In the present study, an attempt has been made to estimate the snow/glacier melt contribution in the head water region of the Beas Basin using a conventional hydrograph approach and a modeling (SNOWMOD) technique. The discharge and other meteorological data from 1996 to 2008 of the Manali site were used for the study. The results of SNOWMOD modeling reveal that snow/glacier melt contribution to the Beas River in the head water region varied between 52 (minimum) and 56% (maximum) with an annual average of 54% during the study period. The results obtained using the conventional approach showed the contribution of snow/glacier melt varied between 48 (minimum) and 52% (maximum) with an annual average of 50%. Results obtained using both techniques corroborate each other. This study reveals that the Beas River is mainly sustained by the snow/glacier melt contribution in the head water region.

INTRODUCTION

Himalayan snow cover has an enormous amount of water embedded in the form of snowpack/glacier (Siva Sankar 2002) and this is also a perennial source of major river systems like the Indus, Ganga and Brahmaputra, all originating from the high glaciers of the Himalayan region. These rivers receive substantial amounts of snow/glacier melt water and are considered as the lifeline of the Indian sub-continent. Snowmelt during summer forms an important component of stream runoff (Kulkarni et al. 2010). Despite their well recognized importance and potential, not many studies have been carried out to assess the detailed contribution of snow/glaciers in these rivers. However, some such studies have been made for a few river basins in the western Himalayan region on an annual basis (Singh et al. 1995; Singh & Kumar 1997; Singh & Jain 2003; Kumar et al. 2005; Prasad & Roy 2005; Haritashya et al. 2006). Singh et al. (1997) estimated the average annual contribution of snow/glaciers to be about 28% at Devprayag (830 m above mean sea level (amsl)) and 49% at Akhnoor (301 m amsl) in the Ganga and Chenab Rivers. Similarly, Singh & Jain (2002) estimated 60% snowmelt contribution at the Bhakra Dam site (for the Satluj River basin). Shashi Kumar et al. (1993) applied snowmelt runoff model for some parts of the study area, viz. the Beas Basin up to Thalot and the Parwati River up to the Phulga dam site. The snowmelt model (SNOWMOD) has been applied to Himalayan basins such as Satluj, Beas and Chenab (Jain et al. (1997); Jain & Singh 2003; Singh & Bengtsson, 2004; Arora et al. 2008; Jain et al. 2010). In addition, Ahluwalia et al. (2013) estimated 50% snow/glacier melt contribution in the Beas River at Manali. However, snow/glacier melt contribution for a small scale catchment and its seasonal variation are still not well studied. The present study is an attempt to estimate the contribution of snow/glacier melt, rainfall and groundwater subsurface flow, using a conventional approach and the SNOWMOD model for the Beas River basin at Manali (1,980 m amsl). This region represents the head watershed of the Beas River basin.

THE STUDY AREA

The upstream of the Beas River basin in the western Himalayas of Himachal Pradesh was selected for snowmelt runoff modeling (Figure 1). Beas Kund (3,505 m amsl) and the Rohtang pass (3,977 m amsl) are the major contributors of the water in the Beas River. The water originating from both places meets at Palchan, which is situated 10 km north of Manali (Figure 1). Beas River is one of the major tributaries of the Indus River system. The Beas Basin covers an area of 362 km2 up to Manali town, with an elevation range of 1,913–5,821 m. The upper reaches of the Beas River basin fall within Kullu district, with an estimated population of 437,903 (Census 2011).
Figure 1

Location map of Beas Basin up to Manali.

Figure 1

Location map of Beas Basin up to Manali.

The study area experiences a severe winter (December–March) characterized by the occurrence of severe snowfall at higher altitudes. The winter season is followed by the pre-monsoon season (April–June), having an average maximum temperature of 14.1 °C to a minimum 0.22 °C, with average rainfall during this period of 106.12 mm. In the summer season, temperature varies from a maximum 24.6 °C to a minimum 8.9 °C, and rainfall during this season averages 86.83 mm in Manali. The majority of the precipitation, over 70% of the annual rainfall, is concentrated in the monsoon months (June–September) (Das 2013). The upper reaches of the basin are snow covered and glaciated, which contributes melt water to the streamflow. The post-monsoon season during October–November is generally the driest, with scanty amounts of rainfall.

