Science of the Total Environment 505 (2015) 748761

Contents lists available at ScienceDirect

Science of the Total Environment

journal homepage: www.elsevier.com/locate/scitotenv

Snow cover trend and hydrological characteristics of the Astore River

basin (Western Himalayas) and its comparison to the Hunza basin
(Karakoram region)
Adnan Ahmad Tahir a,b,c,, Pierre Chevallier a, Yves Arnaud d, Muhammad Ashraf e, Muhammad Tousif Bhatti f
a

Laboratoire Hydrosciences UMR 5569 (CNRS, IRD, Montpellier University 1&2), CC57, Universit Montpellier 2 Sciences et Techniques, Place Eugne Bataillon, 34095 Montpellier Cedex 5,
France
b
Department of Environmental Sciences, COMSATS Institute of Information Technology, 22060 Abbottabad, Pakistan
c
Institute of Earth Surface Dynamics, Faculty of Geosciences and Environment, University of Lausanne, Switzerland
d
Laboratoire des Transferts en Hydrologie et Environnement UMR 5564 (CNRS, IRD), Universit Joseph Fourrier, Grenoble INP), LGGE, 54 rue Molire, Domaine Universitaire, BP 96, 38402 Saint
Martin d'Heres Cedex, France
e
Centre of Excellence in Water Resources Engineering (CEWRE), University of Engineering & Technology (UET), Lahore, Pakistan
f
International Water Management Institute, 12 km Multan Road, Chowk Thokar Niaz Baig, 53700 Lahore, Pakistan

H I G H L I G H T S
Constant trend in snow cover area in Astore basin during last decade.
Increasing trend of river discharge during last 34-years in the Astore River basin.
Increasing winter precipitation and decreasing summer temperature at Astore.

a r t i c l e

i n f o

Article history:
Received 15 May 2014
Received in revised form 13 October 2014
Accepted 19 October 2014
Available online 5 November 2014
Editor: Simon Pollard
Keywords:
Spatio-temporal snow cover trend
Upper Indus River Basin
Central Karakoram & Western Himalayas
Hydrological regime
Climate trend
Water resources management

a b s t r a c t
A large proportion of Pakistan's irrigation water supply is taken from the Upper Indus River Basin (UIB) in the
HimalayaKarakoramHindukush range. More than half of the annual ow in the UIB is contributed by ve of
its snow and glacier-fed sub-basins including the Astore (Western Himalaya south latitude of the UIB) and
Hunza (Central Karakoram north latitude of the UIB) River basins. Studying the snow cover, its spatiotemporal change and the hydrological response of these sub-basins is important so as to better manage water resources. This paper compares new data from the Astore River basin (mean catchment elevation, 4100 m above
sea level; m asl afterwards), obtained using MODIS satellite snow cover images, with data from a previouslystudied high-altitude basin, the Hunza (mean catchment elevation, 4650 m asl). The hydrological regime of
this sub-catchment was analyzed using the hydrological and climate data available at different altitudes from
the basin area. The results suggest that the UIB is a region undergoing a stable or slightly increasing trend of
snow cover in the southern (Western Himalayas) and northern (Central Karakoram) parts. Discharge from the
UIB is a combination of snow and glacier melt with rainfall-runoff at southern part, but snow and glacier melt
are dominant at the northern part of the catchment. Similar snow cover trends (stable or slightly increasing)
but different river ow trends (increasing in Astore and decreasing in Hunza) suggest a sub-catchment level
study of the UIB to understand thoroughly its hydrological behavior for better ood forecasting and water resources management.
2014 Elsevier B.V. All rights reserved.

1. Introduction and background

Corresponding author at: Institute of Earth Surface Dynamics, Faculty of Geosciences

and Environment, University of Lausanne, Switzerland. Tel.: +41 78 814 18 37.
E-mail addresses: uaf_adnan@hotmail.fr (A.A. Tahir), pierre.chevallier@ird.fr
(P. Chevallier), yves.arnaud@ird.fr (Y. Arnaud), ashraf_uaf@hotmail.com (M. Ashraf),
tousif.bhatti@cgiar.org (M.T. Bhatti).

http://dx.doi.org/10.1016/j.scitotenv.2014.10.065
0048-9697/ 2014 Elsevier B.V. All rights reserved.

The agriculture-based economy of Pakistan is dependent on irrigation waters supplied by the Indus River and its tributaries (SIHP,
1990). Most of the ow abstracted from the Indus River at Tarbela
(Fig. 1a) is contributed by snow and glacier melt (Bookhagen and
Burbank, 2010; Immerzeel et al., 2012, 2013) of the Karakoram,
Himalaya and Hindukush mountains. The Indus catchment area

A.A. Tahir et al. / Science of the Total Environment 505 (2015) 748761

749

Fig. 1. (a) Map of Pakistan including the Upper Indus River Basin (UIB) area, its sub-basins, Tarbela dam and study sites (Astore and Hunza sub-basins with their Global Digital Elevation
Models GDEM) and (b) mean monthly recorded precipitation totals (19742010) and mean monthly runoff (19742008) of both study sites. Precipitation records are the average totals of
all three climate stations (Astore, Rama and Rattu CS for Astore basin; Khunjerab, Naltar and Ziarat CS for Hunza basin) situated in the both catchments.

upstream of Tarbela, Upper Indus River Basin (UIB) (Fig. 1a), is the main
contributor to the Indus River. Tarbela reservoir was constructed in
1974 on the Indus River to supply irrigation water to the Indus Irrigation
System through a large network of canals, and so to supply the agricultural lands of Punjab and Sindh (Pakistan) (Fig. 1a), the main producers
of agricultural products in the country. The mean annual inow to
Tarbela is 2410 m3s1 309 m3s 1, as estimated from the 40year (19692008) SWHP (Surface Water Hydrology Project) ow
record.

The Upper Indus River Basin has a total catchment area of

N200,000 km2 (Bookhagen and Burbank, 2010; Forsythe et al., 2012b;
Immerzeel et al., 2009; Tahir et al., 2011a) and almost 12% of the total
area ( 22,000 km2) is covered by glacier ice (Hewitt, 2001; Hewitt,
2007). There are many discrepancies in the estimated catchment area
reported by Khan et al. (2014) that may be a result of different datasets,
methodologies and assumptions used to estimate this area by different
authors. Within the UIB, ve catchments, the Shyok, Shigar, Astore,
Gilgit and Hunza rivers (Fig. 1a), contribute more than 60% of the total

750

A.A. Tahir et al. / Science of the Total Environment 505 (2015) 748761

Upper Indus ow (Mukhopadhyay and Khan, 2014a; Tahir et al., 2011a)

