Impact of warmer climate on melt and evaporation for the rainfed,

snowfed and glacierfed basins in the Himalayan region

Pratap Singha,*, Lars Bengtssonb

aHead, Mountain Hydrology Division, National Institute of Hydrology, Roorkee 247 667 (U.A.), IndiabDepartment of Water Resources Engineering, Lund University, SE-221 00 Lund, Sweden

Received 21 January 2003; revised 3 June 2004; accepted 3 June 2004

Abstract

The impact of warmer climate on melt and evaporation was studied for rainfed, snowfed and glacierfed basins located in the

western Himalayan region. Hydrological processes were simulated under current climatic conditions using a conceptual

hydrological model, which accounts for the rainfall–runoff, evaporation losses, snow and glacier melt. After simulations of

daily observed streamflow ðR2 ¼ 0:90Þ for 6 years, the model was used to study the impact of warmer climate on melt and

evaporation. Based on the future projected climatic scenarios in the study region, three temperature scenarios (T þ 1; T þ 2 and

T þ 3 8C) were adopted for quantifying the effect of warmer climate. The comparison of the effect of warmer climate on

different types of basins indicated that the increase in evaporation was the maximum for snowfed basins. For a T þ 2 8C

scenario, the annual evaporation for the rainfed basins increased by about 12%, whereas for the snowfed basins it increased by

about 24%. The high increase of the evaporation losses would reduce the runoff. It was found that under a warmer climate, melt

was reduced from snowfed basins, but increased from glacierfed basins. For a T þ 2 8C scenario, annual melt was reduced by

about 18% for the studied snowfed basin, while it increased by about 33% for the glacierfed basin. Thus, impact of warmer

climate on the melt from the snowfed and glacierfed basins was opposite to each other. The study suggests that out of three types

of basins, snowfed basins are more sensitive in terms of reduction in water availability due to a compound effect of increase in

evaporation and decrease in melt. For a complex type of basin, the decrease in melt from seasonal snow may be counterbalanced

by increase in melt from glaciers. However, on long-term basis, when the areal extent of glaciers will decrease due to higher

melt rate, the water availability from the complex basins will be reduced.

q 2004 Elsevier B.V. All rights reserved.

Keywords: Rainfed basin; Snowfed basin; Glacierfed basin; Warmer climate

1. Introduction

Concentration of carbon dioxide (CO2) and other

trace gases in the atmosphere has increased over

the last century, resulting in global warming.

Recently, IPCC (2001a) has indicated that the average

global surface air temperature has increased by

0.6 ^ 0.2 8C since the late 19th century and it is

projected to increase by 1.4–5.8 8C over the period

1990–2100. The land areas will warm more rapidly

than the global average. The warming in the northern

regions of North America, and northern and Central

Asia, may exceed global mean warming by more than

Journal of Hydrology 300 (2005) 140–154

www.elsevier.com/locate/jhydrol

0022-1694/$ - see front matter q 2004 Elsevier B.V. All rights reserved.

doi:10.1016/j.jhydrol.2004.06.005

* Corresponding author. Fax: þ91-1332-72123.

E-mail address: pratap@nih.ernet.in (P. Singh).

40%. Studies show that the extent of global snow

covered and glacierized area has been reduced due to

climatic changes over the last century. It is observed

that the extent of snow cover has decreased by 10%

since late 1960s (IPCC, 2001a). Long-term records on

glacier fluctuations on the global scale indicating the

mass loss/retreat of mountain glaciers support the

change in the climate in the past century (Letreguilly

and Reynaud, 1990). However, in contrast, some

glaciers like Nigardsbreen (Norway) and Franz Josef

Glacier (New Zealand) have shown advancing trend

(Oerlemans et al., 1998). On the whole, a widespread

retreat of mountain glaciers has been observed in non-

polar regions during the 20th century. These glaciers

and ice caps are projected to continue their wide-

spread retreat during the 21st century. Climatic

changes, which occurred during the 20th century,

have had a pronounced effect on the glacial and

periglacial extent of the Alps. Since the middle of the

past century, the areal extent of glacierization in

the European Alps is reduced 30–40%, whereas the

volume of ice has been reduced by 50% (Haeberli and

Beniston, 1998). IPCC (2001b) indicated that half of

Europe’s alpine glaciers could disappear by the end of

21st century. Under the warmer climate, the amount

of solid precipitation will reduce in total precipitation,

resulting in a decrease in snow accumulation. Jones

(1999) reported that the climate change will affect the

seasonal altitudinal and latitudinal shift in the freezing

line. The result is likely to be decreased snowfall over

most regions, combined with more rapid melting at

lower altitudes and latitudes, reducing snow cover

durations. Dyurgerov (2002) reported that existing

trend of changes in the volume of glaciers shows that

melting of glaciers will accelerate in continental

regions, North America, South America, Central Asia,

sub-polar glaciers and will contribute to sea level rise.

Recent global climate analysis (IPCC, 2001b) has

indicated that the climate change is likely to change

streamflow volume, as well as the temporal distri-

bution throughout the year over Asian region,

imposing significant stress on the water resources in

the region. An examination of the possible effects

of climate change in the design and management

of water resources systems was suggested by WMO

et al. (1991).

The major river systems of the Indian sub-

continent, Brahamaputra, Ganga and Indus, which

originate in the Himalayas are expected to be much

vulnerable to climate change because of substantial

contribution from snow and glaciers (Singh et al.,

1997a; Singh and Jain, 2002). During winter, a large

extent of mountainous area of Himalayan river

basins is covered by snow. For example, about 65%

of the present study basin is covered by snow by the

end of winter and about 15% remains under

perpetual snow/glaciers. The response of hydrologi-

cal systems, erosion processes and sedimentation in

this region could alter significantly due to climate

change. It is understood that the global warming and

its impact on the hydrological cycle and the nature

of hydrological events would pose an additional

threat to the Himalayan region. Extreme precipi-

tation events in the Himalayas may cause wide-

spread slope failures (Ives and Messerli, 1989).

Possible impacts of climatic changes on various

aspects of hydrological cycle are not much studied

(Divya and Mehrotra, 1995; Singh and Kumar,

1997a).

