Introduction

NO₃ ^– in polar snow and firn layers is "reversibly" deposited, and undergoes significant post-depositional exchange between snow/firn and the atmosphere, provided decreasing atmospheric concentrations result in supersaturated concentrations in the snow (Reference De Angelis, Legrand and DelmasDe Angelis and Legrand, 1995). This loss by re-equilibration is moderated by the rate of snow accumulation, with greater losses possible where low accumulation keeps a given layer of snow in "contact" with the atmosphere for longer periods (Reference De Angelis, Legrand and DelmasDe Angelis and Legrand, 1995). However, it is still unclear how such processes would differ in alkaline snow layers from the Qinghai-Tibetan Plateau (QTP). Ice cores drilled from the polar regions and the QTP for paleo-reconstruction necessitate a better understanding of the relationship between the chemical concentrations in the atmosphere and in the ice.

In this paper, we compare the NO₃ ^– profiles of the snow pits in central East Antarctica (Fig. 1) with those in the northern QTP (Fig. 2), and try to account for the post-depositional differences between these two very different areas with very different chemical and physical conditions in the local atmosphere.

Fig. 1. Map of Antarctica showing locations of snow pits.

Fig. 2. Map of the head of Ürümqi River, Tien Shan, China, showing sampling locations.

Post-Depositional Modification of NO³− in East Antarctica

Figure 3 shows NO₃ ^– profiles from snow and ice cores from South Pole (accumulation rate 80 kg m-1 ^a-1), Vostok (23 kg m⁻² a⁻¹) and Dome C (34 kg m⁻² a⁻¹), Antarctica. Compared to the mean, and fairly constant, NO₃ ^– concentration of 100 ng g⁻¹ observed at South Pole during the last several decades, cores from Vostok and Dome C exhibit lower NO₃ ^– concentrations before 1975 and a sharp increase since then.

Fig. 3. NO₃ ^– profiles from ice drilled at South Pole, Vostok and Done C,Antartica adapted from Reference Mayewski and LegrandMayewski and Legrand, 1990).

Since all the sites are located on the high Antarctic plateau, it is unlikely that the difference by a factor of 5 between NO₃ ^– levels at these sites before 1975 really reflects different atmospheric NO₃ ^– levels (Reference Legrand, Leopold, Domine, Wolff and BalesLegrand and others, 1996). Instead, it has been proposed that low NO₃ ^– concentrations in deep snow layers at sites like Dome C and Vostok reflect a re-emission of NO₃ ^– from the snow after deposition (Reference Mayewski and LegrandMayewski and Legrand 1990; Reference De Angelis, Legrand and DelmasDe Angelis and Legrand, 1995). Reference Legrand, Leopold, Domine, Wolff and BalesLegrand and others (1996) further suggested that a snow-accumulation rate in the order of 80 kg m–2 a–1 (the accumulation rate at South Pole) is large enough to strongly limit post-depositional effects.

Another three 2 m snow pits dug at South Pole in 1994 provide detailed (2 cm resolution) NO₃ ^– profiles for the period 1987-94 (Reference Dibb and WhitlowDibb and Whitlow, 1996). The annual spring NO₃ ^– peaks increase progressively toward the surface, although Reference Mayewski and LegrandMayewski and Legrand (1990) noted that the highest NO₃ ^– peak in the South Pole snow corresponded to spring 1987, which was the year of the deepest ozone hole measured up to that point. These three new NO₃ ^– profiles, however, show no correspondence to the severity of ozone depletion, and the very high NO₃ ^– concentrations in the 1987 snow layer at 20 cm depth in the 1988 snow pits are not apparent in the 1987 layer in any of the 1994 snow pits. This implies that even at a relatively high-accumulation area such as South Pole, interpretation of NO₃ ^– profiles must still account for post-deposition losses. In addition, it is noticed that the mean NO₃ ^– concentration of these snow pits is around 100 ng g⁻¹, similar to the previous results (Reference Mayewski and LegrandMayewski and Legrand, 1990).

Two 2 m snow pits were sampled in central East Antarctica along the route of the 1990 International Trans-Antarctic Expedition (ITAE). Snow pit 1 (85°53’S, 105°29’E; 3150 ma.s.l.) was sampled on 27 December 1989. The annual mean temperature is –51.6°C and the annual accumulation is 44 kg m⁻² a⁻¹. Snow pit 2 (81 °50’ S, 106°28’ E; 3310 m a.s.l.) was sampled on 8 January 1990. The annual mean temperature is –53.9°C and the annual accumulation is 33 kg m⁻² a⁻¹. The detailed sampling and analysis processes are described elsewhere (Reference Dahe, Zeller and DreschhoffQin and others, 1992). Figure 4 shows NO₃ ^– profiles of the two snow pits with significant NO₃ ^– spring peaks in the surface snow layers. The maximum NO₃ ^– concentrations in pits 1 and 2 are 575 and 198 ng g⁻¹, respectively, and the mean NO₃ ^– concentrations are 67 and 17 ng g⁻¹, respectively.

