Introduction

Deconvoluting the biogeochemical cycling of atmospheric sulfur is important for understanding climate change. Sulfate (SO₄ ²–) aerosols play a key role in moderating the Earth’s climate. They function as cloud condensation nuclei and can enhance the lifetime of clouds and their ability to reflect incoming solar radiation. Additional aerosol mass in the atmosphere results in an increase in optical thickness and cloud albedo (Reference Anderson, Wolfe, Warren and DelmasAnderson and others 1995). This effect could in turn cause a change in oceanic productivity (which is a source of SO₄ ²– aerosols), thereby creating a climate feedback loop that is susceptible to climate change (Reference Bates, Charlson and GammonBates and others 1987; Reference Charlson, Lovelock, Andreae and WarrenCharlson and others 1987). Sulfate aerosols also contribute to the formation of SO₄ ²– haze. The size of SO₄ ²– particles determines their effect on climate. Sulfate particles derived from sea-salt particles are generally >1 μm in diameter and play an important role in the marine boundary layer (Reference MurphyMurphy and others 1998). Sulfate aerosols play various other roles in atmospheric chemistry that could be affected by or affect climate change, including an alteration of oxidation processes in the marine boundary layer (Reference Andreae and CrutzenAndreae and Crutzen, 1997).

The Antarctic displays high sensitivity to climate change and in turn greatly impacts the climate of low latitudes (Reference Bromwich and ParishBromwich and Parish, 1998). One of the limitations to understanding the sulfur cycle is the assessment of how much each potential SO₄ ²– source contributes and the accompanying effect on atmospheric processes and climate. There are three recognized sources of SO₄ ²– in Antarctica: sea-salt, marine biogenic and volcanic (Reference DelmasDelmas, 1982; Reference LegrandLegrand, 1997). Anthropogenic source contributions to Antarctica are believed to be negligible (Reference ShawShaw, 1982; Reference Legrand and MayewskiLegrand and Mayewski, 1997). Unlike ice-core records from Greenland, which clearly show an increase in background SO₄ ²– concentration since the industrial revolution (Reference MayewskiMayewski and others 1993), Antarctica shows no increase in total SO₄ ²– concentration with time (Reference LegrandLegrand, 1997; Reference Legrand and MayewskiLegrand and Mayewski, 1997). The relatively few sources of SO₄ ²–, when compared to Greenland, make Antarctica a useful environment to observe the natural sulfur cycle.

Estimates of the Antarctic SO₄ ²– budget are based on aerosol measurements, analysis of snow and ice samples, and an understanding of volcanism and meteorology of the region. Aerosol studies indicate that background xsSO₄ ²– in the Antarctic comes primarily from marine biogenic sources (Reference Pszenny, Castelle, Galloway and DucePszenny and others 1989; Reference MinikinMinikin and others 1998). Mount Erebus, the only active volcano on the continent, contributes to the Antarctic sulfur cycle, but the strength of this source varies temporally depending on the size of the lava lake (Reference Kyle and MeekerKyle and Meeker, 1990).