The changes in the hydrological response of a basin

will depend on the sources of runoff, climatic

conditions, physical characteristics of the basin and

the magnitude of projected climatic scenarios. Thus,

basins located in different regions will experience

different impact of the variability in the climate (Chiew

et al., 1995). Cayan et al. (1993) examined the

influence of climate parameters on seasonal stream-

flow in watersheds over a range of elevations in

California and Oregon. They studied the effect of

precipitation, temperature and snow water content on

streamflow using linear regression models. They

reported a large difference in the lower and high altitude

basins because of changes in their runoff distribution

and climatic regime. Precipitation had the greatest

influence on streamflow variations in spring, while

temperature was important in spring in mid and high

elevations. By late spring, snow water content

accounted for nearly all summer streamflow variations

at mid and high elevations. In the Himalayan region, a

variety of basins exist and broadly, in the terms of

source of runoff, these can be categorized in there types

of basins: (1) rainfed basins: runoff is generated

exclusively from rainfall; the altitude of such basins

varies from about 500 to 2000 m, (2) snowfed basins:

runoff is generated both from rainfall–runoff; snow-

melt; the altitude varies from 2000 to about 4000 m

P. Singh, L. Bengtsson / Journal of Hydrology 300 (2005) 140–154 141

altitude. The contribution from snowmelt increases

with altitude. The precipitation and temperature

patterns in this type of basins are such that snowfall

occurring during the preceding winter is completely

melted away during next spring and summer months,

i.e. such basins receive seasonal snow, (3) glacierfed

basins: runoff from such basins is primarily generated

from the melting of permanent snow fields and

glaciers. Direct rainfall contribution is not significant.

These are high altitude basins and cover an elevation

range about 4000 to 6000–7000 m. A schematic

diagram showing the form of precipitation for different

types of basins located at different altitudes is shown in

Fig. 1. A basin, which receives the contribution to

streamflow from all types of sources like rainfall,

snowmelt and glacier melt, is defined as a complex type

of basin. It is understood that impact of climate change

will be different for different types of basins. In the

present study, attempts have been made to investigate

the impact of warmer climate on the melt and

evaporation for these three types of basins located in

the same region but having different hydrological

characteristics.

2. Projected climatic changes over Indian

sub-continent and India

Warming of the Indian sub-continent by 0.4 8C

over the period 1901–1982 was reported by Hingane

et al. (1985), indicating that this warming since 1900

is broadly consistent with observed global warming

over the last century. IPCC (1990) reported that for

the Indian sub-continent the warming will be between

1 and 2 8C by 2030. Recently, Lal and Singh (2001)

and Lal (2001) have studied the impact of increase in

greenhouse gases on temperature and precipitation

using the different GCMs. They have reported that the

average annual mean surface temperature over Indian

sub-continent is likely to increase by about 2.7 and

3.8 8C during the decades of 2050s and 2080s,

respectively. Moreover, it is likely that over inland

regions of the Indian sub-continent, the mean surface

temperature may rise between 3.5 and 5.5 8C by 2080

(Lal, 2001). On seasonal basis, the projected surface

warming is higher in winter than in summer. The

increase in annual mean precipitation over the

Indian sub-continent is projected to be 7 and 11%,

Fig. 1. A schematic presentation of the rainfed, snowfed and glacierfed basins.

P. Singh, L. Bengtsson / Journal of Hydrology 300 (2005) 140–154142

respectively, during the decades of 2050s and 2080s.

Over the Indian sub-continent as a whole, no

significant changes in winter precipitation are pro-

jected in any of the four GCMs. IPCC (1990)

projected that precipitation will change little in winter

and will generally increase throughout the region by

5–15% in summer. The spatial distribution of surface

warming suggests that north India, the region where

the present study area is located, may experience an

annual mean surface warming of 3 8C or more by

2050s. GCM models simulate peak warming of 3 8C

over north and central India in winter. Simulation

studies made by Lal et al. (1992) using Hamburg

global coupled-atmosphere–ocean circulation model

indicated the possibility of an increase of rainfall in

parts of northern India while decrease in rainfall in

southern parts of peninsular India. As such under

warmer climate, variability in Asian summer mon-

soon is expected to increase along with changes in the

frequency and intensity of extreme climate events in

this region.

3. Application of hydrological models for studying

the impact of climatic changes

The studies related to the impacts of climate

change on various components of the, hydrologic

cycle may be classified broadly into two categories;

(i) studies using GCMs directly to predict impact of

climate change scenarios (Cohen, 1986; Gleick, 1987;

Sausen et al., 1994; McCabe, 1994; Loaiciga et al.,

1996), and (ii) studies using hydrological models with

assumed plausible hypothetical climatic inputs

(Nemec and Schaake, 1982; Nemec, 1989; McCabe

and Ayers, 1989; Sanderson and Smith, 1990;

Thomsen, 1990; Rango, 1992; Cayan et al., 1993;

Burn, 1994; Rango and Martinec, 1994; Chiew et al.,

1995; Singh and Kumar, 1997a). Although GCMs are

considered as invaluable tools for identifying climatic

sensitivities and changes in global climate character-

istics, their application for studying the impact of

climate change on the hydrological response on the

basin scale is limited due to their poor spatial

resolution. A single grid of GCM may encompass

hundreds of square kilometres, including mountai-

nous and desert terrain, oceans and land areas.

Usually, the output of GCMs is given for much larger

scale than the scale of even a large watershed.

Interpolation and disaggregation schemes have

been used to overcome the spatial resolution limi-

tation of the GCMs (Epstein and Ramirez, 1994;

Bardossy and Van Mierlo, 2000). Jones (1999)

suggested for further research linking GCM output

with hydrological simulation models at all scales of

catchments.

Application of conceptual hydrological models is

understood to be a suitable approach for assessing the

expected changes in the hydrological response due to

changes in temperature and precipitation and other

climatic variables on the basin scale. The input

variables to the hydrological models can be either

hypothetical climate scenarios or the output of GCMs.

