1. Introduction
Glaciers and ice sheets play an important role in the understanding of present and past environmental changes, as well as providing information on the areal deposition patterns of nuclear and volcanic events. Polar glaciers are known to preserve these records systematically for many thousands of years due to the minimal melting of ice in Arctic and Antarctic regions (Reference Delmas, Briat and LegrandDelmas and others, 1982; Reference Von Gunten, Rössler and GäggelerVon Gunten and others, 1983; Reference JouzelJouzel and others, 1987). In contrast, temperate glaciers (located in high altitudes) hold potential records of environmental changes covering time-scales of a few centuries (Reference ThompsonThompson and others, 1997). In this context, radioactive isotopes serve as excellent time markers, while stable isotopes are sensitive indicators of climatic change. The natural radioisotopes of both short and long half-lives (3H = 12.3 years; 210Pb = 22.3 years; 32Si = 178 years (as revised by Reference Nijampurkar, Rao, Oldfield and RenbergNijampurkar and others, 1998); 14C = 5730 years) and the artificial radioisotope 137Cs (half-life = 30 years) have been used to derive the accumulation rates and ages of young and older ice in polar and temperate regions. The stable-isotope composition of snow (δ 18O and δD) deposited in the polar and mid-latitude regions depends mainly on the air temperature and sources of moisture (Reference Nijampurkar and BhandariNijampurkar and Bhandari, 1984; Reference Peel, Mulvaney and DavisonPeel and others, 1988). The δ 18O and δD measurements, therefore, can be used to understand the depositional history of ice, past climatic variations, sublimation and homogenization processes occurring in the upper layers of glacier ice.
During the past 30 years, several studies have been carried out in polar and temperate regions, particularly Antarctic, Arctic and Alpine regions, using radioactive and stable isotopes (Reference Dansgaard, Johnsen, Moller and LangwayDansgaard and others, 1969; Reference JouzelJouzel and others, 1987). Studies from the Indian and Nepal Himalaya are relatively sparse (Reference YasunariYasunari 1976; Reference Grabczak, Niewodniczanski and RózanskiGrabczak and others, 1983; Reference Nijampurkar and BhandariNijampurkar and Bhandari, 1984; Reference Mayewski, Lyons, Spencer and ClaytonMayewski and others, 1986; Reference WakeWake, 1989; Reference Nijampurkar and RaoNijampurkar and Rao, 1993). Recent studies based on δ 18O in snow/ice and ice core from the Tibetan (Xizang) Himalaya have addressed the seasonal relationship between δ 18O in snow/ice and air temperature and moisture sources (Reference Aizen, Aizen, Melack and MartmaAizen and others, 1996; Reference Tandong, Thompson, Mosley-Thompson, Zhihong, Xinping and Ping-NanYao and others, 1996; Reference ThompsonThompson and others, 1997). Reference Tandong, Thompson, Mosley-Thompson, Zhihong, Xinping and Ping-NanYao and others (1996) demonstrated that the δ 18O–T relationship has a positive slope over the northern part of the Tibetan Plateau and suggest that the slope “inverts” over a narrow band between the Himalaya and the central plateau to the north.
The climate of the glaciated central Himalayan region is influenced by precipitation during the summer monsoon which develops over the Asian countries. Over India, the monsoonal rains begin by mid-June. The southwest monsoon also plays an important role during the ablation period of Himalayan glaciers. Precipitation in the form of snow occurs only during October to March, when the primary source of moisture is related to the winter monsoon and western disturbances. The accumulation is maximum when the snowline recedes to about 1500 m in the western Himalaya and to about 3000 m in the eastern Himalaya.
The main objectives of this study are: (i) to obtain information on the ice dynamics and snow-accumulation rates of Dokriani Bamak glacier, and (ii) to decipher the relationship, if any, between the annual mean δ 18O content in surface snow and elevation with respect to glacier orientation. The isotopic data on snow/ice and shallow ice core, thus collected, are evaluated in terms of their implications for short-term climatic changes and the depositional history of the glacier ice.
