SEARCH

SEARCH BY CITATION

Keywords:

  • frost flowers;
  • ice cores;
  • aerosol;
  • sea salt;
  • tropospheric ozone

Abstract

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Sampling Sites and Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusions
  8. Acknowledgments
  9. References

[1] This paper discusses the chemical composition of frost flowers and their accompanying slush layers and the evidence for their role as a salt source in processes important to atmospheric chemistry and ice core interpretation. Analysis of Antarctic frost flowers shows that they are highly saline and fractionated in sea-salt ions, with sulfate being depleted strongly relative to sodium. Because frost flowers give a bright return on satellite scatterometer images, the times and places of their formation can be identified. When winds blow towards an aerosol sampling station from areas identified by the scatterometer as covered with flowers, the collected aerosol is also depleted in sulfate. Because the flowers have a large salinity, bromide concentrations are elevated in frost flowers relative to seawater. With their high surface area, it is possible that bromine is released to the atmosphere from frost flowers, with consequent implications for tropospheric ozone depletion. The finding that quantities of fractionated sea salt are available at the sea–ice interface in the winter months and may be transported inland as aerosol also has implications for the interpretation of ice core records. Analysis of one near-coastal core shows that the majority of the sodium comes from a fractionated source rather than from open water. Hitherto, strong sea-salt signals in ice cores have been attributed to increased open water and more efficient transport inland, perhaps due to stormier weather. At least in coastal regions, however, these signals may be related instead to the increased formation of sea ice and frost flowers.

1. Introduction

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Sampling Sites and Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusions
  8. Acknowledgments
  9. References

[2] Frost flowers have not been extensively studied; a few papers discuss their physical properties, especially as related to remote sensing [Hallikainen and Winebrenner, 1992; Onstott, 1992; Martin et al., 1995, 1996; Nghiem et al., 1997], but only one paper to date has briefly considered their chemistry [Rankin et al., 2000]. Yet the flowers are widespread; the sea ice extent around Antarctica increases by 15 million km2 over the winter each year [Gloerson and Campbell, 1991], and frost flower formation is a normal part of the sea ice formation process. Because these flowers grow at the interface between the atmosphere and ocean, it is important to understand the effect that these fragile crystals have on air–sea chemical exchange.

[3] The freezing of seawater was investigated by Richardson [1976]. With the aid of Nuclear Magnetic Resonance spectroscopy, he looked at the partitioning of sea-salt ions between the solid and liquid phases with respect to temperature. Sea ice is fresher than the seawater from which it is formed, and as temperature decreases below the freezing point, the volume ratio of ice to brine increases and the brine becomes more concentrated. However, at low temperatures not all ions behave in the same way. Below −8°C, sodium sulfate begins to precipitate out as mirabilite (Na2SO4 · 10 H2O). The loss is exponential with temperature: at −10°C approximately half the sulfate has been lost from the liquid phase, and by −20°C only a tenth remains. At the temperatures commonly experienced at the air–ice interface in winter, this fractionation may remove most of the sulfate and up to 13% of the sodium from the brine. Both of these effects are important in our ice core interpretation.

[4] Below −22°C sodium chloride begins to be precipitated, and at lower temperatures other salts are also lost from the brine. Because ice surface temperatures this cold are generally associated with multiyear ice on which flowers do not form, we do not expect to see flowers depleted in NaCl.

[5] The nature of the frost flower formation and the brine segregation depends on the temperature structure within the newly formed sea ice and in the atmospheric boundary layer above the ice. Figure 1 shows a schematic drawing of frost flower formation on young sea ice, along with an example of a temperature profile in the ocean, ice, and atmosphere. Because the ice surface is warmer than the overlying atmosphere, the temperature profile above the ice surface is determined by a turbulent boundary layer, so that the air temperature decreases rapidly with height over distances of a few cm.

image

Figure 1. Schematic drawing of frost flower formation on young sea ice, with an example of a typical temperature profile in the ocean, ice, and atmosphere. The numbers on the temperature profile are for illustration purposes only (see the work of Martin et al. [1996] for additional examples).

Download figure to PowerPoint

[6] Frost flower formation takes place in four steps [Martin et al., 1996]. First, within thin porous young ice, the thermomolecular pressure gradient transports brine from the ice interior toward the relatively colder surface [Wettlaufer and Worster, 1995]. At the surface, the brine accumulates as both a liquid and a slushy layer on top of the thin newly formed sea ice [Perovich and Richter-Menge, 1994]. Second, the surface brine evaporates into the colder atmospheric boundary layer. This creates a 1–3 cm thick water vapor layer just above the surface that is supersaturated with respect to ice. Thus any crystal growth into this supersaturated layer is enhanced, yielding the growth of frost crystals. Third, beneath the crystals, a slush layer consisting of saturated brine forms and increases in thickness up to characteristic values of 2–4 mm. Fourth, surface tension effects draw up the surface brine onto the frost crystals, yielding the large observed salinities.

[7] For air temperatures of about −20°C, both the slush layer and crystals have salinities of about 100 psu, almost three times as concentrated as seawater [Drinkwater and Crocker, 1988; Perovich and Richter-Menge, 1994; Martin et al., 1995]. Also, as illustrated by Figure 2 and by the figures of Nghiem et al. [1997], the areal distribution of the flowers and slush layers is patchy, with typical length and separation scales of 10 cm. Between these patches the ice surface is solid, with no slush layers. These patchy conductive slush layers beneath the flowers mean that when frost flowers are present on sea ice, the surface presents a greater roughness and backscatter coefficient at radar frequencies than snow or an ice surface without flowers [Nghiem et al., 1997]. This increase in radar backscatter and corresponding bright signal in synthetic aperture radar (SAR) and scatterometer imagery makes it possible to identify flower regions by satellite [Hallikainen and Winebrenner, 1992; Onstott, 1992; Nghiem et al., 1997].

image

Figure 2. A patch of frost flowers in the Weddell Sea, near Halley base.

Download figure to PowerPoint

[8] The decrease in air temperature with distance above the surface has consequences for the flower and slush layer chemistry. Suppose that the air temperature is −20°C, while because of heat flux through the ice, the ice surface is warmer than the air temperature, but still well below freezing. The growth region of a commonly observed dendritic flower occurs between temperatures of −12°C and −16°C [Martin et al., 1996]. This means that for these flowers, the mirabilite is precipitated out either within the slush layer, or, if the ice is cold enough, within the ice interior. The flux into the flowers of brine that is depleted in sodium sulfate means that depending on the ice surface temperature, the mirabilite either precipitates out within the slush, enhancing it with respect to sodium sulfate, or within the sea ice interior, yielding both depleted slush and flowers. Thus while we expect that the flowers are generally depleted in sodium sulfate, the slush layer concentrations may vary greatly.

