CC BY-NC-ND
0Contents lists available at ScienceDirect
Journal of Hydrology: Regional Studies
journal homepage www.elsevier.com/locate/ejrh
HYDRDLDGY: I :?--?.'! !.''.'•" kH'J ll-ll
Analysis of streamflow variations in the Heihe River Basin, northwest China: Trends, abrupt changes, driving factors and ecological influences
Aijing Zhang3, Chunmiao Zhenga b *, Sai Wanga, Yingying Yaoa
a Center for Water Research, College of Engineering, Peking University, Beijing, China b Department of Geological Sciences, University of Alabama, Tuscaloosa, AL, United States
ARTICLE INFO
Article history:
Received 13 April 2014
Received in revised form 20 October 2014
Accepted 25 October 2014
Keywords:
Streamflow changes Trends test Abrupt change test Heihe River Basin Inland River Basin
ABSTRACT
Study Region: The Heihe River Basin (HRB) is the second largest inland river basin of China located in northwest China with a minor portion in Mongolia. Study Focus: With increasing water demands from domestic, agricultural, and industrial sectors, the HRB has been increasingly undergoing water resources shortage and eco-environmental degradation, especially in the lower HRB. Discerning the trends and any abrupt changes in the streamflow over the river basin could help unravel the causes and effects of historical variations of the water resources. Statistical methods for detecting trends, abrupt and gradual changes were applied to the long-term streamflow, precipitation and temperature data over the HRB to analyze and understand annual and seasonal hydroclimatic variations over the past five decades. New Hydrological Insights for the Region: The findings of this study indicated that although the streamflow coming from the upper reaches have risen, those flowing to the lower reaches have declined significantly. Analysis ofthe correlation between climatic factors and streamflow variations and the assessment of the development of socioeconomics revealed that: (1) rising temperature and precipitation are the main cause to explain the increases in streamflow in the upper HRB; and (2) human activities in the middle reaches of the HRB rather than climate changes were primarily responsible for the water shortage and ecological deterioration ofthe lower HRB. © 2014 The Authors. Published by Elsevier B.V. This is an open access article
under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).
* Corresponding author at: Center for Water Research, Peking University, Beijing 100871, China. Tel.: +86 10 6276 7687. E-mail address: czheng@pku.edu.cn (C. Zheng).
http://dx.doi.org/10.1016Zj.ejrh.2014.10.005
2214-5818/© 2014 The Authors. Published by Elsevier B.V. This is an open access article underthe CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).
1. Introduction
Inland river basins in China take up approximately one third of the national territory. They are mainly distributed in the northwest with an arid or semi-arid climate and fragile ecosystem (Wang and Cheng, 2000; Cheng et al., 2006). For tens of thousands of years, these inland rivers provide people with water, food, shelter and spiritual connection. However, in recent decades, water problems have become a principal challenge that threatens socioeconomic development and ecological health due to over exploitation and unreasonable use of water resources (Wang and Cheng, 2000; Cheng et al., 2006; Xu et al., 2010; Zhang et al., 2012a,b; Chen et al., 2013).
As the second largest inland river basin of China, the Heihe River Basin (HRB) (as shown in Fig. 1) is under constant water and ecological stresses with terminal lakes drying-up, water table decline, grassland degeneration, and widespread desertification, due to the impact of climate change and human activities (Zhu et al., 2005; Hu et al., 2007; Zhang et al., 2011; Wang et al., 2013; Min et al., 2013). Specific measures have been undertaken over years to protect and restore the deteriorated ecosystems in the HRB. For example, ecological protection projects such as returning grazing land to grassland and conserving public forests have been carried out over the HRB since the late 1990s; Stringent water conservation measures have been implemented in the Zhangye area as a pilot project since 2002 (Kang et al., 2007). In particular, an Ecological Water Diversion Project (EWDP) was initiated by the Chinese government since 2000 to ensure the delivery of a minimum amount of water supply to lower reaches for ecological water needs. The EWDP project has enjoyed some success with the return of water to the terminal lakes and recovery of the riparian vegetation (Zhao et al., 2007; Wang et al., 2011a,b; Zhang et al., 2011).
Fig. 1. Location of the Heihe River Basin.