2. Location
Dokriani Bamak (DB) glacier, located at 31°49′ N, 78°47′ E and at an altitude of 3900–5000 m, is a well-developed valley glacier of the Gangotri group in the central Himalaya. It lies southwest of Gangotri glacier (Fig. 1). DB glacier is 6 km long and its width varies from 0.86 km near the snout to 1.5 km in the accumulation zone. Within its catchment area of 15.2 km2, 9.8 km2 is covered with ice and 4.1 km with permanent snow-fIelds. DB glacier is formed by two cirque glaciers: one on the northern slopes of Draupadi ka Danda (5614 m a.s.l.) and the other on the southwestern slopes of Janoli (6632 m a.s.l.) (Fig. 1). It then flows from the accumulation zone with a gradient of 12°. It terminates at an altitude of 3882 m and flows between the eroded lateral moraines formed during the past glacial periods. Lateral moraines are prominent glacial features of Din Gad Valley (Fig. 1). Three-quarters of the ablation zone of the glacier is overlain by a thick cover of supraglacial debris. Marginal and transverse crevasses are well developed in the ablation zone, particularly in the altitude range 4200–4300 m; longitudinal and splaying crevasses are developed on the northeast flank of the icefall. Avalanches occur very frequently in the accumulation zone of the glacier. The geomorphology of the glacier is discussed in detail elsewhere (Gergan and Dhobal, 1996). DB glacier is one of the few glaciers in the Bhagirathi basin that is easily accessible throughout the year. Located on the southern slopes of the Himalaya, it faces southward with respect to the Himalaya. The major source of moisture is the rainfall that occurs during the summer monsoon (June-September), passing over the glacier from west to east and giving an average precipitation of 1000–1200 mm below the equilibrium line. The second phase of precipitation occurs as snowfall from October to March (at times it extends to April), when western disturbances move eastward over northern India.
Recession of glaciers is observed worldwide, and Himalayan glaciers are no exception. DB glacier receded at an average rate of 17.5 m a−1 during the period 1990–95. Mass-balance studies carried out on the glacier during the sampling period 1992–95 show a negative trend. The equilibrium-line altitude calculated from mass-balance data shows an increase in height in successive years, which indicates that the accumulation-area ratio of DB glacier decreased from 0.68 to 0.62 during 1992–95 (Gergan and Dhobal, 1996). A rise in the atmospheric temperature is indicated by an increase in the height of the equilibrium line.
3. Sample Collection and Analytical Techniques
3.1. Sample collection
The snow and surface ice samples were collected at altitudes of 3800–4900 m along the central flowline of the glacier during 1992–94. The snow samples from four different pits (Fig. 1) were collected during May–June 1994 at varying altitudes and depths along the central axis of the glacier flow. A 6 m long ice core was raised during October–November 1993, at an altitude of 4863 m in the accumulation zone, and was subsampled at an interval of 20 cm. During the 1992 expedition, large quantities (∼1000 kg) of surface snout ice were collected and processed for determination of 32Si activity and dating of the snout ice. The chemical procedures adopted in the field as well as in the laboratory for the extraction, purification and estimation of 32Si and 137Cs are described in our earlier papers (Reference Nijampurkar, Bhandari, Vohra and KrishnanNijampurkar and others, 1982; Reference Nijampurkar and RaoNijampurkar and Rao, 1992); 210Pb was assayed by alpha spectrometry (Reference Sarin, Bhusan, Rengarajan and YadavSarin and others, 1992).
3.2. Analytical techniques
32Si analysis
About 1000 kg of meltwater was collected from snout ice in plastic drums and spiked with Na2SiO3 and FeCl3 solutions. After equilibration, silica was co-precipitated along with Fe(OH)3 at a pH of ∼8. In the laboratory, silica (along with 32Si) was separated from Fe(OH)3 and allowed to equilibrate with its daughter isotope, 32P (t 1/2 = 14.3 days). After 3–4 months, 32P was recovered as Mg2P2O7 and its activity was assayed on a low-background Geiger–Müller counter operated in anti-coincidence with a NaI(Tl) detector (Reference Nijampurkar, Bhandari, Vohra and KrishnanNijampurkar and others, 1982). The counting system had an efficiency of 27%. The in situ 32Si concentration in the snout ice was ascertained from the measured 32P activity.
137Cs analysis
Samples of snow, surface ice and those from the shallow ice core were filtered to remove suspended dust, if any, and acidified with 8M HNO3. The filtrate was evaporated to dryness, and the residue was counted on a high-resolution, low-background HPGe well detector for assessing 137Cs activity under the 661.5 keV photopeak.
δ 18O measurements
About 10 g samples of meltwaters from the snow/ice, pits and shallow ice-core samples were collected in airtight plastic bottles, transported to the laboratory and kept frozen until analysis. Oxygen isotope analysis was carried out using an auto-mass spectrometer at the Glaciology Laboratory of the Niels Bohr Institute, Copenhagen, Denmark.
4. Results and Discussion
The results of the isotope analysis on snow/ice, pits and shallow ice-core samples from DB glacier are given in Tables 1–3 and Figures 2–6, and are discussed in relation to glacier ice flow and accumulation rates and short-term climatic variations.
