998), those deep-winter fluxes were probably exchanges with the snow and brine above the ice, and possibly the that is dense enough to sink beneath concerted research effort to build more top layer of the ice, but not with the ice the mixed layer and enter the abyssal conclusive evidence. interior. However, Miller et al. (2011) did system (Figure 2). This kind Anderson et al. (2004) proposed a find evidence that ice brines contribute of deep convection typically occurs in very interesting alternative mechanism inorganic carbon to the water under the polynyas (Orsi et al., 1999), but is not for ice-facilitated CO2 export that does ice. As spring approached, Miller et al. necessarily descriptive of sea ice forma- not invoke CaCO3 precipitation. They (2011) saw upward fluxes, increasing in tion over the entire sea ice zone where suggested that the process of frazil magnitude as the temperature increased. sea ice and the mixed layer depth are ice formation (first stage of “new” ice Other springtime studies also saw coupled (Martinson, 1990). Proof of when slender spikes or plates of ice upward fluxes at the beginning of their the TIC pump has also been substan- are suspended in turbulent water) at deployments, but the fluxes very quickly tially hampered by the fact that CaCO3 the ocean surface hydrodynamically switched, continuing strongly downward precipitation had not been confirmed in enhances air-sea gas exchange, allowing into the ice, as primary production and natural sea ice. That is, CaCO3 had only the cold surface waters to equilibrate melt (and therefore, possibly CaCO3 been recovered from sea ice in laborato- with the atmosphere (or nearly so) dissolution) advanced (Delille, 2006; ries. Dieckmann et al.’s (2008) successful before the ice completely covers the sea Nomura et al., 2010a; Papakyriakou isolation and identification of CaCO3 surface. Thus, the polar mixed layer and Miller, 2011). precipitates from Antarctic sea ice is could become saturated with CO2 before an intriguing development in the sea contributing to any deep convection. Other Gases ice TIC pump hypothesis, but it is only In summary, recent work has indi- After CO2, DMS is the most extensively one component of a complicated physi- cated that all these conjectured sea-ice studied gas in sea ice (Thomas and cochemical process, which requires a CO2 transport mechanisms may be valid, Dieckmann, 2010). An important link in

206 Oceanography | Vol.24, No.3 the global sulfur cycle (Charlson et al., a correction is made to account for quantitatively drawing CO2 out of the 1987; Ayers and Cainey, 2007), DMS is the gas in brine. This approach works solution and giving unrealistically high a microbial decay product of dimethyl- well for nonbuffered gases such as 2N pCO2 estimates in the ice. Therefore, sulfonioproprionate (DMSP), which is and O2, but it is problematic for CO2. further technique development has produced by a number of algal species. Exposing a sea ice sample (or seawater) focused on crushing the ice in smaller Thus, DMS is one of the primary vectors to a vacuum disrupts the equilibria volumes of head space so that there is no

(after sea salt aerosols) for transporting between the gas (CO2) and its hydrolysis need to draw a vacuum in order to get – 2– sulfur from the ocean to the atmosphere, products (H2CO3, HCO 3, and CO 3 ), an accurate estimate of the gas mixing where sulfur plays a significant role in cloud condensation and climate regula- tion. A number of studies have reported A. TIC PUMP SCENARIO high DMS and DMSP concentrations Onshore O­shore in sea ice (e.g., Trevena et al., 2000, Wind 2003; Trevena and Jones, 2006) and Fall Winter Spring Summer significant DMS fluxes out of sea ice (Zemmelink et al., 2008). Other gases of biogeochemical interest that are associated with sea ice include ALK trapped ALK dissolves the halogens (mainly alkyl bromides), DIC in brine in ice in ML which strongly influence polar atmo- spheric chemistry (Simpson et al., 2007), and methane, an ultrastrong “green- Continental shelf DIC sinks into abyssal water house” gas that is produced in anoxic sediments and is often found in bubbles trapped under sea ice (Kvenvolden et al., 1993; Shakhova et al., 2009). However, as yet, little work has been done on the behavior of these gases within sea ice. B. ZERO SUM CHANGE SCENARIO

MEASURING GAS Fall Winter Spring Summer DYNAMICS IN SEA ICE Matsuo and Miyake’s (1966) measure- ments of gas concentrations in sea ice were based on collecting and analyzing ALK trapped DIC in brine the gas released from ice through in ice ALK and DIC successive melting and refreezing cycles Brine detrains in ML mix in ML under a vacuum. Adapted from the field of glacial ice geochemistry, the first “modern” studies of gases in sea Figure 2. Schematic of the total inorganic carbon (TIC) pump. In panel A, the scenario depicts abyssal water formation in polynyas and the offshore transport of sea ice so that freeze and melt occur in ice generally involved crushing the different regions of the ocean. Under such a scenario, if carbonate alkalinity (ALK) remains trapped in ice in a vacuum and collecting the gas the ice as salt crystals and dissolved inorganic carbon (DIC) drains with the brine, then a “TIC pump” released for analysis (Tison et al., 2002). might result in a net transport of inorganic carbon to the deep ocean. Panel B depicts a scenario where ice forms and melts over a water column with a homogeneous mixed layer (ML). In this case, brine This approach extracts gas from the gas from sea ice is likely to detrain in the mixed layer and subsequently recombine with carbonate alka- pockets and from the brine, and then linity in the ice. This scenario depicts a zero-sum change in the mixed layer inorganic carbon budget.

Oceanography | September 2011 207 b

a

Figure 3. In situ pCO2 peeper array for vertical profile measurements (Amundsen Gulf, December 2007). (a) Just before deployment into the hole at the right, and (b) after deployment, with sampling ports available for connection to an infrared detector. Credit: N. Sutherland

ratios within the ice. In contrast, recent soil sciences (Holter, 1990; Jacinthe and Nomura et al., 2009, 2010a; Miller et al., measurements of DMS, a relatively Dick, 1996; Kammann et al., 2001), 2011). Ice cores are collected, sectioned, nonbuffered gas, in sea ice have been and are simply gas-permeable silicone and melted, then the melts are analyzed relatively straightforward, involving chambers that are frozen into the ice. for TIC and AT (Rysgaard et al., 2007; melting the ice in a closed vessel that is Narrow-bore tubing leads from the Miller et al., 2011). Those values are then purged with N2 to extract the DMS peepers to the surface and is connected then corrected to brine salinity from from solution (Trevena and Jones, 2006). to a portable gas analyzer (Figure 3). To the in situ temperature (Petrich and

Destructive methods, requiring date, this approach has only been used Eicken, 2010), and pCO2 is calculated removal and melting or crushing of ice for CO2 measurements, although results from those extrapolated “brine” TIC cores, are not ideal when attempting to from soils studies have indicated that and AT concentrations. This approach study a natural system, and researchers peepers should also be appropriate for requires the assumption that all the have had some success freezing in situ other gases in sea ice (Jacinthe and Dick, TIC and AT measured in the bulk ice probes into the ice to measure gases. 1996; Kammann et al., 2001). When melt initially resided in the brines and

Most notably, oxygen-sensing optodes measuring CO2 pressures in sea ice using that the carbonate equilibrium equa- have generated some very intriguing peepers, they must be connected to the tions for seawater are applicable to sea data, although their field deployment gas analyzer (generally a nondispersive ice brines, which may not be the case has been limited and not much has been infrared detector) in a closed loop so that due to ion-ion interactions that take published to date (Mock et al., 2002, the gas is recycled back into the peeper, place in such concentrated 2003; Rysgaard et al., 2008). The one assuring that a vacuum is not drawn on (Marion, 2001). Collecting and analyzing drawback to measuring oxygen in sea the ice (see Figure 1 for a summary of the brines allows pH to be analyzed, ice using optodes is that their response pCO2 in sea ice analyzed using peepers). as well as TIC and AT, but collecting time is quite slow. Slow response times Finally, many of the scientists working a sufficient quantity of brine for such are also a problem for in situ CO2 probes on the problem of CO2 cycling in sea analyses requires allowing brines to that have been used in sea ice, which can ice are oceanographers and, therefore, drain either into sackholes drilled into take about 24 hours to equilibrate with have naturally attempted to calculate the ice (Delille et al., 2007) or from their surroundings at subzero tempera- pCO2 from total inorganic carbon (TIC), cores extracted from the ice (Nomura tures. These probes, dubbed “peepers,” total alkalinity (AT), and/or pH in bulk et al., 2009). The brine drainage process were adapted from methods used in sea ice or brines (Delille et al., 2007; can take some time, particularly at low