Ice-core records provide an historical record of atmospheric chemistry, including SO₄ ²– deposition (Reference Delmas and BoutronDelmas and Boutron, 1977; Reference DelmasDelmas, 1982; Reference Legrand, Feniet-Saigne, Saltzman, Germain, Barkov and PetrovLegrand and others 1991; Reference LegrandLegrand, 1997) and have been used to assess the sulfur cycle in Antarctica (Reference Delmas and BoutronDelmas and Boutron, 1980; Reference Legrand and FenietLegrand and Feniet-Saigne, 1991; Reference Meyerson, Mayewski, Kreutz, Meeker, Whitlow and TwicklerMeyerson and others 2002). Snow-pit and ice-core data demonstrate that the SO₄ ²– production operates seasonally, with concentration minima in the winter and maxima in the summer (Reference Legrand and PasteurLegrand and Pasteur, 1998; Reference MinikinMinikin and others 1998). Excess SO₄ ²– concentrations are independent of accumulation rate and elevation on a continent-wide scale (Reference Mulvaney and WolffMulvaney and Wolff, 1994; Reference Kreutz and MayewskiKreutz and Mayewski, 1999). Distance inland correlates to a decrease in the concentration of xsSO₄ ²–, which may be due to the portion of SO₄ ²– that is derived from marine productivity. The 1998 Italian ITASE (International Trans-Antarctic Scientific Expedition) traverse in East Antarctica demonstrated that the decrease inland of xsSO₄ ²– extended only 250 km and then xsSO₄ ²– increased in concentration from 250 to 770 km (Reference PropositoProposito and others 2002). In surface snow measurements, there is a negative correlation between both distance inland and elevation and methanesulfonic acid (MSA) concentration. As with xsSO₄ ²–, accumulation rate does not affect MSA concentration, except perhaps at low-accumulation-rate sites (Reference Legrand, Feniet-Saigne, Saltzman and GermainLegrand and others 1992; Reference Wagnon, Delmas and LegrandWagnon and others, 1999). A seasonal variation of MSA occurs in several cores. The presence of the seasonal signal depends on location, with only an interannual signal at South Pole and a stronger seasonal signal in coastal areas (Reference Legrand, Feniet-Saigne, Saltzman and GermainLegrand and others 1992; Reference Meyerson, Mayewski, Kreutz, Meeker, Whitlow and TwicklerMeyerson and others 2002). In other Antarctic studies, there is no seasonal variation, unlike both sea-salt SO₄ ²– and xsSO₄ ²– (Reference Ivey, Davies, Morgan and AyersIvey and others 1986). Recent aerosol studies and data from snow pits and firn cores indicate that MSA and xsSO₄ ²– correlate on a seasonal basis, with summer peaks and winter lows (Reference Legrand, Feniet-Saigne, Saltzman and GermainLegrand and others, 1992; Reference StenniStenni and others, 2000; Reference Arimoto, Nottingham, Webb and SchloessinArimoto and others 2001).

The SO₄ ²– sources and emission rates are important indicators of the vertical distribution of SO₄ ²– aerosols over Antarctica, and therefore the radiative effects. Stable sulfur isotopes provide a potentially useful tool for estimating SO₄ ²– sources in ice cores. This technique has previously been used to determine the source of SO₄ ²– in precipitation and aerosol studies (Reference Calhoun, Bates and CharlsonCalhoun and others 1991; Reference Nriagu, Coker and BarrieNriagu and others 1991; Reference Wadleigh, Schwartz and KramerWadleigh and others 1996). Sulfur isotopes are useful in studying atmospheric transport mechanisms and chemistry during volcanic eruptions (Reference Castleman, Munkelwicz and ManowitzCastleman and others 1973) and during anthropogenic emissions (Reference Newman, Forrest and ManowitzNewman and others 1975). Differences in chemical reaction rates can cause isotope fractionation, resulting in an isotopic ‘signature’ that can be used to identify the chemical process (such as incorporation into a biologic system) or source of the atmospheric SO₄ ²–. The isotopic signature of sulfur can be used to indicate the source of the SO₄ ²– if there are a few sources with distinct signatures, as is the case in Antarctica (Reference NielsenNielsen, 1974; Reference McArdle and LissMcArdle and Liss, 1995). The resulting δ³⁴S can be deconstructed through a mixing equation. The more that is understood about the sources of atmospheric sulfur in a region, the easier it becomes to use sulfur isotopes to quantify SO₄ ^2– sources.

Recently, sulfur isotopes were used in an ice-core study in central Antarctica (Reference Patris, Jouzel and DelmasPatris and others 2000). A quantitative assessment of SO₄ ²– at South Pole was made based on sulfur isotopic measurements on sections of a firn core. Reference Patris, Jouzel and DelmasPatris and others (2000) determined that marine biogenic emissions are the dominant background source of SO₄ ²–. Sulfur isotopes in a Greenland ice core were also used to examine the Arctic sulfur cycle during both pre-industrial and industrial times (Reference Patris, Jouzel and DelmasPatris and others 2002). Applying sulfur isotope studies to ice-core records can greatly improve understanding of the sulfur cycle over long periods of time by providing evidence for SO₄ ²– source regions and transport.

The δ³⁴S measurements presented here are the first sulfur isotopic measurements from the West Antarctic ice sheet. These measurements are used to infer differences between the sources of SO₄ ²– in West Antarctica and central Antarctica. These data indicate the potential of sulfur isotopes as a tool for partitioning the SO₄ ²– sources (marine biogenic, sea-salt and volcanic) to West Antarctica.