and the management of Tarbela reservoir depends to a large extent on
this inow. The tributaries are principally fed by snow and glacier
melt; the effect of the monsoon rainfall system is very limited because
of the Western Himalaya mountain barrier.
The active hydrological zone for the Upper Indus River Basin is situated in the high-altitude Karakoram ranges. Liniger et al. (1998) noted
that more than 80% of the ow of the Indus River System into the plains
originates in the Hindukush, Karakoram and Western Himalaya mountain areas. Various authors (Archer, 2003; Hewitt et al., 1989;
Mukhopadhyay and Khan, 2014a; Wake, 1989; Young and Hewitt,
1990) reported that more than 70% of the ow of the Upper Indus
River comes from the zones of heavy snowfall and glacierized basins
above 4000 m in altitude. Archer (2003) found a strong correlation between summer runoff and summer mean temperature in the highelevation tributaries of the UIB, fed by snow and glacier melt. Archer
(2003) and Forsythe et al. (2012b) have indicated that a single degree
rise in mean summer temperature might result in a 1020% increase
in summer runoff into the Hunza River. Tahir et al. (2011b) found that a
1 C rise in mean annual temperature (by keeping the other climate variables constant) may force the mean annual discharge in the Hunza
River to increase by 32% in next few years. If this scenario holds, the
glaciers may retreat, generating a high volume of water over a few decades but then leading to longer-term water shortages.
The UIB includes a part of the Hindukush, Karakoram and Himalayan ranges (Western Himalayas) and so has a large climate variation
within the basin area. The UIB receives maximum annual precipitation in winter and spring originating from the west (Dimri and
Chevuturi, 2014; Dimri et al., 2013; Young and Hewitt, 1990). The
climatic records in the UIB are different to the Eastern Himalayas because the high mountains of HindukushKarakoramHimalaya
(HKH) decrease the effect of the monsoon in northern parts of the
UIB (Fowler and Archer, 2005; Young and Hewitt, 1990). Besides
this southnorth gradient, there is a strong eastwest gradient in
monsoonal strength as it migrates along the Himalayan front
(Bookhagen and Burbank, 2006; Bookhagen and Burbank, 2010).
Two precipitation types inuence the UIB discharge: (i) westerly circulations (which produce snow) which dominantly inuence the
discharge in the northwest (Central and West Karakoram) and
southwest (Western Himalayas) (Dimri et al., 2013) of UIB and (ii)
winter rainfall at lower elevations which inuences discharge in
the southwest part of UIB. Understanding the spatio-temporal patterns of snow cover in these regions is needed to better understand
the hydrology of the UIB.
These regions, generating the majority of the UIB ow, on the one
hand have no or scarce climate data; particularly above 4000 m. On
the other hand, there is a lack of detailed investigation of snow and ice
processes and their relevance to the hydrological regime in the highaltitude areas of the UIB except for a few studies (e.g. Akhtar et al.,
2008; Bookhagen and Burbank, 2010; Immerzeel et al., 2012). Satellite
remote-sensing techniques may be the only way to analyze the spatial
variation of snow cover in these remote regions of the UIB (Tahir
et al., 2011a; Sirguey et al., 2009). Many researchers (Immerzeel et al.,
2009; Lee et al., 2005; Maurer et al., 2003) in different regions of the
world have used MODIS snow products to estimate the spatial variation
of snow and for snowmelt runoff modeling. Tahir et al. (2011a) and
Tekeli et al. (2005) compared the coarse-resolution (500 m 500 m)
MODIS data with ne-resolution (30 m 30 m) ASTER data and ground
observations and reported that MODIS estimated the snow cover with
75%, 95% and 99% accuracy at lower, middle and higher altitudes, respectively, when compared with ASTER and 62%82% accuracy when
compared to ground observations. The ndings of Forsythe et al.
(2012a) support the use of MODIS snow and temperature spatial data
products as analogues for the climate drivers that lead the runoff generation in various UIB sub-catchments. The validity of MODIS encourages
the use of these data for the estimation of spatial changes in snow cover

at varying altitudes and different climate regions of the Upper Indus

River Basin.
The relief of the catchment (455 m at Tarbela, 8611 m at K2) (Fig. 1a)
makes it difcult to study the snow cover spatial variation and hydrological regime using a global (basin-wide) approach. It is therefore necessary to understand how different mountain ranges within the UIB
(Fig. 1a) respond to climatic variability to allow better ood forecasting
and water resources management with regard to irrigation demands,
especially for the agricultural lands downstream of Tarbela. This study
is different from the previous studies on UIB (e.g. Archer and Fowler,
2004; Forsythe et al., 2012b; Fowler and Archer, 2006; Hasson et al.,
2013) in the sense that we focus upon the spatio-temporal trends of
snow cover in three different altitudinal zones (contrary to the basinwide approach adopted by these previously mentioned authors) within
a southern sub-catchment of UIB (Astore; Fig. 1a). These trends and the
hydroclimatic characteristics of this sub-catchment (Astore) are then
compared to a northern sub-catchment (Hunza; Fig. 1a). These two
sub-catchments, Astore and Hunza, are selected because these are dominated by two types of hydrological regimes (nival and glacial regime,
respectively; Archer, 2003).
The main scientic goals addressed in this study are:
a) to estimate the spatio-temporal trends in snow cover in a Western
Himalayan River basin close to Central Karakoram;
b) to analyze the annual and seasonal trends of the hydrometeorological variables (river runoff, precipitation and mean temperature) in
the study area;
c) to establish the annual and seasonal correlation between snow cover
area, mean temperature, precipitation and stream ow in the catchment area (hydrological behavior of the catchment); and
d) to compare spatio-temporal trends in snow cover and hydrological
processes in a Western Himalayan catchment and a previously studied high-altitude Central Karakoram catchment (Hunza River basin).
2. Study area
The snow-fed sub-catchment of the Astore (Fig. 2) (sub-basin of
UIB) was selected for the analysis of snow cover trends and hydrological
regime. These characteristics of the Astore basin were then compared to
a previously studied UIB sub-catchment (the Hunza River basin)
(Fig. 1a). The Astore basin was selected because it has a contrasting geographical location (southern foothills of the Western Himalayas)
(Fig. 1a) to the Hunza. These two sub-catchments also differ from
each other on the basis of known features, for example, the Astore
basin has a north-facing orientation, snow-fed regime, lower latitude
and mid-altitude as compared to south orientation, glacier-fed regime,
high altitude and latitude of the Hunza basin. The Hunza and Astore
sub-catchments are inuenced by the same climate regime (Westerlies)
but in slightly different ways (Fig. 1b), despite their proximity. The
Hunza River discharge is dependent on the westerly circulations that
bring maximum precipitation in winter in the form of snow (which
melts in summer to generate high discharges), whereas the Astore
River discharge is inuenced by the winter rainfall (at lower elevations)
which combines with the winter solid precipitation forced by
Westerlies.
The main differentiating features of the Astore and Hunza River basins are summarized in Tables 1 and 2. The Astore River basin (drainage
area, 3990 km2) (Fig. 2) is situated in the mid-altitude extreme Western Himalaya region, with a mean catchment elevation of 4100 m asl.
The Astore River basin's drainage area was estimated using data from an
Advanced Spaceborne Thermal Emission and Reection Radiometer
(ASTER-version 2) processed through a Global Digital Elevation Model
(GDEM) at a 1 arc-second resolution (approx. 30 m). The hypsometric
curve and the percentage area under each 500-m altitudinal layer estimated from the ASTER GDEM for the Astore and Hunza River basins
are presented in Fig. 3. The Astore basin has a lower mean altitude

A.A. Tahir et al. / Science of the Total Environment 505 (2015) 748761

751

Fig. 2. Global Digital Elevation Model (GDEM) of the Astore River basin showing three altitudinal zones, glacier coverage across them (estimated from RGI), and superposed by the gauging
stations.

(4100 m asl) than the Hunza basin (4650 m asl). Only 5% area of
the Astore basin is above 5000 m asl, while this percentage is 33% in
the Hunza basin (Fig. 3 & Table 1).
Three climate stations (Astore, Rama and Rattu) with precipitation
gauges are installed at different altitudes within the catchment area of
the Astore River (Figs. 2 and 3). The mean total annual recorded precipitation is 794 mm at Rama (3179 m asl), 723 mm at Rattu
(2718 m asl) and 501 mm at Astore (2168 m asl) according to
the available data records of these climate stations. The Astore River has
a mean annual ow of 137 m3s1 (i.e., 1084 mm of water depth
equivalent) gauged at Doyian hydrometric station ( 1583 m asl)
(Fig. 2).
The present precipitation records in the Astore catchment are not
representative of the runoff at the outlet because of the lack of climate
data records at altitudes above ~ 3200 m asl and underestimation of

winter precipitation (snow) in addition to the well-known gauging errors (wind-induced, wetting loss, trace precipitation, blowing and
drifting snow, and systematic mechanical errors, etc.) in highmountain environments described by many authors, e.g. Sevruk
(1985), Sevruk (1989) and Frland et al. (1996). Another reason for
this underestimation is that the spatially integrated precipitation value
is largely inuenced by orographic precipitation (a function of elevation
and terrain characteristics) whereas the value recorded at a meteorological station is simply a point measurement.
3. Materials and methods
Preprocessed satellite data products (ASTER GDEM V2 and MODIS
snow cover data) and ground observation data (stream ow, temperature and precipitation) sources and their treatment for the Hunza