Fig. 4. NO₃ ^– profiles from two snow pits on the ITAE route in central East Antarctica.

The three independent experiments mentioned above at least validate the facts that (1) the annual spring NO₃ ^– peaks in the uppermost surface snow layers are distributed widely in central East Antarctica, and (2) high NO₃ ^– levels are present in the South Pole region (snow pit 1), while low NO₃ ^– levels are evident in the Vostok-Dome C region (snow pit 2). Similar NO₃ ^– snow-pit profiles from a wide spatial and temporal scale clearly demonstrate the post-depositional modification of NO₃ ^– in snow layers of East Antarctica.

Candidate processes of NO₃ ^– loss include photochemical destruction of NO₃ ^–, wind-scouring of the surface snow layer, loss of HNO_s in the vapour phase, and scavenging of NO₃ ^– by other wind-blown particulates or gases, probably aided by windpumping of the snowpack (Reference Pomeroy, Jones, Wolff and BalesPomeroy and Jones, 1996). However, the difference by a factor of about 4 between NO₃ ^– levels in these two regions of East Antarctica might be accounted for by atmospheric ionization products (Reference Dahe, Zeller and DreschhoffQin and others, 1992). Thus the annual mean NO₃ ^– concentration in snow might be proportional to its annual mean concentration in the atmosphere despite the post-depositional process, as long as the depositional environment does not vary significantly (Reference YangYang and others, 1995).

Post-Depositional Processes at the Northern QTP

Four snow pits were sampled at the east branch of Urumqi Glacier No. 1 at the head of Urumqi River, Tien Shan, China (shown as SP in Fig. 2), on 25 December 1995 and 8,19 and 29 January 1996. They are designated snow pits A, B, G and D, respectively. In addition, surface snow sampling was conducted during the period 12-29 January 1996 (shown as Dl–D5 in Fig. 2). During sampling, personnel wore disposable polyethylene gloves to minimize contamination. After sampling, the snow pit was partially refilled, and sampling on successive days involved digging out the pit and refacing the sampling surface by at least 1 m. The same strata were resampled on each occasion after allowing for accumulation or ablation. Snow samples were transferred into pre-cleaned polyethylene bags with plastic scoops, then transferred into a pre-cleaned polyethylene container after melting. For surface snow sampling, the uppermost stratigraphic layer was cut and scraped into a cubic block with a clean stainless-steel knife for further processing as discussed above. Because of the surface irregularities mainly caused by snowdrifting, the depth of the uppermost stratigraphic layer varied from 5 to 10 cm. Each sample was taken from a fresh surface slightly upwind of the previously disturbed area.

Bottled samples were returned frozen to the laboratory and kept in a cold room at –20°C until chemical analysis was performed in a class 100 clean room. The cations were determined by a PE-2380 atomic absorption spectrometer, and the anions by Dionex DX-100 ion chromatography using the IonPac AS4A-SG column system and an isocratic carbonate/bicarbonate eluent. Precision is estimated to be within <5%, with a detection limit of 5 ng g⁻¹ for all the ions discussed (Reference Cuilan, Jianchen, Li and ShuguiHuang and others, 1998).

The NO₃-, SO₄ ²- and Ca²⁺ profiles of the snow pits are shown in Figure 5. Similar trends are observed among these profiles. The peak ion concentrations below 20 cm depth correspond to the dirty layer formed the previous autumn. Extremely high NO₃ ^– and SO₄ ²- concentrations in snow pit C might reflect the spatial variance, but ion elution processes may also contribute since snowmelt percolation is involved in the firnification and densification processes during autumn. However, all the ions considered remain rather stable in the low parts of the snow pits, while they increased approximately 1.6-, 2.5- and 6.1-fold for NO₃- SO₄ ²- and Ca²⁺, respectively, in the uppermost snow layers between 25 December 1995 and 29 January 1996.

Fig. 5. NO₃ ^–, SO₄₂ ^–andCa2⁺ snow-pit profiles from samples collected at Ürümqi Glacier No. 1 at the head of Ürümqi River.

The NO₃ ^– and SO₄₂ ^– (no data available for Ca²⁺) profiles of the surface snowpack from the D1-D5 sites are shown in Figure 6. Though spatial variation of the NO₃ ^– concentrations can be identified at the five sampling sites, it is also apparent that the NC3- and SO4₂- concentrations increase during the sampling period, and they display similar trends. No apparent precipitation event occurred over the period considered, but this does not rule out possible changes in the surface snow layer that may be taking place due to slight scouring by the wind, or accumulation of light snowfall.