Methods

Core processing

The 147m RIDSA ice core (78.73˚ S, 116.33˚W; 1740m a.s.l.) (Fig. 1) was retrieved in West Antarctica in 1995 (Kreutz and Mayewski, 1999). Core processing was performed in clean conditions, with samples collected into pre-cleaned containers and stored below –15˚C until melting and analysis. Processing and chemical analysis were done at 3 cm intervals down to 60 m, yielding high-resolution (sub-annual) chemical data. From meter 60 to 147, processing was done at 60 cm intervals. One-meter sections of core from 10 to 26m were used for sulfur isotopic analysis, covering the period 1935–76. The time covered by individual samples in this dataset (RIDSA 2001) was 2–3 years, depending on accumulation rate.

Fig. 1. RIDSA location map. Several other cores are denoted in white text. Volcanoes that are known to have been active during the time period covered by isotopic measurements presented in this paper are represented with black text.

Sections of the RIDSA core from 67.2 to 69.6 m, covering the years 1796–1805, and from 13.8 to 15.9 m, covering the years 1961–67, were used for isotopic method development. These samples are designated as RIDSA 2000.

Ion analysis

Major-ions (Na⁺, K⁺, NH₄ ⁺, Mg²⁺, Ca²⁺, Cl^–, NO₃ ^–, SO₄ ²–) and MS (measured as methanesulfonic acid or MSA) measurements were made by ion chromatography at the (parts per billion) ppb level. Anions were analyzed on a Dionex DX-500 ion chromatograph with an AS-11 column using 6mM NaOH eluent. Cations were analyzed on a Dionex DX-500 ion chromatograph with a CS-12A column using 25 mM MSA eluent. A β-activity profile was made using 20cm samples (Reference Kreutz, Mayewski, Meeker, Twickler and WhitlowKreutz and others 2000). The β-activity maximum is assumed to indicate the austral summer 1964/65, based on the global peak from the Atmospheric Test Ban Treaty (1963) and the transport lag to Antarctica (Reference Picciotto and WilgainPicciotto and Wilgain, 1963). It is possible to count back to the global horizon using summer peaks in nssSO₄ ²–. Dating is therefore based on seasonal nssSO₄ ²– cycles (Reference Kreutz, Mayewski, Meeker, Twickler and WhitlowKreutz and others 2000). These dating techniques show that the entire core covers the time period 1506–1995.

The excess SO₄ ²– is calculated from the Na⁺ concentration of the samples using the following equation:

(1)

where k is equal to the sea-water (SO₄ ²–/Na⁺) mass ratio, 0.251 (Reference Wilson, Riley and SkirrowWilson, 1975).

Isotope analysis

One hundred micrograms of SO₄ ²– is necessary for isotopic analysis, making high-resolution analysis difficult in Antarctic cores due to low accumulation rates (Reference Patris, Jouzel and DelmasPatris and others 2000). The RIDSA 2000 samples are aggregated residual from samples taken from the RIDSA core that were previously aliquoted for chemical analysis. Sampling was initially performed at 3 cm intervals for analysis of major ions. Two time periods, one covering the Agung (Indonesia) period (1963 eruption) and one covering a pre-industrial period, were chosen, and sample remaining from the initial chemical analysis was combined until a volume of 1 L was attained. The years covered by these two measurements were dependent on the volume needed for isotopic analysis (1 L). Additional isotopic analysis was performed on 1m sections of the RIDSA core, including the Agung period. All core sections were melted in sterile conditions and then transferred to 1 L bottles for analysis at the Memorial University of Newfoundland.

