Professional Documents
Culture Documents
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.
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.
750
A.A. Tahir et al. / Science of the Total Environment 505 (2015) 748761
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
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).
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
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
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
Trend analysis:
Stream ow
Seasonal precipitation
Summer mean temperature
Snow cover area
Constancy/
slight increase
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.
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).
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