[9] Frost flowers are typically a few centimeters in diameter and have a delicate crystalline form. The precise morphology of the crystals depends on the temperature at which they form [Perovich and Richter-Menge, 1994; Martin et al., 1996]. Figure 2 shows the typical appearance of frost flower fields in the vicinity of Halley.

[10] There are at least two reasons for regarding frost flowers as more than an interesting curiosity. The first comes from glacial ice core interpretation, the second from the role of bromine in tropospheric ozone depletion. First, examination of winter deposited layers in near-coastal ice cores, and also in aerosol collected during the winter months in coastal locations, shows that they are often depleted in sulfate relative to the other sea-salt ions: the same fractionation as is seen in seawater as it freezes. Although we have no direct observations of windblown frost flowers, given that the flower and slush layer growth create a segregation of the salts and that the sail-like structure of the flowers should enhance their aerodynamic drag, this means that the flowers, rather than open seawater, may be a significant source of windblown sea salt in winter [Wagenbach et al., 1998]. An analysis of one frost flower sample from the Weddell Sea did show close agreement between the ratio of sulfate/sodium found in the sample, and that found in aerosol collected in winter at the British Antarctic Survey station, Halley [Rankin et al., 2000]. A simple calculation in this paper also suggested that frost flowers should be responsible for the majority of sea-salt aerosol produced within the sea–ice zone in winter.

[11] Second, the role of bromine in ozone depletion in the polar troposphere has been demonstrated over the last decade or so. Barrie et al. [1988] showed that ozone concentrations are anticorrelated with filterable bromide during the Arctic spring; more recently measurements of reactive bromine species such as BrO [e.g., Tuckermann et al., 1997] have given insight into the reaction pathways involved. Bromide ions are present in seawater at relatively low concentration, but frost flower salinity is considerably higher. Large quantities of sea salt and thus bromide can therefore be held within frost flowers at the ocean–atmosphere boundary, and due to the large surface area of frost flowers could potentially be easily liberated to the atmosphere.

[12] In this paper, we demonstrate that the seawater fractionation found in frost flowers is consistent in samples collected in different years and on opposite sides of the Antarctic continent. We discuss briefly the potential of satellite-borne radar scatterometers to give estimates of flower production, and look at the implications that flowers may have for tropospheric chemistry and ice core interpretation.

2. Sampling Sites and Methods

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Sampling Sites and Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusions
  8. Acknowledgments
  9. References

[13] Our samples are from two sites: the Weddell Sea near the coastal Antarctic station Halley (75°S, 26°W, where samples were taken over 2 consecutive years), and near the Mertz Glacier (67°S, 145°E). Figure 3 shows the location of the sample sites and other locations mentioned in the text. Samples at Halley were collected in the austral spring, while those at Mertz Glacier were collected in the middle of winter; but as frost flower formation depends only on the temperatures and wind conditions prevailing at the time we would not expect any differences between the sample sets for this reason.

image

Figure 3. Map of Antarctica showing sites mentioned in the text.

Download figure to PowerPoint

2.1. Weddell Sea Samples 1998

[14] Frost flowers were collected on the 2 October 1998 from a site in the Weddell Sea near the British Antarctic Survey station Halley. The monthly mean surface air temperature near Halley is around −5°C in January and drops to −30°C in August, and thus fractionation may be expected during the winter months at the sea ice surface [Richardson, 1976], which will have a temperature somewhere between that of the air and that of the seawater.

[15] The site chosen was an extensive area of new sea ice, 90% covered with frost flowers. The ice was approximately 15 cm thick and uniform, and an open water lead was present 20 m to the north of the sample site. The flowers were approximately 3 cm tall, and appeared similar to thick rime. A metal spatula was used to push a number of flowers into a plastic sample tube. The sample tube was kept frozen at −20°C until analysis in Cambridge around 1 year later. For comparison with the frost flowers, two samples of surface seawater collected from RRS Bransfield in the Weddell Sea, approximately 250 km from Halley, were also analyzed. These results were previously presented by Rankin et al. [2000].

2.2. Weddell Sea Samples 1999

[16] Several samples of frost flowers and surface slush were taken between 29 September and 3 October 1999 from sites near Halley station. Samples were taken from sea ice less than a week old, and were collected under similar conditions to the previous year's samples.

2.3. Mertz Glacier Samples

[17] Samples of frost flowers, surface slush, sea ice and seawater were collected from the Australian research vessel “Aurora Australis” between 22 and 24 August 1999. The location of the vessel was around 67°S, 145°E, off the Mertz glacier on the opposite side of Antarctica to Halley station.

2.4. Analysis

[18] The melted frost flowers, slush, ice and seawater samples were diluted by a factor of 103 or 104 with ultrahigh purity water, and analyzed on Dionex ion chromatographs. Anions were analyzed using gradient elution with sodium hydroxide or potassium hydroxide eluent, and cations were analyzed with isocratic methane sulfonic acid (MSA) eluent. Accuracies for ion chromatography measurements are typically around ±5%. However, some of the bromide measurements were close to the detection limits and have slightly higher uncertainties of around ±10%.

3. Results

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Sampling Sites and Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusions
  8. Acknowledgments
  9. References

[19] Figure 4 shows the sample salinities. These are calculated from the sum of all the ions measured with the ion chromatographs, rather than by the conductivity measurements normally used to calculate salinity by oceanographers. However, the two methods should give similar results [Riley and Chester, 1971]. The flower salinities are typically around three times that of seawater, consistent with previous measurements such as those of Perovich and Richter-Menge [1994]. The single sample collected near Halley in 1998 has a much lower salinity; although we cannot rule out an analytical error, one possible explanation is that the flowers were covered in snow, diluting the salt concentration. Perovich and Richter-Menge [1994] noted that the salinity of the patch of frost flowers they investigated evolved over time, and suggested collection of blowing snow by the flowers as one reason for the changes. Collection of snow on a frost flower cover field was also observed by Ulander et al. [1995].

image

Figure 4. Salinities of samples of frost flowers (dark gray), slush (white), ice (black), and seawater (light gray) analyzed. The line shows the salinity of Standard Mean Ocean Water.

