Seasonal variations of stable isotopes, density and ion concentrations

Figure 2 shows the vertical profiles of the stable isotopes of water (δ¹⁸O, δD) and deuterium-excess (d-excess; d=δD– 8δ¹⁸O) obtained from the pit. The δ¹⁸O and δD varied synchronously and showed distinct seasonal variations, with the most recent peak occurring near the snow surface of 5 July 2009. Reference SteffensenSteffensen (1985) reported that the δ¹⁸O values of snow vary in parallel with the temperatures at Dye 3, Greenland. A similar seasonality in δ¹⁸O has also been reported at other Greenland sites (e.g. Reference Finkel, Langway and ClausenFinkel and others, 1986; Reference BeerBeer and others, 1991). Therefore, we can reasonably assume that the layers with δ¹⁸O and δD maxima and minima were summer and winter layers, respectively. We dated the pit using the seasonal variation of δ¹⁸O and δD. The 2m deep pit contained snow deposited over a 4 year period. The seasonal variation of d-excess, which depends on the sea surface temperature and evaporation processes of sea water in the vapor source region (e.g. Reference UemuraUemura, 2007), differed from that of δ¹⁸O and δD, as it showed maxima in autumn and minima from spring to early summer. The seasonal variability of d-excess observed at NEEM is similar to that reported at other Greenland ice-coring sites (Reference Johnsen, Dansgaard and WhiteJohnsen and others, 1989).

Fig. 2. Vertical profiles of stable isotopes of water (δ¹⁸O and δD) and d-excess. Summer and winter were defined from maximum and minimum values of δ¹⁸O and δD, respectively.

The snow density also showed seasonal variation, which is superimposed on the general trend of increasing density with depth (Fig. 3). Low-density layers were observed at depths of approximately 0.05, 0.50, 0.90, 1.40 and 2.00 m. A comparison with the stable-isotope profiles (Fig. 2) indicates that these were summer layers. The seasonality of snow density at NEEM was similar to that observed at Summit, Greenland (Reference Albert and ShultzAlbert and Shultz, 2002), which was caused by a wind-pack effect resulting in a higher density of snow in winter. Although the wind speed at NEEM appears to be greater in winter than in other seasons, which may account for the increased snow density, this speculation needs to be confirmed using the data from the automatic weather station that was installed at NEEM during the 2009 season. The snow temperature (Fig. 3) continued to decrease from the surface to the bottom of the pit (–23.7˚C at 2.01m depth). As the visual stratigraphy of the pit showed little horizontal variability, we assume that snow and ion depositions were not seriously disturbed by wind scouring.

Fig. 3. Vertical profiles of snow temperature and density.

We found thin ice layers with thickness of ~1mm or less in the pit, at depths of approximately 0.11, 0.18 and 0.23 m, which were formed by minor surface snowmelt during the summer months at NEEM. The occurrence of slight summer surface melting did not have a noticeable impact on the profiles of stable isotopes and ions, since both showed regular variations (Fig. 2 and 4).

Fig. 4. Vertical profiles of electric conductivity, pH, and ion concentrations. Black and red curves indicate total concentrations and non-sea-salt (nss) fractions, respectively.

The vertical profiles of pH, electric conductivity and concentrations of ions (Cl^–, N₃ ^–, S₄ ^2–, MSA, Na⁺, K⁺, Mg²⁺ and Ca²⁺) are shown in Figure 4. Non-sea-salt (nss) components Oof Cl^–, SO₄ ^2–, K⁺, Mg²⁺ and Ca²⁺ were calculated using sea-water ratios of these ions with respect to Na⁺, assuming that Na⁺ is solely of sea-salt origin, and are plotted in Figure 4. For instance, concentration of nssCl^–, [nssCl^–], was calculated with the following equation:

where (Cl^–/Na⁺)_sea is the equivalent concentration ratio of Cl^–/Na⁺ in the sea water, which is 1.17. Similarly, concentrations of nssSO₄ ^2–, nssK⁺, nssMg²⁺ and nssCa²⁺ were calculated using sea-water ratios 0.12, 0.021, 0.23 and 0.044, respectively. On average, non-sea-salt fractions of Cl^–, SO₄ ^2–, K⁺, Mg²⁺ and Ca²⁺ were 26%, 96%, 66%, 47% and 92%, respectively.

