Introduction
Primary and secondary marine aerosols constitute the main contributions to the snow composition in coastal regions in Antarctica and may play an important role in the climate control of remote marine regions, as principal factors for cloud-condensation-nuclei formation (Reference Charlson, Lovelock, Andreae and WarrenCharlson and others, 1987; Reference Ayers and GrasAyers and Gras, 1991). Antarctic coastal areas are directly exposed to marine air masses and are characterised by high annual snow accumulation. Chemical analysis of samples from snow pits and firn cores gives high-resolution records of the environmental and climatic history of the Southern Ocean (Reference Wagenbach, Wolff and BalesWagenbach, 1996). Snow sampling at stations located in different geographic positions (altitude and distance from the sea) in the area surrounding Terra Nova Bay can give useful information about aerosol sources and transport processes (such as short- and long-range transport and fractionating effects). Variations in marine productivity can also be investigated by means of the seasonal pattern and the temporal and spatial distribution of some biogenic markers, such as non-sea-salt sulphate (nssSO42-) and methanesulphonic acid (MSA). Since these sulphur-cycle compounds are quite stable in glacial ice, the climatic conditions and phytoplanktonic-activity variations can also be evaluated for the past (Reference Legrand, Feniet-Saigne, Saltzman, Germain, Barkov and PetrovLegrand and others, 1991; Reference Saltzman and DelmasSaltzman, 1995).
CI" Na+, Mg2+ and, partially, SO,2- are the typical components of primary marine aerosol (sea spray) which is the dominant component in snow collected in low-altitude coastal areas. When the distance from the sea and/or more importantly the altitude increases, the contribution of secondary atmospheric aerosol becomes important. In unpolluted marine areas such as Antarctica, where the impact of other aerosol sources is low, the secondary marine aerosol is closely related to biologic processes. The main compound emitted by the oceanic phytoplanktonic activity is dimethylsulphide (DMS) which is oxidised in the atmosphere mainly to H2SO4 (determined as nssSO4 2-) and MSA (Reference Andreae and RaemdonckAndreae and Raemdonck, 1983; Reference Saltzman and DelmasSaltzman, 1995). MSA is exclusively produced by marine life (Reference Savoie and ProsperoSavoie and Prospero, 1989; Reference Legrand, Feniet-Saigne, Saltzman, Germain, Barkov and PetrovLegrand and others, 1991; Reference Wagenbach, Wolff and BalesWagenbach, 1996), whereas ussSO4 2– has other potential sources such as volcanic, anthropogenic and crustal inputs (Reference Legrand and DelmasLegrand and Delmas, 1987; Reference ShawShaw, 1988). In Antarctic coastal areas, however, the ussSO4 2–,2" biogenic source is dominant, especially during summer (Reference Legrand and DelmasLegrand, 1995; Reference Wagenbach, Wolff and BalesWagenbach, 1996; Reference Udisti, Traversi, Becagli and PiccardiUdisti and others, 1998). The seasonal pattern of DMS emissions gives an analogous seasonal character to the MSA and nssSO42- content in aerosol and snow, with concentration maxima during the phytoplanktonic bloom periods and immediately thereafter (Reference Prospero, Savoie, Saltzman and LarsonProspero and others, 1991; Reference Saltzman and DelmasSaltzman, 1995; Reference Wagenbach, Wolff and BalesWagenbach, 1996; Reference de Mora, Wylie and Dickde Mora and others, 1997). Previous studies on the distribution of marine components in snow and firn samples collected in northern Victoria Land (Reference Udisti, Casella, Piccardi, Restelli and AngelettiUdisti and others, 1993, Reference Udisti, Traversi, Becagli and Piccardi1998; Reference Piccardi, Becagli, Traversi, Udisti, Colacino, Giovannelli and StefanuttiPiccardi and others, 1996a, b) revealed a sharp seasonal pattern for MSA and nssSO42- (so their summer maxima can be used as reliable seasonal markers in firn-core dating; Reference UdistiUdisti, 1996). There is a less evident seasonal trend for sea-spray components (CF, Na+, Mg2+), but higher concentrations are found in winter snowfalls. A relevant influence of altitude on the chemical composition of snow and on aerosol fractionating effects was also observed.