Download figure to PowerPoint

[20] The slush samples also have salinities considerably higher than seawater, and the ice samples are less saline than seawater, as we would expect from the expulsion of brine during the sea ice production process.

[21] Figure 5 shows the sodium/sulfate weight ratios for the samples, and for a standard seawater sample. The horizontal line shows the standard seawater ratio of 0.25. The figure shows that the flower samples are consistently depleted in terms of sulfate, with characteristic values of 0.1. In contrast, the slush values are highly variable, ranging from 0.1 to greater than 0.5. The sea ice samples from near the Mertz glacier show a slight but similar variability.

image

Figure 5. Sulfate to sodium ratios in samples analyzed. The lines are the values found in Standard Mean Ocean Water (solid) and in winter aerosol at Halley (dashed).

Download figure to PowerPoint

[22] For the flowers, the samples analyzed in this experiment show similar fractionation to that found in the first sample analyzed, with consistent loss of sulfate relative to sodium. The samples taken in both years near Halley and those taken from near the Mertz glacier on the other side of Antarctica all show the same depletion, demonstrating that the effect is temporally and spatially consistent. Typically the sulfate/sodium weight ratio in flowers is between 0.05 and 0.10, compared to a seawater ratio of 0.25. Fractionation in all samples is similar to that observed in aerosol samples at Halley base in winter [Wagenbach et al., 1998] and in winter layers in ice cores [Gjessing, 1989].

[23] Bromide concentrations, measured for the first time in frost flowers, are given in Figure 6. Only the samples collected from near the Mertz glacier were analyzed fully for bromide and data are presented for this set; a few samples from those collected near Halley in 1999 were also analyzed and show similar results. The samples have concentrations of bromide around three times those found in seawater, in line with the increased salinities of the samples. The ratios of Br/Na are similar to those found in seawater. This effect is independent of the segregation described above, and is determined by the concentrating effect of the cold temperatures: the enhanced Br levels are a consequence of sea-salt concentration only.

image

Figure 6. Bromide ion concentrations in samples taken from near the Mertz glacier.

Download figure to PowerPoint

4. Discussion

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Sampling Sites and Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusions
  8. Acknowledgments
  9. References

4.1. Scatterometer Images and Aerosol Chemistry

[24] In order to assess the extent to which frost flowers may be important in affecting atmospheric chemistry in the polar regions, it is important to have an idea of how widespread they are. There are few data available. However, most sea ice in the Antarctic thaws and refreezes every year, and even in mature sea ice new leads are constantly opening and refreezing. The fractionation process can be expected to occur wherever the ice surface temperatures are below −8°C, likely to be the case for almost all sea ice production during the winter in Antarctic seas. Frost flowers typically last for a number of days before being blown away or obscured by snow [Perovich and Richter-Menge, 1994].

[25] The greatest areas of new ice production are often found in coastal leads when offshore breezes blow the ice away from the coast. Although no rigorous studies have been made of the extent to which frost flowers are present at the coast, eyewitness accounts suggest that in the area around Halley, frost flowers may be present any time during the winter from early March to October, and at times can cover new sea ice as far as the eye can see (R. Ladkin, personal communication, 2002). The scenario is likely to be typical for the coast around the whole Antarctic continent, where offshore winds predominate.

[26] Correlation between QSCAT Images and Aerosol Filters: One tool that may be used to gain a better idea of frost flower coverage is images from satellite-borne radar scatterometers. Because of the accompanying slush layers and the rough surface they present to the radar beam, frost flowers are good scatterers. Although old sea ice with pressure ridges can also give a bright return, frost flowers can be distinguished by the speed with which they form; a frost flower bloom forms within 24 hours [Nghiem et al., 1997], and will last only a few days in sequential images, while the signal from older ice evolves over a period of months [Hallikainen and Winebrenner, 1992]. Ulander et al. [1995] compared backscatter images with concurrent surface observations of areas of frost flowers and surface slush from a cruise in the Arctic, noting a rapid, strong increase in return from the surface as the new ice formed in good agreement with model predictions.

[27] For days 130–300 of the year 2000, Figure 7 shows the time series when bright patches, indicative of frost flowers, are present in scatterometer images of the coastal waters around Halley. The six periods when frost flowers are present are marked by the horizontal black bars and labeled A to F. The figure also shows the times of filter changes on a high volume air sampler that operates year-round at Halley. For filters depleted in sulfate relative to sodium, the sampling period is shaded in gray. Sodium concentrations in the air derived from the filters are shown as a bar chart above the time series.

image

Figure 7. Comparison of the times of filter changes (vertical lines) at an aerosol sampling instrument at Halley with the periods of frost flower blooms (black horizontal bars) observed in QSCAT images of the sea ice around Halley. The gray areas indicate filters that are depleted in sulfate; the letters identify different periods of frost flower formation. Sodium concentrations derived from the filters are shown as a bar chart above the time series.

Download figure to PowerPoint

[28] As the figure shows, the average concentration of sodium during periods of negative non-sea-salt sulfate (nss-SO4) in the period studied was considerably higher than during periods of positive nss-SO4. Although fractionated aerosol is present less often than unfractionated aerosol, it may nonetheless dominate the annual sea-salt budget.

[29] For each depleted filter, the figure shows that at some point during the run there is a patch of frost flowers nearby. However, the reverse is not true: there are no depleted filters associated with flower episodes B and E, and other episodes extend into filter sampling periods when again the filters show no sulfate depletion.

[30] To understand why a particular episode leads to depleted aerosol being present at Halley, it is necessary to look at the meteorology of the frost flower episodes. For new sea ice to form, open water must be present, which occurs most often when winds are offshore, creating a shore lead. However, for aerosols generated from frost flowers to be blown inland, the wind direction must change to blow onshore. This sequence of offshore and onshore winds is consistent with the previous work of Hall and Wolff [1998], who studied periods with high sodium content on daily aerosol filters at Halley. Over a period of 2 years they identified at least eight occasions where a change in the wind direction from easterly to westerly was followed by high fractionated sea-salt aerosol loadings, and surmised that this corresponded to periods of new sea ice formation (when offshore winds created new shore leads that then froze over) followed by transport of frost flowers towards Halley.