The ability of hydrologic models to incorporate

projected variations in climatic variables makes

them especially attractive for climatic change impact

studies related to water resources. Depending on the

purpose of the study and data availability, various

hydrologic models have been used to study the

impacts of climate change for snowfed basins. Gleick

(1987) used a water balance model to estimate the

impact of climate on monthly water availability.

Detailed studies using a deterministic National

Weather Service River Forecasting System

(NWSRFS) model were carried out in mountain

basins (Lettenmaier and Gan, 1990; Cooley, 1990;

Nash and Gleick, 1991; Panagoulia, 1991). However,

NWSRFS model consists of several components, but

in most of the above-mentioned studies, a coupling of

soil moisture accounting model that calculates gains

and losses of water in the soil through various

processes (e.g. evaporation, transpiration, infiltration)

and a snow accumulation and ablation model that

calculates the accumulation of snow and the contri-

bution of snowmelt to soil moisture and runoff, have

been used. Rango (1992) used the snowmelt runoff

model (SRM) for Rio Grande and Kings river basins

to study the changes in snowmelt runoff under warmer

climate scenarios. Rango and Martinec (1994)

examined the influence of changes in temperature

and precipitation on the snow cover using SRM. Singh

and Kumar (1997a) used University of British

Columbia (UBC) watershed model (Quick and

Pipes, 1977) for studying such effects for a high

altitude river. In this study, a conceptual snowmelt

model has been used to assess the impact of global

P. Singh, L. Bengtsson / Journal of Hydrology 300 (2005) 140–154 143

warming on the different types of basins in the

Himalayan region. The model was tested for simulat-

ing runoff from the basin before using it for changed

climatic scenarios.

4. Brief description of the structure of snowmelt

model used

The streamflow of the Himalayan rivers consists of

contributions from melting of snow and ice, and

rainfall-induced runoff. Data on precipitation, tem-

perature and river flow are scarce because of

inadequate hydrometeorological network in the high

altitude regions with rugged terrain and poor acces-

sibility. A hydrological model, which can handle both

snowmelt and rainfall runoff efficiently, using limited

data, is considered most suited for the simulation of

streamflow generated from the Himalayan basins. A

conceptual snowmelt model (SNOWMOD) was used

for simulation of daily streamflow for study basin.

A detailed description of the model has been given by

Singh and Jain (2003). The structure of the model is

shown in Fig. 2. Because of poor availability of

snowfall data in the Himalayan basins, snow-covered

area (SCA) derived from satellite data was used in the

model. However, SCA can be obtained at different

temporal and spatial resolutions, but the frequency of

data should be such that (,15 days) the depletion

trend is well established for the study basin. The

basic inputs to the model were temperature, rainfall,

SCA and glacier-covered area. In addition to the

meteorological data, physiographical information of

the basin, which includes total area of the basin, its

altitudinal distribution through elevation bands and

their respective areas, altitude of precipitation and

temperature stations are also required. Jain (2001)

simulated the streamflow of Satluj Basin using the

model used in this study, but impact of warmer

climate was not studied.

The model considers the mountain basin as a

number of elevation zones depending upon

Fig. 2. Flow chart of the snowmelt model.

P. Singh, L. Bengtsson / Journal of Hydrology 300 (2005) 140–154144

the topographic relief and computes runoff from each

elevation zone separately. The model deals with

snowmelt and rainfall runoff by performing the

following three operations at each time step:

(i) extrapolate available meteorological data to the

different elevation zones, (ii) calculate melt rates

and/or rainfall at different points, and (iii) integrate

the melt runoff from SCA and rainfall runoff from

snow free area (SFA), and route these components

separately with proper accounting of subsurface flow

to the outlet of the basin. For the purpose of

computation of different components of runoff, each

elevation zone is treated as a separate watershed with

its own characteristics. Total streamflow for the whole

basin is obtained by synthesizing the runoff from all

elevation zones.

Estimation of melt using the energy balance

approach requires much climatic data such as

radiation, cloudiness, wind speed, etc. and such

meteorological data are hardly available in the

Himalayan region. Therefore, the following tempera-

ture index or degree-day approach has been used to

compute the melt in the basin.

M ¼ DðTi 2 TbÞ ð1Þ

where M is the depth of melt water (mm) produced

in a unit time (1 day in present case), D is the

degree-day factor (mm 8C21 d21), Ti is the index air

temperature (8C), Tb is the base temperature

(usually, 0 8C). Clearly, D is used to convert the

degree-days to melt expressed in depth of water. D

is influenced by the physical properties of snowpack

and changes with time because properties of snow

change with time. D is lower in the beginning of

melt season and higher towards the end of melt

season. Daily mean temperature is the most

commonly used index of temperature for snowmelt.

The model also computes the melt occurred due to

rainfall over the snowpack using the following

equation (Singh and Singh, 2001).

Mr ¼ 4:2TrPr=325 ð2Þ

where Mr is the melt caused by the energy supplied

by rain (mm d21), Tr is the temperature of rain or

air temperature during rain (8C), Pr is the depth of

rain (mm d21). It is to be noted that only high

rainfall events occurring at higher temperatures

would cause the melting due to rain, otherwise

this component would not be so significant (Singh

et al., 1997b).

In general, air temperatures are available at few

locations in the basin. These point values are

extrapolated or interpolated to mid elevation of each

elevation zone using a predefined temperature lapse,

as given below.

Tij ¼ Ti;base 2 dðhj 2 hbaseÞ ð3Þ

where Ti;j is daily mean temperature, on ith day in jth

zone (8C), Ti;base is daily mean temperature (8C) on ith

day at base station, hj is zonal hypsometric mean

elevation (m), hbase is elevation of base station (m) and

d is temperature lapse rate (8C/100 m). The tempera-

ture lapse rate of 0.65 8C/100 m was adopted for the

present study basin (Singh, 1991). The temperature in

a particular elevation zone decides the form of

precipitation and model handles it accordingly. A

critical temperature, Tc (2 8C), was specified in the

model to determine whether the measured precipi-

tation was rain or snow.