DATA AND METHODOLOGY

Physiographical and hydro-meteorological data are required for computing the streamflow from the basin. Physiographical data represent physical features of the basin, including its total area, its altitudinal distribution through elevation zones, the areas of these zones, and the altitude of precipitation and the temperature station. Hydro-meteorological data including daily precipitation, mean air temperature, the snow-covered area and streamflow data are needed at the beginning of simulation. To simulate the streamflow, the daily rainfall data and temperature for Manali station were used. Daily mean temperatures were computed by taking the average of available daily maximum and minimum temperatures. In the case of a snow-covered area, snow accumulation starts from October, therefore the period from October to September was considered as the hydrological year (Clow 2010).

Hydro-meteorological data for the Beas River basin in the western Himalayas were acquired from 1996 to 2008 from the Bhakra Beas Management Board, Pandoh, Himachal Pradesh. Daily discharge, temperature and precipitation data at Manali station (32 ° 09′ 35″ N and 77 ° 05′ 59″E at 1,980 m amsl) in Kullu district were used for the hydrological modeling (Figure 1).

In the present study, a digital elevation model generated using an advanced spaceborne thermal emission and reflection radiometer sensor (ASTER) has been used for obtaining the elevation zone for snowmelt runoff, due to the non-availability of topographic maps. ASTER is an imaging instrument that is flying on the TERRA satellite launched in December 1999 as part of NASA's earth observing system. ASTER resolution ranges from 15 to 90 m, depending on the wavelength. The instrument records the information in three bands: the visible and near infrared, the shortwave infrared and the thermal infrared, oriented on the nadir and backward looking. The total basin area has been sliced in to 7 zones of equal intervals of 500 m with ranges from 1,800 to greater than 5,400 m. The maximum area (91.13 Km2) is lying between the zones 3,600 and 4,200 m, whereas the minimum area, 1.12 km2, is in the zone greater than 5,400 m.

In this study, moderate resolution imaging spectroradiometer (MODIS) snow cover products have been used for snow cover analysis. This delivers public domain data in raster format. The MODIS snow cover product is freely available through the Distributed Active Archive Center located at the National Snow and Ice Data Center. The MOD10A2 (MODIS/TERRA SNOW COVER 8 days composite L3 GLOBAL 500 m SIN GRID V005) snow product with 500 m spatial was obtained for the study basin for the years 2004–2009.

Hydrograph separation

The process of separating the time distribution of different components (snow/glacier melt, baseflow and rainfall runoff) from the total runoff hydrograph is referred to as hydrograph separation. In the present study, a conventional (graphical method) and a modeling approach have been applied to estimate the contribution of snow/glacier melt runoff to the Beas River as the rainfall amount is negligible during the study period.

Graphical method

The graphical method prevailed from 1930 to 1960. Different methods have been proposed for the computation of baseflow and other components to streamflow. These conventional methods are constant discharge baseflow, constant slope baseflow, concave baseflow and master depletion curve methods. The selection of the appropriate conventional methods for watershed analysis depends highly on the type and amount of available measured data, the required accuracy for the design problem, and computational effort. In this study, the constant discharge baseflow method (Figure 2) has been used to separate out the baseflow and snowmelt/glacier contribution from the negligible amount of rainfall.
Figure 2

Constant slope method for hydrograph separation.

Figure 2

Constant slope method for hydrograph separation.

In this method, baseflow and direct runoff are separated by a straight line beginning at the point of the lowest discharge rate at the start of the event runoff, and extending at a constant discharge rate until it intersects the recession limb of the hydrograph. In this study, a straight line plot is used which starts from the lowest discharge value in the rising limb and extends with a constant slope until it intersects the recession limb of the hydrograph. This method is still used by hydrologists as a basis for comparing runoff in different watersheds, but the limitation of this method is that it does not reveal much about the hydrological process. Some assumptions were made before applying this techniques to separate out the snowmelt runoff in Beas Basin: (1) the baseflow is defined as the season with least frost; and (2) the contribution from rain is negligible.