4.1. Radioisotopic studies for ice dynamics
Dating of snout ice of the glacier by 32Si and 210Pb
The specific activities of 32Si and 210Pb in the snout-ice samples collected during August 1992 are calculated to be 3.7 ± 0.5 μBq kg−1 and 13.2 ±1.3 mBq kg−1, respectively (Table 1). Using a two-component box model (Reference Nijampurkar, Bhandari, Vohra and KrishnanNijampurkar and others, 1982), the corrected value of 32Si activity in the snout ice is calculated to be 2.5 μBq kg−1. Using this value and assuming the average 32Si concentration of the snowfall in the Himalayan region to be 11.7 μBq kg−1, a radiometric age of 400 years is obtained for the snout ice of DB glacier. The basic assumption made in the 32Si dating method is that the production rate of 32Si, as well as its atmospheric fallout over the Himalayan region, has remained constant over the past few centuries. Based on a radiometric age of 400 years for the snout ice and a glacier length of 6 km, the average flow rate of ice over the last four centuries is estimated to be 14 m a−1. The 32Si fallout concentration in snow, 11.7 μBq kg−1, is an average value based on experimental data from two different glaciers, Nehnar in Kashmir, and Changme-Khangfu in Sikkim, which range between 5 and 16.7 μBq kg−1 (Reference Nijampurkar, Bhandari, Vohra and KrishnanNijampurkar and others, 1982). A 10% change in the fallout concentration would change the snout-ice age by ∼7%.
Accumulation rates and surface flow rates
The specific activities of 137Cs in snow and surface ice samples collected during 1993 are given in Table 2. The concentrations of Cs in snow samples at different altitudes are generally < 1.5 mBq kg−1. These activities represent the background level of 137Cs concentration, as there was no significant emission of artificial radioactivity during the period of sample collection. However, the 137Cs activity measured in the surface ice collected at 4380 m (∼500 m below the equilibrium line) is 22 ± 3.8 mBq kg−1. The enhanced activity of 137Cs is attributed to its peak fallout resulting from nuclear bomb tests during the early 1960s. Such an observation suggests that the snow deposited during 1963–64 in the accumulation zone (∼500 m above the equilibrium line) could have travelled all along the glacier flowline and emerged at 4380 m altitude after nearly 30 years, traversing a total distance of ∼1000 m. This linear movement allows us to estimate a glacier ice-flow rate of about 32 m a−1 near the equilibrium line, which is higher than the average flow rate of ice (14 m a−1 as determined from snout age) all along the glacier. It could be argued that 137Cs activity in the surface ice at 4380 m is not related to 1963 fallout and instead is associated with the Chernobyl accident in April 1986. However, the 137Cs activity arising from the former event is further supported by the measured 210Pb activity of 35 mBq kg−1 in the same sample. The constant fallout activity of 210Pb in the Himalayan glacier is ∼100 mBq kg−1 (Reference Nijampurkar, Bhandari, Vohra and KrishnanNijampurkar and others, 1982). This buried activity after a time-span of 30 years could correspond to ∼38 mBq kg−1, quite similar to that observed in the surface ice at 4380 m. Thus, the flow rate of ice as deduced by us is quite representative near the equilibrium line. In Table 3, the snout-ice ages and time-averaged flow rates of the surface ice of some temperate Himalayan glaciers are compared (Reference Nijampurkar and RaoNijampurkar and Rao, 1993). The snout ages vary from 160 to 840 years, whereas the ice-flow rates vary from 4 to 23 m a−1.
The depth profile of 137Cs specific activity in a 6 m ice core (raised at 4863 m in the accumulaton zone) shows a maximum value of 19.2 mBq kg−1 at ∼3 m depth (Fig. 2) compared to an average value of 3.3 ± 2 mBq kg−1 up to 2.5 m depth and of 5 ± 3 mBq kg−1 at 4–6.25 m depth. The maximum activity is higher by a factor of 4 than the average value in snow samples above and below 3 m depth. This episode of enhanced 137Cs activity in the ice core is attributed to Chernobyl fallout, similar to that observed on Chhota Shigri glacier (Reference Nijampurkar and RaoNijampurkar and Rao, 1990). Using this activity as a time marker, the accumulation rate of ice at 4863 m in the accumulation zone of DB glacier is estimated to be 0.43 m a−1, which is comparable with that observed on neighbouring glaciers in the central Himalaya.