In support of eco-environmental protection and restoration, numerous studies have been carried out in the HRB in recent years. These studies contain quantity and quality analysis on the surface water and groundwater resources (Qin et al., 2011; Cao et al., 2012; Wu et al., 2014), evaluation of the human activity and climate change impacts on the eco-hydrological processes of the HRB (Wang et al., 2005a,b; Zanget al., 2013; Qin et al., 2013), elucidation of effective water resources management policies (Chen et al., 2005), integrated remote sensing for comprehensive watershed observations (Li et al., 2013), development of hydrological models for understanding the water cycle and associated ecological processes in the inland basin (Hu et al., 2007; Zhou et al., 2011; Guo et al., 2012; Yin et al., 2012; Wei et al., 2013; Zheng et al., 2013). Since 2010, a major research initiative has been launched for an integrated ecological-hydrological-economic study of the HRB to provide a stronger scientific underpinning for sustainable water management (Zheng et al., 2012; Yao et al., 2014).
Trend and abrupt change detection of the hydrologic time series can help us understand the causes of historic changes (Rouge et al., 2012) and offer more insights to water resource management and ecological conservation. Many studies have discussed the streamflow changes in the HRB over the last half century (Li et al., 2012; Zou and Zhang, 2012). However, there are some deficiencies for the existing studies: (1) most of the previous researches focused only on the streamflow changes at two gaging stations (Yingluoxia and Zhengyixia; see Fig. 1) on the main stream of Heihe River with few, if any, detailed analysis on the streamflow variations at other stations or along tributaries; (2) streamflow series data have not been updated such that streamflow changes before and after the Ecological Water Diversion Project could not be analyzed; and (3) driving factors and ecological influences of the streamflow variations were not fully explored.
Thus, the primary aim of this study is (1) to analyze temporal variations of the streamflow over the HRB, detect abrupt changes and trends if present; (2) to discern the main driving factors for the observed streamflow changes; and (3) to elucidate the ecological and environmental problems caused by over exploitation of water resources in the past. The paper is structured as follows. After this introduction, Section 2 describes the study site and datasets available for this study. Section 3 discusses the methodology used in the analysis. Section 4 presents the results of streamflow analysis in terms of trends and abrupt changes. Section 5 provides a discussion of the results in the context of climate change and human activities. Finally, Section 6 concludes with a summary of the main findings and insights from this study.
2. Study area and datasets
2.1. Study area
The Heihe River Basin (HRB) is located in the northwest of China with a minor portion in Mongolia (Fig. 1). The core drainage area is approximately 130,000 km2 with a mainstream length of 821 km. Its geographical range extends from 37°41' to 42°42'N and 96°42' to 102°00' E. The HRB includes three sections from south to north: upstream from the Qilian Mountains to the Yingluoxia Canyon (outlets of the mountains), midstream running from the Yingluoxia Canyon to Zhengyixia Canyon, and downstream terminating in the Juyan Lakes (east and west branches, respectively). This region is characterized by a continental climate. Depending on the location, the average annual air temperature is 2-3 °C in the upper HRB, 6-8 °C in the middle HRB, and 8-10 °C in the lower HRB. The average annual precipitation is 200-500 mm, 120-200 mm and less than 50 mm in the upstream, midstream and most downstream regions, respectively (Qi and Luo, 2005). From southern mountain region to the northern Gobi desert, potential evapotranspiration ranges from 500-4000 mm per year.
The HRB has a distinct landscape, ecological and climate gradient from the upstream to downstream. The upstream is characterized by the mountainous terrains from Qilian Mountains to Yingluoxia Canyon. Most of the streamflow in the Heihe River and its tributaries are generated from rainfall and ice-snow melting in the upstream mountainous area (Wang et al., 2010). The midstream, from Yingluoxia Canyon to Zhengyixia Canyon, is characterized by oases with irrigated agriculture. It is the major zone of water consumption by human and agriculture. The downstream is characterized by a vast Gobi desert where the runoff is greatly reduced or disappears through evapotranspiration and river leakage.
Table 1
Information of the hydrological stations.