Modeling approach

In the current study, the SNOWMOD model is applied to estimate the different components in the Beas River. The SNOWMOD model (Jain et al. 2001; Singh & Jain 2003) is unique as it simulates all components of runoff, i.e. snow/glacier melt runoff, rainfall-induced runoff and baseflow, using limited data. The SNOWMOD is a temperature index model, which is designed to simulate daily streamflow for a mountainous basin having contributions from both snowmelt and rainfall. The generation of streamflow from such basins involves the determination of the input derived from snow/glacier melt, rain, and its transformation into runoff. The basin is divided into a number of elevation zones and various hydrological processes relevant to snowmelt and rainfall runoff are evaluated for each zone. The model achieves three operations at each time step. At first, the available meteorological data are extrapolated at different altitude zones. Then, the rate of snowmelt is calculated at each time step. Finally, the snowmelt runoff from the snow cover area (SCA) and the rainfall runoff from the snow-free area are integrated, and these components are routed separately with proper accounting of baseflow to the outlet of the basin. The model (Singh & Jain 2003) optimizes the parameters used in routing the snowmelt runoff and rainfall runoff. To execute the SNOWMOD model, the following input data are required: (1) the physical features of the basin, which include snow cover area, the elevation zones and their areas, the altitude of the meteorological stations and other watershed characteristics affecting runoff; (2) time variable data including precipitation, air temperatures, the snow cover area, streamflow data and other parameters determining the distribution of temperature and precipitation; and (3) miscellaneous job control and time control data, which specify such items as total computation period, routing intervals, etc.

The information on SCA (Figure 3) is determined from the MODIS satellite data. The satellite data were processed using ERDAS Imagine. SCA was estimated and plotted against the elapsed time to construct the depletion curves for the various elevation zones in the basin.
Figure 3

Snow cover map of Beas Basin up to Manali, Himachal Pradesh.

Figure 3

Snow cover map of Beas Basin up to Manali, Himachal Pradesh.

To simulate daily runoff, the daily SCA for each zone was calculated as input to the model. Daily values of SCA were obtained by interpolating/extrapolating the derived depletion curves. Because the amount of snowfall/SCA and temperature conditions fluctuate from year to year, SCA and depletion trends were also estimated during the study period. The parameter values used in calibration of the model are given in Table 1.

Table 1

Parameter values used for the calibration of the model

S. no. Parameter Symbol Value 
1. Degree-day factor 3.0–7.0 mm oC−1 day−1 
2. Runoff coefficient for rain Cr 0.40–0.70 
3. Runoff coefficient for snow CS 0.50–0.80 
4. Temperature lapse rate δ Seasonally varying 
5. Critical temperature Tc 2 o
S. no. Parameter Symbol Value 
1. Degree-day factor 3.0–7.0 mm oC−1 day−1 
2. Runoff coefficient for rain Cr 0.40–0.70 
3. Runoff coefficient for snow CS 0.50–0.80 
4. Temperature lapse rate δ Seasonally varying 
5. Critical temperature Tc 2 o

RESULTS AND DISCUSSION

Precipitation trend and variability

Precipitation data in terms of rainfall have been collected from 1996 to 2008 at Manali, Himachal Pradesh. Figures 4 and 5 reveal the rainfall pattern at Manali on a 10-daily basis during 1996–2008. Annual rainfall during 1996, 1997, 1998, 1999 and 2000 was observed to be 1,766, 1,882.81, 2,343.2, 2,053.6 and 1,462.2 mm, respectively. While during the years 2001, 2002, 2003, 2004, 2005, 2006, 2007 and 2008, total rainfall was recorded as 883.3, 626.0, 584.2, 466.1, 821.9, 547.4, 468.4 and 962.1 mm, respectively, and 10 days' accumulative rainfall events of less than 50 mm have also been observed. On the basis of rainfall amount, the years 2002, 2003, 2004, 2006, 2007 and 2008 have been identified as the lowest rainfall years. It is obvious that rainfall of less than 50 mm (10 daily rainfall amount) does not significantly contribute to river discharge as surface runoff due to several initial losses.
Figure 4

Discharge pattern of Beas River during 1996–2000.

Figure 4

Discharge pattern of Beas River during 1996–2000.

Figure 5

Discharge pattern of Beas River from 2000 to 2008 up to Manali.

Figure 5

Discharge pattern of Beas River from 2000 to 2008 up to Manali.