208 Oceanography | Vol.24, No.3 temperatures, during which it is difficult on less-destructive measurements meteorological tower-based technique to limit CO2 exchange with the atmo- employing some variation on either eddy that relates a flux (e.g., heat, trace sphere, which affects both the TIC and covariance or bell-jar chambers. gas) to the time-averaged covariance pH measurements, and again, pCO2 is Enclosures (or chambers) act to between high-frequency fluctuations of underestimated. This problem of gas control the volume of air available for vertical wind velocity and a scalar exchange during brine collection also exchange across a covered surface so (e.g., temperature or gas ; applies to the measurement of any gas, that emission or uptake can be measured see Figure 4; Bakan, 1978; Edson et al., including O2 and DMS (Delille et al., as a concentration change (see photo 1998). Typically, measurements are made 2007), although Papadimitriou et al. of surface flux chamber on article title with fast-response sensors at a frequency (2007) calculated that when sufficient page). The validity of the flux measure- between 10 Hz and 20 Hz over a period brine can be collected quickly (i.e., in ment depends on the extent to which of approximately 30 to 60 minutes. Like about 10 minutes), the sheltered condi- the enclosure impacts gas transport, the enclosure method, the EC tech- tions and low temperatures of a capped production, or consumption in the snow nique also has limitations, being most sackhole limit gas exchange, and the or upper sea ice. Further, enclosures applicable over flat and homogeneous analyses should not be significantly sample over only a very small area, surfaces under steady or slowly changing impacted. Regardless of whether data are which substantially limits the utility of atmospheric conditions (Baldochi, from melted ice cores or brines, calcu- the data in a heterogeneous environ- 2003). A suite of corrections is also lating pCO2 from the other parameters ment. Enclosures modify the surface’s required to account for, among other is also fundamentally dubious for sea ice thermal regime by affecting energy things, the separation distance between because the available stability constants exchange via radiation and convection, the measurement of vertical wind have only been defined for conditions possibly affecting carbon equilibrium velocity and the scalar, and to compen- of low salinity and high temperature relations and restricting measurement sate for our inability to sample all signifi- (i.e., seawater). The one published study to a diffusive flux. Other problems relate cant scales of turbulence contributing to that collected the data necessary to over- to pressure changes in the enclosure the turbulent transfer of CO2. Further determine the system found that calcu- associated with wind fluctuations and details on the technique are available lated and measured pCO2 in ice brines modified gas concentration in the closed from numerous sources (e.g., Massman agreed well, at least in spring ice (Delille head space, which impact the surface- and Lee, 2002; Baldochi, 2003). et al., 2007), indicating that the constants to-chamber concentration gradient The technique for trace gas studies can be applied at least a bit beyond their and hence the diffusive flux (Nomura over sea ice has thus far been limited to empirical range. et al., 2010b). Benefits of the technique fluxes of H2O and CO2 (Semiletov et al., are that enclosures are inexpensive and 2004; Miller et al., 2011; Papakyriakou The Ins and Outs of relatively simple to operate, allowing for and Miller, 2011), and these studies use Flux Measurements frequent deployments in numerous loca- fast-response nondispersive infrared To date, three different methods have tions. Results provide, if not an accurate analyzers (NDIR) of open-type (OP) been employed to measure gas flux in measure of the flux, a feature-driven architecture, meaning the sensor’s and out of sea ice. The first to make indication of relative uptake and emis- optical path is in the free atmosphere. such measurements were Gosink sion activity. Valuable insight on the CO2 Alternatively, measurements can be et al. (1976), who charged boreholes source-sink character of sea ice has been made within a confined cell using a with gases, sealed the boreholes, then acquired through the use of enclosures closed-path NDIR (CP). The OP system measured the gas concentrations in cores in the Arctic (Semiletov et al., 2004; has the advantages of the optical path taken from the surrounding ice over Nomura et al., 2006, 2010b; Miller et al., being very near to the vertical wind the subsequent hours. This approach in press) and in the Antarctic (Delille, velocity measurement, sampling air with generated extremely interesting results, 2006; Delille et al., 2007). minimal aerodynamic disturbances, and but most efforts since have focused Eddy covariance (EC) is a micro- of having modest power requirements.

Oceanography | September 2011 209 The CP NDIR, on the other hand, is incorporated in EC systems for DMS SEA ICE MICROSTRUCTURE: typically located several meters from over sea ice; rather, existing measure- A TRANSPORT PATHWAY AND the vertical wind velocity measurement, ments of DMS emission are based on a A HOME AWAY FROM HOME requiring that air be drawn through a modification of the EC method, in which One of the remarkable capacities of tube to the sensor using a high-capacity the average gas concentration over the sympagic (ice-associated) organisms is pump (McGillis et al., 2001). Cold- period required for the detector response to maintain viability inside the porous region measurements of the CO2 flux is compared to the average air move- microstructure of the ice throughout using a CP EC system do not contain ment over the same period (relaxed eddy the winter and subsequently to multiply biases that have been recently reported accumulation; Zemmelink et al., 2008). in spring. These organisms inhabit for open-path EC systems (Amiro, In many ways, sea ice is an ideal environ- the interstitial spaces (brine channels 2010; Prytherch et al., 2010). Although ment for EC measurements because the and gas pockets) within the ice crystal corrections are available (Webb et al., surface can be relatively smooth and structure (Burkholder and Mandelli, 1980; Burba et al., 2008) to OP data uniform (at least for first-year ice) and 1965; Arrigo et al., 2010; Deming, 2010). sets, and have been used in CO2 flux the platform is stable. On the other hand, Diatoms comprise an important part of studies over sea ice (Papakyriakou and sea ice is an extremely variable envi- the ice algal community and appear to Miller, 2011), they can be large in rela- ronment on the mesoscale, with cores have the capacity not only to adapt from tion to the observed fluxes, and some collected only a couple of meters apart surface ocean to interstitial ice condi- are empirically based (Burba et al., 2008) often displaying quite different charac- tions but also to modify the physical and untested over polar snow and sea teristics (Miller et al., 2011) that result in structure of the ice to suit their needs ice surfaces. Further work is required to spatial variations at a scale not resolved (Krembs et al., 2002, 2011). Here, we confirm the results of previous studies. by eddy covariance measurements. discuss what is known about the rates of Fast-response detectors have not been material fluxes through the ice micro- structure and their mechanisms. The porous microstructure of sea ice is the central factor controlling biogeo- chemical cycling and fluxes in the ice. The fluid phase of sea ice is composed of brine inclusions that range from 0.1 to 10 mm in length (Light et al., 2003) and air inclusions spanning sizes of 0.01 to 1 mm in first-year sea ice (Light et al., 2003), to tens of millimeters in multiyear hummock ice (Perovich and Gow, 1996). Porosity, defined as the volume of the voids divided by the total volume of ice, can range from 2 to 40% (Perovich and Gow, 1996). Much of this internal volume is filled with brine, but in first-year ice, some 1 to 5% is filled with gas, and the gas-filled voids are known to increase significantly in multiyear, as well as springtime, ice. The Figure 4. Eddy covariance system for measurements of CO fluxes deployed over the sea ice (Franklin 2 connectivity of these individual inclu- Bay, winter 2004). The sonic anemometer and open-path CO2 detector are at the top of the picture. Credit: T. Papakyriakou sions is difficult to measure, although

