Journal of Hydrology (2007) 344, 171– 184
available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/jhydrol
Historical temporal trends of hydro-climatic variables
and runoff response to climate variability and their
relevance in water resource management in the
Hanjiang basin
Hua Chen a, Shenglian Guo
a,*
, Chong-yu Xu
b,d
, Vijay P. Singh
c
a
State Key Laboratory of Water Resources and Hydropower Engineering Science, Wuhan University, Donghunanlu 8, Wuhan,
Hubei 430072, China
b
Department of Geosciences, University of Oslo, P.O. Box 1047, Blindern, NO-0316 Oslo, Norway
c
Department of Biological and Agricultural Engineering, Texas A&M University, Scoates Hall, 2117 TAMU, College Station,
TX 77843-2117, USA
d
Department of Earth Sciences, Uppsala University, Sweden
Received 14 January 2006; received in revised form 31 May 2007; accepted 30 June 2007
KEYWORDS
Trends analysis;
Mann–Kendall;
Climate variability;
Water balance model;
Danjiangkou reservoir;
South-to-North Water
Diversion Project
Summary The Danjiangkou reservoir lies in the upper Hanjiang basin and is the source of
water for the middle route of the South-to-North Water Diversion Project (SNWDP) in
China. Any significant change in the magnitude or timing of runoff from the Danjiangkou
reservoir induced by changes in climatic variables would have significant implications for
the economic prosperity of the area in the Hanjiang basin as well as for the South-to-North
Water Diversion Project. In this paper the following issues are investigated: (1) Temporal
trends of annual and seasonal precipitation and temperature from 1951 to 2003 in the
Hanjiang basin are analyzed using the Mann–Kendall and the linear regression methods;
spatial distributions of precipitation and temperature are interpolated by the inverse distance weighted interpolation method. (2) Temporal trends of runoff, precipitation and
temperature from 1951 to 2003 in the Danjiangkou reservoir, an upper stream basin of
the Hanjiang River, are further tested. (3) To assess the impact of climate change on
water resources and predict the future runoff change in the Danjiangkou reservoir basin,
a two-parameter water balance model is used to simulate the hydrological response for
the climate change predicted by GCMs for the region for the period of 2021–2050.
* Corresponding author. Tel.: +86 27 68772765; fax: +86 27 68773568.
E-mail address: slguo@whu.edu.cn (S. Guo).
0022-1694/$ - see front matter ª 2007 Published by Elsevier B.V.
doi:10.1016/j.jhydrol.2007.06.034
172
H. Chen et al.
The results indicate that (1) at the a = 0.05 significance level precipitation in the Hanjiang basin has no trend, but the temperature in the same region has significant upward
trends in most parts of the Hanjiang basin. (2) The mean annual, spring, and winter runoffs in the Danjiangkou reservoir basin have decreasing trends. (3) The results simulated
for the period 2021–2050 show that runoff of the Danjiangkou reservoir would increase in
all the seasons, mainly in response to the predicted precipitation increase in the region.
Sensitivity analysis shows that a 1 C and 2 C increase in temperature would reduce the
mean annual runoff to about 3.5% and 7%, respectively. A decrease/increase of the mean
monthly precipitation of 20% and 10% would decrease/increase the mean annual runoff to
about 30% and 15%, respectively. The results of this study provide a scientific reference
not only for assessing the impact of the climate change on water resources and the flood
prevention in the Hanjiang basin, but also for dimensioning the middle route of the SNWDP
in China.
ª 2007 Published by Elsevier B.V.
Introduction
Investigations of regional and global climatic changes and
variabilities and their impacts on the society have received considerable attention in recent years. According
to the Third Assessment Report of the Intergovernmental
Panel on Climate Change (IPCC, 2001), the global average
surface temperature has increased by about 0.6 ± 0.2 C
over the 20th century, and rainfall has decreased over
much of the Northern Hemisphere sub-tropical regions
by about 0.3% per decade during the 20th century. In
some regions, such as parts of Asia and Africa, the frequency and intensity of droughts have been observed to
increase in recent decades. Much attention had been paid
to analyze climatic changes in China (Guo et al., 2002;
Gemmer et al., 2004; Wang et al., 2004; Xiong and Guo,
2004; Guo et al., 2005; Zhang et al., 2005; Huang
et al., 2005; Xu et al., 2006; Zhou and Yu, 2006). Guo
et al. (2002) studied the impact of climate change on
water resources of the Hanjiang basin based on a semidistributed monthly water balance model and their results
showed that the precipitation change is the main factor
for the change in runoff. Wang et al. (2004) demonstrated
that an increasing trend in precipitation variations was observed during the second half of the 20th century in West
China, but a similar trend was not found in East China,
where the 20- to 40-year periodicities were predominant
in the precipitation variability. Huang et al. (2005) showed
that the temperature rise in winter was shown to be
linked to the presence of an anomalously strong zonal circulation in Eurasia and a weak polar vortex since the
1980s. Xu et al. (2006) analyzed the future climate change
responses in the time-slice of 2071–2100 (2080s) under
SRES B2 scenario over China. According to their results,
there would be an obvious surface air temperature increase in the north of China relative to that in the south
of China, and there would be an overall increase of the
simulated precipitation in the 2080s under SRES B2 scenario over most areas of China. At the same time there
would be significant precipitation decreases in South China
in winter and obvious precipitation decreases in Northeast
China and North China in summer with high surface air
temperature increases. Zhou and Yu (2006) examined variations of the surface air temperature over China and the
globe in the 20th century simulated by 19 coupled climate
models driven by historical natural and anthropogenic
forcings. These studies will be very helpful to analyze
and assess the impact of climatic change on the water resources in China.
Observational and historical hydro-climatic data are
generally used for planning and designing water resources
projects. There is an implicit assumption of so-called stationarity, implying time-invariant statistical characteristics
of the time series under consideration, in virtually all
water resources engineering works. Such an assumption
can no longer be valid if the global climate changes as a
result of the increase of greenhouse gases in the atmosphere. This, of course, results in major problems (e.g.,
dislocation and inefficiencies) in regional water resources
management (Kahya and Kalaycı, 2004). Therefore, the
need to study the climatic change trends in the Hanjiang
basin, such as precipitation, temperature and runoff, is
urgent.