Table 1
Key features of the study areas (Astore and Hunza River basins). Altitudes of the high-altitude climate stations are rechecked by the WAPDA and SIHP (Snow and Ice Hydrology Project)
and corrected altitudes are given in the table (these altitudes are different than published previously by Archer, 2003 and Tahir et al., 2011a). Moreover the glacier cover area for Astore and
Hunza basins is estimated using latest RGI (version 3.2) datasets and it differs than presented previously by Tahir et al. (2011a).
Catchment

Astore

Hunza

River ow gauging station

Latitude
Longitude
Elevation of gauging station
Drainage area
Glacier-covered area
Glacier cover percentage
Mean elevation (computed from hypsometric curve)
Median elevation (computed from GDEM)
Area above 5000 m
No. of meteorological stations
Name and elevation (m asl) of meteorological stations

Doyian
35 33 N
74 42 E
1583 m asl
3990 km2
~543 km2
~13.6%
~4100 m asl
~4594 m asl
(~200 km2) ~5%
3 (1 installed by PMD & 2 installed by WAPDA)
Astore 2168 m
Rama 3179 m
Rattu 2718 m

Dainyor bridge
35 56 N
74 23 E
1450 m asl
13,733 km2
~3840 km2
~28%
~4650 m asl
~4631 m asl
(~4463 km2) ~32.5% (Tahir et al., 2011a)
3 (installed by WAPDA)
Khunjerab 4440 m
Ziarat 3020 m
Naltar 2898 m

752

A.A. Tahir et al. / Science of the Total Environment 505 (2015) 748761

Table 2
Main features of the elevation zones extracted from the ASTER GDEM of the Astore River basin (in black) and Hunza River basin (in dark gray) and their corresponding climate station.
Zone

A
(Low-altitude zone)

Basin
Elevation band (m asl)
Mean elevation (m asl)
Median elevation (m asl)
Area (%)
Area (km2)
Climate stations

Astore
12133300
~2950
~2263
16
638
Astore, Rama, Rattu

B
(Mid-altitude zone)
Hunza
14323300
~2850
~2366
11
1541
Naltar, Ziarat

Astore
33014300
~3910
~3800
50
1995

C
(High-altitude zone)
Hunza
33014300
~3850
~3800
25
3413

Astore
43018069
~4600
~6132
34
1357

Hunza
43017849
~5000
~6067
64
8779
Khunjerab

Elevation band area and mean elevation for each zone were computed from the hypsometric curve that is estimated from ASTER Global Digital Elevation Model (GDEM).

River basin are explained in detail in two previously published studies

by Tahir et al. (2011a) and Tahir et al. (2011b). The same data sets are
used in this study. A brief description of these data sets is given below.

hypsometric curves and median elevations, were also estimated with

the ASTER-GDEM (Tables 1 and 2; Fig. 3). The hypsometric curves
were then used to calculate the mean elevations of the river basins
and their respective altitudinal zones (Tables 1 and 2).

3.1. Topography
3.2. Hydrometeorology
The Advanced Spaceborne Thermal Emission and Reection Radiometer (ASTER), Global Digital Elevation Model version 2 (GDEM V2)
was used to delineate the river basin boundary. The coordinates of outlet points of Astore and Hunza rivers for the basin boundary delineation
provided by WAPDA were slightly re-adjusted during the process so
that these points completely overlap the stream line. The coordinates
given by WAPDA for Astore basin were 74 42 15.12 (longitude) and
35 32 42 (latitude) and these were re-adjusted to 74 39 7.62 (longitude) and 35 34 20.7 (latitude) during the watershed delineation
process in this study. The coordinates given by WAPDA for Hunza
basin were 74 22 54.17 (longitude) and 35 55 38.6 (latitude) and
these were re-adjusted to 74 22 15 (longitude) and 35 55 7 (latitude). The ASTER GDEM is a product of METI and NASA and is available
for high-latitude and steep mountainous areas not covered by SRTM3
(METI and NASA, 2011). Three different altitudinal zones were extracted from the GDEM of the study area (Fig. 2 and Table 2) for detailed
analysis of snow cover spatiotemporal change in the Astore and
Hunza River basin. The threshold values for the altitudinal zonal boundaries of Astore basin were kept similar as for the previously studied
Hunza basin so that the basin characteristics at the same elevations
can be compared. Total catchment and zonal areas, along with

The SWHP of the Water and Power Development Authority (SWHP

WAPDA) carries out stream ow measurements in Pakistan with the
earliest observations dating from 1960. A 34-year (19742007) database of daily discharges for the Astore River gauged at Doyian (Fig. 2)
was made available for this study. Meteorological data (daily mean temperatures and precipitation) available at two high-altitude Automatic
Weather Stations (AWS), Rama and Rattu (Fig. 2), was provided by
WAPDA and for a valley station, Astore (Fig. 2) by the PMD (Pakistan
Meteorological Department). The data (temperature, precipitation, relative humidity, solar radiation, snow water equivalent, wind speed and
direction) from WAPDA climate stations is available for 14 years
(19952008), whereas the climate data (temperature, rainfall, relative
humidity, wind speed and direction) from the station installed and
managed by the PMD (Archer, 2003) is available for over 50 years
(19542012). Tipping buckets are used by both the departments
WAPDA and PMD for the precipitation recording and the snow pillows
are used for snow water equivalent estimation by WAPDA.
Climate (seasonal precipitation and mean temperature) and stream
ow data were analyzed by applying the regression analysis to investigate the climate change trend in the Astore catchment area and to

Fig. 3. Hypsometric curves and the distribution of area under 500-m elevation bands for the Astore and Hunza River basins (estimated from ASTER-GDEM). The three altitudinal zones and
the location of the respective weather stations are presented on the right hand side y-axis.

A.A. Tahir et al. / Science of the Total Environment 505 (2015) 748761

compare these trends with Hunza basin climate trends. The trend analysis was performed on the long data series (34-year time series;
19742007, this period was selected because it is the common duration
with the river discharge data) available from the Astore climate station.
Trends on the standardized values of annual runoff (Astore and Hunza
rivers) and the total annual precipitation (recorded at the Astore CS)
were applied. The values were standardized using the normal deviate
formula, i.e., the distance of one data point from the mean, divided by
the standard deviation of the distribution. A seasonal analysis of the precipitation and temperature data was performed for the winter (December to February; DJF) and summer (July to September; JAS) periods at
Astore climate station. The nonparametric MannKendall (MK) trend
test (signicance level, 5%) for periodic data (Hirsch and Slack, 1984;
Hirsch et al., 1982; Kendall, 1975; Mann, 1945) and TheilSen estimator
(also known as Sen's slope) (Gilbert, 1987; Sen, 1968) were applied to
identify the trends in time series data; the values of the trend test and
slope are represented by Kendall's tau () coefcient and Sen's slope
(S), respectively. The MK test is widely adopted to assess signicant
trends in time series (Hirsch and Slack, 1984; Pellicciotti et al., 2010).
It is a nonparametric test, less sensitive to extreme sample values, and
independent from the hypotheses about the nature of the trend, either
linear or nonlinear. Sen's slope is a method for robust linear regression
that chooses the median slope among all lines through pairs of twodimensional sample points.
An analysis of the annual and seasonal relationship between different variables (total recorded precipitation, snow cover, mean temperature and discharge) on different seasons (winter and summer) was also
carried out. The Pearson product moment correlation (Rodgers and
Nicewander, 1988), Spearman rank order correlation (Spearman,
1904) and Kendall's rank correlation (Kendall, 1975; Kendall and
Gibbons, 1990) tests were performed (signicance level, 5%) to evaluate
the relationship between these variables at various gauging stations
mentioned above to investigate the main controlling factors of runoff
for the Astore River at Doyian.
3.3. Snow cover
The Moderate Resolution Imaging Spectroradiometer (MODIS)
snow products were selected to estimate the spatio-temporal trends
in snow cover. The MODIS/Terra Snow Cover 8-Day L3 Global 500-m
Grid (MOD10A2), used for this study, contains data elds for maximum
snow cover extent over an 8-day repeated period (Hall et al., 2006, updated weekly) and has a resolution of approximately 500 m (precisely,
15) completely covering the Astore River basin. The US National Snow
and Ice Data Center (http://nsidc.org/cgi-bin/snowi/search.pl) provided
a data set of 600 processed MOD10A2 (V005) images available from
March 2000 to December 2012. Daily data is also available from the
MOD10A1 dataset but we used 8-day MOD10A2 data so that to be coherent when comparing the trends with the case of previously studied
Hunza basin.
The Astore River basin area was then extracted to estimate the snow
cover (%) in the study area over a 13-year period. MODIS snow products
may also contain a signicant percentage of cloud covers that, if not
treated, may lead to possible uncertainties about actual amount of
snow cover. In our data set, if the cloud cover exceeded 15% on a specic
image date then we eliminated this image from our time series data. Almost 14% of total available images (i.e. 87 images) were discarded from
the data time series due to cloud cover obscurity. The snow cover on
that eliminated image date was then estimated by linearly interpolating
between the previous and the next cloud free snow cover images. The
snow cover was also estimated for the different elevation bands
(Table 2) to examine the spatiotemporal trend of snow cover in these
zones.
The entire data series was treated for seasonal adjustment prior to
trend analysis. Seasonality can be dened as a pattern in a time series
that repeats at regular intervals, in this case every year. Seasonal