210 Oceanography | Vol.24, No.3 computed tomography scans and convects away from the freezing interface sea ice where porosity was low, especially network algorithms have made some and into the water column. However, new near the ice surface, as a result of cold progress in quantifying this connectivity studies using the “mushy layer” theory temperatures. To date, the rate of gas (Golden et al., 2007). Pore connec- indicate that brine drainage requires a in warm, springtime ice has tivity is certainly the controlling factor critical ice thickness and that drainage not been documented in the field or the determining transport within the sea by gravity alone is the primary process laboratory. The progressive freshening ice microstructure; however, transport leading to bulk desalination (Feltham and drainage of springtime sea ice will is usually related to porosity as that is et al., 2006; Notz and Worster, 2009). likely yield much larger gas diffusion easier to quantify. In comparison to the permeation rates, although gas transport by brine Brine drainage from sea ice is by far of brine, there has been relatively little convection will complicate this picture. the best-studied result of solute segrega- consideration given to the process of tion and internal sea ice microstructure gas permeation through sea ice, but INSIGHTS FROM THEORY (Cox and Weeks, 1988; Golden et al., it is conceptually similar to gas diffu- AND EXPERIMENTATION 1998, 2007; Freitag and Eicken, 2003; sion through the unsaturated (vadose) Sea Ice and the Mushy Layer Oertling and Watts, 2004). Brine zone in soils. From those empirical To determine the physical properties of drainage plays a role in forming deep- relationships, we can infer that sea ice sea ice, all models, including those based water masses throughout the global permeability would range from about on first physical principles, require, at ocean and establishing the high salinities 10–7 to 10–4 cm2 s–1 in first-year sea minimum, profiles of either temperature required for deep convection (Steele and ice with 1 to 5% gas porosity, based or salinity through the ice (Feltham et al., Morison, 1995; Orsi et al., 1999; Tamura on the Penman-Millington Quirk soil 2006). Beyond this constraint, models et al., 2008). As discussed above, several gas diffusion model (Moldrup et al., based on sea ice thermodynamics hypotheses exist for sea ice enhance- 2004; Kawamoto et al., 2006). The require nothing else. The greatest recent ment of the CO2 solubility pump in few laboratory and field estimates theoretical advance in sea ice simula- these deep- and intermediate-water that have been made of gas diffusion tions has come from the recognition formation regions. through sea ice fall within this range, that sea ice can be treated as a “mushy The liquid permeability of sea ice is exhibiting large differences in magni- layer” (Worster, 1991). This constitutive well established and is similar to fine- tude, depending on the specific gas. formulation describes the evolution of a –13 2 grain sand (on the order of 10 m ; For CO2, gas diffusivity ranges from pure crystalline phase bathed within its Golden et al., 2007). Below 5% porosity, 18 to 90 x 10–5 cm2 s–1 (Gosink et al., impure fluid. The advantage of mushy brine drainage is thought to virtually 1976; Nomura et al., 2006). Diffusion layer theory is that it presents the most cease (Cox and Weeks, 1988). Golden ranges from 3 to 6 x 10–5 cm2 s–1 elegant, and in some cases analytical, et al. (2006) established a theoretical for O2 (Loose et al., 2010) and from solutions to the equations describing –5 2 –1 upper bound for liquid permeability 0.4 to 20 x 10 cm s for SF6 (Gosink the physical evolution of sea ice and, that reproduces this behavior using et al., 1976; Loose et al., 2010). Much of therefore, represents a principal tool in cylindrical capillaries whose internal this variability can probably be attributed describing the behavior of growing sea radii are representative of brine porosity. to differences in sea ice porosity, and/ ice and associated brine drainage. However, some debate exists on precisely or gas permeation through sea ice may There are at least two important what stage of ice growth leads to brine be fundamentally different from liquid implications of mushy layer theory for drainage and what processes are respon- permeation. However, we lack sufficient biogeochemical processes. First, desalina- sible for transporting brine out of the research to link porosity and gas diffu- tion of sea ice does not occur by solute ice. Solute segregation during ice growth sivity in sea ice in a quantitative fashion. segregation at the freezing interface leads to solute accumulation at the It is important to point out that (Notz and Worster, 2009). Instead, freezing interface. The canonical view all studies of gas diffusion have been gravity-driven brine drainage is the has been that this brine subsequently conducted in cold, columnar first-year principal drainage mechanism in young

Oceanography | September 2011 211 sea ice, and this process begins only after understanding of the strength and timing PROGRESS TOWARD SEA ICE the ice has reached a certain thickness of water column convection beneath BIOGEOCHEMISTRY IN MODELS (Notz and Worster, 2009). This outcome sea ice, the processes that take place in OF THE POLAR OCEAN is supported by the results of Wettlaufer polynyas, and even the carbonate compo- Climate system models1 to date treat sea et al. (1997), but it is seemingly at odds sition of brine as it drains. ice in a unimodal fashion; heat, radia- with experimental evidence that has been The second important implication tion, and gas fluxes are all influenced by used to construct empirical equations for from mushy layer theory is that gravity- sea ice cover, but the biogeochemistry of sea ice growth, including those of Cox driven brine drainage is accompanied by ice itself is only beginning to take shape and Weeks (1983, 1988), as well as the a return flow of seawater that penetrates within “system” codes. A Web of Science results of Loose et al. (2009), who show into the ice matrix to occupy the volume search of the terms “carbon dioxide, sea a nearly instantaneous response in both recently vacated by the brine. This ice, model” will yield dozens of results oxygen and salt during the first stages seawater pumping mechanism might be related to climate simulations that focus of ice crystal nucleation. However, the an important source of nutrients to the on the loss of ice coverage. Carbonate apparent discrepancies may be a function algal community living near the bottom chemistry within the sea ice matrix, of the water conditions beneath the ice of the ice. Indeed, the high concentra- however, is difficult to find. during formation. In mushy layer theory, tions of iron observed in sea ice (an the solute boundary layer governs the order of magnitude higher than in the Representing Organic and transition region from ice crystal lattice underlying water) have been explained Inorganic Gas Production in to pure liquid, and transport is assumed as resulting from effective “filtering” of System Models to proceed at the rate of molecular diffu- particle-reactive iron from seawater onto The cycling of climate-active gases is sion, therefore permitting a continuous the solid ice surfaces, as large volumes distinct within different sea ice types, salinity gradient from the water into the of surface seawater are flushed through and, in turn, sea ice types are a function ice (Notz and Worster, 2009). However, Antarctic sea ice, particularly as spring of age and their formation conditions turbulence beneath the ice is a critical melt advances (Lannuzel et al., 2010). (Petrich and Eicken, 2010). One strategy control on solute boundary layer thick- However, mushy layer theory does not for representing overall biogeochemical ness and, hence, on the transport of account for gas processes in sea ice, as cycling in sea ice has been to incorporate solutes away from that interface (Killawee gas is actually a third phase—neither source, sink, and transport terms directly et al., 1998). The experiments of Loose pure solid nor impure liquid. The addi- into dynamic pack models themselves et al. (2009) involved a water bath that tion of a third phase may effectively (Elliot et al., 2010; Vancoppenolle et al., was mixed using submersible pumps, remove the seawater infiltration mecha- 2010). The vertical transport of solutes potentially disrupting the boundary nism by allowing drained brine channels is thus handled as either simple eddy dictated by mushy layer theory. In the to become gas-filled void spaces that diffusion, or through mixing length polar ocean, sea ice formation occurs draw gas into the ice from the atmo- theory coupled to multiphase porous under a variety of turbulence conditions, sphere, or from the brine solution. The flow through the evolving brine. In but the solute boundary layer beneath subject of brine drainage mechanisms either case, thresholds such as “the natural sea ice may still be governed by (solute segregation versus gravity rule of fives” (Golden et al., 1998) and molecular rates, as indicated by mushy drainage), the range of applicability of pulse freezing enhancements (Fritsen layer theory. These phenomena may only mushy layer theory, and the mechanisms et al., 1998; Vancoppenolle et al., 2010) be valid for the bottom centimeters of that lead to seawater infiltration are all can be represented as an easy means the ice and/or for quiescent waters, but topics that remain to be conclusively to parameterize more complex porous if they prove to be more pervasive, they resolved by theoretical and data-based microstructure dynamics. have tremendous implications for our representations of sea ice. As in the open ocean (contrast Moore

1 System models are defined as the class of codes that explicitly represent ecosystem dynamics and all the principal biogeochemical cycles as well as the dynamics that determine transport.