The aim of this paper is to confirm and complete such observations with a large new dataset derived from chemical analysis of four snow pits sampled at different positions in coastal areas of Terra Nova Bay. A spatial evaluation of the summer marine contribution was also performed by analyzing fresh summer snow samples collected in vertical transects (stations located very close together but at different altitudes) and horizontal transects (stations at different distances from the coastline arranged along a sea-inland direction).
Sampling and Analytical Methods
Sampling stations
Terra Nova Bay is located between the Ross Sea and the Transantarctic Mountains, in a region affected by intense katabatic winds (Reference Bromwich and KurtzBromwich and Kurtz, 1984). Study of the seasonal pattern of snow components from snow-pit or firncore samples calls for undisturbed annual snow deposition. To avoid areas where wind effects on the snow cover (e.g. ablation or redistribution) can perturb the snow-layer deposition, a detailed analysis of satellite photos was performed (Stenni and others, unpublished). In this way, it was possible to choose stations where wind morphologies (e.g. sastrugi and snowdrift) were missing, so that an undisturbed snow deposition was highly probable.
Samples were obtained from four snow pits (2.5-3.5 m deep) dug at stations located within about 200 km of the Terra Nova Bay Italian base, during the 1993-94 Italian Antarctic Campaign.
During the Italian Antarctic Campaigns from 1993-94 to 1996-97, about 100 samples of surface snow were collected in summer at different stations in the same area. In particular, surface snow from coast-inland transects was obtained along the Reeves-Priestley Névé and Rennick Névé. The surface snow was collected as soon as possible after deposition (maximum 2 days), so the samples can be considered as "fresh snow" samples, little affected by post-depositional processes.
Sampling procedures and analytical methods
All samples were collected using a sampling protocol to avoid contamination (Reference Udisti, Barbolani and PiccardiUdisti and others, 1991, Reference Udisti, Bellandi and Piccardi1994; Reference Piccardi, Udisti and CasellaPiccardi and others 1994). The snow pits were hand-excavated by personnel wearing clean-room clothing. Pit walls were cleaned by removing a 5-10 cm snow layer with a stainless-steel scraper just before sampling. Pre-cleaned polypropylene vials (inside diameter 35 mm) were inserted into the vertical snow walls, and after removal each sampling vial was sealed in a double polyethylene bag. The vials were kept frozen (-20°C) until analysis (usually a few months after sampling).
Snow components were determined using an integrated system of three ion chromatographs (ICs)with conductivity detectors. The separation columns were Dionex AS12A (CF, Br", NO3 –, PO4 3- and SO4 2–), AS11A (F-, acetate, propionate, formate, MSA and pyruvate) and Dionex CS12A (Na+, NH4 +, K+, Mg2+ and Ca2+), followed by electrochemical micro-membrane conductivity suppressors. All components were analyzed as soon as possible after melting (usually 30 min). The samples were melted under a class 100 laminar hood and filtered on a 0.45 μm Teflon membrane just before analysis. A lmL sample loop was sufficient to determine the lowest concentrations of the components without pre-concentration. Blank control procedures, sample treatment and IC methods are described in Reference Udisti, Barbolani and PiccardiUdisti and others, (1991, Reference Udisti, Bellandi and Piccardi1994), Reference Piccardi, Udisti and CasellaPiccardi and others, (1994) and Reference UdistiUdisti (1996).
04, Mount Melbourne. Surface snow: six sites between 74°32’ S, 164°46’E and 74°21’S, 164° 41’E (360-2670 m a. s. I).
19, Styx Glacier plateau. Snow pit: 73°52’ S, 63° 42 E (1660 m a. sl).
