Hydrological characteristics of the Dokriani Glacier in the Garhwal Himalayas

Transcription

1 Hydrological Sciences -Journal- des Sciences Hydrologiques,<l0,2, April Hydrological characteristics of the Dokriani Glacier in the Garhwal Himalayas PRATAP SINGH, K. S. RAMASASTRI & U. K. SINGH National Institute of Hydrology, Roorkee , India J. T. GERGAN & D. P. DOBHAL Wadia Institute of Himalayan Geology, Dehradun , India Abstract Observations of discharge, temperature and suspended sediment made at a gauging site established near the snout of the Dokriani glacier in the western Himalayan region are presented. These observations were made during a scientific expedition to this glacier over 21 days ( ). Because of harsh weather conditions, observations could not be made for a longer period. The minimum streamflow in the glacier melt stream was observed at 0700 h whereas the maximum was observed at 1800 h. The ratio of maximum to minimum flow was computed to be 1.81 from the continuous hourly observations. Based on an analysis of the recession of the hydrograph, it was found that the meltwater time lag from the accumulation zone of the glacier was more than seven times higher than that from the ablation zone. No specific relationship was observed between suspended sediment and discharge. The average values of the suspended sediment concentration and load were found to be 350 ppm and 180 t day" 1, respectively, for the study period. Weathering processes in different zones of the glacier were also studied to find out the source of the sediment transported by the meltwater into the melt stream. A high correlation coefficient (r = 0.89) was found between the glacier specific runoff and the air temperature at the gauging site. It showed that temperature alone can represent the melting of the glacier and may be considered for the hydrological modelling of glacier melt runoff. Based on observations over three days under clear weather conditions for isolated snow blocks, the average snowmelt factor was computed to be 5.4 mm /( C. 6 h). Caractéristiques hydrologiques du glacier Dokriani, Garhwal Himalaya Résumé Les observations de débit, de température et de matières en suspension réalisées à la station de jaugeage établie près de la langue du glacier Dokriani dans l'ouest de la région himalayenne sont présentées ici. Ces observations ont été réalisées dans le cadre d'une expédition scientifique d'une durée de 21 jours ( ). En raison de rudes conditions météorologiques, il n'a pas été possible de poursuivre les observations plus longtemps. Le débit minimum de fonte a été observé à 0700 heures, cependant que le maximum de ce débit a été observé à 1800 heures. Le rapport du maximum de débit au minimum de débit a été estimé à 1.81 à partir d'observations horaires. L'analyse des Open for discussion until I October 1995

2 P. Singh et al. hydrogrammes de décrue a permis de montrer que le temps de réponse de la zone d'accumulation était supérieur à sept fois le temps de réponse de la zone d'ablation. Aucune relation entre le débit et les matières en suspension n'a été observée. Durant la période étudiée, les valeurs moyennes de la concentration des matières en suspension et du transport solide ont été respectivement égales à 350 ppm et à 180 tonnes par jour. Les processus d'effritement ont été étudiés en différentes zones du glacier afin de déterminer l'origine des sédiments entrainés par les eaux de fonte vers le torrent. Il existe un coefficient de corrélation élevé (r = 0.89) entre l'écoulement spécifique du glacier et la température de l'air à la station de jaugeage. Cela montre que la température peut à elle seule expliquer la fonte du glacier et peut être seule retenue pour la modélisation hydrologique de l'écoulement de fonte. Sur la base de trois journées d'observation par temps clair et pour des blocs de neige isolés, le facteur de fonte de la neige a été estimé à 5.4 mm/( C. 6 h). INTRODUCTION High mountain areas featuring glaciers and snow are widely distributed throughout the world. Himalaya is one of the regions where numerous glaciers exist. The water flowing in the Himalayan rivers is a combined drainage from rainfall, snowmelt and glacier melt runoff. Snow and glacier runoff play a vital role in making all the northern rivers of India perennial. The glacier contribution to the rivers originating from the Himalayas starts in July/August when seasonal snow cover melts and continues till October/November depending upon the prevailing climatic conditions in that region. In the regions where monsoon rain penetrates into the high altitude valleys containing glaciers and coincides with glacier melt runoff, the flow in the rivers is augmented suddenly and sometimes may cause havoc downstream. During winter these rivers consist of contributions from sub-surface flow and winter rains in the lower parts of the basins. Glacier melt runoff is negligible in the winter season because of very low temperatures over the glaciers. Assessments of the melting rates and determinations of the drainage characteristics of glaciers in the high mountains have importance in harnessing the Himalayan river systems for hydroelectric power generation and irrigation, and lead to other benefits such as reservoir operation and the design of hydraulic structures. Himalayan glaciers provide a significant contribution to the flow in Himalayan rivers, but very limited studies have been carried out to understand the melt processes and other hydrological characteristics of the glaciers in the Himalayas. One of the basic reasons for such limitation is the difficulty in collecting data at high altitude for a long period because of inaccessibility and harsh weather conditions. Vohra (1980) has highlighted the problems associated with Himalayan glaciers. In the present study, results based on the analysis of data collected for a short period (August/September 1992) during an expedition to the Dokriani Glacier in the western Himalayas are presented. Efforts have been made to understand the hydrological behaviour of this glacier.

