Introduction

Sulphur-rich gases, principally sulphur dioxide (SO₂) and hydrogen sulphide (H₂S), emitted by volcanic eruptions have the capability of cooling global climate by a few tenths of a degree for several years following the eruption (Reference ZielinskiZielinski, 2000, and references therein). for cooling to occur, these gases must be injected into the stratosphere where they are rapidly (within a month of the eruption (Reference Bluth, Schnetzler, Krueger and WalterBluth and others, 1993)) oxidized to sulphate (SO₄ ^2– ) and sulphuric acid (H₂SO₄).These aerosols are then rapidly advected around the globe with a time-scale of 2–3weeks. Meridional transport is much slower, with transport to the polar regions taking about 1–2 years (Reference RobockRobock, 2000, and references therein). the spatial distribution of volcanic aerosols depends upon the location of the eruption, the time of year, the location of the intertropical convergence zone and the quasi-biennial oscillation (Reference ZielinskiZielinski, 2000). Equatorial eruptions have the greatest potential to influence climate as their aerosols can be transported to both hemispheres. for example, the Mount Pinatubo (Philippines) eruption in June 1991 injected into the atmosphere about 18±2610⁹ kg of SO₂ (Reference KruegerKrueger and others, 1995). This volcanic plume had spread around the Earth by mid-1992, reducing global tropospheric and surface temperatures by 0.2– 0.7˚C (Reference McCormick, Thompson and TrepteMcCormick and others, 1995; Reference Jones, Kelly, Fiocco, Fu’a and ViscontiJones and Kelly,1996).

Fallout from explosive eruptions like Mount Pinatubo is captured in the polar and tropical ice caps around the globe, as first studied by Reference HammerHammer (1977). Ice-core volcanic signatures have been used to estimate the optical depth of eruptions prior to instrumental records (Reference ZielinskiZielinski, 1995, Reference Zielinski2000) and the H₂SO₄ flux produced by a given eruption (e.g. Reference Moore, Narita and MaenoMoore and others, 1991; Reference Delmas, Kirchner, Palais and PetitDelmas and others, 1992; Reference Clausen and HammerClausen and others, 1997; Reference Cole-Dai, Mosley-Thompson and ThompsonCole-Dai and others, 1997, 2000). These calculated fluxes can then be compared to the flux of known volcanic signatures such as El Chichόn, Mexico (AD1982), and Mount Pinatubo.

Volcanic signals in ice cores have been detected by electrical conductivity (e.g. Reference ClausenClausen and others, 1997), dielectric profiling (Reference Moore, Narita and MaenoMoore and others, 1991) and trace chemical analysis (e.g. Reference Delmas, Kirchner, Palais and PetitDelmas and others, 1992; Reference ZielinskiZielinski, 1995; Reference Cole-Dai, Mosley-Thompson, Wight and ThompsonCole-Dai and others, 2000). the first two methods measure the total acidity of the ice including nitric acid, and other minor acids such as hydrochloric and hydrofluoric acid. the trace chemistry method gives a direct measure of the quantity of SO₄ ^2– in the ice.

The SO₄ ^2– records presented here were obtained by chemical analysis (Reference Curran and PalmerCurran and Palmer, 2000) of three ice cores drilled near the summit of Law Dome, a small ice cap abutting the edge of the main East Antarctic ice sheet around 112˚ E (Fig. 1). the Law Dome summit region is characterized by high accumulation (0.7ma^–1 ice equivalent (IE)), relatively low wind speeds (8.3ms^–1) and low mean temperatures (–21.8˚C; Reference Morgan, Wookey, Jun, van Ommen, Skinner and FitzpatrickMorgan and others, 1997). Reference Palmer, van Ommen, Curran, Morgan, Souney and P.A. MayewskiPalmer and others (2001) used multi-parameter annual-layer counting to precisely date the three ice-core records with an age uncertainty of ±1 year at AD 1301. Here, we investigate the magnitude of the deposited volcanic H₂SO₄ fluxes in the Law Dome ice cores and other previously published Antarctic ice-core records.

Fig. 1 Map of the Antarctic ice cores used in this study, with the Tambora flux values in parentheses (units of kg km^–2).

The Volcanic Signal

The main sources of SO₄ ^2– in Antarctic ice include sea salt, marine biogenic activity and volcanic fallout. Sea-salt input peaks in late winter with the maximum in storm activity. This noisy signal was removed using the sodium (Na⁺) concentration as an indicator of sea-salt input. the non-sea-salt SO₄ ^2– (nssSO₄ ^2–) component was calculated using the total SO₄ ^2– and Na⁺ concentrations (units μeq L^–1) and the sea-salt ratio:

where r= 0.0865 is the sea-salt ratio of SO₄ ^2– to Na⁺ corrected for sea-salt fractionation processes at Law Dome (personal communication from M. Curran, 2001). This correction was calculated using a similar technique to that described by Reference Hall and E.W.Hall and Wolff (1998).