Seasonal variation of discharge

Seasonal variation of runoff shows that the major part of the runoff is generated from the catchment area during the monsoon season followed by pre-monsoon, winter and post-monsoon (Figure 6). The minimum discharge during post-monsoon months is due to negligible amounts of rainfall and negligible contribution from snow/glacier melt runoff. Therefore, the lean flow period in the Beas River is during the post-monsoon and winter months. The discharge of the pre-monsoon months are comparatively higher than those of post-monsoon and winter months due to the snowmelt contribution from lower reaches resulting from an increase in air temperature.
Figure 6

Seasonal variation of rainfall in the Beas Basin.

Figure 6

Seasonal variation of rainfall in the Beas Basin.

Discharge pattern

The daily discharge data of the River Beas at Manali from 1996 to 2008 is plotted with rainfall and temperature (Figure 7). The hydrographs shows the rising trend from March onward and the first peak of river discharge is recorded in June (in almost each year), with the second highest discharge recorded in July/August (Figures 4 and 5).
Figure 7

Variation of daily discharge data with temperature rainfall.

Figure 7

Variation of daily discharge data with temperature rainfall.

The results clearly reveal that the rise in discharge during the pre-monsoon period (March–June) is mainly contributed to by the melting of snow. The temperature-induced melting plays a major role during the pre-monsson period. Since precipitation is less and rain events are mostly less than 20 mm for daily measurement (Figure 7) and less than 50 mm for 10 daily (Figure 8), the increase in discharge is controlled by increase in air temperature.
Figure 8

Variation of 10 daily discharge data with temperature and rainfall.

Figure 8

Variation of 10 daily discharge data with temperature and rainfall.

However, the second peak of the hydrograph is observed usually in the months of July and August, which fall under the monsoon period (July–September). During this period, the basin experiences more rainfall (Figure 6). Thus, the hydrograph peak developed in the months of July and August is a combined effect of snow/glacier melt from higher reaches and rainfall-derived runoff during July and August. Therefore, peak flows in the months of July and August generally correspond to temperature and rainfall events. The discharge data reveal that runoff is minimum during December–February in each year (Figure 8). During post-monsoon months, discharge becomes minimum due to insignificant rain events and a fall in air temperature lessening the melting of snow/glaciers.

An attempt has been made to establish the relationship between temperature and discharge at the Manali site during the study period. The monthly data of discharge and temperature from 2001 to 2008 have been used to derive the relationship at the Manali site. It is found that discharge is around 100 cusecs (∼2.83 m3s−1) when the temperature is near to 5 °C during January. The discharge of the river at Manali is approximately similar up to mid-February (Figure 9). However, from March onward as the temperature start to rise, the discharge of the river shows an increasing trend. The analysis of data for the year 2006 shows a good relationship between temperature and discharge due to much less rainfall in 2006 compared with other years.
Figure 9

Variation of discharge along with rainfall and temperature.

Figure 9

Variation of discharge along with rainfall and temperature.

In 2006, the rising and declining trend of the hydrograph follows the trend of the temperature (Figure 9). This relationship between temperature and discharge clearly indicates the contribution from the melt is the major component. The rainfall contribution is observed when a large amount of rain takes place, for example in 2005. However, the discharge increase from 100 cusecs (∼2.83 m3s−1) to 2,090 cusecs (59.18 m3 s−1) in January and May/June, respectively, is only due to the snowmelt contribution during 2006. The relationship between discharge and temperature has been developed for 2000–2008 (Figure 10). It reveals that temperature is the main controlling factor for the variation in runoff of the Beas River at Manali.
Figure 10

Relationship between temperature and discharge for the Beas River from 2000 to 2008 up to Manali.

Figure 10

Relationship between temperature and discharge for the Beas River from 2000 to 2008 up to Manali.

Estimation of snowmelt contribution using a conventional approach

A conventional separation technique (graphical separation technique) was applied in the Beas Basin up to Manali. The constant discharge method assumes that baseflow is constant during the storm hydrograph. The minimum streamflow immediately prior to the rising limb was used as the constant value. This technique is used to separate out the baseflow from the hydrograph of the Beas River. The discharge for the years 2002, 2003, 2006, 2007 and 2008 has been used to estimate the snowmelt contribution using the baseflow separation techniques. In these years, the rainfall in a day has been found to be less than 20 mm except for three events of approximately 100 mm. Figure 8 shows the snowmelt contribution in the Beas River, which is separated out using baseflow separation techniques. Thus, the snow and glacier melt contribution can be estimated using a simple equation. 
formula
 
formula
On average, annual snowmelt contribution into the Beas river at Manali was found to be 48%, 48%, 50%, 50% and 52% during the years 2002, 2003, 2006, 2007 and 2008, respectively.