Concentrations of Na⁺ and Cl^– showed very similar profiles, with pronounced seasonal variations. Both ions peaked in winter to early spring. Concentrations of Mg²⁺ also peaked in winter to early spring, while those of Ca²⁺ peaked in late winter to spring, slightly later than Na⁺, Cl^– and Mg²⁺. The concentrations of K⁺ showed a peak in winter to spring and possibly a secondary peak in summer. Concentrations of NO₃ ^– showed double peaks, one in summer and the other in winter to spring, whereas those of SO₄ ^2– peaked in winter to spring. Concentrations of MSA seem to show double peaks, one in spring and the other in late summer to autumn. Seasonal variations of pH and electric conductivity were not very clear in the present-day snow at NEEM.

The winter-to-early-spring peak observed for Na⁺ and Cl^– is in agreement with previous studies carried out on the Greenland ice sheet (Reference SteffensenSteffensen, 1988; Reference Whitlow, Mayewski and DibbWhitlow and others, 1992; Reference Candelone, Jaffrezo, Hong, Davidson and BoutronCandelone and others, 1996; Reference Dibb, Whitlow and ArsenaultDibb and others, 2007). In order to examine seasonal changes of Na⁺ and Cl^– sources, we computed Cl^–/Na⁺ ratios (Fig. 5). The ratios showed clear seasonal variations, with peaks occurring in summer when Cl^– concentrations were higher than Na⁺ concentrations with respect to the sea-water ratio. The Cl^–/ Na⁺ peaks can thus be used as an indicator of summer layers at NEEM. The high Cl^–/Na⁺ ratios together with elevated nssCl^– concentrations (Fig. 4) in summer suggest that Cl^– ions in summer Greenland snow were supplied from sources other than sea-salt aerosols, such as gaseous HCl due to sea-salt dechlorination (e.g. Reference Legrand and DelmasLegrand and Delmas, 1988; Reference Legrand, Preunkert, Wagenbach and FischerLegrand and others, 2002). On the other hand, the Cl^–/Na⁺ ratios in winter to early spring were very close to that of sea water, suggesting that in winter to early spring, Na⁺ and Cl^– mainly originated from sea-salt aerosols.

Fig. 5. Vertical profiles of Cl^–/Na⁺ and MSA/nssSO₄ ^2–. The dashed line represents the Cl^–/Na⁺ ratio in sea water (1.17).

The seasonal variation of Ca²⁺ concentrations was similar to that observed by Reference Dibb, Whitlow and ArsenaultDibb and others (2007), who collected and analyzed surface snow every day for 3 years at Summit, Greenland. They reported that Ca²⁺, which mainly originates from dust, showed a strong peak in April. As the sea-salt fraction of Ca²⁺ is small (Fig. 4), the major source of Ca²⁺ at the NEEM site would be mineral dust. Dust layers have been observed in late-winter-to-spring snow layers in Yukon, Canada (Reference Goto-Azuma, Koerner, Demuth and WatanabeGoto-Azuma and others, 2006), and over wide areas of the Greenland ice sheet (e.g. Reference SteffensenSteffensen, 1988; Reference Whitlow, Mayewski and DibbWhitlow and others, 1992; Reference Mosher, Winkler and JaffrezoMosher and others, 1993; Reference Drab, Gaudichet, Jaffrezo and ColinDrab and others, 2002). Although these sites are distant from Asia, their major sources of dust were attributed to those in Asia. Therefore, nssCa²⁺/dust transported to the NEEM site in spring would have also likely originated from Asian sources. Apportionment of Mg²⁺ by calculation of sea-salt and nss components (Fig. 4) suggests that the rising leg of the Mg²⁺ peak in winter to early spring is of sea-salt origin and that the later part of the peak is due to the influx of mineral dust.