There are many parametric and non-parametric methods that have been applied for detection of trends (Zhang
et al., 2006). Parametric trend tests are more powerful
than non-parametric ones, but they require data to be
independent and normally distributed. On the other hand,
non-parametric trend tests only require the data be independent and can tolerate outliers in the data. One of the
widely used non-parametric tests for detecting a trend in
hydro-climatic time series is the Mann–Kendall (MK) test
(Hirsch et al., 1982; van Belle and Hughes, 1984; Zetterqvist, 1991; Zhang et al., 2001; Burn and Elnur, 2002; Yue
et al., 2002; Yue and Wang, 2002; Yue and Pilon, 2004;
Burn et al., 2004; Zhang et al., 2005; Arora et al.,
2005; Aziz and Burn, 2006; Gemmer et al., 2004; Zhang
et al., 2006; Zhu and Day, 2005). Burn and Elnur (2002)
developed a trend detection framework which utilized
the MK test to identify trends in hydrological variables.
Kundzewicz and Robson (2004) outlined and presented a
brief overview of trend detection tests. Yue and Pilon
(2004) applied Monte Carlo simulation to compare the
power of the statistical tests: the parametric t-test, the
non-parametric MK, bootstrap-based slope, and bootstrap-based MK tests to assess the significance of monotonic trends. Their simulation results indicate that the
t-test and the BS-slope test (slope-based tests) have the
same power and the MK and BS-based MK tests (rank
based tests) have the same power. For normally distrib-
Historical temporal trends of hydro-climatic variables and runoff response to climate variability
uted data, the power of the slope-based tests is slightly
higher than that of the rank-based tests, and vice versa.
Zhang et al. (2005) applied the MK test to examine trends
in the precipitation and temperature data in the Yangtze
River basin and detected significant positive and negative
trends at the a = 0.1 significance level. Becker et al.
(2006) analyzed precipitation trends in the Yangtze River
basin for the past 50 years by applying the MK trend test
and geospatial analyses. Significant positive trends at
many stations were observed for summer months, which
naturally show precipitation maxima. Zhang et al. (2006)
detected the temporal trends and frequency changes at
three major hydrological stations at Yangtze River, i.e.,
Yichang, Hankou and Datong, representing the upper,
middle and lower reaches, respectively, with the help of
parametric t-test, MK analysis and wavelet transform
methods. The MK test has been approved as a powerful
tool to detect trends of the hydro-meteorological time
series.
Using both statistical methods and a well-tested hydrological model, this study examines both the historical
variations and the future changes in the hydro-climatic
variables in the Hanjiang River basin and sub-basins thereof. More specifically, the objectives of the study are
threefold: (1) to examine temporal trends and their spatial distribution of historical annual and seasonal precipitation and temperature series in the Hanjiang basin, (2)
to detect temporal trends of precipitation, temperature,
and runoff in the Danjiangkou reservoir basin, an upstream sub-basin of the Hanjinag, and analyze mutual
relationships between the three factors in order to clarify
the impact of climatic change on water resources, and (3)
to predict future runoff changes based on future scenarios
obtained from GCMs as input to a two-parameter water
balance model.
Figure 1
173
The Hanjiang basin and the South-to-North
Water Diversion Project
The Hanjiang basin
The Hanjiang River, located between 106–114 E and
30–34 N and shown in Fig. 1, is the biggest tributary of
the Yangtze River. It passes through the provinces of Shanxi,
Sichuan, Henan and Hubei; and merges into the Yangtze River at Wuhan. The river’s length is 1570 km and the basin
area is 170,400 km2. The basin has a sub-tropical monsoon
climate and has, as a result, dramatic diversity in its water
resources. Since there are many high and large mountains in
the basin, the study further divides the basin into different
climatic zones, i.e., sub-tropical, warm temperate, mountainous temperate and mountainous sub-temperate zones,
etc. For this study, the whole Hanjiang basin is divided into
three regions, as shown in Fig. 2: the Danjiangkou reservoir
sub-basin (upper sub-basin), the middle and the lower subbasins according to the differences in longitude and altitude, which give rise to sharp changes in precipitation and
temperature within the region. The basin’s annual precipitation is about 700–1000 mm, which gradually increases
from the upper to the lower basin and decreases from the
south to the north in the upper basin.
The Danjiangkou reservoir is the source of water for the
middle route of the well-known South-to-North Water Diversion Project (SNWDP) in China. The drainage area of the reservoir is 96,000 km2. As the Danjiangkou reservoir will
transfer part of its water for the SNWDP, which will alter
the reservoir’s discharge and its distribution. This will have
an impact on the socio-economic development and the environment in the middle and lower sub-basins which are the
bases of agriculture and industry in the Hubei province.
Location of the Hanjiang basin and the middle route of the SNWDP in China.
174
H. Chen et al.
Figure 2
The locations of the meteorological stations in the Hanjiang basin.
The South-to-North Water Diversion Project
In North China, economic growth and population increase
have principally been responsible for excessive extraction
of groundwater as well as utilization of untreated urban
sewage water. This situation is to be improved by implementing, step by step, the long distance SNWDP, together
with intensive water conversion, pollution control and the
rational use of local water resources. As its importance to
the Chinese economic and societal development, the project has attracted more and more attention and has therefore been a subject of significant discussion (Wang and Ma,
1999; Chen et al., 2002; Liu and Zheng, 2002; Shao et al.,
2003; Yang and Zehnder, 2005). Wang and Ma (1999) pointed
out that the middle route would be affected by several environmental–geological problems, such as the slope stability
of swelling clay and rock, soil salinization as a result of
the rise of the groundwater table due to channel leakage,
the settlement of ground surface in the coal mining area,
liquefaction of sand, drainage through the left bank of the
canal and frozen heave problems. The south-to-north water
transfer schemes are considered and discussed by Liu and
Zheng (2002) in the context of resolving water shortage
problems in the north of China. As a substantial amount of
water will be diverted from the Hanjiang River in the Middle
Route Project, causing reductions of runoff in the downstream, which may lead to the worsening of existing eutrophication problem there. Shao et al. (2003) argued for a
detailed study and precautions must be made to prevent
the environmental and ecological problems on the downstream of the Hanjiang. Yang and Zehnder (2005) examined
the decision-making process of the project from the viewpoint of the country’s transition from a centrally planned
economy to a market economy, rapid economic development, and severe environmental degradation; and suggested a high degree of uncertainty in the future water
demand.