The model computes surface runoff from SCA

and SFA. The response function (runoff coefficient)

is different for SCA and is determined by the ratio

of available soil moisture to the field capacity. The

runoff from SCA consists of: (i) melt caused due

to prevailing air temperature, (ii) under rainy

conditions, snowmelt due to heat transferred to the

snow from rain, and (iii) runoff from rain itself

falling over SCA. Daily total runoff from these three

sources of runoff from SCA, QSCA is computed by

adding contribution from each elevation zone and is

given by

QSCA ¼Xn

j¼1

ðMs;i;j þ Mr;i;j þ Rs;i;jÞ ð4Þ

where n is the total number of zones. Different

components of runoff from each elevation zone of

SCA are computed using the following equations,

Ms;i;j ¼ Cs;i;jDi;jTi;jSc;i;j ð5Þ

Mr;i;j ¼ 4:2Ti;jPi;jSc;i;j=325 ð6Þ

Rs;i;j ¼ Cs;i;jPi;jSc;i;j ð7Þ

where

P. Singh, L. Bengtsson / Journal of Hydrology 300 (2005) 140–154 145

Ms;i;j melt runoff depth on ith day for jth zone

(mm d21)

Mr;i;j melt runoff depth due to rain on snow on ith

day for jth zone (mm d21)

Rs;i;j runoff from rainfall occurred on SCA on ith

day for jth zone (mm d21)

Pi;j rainfall ith day for jth zone (mm d21)

Cs;i;j response function (runoff coefficient) for

melt on ith day for jth zone

Di;j degree-day factor on ith day for jth zone

(mm 8C21 d21)

Ti;j temperature on ith day for jth zone (8C)

Sc;i;j ratio of SCA to the total area of jth zone on

ith day.

As discussed above, the response function or the

runoff coefficient is the function of available soil

moisture and field capacity. Therefore, moisture input

(melt, rainfall) to the soil controls the value of runoff

coefficient. Thus, runoff coefficient is not constant and

varies with time. Similarly, the runoff from SFA,

QSFA; for all the zones is computed using the

following equation,

QSFA ¼Xn

j¼1

Rf ;i;j ð8Þ

It is to be noted that SCA and SFA are

complimentary to each other, therefore, Sf ;i;j can be

directly calculated as ð1 2 Sc;i;jÞ: The source of

surface runoff from the SFA is only rainfall. Like

melt runoff computations, runoff from the SFA was

also computed for each zone using the following

expression.

Rf ;i;j ¼ Cr;i;jPi;jSf ;i;j ð9Þ

where

Pi;j rainfall on snow on ith day for jth zone

(mm d21)

Cr;i;j response function (runoff coefficient) for rain

on ith day for jth zone

Sf ;i;j ratio of SFA to the total area of jth zone on

ith day.

The subsurface flow, Qb; represents the runoff from

the unsaturated zone of the basin to the streamflow.

After accounting for evaporation losses and direct

surface runoff from the melt and rainfall, the

remaining water contributes to the groundwater

storage through infiltration and appears at the outlet

of the basin with much delayed effect as subsurface

flow. The depletion of this groundwater storage also

takes place due to evaporation and percolation of

water to deep groundwater zone. Evaporation losses

from the SFA are governed by temperature and

available soil moisture. Evaporation losses from the

SCA are very small and therefore, not considered in

the present computation (Bengtsson, 1980). The

subsurface flow from the basin was computed using

the following equation.

Qb ¼Xn

j¼1

Rb;i;j ð10Þ

where Rb;i;j is the depth of runoff contributing to

subsurface flow from each zone and is given by

Rb;i;j ¼ b½ð1 2 Cr;i;jÞRf ;i;j þ ð1 2 Cs;i;jÞMt;i;j� ð11Þ

where Mt;i;j ¼ Ms;i;j þ Mr;i;j þ Rs;i;j; and b is factor

controlling the deep water percolation.

Different components of streamflow were routed

separately because of differences in their hydrological

response. For a basin, which gets runoff from both

melt and rainfall, the hydrological response of the

basin at a particular time depends upon the areal

extent of SCA and SFA in the basin. In the beginning

of the melt season, the higher extent of SCA provides

much delayed response of melt as compared to the

later part of melt season. Such delayed response is

generated because of time taken by the melt water first

to infiltrate through snowpack and then to flow as

overland flow under the snowpack. The response of

snowmelt to streamflow improves with the advance-

ment of the melt season because the areal extent of

SCA reduces with time due to the melting of snow.

Further, SFA is complimentary to SCA and therefore,

as the melt season progresses, the SCA decreases and

SFA increases. The extent of SFA is minimum in the

beginning of melt season and therefore, rainfall has

lowest time of concentration during this period. But,

with progress in melt season, the time of concen-

tration for rainfall increases due to increase in SFA.

Thus, on the basin scale, the response of melt

from SCA improves with melt season, whereas for

rainfall from SFA it becomes relatively slower.

P. Singh, L. Bengtsson / Journal of Hydrology 300 (2005) 140–154146

Therefore, keeping in view the different responses of

snowmelt and rain and their variations with time, both

components were routed separately by splitting the

basin into two parts, namely SCA and SFA. The linear

cascade reservoir approach was used for routing of

different components from each part of the basin. The

storage coefficients for SCA and SFA were considered

as function of their respective areas and were

optimized. Storage coefficient for the subsurface

flow was computed using streamflow data of recession

periods (winter season) and was used for routing of

subsurface flow.

Finally, daily total streamflow, Q; emerging out

from the basin is calculated by adding the different

routed components of discharge for each day, as given

below:

Q ¼ QSCA þ QSFA þ Qb ð12Þ

5. Study area and its partition based on sources

of runoff

The Satluj River basin upstream the Bhakra

Reservoir (Indian part) located in the western

Himalayan region was selected for the present study

(Fig. 3). The Satluj River rises in the lakes of

Mansarover and Rakastal in the Tibetan plateau at

an elevation of more than 4500 m and forms a part of

the Indus River system. Because of suitable topo-

graphy and availability of abundant water, this basin

has a huge hydropower potential. Several hydropower

schemes exist and more are planned for this river. The

total area of the study basin is about 22,275 km2 and

elevation varies from about 500 to 7000 m. About

65% of the basin area is covered with snow during

winters (Singh and Jain, 2002).