4.2. δ 18O concentrations in snow, pits, surface ice and a shallow ice core
δ 18O concentrations in snow and ice
The results of the analysis of the fresh snow and surface ice samples collected at different altitudes (in both the accumulation and ablation zones of the glacier) are shown in Figure 3. The δ 18O in snow samples varied from −5‰ to −9‰; however, there is no systematic variation of δ 18O with increasing altitude, as these samples represent only a single snowfall event and not the annual precipitation. These oxygen isotope ratios are similar to those reported in our earlier studies on Himalayan glaciers (Reference Nijampurkar and BhandariNijampurkar and Bhandari, 1984; Reference Nijampurkar and RaoNijampurkar and Rao, 1992). The δ 18O values of surface ice samples at different altitudes range from −11‰ to −13.4‰. The δ 18O value of snout ice is −13.4‰, the most depleted oxygen isotope ratio on the entire glacier. Comparison with the observed mean value of −9.2‰ in a 6 m ice core from 4863 m elevation suggests that the depleted value of δ 18O in snout ice (−13.4‰) could be associated with the cooler climatic conditions prevailing a few centuries ago during the Little Ice Age (LIA).
δ 18O concentrations in snow pits
The δ 18O composition of snow samples, collected at varying depths from four pits at altitudes of 3826–4695 m, shows a large variation from −4‰ to −23‰ (Fig. 4). For pit P1 (3826 m), δ 18O values range from −4.5‰ to −7.2‰, whereas for samples collected from pit P5 (4695 m) they range from −5‰ to −23‰ (Fig. 4) for the same depth range of 0–1.6 m. A decreasing trend in δ 18O concentrations with increasing elevation is thus observed in the snow pits (Fig. 4), with the most depleted values occurring in pit P5 from the highest elevation (4695 m). Snow samples within the same pit also show a seasonal variation in δ 18O with depth (Fig. 4), suggesting that the earliest snow deposited during October–November is more depleted in δ 18O than the snowfall during later months (February–April). The larger degree of isotopic enrichment in subsequent precipitation at successively higher elevations is attributed to the orientation of the glacier (south with respect to Himalaya, windward). These results clearly demonstrate the varying sources of moisture/precipitation during the winter monsoon and those related to western disturbances moving over the central Himalaya. The “altitude effect” for the annual mean δ 18O in snow is 0. 9‰ per 100 m (Fig. 5) for this glacier.
The radiometric age of 400 years for snout ice, associated with a δ 18O value of −13.4‰ and mean δ 18O (−9.2‰) in a 6 m ice core, yields a difference of 4.2‰, suggesting that air temperatures in this region during the LIA were cooler by at least a few °C. The empirical relationship between δ 18O in annual precipitation and the mean surface air temperature (at the sample location sites) for the polar regions has been established by earlier workers (Reference DansgaardDansgaard, 1964). However, in this study it has not been possible to establish a δ 18O–T relation, due to lack of data on mean surface air temperatures.
δ 18O concentrations in shallow ice core
The δ 18O measurements on samples from a 6 m ice core, from the accumulation zone of the glacier, show a highly depleted value of −15.2‰ at 0.25 m depth (Fig. 6). A likely explanation for this extreme negative value is isotopic variance and enrichment in the δ 18O values at greater depth that may have been modified via repeated freezing and thawing; and that there has not been enough time for these processes to smooth the isotopic signal in the near-surface ice. In addition, the highly depleted δ 18O value could be associated with the episodic stormy conditions and the dominance of one of the varying sources of moisture.
5. Conclusions
The main conclusions that can be drawn from our study are:
-
1. Based on the 137Cs activity in a surface ice sample collected at 4380 m altitude, the flow rate (ice transfer) of surface ice is estimated to be 32 m a−1 near the equilibrium line. This is in contrast to the time-averaged flow rate of 14 m a−1, along the entire glacier length, during the past four centuries.
-
2. The peak concentrations of 137Cs (19.2 mBq kg−1) observed at 3 m depth in an ice core confirm the presence of Chernobyl fallout activity, based on which the accumulation rate of snow at ∼4900 m altitude is calculated to be 0.43 m a−1.
-
3. An “altitudinal effect” of 0.9‰ in δ 18O per 100 m elevation on DB glacier is observable in the annual snowfall.
-
4. A significantly depleted δ 18O value (−13.4‰) in the snout ice compared to the mean δ 18O (−9.2‰) value in a shallow ice core suggests that climatic conditions during the LIA period were a few °C cooler than the present-day environment.
Acknowledgements
This study was funded by the Department of Science and Technology, New Delhi. We thank S. S. Das and K. Soni and members of the ice-core drilling team for their help in the field, H. B. Clausen, of the Geophysical Institute of Copenhagen for providing δ 18O measurements in the snow/ice and ice-core samples, and S. Krishnaswami for help and encouragement during the course of this study. We also benefited from the comments of two anonymous referees and of the Scientific Editor, D. A. Peel.