27, McCarthy Ridge. Snow pit: 74°36 S, 163°03’E (870 m a.sl).
36, Hercules Névé. Snow pit: 73°06 S, 165°28’ E (2960 m a.sl).
44, Pilot Glacier. Snow pit: 73° 16′ S, 165° 31’ E (2100 m a.sl).
Transect 42, Reeves-Priestley Plateau. Surface snow: eight sites spaced 10-20 km apart from 74°29’ S, 161°00’E (1540 m a.sl) to 73°37’ S, 157°54’ E (2320 m a.sl).
Transect 43, Rennick Glacier plateau. Surface snow: nine sites spaced 10-20 km apart from 73°20’ S, 162°30’E (2210 m a.sl) to 72°20’ S, 162°26 E (1350 m a.sl). The elevation decreases away from the sea.
Not all the stations are visible, because some sampling points are obscured by the dark patches indicating the Na+concentrations.
For each sample, the nssSO4 2- concentration was calculated (Reference Maupetit and DelmasMaupetit and Delmas, 1992) by
Previous nssSO42- measurements on samples collected in the same region indicated that Na+ is the best sea-spray marker. In the lowest-elevation station (station 27, McCarthy Ridge, 870ma.s.l.) some negative nssSO42- values were found in winter samples, when the biogenic contribution was very low. For a reliable comparison of all snow-pit data, only positive nssSO4 2- values were used for the statistical treatments. A detailed discussion of the nssSO4 2- calculation can be found in Reference UdistiUdisti (1996).
The snow-pit dating was performed using a multiparametric method based on normalisation of the concentration/depth profiles of H2O2, nssSO4 2- and MSA (Reference UdistiUdisti, 1996). The dating results are reported in Reference Piccardi, Barbolani, Becagli, Traversi, Udisti and CseconPiccardi and others (1995). The normalised profiles of the three seasonal indicators gave a new, more reliable seasonal profile (designated the "normalised sum"), with a sinusoidal pattern, where the maximum values are related to summer samples and the minimum values to winter samples. Summer samples are those with a normalised sum value higher than half of the relative annual peak height, and winter samples those with lower normalised sum values.
Results and Discussion
Sea-salt concentrations in Antarctic snow show the largest spatial and temporal variations, compared to other source components. In aerosol measurements, the mean concentrations range from 0.062 μg m−3 at the South Pole (Reference Tuncel, Aras and ZollerTuncel and others, 1989) to 3.2 μg m−3 on the Antarctic Peninsula (Palmer station; Reference Wagenbach, Wolff and BalesWagenbach, 1996). Coastal sites influenced by katabatic winds, which are able to blow away the marine air masses from the coastal areas, show an intermediate sea-salt aerosol content (0.27 μgm−3 at Mawson; Reference Wagenbach, Wolff and BalesWagenbach, 1996).
Reference Allen, Mayewski, Lyons and SpencerAllen and others (1985), analyzing two ice cores collected at two stations located in northern Victoria Land (Evans Névé and Morozumi Range), found the following respective median values for some of the compounds discussed here: chloride 147 and 163 μg L-1; sodium 81.6 and 87.6 μgL-1; sulphate 68.2 and 73 μg L-1;
Figure 1 shows the firn-core and surface-snow sampling stations. For each station, the mean concentration of Na+ (considered as a specific sea-spray marker) is reported (proportional to the symbol size). The importance of location in examining the contribution of sea spray to the snow composition is shown by the fact that the highest mean Na+ contents are at the lower-altitude coastal stations (Fig. 1). In particular, the importance of distance from the sea and altitude is highlighted by the Reeves-Priestley (horizontal) and Melbourne (vertical) transects, respectively. For the Reeves-Priestley transects, the sampling sites were spaced about 10-20 km apart and located at high altitudes (1540-2320ma.s.L). A progressive decrease of the Na+ mean concentration is evident for the low absolute values. The Melbourne transect gives very useful information concerning the altitude effect. In fact, the samples were collected at closely spaced stations but at very different altitudes. The Na+ decrease is very sharp for the first 1000 m. At higher altitudes, the decreasing trend is less evident.