3 Hydrological characteristics of the Dokriani Glacier 245 SALIENT FEATURES OF THE DOKEIANI GLACIER The Dokriani Glacier is a valley type of glacier located in the western Himalayas. This section of the Himalayas is known as the Garhwal Himalayas (Fig. 1). It lies at about latitude 31 N and longitude 79 E and originates in the vicinity of the Janoli (6633 m) and Draupadi ka Danda (5716 m) peaks. The glacier elevation ranges from 3950 to 6400 m. The maximum length of the glacier is about 5.5 km and its width varies from 0.1 to 2.0 km from the snout to the accumulation zone. The glacier snout is situated at an altitude of 3950 m and is covered by huge boulders and debris. The lower portion of the glacier is almost totally covered by debris. The material of these moraines is derived from the valley sides through various processes such as debris slides, avalanches and weathering processes. The middle part of the glacier is highly fractured and consists of many glacial features such as crevasses, moulins, glacier tables and ground moraines. The crevasses are mainly of the transverse type and are wide and long. Sometimes longitudinal crevasses are also seen along the sides of the glacier. The glacier is bounded by two large lateral moraines about m in height. Besides these lateral moraines, there are four other lateral moraines observed at different heights. These different levels of the moraines indicate past extensions of the glacier. The melt stream originating from the Dokriani Glacier is known as Din Gad which follows a narrow valley and meets the Bhagirathi River near Bhukki. 32' HIMACHAL Fig. 1 Location of the Dokriani Glacier in the Garhwal Himalayas.

4 246 P. Singh et al. Geomorphological features The geomorphological characteristics of a glacierized catchment influence the melting of the glacier and the drainage pattern of the glacier melt runoff. The primary characteristics of the catchment are its area, length, shape, elevation, slope, orientation, etc. The main geomorphological parameters of the Dokriani Glacier are given in Table 1 and the area-elevation curve of the glacier is illustrated in Fig n : UJ 4500 i / Cumulative area (sq. km) Fig. 2 Area-elevation curve of the Dokriani Glacier Table 1 Geomorphological parameters of the Dokriani Glacier Serial no. Parameter Value Glacier area (km') Drainage area (km 2 ) Glacier perimeter (km) Glacier length (km) Eccentricity of glacier Circularity ratio of glacier Elongation ratio of glacier Glacier relief (km) Average gradient of glacier (%)

5 Hydrological characteristics of the Dokriani Glacier 247 FIELD OBSERVATIONS Glacier melt runoff observations A survey of the melt stream starting from the snout to about 2 km downstream was made and finally a suitable gauging site was selected and established for streamflow measurements. This gauging site was about 800 m downstream from the snout of the glacier and about 1 km upstream of the confluence of several small nallahs joining the main stream. The flow was not greatly turbulent at this site and the total meltwater was confined to a single channel. A bridge over the melt stream was made with the help of wooden logs and using boulders as a bridge support at both ends. A view of the gauging site is shown in Fig. 3. Fig. 3 Gauging site on the Dokriani melt stream. The velocity-area method was used to estimate the discharge in the melt stream. Wooden floats were used to measure the velocity of flow. The time of travel of the floats was determined with a stopwatch. The straight reach at the gauging site was only 6 m long, forming the best possible straight reach which could be found along the channel. In fact, to get a long straight reach of flow is a great problem in this type of terrain. For accuracy, the velocity readings were repeated at least three times and an average value adopted for further computations. It was assumed that there was no change in the cross-sectional area of the gauging site during the study period. Atmospheric and water temperatures were also observed at the gauging site when flows were measured. The observations of discharge, suspended sediment and air temperature were made for a period of 21 days ( ). For this period hourly observations were made for all the parameters only during daytime and