The study area is characterized by diversified

climatic patterns. The westerly weather disturbances

deposit nearly all the precipitation in the form of snow

during the winter months in the upper part and middle

part of the basin. Singh and Jain (2002) have reported

that annual flow of the study basin is about 550 mm, in

which about 60% contribution derived is from

the melting of snow and glaciers. The mean

annual rainfall in the outer, middle and greater

Himalayan ranges of the basin is about 1300, 700

and 200 mm, respectively (Singh and Kumar, 1997b).

The distribution of rainfall indicates that mostly

rainfall is concentrated in the lower and middle

parts of the basin and significantly reduces in the

middle and upper part of the basin. There are very

limited records available on the evaporation. Distri-

bution of precipitation, development and depletion of

snow, and wide range of basin relief allowed to divide

the study area into rainfed, snowfed and glacierfed

Fig. 3. Location of Satluj Basin (Indian part) up to Bhakra Dam along with network of hydrometeorological stations having rainfall ðPÞ;

temperature ðTÞ and discharge ðQÞ data.

P. Singh, L. Bengtsson / Journal of Hydrology 300 (2005) 140–154 147

basins. As shown in Fig. 4, the total basin was divided

into 10 elevation zones for the modelling of sream-

flow. Zones 1–3, 4–8, and 9 and 10 represent rainfed,

snowfed and glacierfed parts of the basin, respect-

ively, and their respective areas are 4295, 14,271 and

3709 km2.

6. Data used and simulation of streamflow

Daily mean temperature of 5 stations and rainfall

data of 9 stations were used in the present study. The

altitude of the meteorological stations varied from 518

to 3639 m. Daily mean river discharge available at

Bhakra Dam was used. SCA for the study period was

obtained analysing the available remote sensing data

from different satellites. Landsat (Multi-spectral

scanner (MSS), 80 m resolution) data for 2 years

(1985/86–1986/87), Indian Remote Sensing Satellite

(IRS) (Linear imaging self-scanning system (LISS-I),

72.5 m resolution) data for 4 years (1987/88–

1990/91) and IRS (Wide field sensor (WiFS), 180 m

resolution) data for 3 years (1996/97–1998/99) were

procured from National Remote Sensing Agency

(NRSA), Hydrabad, India and, processed and used

in the study. In order to obtain extent of SCA of the

basin, the image analysis was carried out using Earth

Resources Data Analysis System (ERDAS IMA-

GINE, 1997) software. SCA for each elevation zone

was plotted against the elapsed time to construct the

depletion curves for the various elevation bands in the

basin. Keeping in view the cost involved in procuring

remote sensing data, the satellite data were obtained at

the frequency of 15 days or 1 month and extrapolated

for intermediate periods.

The model was calibrated for simulating the

streamflow using data of 3 years (1985/86 –

1987/88). After calibration of the model, the model

was used to simulate daily streamflow using indepen-

dent data for 3 years (1988/89–1990/91). The

comparison of daily observed and simulated stream-

flow for 1985/86–1987/88 and 1988/89–1900/91 is

shown in Fig. 5(a) and (b), respectively. Table 1 also

shows a comparison of mean monthly and mean

annual observed and simulated discharge for cali-

bration and validation periods. A good agreement

with respect to peak flows as well as with respect to

low flows was observed for all the 6 years of

simulation. These results indicate that the model

was able to simulate mean monthly flows with

reasonably good accuracy. The overall efficiency of

the model, explained variance, R2 (Nash and Sutcliffe,

1970) over the study period of 6 years was about 0.90.

The difference in total volume ðDvÞ of computed and

observed streamflow was about 2.0% and root mean

square error (RMSE) was about 0.33 min.

In regions like the Himalayas, climatic conditions

and hence the sources of runoff and land use pattern

change with altitude. There are water resources

projects on the rivers, which have runoff either only

from rain, or a combination of rain and melt from

seasonal snow/glaciers. Therefore, it becomes import-

ant to study the effect of climate change on each type

of basin in the same region. Separate streamflow data

are needed for each type of basin for simulating the

streamflow and then to study the climate change

impact. However, in practice, streamflow data for

such individual basins are not available for the

Himalayan rivers. Mostly the gauging sites are located

at lower altitudes and streamflow observed at

these sites has contributions from rainfall, seasonal

Fig. 4. Area of different elevation zones as percentage of total basin

area and area-elevation curve for the Satluj Basin upstream of

Bhakra Reservoir (Indian part).

P. Singh, L. Bengtsson / Journal of Hydrology 300 (2005) 140–154148

snowmelt and glacier melt. In such a case, it is

possible to simulate the hydrological processes for the

whole basin contributing to runoff and then to study

the expected changes on the different parts of the

basin representing different hydrological character.

This approach has been used in the present study.

7. Adopted future climatic scenarios

Following the projected increase in temperature in

the study region, the present study was carried out for

three temperature scenarios (T þ l; T þ 2 and

T þ 3 8C). The adopted increase in temperature from

1 to 3 8C represents the estimates of expected changes

in climatic variables over the study region (IPCC,

1990; Lal et al., 1992; Lal and Singh, 2001; Lal,

2001). Since large uncertainty is associated with

projected change in precipitation and its distribution,

especially for the snow, the investigations in present

study were restricted to assess the impact of warmer

temperatures. Singh and Kumar (1997a) also adopted

similar temperature scenarios. Like in most other

reported studies, temperatures were uniformly

varied by the projected amount of changes over

the simulation period (Nemec and Schaake, 1982;

Fig. 5. (a) Observed and simulated daily streamflow for the Satluj River at Bhakra for 3 years (Calibration period: 1985/86–1987/88). (b)

Observed and simulated daily streamflow for the Satluj River at Bhakra for 3 years (Simulation period: 1988/89–1990/91).

P. Singh, L. Bengtsson / Journal of Hydrology 300 (2005) 140–154 149

Ng and Marsalek, 1992; Rango and Martinec, 1994;

Singh and Kumar, 1997a).