Modeling of the streamflow of the Beas River

The flow data for the year 2005–2006 have been used to calibrate the model, whereas the years 2006–07, 2007–08 and 2008–09 have been used for validating the model. The efficiency of the model has been computed based on the daily simulated and observed flow values for 2 years. The values of the model efficiencies (R2) are 82%, 81%, 83% and 87% for the years 2005–06, 2006–07, 2007–08 and 2008–09, respectively. The performance of the model in preserving the runoff volume has been tested by criteria computed as the percentage difference in observed and simulated runoff (D) during the entire years. These values are 6.52%, 5.41%, 6.51% and 5.16% for the years 2005–06, 2006–07, 2007–08 and 2008–09, respectively. The comparison on a daily basis is shown in Figure 11 for the year 2006–07 and Figure 12 for the year 2007–08. The model has separated out the simulated and observed flow hydrographs for all contributions of rainfall, snow and glacier melt and baseflow from the simulated flows. It has been also observed that the model has simulated the daily flows reasonably well generally showing a good match with the daily observed flows. The trends and peaks of the daily flow hydrographs for the entire period are very well simulated by the model. The percentage difference in volume, model efficiency and contributions of rain, snow and baseflow computed by the model for 4 years are given in Table 2. Most of the high peaks observed in the daily flow hydrographs are generally during June and August, which shows the contribution of the glacier melt in June and also due to rain in August. Thus, these months are considered as the peak melting season in the western Himalayan region. However, sometimes flow resulting from high rainfall also reflects the peaks in the daily flow hydrographs. The simulation of baseflow indicates that the baseflow contribution to the streamflow increases as the season advances, being at a maximum during the peak melting season, and then starts decreasing.
Table 2

Difference in volume, model efficiency and contributions of rain snow and baseflow computed by the model

Year Difference in volume (%) Model efficiency (%) Rain (%) Snow (%) Baseflow (%) 
2005–06 6.52 82 6.00 54 40 
2006–07 5.41 81 5.40 55 39.6 
2007–08 6.51 83 5.10 54 40.9 
2008–09 5.16 87 5.00 56 39.0 
Year Difference in volume (%) Model efficiency (%) Rain (%) Snow (%) Baseflow (%) 
2005–06 6.52 82 6.00 54 40 
2006–07 5.41 81 5.40 55 39.6 
2007–08 6.51 83 5.10 54 40.9 
2008–09 5.16 87 5.00 56 39.0 
Figure 11

Observed and simulated daily streamflow for the Beas River at Manali for the period 2006–07.

Figure 11

Observed and simulated daily streamflow for the Beas River at Manali for the period 2006–07.

Figure 12

Observed and silmulated daily streamflow for the Beas River at Manali for the period 2007–08.

Figure 12

Observed and silmulated daily streamflow for the Beas River at Manali for the period 2007–08.

CONCLUSIONS

The variation in climatic conditions is responsible for the variation in streamflow discharge and also affects the contribution of different components to the streamflow discharge. Snow accumulation in the Himalayas occurs generally from December to March, sometimes extending from October to April. On other hand, snowmelt is observed from April to June, which are the pre-monsoon months under Indian sub-continental conditions.

The streamflow discharge is minimum during the winter months. It starts rising from the month of March and continues an increasing trend up to September. In these months, the rise in discharge is mainly due to snow/glacier melt during the pre-monsoon months. However during the monsoon months (July–September), the river receives a contribution from the high reaches ice melt and rainfall-induced runoff. Therefore, the river gains maximum discharge during this period, whereas it receives minimum discharge during the winter season.

The results indicate that the estimated annual average contribution of snow/glacier melt water in the Beas River comes in at around 50% (using a conventional approach) and approximately 54% using the streamflow modeling technique (SNOWMOD).

However, these techniques are used to understand the annual contribution of snowmelt in the Beas River. Further investigations are required for more precise quantification of the snow/glacier contribution on a temporal and spatial level using other techniques such as isotopic techniques. It is always recommended to separate out the different components to river flow with more than one method.