With Beijing city as its uppermost destination, the middle route of SNWDP will supply water for North China,
including the Tangbaihe plain and the middle and western
parts of the Huang–Huai–Hai plain with a total area of
about 155,000 km2, as shown in Fig. 1. The first task of
the middle route of SNWDP is to heighten the dam of
the Danjiangkou reservoir from 157 m3 to 170 m to enlarge
the storage capacity to 29.05 billion m3 from the current
level of 11.6 billion m3. Based on the completion of Danjiangkou reservoir extension project, the reservoir area
will then extend by another 324 km2, affecting more than
250,000 people in Shiyan of Hubei Province and Xichuan
County in Henan Province. The expanded Danjiangkou Reservoir will provide 9.5 billion m3 of water annually to Beijing, Tianjin, and cities in Hebei, Henan and Hubei
provinces. It will be able to provide 13–14 billion m3 by
2030. The water transfer project, after being diverted to
the North, will impact on the water utilization along the
middle and lower reaches of Hangjiang River. In accordance with the development level in 2020, some compensatory projects will be built on the middle and lower
Hanjiang to ensure the development of industry and agriculture, and navigation and preservation of the environment of the water exporting region. The project will be
an important and a basic facility for mitigating the existing
crisis of water resources in North China.
The advantages of this project lie mainly in the good
quality of water to be diverted and greater water-supply
coverage available, and that the water can be conveyed
by gravity through canals to be built along Funiu and Taihang Mountains. The undesirable point of the middle route
is that the amount of water allowable for diversion from
the Danjiangkou reservoir would be subject to certain limitations. If successive dry years occur, the water available
for diversion would be insufficient. It is therefore of vital
importance to study the variability of discharge in the basin
in relation to the climate variability in the region.
Data and methodology
Data
The following hydro-meteorological data are used in the
analysis. Daily precipitation and temperature data (1951–
2003) from 14 National Meteorological Observatory (NMO)
Historical temporal trends of hydro-climatic variables and runoff response to climate variability
Table 1 The latitude and longitude of the meteorological
stations in the Hanjiang basin
No.
Station
name
Latitude
(N)
Longitude
(E)
Elevation
(m a.s.l)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Hanzhong
Foping
Shangzhou
Xixia
Nanyang
Shiquan
Ankang
Yunxian
Fangxian
Laohekou
Zaoyang
Zhongxiang
Tianmen
Wuhan
33.04
33.32
33.52
33.18
33.02
33.03
32.43
32.51
32.03
32.23
32.09
31.10
30.40
30.37
107.02
107.59
109.58
111.30
112.35
108.16
109.02
110.49
110.44
111.40
112.45
112.34
113.10
114.08
508.4
1087.7
742.2
250.3
129.2
484.9
290.8
201.9
434.4
90.0
125.5
65.8
34.1
23.3
stations were provided by the National Climatic Centre of
China. The location of the stations in the basin is shown in
Fig. 2, and their altitudes and longitudes are listed in Table
1. The monthly runoff (1951–2003) of the Danjiangkou reservoir was provided by the Bureau of Hydrology of the
Changjiang Water Resources Commission. The series of the
annual temperature, precipitation and runoff are tested to
be normally distributed by using the Kolmogorov–Smirnov
method with the only exception being of the precipitation
series of the Nanyang at the a = 0.1 significance level. The
monthly precipitation and temperature data of the GCMs
were downloaded from the website of the IPCC Data Distribute Center (http://ipcc-ddc.cru.uea.ac.uk/).
Lying in a sub-tropical monsoon climate zone, the basin
climate has dramatic diversity. The mean annual temperature is about 15–17 C, the maximum mean monthly temperature is about 22–24 C in July. The extreme maximum
temperature in July and the extreme minimum temperature in January are 43.4 C and 20 C, respectively. Both
precipitation and runoff have strong inter-annual and intra-annual variabilities. The mean annual precipitation is
about 700–1100 mm and 80% of it happens in the period
between May and October. The seasonal variation of runoff
is similar to that of precipitation. The runoff between July
and October is 65% of the annual runoff. The mean runoff
coefficient of the basin is 0.48. The long-term mean runoff
of the Danjiangkou reservoir basin is 1200 m3/s and the
maximum runoff is about six times the minimum runoff
in one year. The coefficient of variation of annual runoff
is 0.4
Hydrological water balance model
Hydrologic models provide a framework in which to conceptualize and investigate the relationship between climate and water resources. The scientific literature of
the past two decades contains a large number of reports
dealing with the application of hydrologic models to the
assessment of potential effects of climate change on a
variety of water resource issues (Xu, 1999). The choice
175
of a model for a particular case study depends on many
factors (Gleick, 1986), amongst which the study purpose,
model and data availability have been the dominant ones
(Ng and Marsalek, 1992; Xu, 1999). For example, for
assessing water resources management on a regional scale,
monthly rainfall-runoff (water balance) models were found
useful for identifying hydrologic consequences of changes
in temperature, precipitation, and other climatic variables
(e.g., Gleick, 1986; Schaake and Liu, 1989; Mimikou et al.,
1991; Arnell, 1992; Xu and Halldin, 1997; Xu and Singh,
1998; Guo et al., 2002).
The hydrological model used in this study is a twoparameter monthly water balance model developed by
Xiong and Guo (1999) for simulating the hydrological impact of the changing climate. The model has only two
parameters to be calibrated and requires as input the
monthly areal precipitation, pan evaporation, air temperature and streamflow (for calibration only). The model
outputs include monthly streamflow, actual evapotranspiration and soil moisture content index. The model has
been tested in more than 100 catchments with different
climates and sizes in China and has, with its simple structure, proved to be quite efficient for simulating monthly
runoff. The two-parameter monthly water balance model
can be easily and efficiently incorporated in the water resources planning program and study of climate change impacts due to its simplicity and high efficiency of
performance. This model has been used as a standard tool
for simulating climate change impacts by different institutions in China, including the National Climate Centre. Guo
et al. (2002) applied this model to simulate climate change
impacts in the Danjiangkou reservoir basin. They had done
the calibration and evaluation of the model and got very
good simulation results of monthly runoff in Hanjiang basin. Based on their work, the two-parameter water balance
model was selected to use in this study. The parameter
values of the model which had been well calibrated and
tested by Guo et al. (2002) for the same basin are directly
used in this study.