Marine biogenic activity results in a summer peak in SO₄ ^2– at Law Dome. This relatively constant signal was removed from the nssSO₄ ^2– record by subtracting a mean seasonal cycle calculated for the entire 695 year record using the technique described by Reference Van Ommen and MorganVan Ommen and Morgan (1996). Periods of known volcanic activity identified by Reference Palmer, van Ommen, Curran, Morgan, Souney and P.A. MayewskiPalmer and others (2001) were neglected in the calculation of the mean seasonal cycle.

The amount of volcanic H₂SO₄ preserved in the Law Dome ice-core record for a volcanic signal was estimated from this residual nssSO₄ ^2– record by summing the residual H₂SO₄ flux for each sample to give a total flux for the duration of the eruption signature. the individual sample fluxes (f_s, in kg km^–2 a^–1) were calculated using

where [nssSO₄ ^2–]_residual is the residual nssSO₄ ^2– concentration in units of mg kg^–1, MWH₂SO₄ and MWSO₄ ^2– are the molecular weights of H₂SO⁴ and SO₄ ^2–, ρ =917 kg m^–3 (density of ice (Reference PatersonPaterson, 1994)), and l is the depth that would have been covered by a sample (in IE) when it was deposited on the surface (typically 0.05 mIE). That is, the sample length has been corrected to account for the apparent reduction in flux due to the spreading of the ice sheet.

The residual annual H₂SO₄ flux record for Law Dome is shown in Figure 2. the 20 identified periods of volcanic activity as determined by Palmer and others (2001) are also shown in Figure 2 as shaded regions. Three periods of volcanic activity, denoted by stars in Figure 2, contain multiple eruption signatures that were unable to be clearly resolved with the time-scale graphed in Figure 2.

Fig. 2 The residual annual H₂SO₄ flux record (AD1301–1995). Shaded regions highlight periods of volcanic activity in the Law Dome record (Reference Palmer, van Ommen, Curran, Morgan, Souney and P.A. MayewskiPalmer and others, 2001), with starred regions containing multiple eruption signatures.

Spatialvariability of Volcanic Fluxes

Volcanic H₂SO₄ fluxes across the Antarctic continent vary considerably. for example, the Tambora (Indonesia) eruption (Fig.1), a reference horizon found in most polar ice-core records, has fluxes estimated between 22.4 kg km^–2(Plateau Remote (Reference Cole-Dai, Mosley-Thompson, Wight and ThompsonCole-Dai and others, 2000)) and 133 kg km^–2 (Siple Station (Reference Cole-Dai, Mosley-Thompson and ThompsonCole-Dai and others, 1997)). the Tambora flux in the Law Dome ice-core record was 79.8 kg km^–2. Spatial variability studies at South Pole calculated only a 20% variation in nssSO₄ ^2– fluxes for the Mount Pinatubo eruption over a calculated 400 km² area (Reference Cole-Dai, Mosley-Thompson and DaheCole-Dai and Mosley-Thompson, 1999). Studies at Greenland using five ice-core records show similar variability (17%) for the Tambora eruption (Reference Clausen and HammerClausen and Hammer, 1988). the variability seen in the South Pole and Greenland studies may be attributed to site characteristics such as surface irregularities (e.g. sastrugi), temperature, wind speed and surface elevation that modulate the local accumulation and hence the flux. It is also possible that some of the variability between the Antarctic ice-core records is a result of the different analytical techniques used to measure the volcanic signal in the ice and calculate the H₂SO₄ flux values (in particular, removal of the non-volcanic background).

¹ This work.

² Siple andDyer denote the Siple Station and Dyer Plateau ice-core records (Reference Cole-Dai, Mosley-Thompson and ThompsonCole-Dai and others, 1997). the unknown volcanic events S9 and D9 (AD1695– 97) and S11 and D11 (AD 1673–75) (Reference Cole-Dai, Mosley-Thompson and ThompsonCole-Dai and others, 1997) are attributed to Unknown (AD1695.8) and Gamkonora, respectively.

³ G15 ice-core record was drilled on the Mizuho Plateau (Reference Moore, Narita and MaenoMoore and others, 1991).

⁴ the PS1 ice-core record is from South Pole (Reference Delmas, Kirchner, Palais and PetitDelmas and others, 1992). the unknown events in AD 1641,1450 and 1340 are attributed to Parker, Kuwae and Unknown (AD1345.1), respectively.

⁵ the NBY89 ice core is from Byrd Station (Reference Langway, Osada, Clausen, Hammer, Shoji and MitaniLangway and others, 1994). the volcanic events dated AD 1648 and 1605 are attributed to Parker and Huaynaputina, respectively.