Station River Lon. (E°) Lat. (№) Areaa (km2) Runoff (108 m3) Data period
Qilian (QL) Heihe River 100.23 38.20 2452 4.57 1968-2010
Zhamashenke (ZM) Heihe River 99.98 38.23 4986 7.16 1957-2010
Yingluoxia (YL) Heihe River 100.18 38.82 10,009 15.84 1945-2012
Gaoai (GA) Heihe River 100.40 39.13 20,299 10.34 1977-2010
Zhengyixia (ZY) Heihe River 99.42 39.79 35,634 10.17 1957-2012
Lijiaqiao (LJ) Xida River 101.13 38.52 1143 0.66 1956-2009
Shuangshusi (SS) Hongshui River 100.83 38.32 578 1.42 1956-2007
Wafangcheng (WF) Dazhuma River 100.48 38.43 229 0.89 1956-2009
Sunan (SN) Liyuan River 99.63 38.84 1080 1.76 1962-2010
Liyuan (LY) Liyuan River 100.00 38.97 2240 2.43 1956-2009
Hongsha(HS) Hongsha River 99.20 39.18 619 1.11 1956-2009
Xindi (XD) Hongshui River 98.42 39.57 1581 2.71 1956-2009
Yuanyangchi (YY) Taolai River 98.84 39.91 12,439 3.18 1978-2009
Shaomaying (SM) Heihe River 99.96 40.75 - - 2000-2012
Langxinshan (LX) Heihe River 100.36 41.08 - - 2000-2012
Juyan lake (JY) Heihe River 101.11 42.21 - - 2003-2012
a Catchment area upstream of the hydrological station.
Over the past half century, with the rapid population growth, socioeconomic development and climate change, ecological and environment problems associated with unimpeded water resource exploitation have continued to worsen from year to year. In the upstream, the quality of grassland resources has declined sharply due to over-grazing; the glaciers and snowpack have been shrinking because of climate warming. Pushed by the traditional economic planting structure and development model that emphasizes GDP growth over eco-environmental quality, the water demand and consumption in the midstream areas have steadily increased, leaving less and less water for the downstream. Consequently, in the lower HRB, due to water shortage, the extent of oasis has shrunk and health of the groundwater dependent ecosystem has deteriorated. The terminal lakes were dried up until 2002, two years after the EWDP was implemented by the government. It is clear that a sound policy for allocation of precious water resources based on hydrological, ecological, socioeconomic, and sociopolitical realities are urgently needed for the HRB. This study will support effective policy formulation by shielding lights on the trends and abrupt changes in the historical streamflow data for the entire HRB.
2.2. Data sets
Daily precipitation and mean temperature data from 17 National Meteorological Observatory stations (Fig. 1), with continuous data from 1960 to 2012 in or around the HRB were used for this study. These stations, which possess high quality data, are maintained and released according to the standards set by the National Meteorological Administration of China (http://cdc.cma.gov.cn/home.do). Monthly observed streamflow data of 16 hydrological stations (Table 1) were collected from Hydro-logical Bureau of Gansu Province and the Inner Mongolia Autonomous Region, which are also of high quality. Streamflow series of the upper and middle HRB (the first 13 stations) are used to analyze streamflow variations, and that of the last three stations are only used to detect the inflow changes to the downstream for lacking of long term records. A few missing data were filled based on nearby stations and a correlation analysis between individual stations.
3. Methodology
This study aims to detect the presence of trends and abrupt changes of the streamflow time series over the HRB. To analyze driving factors of the streamflow change, trends and abrupt changes in the data of the meteorological series were also tested. Throughout this study, two types of statistical
analysis methodology were used: a trend test (Mann-Kendall test) and a change-point test (Pettitt test).
3.1. Trend test
A trend test is performed on the hydrological and meteorological data to analyze gradual changes or tendencies. The Mann-Kendall test (Mann, 1945; Kendall, 1975) is one of the most popular trend detection method used in the world. It is a non-parametric test which can cope with missing values and values below a detection limit. For an independently distributed time seriesX(n), null hypothesis (H0) of the Mann-Kendall (MK) test is no trend.