Methods
Four statistical methods are used in this study to analyze the
spatial variations and temporal trends of the hydro-climatic
series: (1) A simple linear regression method, which is a
parametric t-test method, is used to test the long-term linear trend. (2) The Mann–Kendall test, which was originally
devised by Mann (1945) as a non-parametric test for detecting trends and the distribution of the test statistic, derived
by Kendall (1975), is used to test the non-linear trend
as well as the turning point. (3) The Kendall s method, which
is a non-parametric method for testing the correlation
(Kendall, 1938; Sprent, 1990), is used to test the correlation
between runoff, precipitation and temperature. (4) An
inverse distance weighted interpolation method (IDW),
which is based on the assumption that the interpolating
surface should be influenced most by nearby points and less
by more distant points (Gemmer et al., 2004), is used to
map the regional distribution of the test statistics.
To predict the hydrological impact of precipitation and
temperature changes in the Danjiangkou reservoir basin,
the delta change method (Hay et al., 2000) and the
176
H. Chen et al.
two-parameter water balance model (Xiong and Guo, 1999;
Guo et al., 2002), which have been tested in the Danjiangkou reservoir basin, are used. The delta change method
calculates the difference between GCMs simulated current
and future climate conditions and adds to the current observed time series of climate variables in the Danjiangkou
basin as input to the two-parameter water balance model
to simulate future runoff.
In order to study the decadal variation of historical annual
and seasonal precipitation, temperature and runoff, the
b
Temperature vatiations (0C)
400
Upper
200
0
-200
1960
Precipitation variation (mm)
Decadal variations of annual and seasonal
precipitation, temperature and runoff in Hanjiang
basin
1980
T(year)
0.8
Upper
0.4
0
-0.4
-0.8
1960
2000
Temperature vatiations (0C)
Precipitation variation (mm)
a
Results and discussion
400
Middle
200
0
-200
1980
T(year)
2000
1
Middle
0.5
0
-0.5
-1
1960
1980
T(year)
1960
2000
1980
T(year)
2000
Temperature vatiations ( C)
Lower
0
Precipitation variation (mm)
1.5
Lower
400
0
1
0.5
0
-0.5
-400
-1
1960
Runoff (108m3)
c
1980
T(year)
2000
1960
1980
T(year)
2000
400
Precipitation , Temperature
or Runoff Variations
Mean Differences of Decadal
Values to the Longterm Mean
Base line
200
0
-200
1960
1980
T(year)
2000
Figure 3 The annual precipitation (a), temperature (b) and runoff (c) variations and their mean differences of decadal values to
the long-term mean in the upper, middle and lower Hanjiang basin.
Historical temporal trends of hydro-climatic variables and runoff response to climate variability
mean annual and seasonal values for the whole study period (1951–2003) and for each decade (1951–1960, 1961–
1970, 1971–1980, 1981–1990, and 1991–2003) were calculated and compared. These results were plotted in
Fig. 3 and shown in Tables 2 and 3. It is observed that:
(1) For the upper basin a wet period (1981–1990) and a
dry (1991–2003) period are clearly seen which have a
mean annual precipitation of 76.11 mm (73.93 mm) higher
(lower) and a mean annual runoff of 67.36 · 108 m3
(81.30 · 108 m3) higher (lower) than the long-term average, respectively. (2) The periods (1951–1960, 1961–
1970 and 1991–2003) are warmer than the long-term
average, while the period (1981–1990) is cooler than the
long-term average. The results reveal that the seasonal
variations of the decadal values are much larger than
those of annual differences, especially for precipitation.
During the period of 1991–2003 precipitation and runoff
in autumn are 32.08 mm and 36.10 · 108 m3 lower than
their long-term average values, respectively. (3) For the
middle basin the mean decadal changes are not remarkable, the biggest difference is found for the period
(1991–2003) which has a mean precipitation of 27.03
mm lower than the long-term average and the period
177
(1991–2003) is 0.417 C warmer than the long-term average, and before the 1980s the decadal variations of temperature are lower than the long-term average. For
seasonal variations, precipitation in autumn has a larger
difference than that of precipitation in other seasons.
(4) For the lower basin, a dry period (1971–1980) and a
wet period (1991–2003) are found which have a mean annual value of 108.53 mm (46.36 mm) lower (higher) than
the long-term average, respectively. The period (1991–
2003) is warmer, while the rest of the periods are cooler
than the long-term average. For the period (1991–2003),
temperature in winter is 1.094 C higher than the longterm average and runoff is much lower than the long-term
average (Fig. 3c).
The above discussion shows that when entering the
1990s there is a very dry period in the Hanjiang basin.
If this situation does not alter in the 21st century it will
have serious implications for agriculture, industry and
drinking water supply in the middle China region. More
importantly, this decrease in runoff will have a direct
impact on the very viability of the middle route of
the South-to-North Water Diversion Project (SNWDP) in
China.
Table 2 Decadal variations of the precipitation (mm) and temperature (C) as compared with the long-term average in the
upper, middle and lower Hanjiang basin
1951–1960
1961–1970
1971–1980
P
T
P
T
P
T
1981–1990
1991–2003
P
T
P
T
Upper basin
Annual
Spring
Summer
Autumn
Winter
15.73
11.37
40.78
19.45
3.64
0.066
0.043
0.193
0.110
0.235
20.47
29.74
41.35
36.32
2.68
0.023
0.102
0.615
0.095
0.221
15.96
0.97
19.80
3.21
1.03
0.027
0.039
0.169
0.061
0.047
76.11
10.84
41.68
21.61
1.24
0.300
0.197
0.702
0.115
0.175
73.93
23.22
16.39
32.08
2.48
0.148
0.167
0.211
0.030
0.449
Middle basin
Annual
Spring
Summer
Autumn
Winter
14.02
9.99
35.75
31.85
7.03
0.065
0.352
0.003
0.060
0.451
7.89
25.28
27.29
24.11
7.47
0.110
0.196
0.392
0.205
0.381
11.15
13.85
5.87
10.46
3.53
0.188
0.203
0.002
0.107
0.172
1.94
16.98
2.38
21.52
3.70
0.179
0.041
0.241
0.074
0.098
27.03
9.35
0.16
18.65
0.22
0.417
0.546
0.120
0.344
0.853
Lower basin
Annual
Spring
Summer
Autumn
Winter
29.72
43.59
26.75
36.72
2.44
0.226
0.453
0.035
0.021
0.353
43.52
29.64
2.34
0.22
4.37
0.207
0.259
0.138
0.235
0.564
108.53
2.81
80.07
13.56
10.03
0.218
0.264
0.148
0.351
0.280
62.05
37.73
27.22
77.48
0.31
0.168
0.001
0.312
0.208
0.225
46.36
16.13
30.03
20.76
9.44
0.630
0.750
0.221
0.626
1.094
Note: P – precipitation; T – temperature.