212 Oceanography | Vol.24, No.3 et al. [2002] with Moore et al. [2004]), detailed representation of the biogeo- et al., 1994; Stefels et al., 2007). In a modeling groups have used community chemical system, TIC must be realisti- dynamic ice code, excess DMS emanating ecosystem codes as a starting point cally fractionated, with some portion from the matrix must be permitted to for sea ice biogeochemistry. Nutrient precipitated as CaCO3 and some trans- exchange with surface waters by spilling simulations in the ice column began as ported along with other brine solutes, into leads or marginal waters. There, box or layered models of photosynthetic as discussed above. The eddy mixing it will concentrate in thin, freshened uptake (Arrigo et al., 1993). Scales then concepts should be directly applicable layers near the surface (Matrai et al., expanded to the pan-polar, with heavy (Vancoppenolle et al., 2010; Jeffrey 2008), and interaction with mixed layer reliance on data assimilation to represent et al., in press). But, models for thermo- becomes critical. In the Southern processes such as phytoplankton abun- dynamic salt effects on solubility, and Ocean, strong DMS producers such as dance and distributed sea ice plus snow acid-base and mineral equilibria all need Phaeocystis populate the pack surface cover (Arrigo et al., 1997). Early algo- to be extrapolated to ice temperature (Stefels et al., 2007). If the increasing rithms for biogeochemical cycling were and salinity conditions. In an additional dominance of first-year ice in the Arctic then adapted to the Arctic bottom layer, complication, gravity drainage leaves Ocean over the coming decades creates usually for landfast ice (Lavoie et al., air-filled channels above the freeboard a marine ecosystem more like that in 2005; Jin et al., 2006). In a few cases, level (Light et al., 2003), and laminar the Antarctic, analogous phenomena transition to full pack dynamics has been layer models for vertical transfer can be may become important in the north. accomplished (Deal et al., 2011). Real applied inside brine tubes at the internal Ultimately, both upper-ice habitats and connections to the climate system have liquid-gas interface. One challenge will brine channel sulfur migration must be lagged, however. Solutes are converted then be parameterization of gas diffusion simulated (Arrigo et al., 1997; Stefels to phytoplankton and particulates with rates through upper-level networks and et al., 2007). Volatile halogens of the some fidelity, but dissolved gas chemistry snow. Another will be to account for hypersaline regime, extending up to the is not explicitly represented in these structural effects of the carbonate solids. sea ice-snow interface, are often present models. For example, Redfield ratios All gases are subject to excess satura- as alkylated organic molecules, but are generally used to couple individual tion as ionic strength (salinity) increases. elemental iodine gas is also included. elements, a highly dubious approach Bubble formation is well known in sea These compounds are transported into in the extreme environment of sea ice (Weeks and Ackley, 1982; Light et al., the polar troposphere where they cata- ice. Meanwhile, production of O2 and 2003), but the volumes are small, and lyze photochemistry (Saiz-Lopez consumption of CO2 by photosynthesis micrographs are only sometimes able et al., 2007; Simpson et al., 2007; Freiss are not simulated. to identify gas bubbles as brine channel et al., 2010). But, they are yet to be dealt The field evidence of precipitating inclusions. Nevertheless, rich headspace with in the models.

CaCO3 polymorphs from natural sea and surface chemistries likely occur The ultrastrong greenhouse forcing ice cores (Dieckmann et al., 2008) and within the ice matrix. Most current ice agent methane may present special the elaboration of the carbonate pump geochemical cycling models ignore these difficulties. Bubble seeps or flares consti- hypothesis (Alekseev and Nagurny, 2007; as secondary effects (Vancoppenolle tute well-known features of the natural Delille et al., 2007; Rysgaard et al., 2007) et al., 2010; Jeffrey et al., in press), cycle at the global scale. Moreover, large provide strong motivation to include but the bubble situation may deserve releases have recently been connected carbonate mineral processes into sea closer examination. with global warming induced degra- ice models (see above for more detail). dation of submarine permafrost and The situation has been modeled several Simulating Volatile clathrates (frozen methane hydrates). times at the analytical and box levels Gas Production Such structures are predominantly found (Jones and Coote, 1981; Rysgaard et al., In Arctic ice, algae residing primarily in in the Arctic (Kvenvolden et al. 1993; 2009), and the connections to ecosystem the bottom layer produce and/or support Shakhova et al. 2009). The molecule can dynamics will be straightforward. In any locally high levels of DMS (Levasseur move rapidly upward from sediments

Oceanography | September 2011 213 in the bubble phase (Obzhirov et al., ecodynamically well developed, but plankton. But, the gases remain explicitly 2004; Shakhova et al. 2009). Therefore, much hard work remains before gases unincorporated across this suite of major concentrations build up beneath the ice will be adequately represented. This state- efforts. The Deal et al. (2011) model is pack, both in solution and as trapped ment applies whether the compounds now being extended with bottom layer pockets of . It may be necessary involved are major or trace, simple or sulfur dynamics so that it will encompass to add yet another phase to gas mixing complex, dissolved or otherwise. For ice-sourced DMSP and DMS. Dissolved inside the sea ice matrix, and numerical example, extant codes treat the Monod inorganic carbon, alkalinity, solid precipi- approaches to brine transport already equation for organism growth, as well tation, and photosynthetic conversion constitute multiphase porous flow as light limitation, in much the same of carbonate to organics/molecular (Jeffrey et al., in press). fashion within their ecosystems, except oxygen should be expected to make their Recent general evolution of the sea for Vancoppenolle et al. (2010) who appearance soon in at least some of these ice biogeochemistry models are char- reverse engineer primary production systems. Investigators in the Deal group acterized through a set of examples from observed rates. Lavoie et al. (2005) further intend to consider methane in Table 1. Approaches being pursued include an additional, hypothetical ice bubbles arriving from below the ice pack around the community tend to be growth rate restriction on the phyto- (Shakhova et al., 2009).

Table 1. A survey of simulations that resolve biogeochemical cycles within sea ice. Box models may be considered zero dimensional. Note that only a few groups have ventured into the realm of horizontal distribution, and that gases are in general yet to be included. An interesting exception is ammonia/um, but this is due to its role in nitrogen remineralization.

Biological Elements Reference Dimensions Physical Domain Ice Dynamics and Cycles Optics Validation N, P, Si, and Heat transfer Snow, ice Local ice, Arrigo et al. Landfast at Redfield to Vertical through 10 mm attenuation, and nutrient, and (1993) McMurdo Sound chlorophyll*, layers self-shading chlorophyll data recycle NH4 N, P, Si, and Snow, ice Arrigo et al. Remotely sensed Redfield to Coarse 3D Southern Ocean attenuation, and Multiple ice cores (1997) ice, snow extent chlorophyll*, self-shading recycle NH4 Heat transfer Snow/ice attenu- Local light, Lavoie et al. Landfast at through one Si and Redfield to Box model ation, biological nutrient, and (2005) Resolute snow, two ice chlorophyll* heating chlorophyll data layers N, Si, and Redfield Landfast at Snow, ice Local ice column Jin et al. (2006) Box model Fixed empirical to chlorophyll*, Barrow attenuation chlorophyll recycle NH4 Drift station Vertical brine plus Si, from C rates Vancoppenolle Hard switch as cores at sites Ice Vertical Weddell Sea eddy drain, data based on empir- et al. (2010) short-wave diode Station Polarstern flush ical chlorophyll (ISPOL) and Liège N, Si, C, and Pan-Arctic (global Snow, ice Multiple Deal et al. 2D bottom Multiple layers Redfield to but Antarctic attenuation, and chlorophyll (2011) layer and thicknesses chlorophyll*, unvalidated) self-shading bottom data recycle NH4 *Ecodynamics are estimated using empirical relationships based on in situ chlorophyll concentrations and the Redfield ratios for sea ice.