In the MK test, the sign (sgn) is used to count the difference between two values (x,- and Xj) from X(n) which is defined as:
(1 if Xj > Xj
0 if Xj =Xj (1)
-1 if Xj < Xj
S which is defined as total sgn of the whole time series is measured as:
S = sgn(Xj -Xj) (2)
1 <i<j<n
For n >8, if H0 is valid, S is approximately normally distributed with the following mean and variance:
E(S) = 0 (3)
(n(n - 1)(2n + 5) - emfmm(m - 1)(2m + 5)) V(S) 18 (4) In Eq. (4), tm is the number of data in a tied group (there is a tie when Xj = x,) and m is the number of tied groups. The standardized test statistics Z follows a standard normal distribution:
( (S -1)/^^ if S > 0 Z =1 0 if S = 0 (5)
1 (S + 1)/yV(S) if s < 0
H0 will be rejected at the significance level of a when the absolute value of Z is bigger than Z(1-(a/2)). Serial correlations can affect the results of MK test (Yue et al., 2002; Hamed, 2009), therefore correlations of the series were computed firstly before a trend test. When only a lag-1 autocorrelation was found to be significant, the MK test of Yue et al. (2002) was used. For all the other significant correlations, the method proposed by Hamed and Rao (1998) was employed. In addition, the magnitude of a trend was also estimated by the method of Hirsch et al. (1982) extended from Sen (1968).
3.2. Abrupt changes test
The Pettitt test (Pettitt, 1979) is also a non-parametric test. It arbitrarily splits a time series into two sub-samples and implement a rank-based comparison between them. For a time series X(n), the separated two sub-samples before and after the date t, Pettitt statistics k(T) can be computed as follows:
k(t) = sgn(Xj - Xj) (6)
i=1 j=t+ 1
/--^^ /«JY ] /-/ "JY
/ jm J 7 ) / fsu J — Boundary — Mainstream IH Lake ■ H-station of downstream
JXD v > YY J*DT V-v
1 ▼ \ ^ HS Irrigation Area
X ijl, < Zjr^v/ By 2000 v I^XFX y Entire series V Significant downward ▼ Downward a Upward A Significant upward
Fig. 2. Results of the MKtest for annual streamflow (a = 0.05).
where sgn is defined as in Eq. (1). The abrupt change most likely takes place at the date t where the absolute value of k(t) reaches the maximum. Therefore, the final Petitt statistics K and time of the abrupt change Tare introduced as follows:
T = arg max(|k(t)|) (7)
K = max(|k(t )|) (8)
The significance probability associated with the rejection of the assumption that there is no change is approximated by:
p *2exp((9)
Pettitt test reports the greatest likely change point in a time series. In this study the two-sample t-test was also used to determine if the two sets, before and after the detected change point, are significantly different from each other. The hydrometeorological series is identified to exhibit a significant abrupt change only when the result of t-test is true.
4. Results
Trends of the seasonal and annual streamflow series from the gaging stations located in the upper and middle HRB were tested using the MK test. To discuss the streamflow response to the change in climate factors, trends ofthe annual and seasonal precipitation and mean temperature series were also analyzed by the MK test. Significance level of a = 0.05 and a = 0.01 were used in the MK test. Abrupt changes of the annual streamflow, precipitation and mean temperature series were detected based on the Pettitt method with a significance level of a = 0.05.
4.1. Trends of the streamflow series
4.1.1. Streamflow trends in the upper and middle HRB
Because the EWDP on the mainstream of Heihe River was initiated in 2000 which significantly altered the streamflow distribution in the middle and lower HRB, we computed the trends of the streamflow series both to 2000 and to the present. Figs. 2 and 3 depict the results of the MK test of annual streamflow data for the two series, one labeled "By 2000" and the other "Entire series".
For the annual streamflow series up to 2000, a significant trend was detected on only two stations located on the mainstream. One is the Qilian station (QL) in the upper stream where a significant upward trend was found (marked as a larger upward triangle in red in Fig. 2) with a Z-value of 2.12 (see Fig. 3), the other is Zhengyixia station (ZY) where a significant downward trend was identified (marked as a larger downward triangle in green in Fig. 2) with a Z-value of-2.87 (see Fig. 3). Trends of annual streamflow for all the other stations are generally insignificant. However, there is an obvious change in the characteristic of the streamflow data from different reaches: the trend is slightly upward
Fig. 3. MK test Z-values for annual streamflow.
for the gaging stations located upgradient to the irrigation area in the middle HRB. However, the trend is slightly downward for the gaging stations down gradient to the irrigation area. This indicates flowing through the irrigation area caused the trend of streamflow to reverse.