6 248 P. Singh et al. mean values of the parameters were computed. To obtain a complete hydrograph at the gauging site, hourly observations of streamflow and air temperature were made for 24 h on / Runoff from isolated snow Mocks In order to determine the snowmelt factor, snow blocks (60 cm x 60 cm x 30 cm) were extracted from the glacier snowpack at 4000 m altitude and studied for their melting rate. The observations were made from 0800 h to 1400 h on three days (21, 22 and ). The air temperature was observed each 15 min at 2.0 m above the surface of the snow block and used to compute the average temperature for each hour. The meltwater was collected in a bucket and measured every hour just after collection to minimize evaporation losses. The amount of meltwater lost via percolation was checked by putting a white plastic sheet under the snow block. This ensured that the total meltwater was collected in the bucket. The density of the snow blocks was measured to be 0.60 g cm" 3. The observations for the isolated snow blocks were restricted to clear weather conditions with a view to estimating the melt factor under fair weather conditions. Observations after 1400 h were not taken because after this time clouds formed and clear weather conditions did not persist. This phenomenon was observed regularly during the whole study period. One of the reasons for the formation of clouds in the afternoon is thought to be the availability of soil moisture for evaporation and évapotranspiration. As soon as sufficient solar radiation is available, évapotranspiration processes occur strongly and clouds are formed in the afternoon. MELT FACTOR The melt factor is also known as the degree-day factor when the time unit is a day. It is an important parameter and is used to convert the degree-days to snowmelt or ice melt expressed in depth of water. It is given in the following form: M = D f (T a -TJ (1) where M = depth of meltwater (mm per unit time) ; T a mean air temperature ( C); T 0 = base temperature (0 C); and Df = melt factor (mm/( C. unit time)). Information on the melt factor is required for glacier melt modelling studies. However, the melt factor is variable throughout the melt period because the changes in snow properties influence the melting process. It is possible to compute the melt factor at a point by measuring temperature and meltwater from the snowpack. The point measurements can be used for infor-

7 Hydrological characteristics of the Dokriani Glacier 249 mation on how well a specific station represents the hydrological characteristics of the given zone. The observed values of average temperature and meltwater along with the computed melt factor for 6 h periods are given in Table 2. Details for the hourly basis are given in Table 3. Values for D^have been given by several investigators. Anderson (1973) reported a seasonal variation in the melt factor and gave a range of mm/( C. 6 h) using the mean air temperature on the same pattern as Male & Granger (1978). From this study, an average value of D^-was computed to be 5.4 mm/( C. 6 h). It is expected that values of D f derived from the snow block measurements may be slightly on the high side because of energy gained at the snow block walls. Very likely, melt rates at an undisturbed site would be lower. There are maximum contributions from glaciers to the Indian river systems in the months of July and August. Therefore, this value of the melt factor may be considered as the maximum melt factor for the glacier snow. Table 2 Mean temperature and snowmelt factors for different dates; melt rates were observed between 0800 and 1400 h Date ( h) Mean i air temp. ( C) Depth (mm) of meltwater D, (mni/cc. 6 h) Table 3 Hourly temperature, snowmelt water and snowmelt factors Date Hours Average T a Runoff from snow Dj CO blocks (mm/cc. h)) (mm ir')