214 Oceanography | Vol.24, No.3 RESEARCH FOR THE FUTURE balance among biological, physical, sea ice zone that account for This article emphasizes the role of sea and chemical processes in the sea ice. (1) sediment transport from the ice processes in the context of polar The interplay among these processes continental shelves to offshore melting ocean and global-scale biogeochemical can be revealed through laboratory and sites, (2) aeolian deposition, and cycles, with particular attention given theoretical study. Ultimately, the impact (3) water-column recycling to biogenic gas fluxes. The retreat of of each process must be validated with • Incorporation of all of the above summer sea ice in the Arctic has brought field studies. Here is a list of critical processes into models of sea ice at urgency to the search for a first-order topics that we hope the future of sea ice the ocean basin scale and in the understanding of carbon, oxygen, sciences will address to push our under- climate system and carbonate cycling in seasonal standing to the level of prediction: sea ice. In the past 50 years, ice algae References cryoadaptations, mineral precipitation, • In situ sensor development for Alekseev, G.V., and A.P. Nagurny. 2007. Role of sea ice in the formation of annual carbon and the permeability of sea ice have all measuring and monitoring gases in dioxide cycle in the Arctic. Doklady Earth been revealed as processes that may sea ice (laboratory, field) Sciences 417A:1,398–1,401. Amiro, B. 2010. Estimating annual carbon impact the polar ocean carbon cycle. • Validation of the timing and the dioxide eddy fluxes using open-path analysers Whereas these processes are now well processes (solute segregation vs. for cold forest sites. Agricultural and Forest Meteorology 150:1,366–1,372. documented—a major step forward— gravity drainage) that lead to desali- Anderson, L.G., and E.P. Jones. 1985. their net effect is far from clear. Many nation of sea ice and water column Measurements of total alkalinity, calcium, and sulfate in natural sea ice. Journal of Geophysical of the studies that we highlighted here convection under ice Research 90:9,194–9,198, http://dx.doi. have yielded conflicting results—often • Determination of the conditions org/10.1029/JC090iC05p09194. Anderson, L.G., E. Falck, E.P. Jones, S. Jutterström, because these studies are scattered under which CaCO3 polymorphs are and J.H. Swift. 2004. Enhanced uptake of throughout the seasonal ice cycle and found in natural sea ice atmospheric CO2 during freezing of seawater: A field study in Storfjorden, Svalbard. Journal of because of the inherent spatial variability • Determination of empirical scaling Geophysical Research 109, C06004, http://dx.doi. in sea ice. Previously oversimplified laws for the rate of bulk gas perme- org/10.1029/2003JC002120. Arrigo, K.R., J. Kremer, and C.W. Sullivan. 1993. processes, such as air-sea gas exchange ability in all sea ice types, including A simulated Antarctic fast ice ecosystem. in sea ice-covered waters, have proven columnar, granular, multiyear, and Journal of Geophysical Research 98:6,929–6,946, important to elemental budgets in simple especially springtime or melting ice http://dx.doi.org/10.1029/93JC00141. Arrigo, K., T. Mock, and M. Lizotte. 2010. Primary calculations, but remain unconstrained (laboratory, field) producers and sea ice. Pp. 282–325 in Sea Ice. by direct measurement. The community • Determination of the veracity of D.N. Thomas and G.S. Dieckmann, eds, Wiley- Blackwell, Oxford, UK. awaits emerging results from the recent mushy layer theory in the study of Arrigo, K.R., D.L. Worthen, M.P. Lizotte, P. Dixon, International Polar Year, and from future real sea ice and G. Dieckmann. 1997. Primary production in Antarctic sea ice. Science 276:394–398, http:// field and laboratory studies, as there is ° Quantification of seawater infiltra- dx.doi.org/10.1126/science.276.5311.394. a pressing need to develop predictive tion as a return flow caused by the Assur, A. 1958. Composition of sea ice and its tensile strength. Pp. 106–138 in Arctic Sea scaling laws that capture the cycling of brine drainage process Ice. Conference held at Easton, Maryland, algae/phytoplankton, nutrients, and, ° Correction for the effects of gas February 24–27, 1958, National Academy of Sciences-National Academy of Engineering, ultimately, carbon between the water as a third phase in ice that is not Washington, DC, Research Publication 598. column, the ice, and the atmosphere. explicitly accounted for in mushy Ayers, G.P., and J.M. Cainey. 2007. The CLAW hypothesis: A review of the major develop- These microscale processes have little layer theory ments. Environmental Chemistry 4:366–374, hope of being represented directly within • Exploration of the effect of organic http://dx.doi.org/10.1071/EN07080. Bakan, S. 1978. Note on the eddy correlation Earth system models, at any dimen- material on the physical properties method for CO2 flux measurements. Boundary- sion. Therefore, simulations will rely on of ice: porosity, permeability, and the Layer Meterology 14:597–600, http://dx.doi. org/10.1007/BF00121898. approximation of these processes, and radiative balance those approximations must be grounded • Development of seasonally varying in a robust understanding of the net nutrient budgets for the seasonal

Oceanography | September 2011 215 Bakker, D.C.E., H.J.W. DeBaar, and U.V. Bathmann. Dieckmann, G.S., G. Nehrke, S. Papadimitriou, Gosink, T.A. 1980. Atmospheric trace gases in 1997. Changes of carbon dioxide in surface J. Göttlicher, R. Steninger, H. Kennedy, D. Wolf- association with sea ice. Antarctic Journal of the waters during spring in the Southern Ocean. Gladrow, and D.N. Thomas. 2008. Calcium United States 15:82–83. Deep-Sea Research Part II 44:91–127, http:// carbonate as ikaite crystals in Antarctic sea ice. Gosink, T.A., J.G. Pearson, and J.J. Kelly. 1976. Gas dx.doi.org/10.1016/S0967-0645(96)00075-6. Geophysical Research Letters 35, L08501, http:// movement through sea-ice. Nature 263:41­–42, Baldochi, D. 2003. Assessing the eddy covari- dx.doi.org/10.1029/2008GL033540. http://dx.doi.org/10.1038/263041a0. ance technique for evaluating carbon dioxide Edson, J.B., A.A. Hinton, K.E. Prada, J.E. Hare, Holter, P. 1990. Sampling air from dung pats by exchange rates of ecosystems: Past, present and and C.W. Fairall. 1998. Direct covariance silicone rubber diffusion chambers. Soil Biology future. Global Change Biology 9:479–492, http:// flux estimates from mobile platforms at & Biochemistry 22:995–997, http://dx.doi. dx.doi.org/10.1046/j.1365-2486.2003.00629.x. sea. Journal of Atmospheric and Oceanic org/10.1016/0038-0717(90)90143-N. Booth, J.A. 1984. The epontic algal community Technology 15:547–562, http://dx.doi. Jacinthe, P.-A., and W.A. Dick. 1996. Use of of the ice edge zone and its significance to the org/10.1175/1520-0426(1998)015<0547:​ silicone tubing to sample nitrous oxide Davis Strait ecosystem. Arctic 37:234–243. DCFEFM>2.0.CO;2. in the soil atmosphere. Soil Biology & Burba, G.G., D.K. McDermitt, A. Grelle, Elliot, S., E.C. Hunke, N. Jeffrey, A. Turner, Biochemistry 28:721–726, http://dx.doi. D.J. Anderson, and L. Xu. 2008. Addressing M. Maltrud, C. Deal, and M. Jin. 2010. Systems org/10.1016/0038-0717(95)00176-X. the influence of instrument surface heat level simulation of sea ice-atmosphere biogeo- Jeffrey, N., E.C. Hunke, and S. Elliot. In press.

exchange on the measurements of CO2 flux chemistry connections. SOLAS News 11:14–18. Modeling the transport of passive tracers in from open-path gas analyzers. Global Change Else, B.G.T., T.N. Papakyriakou, R.J. Galley, sea ice. Journal of Geophysical Research, http:// Biology 14:1,854–1,876. W.M. Drennan, L.A. Miller, and H. Thomas. dx.doi.org/10.1029/2010JC006527. Burkholder, P.R., and E.F. Mandelli. 1965. In press. Eddy covariance measurements of Jin, M., C. Deal, J. Wang, K. Shin, N. Tanaka,

Productivity of microalgae in Antarctic Sea Ice. wintertime CO2 fluxes in an arctic polynya: T. Whitledge, S. Lee, and R. Gradinger. Science 149:872–874, http://dx.doi.org/10.1126/ Evidence for enhanced gas transfer during ice 2006. Controls of the landfast ice-ocean science.149.3686.872. formation. Journal of Geophysical Research, ecosystem offshore Barrow, Alaska. Annals Carey, A.G. Jr. 1987. Particle flux beneath fast http://dx.doi.org/10.1029/2010JC006760. of Glaciology 44:63–72, http://dx.doi. ice in the shallow southwestern Beaufort Feltham, D.L., N. Untersteiner, and J.S. Wettlaufer. org/10.3189/172756406781811709.