Difference appeared in the MK test results for the time series to the present. The significant upward trend was found in Qilian, Yingluoxia and Sunan stations with Z-values of 2.37, 2.87 and 2.78, respectively. It is an indication that a rising trend of the streamflow in the upper HRB is becoming more pronounced. On the contrary, the declining trend of the streamflow in Zhengyixia station is lessened with a Z-value of -1.58. EWDP on the mainstream should be the reason induced decreasing trend of the streamflow in Zhengyixia station slow down. In the eastern tributaries, declining trend of the streamflow for Lijiaqiao station becomes more significant. For other stations, the trend of streamflow data does not change much.
Results of the MK test for seasonal streamflow data are shown in Fig. 4. Fig. 4(a) depicts the trends for the data series up to 2000, while Fig. 4(b) depicts the trends up to the present. In both Fig. 4(a) and (b), the four panels, from left to right, show the data for spring, summer, autumn, and winter streamflow, respectively.
For the time series up to 2000, the general pattern of streamflow changes remains fairly consistent from season to season. That is, most of the upper stream stations show an upward trend; while most of the middle stream stations show a downward trend (see Fig. 4(a)). However, it appears that the more significant increase, in terms of both the number of stations and the magnitude of increase, occurred in the summer and winter.
For the time series up to the present, the overall upward and downward trends from season to season are similar to those shown by the time series up to 2000. One obvious difference is more significant upward trend for upstream stations (Fig. 4(b)). This may be an indication of more flow generation in the upper stream due to more glacier thawing and snowmelt as a result of climate warming.
4.1.2. Inflow to the downstream of HRB
Only a linear trend was used to test changes of the streamflow released to the lower HRB (as measured at the Zhengyixia gaging station). Generally speaking, from 2000 to 2012, streamflow released to the downstream has been increasing (see Fig. 5), and annual streamflows ofZhengyixia (ZY), Shaomay-ing (SM), and Langxinshan (LX) stations have the same pattern of variation. In the process of increasing, there is a low point in 2004 for the ZY, SM, and LX stations because of the dry year. After 2005, the water quantity released to the downstream has remained relatively stable. As a result of water use in the Ejina Oasis and the EWDP project, streamflow of the Juyan Lake (JY) station, representative of the surface water entering the East Juyan Lake, is changing in a different way than other stations. For example, in the wet year of 2008, only 0.11 x 108 m3 of water was delivered to the East Juyan Lake although the water flow through the LX station was at a relatively high rate. The mean annual streamflow of ZY, SM, LX and JY stations, progressively further downstream, is 10.1 x 108, 6.5 x 108, 5.3 x 108 and 0.5 x 108 m3, respectively.
Boundary — mainstream H Lake Irrigation Area V Significant downward »Downward A Upward A Significant upward
(b) Entire series Fig. 4. Results of the MK test for seasonal streamflow (a = 0.05).
On a monthly scale, the streamflows at Zhengyixia, Shaomaying and Langxinshan stations also have a similar temporal distribution (Fig. 6). Streamflow is concentrated during July to October, taking up more than 50% of the annual total. Streamflow for May, June and November are very low. Almost only from July to October, the flow can reach the East Juyan Lake.
4.2. Abrupt change of the streamflow
A change point indicates the starting time of the abrupt change in streamflow. Those of the annual streamflow series in the upper and middle HRB were first detected based on the Pettitt method. Then
—^—Zhengyixia —o— Shaomaying Langxinshan Juyan Lake
---Trend line(ZY)
---Trendline(SM)
---Trendline(LX)
---Trendline(JY)
Fig. 5.
Variations of the annual streamflow transported to the lower HRB.
Fig. 6. Variations of the monthly streamflow for the downstream.
two-sample t-test was used to determine if the means of the two populations before and after the change point are significantly different.