ACKNOWLEDGEMENTS

The authors are grateful to Bhakra Beas Management Board (BBMB) for providing the data used in this study. The authors are also grateful to the Director NIH, Roorkee, and Director, WIHG, for providing the necessary facility to publish this work.

REFERENCES

REFERENCES
Ahluwalia
R. S.
Rai
S. P.
Jain
S.
Bhisham
K.
Dobhal
D. P.
2013
Assessment of snowmelt runoff modeling and isotope analysis: a case study from the western Himalaya, India
.
Annals of Glaciology
54
(
62
),
299–304
.
Arora
M.
Singh
P.
Goel
N. K.
Singh
R. D.
2008
Climate variability influences on hydrological responses of a large Himalayan basin
.
Water Resources Management
22
(
10
),
1461
1475
.
Census of India
2011
http://censusindia.gov.in.
Das
P. K.
2013
The Monsoons, 12th
edn
.
National Book Trust
,
New Delhi, India
.
Haritashya
U. K.
Singh
P.
Kumar
N.
Gupta
R. P.
2006
Suspended sediment from the Gangotri Glacier: quantification, variability and associations with discharge and air temperature
.
Journal of Hydrology
 
321
(
1–4
),
116
130
.
Jain
S. K.
Rathore
D. S.
Chaudhary
A.
2001
Snow cover estimation in Himalayan basins using remote sensing
.
Report no. CS (AR)-10/2000-01
,
National Institute of Hydrology
,
Roorkee
,
India
.
Jain
S. K.
Singh
P.
Kumar
N.
1997
Estimation of snow and glacier-melt contribution to the Chenab river, Western Himalaya
.
Mountain Research and Development
17
(
1
),
49
56
.
Jain
S. K.
Goswami
A.
Saraf
A. K.
2010
Assessment of snowmelt runoff using remote sensing and effect of climate change on runoff
.
Water Res. Mgt.
24
,
1763
1777
.
Kulkarni
A. V.
Rathore
B. P.
Singh
S. K.
Ajai
.
2010
Distribution of seasonal snow cover in central and western Himalaya
.
Annals of Glaciology
51
(
54
),
123
128
.
Kumar
V.
Singh
P.
Jain
S. K.
2005
Rainfall trends over Himachal Pradesh, Western Himalaya, India
. In:
Proceedings Conference on Development of Hydro Power Projects – A Prospective Challenge, Shimla, April 20–22
,
Central Board of Irrigation and Power
,
New Delhi
,
India
.
Prasad
V. H.
Roy
P. S.
2005
Estimation of Snowmelt Runoff in Beas Basin
.
India Geocarto International
20
(
2
),
41
47
.
Shashi Kumar
V.
Paul
P. R.
Ramana Rao
Ch. L. V.
Haefner
H.
Seibel
K.
1993
Snowmelt Runoff Forecasting Studies in Himalayan Basin Snow and Glacier Hydrology
. In:
Proceedings of the Kathmandu Symposium, Kathmandu, November 1992
,
IAHS Publication No. 218
, pp.
85
94
.
Singh
P.
Bengtsson
L.
2004
Hydrological sensitivity of a large Himalayan basin to the climate change
.
Hydrol. Processes
,
18
,
2363
2385
.
Singh
P.
Jain
S. K.
Kumar
N.
1997
Estimation of snow and glacier contribution to the Chenab River, Western Himalaya
.
Mountain Research and Development
17
(
1
),
49
56
.
Singh
P.
Jain
S. K.
2002
Snow and glacier melt in the Satluj River at Bhakra Dam in the western Himalayan region
.
Hydrological Sciences Journal
47
(
1
),
93
106
.
Singh
P.
Jain
S. K.
2003
Modelling of streamflow and its components for a large Himalayan basin with predominant snow melt yields
.
Hydrological Sciences Journal
48
(
2
),
257
275
.
Singh
P.
Ramasastri
K. S.
Kumar
N.
1995
Topography influence on precipitation distribution in different ranges of western Himalaya
.
Nordic Hydrology
26
(
4–5
),
259
289
.
Siva Sankar
E.
2002
Spatial snowmelt modelling in Sutlej river basin, Western Himalayas
.
IAPRS and SIS
Vol. 34, Part 7
Resource and Environmental Monitoring
,
Hyderabad
.