Sea, Arctic Ocean. Marine Ecology Progress 2006. Sea ice is a mushy layer. Geophysical Jones, E.P., and A.R. Coote. 1981. Oceanic CO2 Series 40:247–257, http://dx.doi.org/10.3354/ Research Letters 33, L14501, http://dx.doi. produced by the precipitation of CaCO3 meps040247. org/10.1029/2006GL026290. from brines in sea ice. Journal of Geophysical Caron, D.A., and R.J. Gast. 2010. Heterotrophic Freiss, U., T. Deutschmann, B. Gilfedder, R. Weller, Research 86:11,041–11,043, http://dx.doi. protists associated with sea ice. Chapter 9 in Sea and U. Platt. 2010. Iodine monoxide in the org/10.1029/JC086iC11p11041. Ice, 2nd ed. D.N. Thomas and G.S. Dieckmann, Antarctic snowpack. Atmospheric Chemistry Kammann, C., L. Grünhage, and H.-J. Jäger. eds, Wiley-Blackwell, Oxford, UK, http://dx.doi. and Physics 10:2,439–2,456. 2001. A new sampling technique to monitor

org/10.1002/9781444317145.ch9. Freitag, J., and H. Eicken. 2003. Meltwater circula- concentrations of CH4, N2O and CO2 in Charlson, R.J., J.E. Lovelock, M.O. Andreae, and tion and permeability of Arctic summer sea ice air at well-defined depths in soils with S.G. Warren. 1987. Oceanic phytoplankton, derived from hydrological field experiments. varied water potential. European Journal atmospheric sulphur, cloud albedo and Journal of Glaciology 49:349–358, http://dx.doi. of Soil Science 52:297–303, http://dx.doi. climate. Nature 326:655–661, http://dx.doi. org/10.3189/172756503781830601. org/10.1046/j.1365-2389.2001.00380.x. org/10.1038/326655a0. Fritsen, C., S.F. Ackley, J. Kremer, and Kawamoto, K., P. Moldrup, P. Schjonning, Cox, G.F.N., and W.F. Weeks. 1983. Equations C.W. Sullivan. 1998. Flood-freeze cycles B.V. Iversen, D.E. Rolston, and T. Komatsu. for determining the gas and brine volumes and microalgal dynamics in Antarctic pack 2006. Gas transport parameters in the vadose in sea-ice samples. Journal of Glaciology ice. Pp. 1–22 in Antarctic Sea Ice: Biological zone: Gas diffusivity in field and lysimeter soil 29:306–316. Processes, Interactions and Variability. Antarctic profiles. Vadose Zone Journal 5:1,194–1,204. Cox, G.F.N., and W.F. Weeks. 1988. Numerical Research Series, vol. 73, M. Lizotte and Kelley, J.J. 1987. Carbon dioxide in the simulations of the profile properties of K.R. Arrigo, eds, American Geophysical Union, Arctic environment. The Journal of Earth undeformed first-year sea ice during the Washington, DC. Sciences 35:341–354. growth season. Journal of Geophysical Gibson, J.A.E., and T.W. Trull. 1999. Annual Kelley, J.J., and T.A. Gosink. 1979. Gases in Sea Ice:

Research 93:12,449–12,460, http://dx.doi. cycle of fCO2 under sea-ice and in open 1975–1979. University of Alaska Fairbanks, org/10.1029/JC093iC10p12449. water in Prydz Bay, East Antarctica. Marine Fairbanks, AK. Deal, C., M. Jin, S. Elliot, E.C. Hunke, M. Maltrud, Chemistry 66:187–200, http://dx.doi. Killawee, J.A., I.J. Fairchild, J.-L. Tison, L. Janssens, and N. Jeffrey. 2011. Large-scale modeling org/10.1016/S0304-4203(99)00040-7. and R. Lorrain. 1998. Segregation of solutes of primary production and ice algal biomass Gitterman, K.E. 1937. Thermal analysis of sea and gases in experimental freezing of dilute within arctic sea ice in 1992. Journal of water. Trudy Solyanoy Laboratorii (Transactions solutions: Implications for natural glacial Geophysical Research 116, C07004, http://dx.doi. of Saline Laboratory) 15:5–23. systems. Geochimica et Cosmochimica Acta org/10.1029/2010JC006409. Golden, K.M., S.F. Ackley, and V.I. Lytle. 1998. 62:3,637–3,655, http://dx.doi.org/10.1016/ Delille, B. 2006. Inorganic Carbon Dynamics and The percolation phase transition in sea S0016-7037(98)00268-3.

Air-Ice-Sea CO2 Fluxes in the Open and Coastal ice. Science 282:2,238–2,241, http://dx.doi. Krembs, C., H. Eicken, and J.W. Deming. 2011. Waters of the Southern Ocean. Docteur en org/10.1126/science.282.5397.2238. Exopolymer alteration of physical properties Sciences thesis, Université de Liège, Liège, Golden, K.M., H. Eicken, A.L. Heaton, J. Miner, of sea ice and implications for ice habitability France. D.J. Pringle, and J. Zhu. 2007. Thermal evolu- and biogeochemistry in a warmer Arctic. Delille, B., B. Jourdain, A.V. Borges, J.-L. Tison, tion of permeability and microstructure in sea Proceedings of the National Academy Sciences of

and D. Delille. 2007. Biogas (CO2, O2, dimeth- ice. Geophysical Research Letters 34, L16501, the United States of America 108:3,653–3,658, ylsulfide) dynamics in spring Antarctic fast ice. http://dx.doi.org/10.1029/2007GL030447. http://dx.doi.org/10.1073/pnas.1100701108. Limnology and Oceanography 52:1,367–1,379, Golden, K.M., A.L. Heaton, H. Eicken, and Krembs, C., H. Eicken, K. Junge, and J.W. Deming. http://dx.doi.org/10.4319/lo.2007.52.4.1367. V.I. Lytle. 2006. Void bounds for fluid transport 2002. High concentrations of exopoly- Deming, J.W. 2010. Sea ice bacteria and viruses. in sea ice. Mechanics of Materials 38:801–817, meric substances in Arctic winter sea ice: Chapter 7 in Sea Ice, 2nd ed. D.N. Thomas http://dx.doi.org/10.1016/j.mechmat.2005.​ Implications for the polar ocean carbon cycle and G.S. Dieckmann, eds, Wiley-Blackwell, 06.015. Oxford, UK.

216 Oceanography | Vol.24, No.3 and cryoprotection of diatoms. Deep-Sea Lyakhin, Y.I. 1970. Saturation of water of the undisturbed soil. Soil Science Society of America Research Part I 49:2,163–2,181, http://dx.doi. Sea of Okhotsk with calcium carbonate. Journal 68:750–759, http://dx.doi.org/10.2136/ org/10.1016/S0967-0637(02)00122-X. Oceanology 10:789–795. sssaj2004.0750. Kuosa, H., B. Norrman, K. Kivi, and F. Brandini. Marion, G.M. 2001. Carbonate mineral solu- Moore, J., S. Doney, J. Kleypas, D. Glover, and 1992. Effects of Antarctic sea ice biota on bility at low temperatures in the Na-K-Mg- I. Fung. 2002. An intermediate complexity

seeding as studied in aquarium experiments. Ca-H-Cl-SO4-OH-HCO3-CO3-CO2-H2O marine ecosystem model for the global domain. Polar Biology 12:333–339, http://dx.doi. system. Geochimica et Cosmochimica Acta Deep-Sea Research Part II 49:403–462, http:// org/10.1007/BF00243104. 65:1,883–1,896. dx.doi.org/10.1016/S0967-0645(01)00108-4. Kvenvolden, K.A., M.D. Lilley, T.D. Lorenson, Martinson, D.G. 1990. Evolution of the Southern Moore, J., S. Doney, and K. Lindsay. 2004. Upper P.W. Barnes, and E. McLaughlin. 1993. The Ocean winter mixed layer and sea ice: ocean ecosystem dynamics and iron cycling Beaufort Sea continental shelf as a seasonal Open ocean deep-water formation and in a global three dimensional model. Global source of atmospheric methane. Geophysical ventilation. Journal of Geophysical Research Biogeochemical Cycles 18, GB4028, http://dx.doi. Research Letters 20:2,459–2,462, http://dx.doi. 95:11,641–11,654, http://dx.doi.org/10.1029/ org/10.1029/2004GB002220. org/10.1029/93GL02727. JC095iC07p11641. Nagurnyi, A.P. 2008. On the role of the Arctic sea Lannuzel, D., V. Schoemann, J. de Jong, B. Pasquer, Massman, W.J., and X. Lee. 2002. Eddy covari- ice in seasonal variability of carbon dioxide P. van der Merwe, F. Masson, J.-L. Tison, and ance flux corrections and uncertainties in concentration in northern latitudes. Russian A. Bowie. 2010. Distribution of dissolved iron long-term studies of carbon and energy Meteorology and Hydrology 33:43–47. in Antarctic sea ice: Spatial, seasonal, and exchanges. Agricultural and Forest Meteorology Nelson, K.H., and T.G. Thompson. 1954. inter-annual variability. Journal of Geophysical 113:121–144, http://dx.doi.org/10.1016/ Deposition of salts from sea water by Research 115, G03022, http://dx.doi. S0168-1923(02)00105-3. frigid concentration. Journal of Marine org/10.1029/2009JG001031. Matrai, P., L. Tranvik, C. Leck, and J. Knulst. Research 13:166–182. Lannuzel, D., V. Schoemann, J. de Jong, J.-L. Tison, 2008. Are high arctic surface microlayers a Nomura, D., H. Eicken, R. Gradinger, and and L. Chou. 2007. Distribution and biogeo- potential source of aerosol organic precursors? K. Shirasawa. 2010a. Rapid physically driven