Modified SCA for each temperature scenario is

needed for each zone for computing the melt under

warmer climate. Following the procedure outlined by

Rango and Martinec (1994), new depletion curves for

all zones were obtained for the warmer temperature

scenarios (T þ 1; T þ 2; T þ 3 8C) by relating SCA to

the cumulative snowmelt corresponding to the

temperature scenario. The model is used to compute

daily melt for each zone using SCA of current climate

as input and then cumulative melt for each zone is

determined. A relationship between SCA and cumu-

lative melt is established for each zone. Then, model

is run for warmer climatic scenario and cumulative

melt is computed for each climatic scenario for each

zone. New daily SCA for each zone for a temperature

scenario is generated using cumulative melt corre-

sponding to that temperature and established relation-

ship between cumulative melt and the SCA for that

zone. Rango and Martinec (1994) found that if

precipitation during the snowmelt period in the

current climate does not include snowfall, changes

in precipitation do not affect the snow cover. More-

over, they found that the effect of increase in snow

precipitation by 10–20% during the summer period

on snow cover depletion was insignificant as com-

pared to the effect of temperature. There is almost no

snowfall during the melting season in the study area.

Therefore, the effect of fresh snow during summer

period was not included in preparing the new

depletion curves.

8. Effect of warmer climate on evaporation

For each type of basin, the changes in annual

evaporation due to warmer climate were estimated

separately for all the 6 years. The average annual

change were computed and presented in this paper.

Because of large difference in the area of rainfed,

snowfed and glacierfed basin, impact assessment was

made for the unit area and compared. Fig. 6(a) shows

a comparison of the annual evaporation occurred in

different types of basins under current temperature

ðT þ 0 8CÞ condition and warmer climate ðT þ 2 8CÞ

scenario. As expected, evaporation increased under

warmer climate scenarios for both rainfed and

snowfed basins. Evaporation losses from the glacier-

ized basin were considered negligible both for current

and warmer climate scenarios, because the vapour

pressure gradient is very small (Bengtsson, 1980).

Fig. 7(a) illustrates that evaporation increased

linearly with an increase in temperature both for

rainfed and snowfed basins. The changes in evapor-

ation from the snowfed basin were found to be larger

than for the rainfed basin. For an increase in

temperature from 1 to 3 8C for the rainfed basin,

Table 1

Comparison of mean monthly observed and simulated discharge for calibration and validation periods

Months Calibration period (1985/86–87/88) Validation period (1988/89–90/91)

Observed mean (mm) Simulated mean (mm) Diff. (%) Observed mean (mm) Simulated mean (mm) Diff. (%)

November 15.5 16.4 5.8 16.9 17.8 5.3

December 13.6 14.5 6.6 16.4 16.0 22.4

January 12.0 12.6 5.0 15.3 14.7 23.9

February 10.1 9.8 23.0 12.5 12.0 24.0

March 12.7 13.6 7.0 17.2 16.0 27.0

April 19.5 18.4 25.6 19.6 18.0 28.1

May 44.2 46.2 4.5 50.3 49.7 21.2

June 82.4 82.1 20.4 94.0 92.3 21.8

July 133.4 140.1 5.0 129.8 131.0 0.9

August 129.3 127.5 21.4 108.0 115.5 6.9

September 66.6 66.4 20.3 64.8 69.0 6.5

October 27.0 29.4 8.8 23.7 23.2 22.1

Yearly 566.3 577.0 1.9 568.5 575.2 1.2

P. Singh, L. Bengtsson / Journal of Hydrology 300 (2005) 140–154150

the annual evaporation increased from 6 to 18%,

while for the snowfed basin corresponding increase

was 13–35%. The increase in evaporation for the

snowfed basin was about two times higher than that of

the rainfed basin. Such a high increase in evaporation

for the snowfed basin is possible primarily because of

two reasons. First, the warmer climate increases

evaporation rate and secondly, the extent of SFA

wherefrom evaporation takes place increases with

time on account of faster disappearance of snow. So,

for the snowfed basin, both increase in temperature

and increase in SFA exert a compound effect and

impact of warmer climate on evaporation for such a

basin becomes relatively larger than the rainfed basin,

resulting in much more soil moisture deficit causing

reduction in runoff. It is to be pointed out that in case

of rainfed basin, the area exposed to evaporation does

not change with time and only increase in temperature

influences the evaporation. The complex type of basin

is the sum of different constitutions. In the present

case, complex basin covered rainfed, snowfed and

glacierfed basins all together and integrated effect of

warmer climate on each type of basin represented the

net effect on the complex basin. For the complex

basin, the evaporation increased by 8–25% for

increase in temperature from 1 to 3 8C.

9. Effect of warmer climate on melt

The average changes in annual melt under warmer

climate ðT þ 2 8CÞ for snowfed, glacierfed and

complex basins are shown in Fig. 6(b). Under a

warmer climate, the melt from snowfed basin

Fig. 6. Impact of warmer climate (T þ 2 8C scenario) on (a) annual evaporation, and (b) annual melt for different types of basins in the

Himalayan region.

P. Singh, L. Bengtsson / Journal of Hydrology 300 (2005) 140–154 151

reduced, whereas from the glacierfed basin it

increased. Reduction in melt from snowfed basin

under warmer climate is possible due to availability of

lesser amount of snow in the basin. Results indicate

that the effect of warmer climate on the melt from

snowfed basin was found to be opposite as compared

to the glacierfed basin. Like evaporation, variations in

annual melt, either decrease from the snowfed basin

or increase from the glacierfed basin, varied linearly

with increase in temperature (Fig. 7(b)). For the

considered range of temperature increase (1–3 8C),

the melt from the glaciated basin increased by

16–50%, for the snowfed basin snowmelt decreased

by 11–23% for the same increase in temperature.

Thus, the magnitude of changes caused by warmer

climate was much higher than the snowfed basin. It is

important to note that for a complex basin, the

decrease in snowmelt is counterbalanced by increased

melt from the glaciers. For example, in the case of

present study, total melt reduced by about 2%, when

the whole basin was taken into account. However, in

the long-term perspective, glaciers may retreat

considerably and their area may reduce, or may

even disappear from the basin. The process is,

however, beyond the scope of the present study.