Significant streamflow abrupt changes were found for eight out of 13 stations in the upper and middle HRB (Table 2). Significant upward abrupt changes are found for five stations located in the upper HRB. Of them, starting times of three are around the year 1980 and two at the year of 2001. Streamflow of three stations in the middle HRB shows downward abrupt changes: one is Zhengyixia on the mainstream (1979), one is Xindi station on one of the western tributaries (1972), and the other is Lijiaqiao station on one of the eastern tributaries (1990). The upper HRB is affected by relatively few human activities, thus the upward abrupt changes of the streamflow are most likely to have been caused by climate change. However, the downward abrupt changes in the middle HRB stations have been caused by both climate change and human activities.
4.3. Streamflow difference between Yingluoxia and Zhengyixia stations
The Yingluoxia station sits at the junction between the upper and middle HRB, whose streamflow represents nearly all the water resources of the entire HRB, since most of the flow is generated in the upper stream of the HRB from precipitation and snowmelt. Streamflow of the Zhengyixia station, which is located at the transition point between the middle and lower HRB, represents the water resources available for the lower HRB. The streamflow difference between Yingluoxia and Zhengyixia stations is close to the total water consumption in the middle HRB. Analysis of water consumption intensity in the middle HRB can yield a better understanding of decreasing streamflow at the Zhengyixia station. The annual streamflow variation and difference of the two stations are shown in Figs. 7 and 8.
MK test results of the streamflow difference between Yingluoxia and Zhengyixia stations for the time series up to 2000 and to the present are nearly the same, with the Z-value of 5.83 and 5.86, respectively. A significant upward abrupt change is found in 1982 for the streamflow difference series. The average annual water consumptions before and after 1982 are 4.01 x 108 m3 and 7.32 x 108 m3,
Table 2
Abrupt change points of the streamflow.
Station Pettitt t-test Station Pettitt t-test
Change point P t Abrupt change Change point P t Abrupt change
QL 1982 0.02 -4.39 S WF 2001 0.30 -2.16 S
ZM 2001 0.31 -2.31 S SN 1980 0.09 -2.65 S
YL 1979 0.04 -2.65 S LY 1989 0.32 1.58 NS
GY 1990 0.40 1.94 NS HS 1972 0.24 1.74 NS
ZY 1984 0.08 2.81 S XD 1972 0.05 2.5 S
LJ 1990 0.00 3.77 S YY 1985 0.39 1.91 NS
SS 1990 0.99 0.67 NS
S: significant abrupt change; NS: no significant abrupt change.
Fig. 7. Variations of annual streamflow for the Yingluoxia and Zhengyixia stations.
respectively. The results indicate that water consumption of the midstream region has been growing significantly, and the abrupt increase started in the early 1980s.
Streamflow difference between Yingluoxia and Zhengyixia stations is characterized by four distinct stages according to the variation of the five year moving average (see Fig. 8), namely, stage 1: steadily decreasing (1957-1974); stage 2: steadily increasing (1975-1999); stage 3: variably decreasing (2000-2005); and stage 4: variably increasing (2006-2012). It is still difficult to give a clear explanation to the decreasing trend for stage 1, but it is possible that the dry period, coupled with the absence of an effective water conservancy project, is the reason. The increasing trend for water consumption in the middle HRB during stage 2 is obviously due to the socioeconomic development. After the initiation of the EWDP on the main stream of Heihe River in 2000, water consumption was controlled in stage 3.
During the third stage, to ensure water supply to the lower HRB in low-flow years, less water is used in the middle HRB such that a valley point can be seen in 2004. In stage 4, water consumption has been rising again, although water use has been restricted due to the EWDP. The EWDP sets rules for the minimal water release to the downstream through the Zhengyixia station but not the amount of water available in the middle HRB. It causes more water to be used in the middle HRB during the wet years, and explains the rising water consumption in stage 4. Drought and wetness is the dominant factor of water consumption in the middle HRB after the implementation of EWDP. In contrast, water released to the downstream through the Zhengyixia station is relatively stable.
4.4. Trends of precipitation and temperature series
The annual precipitation and temperature time series and their MK test results in the upper, middle and lower HRB for the last 53 years (1960-2012) are shown in Fig. 9. The graphs on the left in Fig. 9 are for precipitation data while those on the right are for temperature data.