Sample aliquots used for stable-isotopic analysis were transferred to acid-washed ‘snap-cap’ vials. Vials were placed on a hot plate in a positive-pressure, High Efficiency Particle Air (HEPA)-filtered air fume hood, and the solutions were slowly evaporated until the volume was reduced to approximately 20 mL from an initial 1 L. The concentrated sample was then used to fill 10×10×10 mM tin capsules and evaporated to dryness. The process was repeated between two and five times according to the measured SO₄ ²– concentration in order to obtain sufficient sample for isotopic analysis. After the final evaporation, the tin capsule was closed and placed in the auto-sampler of a Carlo Erba 1500 elemental analyzer interfaced by ConFlo II to a Finnigan MAT 252 stable-isotope ratio mass spectrometer.The sulfur isotopic ratio is expressed in delta notation with respect to the standard, Vienna Canyon Diablo Troilite (VCDT), and is represented by:

(2)

Calibration to VCDT was performed using two Newfoundland internal standards, NZ-1 and NZ-2, with reported values of –0.30 ± 0.3‰ and +21.0 ± 0.3‰ respectively. The overall error on isotopic analysis based on duplicates and internal standards is ±0.5‰.

RESULTS Ion data

Ion averages for each of the meters taken for isotopic analysis (10–26 m; RIDSA 2001) and the RIDSA 2000 samples are presented in Table 1. The one sample of overlap between the two RIDSA datasets is consistent with the other.

Table 1. Ion concentrations for each RIDSA sample. Units are ng g^–1 except for Cl/Na mass ratio and R value, which have no units. The R value is the ratio MSA/xsSO₄ ^2– and is often used to assess the marine biogenic portion of total SO₄ ^2–

Sodium and xsSO₄ ²– have strong seasonal cycles, consistent with other studies in the Antarctic (Fig. 2) (Reference Murozumi, Chow and PattersonMurozumi and others 1969; Reference Herron and LangwayHerron and Langway, 1979; Reference HammerHammer, 1980; Reference Legrand, Feniet-Saigne, Saltzman and GermainLegrand and others 1992; Reference Whitlow, Mayewski and DibbWhitlow and others 1992; Reference MinikinMinikin and others 1998; Kreutz and Mayewski, 1999; Reference Reusch, Mayewski, Whitlow, Pittalwala and TwicklerReusch and others 1999).

Fig. 2. High-resolution Na⁺, xsSO₄ ²– and MSA data from the RIDSA core. These data show the portion of the core that was analyzed for sulfur isotopes and extend to the top of the core. A clear seasonal cycle can be seen for all species.

Na⁺ peaks in the winter due to increased storm activity, while xsSO₄ ²– has a summer peak due to its marine biogenic component (Reference Whitlow, Mayewski and DibbWhitlow and others 1992). During the summer months, sea ice retreats as solar radiation increases, leaving more open ocean in which primary productivity can thrive. One sign of increased productivity might be the increase in SO₄ ²– concentration during the summer months. The MSA profile shows a subdued seasonal signal relative to the xsSO₄ ²– signal (Fig. 2). In other glaciochemical studies performed in the Antarctic, MSA has also been found to correlate with xsSO₄ ²–, with both showing a summer maximum (Reference Kreutz and MayewskiKreutz and Mayewski 1999). The xsSO₄ ²– is a combination of primarily, or perhaps exclusively, marine biogenic and volcanic SO₄ ²–. Table 2 gives concentration data for the three possible contributors to total SO₄ ²– for the RIDSA samples.

Table 2. Sulfur data from the RIDSA core. f _ss indicates the fraction of the total SO₄ ^2– concentration that comes from sea salt, δ_tot is the isotopic signature for total SO₄ ^2– and δ_nss is the isotopic signature for the non-sea-salt portion of total SO₄ ²–

The Agung eruption (17 March 1963; Reference Devine, Sigurdsson, Davis and SelfDevine and others, 1984) is apparent in the xsSO₄ ²– profile (Fig. 2) and can be seen as an increase in the seasonal minima of SO₄ ²– from late 1963 to 1966. The eruption is recorded in the RIDSA core starting in late 1963, about 6 months after the eruption. This time difference of around 6 months between the eruption and its subsequent deposition in the Antarctic is due to the long distance the SO₄ ²– load is transported. The lag time varies depending on the location of the volcano, but other work done in the Antarctic shows a similar 6 month time lag for the Agung eruption (Reference LegrandLegrand, 1997). Other Antarctic cores show volcanic peaks as a SO₄ ²– spike above all other seasonal maxima. In the RIDSA core, however, volcanic input is typically represented by an increase in background SO₄ ²–, as is the case with the Agung eruption (Fig. 2). The other volcanic period shown in Figure 2, the 1991 Pinatubo–Hudson eruption, is an exception. This may be related to seasonal changes in atmospheric transport or accumulation rate.