chemical behaviour of iron in the East Antarctic Marine Chemistry 108:109–122, http://dx.doi. inversion of the air-sea ice CO2 flux in the sea ice. Marine Chemistry 106:18–32, http:// org/10.1016/j.marchem.2007.11.001. seasonal landfast ice off Barrow, Alaska dx.doi.org/10.1016/j.marchem.2006.06.010. Matsuo, S., and Y. Miyake. 1966. Gas composi- after onset of surface melt. Continental Lavoie, D., K. Denman, and C. Michel. 2005. tion in ice samples from Antarctica. Journal of Shelf Research 30:1,998–2,004, http://www. Modeling ice algal growth and decline in a Geophysical Research 71:5,235–5,241. sciencedirect.com/science/article/pii/ seasonally ice-covered region of the Arctic. McGillis, W.R., J.B. Edson, J.E. Hare, and S0278434310003067. Journal of Geophysical Research 110, C11009, C.W. Fairall. 2001. Direct covariance Nomura, D., T. Takatsuka, M. Ishikawa,

http://dx.doi.org/10.1029/2005JC002922. air-sea CO2 fluxes. Journal of Geophysical T. Kawamura, K. Shirasawa, and H. Yoshikawa- Legendre, L., S.F. Ackley, G.S. Dieckmann, Research 106:16,729–16,745, http://dx.doi. Inoue. 2009. Transport of chemical components B. Gulliksen, R. Horner, T. Hoshiai, org/10.1029/2000JC000506. in sea ice and under-ice water during melting in I.A. Melnikov, W.S. Reeburgh, M. Spindler, and Miller, L.A., G. Carnat, B.G.T. Else, N. Sutherland, the seasonally ice-covered Saroma-ko Lagoon, C.W. Sullivan. 1992. Ecology of sea ice biota. and T.N. Papakyriakou. In press. Carbonate Hokkaido, Japan. Estuarine, Coastal and Shelf 2. Global significance. Polar Biology 12:429–444, system evolution at the Arctic Ocean Science 81:201–209, http://dx.doi.org/10.1016/​ http://dx.doi.org/10.1007/BF00243114. surface during autumn freeze-up. Journal j.ecss.2008.10.012. Levasseur, M., M. Gosselin, and S. Michaud. 1994. of Geophysical Research, http://dx.doi. Nomura, D., H. Yoshikawa-Inoue, and A new source of dimethylsulfide (DMS) for org/10.1029/2011JC007143. T. Toyota. 2006. The effect of sea-ice growth

the arctic atmosphere: Ice diatoms. Marine Miller, L.A., R.E. Collins, J.W. Deming, J.K. Ehn, on air-sea CO2 flux in a tank experiment. Biology 121:381–387, http://dx.doi.org/10.1007/ R.W. Macdonald, A. Mucci, O. Owens, Tellus 58B:418–426, http://dx.doi.org/10.1111/​ BF00346748. M. Raudsepp, and N. Sutherland. 2011. Carbon j.1600-0889.2006.00204.x. Light, B., G.A. Maykut, and T.C. Grenfell. 2003. dynamics in Sea Ice: A winter flux time series. Nomura, D., H. Yoshikawa-Inoue, T. Toyota, and Effects of temperature on the microstruc- Journal of Geophysical Research 116, C02028, K. Shirasawa. 2010b. Effects of snow, snow- ture of first-year Arctic sea ice. Journal of http://dx.doi.org/10.1029/2009JC006058. melting and refreezing processes on air-sea-ice

Geophysical Research 108, 3051, http://dx.doi. Miyake, Y., and S. Matsuo. 1963. A role of sea ice CO2 flux. Journal of Glaciology 56:262–270, org/10.1029/2001JC000887. and sea water in the Antarctic on the carbon http://dx.doi.org/10.3189/002214310791968548. Lizotte, M.P. 2001. The contributions of sea ice dioxide cycle in the atmosphere. Papers in Notz, D., and M.G. Worster. 2009. Desalination algae to Antarctic marine primary produc- Meteorology and Geophysics 14:120–125. processes of sea ice revisited. Journal of tion. American Zoology 41:57–73, http:// Mock, T., and D.N. Thomas. 2005. Recent advances Geophysical Research 114, C05006, http://dx.doi. dx.doi.org/10.1668/0003-1569(2001)041[0057:​ in sea-ice microbiology. Environmental org/10.1029/2008JC004885. TCOSIA]2.0.CO;2. Microbiology 7:605–619, http://dx.doi. Obzhirov, A., R. Shakirov, A. Salyuk, E. Suess, Loose, B., W.R. McGillis, P. Schlosser, D. Perovich, org/10.1111/j.1462-2920.2005.00781.x. N. Biebow, and A. Salomatin. 2004. Relations and T. Takahashi. 2009. The effects of freezing, Mock, T., G.S. Dieckmann, C. Haas, A. Krell, between methane venting, geological structure growth and ice cover on gas transport processes J.-L. Tison, A.L. Belem, S. Papadimitriou, and and seismo-tectonics in the Okhotsk Sea. in laboratory seawater experiments. Geophysical D.N. Thomas. 2002. Micro-optodes in sea ice: Geo-Maring Letters 24:135–139, http://dx.doi. Research Letters 36, L05603, http://dx.doi. A new approach to investigate oxygen dynamics org/10.1007/s00367-004-0175-0. org/10.1029/2008GL036318. during sea ice formation. Aquatic Microbial Oertling, A.B., and R.G. Watts. 2004. Growth of

Loose, B., P. Schlosser, D. Perovich, D. Ringelberg, Ecology 29:297–306, http://dx.doi.org/10.3354/ and brine drainage from NaCl-H2O freezing: D.T. Ho, T. Takahashi, J. Richter-Menge, ame029297. A simulation of young sea ice. Journal of C.M. Reynolds, W.R. McGillis, and Mock, T., M. Kruse, and G.S. Dieckmann. 2003. A Geophysical Research 109, C04013, http://dx.doi. J.-L. Tison. 2010. Gas diffusion through new microcosm to investigate oxygen dynamics org/10.1029/2001JC001109. columnar laboratory sea ice: Implications at the sea ice water interface. Aquatic Microbial Orsi, A.H., G.C. Johnson, and J.L. Bullister.

for mixed-layer ventilation of CO2 in the Ecology 30:197–205, http://dx.doi.org/10.3354/ 1999. Circulation, mixing and produc- seasonal ice zone. Tellus B, http://dx.doi. ame030197. tion of Antarctic Bottom Water. Progress org/10.1111/j.1600-0889.2010.00506.x. Moldrup, P., T. Olesen, S. Yoshikawa, T. Komatsu, in Oceanography 43:55–109, http://dx.doi. and D.E. Rolston. 2004. Three-porosity model org/10.1016/S0079-6611(99)00004-X. for predicting the gas diffusion coefficient in

Oceanography | September 2011 217 Papadimitriou, S., D.N. Thomas, H. Kennedy, Smith, W.O., and D.M. Nelson. 1985. Antarctic sea ice: The case of dissolved silica. C. Haas, H. Kuosa, A. Krell, and Phytoplankton bloom produced by a receding Journal of Geophysical Research 115, C02005, G.S. Dieckmann. 2007. Biogeochemical compo- ice edge in the Ross Sea: Spatial coherence with http://dx.doi.org/10.1029/2009JC005369. sition of natural sea ice brines from the Weddell the density field. Science 227:163–166, http:// Webb, E.K., G.I. Pearman, and R. Leuning. 1980. Sea during early austral summer. Limnology and dx.doi.org/10.1126/science.227.4683.163. Correction of flux measurements for density Oceanography 52:1,809–1,823, http://dx.doi. Smith, W.O., and D.M. Nelson. 1986. Importance effects due to heat and water vapour transfer. org/10.4319/lo.2007.52.5.1809. of ice edge phytoplankton production. Quarterly Journal of the Royal Meteorological Papakyriakou, T., and L. Miller. 2011. Springtime Bioscience 36:251–257, http://dx.doi. Society 106:85–100, http://dx.doi.org/10.1002/