[31] As a demonstration that these meteorological conditions can occur and transfer aerosols from the ice to Halley, Figure 8 shows the sequence of scatterometer images from the NASA QSCAT satellite scatterometer for the area around Halley during episode C (days 197–202), along with back trajectories arriving at the station during that period. Figure 9 shows the same information for episode E (days 253–263). The back trajectories were run on the Web interface to the British Atmospheric Data Centre back trajectory model, which utilizes ECMWF reanalysis data.

image

Figure 8. Daily maps of back trajectories arriving at Halley and daily scatterometer images covering the same area as the maps during frost flower bloom C on Figure 7 (days 197–202). On the scatterometer images, the land and ice shelves are masked in gray. The back trajectories are plotted versus their arrival times at 0000 (•), 0600 (○), 1200 (▵), and 1800 (□), with the symbols along the trajectories showing the air parcel position at hourly intervals. During this period, a patch of frost flowers that generates a bright scatterometer return can be seen forming and decaying in Precious Bay (marked by the white arrow) to the southwest of Halley.

Download figure to PowerPoint

image

Figure 9. Daily maps of back trajectories arriving at Halley and daily scatterometer images taken during frost flower bloom E (days 253–263). During this period, frost flowers cover a large area of the sea to the north and west of Halley.

Download figure to PowerPoint

image

Figure 9. (continued)

Download figure to PowerPoint

[32] For episode C, Figure 7 shows that the filter removed on day 200 is depleted in sulfate relative to sodium, while the following filter is not. To explain this difference in filter properties, on day 198, the white area in Precious Bay indicates a transient region of enhanced backscatter, such as would be generated by frost flowers. This region extends outwards and becomes brighter on days 199 and 200. For this period, the back trajectories show that on days 198 and 199, material from this region would be transported to the station. This result is consistent with the fact that the filter removed on day 200 is sulfate depleted. On day 200, back trajectories show the air comes from continental areas and by day 201 the bloom has faded. This result is consistent with the filter of days 200–205 not showing sulfate depletion.

[33] Frost flower episode E, by contrast, has no sulfate-depleted filters associated with it; although, as the sequences of images in Figure 9 show, the bloom covers a large area and lasts considerably longer than episode C. Back trajectories for days 253–260 all show air coming from the east, and although on day 261 the winds become southwesterly and pass over Precious Bay, that region by then is free of frost flowers. On days 262–263, the wind veers to the north, briefly passing over the region where frost flowers are still present. By day 262, the frost flower patch has begun to fade, possibly suggesting that the most mobile parts of the flowers have already been blown away, or that the flowers have become covered with snow. The patch is also further away from Halley than that in bloom C, allowing more aerosol that has been generated from the flowers to be lost through deposition. It is therefore perhaps not surprising that insufficient aerosol is generated during the brief sweep of the winds over the frost flower patch to dominate the aerosol collected in the filter that ran from days 249 to 263.

[34] Similar conclusions can be drawn by looking at the other episodes. Although we cannot say that the presence of frost flowers near Halley will necessarily lead to filters collected there being depleted in sulfate, it does appear to be the case that sulfate-depleted filters are only collected when two conditions are satisfied; first that frost flowers are present nearby; second that meteorological conditions are favorable to generate aerosol from the frost flowers and transport it inland.

4.2. Tropospheric Bromine Chemistry

[35] There has been considerable interest in tropospheric bromine chemistry recently due to its role in ozone depletion during the polar spring. Observations at the Canadian Arctic station Alert show depletion events where ozone is reduced to near-zero concentrations in the boundary layer in a matter of hours [Barrie et al., 1988], and observations of bromine collected on cellulose filters show a clear anticorrelation with ozone concentrations at these times. The depletion has been observed at other locations in the Arctic, for example at Point Barrow in Alaska [Sturges et al., 1993] and Spitzbergen [Solberg et al., 1996], and it has also been recently seen in the Antarctic [Kreher et al., 1997; Wessel et al., 1998]. Aircraft flights during one ozone depletion event in the Arctic found the depleted layer was between 200 and 400 m thick [Hopper et al., 1998].

[36] In its simplest form, the ozone destruction cycle is the reaction of Br + O3 to give BrO + O2, followed by the reaction of BrO + BrO to return 2 Br + O2 (or Br2 + O2 followed by the rapid photolysis of Br2 to 2 Br). The rate limiting step is normally the BrO + BrO reaction. Other pathways may be involved: BrO may react with NO2 or HO2 to give BrONO2 or HOBr respectively, but both these products are rapidly photolyzed to return Br atoms. Some Br is lost by reaction with oxidized organics or with the HO2 radical, giving HBr [Fickert et al., 1999].

[37] To sustain the rates of reaction needed for the observed ozone depletions an efficient initial source of bromine is required. Reactions involving gas phase reservoirs, as originally proposed, are insufficient to account for the bromine observed [Barrie et al., 1988]. A number of heterogeneous reactions have therefore been put forward. Fan and Jacob [1992] suggest the reaction of HBr with HOBr in sea-salt particles as a source of Br2, and this has been further investigated by Fickert et al. [1999]. The Br2 may diffuse from the solution and be subsequently photolyzed to provide a source of Br atoms in the atmosphere. Mozurekewich [1995] proposed various reactions involving free radicals and peroxymonosulfuric acid that could produce bromine, and Oum et al. [1998] demonstrated that the reaction of ozone with seawater ice was also a mechanism by which Br2 could be produced. The heterogeneous release of Br was quantitatively modeled by Vogt et al. [1996].

[38] In all these mechanisms both the salt concentration in the liquid phase, and the surface area over which reactants may diffuse into the solution and products diffuse out, will be important factors in determining the rate of bromine production. Potential sources of salt include the sea surface itself, sea-salt aerosol, both in the air and after deposition in the snowpack, and frost flowers.

[39] We find that bromide ions are present in high concentrations in frost flowers and samples of surface brine, as shown in Figure 6, with concentrations typically two to three times those found in seawater. This is in line with the high salinity of these samples (Figure 4). Although other ions are fractionated during the formation of frost flowers, this is not expected to be the case for bromide ions [Koop et al., 2000], and is indeed not seen in our samples: the Br concentration is proportional to the salinity of the samples.

[40] Studies of the surfaces of mixed NaBr/NaCl crystals grown from aqueous solution have shown that bromide ions are segregated to the surface [Zangmeister et al., 2001]: this occurs for crystals grown from the melt also [Ghosal et al., 2000]. It is possible therefore that if temperatures are low enough in the brine or flowers that NaCl also freezes out, the bromide may be more available.