Isotope data

The δ³⁴S of each meter of core analyzed is given in Table 2. The values range from 3.1‰ to 9.9‰, which is lower than those reported for central Antarctica, which range from 9.3‰ to 18.1‰ (Reference Patris, Jouzel and DelmasPatris and others 2000). The measured isotopic signature is plotted against the inverse of total SO₄ ²– concentration to highlight isotope trends with concentration (Fig. 3).

Fig. 3. Sulfur isotope measurements from the RIDSA core plotted with published data from central Antarctica (Reference Patris, Jouzel and DelmasPatris and others, 2000). The linear regression is shown for each dataset. The δ³⁴S values are plotted against the inverse of the total sulfate concentration.

The central Antarctic data show a clear trend towards lower δ³⁴S values with increased concentration, and a linear regression (r = 0.997) yields a value of 2.6‰ for the intercept, which is assumed to be predominantly volcanic SO₄ ²–.

In West Antarctica, there is no discernible trend with SO₄ ²– concentration for the raw data. The Agung period, which was isotopically distinct in the central Antarctic data, is not distinct in West Antarctica. The variability, or scatter, in the RIDSA data may be due in part to the sampling technique. The core was sampled continuously at 1m intervals. This resulted in ten samples containing two summer sections, and six samples containing three summer sections.

The isotopic signature of the samples can be broken down into its components (Reference Patris, Jouzel and DelmasPatris and others 2000):

(3)

where δ_mes is the measured isotopic signature as given in Table 2, δ_ss, δ_vol and δ_mb are the isotopic signatures of the sea-salt, volcanic and marine biogenic components of the sample, and f _ss, f _xs, f _vol and f _mb are the mass fractions of the sea-salt, excess, volcanic and marine biogenic components, respectively. f _ss and f _xs can be determined using the concentration for xsSO₄ ²–, which was computed using the Na⁺ concentration. The following equation represents the sea-salt fraction of a sample:

(4)

The non-sea-salt fraction is (1 – f _ss). The isotopic signature of sea-salt SO₄ ²– is well constrained through field study to be 21‰ (Reference Rees, Jenkins and MonsterRees and others 1978). This reflects an enrichment in ³⁴S in ocean water compared to terrestrial sulfur. Rearranging Equation 1 yields:

The calculations are explained in the text.

(5)

Substituting the established value for δ_ss into the equation, along with the calculated f _ss and f _xs, yields an isotopic signature for the excess component of the samples (Table 2). These values are plotted against the inverse of the xsSO₄ ²– concentration to yield a mixing diagram for the xs component in Figure 4. No significant correlation exists (r = 0.38).

Fig. 4. The δ³⁴S measurements for the xsSO₄ ^2– portion of the total sulfate from the RIDSA core plotted against the inverse of the xsSO₄ ^2– concentration.

The highest δ³⁴S value (in the RIDSA time series) occurs in the sample covering the years 1947–50 (Fig. 5a). Mount Erebus eruptions are noted on the time series. The original xsSO₄ ²– data were smoothed using a high-tension robust spline to produce the data in Figure 5b. The robust-spline technique differs from other statistical procedures because it is more resistant to outliers, providing a smoothing technique that yields useful information without underestimating sporadic processes (Reference Meeker, Mayewski, Bloomfield and DelmasMeeker and others 1995). The peaks in the smoothed data reflect changes in volcanic SO₄ ²–. There appears to be a relationship between increases in background xsSO₄ ²– and δ³⁴S.

Fig. 5. (a) Time-series δ³⁴S measurements for xsSO₄ ^2– from the RIDSA core, along with Erebus eruptions (as noted by dashed lines with the year). The x-axis error bars depict the years that the data point represents. The variation on the y axis represents the analytical precision of the sample. (b) The smoothed xsSO₄ ^2– time series from the RIDSA core. Smoothing was done with a robust spline. The peaks represent increases in the background SO₄ ^2– concentration.