8 250 P. Singh et al. RELATIONSHIP BETWEEN GLACIER SPECIFIC RUNOFF AND TEMPERATURE The diurnal variations observed in the glacier melt runoff on / are illustrated in Fig. 4. It can be inferred that the maximum flow is observed at about 0700 h whereas the minimum flow is observed at about 1800 h. The ratio of maximum to minimum streamflow derived from the hourly discharge observations is 1.81 and shows the typical storage and release characteristics of the glacier S" "5» 3.0 r^i rrp xnjtti Ô"""25'" Time (Hours") Fig. 4 Hydrograph observed on / at the gauging site 800 m downstream of the Dokriani Glacier snout. To establish a relationship between glacier specific runoff and air temperature observed at the gauging site, specific runoff has been computed using streamflow data and the total area of glacier. Specific runoff is defined as the melt derived per unit area of glacier for a particular day. The relationship between daily specific melt and air temperature at the gauging site is shown in Fig. 5. The following regression equation was obtained for this graph: 0.497^-0.24 (2) where G m is the specific runoff from the glacier (cm day" 1 ) and T is the mean air temperature at the gauging site ( C). A good correlation coefficient (r = 0.89) was obtained between these two variables. It shows that temperature alone can be used to estimate the glacier melt runoff. Temperature data at the glacier surface were not available and so this type of relationship could not be attempted using glacier air temperature. However, it would be possible to relate the temperature observed at the snout to that at the mid-elevation of the glacier to get an average

9 Hydrological characteristics of the Dokriani Glacier 251 8t, >s7 D 3 ~ 3 o5 3 1 r=0.s9 a) g JO o Temperature ( C) Fig. 5 Correlation between specific glacier runoff and mean air temperature at the gauging site from / Table 4 Some salient features of the observations made from to at the gauging site on the glacier melt stream Feature Maximum flow velocity (m s ) Minimum flow velocity (m s" 1 ) Maximum mean daily discharge (m 3 s"') Minimum mean daily discharge (m 3 s" 1 ) Maximum suspended sediment transportation load (ppm) Minimum suspended sediment transportation load (ppm) Maximum mean daily air temperature at gauging site ( C) Minimum mean daily air temperature at gauging site ( C) Maximum melt stream water temperature ( C) Minimum melt stream water temperature ( C) Value temperature over the glacier using a standard temperature lapse rate (6.5 C 1000 m" 1 ). In the present study this was not done because significant variations are found between the actual lapse rate and the standard temperature lapse rate in the various sections of the Himalayas (Bhutiyani, 1989; Singh, 1991) and therefore, development of this relationship was restricted to air temperature data observed at the gauging site. The ranges of observed flow and temperature are given in Table 4.

10 252 P. Singh et al. RECESSION COEFFICIENTS AND TIME LAG To investigate the diurnal variations in the flow of the melt stream, hourly observations were also made. The hydrograph observed on / is shown in Fig. 4. The recession coefficients of those curves have been determined using the following methodology. If a basin is considered as a linear reservoir, the discharge Q(t) from the reservoir at time t will be a linear function of the volume of water stored in the reservoir, S(t), and can be represented by the following equation: Q(t) = ks(f) ( 3 ) where k is a recession coefficient for the drainage from the glacier reservoir. The continuity equation for the reservoir can be written as: or: I-Q = (4) at I = Q + (5) at where / is the input in the form of melt or liquid precipitation or both. From equations (3) and (5) one obtains: / = Q + 1^2. (6) k at During a recession 1=0, and therefore integration yields: Q(t) = Q 0 e- kt 0) where Q 0 is the discharge at time t 0. The recession coefficient k (h" 1 ) can be determined from a semi-logarithmic plot of discharge Q(t) against time t for which equation (7) describes a straight line. The slope of this line is the value of the recession coefficient, k. A semi-logarithmic plot of the recession hydrograph is presented in Fig. 6. The recession hydrograph can be separated into two arcs with different slopes. This indicates that water is derived from two reservoirs which can be considered as, relatively, fast and slow reservoirs. The fast reservoir represents the ablation area (snow free) whereas the slow one represents the accumulation area of the glacier. Consequently, the slope of the first arc, k a, represents the recession from the ablation area and the second arc, k c that of the accumulation area. The values of k a and k c were found to be h" 1 and If 1 for The value of kjk c is 7.6. It should be pointed out that contributions into a glacier melt stream can arise from several reservoirs such as accumulation, ablation, shallow groundwater and deep groundwater reservoirs. However, once the diurnal recessions of glacier melt runoff are studied, the recession trends of the water from the accumulation and ablation areas play a significant role in the rapid recession of the meltwater.