[41] Although surface areas of frost flowers have not to our knowledge been measured, the specific surface area of snow has been measured both by nitrogen and methane adsorption techniques [Hoff et al., 1998; Hanot and Domine, 1999] and by photographic techniques [Fassnacht et al, 1999]. Measurements range between 0.182 and 2.25 m2 g−1 for fresh snow, decreasing with age; measurements of older snow gave 0.06–0.37 m2 g−1 in one study, and 0.25 m2 g−1 in another. The crystalline nature of frost flowers is similar to snow; until measurements of their surface area are made, values obtained from fresh snow are the best estimates available.

[42] Perovich and Richter-Menge [1994] estimated the mass per unit area of frost flowers on a newly frozen lead to be 0.025–0.050 g cm−2. If we estimate the surface area of the frost flowers to be similar to fresh snow, i.e., between 0.2 and 2 m2 g−1, this gives a surface area of the frost flowers relative to the area they cover of between 50 and 1000 m2 m−2, probably nearer the upper limit given the delicate nature of the frost flower crystals.

[43] It is not yet known accurately what fraction of the sea surface in polar regions is covered with frost flowers at any one time. However, the majority of sea ice in the Antarctic melts and reforms each year, and wherever there is new ice formation we can expect frost flowers to grow except where winds are too strong, typically existing for a few days before being blown away or covered in snow. If we assume all sea ice refreezes during a period of 180 days in the winter, and the flowers last on average 5 days this gives an very rough estimate that on average 2.7% of the surface is covered at any one time. This area is likely to be concentrated near the coast where sea ice production is highest due to the offshore winds. It therefore seems likely that during winter, frost flowers present a considerably greater surface area over which exchange may take place than would occur for open seawater.

[44] It has been suggested that sea-salt aerosols could be a source of bromine to the atmosphere [Finlayson-Pitts et al., 1990; Mozurekewich, 1995]. Tang and McConnell [1996], however, point out there is insufficient bromide contained in airborne sea-salt aerosol to account for the observed concentration of reactive bromine species. They suggest that aerosols that have been deposited and incorporated in the snowpack may be a source. However, while the surface area of snow in the snowpack is very large (Waddington et al. [1996] suggest a value of 1500 m2 per m2 of cover for a 10 cm layer of snow, which might be a typical depth of snow accumulated on sea ice), the concentration of bromide in snow is very small. Legrand et al. [1998] find a mean value of 530 ng g−1 of sodium in surface snow at Neumayer station in coastal Antarctica, and assuming the ratio of sodium to bromide to be 161 (that of bulk seawater) [Fegley, 1995], the concentration of bromide is only 3.29 ng g−1, a factor of over 5 × 104 less than that found in the frost flowers we analyzed.

[45] Further work is needed to distinguish between these two potential bromine sources; at the moment there is insufficient data on the true surface area of frost flowers, their coverage, and the dependence of the reactions that produce reactive bromine from sea salt on concentration and surface area to determine the contribution they may be able to make.

[46] The observation however that ozone depletion is associated with the passage of air masses over areas of sea ice [Hopper, 1998] lends support to the hypothesis that flowers are a source of the bromine. Satellite measurements of tropospheric BrO from GOME also show that enhanced BrO is associated with sea ice [Richter et al., 1998; Wagner et al., 2001], and that in the Arctic the regions with high levels move north as the sea ice retreats. Release from aerosol in the snowpack might also be expected to be slow and sustained in nature; the rapid changes in filterable bromine observed are more consistent with the episodic nature of frost flower blooms.

4.3. Ice Core Interpretation

4.3.1. Sodium in Ice Cores

[47] Sea salt is a major constituent of the aerosol present in the Antarctic atmosphere, particularly in the coastal regions. Sodium in Antarctic aerosol is thought to have no other significant source, and after deposition to the snowpack, sodium behaves conservatively [Wolff, 1996]. Sodium concentrations retrieved from ice cores have therefore been used to infer changes in the amount of sea salt reaching the snowpack.

[48] In the past, sea salt has been assumed to come only from open water. Higher levels of sea salt measured in aerosol during the winter have therefore been attributed to increased storminess, leading to higher aerosol production and more efficient transport inland [e.g., Curran et al., 1998]. High levels of sea salt in ice cores from glacial periods have also been attributed to more stormy weather [Petit et al., 1981; Delmas, 1992], and even increases during the Little Ice Age have been attributed to changes in weather patterns [Peel and Mulvaney, 1992; Fischer et al., 1998]. This is despite the fact that sea ice cover increases during cold periods, increasing the distance aerosol must travel from open water to reach the ice sheets.

4.3.2. Sulfate in Ice Cores

[49] Sulfate aerosol is also incorporated into snow, and sulfate concentrations are routinely retrieved from ice cores. A proportion of this sulfate can be attributed to sea salt, but volcanoes, atmospheric oxidation of biogenically produced methane sulfonic acid, and, on more recent time scales, anthropogenic emissions, all also contribute to the sulfate aerosol burden. It is useful to be able to differentiate between sea-salt and non-sea-salt sulfate, and in the past the sea-salt fraction has been calculated by assuming all sodium in the aerosol or ice core to come from sea salt, and multiplying the sodium concentration by the ratio of sulfate/sodium in bulk seawater to give the sea-salt component. The total sulfate concentration minus the sea-salt component is designated nss-SO4.

[50] An unresolved problem in recent years has been that precipitation falling in the winter at coastal Antarctic sites frequently has negative nss-SO4 concentrations [Legrand and Delmas, 1985; Gjessing, 1989; Minikin et al., 1994]. Wagenbach et al. [1998] were the first to suggest that fractionated brine on the surface of newly forming sea ice might be a significant source of sea salt in the winter, leading to the depleted sulfate/sodium ratios observed.

4.4. Dolleman Core

[51] Data from an ice core drilled on Dolleman Island, Antarctica and sampled at approximately monthly resolution have been plotted according to the number of data points with a particular sodium and sulfate concentration (Figure 10). The figure shows that the greatest density of data points lies in a mode centered around 4 microequivalents sulfate, 3 microequivalents sodium. This mode lies to the right of the line indicating the ratio between sulfate and sodium in bulk seawater. This is consistent with an added input of sulfate from marine biogenic sources, as seen in aerosol in coastal locations in the summer [Minikin et al., 1998].

image

Figure 10. Density of samples with particular sodium and sulfate concentrations in an ice core from Dolleman Island. The line shows the ratio between sodium and sulfate in bulk seawater.