Fig. 6. NO₃ ^– and SO42- I profiles of the surface snowpack at the head of Ürümqi River.

Discussion and Conclusions

From rocket measurements it has been reported that the principal ions at very high altitudes (80-90 km) appear to be hydrated NO,-, possibly in the form of small agglomerates (Reference Narcisi, Bailey, Delia Lucca, Sherman and ThomasNarcisi and others, 1971). It has also been found that polar stratospheric clouds contain substantial quantities of nitric acid (Reference McElroy, Salawitch and WofsyMcElroy and others, 1988). Reference Wofsy, Molina, Salawitch, Fox and McElroyWofsy and others (1988) show that nitric acid can be removed from the gas phase during very low stratospheric temperatures in the polar night and early spring. Thus, NO₃ ^– in polar snow is present mainly in the form of HNO₃ (in the gas phase) rather than as aerosol-associated NO₃ ^– (p-NCV) (Reference Legrand and KirchnerLegrand and Kirchner, 1990).

By contrast, Reference Lyons, Mayewski, Spencer and TwicklerLyons and others (1990) report that crustally derived NO," from soils appears to be the main source of NO₃ ^– m snow samples from central Asia. Reference Williams, Tonnessen, Melack and DaqingWilliams and others (1992) performed comprehensive research on the sources and spatial variation of the chemical content of winter snowpack at the head of Ürümqi River, and confirmed that terrestrial dust may be the primary source of NO₃ ^– in the snowpack. Their analysis of variance tests indicated that NO₃ ^– was grouped with Ga²⁺ (mainly in the form of CaCO₃ in the atmosphere of China (Reference Galloway, Dianwu, Jiling and LikensGalloway and others (1987)) and ANG (acid-neutralizing capacity, representing primarily carbonate/bicarbonate). Since the primary source of Ga²⁺ and ANG in the snowpack may be aeolian dust, the grouping of NO₃ ^– with Ga²⁺ and ANG suggests that NO₃ ^– should have a similar source at the head of Ürümqi River.

In addition, crustal dust in the form of CaS0⁴ is a potential source of SO₄ ²-, which is consistent with the high correlation between Ca₂ ⁺ and SO₄ ²- (Reference Williams, Tonnessen, Melack and DaqingWilliams and others, 1992). Therefore similar trends for the N0₃- S0₄2^- and Ca²⁺ snow-pit profiles (Fig. 5) and for the NO₃ ^– and SO4₂ ^- variances in the surface snowpack (Fig. 6) confirm the co-deposition of NO₃ ^– with neutral associated species.

Reference JunyingSun and others (1998) suggest that anthropogenic emissions are a more important source of NO₃ ^– than dust, based on simultaneous sampling of aerosol and snow at Ürümqi Glacier No. 1 between 19 May and 29 June 1996. However, they also propose the presence of NH₄NO₃ in snow samples because the equivalence ratio of ammonium to sulfate exceeds 1 in some cases. Neutralization of nitric acid by ammonia was observed to have the effect of fixing NO₃ ^– and preventing re-emission at several coastal Antarctic sites, though the efficiency of this process has yet to be determined (Reference Mulvaney, Wagenbach and WolffMulvaney and others, 1998). Unfortunately, NH₄ ⁺ measurements were not performed on either our snow-pit or surface-snow samples; but under an alkaline atmospheric environment, gaseous HNO₃ can be absorbed on the surface of alkaline particles and react to form salts (Reference Mamane and GottliebMamane and Gottlieb, 1992), which could cause NO₃ ^– to conform to the seasonality of the fixing agent (Reference Mulvaney, Wagenbach and WolffMulvaney and others, 1998).

Reference Legrand, Lorius, Barkov and PetrovLegrand and others (1988) reported enhanced NO₃ ^– concentrations in ice deposited during the last glacial period. Though based on a good correlation between calcium and NO₃ ^– concentrations along the Vostok ice core, it is suggested that the NO₃ ^– increase corresponds to an additional input of neutral associated species. However, it is still uncertain whether or not the observed NO₃ ^– increase during cold stages reflects higher atmospheric levels of NO₃ ^– related to the scavenging of HNO₃ by dust particles (e.g. CaCO3) or less efficient losses due to the enhanced presence of dust particles in the snow grains (Reference Legrand, Leopold, Domine, Wolff and BalesLegrand and others, 1996). Evaluation of this problem might be facilitated by a better understanding of the NO₃ ^– depositional mechanism in the northern QTP.