753

uctuations in data make it difcult to analyze whether changes in

data for a given period reect substantial increases or decreases in the
magnitude of the data or are due to regularly occurring variation
(Mukhopadhyay and Khan, 2014b; Pellicciotti et al., 2010). Seasonal adjustment is intended to estimate the seasonal variation and to eliminate
its impact on a time series. For the trend analysis of our monthly snow
cover data with seasonality of 12 months, we did not nd out if there
is a trend in the overall series, but if from one month of January to another, and from one month of February to another, and so on. This procedure was repeated for all the months of all the years to remove
seasonality from the entire data series. Alternatively, this trend can be
detected by using the seasonal MannKendall (MK) trend test. This
test automatically analyzes if there is any autocorrelation in the time series and then removes this autocorrelation according to the procedure
described above before applying the trend test. After removing the seasonal climatology from the data time series, trends in snow cover timeseries were determined using Kendall's tau () and Sen's slope (S), as
described above.
All the trends and correlation values presented for the analysis of
hydro-meteorological and snow cover data in this study are tested at a
signicance level of 5%.
4. Results and discussion
4.1. Changes in snow cover in the Astore River basin (Western Himalaya
region)
The average percentage snow cover varies from 95% in winter to
7% in summer (Fig. 4) as estimated from MODIS snow images over
a 13-year period. The snow accumulation period starts in October and
the maximum snow cover reaches a range of 9095% in January
(Fig. 4). The snowmelt period starts in early April and the minimum
snow cover is observed during August and September when the snow
cover drops to a range of 510% (Fig. 4). Approximately 200 km2 of
the Astore catchment area is at an elevation above 5000 m asl (as calculated from the GDEM shown in Fig. 3) and 14% is glaciated, as extracted from the GDEM of Astore River basin (Fig. 2) by using the RGI-GLIMS
version 3.2 (Randolph Glacier Inventory-Global Land Ice Measurements
from Space) dataset (Arendt et al., 2012). The average change in snow
cover area for the entire data period for Astore and Hunza River basins
is presented in Fig. 4.
The temporal trend of monthly snow cover is estimated from a 13year period in three altitudinal zones of the Astore River basin and is
presented in Fig. 5. MK trend analysis shows a stable (tends to slight increase) trend (trends are statistically signicant at a signicance level of
5%) of the snow cover (after excluding seasonal climatology) in all altitudinal zones (Fig. 5) from 2000 to 2012. Kendall's tau () coefcient
values of + 0.066, + 0.088 and + 0.058 were found for the zones A, B
and C, respectively (Fig. 5). Sen's slope (S) values of snow cover trends
are also indicating a stable trend (tends to slight increase) for all the
zones with the values (%year1) of +0.132, +0.270 and +0.090 for
zones A, B and C, respectively (Fig. 5).
Zone A is covering ~ 16% (i.e. the lowest elevation class
12133300 m asl) of the total area of the Astore catchment (Fig. 2 and
Table 2) and the precipitation in this zone is mostly in the form of rainfall. Zone C covers ~ 34% (i.e. the highest elevation class
43018069 m asl) of the total area of Astore catchment and it is largely
covered with the glaciers especially above 5000 m asl (Fig. 2 and
Table 2). This may be the reason for weaker snow cover variation trends
( = +0.066 and +0.058; S = +0.132 and +0.090%year1, respectively) in these two zones and therefore these two may be less sensitive
towards overall variation of the snow cover in the catchment area. Zone
B covers ~50% (i.e. the middle elevation class 33014300 m asl) of the
total area of Astore catchment and receives a large proportion of the
precipitation in the form of snowfall. It may be therefore the zone B
of Astore River basin that is more sensitive towards the overall variation

754

A.A. Tahir et al. / Science of the Total Environment 505 (2015) 748761

Fig. 4. Annual cycle in snow cover area as percent of total area estimated through the analysis of 600 MODIS snow images (scatter plot) for 13 individual years (20002012) in the Astore
River basin and average snow cover area curves for both the Astore and Hunza basins.

of the catchment snow cover area. Zone B is also the proportion of the
catchment area where the snow cover trends are more signicant ( =
+ 0.088; S = + 0.270%year 1) and therefore it may inuence the
catchment snow cover trends signicantly.
Including seasonal climatology (without seasonal adjustment), a
similar stable trend (slight increase, though not signicant) was observed but with different values (i.e. = + 0.032, + 0.001 and
+0.035 for zones A, B and C, respectively). This result is similar to the
snow cover trend in the Hunza River basin (the trends in the Hunza
basin were found on a data series without seasonal adjustment)
(Tahir et al., 2011a), where almost a stable trend with = + 0.01 in
zone A and = + 0.03 in zone C was found. The snow cover trend
value differs slightly in zone B of the Astore River basin from the
Hunza River basin. A Kendall tau () coefcient value of = + 0.088
and Sen's slope value (%year 1) of + 0.270 was found in Zone B of
the Astore River basin (Fig. 5), while this trend value was = 0.01
in zone B of the Hunza River basin (Tahir et al., 2011a), but this value
does not have a signicant effect on the overall snow cover trend of the
Hunza basin since nearly 75% of the Hunza catchment area is not within
this elevation zone (Fig. 3). The stable (or slightly increasing) trends of
the snow cover in zones A, B and C of the Astore basin are the key factors
for overall consistent trend of the snow cover area in the catchment.
A global (basin-wide) approach to analyze the snow cover trend indicates a stable (or slightly increasing) trend in the Astore and Hunza
River basins with a Kendall tau () coefcient value of = +0.03 and
= +0.02, respectively (Table 3).
This increasing trend is observed both in the maximum snow (snow
accumulation) periods from December to February ( = + 0.06 for
Astore; +0.25 for Hunza) and in the minimum snow cover (snowmelt)
periods from July to September ( = +0.07 for the Astore; +0.01 for
Hunza) (Table 3). The trend analysis of climate variables (mean temperature and precipitation) over the same period is presented in Fig. 5,
which may help determine the factors for this consistent snow cover
trend. A Kendall tau () coefcient value of = +0.13 for precipitation