The summer NO₃ ^– peak (Fig. 4) was consistent with previous studies carried out on the Greenland ice sheet (Reference SteffensenSteffensen, 1988; Reference BeerBeer and others, 1991; Reference Whitlow, Mayewski and DibbWhitlow and others, 1992; Reference Fischer, Wagenbach and KipfstuhlFischer and others, 1998; Reference MotoyamaMotoyama and others, 2001; Reference Dibb, Whitlow and ArsenaultDibb and others, 2007). The winter-to-early-spring peak in the NO₃ ^– concentration observed at NEEM was also observed in modern snow at Summit, Greenland (Reference Whitlow, Mayewski and DibbWhitlow and others, 1992; Reference Dibb, Whitlow and ArsenaultDibb and others, 2007), and on Agassiz Ice Cap in the Canadian High Arctic (Reference Goto-Azuma, Koerner, Nakawo and KudoGoto-Azuma and others, 1997), which was attributed to anthropogenic sources. Though Reference Finkel, Langway and ClausenFinkel and others (1986) found a major single summer peak in NO₃ ^– concentrations both in pre- and post-industrial south Greenland snow, they reported an increase of NO₃ ^– concentrations in late winter/early spring since the 1950s due to anthropogenic inputs. To determine the sources of NO₃ ^–, Reference Hastings, Steig and SigmanHastings and others (2004) analyzed nitrogen and oxygen isotopes of NO₃ ^– in snow at Summit, Greenland. Based on seasonal variations of δ¹⁵N, they reported that the predominant source of NO₃ ^– in summer was natural NO_x produced by biomass burning, biogenic soil emissions and lightning, whereas that in winter was anthropogenic NO_x from fossil fuel combustion. Since the seasonal variation of NO₃ ^– at NEEM resembled those documented at other sites in the Arctic, we suggest that the summer and winter-to-early-spring peaks of NO₃ ^– at NEEM are also mainly of natural and anthropogenic origin, respectively. For confirmation, further studies using isotopes, etc., are necessary at NEEM.

The seasonal variations of MSA in the snow at NEEM coincided with those in the air and snow at Dye 3 and Summit, Greenland (Reference Li, Barrie, Talbot, Hariss, Davidson and JaffrezoLi and others, 1993b; Reference Jaffrezo, Davidson, Legrand and DibbJaffrezo and others, 1994), and with those in the air at Alert in the Canadian High Arctic (Reference Li, Barrie and SiroisLi and others, 1993a). MSA is an oxidization product from dimethylsulfide (DMS), which is mainly produced by marine phytoplankton (e.g. Reference Legrand and MayewskiLegrand and Mayewski, 1997). The spring peak in MSA may have been caused by enhanced photochemical activity after the polar sunrise, when solar radiation becomes available to oxidize the winter reservoir of DMS (Reference Li, Barrie, Talbot, Hariss, Davidson and JaffrezoLi and others, 1993b). Alternatively, it could have been caused by long-range transport from lower-latitude oceans (Reference Li, Barrie, Talbot, Hariss, Davidson and JaffrezoLi and others, 1993b). The MSA peak in late summer to autumn, when sea ice around Greenland has retreated, could have been produced by regional DMS emissions.