The Rennick transect shows an apparent anomaly. The Na+ concentration decreases as the altitude decreases. This glacier is located behind the Transantarctic Mountains, and therefore the altitude decreases in the inland direction. The marine air masses coming from the Ross Sea and up Campbell Glacier must rise over 2500 m before they reach the Rennick Glacier plateau, and so the sea-spray aerosol content is deposited preferentially at the highest Rennick site (first site at 2210 m a.s.l.; see also Fig. 1 caption).
Figure 2 shows the altitude effect and the seasonal pattern of the primary (Na+; Fig. 2a) and secondary (nssSO4 2- and MSA; Fig. 2c and d) marine aerosol in the four snow pits. The Na+ concentration is high in the lower station, with a maximum mean value in winter, but decreases quickly in the first altitude step (870-1660 m). Above 1500 m, the concentrations are very low and the seasonal differences less evident. At Hercules Névé station (2960 m a.s.l.), a seasonal trend is lacking. A similar pattern is shown by GT and Mg2 + . On the other hand, nssSO4 2- concentrations are higher in summer at all stations, and the tendency to decrease as a function of altitude is more gradual. For altitudes above 2000 m, the concentration approaches 60 μg L−1. This pattern is confirmed by the MSA concentration trend (Fig. 2d), but a lower seasonal variation is observed for the highest station, where summer and winter mean concentration values are quite similar, ranging from 10 to 20 μgL-1. Figure 2b shows the behaviour of total sulphate. At the lowest station, sulphate follows the same trend as Na+, with high winter values. From 870 to 1660 m, the concentration drops quickly, while the decrease in percentage is lower than for Na+. From 1660 to 2960 m, the sulphate shows an opposite seasonal pattern, with maxima in summer. As for nssS04 2-, a seasonal distribution is also revealed by totS04 2- for the highest station. This shows that the SO4 2- contribution from sea spray is partially counterbalanced by the presence of biogenic sources, which gain importance as altitude increases.
The different contributions of primary and secondary marine aerosols as a function of altitude and seasonality are shown in Figure 3. In winter, at the lowest station (station 27), the sea-spray content (Na+, GT, Mg2 + and swSO4 2-) is more than 95% of the total marine contribution. On the other hand, in summer at the highest station (station 36) the secondary marine aerosol (nssSO42- and MSA) constitutes about 50% of the marine input. These data are in agreement with long-term aerosol observations at Neumayer station (Reference Wagenbach, Wolff and BalesWagenbach, 1996). In winter, 90% of the dry aerosol mass concentration consists of sea salt, and the other 10% consists of nssSO42- and NO3 – in a ratio of approximately 1:1. In summer, the sea-spray content is reduced to 50%, while the marine biogenic compounds (nssSO42- and MSA) constitute the remaining 50%.
The seasonal variations of the sea-salt load are relatively weak (at least for stations above 1000 m a.s.l.) and are not in phase with the annual sea-ice cycle. Minimum sea-spray values in aerosol and snow samples are found in summer when the sea surface is open, and maximum concentrations are measured in autumn and spring when the sea-ice extent is usually greater. Sudden high concentration peaks of sea spray are evident in winter, with values 10-100 times higher than mean seasonal values, especially at the lowest station (Reference Piccardi, Udisti and CasellaPiccardi and others, 1994). A similar pattern is shown by aerosol measurements from other coastal stations (Mawson, Neumayer and Dumont D’Urville; Reference Wagenbach, Wolff and BalesWagenbach, 1996 and references therein). The sea-spray source areas and the transport pathways of such highly enriched sea-spray aerosol are not well known. Heavy salt storms, which are more probable in winter, could be the primary cause, although local sea-ice conditions must also be considered. A winter polynya in Terra Nova Bay is well documented (Reference Bromwich and KurtzBromwich and Kurtz, 1984) and a combination between such a polynya and a heavy storm with marine winds seems to be a reliable explanation for sea-spray winter maxima.