⁶ PR denotes the Plateau Remote ice core (Reference Cole-Dai, Mosley-Thompson, Wight and ThompsonCole-Dai and others, 2000). the volcanic event PR7 (AD1671) is attributed to Gamkonora. the Krakatau flux ratio was reported by Reference Cole-Dai, Mosley-Thompson, Wight and ThompsonCole-Dai and others (2000) as a combination of two volcanic events in the PR ice core. Hence this and the Kuwae flux (due to its magnitude) were not included in the mean flux calculations.

The volcanic signals observed in the H₂SO₄ flux records from the various Antarctic sites were compared using the ratio of a volcanic signal to the well-documented volcanic eruption of Tambora (e.g. Reference Moore, Narita and MaenoMoore and others, 1991; Reference Cole-Dai, Mosley-Thompson and ThompsonCole-Dai and others, 1997). the use of ratios removes much of the inter-site and analytical variability described above.

Seven Antarctic ice-core records (Fig.1) were compared for 13 volcanic eruptions where the calculated flux ratios (f/f_T) are reported in Table 1. the greatest f/f_T variability was observed for the Kuwae (Vanuatu) eruption for which the Plateau Remote f/f_T value appears to be an outlier imposing the need for caution when using this record, as discussed by Reference Cole-Dai, Mosley-Thompson, Wight and ThompsonCole-Dai and others (2000).

Table 1. The volcanic H₂SO₄ flux ratio with respect to the Tambora eruption for seven Antarctic ice cores. the Law Dome ice date for each eruption (Reference Palmer, van Ommen, Curran, Morgan, Souney and P.A. MayewskiPalmer and others, 2001) appears in parentheses. the total Law Dome H₂SO₄ flux for the Tambora eruption was 79.8 kg km^–2

Seven eruptions identified in the Law Dome record (Mount Pinatubo; Cerro Hudson, Chile; El Chichόn; Galunggung, Indonesia;Raung, Indonesia;Kelut, Indonesia; and Billy Mitchell, Bougainville) are not included in Table 1, as there were no comparative f/f_T values in the other records examined. However the Mount Pinatubo and Cerro Hudson f/f_T may be estimated from two South Pole ice cores (0.16 and 0.04, respectively (Reference Cole-Dai, Mosley-Thompson and DaheCole-Dai and others, 1999)), where the Law Dome f/f_T was 0.31 and 0.10, respectively. Thetrends between these ratios are similar even though the ratios are quite different for the two sites.

The average flux ratio for the volcanic events (excluding Kuwae due to its magnitude) in Table 1 is 0.32 (σ = 0.14). That is, the magnitude of the low-latitude eruption signatures preserved in the Antarctic ice cap is about one-third that of the Tambora H₂SO₄ flux. Contrasting this, the Kuwae signature had a flux similar to or greater than that ofTambora, depending upon the site, so this eruption might have had an impact on the Earth’s climate similar to that of the Tambora eruption, the year following which (AD1816) was referred to as the ``year without a summer’’. However, there are few historical records supporting a similar global climatic change following the Kuwae eruption as dated by Reference Palmer, van Ommen, Curran, Morgan, Souney and P.A. MayewskiPalmer and others (2001).

Further evidence for a complex relationship between volcanic aerosols and climate is provided by the Huaynaputina (Peru) eruption in AD 1600. the average f/f_T for this eruption from seven Antarctic ice cores is 0.30 (σ = 0.08), about one-third of the H₂SO₄ flux deposited following the Tambora eruption. A temperature time series from a composite of Northern Hemisphere tree-ring records reports the severest short-term cooling event of the past 600 years in AD 1601, the year following the Huaynaputina eruption (Reference Briffa, Jones, Schweingruber and OsbornBriffa and others, 1998). This suggests that this volcanic eruption had a much larger impact on the Earth’s climate than suggested by the Antarctic ice-core H₂SO₄ fluxes. One explanation for this could be that atmospheric circulation patterns favored the dispersal of volcanic aerosols north of the Equator, creating a larger climatic impact in the Northern Hemisphere.

Concluding Remarks

Of the 20 volcanic eruptions preserved in the ~700 year Law Dome record, the Tambora and Kuwae signatures had the largest H₂SO₄ fluxes. Spatial variability studies using 13 volcanic signatures from seven Antarctic sites showed that, on average, low-latitude volcanic eruptions of global importance emit one-third of the H₂SO₄ produced during the explosive Tambora eruption. However, the H₂SO₄ volcanic fluxes determined from Antarctic ice cores do not provide a complete picture of the climatic impact of explosive eruptions, and further work comparing various paleoenvironmental records is required to achieve this.