How the extent of glaciated area will reduce for the

Himalayan region is very uncertain and needs more

investigations.

10. Conclusions

Runoff generating processes differ from lower

altitude basins to middle, or high altitude basins in the

Himalayan region because of difference in their

sources of runoff and climatic conditions. For the

rainfed basins evaporation losses and rainfall–runoff

are the important hydrological processes for govern-

ing the outflow. In case of snowfed basins, melting

rate of seasonal snow and conversion of SCA to SFA

becomes more important than evaporation and rainfall

runoff. For glaciated basins, melt from glacier

dominates the runoff processes. In the present study,

attempts have been made to investigate the influence

of warmer climate on evaporation and melt for

different types of basins. It is found that the

evaporation from snowfed basins is more sensitive

to warmer climate as compared to rainfed and

glacierfed basins. For an increase in temperature

from 1 to 3 8C for the rainfed basin, the increase in

annual evaporation was computed to be about 6–18%

for the rainfed basin and 13–35% for the snowfed

basin. In addition to warmer temperature, increase in

SFA due to warmer climate also becomes an

important factor for influencing the evaporation for

the snowfed basin. Such impact will generate higher

soil moisture deficit in the snowfed basins attributing

to the reduction in runoff.

Under a warmer climate, the melting from the

snowfed basin was reduced while, in contrast, it

increased from the glacierfed basin. For the con-

sidered temperature scenarios (1–3 8C), the reduction

in snowmelt was observed to be 11–23% for the

snowfed basin while for the glaciated basin melt

increased from 16 to 50%. The magnitude of increase

in melting from the glaciated basin was much more

than the reduction in the seasonal snowmelt.

Fig. 7. Changes in (a) annual evaporation, and (b) annual melt with

increase in temperature for different types of basins in the

Himalayan region.

P. Singh, L. Bengtsson / Journal of Hydrology 300 (2005) 140–154152

The results presented in this paper are not the

predictions, but rather model simulations of the

plausible changes in evaporation and melt. These

results are useful guides for studies of implications on

society and adaptive requirements in land use and

water management practice. In future, land use and

cropping pattern have to be made keeping in view the

projected water stressed conditions, specially, for the

snowfed basins. Moreover, under warmer climatic

conditions, water managers have to plan for storing

more water during summer when the demands are at

their peak. There is need for assessing the effect of

climate change on the seasonal scale for different

types of basins located in complex geographic and

climatic regions, but also specifically to investigate

potential impacts of projected climate warming on the

hydrology and water resources in the mountain

regions of the world. Such studies show a direct

impact of climate change on the different components

of mountain hydrology.

Acknowledgements

This study was carried out during stay of Dr Pratap

Singh in Sweden for which financial support was

provided by the Swedish Institute. The support

obtained from Swedish Institute is thankfully

acknowledged.

References

Bardossy, A., Van Mierlo, J.M.C., 2000. Regional precipitation and

temperature scenarios for climate change. Hydrological

Sciences Journal 45, 559–575.

Bengtsson, L., 1980. Evaporation from a snow cover. Nordic

Hydrology 11, 221–234.

Burn, D.H., 1994. Hydrologic effects of climate change in west-

central Canada. Journal of Hydrology 160, 53–70.

Cayan, D.R., Riddle, L.G., Aguado, E., 1993. The influence of

temperature and precipitation on seasonal streamflow in

California. Water Resources Research 29, 1127–1140.

Chiew, F.H.S., Whetton, P.H., McMahon, T.A., Pittock, A.B., 1995.

Simulation of the impacts of climate change on runoff and soil

moisture in Australian catchments. Journal of Hydrology 167,

121–147.

Cohen, S.J., 1986. Impact of CO2-induced climatic change on water

resources in the Great Lakes basin. Climatic Change 8,

135–153.

Cooley, K.R., 1990. Effects of CO2-induced climatic changes on

snowpack and streamflow. Hydrological Sciences Journal 35,

511–522.

Divya, Mehrotra, R., 1995. Climate change and hydrology with

emphasis on the Indian sub-continent. Hydrological Sciences

Journal 40, 231–242.

Dyurgerov, M., 2002. Glacier mass balance and regime: data of

measurements and analysis. Occasional Paper No. 55. Institute

of Arctic and Alpine Research, University of Colorado,

Boulder, CO.

Earth Resources and Data Analysis System (ERDAS IMAGINE,

Ver. 8.3), 1997. ERDAS, Inc., 2801 Buford Highway, NE, Suite

300, Atlanta, GA 30329-2137, USA.

Epstein, D., Ramirez, J.A., 1994. Spatial disaggregation for studies

of climatic hydrologic sensitivity. Journal of Hydrological

Engineeering 120, 1449–1467.

Gleick, P.H., 1987. Regional hydrological consequences of

increases in atmospheric CO2 and other trace gases. Climate

Change 10, 137–161.

Haeberli, W., Beniston, M., 1998. Climate change and its impact on

glaciers and permafrost in the Alps. AMBIO 27, 258–265.

Hingane, L.S., Rup Kumar, K., Ramanamurthy, B.V., 1985. Long-

term needs of surface air temperature in India. Journal of

Climatology 5, 521–528.

IPCC, 1990. In: Houghten, J.T., Jenkins, G.J., Ephraums, J.J. (Eds.),

Report: Climate change—The IPCC Scientific Assessment.

Cambridge University Press, Cambridge, UK, p. 365.

IPCC, 2001a. In: Houghton, J.T., Ding, Y., Griggs, D.J., Noguer,

M., van der Linden, P.J., Dai, X., Maskel, K., Johnson, C.A.

(Eds.), Report: Climate Change—The Scientific Basis..

IPCC, 2001b. In: McCarthy, J.J., Canziani, O.F., Leary, N.A.,

Dokken, D.J., White, K.S. (Eds.), Report: Climate Change—

Impacts, Adaptations and Vulnerability..