By using Na+ concentrations as a sea-spray marker, it is possible to demonstrate the fractionation phenomena related to the increase in altitude. Figure 4 a-f show the distribution diagrams (box plots) of the Cl-/Na+, Mg2+/Na+ and totSO4 2-/Na+ concentration ratios for the four snow pits sampled as a function of seasonality and altitude. Each box indicates the concentration range of 50% of the samples, and its height is correlated to the variability in the data. The lines in the boxes indicate the median values, and the dashed lines represent the sea-water ratios. In winter, the data variability is generally low and the median ratio values are similar to those of sea water. In summer on the other hand, only the Mg2+/Na+ ratio remains similar to the sea-water composition (except at station 19, where a Mg2 + crustal contribution occurs; Reference Piccardi, Becagli, Traversi, Udisti, Colacino, Giovannelli and StefanuttiPiccardi and others, 1996a), while Cl-/Na+ and, above all, totSO4 2-/Na+ ratios are higher. This difference, relative to the sea-water composition, increases with altitude. For the totSO4 2-/Na+ trend, the explanation is clear. The biogenic contribution, which is very important in summer, increases the ratio with respect to the sea-water content. For CF, fractionation of the aerosol during transport must be considered. So too must other non-sea-salt CF sources such as volcanic inputs, long-range HG1 transport or HG1 formation from a reaction between NaCl and H2S04 in aerosol. To illustrate the Na+ effect, the enrichment factors for CF, Mg2 + and totSO4 2- in summer samples are reported as a function of Na+ concentration (Fig. 5a-c). The enrichment factor (EF) for compound X is calculated as:
The Na+ scale is logarithmic to point out the trend at the lowest Na+ concentrations. A general increase in CF and totSO4 2- EF when the Na+ concentration decreases is evident in all snow-pit stations. The increase in the Cl-/Na+ ratio, which is more evident in summer, may be related to seasonality (aerosol transport processes or seasonal pattern of other CF sources). On the other hand, the non-sea-salt contribution to the CF concentration could be quite constant all year round, but with the CF excess (with respect to the sea-water Cl-/Na+ ratio) becoming evident only in summer when the sea-spray contribution is lower (low values of Na+). This second hypothesis is supported by the experimental observation that at the highest elevation, where the Na+ content is always low (low contribution of sea spray), the CF enrichments are similar both in summer and in winter (Fig. 4a and b).
Figure 5c shows high sulphate enrichment with respect to the sea-water composition in all stations. Enrichment factors in the range 20-100 are found for Na+ concentrations of <10 μg L−1. For this component both the above hypotheses are true. The seasonality of the biogenic source of nssSO4 2- is clear, and its contribution to the total sulphate budget is even more evident when the sea-spray contribution is low, as it is in summer and at high-altitude stations.
Just as Na+ was used for the evaluation of the sea-spray contribution, MSA can be used as a biogenic marker to explain the biogenic contribution of nssSo4 2- and to estimate the distribution processes between MSA and H2SO4 coming from DMS. Such evaluations have to be performed on summer samples, where the biogenic contribution is higher, and in the absence of any post-depositional processes which can change the snow composition. For this reason, samples of fresh summer snow were collected just after deposition within about a 200 km radius of the Terra Nova Bay Italian base. Figure 6a shows the relationship between nssSO4 2- and MSA. There is a good correlation (R = 0.82, n = 95), with a slope (indicating the summer mean nssS04 2-/MSA ratio) of 2.7. Figure 6b shows the trend of the nssS04 2-/MSA (w/w) ratio as a function of MSA concentration. In the same figure, the dashed line indicates the ratio value of 2.7. For MSA concentrations as high as 50 μg L−1, the nssSO4 2-/MSA clearly tends to a value very near to 2.7. When the MSA concentration decreases, the ratio progressively increases to values even higher than 20. The highest values of the nssS04 2-/MSA ratio are mainly obtained at the highest-elevation stations. The trend in the nssSO4 2-/MSA ratio could be explained by one or more of the following:
-
(1) the relative importance of non-DMS sources to the nssSO42- budget;
-
(2) the effect of aerosol fractionation during transport progressively enriching nssSO42- (with respect to the original nssS04 2-/MSA ratio from DMS), as distance from the sea and altitude increase;
-
(3) the contribution of marine air masses coming from lower latitudes (long-range aerosol transport) with an original nssSO4 2–/MSA ratio higher than that obtained in the Antarctic oceans.