and = 0.179 for the mean temperature trend is found over the 13year period (20002012) (Fig. 5). The increasing trend of precipitation
and decreasing trend of mean temperatures over the same period
(20002012) indicate the possible reasons for the snow cover stability
trend. Rising precipitation continues to feed the high altitudes and
result in stable or slightly expanding snow cover, very similar to the
Hunza River basin. Seasonal trends of climate variables (seasonal precipitation and temperature) for a long time series of ~ 34 years
(19742007) are presented in Fig. 6 and are explained in Section 4.2.
The current hypothesis for this phenomenon of stable or slightly increasing snow cover is an increase in winter precipitation (snow-depth)
(though not signicant in this study) presented in Fig. 6 and explained
by Hewitt (2005) and Archer and Fowler (2004). This increased amount
of snow at lower altitudes may not have been melted away by the end of
spring season and hence caused the snow cover area to increase at these
lower altitudes (Fig. 5). Increasing trends of annual and seasonal precipitation in the Asian region are also conrmed by the IPCC's 5th synthesis
report (Hartmann et al., 2013). Using a regional climate model to assess
changes in severity and frequency of snowfall events due to weather
patterns including Westerlies, Ridley et al. (2013) indicated an increase
in frequency of occurrence up to 2100. Ridley et al. (2013) explained
that the Karakoram glaciers may continue to grow, or decline at a slower
rate, compared with those across the rest of the Himalayas due to increasing snowfall events. Another explanation for the phenomenon of
stable snow and glacier cover in the Karakoram is summer cooling (reduced energy inputs in summer months) in high elevation catchments
as described by Fowler and Archer (2006). These climate change trends
in the UIB are playing an important role in the stability or minor
expanding of the snow cover, especially in the Central Karakoram and
Western Himalaya regions of Pakistan. A consistent or slight increasing
trend of snow cover in the two sub-catchments (the Hunza and Astore,
situated at different latitudes and subjected to partially different climates) of the UIB indicates that the trend may be similar in the rest of
Upper Indus River catchment.

A.A. Tahir et al. / Science of the Total Environment 505 (2015) 748761

755

Fig. 5. Trend of MODIS-data-derived snow cover area (seasonally adjusted time series) in three altitudinal zones of the Astore River basin over a 13-year period (20002012) (bottom three
graphs). Trend analysis of total recorded precipitation and mean temperature (average of all three climate stations of Astore basin) for the same period (20002012) (top graph). Trend is
analyzed using a linear trend line equation, MannKendall's trend test and Sen's slope estimator S. Bold values are signicant (P b 0.05).

4.2. Trend analysis of stream ow, winter and summer total precipitation
and mean temperature of the Astore River basin
Trend analysis of the standardized values of annual runoff measured
at the Doyian hydrometric station of the Astore River is presented in
Fig. 7. Analysis of 34 years of ow record (19742007) for the Astore
River at Doyian indicates that the annual ow increases with time, as
shown in Fig. 7.
The Astore River trend of increasing discharge was signicant with a
Kendall tau () coefcient value of = +0.282 (P b 0.05) and a Sen's
slope value of +9.26 mmyear1 (Fig. 7). This contrasts with the ow

trend in the Hunza River where the stream ow has been decreasing
with time, possibly due to cooling temperatures at higher altitudes,
which decelerate snow and glacier melt, especially in summer. The
snow cover in the Astore catchment shows a stable or slightly rising
trends (Fig. 5), thus the trend towards increasing annual discharge in
the Astore River may be attributed to the trend of increasing summer
precipitation as analyzed in this study (Fig. 6) and reported by IPCC's
5th synthesis report (Hartmann et al., 2013) and Archer and Fowler
(2004).
The instrumental climate variables (seasonal precipitation and mean
temperature) for Astore River discharge and snow cover were analyzed

756

A.A. Tahir et al. / Science of the Total Environment 505 (2015) 748761

Table 3
Analysis of spatiotemporal change in snow cover area, hydrological regime and climatic
parameters for the Astore (current study) and Hunza (previous study) River basins. Significant (P 0.05) trends/correlations are indicated in bold.
Astore River
basin

Hunza River
basin

795%
= +0.03

3080%
= +0.02

= +0.06

= +0.25

= +0.07

= +0.01

1084-mm
501-mm at
Astore
723-mm at
Rattu
794-mm at
Rama

742-mm
680-mm at
Naltar
225-mm at
Ziarat
170-mm at
Khunjerab

Correlations (annual):
Q vs SCA
Q vs Tavg
Q vs P
P vs SCA
SCA vs Tavg

0.56
+0.75
0.28
+0.52
0.89

0.89
+0.85
+0.08
0.07
0.80

Correlations (seasonal):
Winter P vs winter SCA
Summer Q vs winter P
Summer Q vs summer Tavg
Summer Q vs summer P
Annual Q vs summer P

+0.44
+0.76
+0.52
0.23
+0.19

+0.01
+0. 1
+0.81
+0.04
+0.005

Increasing
trend
Increasing
trend
Decreasing
trend
Constancy/
slight increase

Decreasing
trend
Increasing
trend
Constancy

Snow cover area:

Annual snow cover area spatial variation
Snow cover trend over 13-years (20002012):
(Kendall's tau () coefcient value)
Snow cover trend on snow accumulation
periods (December to February)
Snow cover trend on snow melt periods
(July to September)
Hydrological regime:
Mean annual runoff
Total annual recorded precipitation

Trend analysis:
Stream ow
Seasonal precipitation
Summer mean temperature
Snow cover area

Constancy/
slight increase

P = precipitation, SCA = snow cover area, Q = discharge, Tavg = mean temperature

to verify the trend reported by Archer and Fowler (2004) and Archer
(2003). A seasonal analysis of precipitation (a) and temperature
(b) data (34-year time series; 19742007) was performed for the winter (December to February) and summer (July to September) periods at
the Astore climate station (Fig. 6). The trends for the seasonal (winter
and summer) precipitation and the summer mean temperature were
not signicant regarding the MK trend analysis, even if they present a
slight increase and decrease, respectively (Fig. 6). An increase in total
precipitation was found during summer and winter seasons with a
Sen's slope value of + 0.263 mmyear1 and + 1.1 mmyear1, respectively. A decreasing mean summer temperature trend was found
with the Sen's slope value of 0.009 Cyear1. These results support
the results of Archer and Fowler (2004) and Archer (2003), with a less
signicant trend (signicance level = 5%) in our analysis. Winter
mean temperature showed an increasing trend with a Sen's slope
value of +0.039 Cyear1. Even if the mean winter temperature is rising, it remains negative in high-altitude zones of snow accumulation,
but it can become striking if it continues to rise at the same rate over
the next few years.
4.3. Correlation between snow cover, climate variables and stream ow in
the Astore River basin (hydrological behavior)
Fig. 8 shows the relationship between standardized values of four
variables (precipitation, snow cover, mean temperature and discharge)