Ives, J.D., Messerli, B., 1989. The Himalayan Dilemma: Reconcil-

ing Development and Conservation. Routledge, London and

New York.

Jain, S.K., 2001. Modelling of Streamflow and Sediment Studies in

the Satluj Basin using Remote Sensing and GIS. Indian Institute

of Technology, Roorkee, India.

Jones, J.A.A., 1999. Climate change and sustainable water

resources: placing the threat to global warming perspective.

Hydrological Sciences Journal 44, 541–557.

Lal, M., 2001. Climatic change—implications for India’s water

resources. Journal of Indian Water Resources Society 21,

101–119.

Lal, M., Singh, S.K., 2001. Global warming and monsoon climate.

Mausam 52, 245–262.

Lal, M., Cubasch, U., Santer, B.D., 1992. Potential changes in

monsoon climate associated with global warming as inferred

from coupled ocean–atmosphere general circulation model.

CAS/JSC Working Group Report No. 17, WMO/TD 467, 66–99.

Letreguilly, A., Reynaud, L., 1990. Space and time distribution of

glacier mass balance in the northern hemisphere. Arctic and

Alpine Research 22, 43–50.

Lettenmaier, D.P., Gan, T.Y., 1990. Hydrologic sensitivities of the

Sacramento–San Joaquin river basin, California, to Global

warming. Water Resources Research 26, 69–86.

P. Singh, L. Bengtsson / Journal of Hydrology 300 (2005) 140–154 153

Loaiciga, H.A., Valdes, J.B., Vogel, R., Garvery, J., Schwarz, H.,

1996. Global warming and the hydrological cycle. Journal of

Hydrology 174, 83–127.

McCabe, G.J. Jr., 1994. Relationship between atmospheric circula-

tion and snowpack in the Gunnison river basin, Colorado.

Journal of Hydrology 157, 157–176.

McCabe, G.J. Jr., Ayers, M., 1989. Hydrologic effect of climate

change in the Delaware river basin. Water Resources Bulletin

25, 1231–1242.

Nash, L.L., Gleick, P.H., 1991. Sensitivity of streamflow in the

Colorodo basin to climatic changes. Journal of Hydrology 125,

221–241.

Nash, J.E., Sutcliffe, J.V., 1970. River forecasting through

conceptual models. 1. A discussion of principles. Journal of

Hydrology 10, 282–290.

Nemec, J., 1989. Implications of Changing Atmosphere on Water

Resources. Proceedings of the WMO Conference, June 27–30,

Toronto, Canada, WMO-710, pp. 211–223.

Nemec, J., Schaake, J., 1982. Sensitivity of water resource systems

to climate variation. Hydrological Sciences Journal 27,

327–343.

Ng, H.Y.F., Marsalek, J., 1992. Sensitivity of streamflow simulation

to changes in climatic inputs. Nordic Hydrology 23, 257–272.

Oerlemans, J., Anderson, B., Hubbard, A., Huybrechts, Ph.,

Johannesson, T., Knap, W.H., Schmeits, M., Stroeven, A.P.,

Wallinga, R.S.W., van de Wal, J., Zuo, Z., 1998. Modelling the

response of glaciers to climate warming. Climate Dynamics 14,

267–274.

Panagoulia, D., 1991. Hydrological response of a medium sized

mountainous catchment to climate changes. Hydrological

Sciences Journal 26, 525–547.

Quick, M.C., Pipes, A., 1977. UBC watershed model. Hydrological

Sciences Bulletin 22, 161–253.

Rango, A., 1992. Worldwide testing of the snowmelt runoff model

with applications for predicting the effects of climate change.

Nordic Hydrology 23, 155–172.

Rango, A., Martinec, J., 1994. Areal extent of seasonal snow cover

in a changed climate. Nordic Hydrology 25, 233–246.

Sanderson, M., Smith, J., 1990. Climate Change and Water in the

Grand River, Ontario. Proceedings of 43rd Annual Conference

of the Canadian Water Resources Association, May 16–18,

Pentiction, British, Columbia, pp. 243–261.

Sausen, R., Schubert, S., Dumenil, L., 1994. A model of river runoff

for use in coupled atmosphere–ocean model. Journal of

Hydrology 155, 337–352.

Singh, P., 1991. A temperature lapse rate study in Western

Himalayas. Hydrology Journal (Indian Association of Hydrol-

ogists) 14, 156–163.

Singh, P., Jain, S.K., 2002. Snow and glacier contribution in the

Satluj river at Bhakra Dam in the Western Himalayan region.

Hydrological Sciences Journal 47, 93–106.

Singh, P., Jain, S.K., 2003. Modelling of streamflow and its

components for a large Himalayan basin with predominant

snowmelt yields. Hydrological Sciences Journal 48, 257–275.

Singh, P., Jain, S.K., Kumar, N., 1997a. Snow and glacier melt

runoff contribution in the Chenab river at Akhnoor. Mountain

Research Development 17, 49–56.

Singh, P., Kumar, N., 1997a. Impact of climate change on the

hydrological response of a snow and glacier melt runoff

dominated Himalayan River. Journal of Hydrology 193,

316–350.

Singh, P., Kumar, N., 1997b. Effect of orography on precipitation in

the western Himalayan region. Journal of Hydrology 199,

183–206.

Singh, P., Singh, V.P., 2001. Snow and Glacier Hydrology. Kluwer

Academic Publishers, The Netherlands.

Singh, P., Spitzbart, G., Huebl, H., Weinmeister, H.W., 1997b.

Hydrological response of snowpack under rain-on-snow event: a

field study. Journal of Hydrology 202, 1–20.

Thomsen, R., 1990. Effect of climate variability and change in

ground water in Europe. Nordic Hydrology 21, 185–194.

WMO, UNEP, UNESCO, IOC, FAO, ICSW, 1991. Climate

Change: Science, Impacts and Policy. In: Jager, J., Ferguson,

H.L. (Eds.), Proceedings of the Second World Climate

Conference (October 1990). Cambridge University Press,

Cambridge, UK, p. 578.

P. Singh, L. Bengtsson / Journal of Hydrology 300 (2005) 140–154154