Table 3
Annual
Spring
Summer
Autumn
Winter
Decadal variations of the runoff (108 m3) of the Danjiangkou reservoir as compared with the long-term average
1951–1960
1961–1970
1971–1980
1981–1990
1991–2003
27.32
2.56
37.39
9.99
2.47
35.61
27.00
28.79
34.03
3.36
24.50
5.69
15.12
2.92
0.77
67.36
6.33
33.58
25.78
1.65
81.30
19.27
20.81
36.10
5.17
178
Spatial distribution of temporal trend in annual and
seasonal precipitation and temperature as detected
by the Mann–Kendall statistics and a linear
regression method
Applying the Mann–Kendall statistics and a linear regression
method to each station in the Hanjiang basin, temporal
trends of annual and seasonal precipitation and temperature are tested at the a = 0.05 significance level. These temporal trends are then interpolated by using the IDW method
to show their spatial distributions.
H. Chen et al.
The spatial distributions of temporal trend in annual and
seasonal precipitation are shown in Fig. 4. It is seen that the
results of both methods are similar and there is no trend in
precipitation in most parts of the Hanjiang basin at the
a = 0.05 significance level.
Fig. 5 demonstrates the spatial distribution of annual and
seasonal trends in temperature in the Hanjiang basin. It is
seen that the results between the two methods are almost
the same and the temporal trend of temperature has rich
spatial patterns as compared with those of precipitation.
The annual temperature has significant upward trends in
Figure 4 Spatial distribution of the temporal trends of annual and seasonal precipitation as measured by the Mann–Kendall
statistics (a) and linear regression (b) at the a = 0.05 significance levels in the Hanjiang basin.
Historical temporal trends of hydro-climatic variables and runoff response to climate variability
179
Figure 5 Spatial distribution of the temporal trends of annual and seasonal temperature as measured by the Mann–Kendall
statistics (a) and linear regression (b) at the a = 0.05 significance levels in the Hanjiang basin.
the middle and lower Hanjiang basin and in the upper part
of the upper basin, while significant downward trends exist
in the lower region of the upper basin. Spring temperature
shows significant upward trends in the middle and lower basin and no trend in the upper basin. In summer there is no
trend in the middle and lower basin and significant downward trends in most parts of the upper basin. In autumn
there are significant upward trends in the lower basin, the
lower part of the middle basin and a small part of the upper
basin; a downward trend is found in a small area near the
Yunxian station. In winter temperature shows significant upward trends in most parts of the basin.
The above discussion shows that temperature has increased dramatically in the middle and lower basins in
all the seasons except for summer which is a rainy season. The significant increase in air temperature has resulted in a dramatic decrease in runoff through
increased evaporation (refer to Fig. 3c). The middle and
lower basins of the Hanjiang River is the economic center
of Hubei Province and rapid economic development experienced during last decades has increased air pollution in
the region, which, in turn, has partly caused temperature
increase and runoff decrease. The increasing trend in air
temperature over the past decades in the region is consis-
H. Chen et al.
0.001
0.115
0.909
0.42
0.02
2.48
0.02
2.32
0.003
0.49
0.629
0.20
0.001
0.27
0.787
0.04
0.07
0.49
0.624
0.81
0.59
0.75
0.46
1.09
Note: Ann. = Annual, Sp. = Spring, Su. = Summer, Au. = Autumn, Wi. = Winter.
b
B ¼ slope of linear regression, T = statistic of t-test, Zmk = statistic of Mann–Kendall.
The critical value Ta/2,n2 = T0.05/2,532 = 2.009.
The critical value Z1a/2 = Z10.05/2 = 1.64.
Significant trends are shown with bold italic values.
0.42
0.47
0.64
0.38
0.45
1.03
0.31
1.11
1.54
1.14
0.26
1.14
0.18
3.54
0.00
3.01
0.87
1.18
0.25
1.62
0.68
1.15
0.26
0.94
Wi.
Au.
Su.
Sp.
0.56
2.13
0.04
2.32
2.3
1.94
0.06
2.03
Wi.
Su.
Temperature
Sp.
Ann.
Wi.
Au.
Su.
Precipitation
Sp.
Ann.
Ann.
Runoff
The availability and variation of runoff in the Danjiangkou
reservoir basin in the future would be a key to the success
of the South-to-North Water Diversion Project. It is therefore necessary to analyze the impact of climate change on
the water resources in the Danjiangkou reservoir basin. To
assess the impact of global warming on water resources
and simulate runoff changes in the Danjiangkou reservoir
basin, monthly precipitation and temperature series are
obtained from different GCMs (HadCM3, CCSRNIES, CSRIO
and GFDL) grids which are nearest to the Danjiangkou reservoir basin. The predicted precipitation and temperature
anomalies from the 1961–1990 average in the Danjiangkou
Results of the trend test for the series of annual runoff, precipitation and temperature in the Danjiangkou reservoir basin
Impact of climate change on water resources in the
Danjiangkou reservoir basin
Table 4
Long-term trends in the runoff of the Danjiangkou reservoir
basin are tested by linear regression and Mann–Kendall
methods. To analyze the relationships among runoff, precipitation and temperature in the Danjiangkou reservoir basin,
Kendall’s s is applied to evaluate correlations among the annual runoff, precipitation and temperature series and the
long-term trends of the area–mean precipitation and temperature in the Danjiangkou reservoir basin are also tested.
^ the test statistics T, the
The values of the regression slope b,
p-value of statistics T and the Mann–Kendall statistics Zmk
are given in Table 4. At the a = 0.05 significance level there
are significant downward trends for annual runoff as seen by
applying the Mann–Kendall statistics: the jZmkj value =
2.03 > Z1a/2 = Z10.05/2 = 1.96, but no trend is detected by
the linear regression as the jTj values 1.94 < Ta/2,n2 =
T0.05/2,532 = 2.009 and the p-value 0.06 > 0.05, with the test
statistic lying close to the limit of the critical region. Both
testing methods show that runoff in spring and winter show
significant downward trends at the a = 0.05 significance le^ values for annual and
vel. Although the regression slope b
seasonal precipitation are all negative, indicating a linear
downward tendency, significant downward trends are not
detected by either of the methods at the a = 0.05 significance level. Both methods find that temperature has significant downward trends in summer and upward trends in
winter at the a = 0.05 significance level.