Discussion

The sulfur isotope ratios in the RIDSA ice core are lower than those in central Antarctica (Reference Patris, Jouzel and DelmasPatris and others 2000). Caution must be used when looking at sulfur isotopic signatures as a source fingerprint because while processes such as atmospheric transport generally lead to similar trends, the δ³⁴S value can still vary greatly for any different source (Reference Castleman, Munkelwicz and ManowitzCastleman and others 1973). For example, δ³⁴S values from volcanic emissions have been estimated to range from –15‰ to +25‰ (Reference McArdle and LissMcArdle and Liss, 1999) and from –5‰ to 5‰ (Reference NielsenNielsen, 1974). Taking this uncertainty into account, there are several explanations for the West Antarctic sulfur isotope data.

Because the δ³⁴S values from West Antarctica fall within the range of newly erupted volcanic material, the background SO₄ ²– source might at first appear to be entirely of volcanic origin. The δ³⁴S values of sulfur from newly erupted basalts range from 2‰ to 7‰ (Reference De Hoog, Taylor and van BergenDe Hoog and others 2001). Studies of volcanic gas emissions show a positive δ³⁴S value (Reference Castleman, Munkelwicz and ManowitzCastleman and others, 1974). There was active volcanism in Antarctica during the time period of the RIDSA core. The Smithsonian database of volcanic eruptions (www.volcano.-si.edu/gvp/) lists Mount Erebus, Deception Island, Candlemas and Bristol Island as known eruptive volcanoes between 1935 and 1976.

It is unlikely that the low δ³⁴S values can be attributed solely to local volcanism. It is clear that there is a strong marine influence on the RIDSA site. Sea-salt SO₄ ²– makes up 11.7–35.3% of the total SO₄ ²– concentration for the samples (Table 2). The isotopic signature of SO₄ ²– from a marine biogenic source is reasonably well constrained at ~18‰ (Reference Calhoun, Bates and CharlsonCalhoun and others 1991; Reference Patris, Jouzel and DelmasPatris and others 2000). This value is slightly lower than that of sea-salt SO₄ ²–, which is close to 21‰ (Reference Rees, Jenkins and MonsterRees and others 1978). The difference between these two marine constituents is the result of fractionation that occurs in the biological use of sea-water SO₄ ²–. Because there is a strong sea-salt component to the total SO₄ ²– at the RIDSA site, it is reasonable to assume that there is a strong marine biogenic component. An isotopic study of the sulfur cycle in central Antarctica determined that the background source of sulfur was almost exclusively of marine biogenic origin (Reference Patris, Jouzel and DelmasPatris and others 2000). In West Antarctica, however, the background isotopic signature is too depleted in ³⁴S to be exclusively of marine origin. The δ³⁴S values for the xs component of the RIDSA samples range from –0.7‰ to 6.8‰. Factoring out the sea-salt contribution reduces the δ³⁴S values, suggesting that the δ³⁴S values from xsSO₄ ²– represent some combination of marine biogenic and volcanic SO₄ ²–.

The heavier isotope (³⁴S) is depleted faster than the lighter isotope (³²S) after volcanic eruptions, due to changes in stratospheric chemistry and the pathway of SO₄ ²– aerosol formation. Plume studies from anthropogenic emissions confirm this mechanism (Reference Castleman, Munkelwicz and ManowitzCastleman and others 1974; Reference Newman, Forrest and ManowitzNewman and others 1975). The lowest δ³⁴S values (–24.4‰) occur higher in the stratosphere (Reference Castleman, Munkelwicz and ManowitzCastleman and others 1974). One possibility in West Antarctica is that a large portion of the sulfur budget comes from a marine biogenic source with a δ³⁴S value around 18‰, and that the rest of the SO₄ ²– can be attributed to aged and fractionated stratospheric air. Stratospheric sulfur with extremely low δ³⁴S values could mix with local air masses with a relatively high δ³⁴S value to produce the intermediate values (Fig. 4). For example, using a 50/50 input from stratospherically transported materials and marine materials (18‰), a calculation can be made from the overall xsδ³⁴S to determine the δ³⁴S of the second contributor. Using the following mixing relationship between recalculated xsδ³⁴S value and its two components

(6)

and substituting an xsδ³⁴S value of 2‰, within the range of the recalculated xsδ³⁴S values, gives the δ³⁴S value of the unknown contributor as –14‰. If stratospheric transport and temperature effects yield δ³⁴S values that are even more negative than –14‰, then the ratio of marine biogenic to stratospheric input would be higher, indicating that local tropospheric deposition is of greater consequence to the West Antarctic sulfur cycle than stratospheric input.