for the Astore River basin. The Astore river ow in summer (JulySeptember) is inuenced signicantly by both summer monsoon rainfall
and snow melt (Fig. 8). The monthly change in snow cover area has a
signicant inverse correlation with the mean temperature and discharge in the Astore River basin, as shown in Fig. 8. This indicates a continuous depletion in the seasonal snow cover in summer as the average
temperature in the area increases, resulting in increasing discharge. The
correlation between monthly precipitation at the Rama high-altitude
(3179 m asl) climate station and the snow cover area is very different
to that found by Tahir et al. (2011a) for the Hunza River basin. A correlation value of +0.36 was found by applying the Kendall rank correlation between the monthly precipitation at Rama and the snow cover
area of the Astore River basin (Fig. 8), while this value was 0.18
for the Hunza River basin (Tahir et al., 2011a). This difference in correlation may occur because the highest climate station of the Hunza
River basin (Khunjerab) is situated at an elevation of 4440 m asl
(Table 1), where the chances of known gauging errors (underestimation
of snowfall) are greater than at Rama ( 3179 m asl), situated at a
lower elevation than Khunjerab.
Important results for the analysis of the seasonal relationship between different variables for different seasons (winter and summer)
in the Astore River basin are given in Fig. 8 and Table 3. A signicant positive correlation was found between the winter precipitation and summer discharge (r = +0.76) (Fig. 8), which indicates the accumulation of
snow during the winter precipitation, which generates the discharge
upon melting in the summer season. Archer (2003) found the same correlation with a higher coefcient value (r = +0.88) on the 19821997
data period. As expected, the winter snow cover area is signicantly correlated to the winter precipitation with a correlation value of +0.44
(Fig. 8). This is due to the fact that the precipitation in winter in these
high-altitude catchments is in the form of snowfall, which increases
the snow cover area.
A correlation between the monthly difference in discharge and
monthly difference in snow cover area for the Astore and Hunza River
basins is presented in Fig. 9. The two variables are inversely correlated
(with signicant coefcient values) for both river catchments (Astore
and Hunza) (Fig. 9), which indicate that the change in river discharge
is dependent largely on the snow cover change in these catchment
areas. The Astore has the correlation values of r = 0.39 (Pearson's correlation coefcient), r = 0.48 (Spearman's rank correlation) and r =
0.36 (Kendall's rank correlation), while these values were r = 0.73,
r = 0.75 and r = 0.56, respectively, for the Hunza basin (Fig. 9). The
correlation values for the Astore are much lower than those for the
Hunza River basin (Fig. 9) because the Astore River discharge is also inuenced by summer rainfall (r = + 0.19) more than the Hunza (r =
+0.005) (Table 3). The discharge reaches its peak from July to September by the increasing snow and glacier melt. It is likely that the substantial snow cover difference has a large impact on discharges because a
larger snow cover area may also result in greater snow depths (at
least in low-slope valley areas).
4.4. Comparison of the physical and hydrological characteristics of the
Astore and Hunza River basins
A summary of the snow cover trend results, hydrological regime
characteristics and the correlation results between different hydrometeorological variables (stream ow, snow cover area, temperature and
precipitation) in the two catchment areas (Astore and Hunza) is presented in Table 3. On the one hand, the Hunza River basin is a highaltitude (mean elevation 4650 m asl) snow- and glacier-fed catchment (mean annual runoff 742 mm) situated in the Central
Karakoram region (with a large concentration of glaciers), which is dependent on temperature seasonality to generate runoff in the spring
and summer seasons. The Astore River basin, on the other hand, is comparatively a mid-altitude (mean elevation 4100 m asl) high-runoff
(mean annual runoff 1084 mm) catchment situated in the Western

A.A. Tahir et al. / Science of the Total Environment 505 (2015) 748761

757

Fig. 6. Trend analysis for climate parameters monitored in Astore River basin (at Astore climate station) (a) total summer (JAS) and winter (DJF) recorded precipitation (19742007) and
(b) mean summer (JAS) and winter (DJF) temperature (19742007).

Himalayas in the south of Karakoram, which has a good correlation between the summer stream ow and the current summer and preceding
winter precipitation amounts.
4.4.1. Concerning the snow cover change
A variation interval of snow cover area from 95% to 7% (Fig. 4)
may be attributed to the small catchment area and a mean elevation
(4100 m asl) lower than the Hunza River basin (4650 m asl). The Astore
River basin has only 5% of its area above 5000 m asl (Fig. 3 and
Table 1), whereas this percentage reaches 33% in the Hunza River
basin, as shown in Fig. 3 and reported by Tahir et al. (2011a). An increase in mean temperature results in snow cover depletion from the
smaller area at a lower altitude more rapidly than in a larger area at a
higher altitude. Similarly, a snowfall event can cover the smaller area
more rapidly than a larger area depending on the topographic characteristics (slope, vegetation and ground surface temperature, etc.) of
the catchment. The relative amplitude of the annual snow cover area

cycles for Astore and Hunza is also attributable to the elevations of the
freezing (0 C) isotherms, particularly the night-time/Tmin isotherm as
described by Forsythe et al. (2012b). Higher magnitude of the snow
cover area is also associated with the fact that the Astore basin receives
more precipitation in the winter and spring (Fig. 1b) that is mostly in
the form of snow.
The snow cover trend is almost stable or increasing slightly (Table 3)
in both catchments due to increasing winter precipitation in the area.
This increasing trend of snow may be a key factor for the positive
mass-balance of glaciers in the Central Karakoram region as reported
in detail by many authors (Hewitt, 2007; Scherler et al., 2011) and is
conrmed by recent reports from Gardelle et al. (2012), Gardelle et al.
(2013) and Kb et al. (2012), who explained that there is a positive
mass balance for the glaciers in the northern and eastern Karakoram regions, unlike in other parts of the world. Considering that the Astore
basin has a behavior that is climatically closer to the Himalayan behavior than the Karakoram region behavior, the increasing snow cover

758

A.A. Tahir et al. / Science of the Total Environment 505 (2015) 748761

Fig. 7. Trend in standardized values of annual ow in the Astore River (monitored at the Doyian hydrometric station) and Hunza River (monitored at the Dainyor hydrometric station), and
total annual recorded precipitation (recorded at Astore climate station) over 34-years (19742007). Signicant (P 0.05) trends are indicated in bold.

trend (Fig. 5) (Table 3) is somewhat surprising because many studies

(Berthier et al., 2007; Bolch et al., 2011; Gardelle et al., 2013; Vaughan
et al., 2013) have conrmed a negative mass balance in the Himalayan

region's glaciers. This may be associated with the fact that the Astore
River basin is situated in the extreme Western Himalayan foothills
and has a different climate from the Eastern Himalaya. The Astore

Fig. 8. Correlation between standardized values (monthly) of total precipitation, snow cover area, mean temperature and discharge for the Astore River basin. Signicant (P 0.05) correlations are indicated in bold.

A.A. Tahir et al. / Science of the Total Environment 505 (2015) 748761

759

Fig. 9. Correlations between monthly difference (difference in snow cover between two consecutive months) in snow cover and monthly difference in discharge in the (a) Astore and
(b) Hunza River basins. All correlation coefcients are signicant (P 0.05).

basin undergoes an intrusion of the Westerlies because it is situated

near the southern Karakoram ranges; this is why the winter precipitation trend (Fig. 6) in the Astore basin is more similar to the Hunza
basin (Tahir et al., 2011a). The Westerlies moisture is predominantly
drained in the HKH region with little moisture left in the far eastern
and southern regions of Himalayas.
The stable or slightly increasing snow cover trend in the Astore River
basin is slightly clearer ( = + 0.03) than in the Hunza River basin
(Table 3). This strength is prominent in the snow accumulation ( =
+ 0.06) and snowmelt ( = + 0.07) periods. This may be associated
with the mean summer temperature in the Astore River basin being in
a declining trend, as observed in Fig. 6, whereas the trend was constant
in the Hunza basin, as analyzed by Tahir et al. (2011a). These results
conrm previous studies by Archer and Fowler (2004), Bookhagen
and Burbank (2010), Hewitt (1998), Hewitt (2005), Hewitt (2007)
and Scherler et al. (2011).
4.4.2. Concerning the hydrological regime
Annual stream ow of the Astore River at Doyian is positively correlated (r = +0.19) (though not statistically signicant) (Table 3) to the
summer monsoon rainfall, which does not inuence the Hunza River
ow to a signicant extent due to reduced intrusion of summer monsoon precipitation into the central and northern Karakoram region.
Mean annual runoff (1084 mm) from the Astore catchment is higher
than the mean annual runoff ( 742 mm) generated from the Hunza
basin (Table 3). This may be associated with the rapid runoff from
spring monsoon rainfalls and winter showers (October to April;
Fig. 1b) and a portion of snowmelt from winter snowfall in the Astore
River basin. The only main source of runoff in the Hunza basin is winter
precipitation (in the form of snow), which starts melting in the spring
season to generate runoff. The observed total annual recorded precipitation value in the Astore River basin is also higher than that in the Hunza
River basin (Table 3).
The correlation results indicate that the annual discharge in the
Astore River is inuenced by the annual mean temperatures and the
current summer and preceding winter precipitation as well as by the
annual and seasonal snowmelt (Table 3). The correlation between summer discharge and summer mean temperature (r = + 0.52) in the
Astore River basin is much lower than that in the Hunza River basin