11 Hydrological characteristics of the Dokriani Glacier 253 \ f \ k = > k c=-019 T-j-rn Time (Hours) Fig. 6 Recession curves of the glacier melt runoff of the Dokriani Glacier observed on / at the gauging site. Assuming that the contribution from the groundwater reservoir is relatively slow compared with these two reservoirs, and that this contribution does not vary over a short duration (one day in the present case), the glacier has been considered in the form of only two reservoirs, namely the accumulation reservoir and the ablation reservoir in the present analysis. There is a time lag between the melting at the glacier surface and the meltwater arriving at the outlet of the glacier. In fact, the basin lag is different for the ablation area and the accumulation area. It may be expected that runoff from the ablation area is faster then that from the accumulation area. The impermeable ice surface and very low roughness of the ice channels in the ablation area contribute to fast runoff of meltwater, whereas in the accumulation area meltwater percolates first into the snowpack and reaches the outlet of the glacier after passing through several englacial and subglacial channels. Assuming recession coefficients as representative of time lag, the ratio of k a and k c can be considered as a ratio of time lags from the ablation and accumulation areas. That ratio shows that meltwater from the accumulation area takes more than seven times as long as that from the ablation area. SUSPENDED SEDIMENT Two sediment samples each day were collected at the times when flow was measured. A known volume of water (500 ml) was scooped from the stream and filtered using Whatman-40 ashless filter paper directly at the site. The filtered samples were promptly packed in polyethylene bags marked with details such as date and time of the sample collection, etc. and analysed later in the NIH Water Quality Laboratory at Roorkee for their dry weights. Because the

12 254 P. Singh et al. 1200n Q. CL 0J c Q. Si 200 J Discharge (m 3 /s) 10 Fig. 7 Observed discharge and suspended sediment at the melt stream gauging site between and amounts of sediment were very small, an accurate balance was used. The suspended sediment reduced to ppm along with flow in the channel is illustrated in Fig. 7. A wide range of suspended sediment concentration was observed in the melt stream although streamflow did not change that significantly. The following explanation might be given for such wide variations in the sediment concentration: (a) occurrence of rain; (b) land slides and/or rock slides over the glacier or along the channel; and (c) emptying of pockets of water, ice dammed lakes, etc. Land slides and rock slides are very common phenomena in such areas. The sediment deposited due to such incidents in the glacier basin is transported to the channel by the rain. For example, the highest value of suspended sediment in the channel is because of rain on that particular day. In some cases the sediment is immediately dumped into the channel due to such incidents and causes a sudden increase in the sediment concentration. Consequently it is difficult to establish a relationship between discharge and sediment. However, it can be broadly concluded that the suspended sediment concentration increases as the flow increases. A similar pattern between flow and suspended sediment has been observed in the melt streams of other glaciers (Singh, 1991). The average value of the suspended sediment concentration was about 350 ppm, whereas the average transportation of the suspended sediment load was estimated to be 180 t day" 1 over the study period. The observed range of the suspended sediment is given in Table 4. Sediments in the melt stream are derived from different parts of the glacier such as the accumulation and ablation zones and the lateral moraines. The major source of debris deposited over these zones is from the valley walls