12 10 8 6 4
—o— Yingluoxia-Zhengyixia
---5 year moving average i- ■ k
J ji //y fy, / *» K^ffe 11
n i A y \
Vvrr^Vv y /K y ^^2004
1957 1962 1967 1972 1977 1982 1987 1992 1997 2002 2007 2012 Year
Fig. 8. Variations of the streamflow difference between Yingluoxia and Zhengyixia stations.
Fig. 9. Precipitation and temperature time series and their MKtest statistics forthe HRB from 1960 to 2012.
For precipitation, it can be seen that there has been a significant increasing trend in the upstream areas (with MK test Z-value of 2.35), a less prominent increasing trend in the midstream areas (with MK test Z-value of 1.63) and essentially no increasing trend in the downstream areas (with MK test Z-value of 0.69). Decadal variability of precipitation indicates that there is a most obvious wet period for the upstream areas during 2003-2012, but none for the midstream and downstream areas.
For temperature, the MK test results show that the climate of the HRB has been getting warmer during the last 53 years. There was an oscillation of the mean annual temperature before 1997, but thereafter the annual temperature was always higher than the long-term mean temperature. The year of 1968 was the coldest year for the last 53 years. The statistically significant increasing trend was found since 1993 in the upper HRB, since 1991 in the middle HRB, and since 1990 in the lower HRB (a = 0.05 level). Over the whole period, the Z-value was 5.6, 5.5 and 5.3 in the upstream, midstream and downstream areas, respectively. These large Z-values imply a high level of warming trend.
The MK test results of seasonal precipitation and temperature variations in the upper, middle and lower HRB from 1960 to 2012 are shown in Fig. 10. In the upstream areas, the MK test analysis shows significant increasing in precipitation for the summer. Therefore, the increase of precipitation in summer was the most important reason for annual precipitation rising in the upstream areas. In the midstream and downstream areas precipitation in the winter shows the most obvious increasing trend compared to other seasons. Temperature increased significantly for all seasons at the a = 0.01 level. The highest increasing trend in the upstream areas occurred in the autumn and winter with Z-value of 5.82, while in the downstream areas the highest increasing trend occurred in the summer with Z-value
Fig. 10. MKtest Z-values of the seasonal precipitation and temperature.
of 6.53. However, in the midstream areas, the Z-values for all four seasons were approximately the same, at 3.55, implying a constant increasing trend within the year.
Fig. 11 (a) shows trends of the annual precipitation and mean temperature spatially. Among the 17 stations, precipitation for only three stations located in the upstream indicates a significant upward trend at the significant level of a = 0.05. Trends of the precipitation are insignificant for the other meteorological stations. Among them, four stations show a slight decreasing trend (one outside the upstream and three in the downstream). For the annual mean temperature, all 17 stations show statistically significant increasing trends with Z-value changes ranging from 3.85 to 6.29.
The magnitude of precipitation and temperature changes is shown in Fig. 11(b). On average, the precipitation has increased by about 6-9 mm/decade in the upper HRB, and 3-6 mm/decade in the middle HRB. In the downstream region, the precipitation has decreased by -0.71 mm/decade in the northwest. For temperature, the magnitude of the increasing tread ranges from 0.30 °C/decade in the southwest to 0.51 °C/decade in the northwest.
4.5. Abrupt change of precipitation and temperature
Change points of the precipitation and temperature were also investigated in this study, and the results are shown in Fig. 12. For precipitation, only three out of 17 stations have a step change point. Two of them exhibited an upward abrupt change occurring in 1981 and 1986, respectively, while the other one exhibited downward abrupt changes occurring in 1997. Unlike precipitation series, all of the annual mean temperature series have an upward abrupt change. Of them, 13 occurred in 1986, three occurred in 1992, and one occurred in 1996.
Precipitation Temperature Precipitation Temperature
0 Downward ® Upward + Significant upward variability of precipitation (mm/lOa) variability of temperature fC/lOa)
(a) Trend (b) Magnitude
Fig. 11. MK trend test and magnitude of the trend for annual precipitation and mean temperature.
Fig. 12. Change point of the annual precipitation and mean temperature.