The variance within the time series (–0.7‰ to 6.8‰; Fig. 5) corresponds to changes in the smoothed background SO₄ ²– signal (Fig. 5). Increases in the δ³⁴S values over time correspond to increases in the xsSO₄ ²– concentration over time. These xsSO₄ ²– peaks do not appear to reflect any global volcanism. Mount Erebus, however, is known to have erupted during the time period covered by each of the samples that has a peak in the time series. The largest xsδ³⁴S value is 6.8‰, in the sample covering the years 1947–50. It coincides with a peak in the background xsSO₄ ²– and a known Erebus eruption. Local volcanism would cause an increase in the δ³⁴S value because newly erupted material has a positive δ³⁴S value and is deposited relatively quickly, without the fractionation associated with transport (Reference Castleman, Munkelwicz and ManowitzCastleman and others 1974). The δ³⁴S peak during the 1947 Erebus event is much larger than the peak associated with the two other local events that occurred during 1957 and 1972. The peak in xsSO₄ ²– during the 1947 event is also much larger than the other peaks and is further evidence of the importance of local volcanism for the West Antarctic sulfur cycle (Reference DelmasDelmas, 1982). There are not unusually high peaks in δ³⁴S associated with the 1957 and 1972 events and this may be because they were weak eruptions relative to the 1947 Erebus event and therefore did not deposit enough SO₄ ²– to impact the isotopic signature in the same way.

The δ³⁴S value during the Agung period could be expected to be very depleted in the heavier (³⁴S) isotope and show a minimum in the dataset based on the assumption that stratospheric SO₄ ²– (which is depleted in ³⁴S) is mixing with SO₄ ^2– from a marine biogenic source. It actually shows a higher value than the samples it precedes and follows. Local volcanism (Mount Erebus) also impacted deposition during this time and may provide an explanation for this discrepancy. If the 1963 Erebus eruption caused a higher δ³⁴S in local precipitation, then the subsequent mixing with precipitation from Agung (which carries a low δ³⁴S) might provide a mechanism by which the overall measured δ³⁴S in the sample might not be as strongly affected by stratospheric air as it would during periods which are free from global volcanic events, such as the data from the 1947–50 sample.

Conclusions

The first sulfur isotope measurements have been made in a West Antarctic ice core and used to assess SO₄ ²– deposition in West Antarctica from 1935 to 1976. The measured δ³⁴S values were recalculated to reflect only the xs component of total SO₄ ²–, which in Antarctica includes only marine biogenic and volcanic contributions. These xsSO₄ ²– δ³⁴S values range from –0.7‰to 6.8‰, lower than those reported in central Antarctica (Reference Patris, Jouzel and DelmasPatris and others 2000). Based on the relatively high percentage of sea-salt SO₄ ²– in the RIDSA core, it can be assumed that there is also a substantial amount of marine biogenic SO₄ ²– in each sample. The xsδ³⁴S values are too low to represent solely a marine biogenic influence. Volcanic plume studies have shown that volcanic emissions into the stratosphere are initially positive with respect to δ³⁴S, and then become progressively lower over time due to a change in the chemical pathway of SO₄ ²– aerosol formation. δ³⁴S values decrease with time and vertical distribution after an eruption (Reference Castleman, Munkelwicz and ManowitzCastleman and others, 1973). Samples taken at the top of the stratosphere show more negative δ³⁴S values than those lower in the stratosphere. This stratospheric air, which is extremely depleted in ³⁴S, could be mixing with local, higher δ³⁴S air to produce the values seen in West Antarctica. Local volcanic emissions, specifically from Mount Erebus, also play a role in the sulfur cycle in West Antarctica, based on the correlation between increased δ³⁴S values, increased background xsSO₄ ²– and Mount Erebus eruptions.

Background peaks in xsSO₄ ²– from West Antarctica can be attributed to local volcanic input. The low resolution of each sample, however, hinders this interpretation of Antarctic SO₄ ^2– production. High-resolution δ³⁴S samples are needed to dissect further the seasonality of the isotopic signature. Once this is determined, it may be possible to apply the use of sulfur isotopes to ice-core studies to more accurately assess the past SO₄ ^2– production.