13 Hydrological characteristics of the Dokriani Glacier 255 and from the reworking of the old moraines. Sediment transported by the glacier follows three basic routes: supraglacial, englacial and subglacial. To understand the weathering processes in different parts of the glacier, samples were collected. Supraglacial samples were collected from the accumulation (4800 m) and ablation (4000 m) zones. Englacial and subglacial samples could not be collected. Samples from the proglacial zone near to the snout were collected and represent a mixture of supraglacial, englacial, subglacial and lateral moraine sediments. Grain size analysis was carried out to separate fine and very fine factions ( fini) from the sample. Quartz grains were identified and picked up from the fractions using an optical microscope. Scanning electron microscopy (SEM) was carried out to study the surface texture of the grains. It was found that the quartz grains from the supraglacial samples were angular in shape whereas those from the proglacial samples were largely subangular. The grains from the lateral moraines were found to vary from angular to subangular. The surface texture of the grains from the supraglacial debris showed radiating fractures (Fig. 8(a)), which is generally caused by mechanical weathering of the source rock (Mahaney et al., 1991). Grains from the proglacial sample had arc shaped steps possibly due to a frost action shattering process and transportation by thick ice (Fig. 8(b)). This could also be considered as a mechanical weathering process. Grains from the lateral moraines samples indicated that they first underwent chemical weathering followed by mechanical weathering (Fig. 8(c)). The SEM study of the various samples from different parts of the glacier demonstrated that mechanical weathering was a dominant process in supra- and englacial debris, whereas chemical weathering was also taking place in the lateral moraines followed by mechanical weathering. From the features of the surface texture of the lateral moraines, it could be expected that chemical and mechanical processes might also produce the lateral moraines sediment into the glacier melt streams. However, it might not be solely responsible for the production of moraine and suspended sediment. Processes like sliding of rocks and debris, snow and glacier avalanches would dominate to produce moraines and sediment. Moreover, sediment derived from the glacier and valley walls is the product of mechanical and chemical weathering processes. This interpretation is based on rather a small number of samples and more data are required to support this interpretation. CONCLUSIONS The results of the present study can be summarized as follows: 1. A relationship between specific glacier melt runoff and air temperature at the gauging site near the snout of the Dokriani Glacier has been established and a high correlation (r = 0.89) found between these two parameters. It indicates that temperature alone can represent melting of glaciers.

14 256 P. Singh et al. (a) * :ÏSW- (W (c) Fig. 8 Surface texture features of quartz grains from (a) supraglacial; (b) proglacial; and (c) lateral moraines.

15 Hydrological characteristics of the Dokriani Glacier 257 Minimum streamflow was observed in the channel at about 0700 h while the maximum flow was recorded at 1800 h under clear weather conditions. No specific relationship was found between discharge and suspended sediment. However, a rainfall event was found to be responsible for the highest value of suspended sediment in the channel. During the study period, the average value of suspended sediment concentration and load were computed to be 350 ppm and 180 t day"', respectively. It was found that sediment arriving in the melt stream was caused by several processes such as debris and rock slides, snow and ice avalanches, and mechanical and chemical weathering processes. However, the former two processes are considered the dominant ones. The snow melt factor derived from the monitoring of isolated snow blocks for a period of six hours in the field on three different days under clear weather conditions was 5.4 mm/( C. 6 h). However, that value may be slightly on the high side because of the extra energy input through the walls of the snow block walls in comparison with the undisturbed snowpack. The density of the snow blocks was 0.6 g cnt 3. Acknowledgements The authors thank the Director, National Institute of Hydrology, Roorkee, and the Director, Wadia Institute of Himalayan Geology, Dehradun, for permitting participation in the glacier expedition. Financial support provided by the Department of Science and Technology, Govt, of India, New Delhi to the expedition under which this study was carried out is gratefully acknowledged. REFERENCES Anderson. E. A. (1973) National Weather Service River Forecast System - Snow Accumulation and Ablation Model. NOAA Tech. Memo NWS HYDRO-17, 219. US Dept. of Commerce. Bhutiyani, M. R. (1989) Modelling of glacier melt during ablation period from energy balance studies on Chhota Shigri Glacier, (HP), India. In: Second National Symp. on Hydrology (4-5 January 1989, Jammu (J&K), India), A116-A126. Mahaney, W. C, Vaikmae R. & Vares, K. (1991) Scanning electron microscopic study of quartz grains in supraglacial debris, Adishy glacier, Caucasus Mountains, USSR. Boreas, 20, Male, 0. H. & Granger, R. J. (1978) Energy mass fluxes at the snow surface in a prairie environment. In: Proc. Meeting on Modelling of Snow Cover Runoff (26-28 September 1978, Hanover, New Hampshire, USA), Singh, P. (1991) A temperature lapse rate study in western Himalayas, Hydrol. J. Indian Assoc. Hydrologists, 14, Singh. P. (1991) Hydrological investigations on Chhota Shigri Glacier (HP), National Institute of Hydrology, Roorkee, India, TR-106. Vohra, C. P. (1980) Some problems of glacier inventory in the Himalayas. In: World Glacier Inventory (Proc. Riederalp Workshop, September 1978), IAHS Pubf. no Received 29 March 1994; accepted 15 November 1994

16