The potential sources of SO₄ ^2– include sea salt, volcanic eruptions, SO₂ produced by fossil fuel combustion, dust, and DMS emissions produced by marine phytoplankton (Reference Legrand, De Angelis, Cachier, Gaudichet and DelmasLegrand and others, 1995; Reference LegrandLegrand, 1997; Reference Legrand and MayewskiLegrand and Mayewski, 1997). As can be seen in Figure 4, sea salt is only a minor component of SO₄ ^2– at the high-elevation site NEEM. Large-magnitude volcanic eruptions produce spikes in SO₄ ^2– concentration (e.g. Reference Legrand and MayewskiLegrand and Mayewski, 1997). They are, however, sporadic and do not make a substantial contribution to the SO₄ ^2– budget in years without large explosive volcanism. The SO₄ ^2– peak in winter to spring (Fig. 4) has also been documented at other ice-sheet sites in north, central and south Greenland (Reference Finkel, Langway and ClausenFinkel and others, 1986; Reference BeerBeer and others, 1991; Reference Whitlow, Mayewski and DibbWhitlow and others, 1992; Reference Jaffrezo, Davidson, Legrand and DibbJaffrezo and others, 1994; Reference Fischer, Wagenbach and KipfstuhlFischer and others, 1998; Reference BiglerBigler and others, 2002; Reference Dibb, Whitlow and ArsenaultDibb and others, 2007) and at an ice-cap site in the Canadian High Arctic (Reference Goto-Azuma, Koerner, Nakawo and KudoGoto-Azuma and others, 1997). A similar seasonal variation was seen in nssSO₄ ^2– in the air in Greenland and the Canadian High Arctic (Reference Barrie and BarrieBarrie and Barrie, 1990; Reference Jaffrezo, Davidson, Legrand and DibbJaffrezo and others, 1994; Reference NormanNorman and others, 1999). Studies in central and south Greenland and in the Canadian High Arctic suggest that the winter-to-early-spring peak was due to the inflow of anthropogenic air pollutants. Similarly, the winter-to-early-spring peak of SO₄ ^2– at NEEM may also be attributed to air pollutants produced by fossil fuel combustion. On the other hand, Reference BiglerBigler and others (2002) reported that a single winter-to-early-spring peak has been observed both in preindustrial and modern snow in north Greenland, though a small shift in the seasonality between them was seen. This would imply that in north Greenland both natural and anthropogenic SO₄ ^2– peak in winter to early spring, in contrast to central and south Greenland and to the Canadian High Arctic, where natural SO₄ ^2– peaks in summer (Reference Finkel, Langway and ClausenFinkel and others, 1986; Whitlow and others, 1992; Reference Goto-Azuma, Koerner, Nakawo and KudoGoto-Azuma and others, 1997; Reference Dibb, Whitlow and ArsenaultDibb and others, 2007). If this is the case, the winter-to-early-spring SO₄ ^2– peak in the modern snow at NEEM could also have both anthropogenic and natural sources.

We now discuss the contributions of mineral dust and oxidation products of DMS to the SO₄ ^2– budget at NEEM. The spring maxima of nssSO₄ ^2– were seen at depths around 0.11, 0.57, 1.14 and 1.71 m (Fig. 4), where nssCa²⁺ also displayed the spring maxima, indicating that CaSO₄ is one of the major sources of both SO₄ ^2– and Ca²⁺ in spring. Earlier studies showed that SO₄ ^2– and Ca²⁺ peaks in spring were associated with episodes of increased atmospheric dust (e.g. Reference BarbanteBarbante and others, 2003). The reaction of CaCO₃ from Asian dust sources with H₂SO₄ (e.g. Reference Legrand and MayewskiLegrand and Mayewski, 1997) during the transport to northwest Greenland may have largely contributed to the spring maximum of SO₄ ^2– and Ca²⁺ concentrations. Nevertheless, nssSO₄ ^2–/nssCa²⁺ ratios at the spring maxima were larger than the stochastic ratio of CaSO₄, which suggests that other nssSO₄ ^2– sources are also important. Anthropogenic input of nssSO₄ ^2– would be an important source particularly during winter and early spring, as was the case for Summit and Dye 3, Greenland (Reference Finkel, Langway and ClausenFinkel and others, 1986; Reference Mann, Shuman, Kelly and KreutzMann and others, 2008), and for Alert, a Canadian High Arctic site (Reference Nriagu, Coker and BarrieNriagu and others, 1991).