The Kendall s for these series are calculated and runoff is
more closely correlated with precipitation (s = 0.746) than
with temperature (s = 0.4) in the Danjiangkou reservoir
basin.
The results show that at the a = 0.05 significance level
the annual runoff has a significant downward trend in the
period of 1953–2003. If the dry years are successive, there
will be insufficient water for the receiving areas of the middle route of the SNWDP. It is therefore of interest to study
how the changes will continue in the future and predict the
impacts of climate changes of the Danjiangkou reservoir
basin.
Au.
Trends of runoff of Danjiangkou reservoir basin
detected by linear regression and Mann–Kendall
statistics
b
B
T
p-Value
Zmk
tent with global warming as reported across most of the
globe.
0.017
2.691
0.01
2.25
180
Historical temporal trends of hydro-climatic variables and runoff response to climate variability
reservoir basin are calculated and smoothed with a 11-year
low-pass filter and plotted in Figs. 6 and 7, respectively. In
the 21st century, temperature in the Danjiangkou reservoir
basin would rise and the results of different GCMs are coincident as shown in Fig. 6. However, precipitation trends of
different GCMs are discordant, as shown in Fig. 7. To test
the relationships between GCMs raw data and observed
data, the coefficient of determination R2 is calculated
and its values are given in Table 5. As the coefficient of
determination R2 for the precipitation of the GFDL model
is 0.43, only the HadCM3, CCSRNIES, CSRIO are selected
for simulating future climate change in Danjiangkou reservoir basin.
The data of the predicted temperature and precipitation changes are recalculated by using the delta change
Anomaly relative to 1961-1990
mean(°C)
8
CCSRNIES
HadCM3
GFDL
CSRIO
Observed
6
4
2
0
-2
1951
1971
1991
2011
2031
T(year)
2051
2071
Figure 6 The GCMs predicted temperature anomalies from
the 1961 to 1990 average in the Danjiangkou reservoir basin
from 1951 to 2100 smoothed with a 11-year low-pass filter.
CCSRNIE
HadCM3
GFDL
CSRIO
Observed
anomaly relative to the 19611990 mean(mm)
1.5
1
0.5
181
method (Hay et al., 2000) as input to the two-parameter
water balance model which had been used in the Danjiangkou reservoir basin and had simulated runoff well
(Xiong and Guo, 1999; Guo et al., 2002). The GCMs climate variables for the period of 1961–1990 with GCMs
simulated future values from 2021 to 2050 are compared
and the differences are added to the observed series from
1961 to 1990 to represent the future climate variables for
the period 2021–2050 as input to the two-parameter
water balance model. The runoff from the Danjiangkou
reservoir basin is predicted for 2021–2050 and the prediction results of mean monthly runoff are plotted in Fig. 8
and the relative mean change in the results are given in
Table 6. It is seen from Fig. 8 that the predicted monthly
runoff will increase in almost all the seasons, except in
autumn for the CCSRNIES predicted scenarios. This is
mainly due to the predicted increase in monthly precipitation by GCMs (see Table 6). The mean annual changes in
runoff are 8.18%, 7.78% and 2.14%, respectively, when
the scenarios predicted by HadCM3, CSRIO and CCSRNIES
are used as input to the water balance model. These runoff changes are the result of the combined effects of the
predicted precipitation and temperature changes. The
mean annual temperature increases by 1.29 C (HadCM3),
0.99 C (CSRIO) and 1.87 C (CCSRNIES) in the Danjiangkou
reservoir basin which reduces runoff by a certain percentage due to the resulting increase in evapotranspiration.
But the reduction in runoff caused by the increase in temperature cannot compensate for the increase in
precipitation.
In order to evaluate the effect of precipitation and
temperature separate on runoff and examine the sensitivity of the hydrological system to the climate change in the
region, a sensitivity analysis is carried out by using six
hypothetical climate change scenarios (DT = 0 combined
with DP = ±20%, ±10%; DP = 0 combined with DT = 1 and
2 C) as input to the water balance model. The results
of sensitivity analysis are plotted in Fig. 9. It is seen that
(1) 1 C and 2 C increases in temperature reduce the
mean annual runoff by about 3.5% and 7%, respectively.
(2) A decrease/increase of mean monthly precipitation
0
100
-0.5
-1
1951
HadCM3
CSRIO
CCSRNIES
Current
80
1971
1991
2011
2031
2051
2071
Figure 7 The GCMs predicted precipitation anomalies from
the 1961 to 1990 average in the Danjiangkou reservoir basin
from 1951 to 2100 smoothed with a 11-year low-pass filter.
Table 5 The determinations coefficient R2 between the
GCMs raw data series and the observed data series from 1951
to 2003 in the Danjiangkou reservoir basin
Precipitation
Temperature
HadCM3
CCSRNIES
CSRIO
GFDL
0.69
0.98
0.70
0.98
0.65
0.97
0.43
0.98
Runoff(108m3)
T(year)
60
40
20
0
1
2
3
4
5
6
7
T (Month)
8
9
10
11
12
Figure 8 Predicted results of mean monthly runoff from 2021
to 2050 in the Danjiangkou reservoir basin.
182
H. Chen et al.
Table 6 Future mean runoff change from 2021 to 2050 simulated by the two-parameter water balance model based on the
future climate change scenarios predicted by GCMs
HadCM3
January
February
March
April
May
June
July
August
September
October
November
December
Ann.
CSRIO
DP (%)
DT (C)
DQ (%)
DP (%)
DT (C)
DQ (%)
DP (%)
DT (C)
DQ (%)
42.22
18.20
10.70
0.42
5.18
14.33
6.83
6.61
12.87
1.77
15.60
4.20
10.48
1.02
0.64
0.77
1.42
1.40
1.27
1.17
1.83
1.70
1.60
1.41
1.24
1.29
13.12
10.96
20.07
7.49
11.54
9.38
13.17
5.15
9.25
3.54
7.04
4.86
8.18
1.89
6.33
10.37
21.44
4.42
0.42
5.74
0.09
10.69
0.41
3.48
3.37
5.39
0.94
1.32
1.42
1.81
1.11
0.70
0.63
0.62
0.32
0.44
1.30
1.21
0.99
0.18
0.06
5.93
20.62
14.07
9.53
14.79
8.41
12.7
4.53
1.53
1.06
7.78
1.38
14.83
19.34
13.56
9.47
0.41
3.47
7.96
9.09
16.38
12.75
22.58
6.08
2.59
1.49
2.81
1.60
1.32
1.73
1.56
1.56
1.82
1.69
2.06
2.25
1.87
4.85
1.64
4.72
16.57
18.29
4.18
3.25
1.86
7.75
10.77
7.74
5.98
2.14
Changes in runoff (%)
40
30
20
10
0
-10
1
2
CCSRNIES
3
4
5
6
7
8
9
10
11
12
13
-20
-30
-40
Months
DP=-20%, DT=0
DP=-10%, DT=0
DP=10%, DT=0
DP=20%, DT=0
DP=0%, DT=1
DP=0%, DT=2
Figure 9 Response of the mean monthly runoff to the
changes in precipitation and temperature in the Danjiangkou
reservoir basin.
of 20 and 10% decreases/increases the mean annual runoff
by about 30% and 15%, respectively. (3) For both changes
in precipitation and temperature the monthly changes of
runoff are quite constant over the year in this study
region.