(r = + 0.80) (Table 3). Summer stream ow in the Astore basin is

also signicantly correlated with the preceding winter's precipitation
(winter snowfall) (Table 3), which melts in summer, whereas the annual discharge of the Hunza River is more strongly correlated to the annual
and seasonal snow cover melt and temperature seasonality (Table 3). A
correlation analysis between the annual discharge and current summer
precipitation indicates a weaker correlation coefcient value in the
Hunza basin (r = + 0.005) than in the Astore basin (r = + 0.19),
which indicates a minimal effect of summer monsoon in the central
Karakoram region. These results support the ndings of Archer (2003)
who indicated (on a shorter time series data than used in this study)
that the Astore and Hunza rivers have nival and glacial regimes, respectively. However, these differences may also stem from the lack of quality
precipitation data recorded at high altitudes.
Overall, the highest correlation values are found between the discharge, temperature and snow cover area (SCA) change (Table 3) in
the Astore and Hunza river basins, which indicates that the Upper
Indus River ow depends mainly on the snow and glacier melt rate,
which increases with the region's temperature seasonality (increased
temperatures in spring and summer). This is in accordance with the results found by Mukhopadhyay and Khan (2014a) and Bookhagen and
Burbank (2010) showing that the Western Himalayan catchments
(e.g., the Indus River catchment) are fed by the snowmelt up to 50%
as a fraction of the total annual discharge.
5. Conclusion
The spatio-temporal trends of snow cover and the hydrological regime of the Astore basin (sub-catchment of the UIB) (situated at the
south latitude of the UIB) were analyzed using satellite data and ground
station observations. The results obtained for the Astore basin were
then compared to the same parameters of the Hunza River basin (situated at the northern latitude of the UIB). The following conclusions
can be derived from the results of this study.
First, the UIB is a region undergoing a consistent/slight increasing
trend of snow cover in the north (Central Karakoram) and south (Western Himalaya) parts, possibly due to increasing winter precipitation and
decreasing or constant summer temperature trends. However, the results of this trend analysis should be veried in other sub-catchments

760

A.A. Tahir et al. / Science of the Total Environment 505 (2015) 748761

of the UIB (Shigar, Shyok and Gilgit) to illustrate this trend more condently. Moreover, the MODIS time series is short (b 15 years) and longer
records would be necessary to draw robust conclusions on snow cover
changes; nevertheless this present analysis provides an important picture of the last decades' snow cover trend.
Second, trend analyses of hydrological and meteorological variables,
calculated for a 34-year period, suggest an increasing trend in stream
ow and seasonal (winter and summer) precipitation and a decreasing
trend in the summer mean temperature in the Astore River basin.
Third, the UIB discharge depends much more on snow and glacier
melt than on the rainfall-runoff in the northern (Hunza) and southern
(Astore) regions, but a combination of the two is present in the southern
region (Astore), where some winter rainfall system overlaps with snow
and glacier melt. The western Himalaya shields both the Astore and
Hunza basins from substantial monsoonal rainfall.
Fourth, in addition to mean annual snow cover, indications of the
stability or slight expansion in summer snow cover in the Astore and
Hunza River basins may possibly be the reason for some glaciers
advancing or surging, as indicated in the previous studies (for
example Hewitt, 2007 and Gardelle et al., 2013). Therefore, the other
sub-catchments of the UIB should be studied in detail for the same parameters (snow cover dynamics and hydrological regime) to thoroughly understand the complex climate conditions of the region for better
stream ow predictions in the peak discharge periods.

Acknowledgments
Adnan Ahmad Tahir was nancially supported by the Higher Education Commission of Pakistan within the framework of a FrancePakistan
collaboration program for overseas studies. This nancial support is
gratefully acknowledged and appreciated. The authors extend their
thanks to the Water and Power Development Authority (WAPDA) and
the Pakistan Meteorological Department (PMD) for sharing the hydrological and meteorological data, respectively. Special thanks to Mr.
Danial Hashmi of WAPDA for providing the corrected coordinates of
the high-altitude weather stations. The authors also wish to thank
NASA and Japan's Ministry of the Economy, Trade and Industry (METI)
for providing ASTER GDEM. The suggestions and comments by the
two reviewers have substantially lifted-up the quality of this paper.
We acknowledge with particular gratitude the reviewers of this paper.
Finally, the authors thank to Prof. Stuart Lane, Director Institute of
Earth Surface Dynamics, University of Lausanne, for providing his services for the correction of English language in this manuscript.

References
Akhtar M, Ahmad N, Booij MJ. The impact of climate change on the water resources of
HindukushKarakorumHimalaya region under different glacier coverage scenarios.
J Hydrol 2008;355:14863.
Archer D. Contrasting hydrological regimes in the upper Indus Basin. J Hydrol 2003;274:
198210.
Archer DR, Fowler HJ. Spatial and temporal variations in precipitation in the Upper Indus
Basin, global teleconnections and hydrological implications. Hydrol Earth Syst Sci
2004;8:4761.
Arendt A, Bolch T, Cogley JG, Gardner A, Hagen J-O, Hock R, et al. Randolph glacier inventory a dataset of global glacier outlines: version 3.2. Global land ice measurements
from space. Boulder Colorado, USA: Digital Media; 2012.
Berthier E, Arnaud Y, Kumar R, Ahmad S, Wagnon P, Chevallier P. Remote sensing estimates of glacier mass balances in the Himachal Pradesh (Western Himalaya, India).
Remote Sens Environ 2007;108:32738.
Bolch T, Pieczonka T, Benn DI. Multi-decadal mass loss of glaciers in the Everest area
(Nepal Himalaya) derived from stereo imagery. Cryosphere 2011;5:34958.
Bookhagen B, Burbank DW. Topography, relief, and TRMM-derived rainfall variations
along the Himalaya. Geophys Res Lett 2006;33:L08405.
Bookhagen B, Burbank DW. Toward a complete Himalayan hydrological budget: spatiotemporal distribution of snowmelt and rainfall and their impact on river discharge.
J Geophys Res 2010;115:F03019.
Dimri AP, Chevuturi A. Model sensitivity analysis study for western disturbances over the
Himalayas. Meteorol Atmos Phys 2014;123:15580.