Download figure to PowerPoint

[52] There is, however, a considerable number of data points with sulfate/sodium ratios less than that of seawater. These are consistent with the fractionation seen in frost flowers, and it seems likely that this mode is from winter aerosol generated from frost flowers on sea ice. Of course, even a few data points with sufficiently high sodium concentrations may contribute more to the total sodium in the core than a large number of data points of low sodium concentration, so if instead of plotting merely the number of data points with given sodium and sulfate concentrations the number of data points multiplied by their sodium concentration is plotted, an idea of the distribution of salt between fractionated and unfractionated modes may be gained. This is shown in Figure 11, with a similar plot for sulfate in Figure 12.

image

Figure 11. Contour plot of the density of data points with given sodium and sulfate concentrations multiplied by their sodium concentrations, indicating the contribution of fractionated and unfractionated samples to the total sodium in the Dolleman core.

Download figure to PowerPoint

image

Figure 12. Contour plot of the density of data points with given sodium and sulfate concentrations multiplied by their sulfate concentrations, indicating the contribution of fractionated and unfractionated samples to the total sulfate in the Dolleman core.

Download figure to PowerPoint

[53] While it is clear that by far the majority of the sulfate in the core is contained in the unfractionated “summer” mode, most of the sodium is contained in the fractionated “winter” mode. In fact, although only 13.1% of the core samples have sulfate to sodium ratios less than that of seawater (indicative of the aerosol having a frost flower source), these samples account for 39.1% of the sodium in the core, and note that this will underestimate sodium due to frost flowers, as samples with salt from frost flowers but also some additional biogenic or volcanic sulfate may have ratios above that of seawater.

[54] That this is likely to be the case for many samples can be seen by considering changes in sodium concentration during fractionation, as aerosol from frost flowers should be depleted in sodium as well as sulfate. As the Introduction discusses, if all the available sulfate in the brine solution precipitates as Na2SO4, prior to frost flower formation, the sodium in aerosol from the flowers should depleted by 13%.

[55] Figure 13 is a ternary plot of sulfate, sodium and chloride in samples from the Dolleman core. Values have been normalized such that a sample having the same ratios of sulfate, sodium, and chloride as seawater will lie in the center of the plot, i.e.,

  • equation image

where [X] is the concentration of an ion in an ice core sample in microequivalents per liter.

image

Figure 13. Ternary plots of sulfate, sodium, and chloride in the Dolleman Island core, normalized such that a sample with the same ratios of Na, Cl, and SO4 as seawater would be plotted in the center of the plot. The inset shows only data points with high Na and Cl concentrations. Also shown on the plots are the values expected from seawater aerosol, from frost flower aerosol completely depleted in sulfate and partially in sodium, and from the two types of aerosol with additional biogenic sulfate.

Download figure to PowerPoint

[56] Aerosol from frost flowers with sulfate completely lost as Na2SO4 would lie at the point 46.6% Na, 53.4% Cl, and 0% SO4. It can be seen that most of the samples lie along a line between this point and the sulfate vertex. Points along this line are consistent with fractionated aerosol with additional sulfate, for example from marine biogenic sources. At least at high salt concentrations, even when nss-SO4 is positive the line is a better fit than the line for nonfractionated aerosol, which passes through the center of the graph. At high sulfate concentrations other processes may affect the data, such as the reaction of sulfuric acid with sea salt to produce volatile HCl. It therefore seems likely that the majority of sea salt in the Dolleman core is from frost flowers, and not unfractionated sea salt.

[57] If this finding is generally true for near-coastal Antarctic cores it suggests that changes in the strength of the sea-salt signal may not be indicative of changes in transport from patches of open water, as previously believed, but instead may be following changes in sea ice production. In coastal regions, this probably corresponds to the production in a limited local area. If the finding also applies further inland, the sodium concentrations may reflect sea ice production in a large region. However, it remains to be seen whether this finding is confined to coastal areas, or whether it applies also to central Antarctic cores.

[58] Unfortunately, it is difficult to see whether the same effect occurs in deep inland cores. At such locations snow accumulation rates are low, so annual cycles cannot be seen. Also, the relative contribution of marine biogenic sulfate is greater as the small particles are transported more efficiently inland than large sea-salt particles. Over a period of a year (a typical amount of accumulation used for each sample in a deep core) any negative nss-SO4 sulfate signal due to frost flowers is likely to be swamped by the positive contribution from biogenic sulfate.

[59] One question that remains to be answered is whether the size distribution of aerosol generated from frost flowers is different from conventional sea-salt aerosol produced by bubble bursting at the air–sea interface, a process that is reasonably well characterized and modeled [e.g., Gong et al., 1997a, 1997b]. No work has yet been done on the entrainment of material from frost flowers into the atmosphere, although given the fragile nature of the flowers mechanical weathering is expected to be an efficient means of generating aerosol. If this produces larger particles than those from seawater, they may not be transported long distances before being deposited. Therefore, although frost flowers appear to be the dominant source of sea salt in coastal sites, this is not necessarily the case further inland.

5. Conclusions

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Sampling Sites and Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusions
  8. Acknowledgments
  9. References

[60] The vertical transport of the interior brine to the surface within newly forming sea ice, and the frost flowers that grow from this brine, yield enhanced concentrations of seawater ions such as bromide, and a depletion in sulfate relative to sodium.

[61] Frost flower blooms can be seen in satellite-borne radar scatterometer images, and are frequently present when offshore winds create new leads near the coast. The presence of aerosol depleted in sulfate at Halley station near the Antarctic coast appears to be correlated with back trajectories arriving from areas where flowers are present.

[62] The bromide concentration in frost flowers is around 3 times higher than in seawater. The combination of high concentrations of bromide present in the liquid phase, and the large surface area presented for exchange, means that frost flowers should be considered as a potential source of the bromine implicated in tropospheric ozone depletion in polar regions.

[63] In an ice core from Dolleman Island, Antarctica, most of the sea salt appears to be fractionated with the loss of mirabilite, consistent with frost flowers rather than open water being the source. We suggest that previous interpretations of high sea salt in ice cores being due to increased storminess or changes in weather patterns may be incorrect: if the majority of sea salt is from frost flowers rather than open water, increased sea salt may rather be an indicator of increased sea ice production.