The previous results and discussion show that
although the period of 1990–2003 was a very dry period
compared with the long-term average (Fig. 3c), runoff in
2021–2050 will increase relative to the mean of 1961–
1990 based on the climate scenarios (Fig. 8 and Table 6).
This is advantageous, especially for the middle route of
the South-to-North Water Diversion Project (SNWDP) in
China.
Conclusions
In this study, temporal trends of precipitation, temperature
as well as their spatial distributions in the Hanjiang basin
were examined. The temporal trends of runoff and the impact of climate change on runoff in the Danjiangkou reservoir basin were also analyzed. The following main
conclusions are drawn:
(1) Historical data show no trend for precipitation in most
parts of the Hanjiang basin at the a = 0.05 significance
level.
(2) Temperature trends indicate a large scale climate
warming, especially during winter when entering the
1990s. Temperature in winter has significantly increased
in the Hanjiang basin since 1991 and the extent of the
increase reaches to 1.09 C in the lower basin. The spatial distribution patterns of temperature trends vary
seasonally. The temperature trends in the middle and
lower basins are similar and are significant at the
a = 0.05 significance level; however the upper basin
behaves differently.
(3) The temporal trends of runoff are the result of the combined effect of precipitation and temperature. Significant decreasing trends are found for mean annual,
spring and winter runoff at the a = 0.05 significance
level.
(4) Climate change has posed a huge challenge to the
existing water resources management practices.
Although the projected changes in precipitation and
runoff through GCMs are scenario- and model-dependent and full of uncertainty, these climate impact
studies are necessary and useful for the final success
in achieving the satisfactory results in assessing the
impacts of climate change on water resources in Hanjiang basin. Potential future climate changes predicted
by GCMs in the 21st century for the period of 2021–
2050 show that both temperature and precipitation will
increase in the Danjiangkou reservoir basin. Simulated
results indicate that during 2021–2050 annual runoff
will increase by 8.18%, 7.78% and 2.14%, respectively,
when the scenarios predicted by HadCM3, CSRIO and
CCSRNIES are used as input to the water balance
model. Sensitivity analysis shows that 1 C and 2 C
increase in temperature reduce mean annual runoff
by about 3.5% and 7%, respectively. A decrease/
increase of mean monthly precipitation by 20 and 10%
decreases/increases mean annual runoff by about 30%
and 15%, respectively.
Historical temporal trends of hydro-climatic variables and runoff response to climate variability
(5) In 2010 the expanded Danjiangkou reservoir will provide
9.5 billion m3 of water annually to Beijing, Tianjin, and
cities in Hebei, Henan and Hubei provinces. It will be able
to provide 13–14 billion m3 by 2030. However the period
of 1990–2003 was a very dry period compared with the
long-term and in future the demand for water utilization
in Hanjiang basin will be more than before. The situation
of the water resources in Hanjiang basin would be critical
in future if the dry phenomenon as experienced in the
period of 1990–2003 persists. The results of this study
show that during 2021–2050 annual runoff will increase
relative to the period 1960–1990. Engineering measures
will be taken to transfer a certain amount of water from
the Yangtze River to the Hanjiang River to keep an adequate discharge during the dry period. Thus, there will
be adequate water resources to keep the sustainable
development of Hanjiang basin and satisfy the water
demand of the middle route of SNWDP. The results of this
study will provide a scientific reference for the water
allocation of the middle route of SNWDP in China.
Acknowledgements
The study is financially supported by the National Natural
Science Fund of China (50679063), International Cooperation Research Fund of China (2005DFA20520) and the Ministry of Education of China (104204). The authors are very
grateful to the National Climate Centre, the Bureau of
Hydrology of the Changjiang Water Resources Commission
and the IPCC for providing valuable climatic and hydrological data. The authors thanks for the very valuable comments from reviewers which greatly improved the quality
of the paper.
References
Arnell, N.W., 1992. Factors controlling the effects of climate
change on river flow regimes in a humid temperate environment.
J. Hydrol. 132, 321–342.
Arora, M., Goel, N.K., Singh, P., 2005. Evaluation of temperature
trends over India. Hydrol. Sci. J. 50 (1), 81–93.
Aziz, O.I.A., Burn, D.H., 2006. Trends and variability in the
hydrological regime of the Mackenzie River Basin. J. Hydrol.
319 (1–4), 282–294.
Becker, S., Gemmer, M., Jiang, T., 2006. Spatiotemporal analysis of
precipitation trends in the Yangtze River catchment. Stochastic
Environ. Res. Risk Assess. 20 (6), 435–444.
Burn, D.H., Elnur, M.A.H., 2002. Detection of hydrologic trends and
variability. J. Hydrol. 255, 107–122.
Burn, D.H., Cunderlik, J.M., Pietroniro, A., 2004. Hydrological
trends and variability in the Liard river basin. Hydrol. Sci. J. 49
(1), 53–67.
Chen, X.Q., Zhang, D.Z., Zhang, E.F., 2002. The south to north
water diversions in China: review and comments. J. Environ.
Plan. Manage. 45 (6), 927–932.
Gemmer, M., Becker, S., Jiang, T., 2004. Observed monthly
precipitation trends in China 1951–2002. Theor. Appl. Climatol.
39, 21–37.
Gleick, P.H., 1986. Methods for evaluating the regional hydrologic
impacts of global climatic changes. J. Hydrol. 88, 97–116.
Guo, S.L., Wang, J.X., Xiong, L.H., Ying, A.W., 2002. A macro-scale
and semi-distributed monthly water balance model to predict
climate change impacts in China. J. Hydrol. 268, 1–15.