Dimri AP, Yasunari T, Wiltshire A, Kumar P, Mathison C, Ridley J, et al. Application of regional climate models to the Indian winter monsoon over the western Himalayas.
Sci Total Environ 2013;468469(Supplement: S36-S47).
Frland EJ, Allerup P, Dahlstrm B, Elomaa E, Jnsson T, Madsen H, et al. Manual for operational correction of Nordic precipitation data. DNMI, editor, DNMI, P.O. Box 43,
Blindern, Oslo, Norway; 1996. p. 66.
Forsythe N, Fowler H, Kilsby C, Archer D. Opportunities from remote sensing for
supporting water resources management in village/valley scale catchments in the
Upper Indus Basin. Water Resour Manag 2012a;26:84571.
Forsythe N, Kilsby CG, Fowler HJ, Archer DR. Assessment of runoff sensitivity in the upper
Indus basin to interannual climate variability and potential change using MODIS satellite data products. Mt Res Dev 2012b;32:1629.
Fowler HJ, Archer DR. Hydro-climatological variability in the Upper Indus Basin and implications for water resources. Regional hydrological impacts of climatic change
impact assessment and decision making. Brazil: IAHS Publ. 295, Seventh IAHS Scientic Assembly at Foz do Iguau; 2005. p. 1318.
Fowler HJ, Archer DR. Conicting signals of climatic change in the Upper Indus Basin. J Climate 2006;19:427693.
Gardelle J, Berthier E, Arnaud Y. Slight mass gain of Karakoram glaciers in the early
twenty-rst century. Nat Geosci 2012;5:3225.
Gardelle J, Berthier E, Arnaud Y, Kb A. Region-wide glacier mass balances over the
PamirKarakoramHimalaya during 19992011. Cryosphere 2013;7:126386.
Gilbert RO. 16.5 Sen's nonparametric estimator of slope. Statistical methods for environmental pollution monitoring. John Wiley and Sons; 1987. p. 2179.
Hall D, Riggs G, Salomonson V. MODIS/Terra Snow Cover 8-Day L3 Global 500 m Grid
V005, [March 2000 to December 2009]. Boulder, Colorado USA: National Snow and
Ice Data Center. Digital media; 2006 (updated weekly).
Hartmann DL, Tank AMGK, Rusticucci M, Alexander LV, Brnnimann S, Charabi Y, et al.
Observations: atmosphere and surface. In: Stocker TF, Qin D, Plattner G-K, Tignor
M, Allen SK, Boschung J, Nauels A, Xia Y, Bex V, Midgley PM, editors. Climate Change
2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge, United
Kingdom and New York, NY, USA: Cambridge University Press; 2013.
Hasson S, Lucarini V, Khan MR, Petitta M, Bolch T, Gioli G. Early 21st century climatology
of snow cover for the western river basins of the Indus River System. Hydrol Earth
Syst Sci Discuss 2013;10:1314590.
Hewitt K. Glaciers receive a surge of attention in the Karakoram Himalaya. EOS Trans Am
Geophys Union 1998;79:1045.
Hewitt K. Hazards of melting as an option: Upper Indus glaciers I & II. Dawn. 20/21May;
2001.
Hewitt K. The Karakoram anomaly? Glacier expansion and the elevation effect,
Karakoram Himalaya. Mt Res Dev 2005;25:33240.
Hewitt K. Tributary glacier surges: an exceptional concentration at Panmah Glacier,
Karakoram Himalaya. J Glaciol 2007;53:1818.
Hewitt K, Wake CP, Young GJ, David C. Hydrological investigation at Biafo Glacier,
Karakoram range, Himalaya; an important source of water for the Indus river. Ann
Glaciol 1989;13:1038.
Hirsch RM, Slack JR. A nonparametric trend test for seasonal data with serial dependence.
Water Resour Res 1984;20:72732.
Hirsch RM, Slack JR, Smith RA. Techniques of trend analysis for monthly water quality
data. Water Resour Res 1982;18:10721.
Immerzeel WW, Droogers P, de Jong SM, Bierkens MFP. Large-scale monitoring of snow
cover and runoff simulation in Himalayan river basins using remote sensing. Remote
Sens Environ 2009;113:409.
Immerzeel WW, Pellicciotti F, Shrestha AB. Glaciers as a proxy to quantify the spatial distribution of precipitation in the Hunza basin. Mt Res Dev 2012;32:308.
Immerzeel WW, Pellicciotti F, Bierkens MFP. Rising river ows throughout the twentyrst century in two Himalayan glacierized watersheds. Nat Geosci 2013;6:7425.
Kb A, Berthier E, Nuth C, Gardelle J, Arnaud Y. Contrasting patterns of early twenty-rstcentury glacier mass change in the Himalayas. Nature 2012;488:4958.
Kendall MG. Rank correlation measures. London: Charles Grifn; 1975.
Kendall MG, Gibbons JD. Rank correlation methods. London: Edward Arnold; 1990.
Khan A, Richards KS, Parker GT, McRobie A, Mukhopadhyay B. How large is the Upper
Indus Basin? The pitfalls of auto-delineation using DEMs. J Hydrol 2014;509:44253.
Lee S, Klein AG, Over TM. A comparison of MODIS and NOHRSC snow-cover products for
simulating streamow using the snowmelt runoff model. Hydrol Process 2005;19:
295172.
Liniger H, Weingartner R, Grosjean M. Mountains of the world: water towers for the 21st
century. Mountain agenda for the commission on sustainable development (CSD),
BO12, Berne; 1998. p. 32.
Mann HB. Nonparametric tests against trend. Econometrica 1945;13:24559.
Maurer EP, Rhoads JD, Dubayah RO, Lettenmaier DP. Evaluation of the snow-covered area
data product from MODIS. Hydrol Process 2003;17:5971.
METI, NASA. ASTER Global Digital Elevation Model (GDEM) version 2. http://reverb.echo.
nasa.gov/reverb/.2013, 2011.
Mukhopadhyay B, Khan A. A quantitative assessment of the genetic sources of the hydrologic ow regimes in Upper Indus Basin and its signicance in a changing climate. J
Hydrol 2014a;509:54972.
Mukhopadhyay B, Khan A. Rising river ows and glacial mass balance in central
Karakoram. J Hydrol 2014b;513:192203.
Pellicciotti F, Bauder A, Parola M. Effect of glaciers on streamow trends in the Swiss Alps.
Water Resour Res 2010;46:W10522.
Ridley J, Wiltshire A, Mathison C. More frequent occurrence of westerly disturbances in
Karakoram up to 2100. Sci Total Environ 2013;468469(Supplement: S31-S35).
Rodgers JL, Nicewander WA. Thirteen ways to look at the correlation coefcient. Am Stat
1988;42:5966.

A.A. Tahir et al. / Science of the Total Environment 505 (2015) 748761
Scherler D, Bookhagen B, Strecker MR. Spatially variable response of Himalayan glaciers
to climate change affected by debris cover. Nat Geosci 2011;4:1569.
Sen PK. Estimates of the regression coefcient based on Kendall's tau. J Am Stat Assoc
1968;63:137989.
Sevruk B. Correction of precipitation measurements. Workshop on the correction of precipitation measurements. Zurich, Switzerland: WMO/IAHS/ETH; 1985. p. 1323.
Sevruk B. Reliability of precipitation measurement. International workshop on precipitation measurement. WMO Tech. Document; 1989. p. 139.
SIHP. Snow and ice hydrology project, Upper Indus River Basin. WAPDA-IDRC-Wilfrid
Laurier University; 1990. p. 179.
Sirguey P, Mathieu R, Arnaud Y, Fitzharris BB. Seven years of snow cover monitoring with
MODIS to model catchment discharge in New Zealand. IEEE International Geoscience
and Remote Sensing Symposium (IGARSS), 2. ; 2009. p. II-8636. [Cape Town, 2009].
Spearman C. The proof and measurement of association between two things. Am J
Psychol 1904;15:72101.
Tahir AA, Chevallier P, Arnaud Y, Ahmad B. Snow cover dynamics and hydrological regime
of the Hunza River basin, Karakoram Range, Northern Pakistan. Hydrol Earth Syst Sci
2011a;15:227590.

761

Tahir AA, Chevallier P, Arnaud Y, Neppel L, Ahmad B. Modeling snowmelt-runoff under

climate scenarios in the Hunza River basin, Karakoram Range, Northern Pakistan. J
Hydrol 2011b;409:10417.
Tekeli AE, Akyrek Z, Arda Sorman A, Sensoy A, nal Sorman A. Using MODIS snow cover
maps in modeling snowmelt runoff process in the eastern part of Turkey. Remote
Sens Environ 2005;97:21630.
Vaughan DG, Comiso JC, Allison I, Carrasco J, Kaser G, Kwok R, et al. Observations:
cryosphere. In: Stocker TF, Qin D, Plattner G-K, Tignor M, Allen SK, Boschung J,
Nauels A, Xia Y, Bex V, Midgley PM, editors. Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. United Kingdom and New York, NY, USA:
Cambridge University Press, Cambridge; 2013.
Wake CP. Glaciochemical investigations as a tool to determine the spatial variation of
snow accumulation in the Central Karakoram, Northern Pakistan. Ann Glaciol 1989;
13:27984.
Young GJ, Hewitt K. Hydrology research in the upper Indus basin, Karakoram Himalaya,
Pakistan. Hydrology of mountainous areas. 1990;190:13952.