[64] Few studies have so far been carried out on frost flowers, and a great deal more could be learned about them. It should be possible to devise algorithms to identify frost flowers on scatterometer images, and thus gain a better idea of the area they cover and how long they last; at present the interpretation of images is rather subjective. Measurements of the flower surface area would assist in understanding their role in production of bromine. Finally, the transport of aerosol from frost flower-covered areas to the continental interior needs to be better understood; measurements of the size distribution of aerosol would assist in this, as would modeling of losses through deposition along back trajectories arriving at ice core drilling sites high on ice caps.

Acknowledgments

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Sampling Sites and Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusions
  8. Acknowledgments
  9. References

[65] We would like to thank Victoria Lytle and her group at the University of Tasmania for collecting frost flower samples for us during a cruise on the Aurora Australis. The satellite scatterometer images were obtained from the NASA-sponsored Scatterometer Climate Record Pathfinder at Brigham Young University through the courtesy of David G. Long, and we are grateful to Mark Drinkwater for help with their interpretation. We thank the British Atmospheric Data Centre for providing access to calculated trajectories using data from the European Centre for Medium Range Weather Forecasts, Anna Jones for her comments on the manuscript, and Nick McWilliam for assistance with mapping the trajectory data. SM gratefully acknowledges the support of the U.S. National Science Foundation under OCE9811097 and of NASA under grant NAG5-11067.