183
Guo, S.L., Chen, H., Zhang, H.G., Xiong, L.H., Liu, P., Pang, B.,
Wang, G.Q., Wang, Y.Z., 2005. A semi-distributed monthly
water balance model and its application in a climate change
impact study in the middle and lower Yellow River basin. Water
Int. 30 (2), 250–260.
Hay, L.E., Wilby, R.L., Leavesley, G.H., 2000. A comparison of delta
change and downscaled GCM scenarios for three mountainous
basins in the United States. J. Am. Water Resour. Assoc. 36 (2),
387–397.
Hirsch, R.M., Slack, J.R., Smith, R.A., 1982. Techniques of trend
analysis for monthly water quality data. Water Resour. Res. 18
(1), 107–121.
Huang, M., Peng, G.B., Leslie, L.M., Shao, X.M., Sha, W.Y., 2005.
SeasonalandregionaltemperaturechangesinChinaoverthe50 year
period 1951–2000. Meteor. Atmos. Phys. 89 (1–4), 105–115.
IPCC Report, 2001. Climate Change 2001: The Scientific Basis.
Cambridge University Press, Cambridge, pp. 140–165.
Kahya, E., Kalaycı, S., 2004. Trend analysis of streamflow in Turkey.
J. Hydrol. 289, 128–144.
Kendall, M.G., 1938. A new measure of rank correlation. Biometrika
30, 81–93.
Kendall, M.G., 1975. Rank Correlation Methods. Charles Griffin,
London.
Kundzewicz, Z.W., Robson, A.J., 2004. Change detection in hydrological records – a review of the methodology. Hydrol. Sci. J. 49
(1), 7–19.
Liu, C.M., Zheng, H.X., 2002. South-to-north water transfer
schemes for China. Int. J. Water Resour. Dev. 18 (3), 453–471.
Mann, H.B., 1945. Non-parametric test against trend. Econometrica
13, 245–259.
Mimikou, M., Kouvopoulos, Y., Cavadias, G., Vayianos, N., 1991.
Regional hydrological effects of climate change. J. Hydrol. 123,
119–146.
Ng, H.Y.F., Marsalek, J., 1992. Sensitivity of streamflow simulation
to changes in climatic inputs. Nordic Hydrol. 23, 257–272.
Schaake, J.C., Liu, C., 1989. Development and applications of
simple water balance models to understand the relationship
between climate and water resources. IAHS Publ. No. 181, pp.
343–352.
Shao, X.J., Wang, H., Wang, Z.Y., 2003. Interbasin transfer projects
and their implications: a China case study. Int. J. River Basin
Manage. 1 (1), 5–14.
Sprent, P., 1990. Applied Nonparametric Statistical Methods.
Chapman and Hall, London.
van Belle, G., Hughes, J.P., 1984. Nonparametric tests for trends in
water quality. Water Resour. Res. 20 (1), 127–136.
Wang, L.S., Ma, C., 1999. A study on the environmental geology of
the Middle Route Project of the south–north water transfer.
Eng. Geol. 51 (3), 153–165.
Wang, S.W., Zhu, J.H., Cai, J.N., 2004. Interdecadal variability of
temperature and precipitation in China since 1880. Adv. Atmos.
Sci. 21 (3), 307–313.
Xiong, L., Guo, S., 1999. Two-parameter water balance model and
its application. J. Hydrol. 216, 111–123.
Xiong, L., Guo, S., 2004. Trend test and change-point
detection for the annual discharge series of the Yangtze
River at the Yichang hydrological station. Hydrol. Sci. J. 49
(1), 99–112.
Xu, C.Y., 1999. From GCMs to river flow: a review of downscaling
techniques and hydrologic modeling approaches. Prog. Phys.
Geog. 23 (2), 229–249.
Xu, C.Y., Halldin, S., 1997. The effect of climate change on
river flow and snow cover in the NOPEX area simulated by a
simple water balance model. Nordic Hydrol. 28 (4/5), 273–
282.
Xu, C.Y., Singh, V.P., 1998. A review on monthly water balance
models for water resources investigation and climatic impact
assessment. Water Resour. Manage. 12, 31–50.
184
Xu, Y.L., Zhang, Y., Lin, E.D., Lin, W.T., Dong, W.J., Jones, R.,
Hassell, D., Wilson, S., 2006. Analyses on the climate change
responses over China under SRES B2 scenario using PRECIS. Chin.
Sci. Bull. 51 (18), 2260–2267.
Yang, H., Zehnder, A.J.B., 2005. The south–north water transfer
project in China – an analysis of water demand uncertainty and
environmental objectives in decision making. Water Int. 30 (3),
339–349.
Yue, S., Pilon, P., Cavadias, G., 2002. Power of the Mann–Kendall
and Spearman’s rho test for detecting monotonic trends in
hydrological series. J. Hydrol. 259, 254–271.
Yue, S., Wang, C.Y., 2002. Applicability of the pre-whitening to
eliminate the influence of serial correlation on the Mann–
Kendall test. Water Resour. Res. 38 (6). doi:10.1029/
2001WR00086.
Yue, S., Pilon, P., 2004. A comparison of the power of the t-test,
Mann–Kendall and bootstrap tests for trend detection. Hydrol.
Sci. J. 49 (1), 21–37.
H. Chen et al.
Zetterqvist, L., 1991. Statistical estimation and interpretation of
trends in water quality time series. Water Resour. Res. 27 (7),
1637–1648.
Zhang, Q., Jiang, T., Gemmer, M., Becker, S., 2005. Precipitation,
temperature and runoff analysis from 1950 to 2002 in the
Yangtze basin, China. Hydrol. Sci. J. 50 (1), 65–80.
Zhang, Q., Liu, C., Xu, C.Y., Xu, Y.P., Jiang, T., 2006.
Observed trends of water level and streamflow during past
100 years in the Yangtze River basin, China. J. Hydrol. 324
(1–4), 255–265.
Zhang, X.B., Harvey, D.H., Hogg, W.D., Yuzyk, T.R., 2001. Trends in
Canadian streamflow. Water Resour. Res. 37 (4), 987–998.
Zhou, T.J., Yu, R.C., 2006. Twentieth-century surface air temperature over China and the globe simulated by coupled climate
models. J. Climate 19 (22), 5843–5858.
Zhu, Y., Day, R.L., 2005. Analysis of streamflow trends and the
effects of climate in Pennsylvania, 1971 to 2001. J. Am. Water
Resour. Assoc. 41 (6), 1393–1405.