CO2 exchange over seasonal sea ice in org/10.2307/1310215. qj.49710644707. the Canadian Arctic Archipelago. Annals Steele, M., and J.H. Morison. 1995. Weeks, W., and S.F. Ackley. 1982. The Growth, of Glaciology 52:215–224, http://dx.doi. water formation in the Barents Sea. Journal Structure and Properties of Sea Ice. Cold Regions org/10.3189/172756411795931534. of Geophysical Research 100:881–894, http:// Research and Engineering Laboratory, ERDC, Perovich, D.K., and A.J. Gow. 1996. A quantitative dx.doi.org/10.1029/94JC02310. Hanover, NH, 140 pp. description of sea ice inclusions. Journal of Stefels, J., M. Steinke, S. Turner, G. Malin, and Weller, G. 1968. Heat-energy transfer through a Geophysical Research 101:18,327–18,343, http:// S. Belviso. 2007. Environmental constraints on four-layer system: Air, snow, sea ice, sea water. dx.doi.org/10.1029/96JC01688. the production and removal of dimethylsul- Journal of Geophysical Research 73:1,209–1,220, Petrich, C., and H. Eicken. 2010. Growth, structure phide and implications for ecosystem modeling. http://dx.doi.org/10.1029/JB073i004p01209. and properties of sea ice. Chapter 2 in Sea Ice, Biogeochemistry 83:245–275, http://dx.doi. Wettlaufer, J.S., M.G. Worster, and H.E. Hupper. 2nd ed. D.N. Thomas and G.S. Dieckmann, eds, org/10.1007/s10533-007-9091-5. 1997. Natural convection during solidifica- Wiley-Blackwell, Oxford, UK, http://dx.doi. Sweeney, C. 2003. The annual cycle of surface tion of an alloy from above with application

org/10.1002/9781444317145.ch2. CO2 and O2 in the Ross Sea: A model for to the evolution of sea ice. Journal of Fluid Prytherch, J., M.J. Yelland, R.W. Pascal, B.I. Moat, gas exchange on the continental shelves of Mechanics 344:291–316, http://dx.doi. I. Skjelvan, and C.C. Neill. 2010. Direct Antarctica. Pp. 295–312 in Biogeochemistry of org/10.1017/S0022112097006022.

measurements of the CO2 flux over the ocean: the Ross Sea. G.R. DiTullio and R.B. Dunbar, Worster, M.G. 1991. Natural convection in a mushy Development of a novel method. Journal of eds, Antarctic Research Series, American layer. Journal of Fluid Mechanics 224:335–359, Geophysical Research 37, L03607, http://dx.doi. Geophysical Union, Washington, DC. http://dx.doi.org/10.1017/S0022112091001787. org/10.1029/2009GL041482. Tamura, T., K. Ohshima, and S. Nihashi. Yager, P.L., D.W.R. Wallace, K. Johnson, Rysgaard, S., J. Bendtsen, L.T. Pedersen, H. Ramlov, 2008. Mapping of sea ice production for W.O. Smith, P.J. Minnett, and J.W. Deming.

and R.N. Glud. 2009. Increased CO2 uptake due Antarctic coastal polynyas. Geophysical 1995. The northeast water polynya as an to sea ice growth and decay in the Nordic Seas. Research Letters 35, L07606, http://dx.doi. atmospheric CO2 sink: A seasonal rectifica- Journal of Geophysical Research 114, C09011, org/10.1029/2007GL032903. tion hypothesis. Journal of Geophysical http://dx.doi.org/10.1029/2008JC005088. Thomas, D.N., and G.S. Dieckmann, eds. 2010. Research 100:4,389–4,398, http://dx.doi. Rysgaard, S., R.N. Glud, M.K. Sejr, J. Bendtsen, Sea Ice, 2nd ed. Wiley-Blackwell, Oxford, org/10.1029/94JC01962. and P.B. Christensen. 2007. Inorganic carbon UK, 640 pp. Zappa, C.J., W.R. McGillis, P.A. Raymond, transport during sea ice growth and decay: Tison, J.-L., C. Haas, M.M. Gowing, S. Sleewaegen, J.B. Edson, E.J. Hintsa, H.J. Zemmelink, A carbon pump in polar seas. Journal of and A. Bernard. 2002. Tank study of physico- J.W.H. Dacey, and D.T. Ho. 2007. Geophysical Research 112, C03016, http://dx.doi. chemical controls on gas content and composi- Environmental turbulent mixing org/10.1029/2006JC003572. tion during growth of young sea ice. Journal controls on air-water gas exchange in Rysgaard, S., R.N. Glud, M.K. Sejr, M.E. Blicher, of Glaciology 48:177–191, http://dx.doi. marine and aquatic systems. Geophysical and H.J. Stahl. 2008. Denitrification activity org/10.3189/172756502781831377. Research Letters 34, L10601, http://dx.doi. and oxygen dynamics in Arctic sea ice. Polar Trevena, A.J., and G.B. Jones. 2006. org/10.1029/2006GL028790. Biology 31:527–537, http://dx.doi.org/10.1007/ Dimethylsulphide and dimethylsulphonio- Zemmelink, H.J., B. Delille, J.L. Tison, s00300-007-0384-x. propionate in Antarctic sea ice and their E.J. Hintsa, L. Houghton, and J.W.H. Dacey.

Saiz-Lopez, A., A. Mahajan, R. Salmon, release during sea ice melting. Marine 2006. CO2 deposition over multi-year ice S. Bauguitte, A. Jones, H. Roscoe, and J. Plane. Chemistry 98:210–222, http://dx.doi. of the western Weddell Sea. Geophysical 2007. Boundary layer halogens in coastal org/10.1016/j.marchem.2005.09.005. Research Letters 33, L13606, http://dx.doi. Antarctica. Science 317:348–351, http://dx.doi. Trevena, A.J., G.B. Jones, S.W. Wright, and org/10.1029/2006GL026320. org/10.1126/science.1141408. R.L. van den Enden. 2000. Profiles of DMSP, Zemmelink, H.J., J.W.H. Dacey, L. Houghton, Semiletov, I., A. Makshtas, S.-I. Akasofu, and algal pigments, nutrients and salinity in pack E.J. Hintsa, and P.S. Liss. 2008. Dimethylsulfide

E.L. Andreas. 2004. Atmospheric CO2 ice from eastern Antarctica. Journal of Sea emissions over the multi-year ice of balance: The role of Arctic sea ice. Geophysical Research 43:265–273, http://dx.doi.org/10.1016/ the western Weddell Sea. Geophysical Research Letters 31, L05121, http://dx.doi. S1385-1101(00)00012-5. Research Letters 35, L06603, http://dx.doi. org/10.1029/2003GL017996. Trevena, A.J., G.B. Jones, S.W. Wright, and org/10.1029/2007GL031847. Shakhova, N., V. Sergienko, and I. Semiletov. 2009. R.L. van den Enden. 2003. Profiles of The contribution of the East Siberian Shelf dimethylsulphoniopropionate (DMSP), algal to the modern methane cycle. Herald of the pigments, nutrients, and salinity in the fast ice Russian Academy of Science 79:237–246, http:// of Prydz Bay, Antarctica. Journal of Geophysical dx.doi.org/10.1134/S101933160903006X. Research 108(C5), 3145, http://dx.doi. Simpson, W.R., R. von Glasow, R. Riedel, org/10.1029/2002JC001369. K. Anderson, P. Ariya, J. Bottenheim, Untersteiner, N. 1964. Calculations of tempera- J. Burrows, L.J. Carpenter, U. Frieß, ture regime and heat budget of sea ice in M.E. Goodsite, and others. 2007. Halogens the Central Arctic. Journal of Geophysical and their role in polar boundary-layer Research 69:4,755–4,766, http://dx.doi. ozone depletion. Atmospheric Chemistry org/10.1029/JZ069i022p04755. and Physics 7:4,375–4,418, http://dx.doi. Vancoppenolle, M., H. Goosse, A. de Montety, org/10.5194/acp-7-4375-2007. T. Fichefet, B. Tremblay, and J.-L. Tison. 2010. Modeling brine and nutrient dynamics in

218 Oceanography | Vol.24, No.3