References

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Sampling Sites and Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  • Barrie, L. A., J. W. Bottenheim, R. C. Schnell, P. J. Crutzen, and R. A. Rasmussen, Ozone destruction and photochemical reactions at polar sunrise in the lower Arctic atmosphere, Nature, 334, 138141, 1988.
  • Curran, M. A. J., T. D. van Ommen, and V. Morgan, Seasonal characteristics of the major ions in the high-accumulation Dome Summit South ice core, Law Dome, Antarctica, Ann. Glaciol., 27, 385390, 1998.
  • Delmas, R. J., Environmental information from ice cores, Rev. Geophys., 30(1), 121, 1992.
  • Drinkwater, M. R., and G. B. Crocker, Modelling changes in the dielectric and scattering properties of young snow-covered sea ice at GHz frequencies, J. Glaciol., 34, 274282, 1988.
  • Fan, S.-M., and D. J. Jacob, Surface ozone depletion in Arctic spring sustained by bromine reactions on aerosols, Nature, 359, 522524, 1992.
  • Fassnacht, S. R., J. Innes, N. Kouwen, and E. D. Soulis, The specific surface area of fresh dendritic snow crystals, Hydrol. Processes, 13(18), 29452962, 1999.
  • Fegley, B., Properties and composition of the terrestrial oceans and of the atmospheres of the Earth and other planets, in Global Earth Physics: A Handbook of Physical Constants, edited by T. J. Ahrens, pp. 320345, AGU, Washington, D. C., 1995.
  • Fickert, S., J. W. Adams, and J. N. Crowley, Activation of Br2 and BrCl via uptake of HOBr onto aqueous salt solutions, J. Geophys. Res., 104, 23,71923,727, 1999.
  • Finlayson-Pitts, B. J., F. E. Livingston, and H. N. Berko, Ozone destruction and bromine photochemistry at ground level in the arctic spring, Nature, 343, 622624, 1990.
  • Fischer, H., M. Werner, D. Wagenbach, M. Schwager, T. Thorsteinnson, F. Wilhelms, J. Kipfstuhl, and S. Sommer, Little ice age clearly recorded in northern Greenland ice cores, Geophys. Res. Lett., 25, 17491752, 1998.
  • Ghosal, S., A. Shbeeb, and J. C. Hemminger, Surface segregation of bromine in bromide doped NaCl: Implications for the seasonal variation in Arctic ozone, Geophys. Res. Lett., 27, 18791882, 2000.
  • Gjessing, Y., Excess and deficit of sulfate in polar snow, Atmos. Environ., 23, 155160, 1989.
  • Gloerson, P., and W. J. Campbell, Recent variations in Arctic and Antarctic sea-ice covers, Nature, 352, 3336, 1991.
  • Gong, S., L. A. Barrie, and J.-P. Blanchet, Modeling sea-salt aerosols in the atmosphere, 1, Model development, J. Geophys. Res., 102, 38053818, 1997a.
  • Gong, S., L. A. Barrie, J. M. Prospero, D. L. Savoie, G. P. Ayers, J.-P. Blanchet, and L. Spacek, Modeling sea-salt aerosols in the atmosphere, 2, Atmospheric concentrations and fluxes, J. Geophys. Res., 102, 38193830, 1997b.
  • Hall, J. S., and E. W. Wolff, Causes of seasonal and daily variations in aerosol sea-salt concentrations at a coastal Antarctic station, Atmos. Environ., 32, 36693677, 1998.
  • Hallikainen, M., and D. P. Winebrenner, The physical basis for sea ice remote sensing, in Microwave Remote Sensing of Sea Ice, edited by F. Carsey, pp. 2946, AGU, Washington, D. C., 1992.
  • Hanot, L., and F. Domine, Evolution of the surface area of a snow layer, Environ. Sci. Technol., 33(23), 42504255, 1999.
  • Hoff, J. T., D. Gregor, D. Mackay, F. Wania, and C. Q. Jia, Measurement of the specific surface area of snow with the nitrogen adsorption technique, Environ. Sci. Technol., 32(1), 5862, 1998.
  • Hopper, J. F., L. A. Barrie, A. Silis, W. Hart, A. J. Gallant, and H. Dryfhout, Ozone and meteorology during the 1994 Polar Sunrise Experiment, J. Geophys. Res., 103, 14811492, 1998.
  • Koop, T., A. Kapilashrami, L. T. Molina, and M. J. Molina, Phase transitions of sea-salt/water mixtures at low temperatures: Implications for ozone chemistry in the polar marine boundary layer, J. Geophys. Res., 105, 26,39326,402, 2000.
  • Kreher, K., P. V. Johnson, S. W. Wood, B. Nardi, and U. Platt, Ground-based measurements of tropospheric and stratospheric BrO at Arrival Heights, Antarctica, Geophys. Res. Lett., 22, 599602, 1997.
  • Legrand, M. R., and R. J. Delmas, Spatial and temporal variations of snow chemistry in Terre Adelie (Antarctica), Ann. Glaciol., 7, 2025, 1985.
  • Legrand, M., F. Ducroz, D. Wagenbach, R. Mulvaney, and J. S. Hall, Ammonium in coastal Antarctic aerosol and snow: Role of the polar ocean and penguin emissions, J. Geophys. Res., 103, 11,04311,056, 1998.
  • Martin, S., R. Ducker, and M. Fort, A laboratory study of frost flower growth on the surface of young sea ice, J. Geophys. Res., 100, 70277036, 1995.
  • Martin, S., Y. Yanling, and R. Drucker, The temperature dependence of frost flower growth on laboratory sea ice and the effect of the flowers on infrared observations of the surface, J. Geophys. Res., 101, 12,11112,125, 1996.
  • Minikin, A., D. Wagenbach, W. Graf, and J. Kipfstuhl, Spatial and seasonal variations of the snow chemistry at the central Filchner-Ronne Ice Shelf, Antarctica, Ann. Glaciol., 20, 283290, 1994.
  • Minikin, A., M. Legrand, J. Hall, D. Wagenbach, C. Kleefeld, E. Wolff, E. C. Pasteur, and F. Ducroz, Sulfur-containing species (sulfate and methanesulfonate) in coastal Antarctic aerosol and precipitation, J. Geophys. Res., 103, 10,97510,990, 1998.
  • Mozurekewich, M., Mechanisms for the release of halogens from sea-salt particles be free radical reactions, J. Geophys. Res., 100, 14,19914,207, 1995.
  • Nghiem, S. V., S. Martin, D. K. Perovich, R. Kwok, R. Drucker, and A. J. Gow, A laboratory study of the effect of frost flowers on C band radar backscatter from sea ice, J. Geophys. Res., 102, 33573370, 1997.
  • Onstott, R. G., SAR and scatterometer signatures of sea ice, in Microwave Remote Sensing of Sea Ice, edited by F. Carsey, pp. 73104, AGU, Washington, D. C., 1992.
  • Oum, K. W., M. J. Lakin, and B. J. Finlayson-Pitts, Bromine activation in the troposphere by the dark reaction of O3 with seawater ice, Geophys. Res. Lett., 25, 39233926, 1998.
  • Peel, D. A., and R. Mulvaney, Time-trends in the pattern of ocean–atmosphere exchange in an ice core from the Weddell Sea Sector of Antarctica, Tellus, Ser. B, 44, 430442, 1992.
  • Perovich, D. K., and J. A. Richter-Menge, Surface characteristics of lead ice, J. Geophys. Res., 99, 16,34116,350, 1994.
  • Petit, J. R., M. Briat, and A. Royer, Ice age aerosol content from East Antarctic ice core samples and past wind strength, Nature, 293, 391394, 1981.
  • Rankin, A. M., E. W. Wolff, and V. Auld, Frost Flowers as a source of fractionated sea salt aerosol in the polar regions, Geophys. Res. Lett., 27, 34693472, 2000.
  • Richardson, C., Phase relationships in sea ice as a function of temperature, J. Glaciol., 17, 507519, 1976.
  • Richter, A., F. Wittrock, M. Eisinger, and J. P. Burrows, GOME observations of tropospheric BrO in Northern Hemispheric spring and summer 1997, Geophys. Res. Lett., 25, 26832686, 1998.
  • Riley, J. P., and R. Chester, Salinity, chlorinity and the physical properties of seawater, in Introduction to Marine Chemistry, pp. 1136, Academic, San Diego, Calif., 1971.
  • Solberg, S., N. Schmidtbauer, A. Semb, and F. Stordal, Boundary-layer ozone depletion as seen in the Norwegian Arctic in spring, J. Atmos. Chem., 23, 301332, 1996.
  • Sturges, W. T., R. C. Schnell, S. Landsberger, S. J. Oltmans, J. M. Harris, and S. M. Li, Chemical and meteorological influences on surface ozone destruction at Barrow, Alaska, during spring 1989, Atmos. Environ., 27, 28512863, 1993.
  • Tang, T., and J. C. McConnell, Autocatalytic release of bromine from Arctic snow pack during polar sunrise, Geophys. Res. Lett., 23, 26332636, 1996.
  • Tuckermann, M., R. Ackermann, C. Gölz, H. Lorenzen-Schmidt, T. Senne, J. Stutz, B. Trost, W. Unold, and U. Platt, DOAS-observation of halogen radical-catalysed Arctic boundary layer ozone destruction during the ARCTOC campaigns 1995 and 1996 in Ny-Alesund, Spitsbergen, Tellus, Ser. B, 49, 533555, 1997.
  • Ulander, L. M. H., A. Carlstrom, and J. Askne, Effect of frost flowers, rough saline snow and slush on the ERS-1 SAR backscatter of thin arctic sea ice, Int. J. Remote Sens., 16(17), 32873305, 1995.
  • Vogt, R., P. J. Crutzen, and R. Sander, A mechanism for halogen release from sea-salt aerosol in the remote marine boundary layer, Nature, 383, 327330, 1996.
  • Waddington, E. D., J. Cunnigham, and S. L. Harder, The effects of snow ventilation on chemical concentrations in ice cores, in Chemical Exchange between the Atmosphere and Polar Snow, edited by E. W. Wolff, pp. 403452, Springer-Verlag, New York, 1996.
  • Wagenbach, D., F. Ducroz, R. Mulvaney, L. Keck, A. Minikin, M. Legrand, J. S. Hall, and E. W. Wolff, Seasalt aerosol in coastal Antarctic regions, J. Geophys. Res., 103, 10,961, 1998.
  • Wagner, T., C. Leue, M. Wenig, K. Pfeilsticker, and U. Platt, Spatial and temporal distribution of enhanced boundary layer BrO concentrations measured by the GOME instrument aboard ERS-2, J. Geophys. Res., 106, 24,22524,235, 2001.
  • Wessel, S., S. Aoki, P. Winkler, R. Weller, A. Herber, H. Gernandt, and O. Schrems, Tropospheric ozone depletion in polar regions: A comparison of observations in the Arctic and Antarctic, Tellus, Ser. B, 50, 3450, 1998.
  • Wettlaufer, J. S., and M. G. Worster, Dynamics of premelted films: Frost heave in a capillary, Phys. Rev. E, 51(5), 46794689, 1995.
  • Wolff, E. W., Location, movement and reactions of impurities in solid ice, in Chemical Exchange between the Atmosphere and Polar Snow, edited by E. W. Wolff, and R. C. Bales, pp. 541560, Springer-Verlag, New York, 1996.
  • Zangmeister, C., J. Turner, and J. Pemberton, Segregation of NaBr in NaBr/NaCl crystals grown from aqueous solutions: Implications for sea salt surface chemistry, Geophys. Res. Lett., 28, 995998, 2001.