The two major final products of the DMS oxidation processes are nssSO₄ ^2– and MSA. Different timings of nssSO₄ ^2– and MSA peaks could imply that DMS is not a predominant source of nssSO₄ ^2– at NEEM. A closer look at the nssSO₄ ^2– seasonality, however, might indicate that the falling leg of the nssSO₄ ^2– peak in spring overlapped at least part of the MSA peak in spring, and that a nssSO₄ ^2– peak in summer to autumn, which was generally very small except in 2007, was seen in the rising leg of the MSA peak in summer to autumn. Hence, we cannot yet rule out the possibility that nssSO₄ ^2– originated from DMS might have shared a noticeable fraction of nssSO₄ ^2– in spring and late summer to autumn. We plot the MSA/nssSO₄ ^2– ratios in Figure 5. The ratio showed a major peak in late summer to autumn, with a minor peak or shoulder in spring. As our study covers only 4 years, and as our pit data are subject to noise, we cannot conclude that the minor peaks or shoulders in nssSO₄ ^2–, MSA and MSA/nssSO₄ ^2– occurred regularly over multiple years. In addition, we cannot yet draw a firm conclusion about the source apportionment of nssSO₄ ^2–. Accordingly, interpretation of the seasonality of MSA/nssSO₄ ^2– is rather complex. Studies using sulfur isotopes and those covering longer periods are needed.

Surface mass balance

We calculated the annual surface mass balance for the years 2006–08 at NEEM (Table 1). Each year was defined using the profile of the stable-isotope ratio, so that the border between the two adjacent years was located at the layer with a stable-isotope minimum. To examine the seasonal distribution of snow accumulation, we divided one year into two parts: one representing the first half of the year (stable-isotope minimum to maximum) and the other representing the second half of the year (stable-isotope maximum to minimum). Here we define the former as ‘winter to summer’ (winter–spring– summer) and the latter as ‘summer to winter’ (summer– autumn–winter). The annual mean surface mass balance during the 3 years was 176 mm w.e., which is substantially less than the 0.26ma^–1 estimated by Reference Buchardt and Dahl-JensenBuchardt and Dahl-Jensen (2008). The surface mass balance in 2006 (2005 winter to 2006 winter) was 203 mm, which was the maximum of the three years. On the other hand, the surface mass balance in 2008 (2007 winter to 2008 winter) was less than in 2006 by ~50mm. We found that the year-to-year variability of surface mass balance at NEEM was >20%. The snow depositions during the winter-to-summer and summer-to-winter seasons were 98 and 78 mm, respectively. Snow deposition during the winter-to-summer half of the year was greater than that during the other half of the year by 10– 20 mm for all three years. As these findings are based only on the past 3 years, they need to be further investigated by examining both shallow and deep ice cores drilled at NEEM.

The depositions of Na⁺, K⁺, Mg²⁺, Ca²⁺, Cl^–, NO₃ ^– and SO₄ ^2– showed large seasonal differences (Table 1). The 2007 winter to 2008 summer season had substantially greater depositions of Na⁺, K⁺, Mg²⁺ and Cl^– than any other seasons. During this season, deposition of these ions was >1.5 times the seasonal mean, which is due to the high peak around 0.6 m depth. Separation of non-sea-salt and sea-salt fractions (Fig. 4) indicates that the sea salt, not dust components, showed a pronounced peak in early spring of 2008. In general, all the ions but MSA show greater deposition during winter to summer than summer to winter. For instance, average depositions of major sea-salt components Na⁺, Cl^– and Mg²⁺ during winter to summer were about two times greater than those during the summer to winter. Ca²⁺ showed an even greater seasonal difference, which was more than threefold. These seasonal differences were greater than those observed for H₂O. Depositions of K⁺, NO₃ ^– and SO₄ ^2– were also greater during the winter to summer. Among all the ionic species, only MSA showed greater deposition in summer to winter. This study demonstrated that the depositions of most chemical components studied were highest in the winter-to-summer half of the year at NEEM.