In fact, the nssSO4 2–/MSA ratio is strongly affected by latitude (Reference Bates, Calhoun and QuinnBates and others, 1992). At medium latitudes, the nssSO4 2- /MSA aerosol ratio can reach values more than ten times greater than those found in the polar regions (Reference BerresheimBerresheim, 1987; Reference Saltzman and DelmasSaltzman, 1995). Moreover, the positive relationship between altitude and nssSO4 2–/MSA ratio leads us to believe that the different sizes of aerosol particles can play a relevant role. For example, the smaller aerosol particles are more involved in long-range aerosol transport from Southern mid-latitudes to Antarctic coastal areas (Reference Wagenbach, Wolff and BalesWagenbach, 1996).
A simpler explanation for the pattern of the nssSO4 2- / MSA ratio as a function of the MSA concentration is obtained by considering a quite constant and globally distributed background nssSO42- contribution (volcanic, crustal and/or from long-range transport), with a superimposed biogenic contribution. When the biogenic source decreases (due to seasonal variation and/or altitude effects) the background nssSO4 2- concentration becomes more important, leading to a progressive increase in the nssSO4 2–/MSA ratio.
Conclusions
The principal contributions to snow composition in northern Victoria Land are primary and secondary marine aerosol, and their relative importance varies as a function of season and position. This phenomenon can be demonstrated by observing the spatial and seasonal distribution of some components chosen as primary (Na+, Mg2+, CI" and tot SO4 2-) and secondary (MSA and nssSO4 2-) aerosol markers.
Altitude and distance from the sea are fundamental parameters in understanding the distribution processes of components coming from sea spray as well as from marine biogenic activity. When altitude and distance from the sea increase, fractionation phenomena related to the aerosol particle size occur. In particular, we found larger GT enrichments in summer samples than in winter samples, but this seasonal difference disappears in the highest stations. This proves the existence of other Cl- sources, evident only when the Na+ concentration is lower (summer precipitation and/or high altitudes, where the relatively large sea-spray particles arrive with difficulty).
For total sulphate, the enrichment with respect to sea-water composition prevalently found in summer is mainly due to the seasonality of the biogenic source (nssSO42- from DMS).
Evaluation of the biogenic contribution to the nssSO4 2- budget as a function of seasonality is difficult because of the contributions of other sulphate sources such as volcanic and crustal inputs.
A nssSO4 2- /MSA ratio of 2.7 was calculated from fresh summer snow samples, and this value could be representative for their distribution from DMS in northern Victoria Land coastal areas. The nssSOt/MSA ratio plotted vs MSA concentration, used as a specific biogenic marker, confirms this ratio for high MSA concentrations, but shows a progressive increase when the MSA concentration decreases. Several factors may explain this increase, e.g. the contribution of a background nssSO42- value from non-DMS sources (volcanic and crustal), the effect of aerosol fractionation during transport (related to altitude and distance from the sea) or the contribution of marine air masses coming from mid-latitudes with different nssSO4 2–/MSA ratios from DMS.
Acknowledgements
This research was carried out in the framework of the Programma Nazionale di Ricerche in Antartide. It was financially supported by Ente per le Nuove Tecnologie, l’Energia e l’Ambiente.