5. Discussion
5.1. Impact of climate change on streamflow
Climate change is the main cause to explain streamflow increasing in the upper HRB for less human activities have occurred in the mountain regions so far. In general, streamflow is positively related with precipitation but negatively correlated with temperature (Fu et al., 2007; Xu et al., 2010). Ice-snow melt-water and precipitation in the high mountain regions are the main water resources of the arid areas in northwest China. A significant increasing trend of the precipitation in the upper HRB, especially during the obvious wet period between 2003 and 2012, may be only part of the reason for headwater increase. Furthermore, increasing air temperature induced more glacier and snow melting during the past decades which contributed significantly to the streamflow increasing in the upper HRB. In the upper HRB, many mountainous terrains are at an elevation of 4000 m or higher, and they are covered in snow or glaciers throughout the year. Melting water of glaciers and snow replenish runoff effectively (Qin et al., 2013). Numerous studies showed a declination of ice and snow cover areas in the HRB during last several decades (Sakai et al., 2006; Wang et al., 2011a,b; Zhang et al., 2012a,b). It resulted in the increase of streamflow in upstream mountainous areas of the HRB (Nakawo, 2009). Related studies have also showed that snowmelt runoff increased obviously from 1970 to present (Wang et al., 2010).
From monthly changes of average streamflow data, the impact of climate change can be seen more clearly in the HRB. Taking the streamflow of the Yingluoxia (YL) station as an example (Fig. 13), streamflow increased in the spring, leading to the flood season in the summer. The spring, March to May, is the snow melting season, and the rising streamflow can be attributed to higher temperature. Increasing precipitation for the summer and autumn (see Fig. 10) can explain streamflow rising in the flood season.
There is hardly any runoff generated in the middle and lower HRB due to low precipitation and high evapotranspiration, and the change in precipitation does not much affect streamflow. In addition, precipitation increases by a very small amount in the middle and lower HRB, and thus the impact of the precipitation change on the streamflow is negligible. Moreover, the relationships between streamflow and air temperature are different in the middle and lower HRB. Higher air temperature leads to higher actual evapotranspiration which resulted in the decrease of streamflow.
200 160
i o E g 80
Fig. 13. Monthly variation of streamflow at the Yingluoxia and Zhengyixia stations.
5.2. Influence of human activities on streamflow
Streamflow is the most important water resource that sustains oases and irrigated agriculture in the HRB. During the last several decades, the hydrological regime of the Heihe River has been strongly affected by extensive human activities. They impacted the streamflow by surface water and groundwater exploitation, land reclamation, engineering project development, and new water-related policy implementation.
5.2.1. Socioeconomic development and its impact on streamflow
The middle reach region of the HRB is the major water consumer, accounting for about 90% of the total water use from the Heihe River. Zhangye City is the main urban center located in the middle HRB on the banks of the main stream of Heihe River. It has 92% of the total population and produces 83% of the total GDP for the entire HRB. In addition, more than 80% of the irrigated oases and 95% of the arable land in the HRB are located in Zhangye City and its vicinity. Economic growth and social development of the city increased the amount of water use in the middle HRB.
Zhangye is an important commodity grains producing region based on irrigated farming, and agriculture is responsible for approximately 90% of the total water consumption in the Zhangye oasis. Agricultural water use resulting from increased farmland areas in the midstream could be the most important factor driving the streamflow decline in the downstream areas. According to the census, the total population of Zhangye City was fewer than 0.6 million in 1950, and then steadily increased to more than 1.3 million in 2010 (Fig. 14(a)). As a result of increasing population, the area of farmland has increased significantly to maintain food supply and economic growth. The irrigated area in Zhangye was about 68,667 ha in the 1950s, and expanded to almost 266,000 ha by 2002, of which 212,000 ha was farmland (Wang et al., 2009).
(a) (b)
—<^1957-1980 (YL) —o—1981-2000 (YL) -2001-2012 (YL)
_-o- 1957-1980 (ZY) -a- 1981-2000 (ZY)--- 2001-2012 (ZY)
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Fig. 14. (a) Population and GDP, (b) Grain output and industry gross output in the past decades in Zhangye city.
Since irrigation sustains the agricultural production, a highly evolved system of irrigation canal networks, pumping stations and hydraulic projects have been constructed in the HRB to expand the irrigation capacity and support the artificial oasis. According to the water conservancy project survey information collected by Ma et al. (2009), there are 159 main canals, 782 branch canals, 5315 lateral canals, 6228 pumping wells and 53 reservoirs in the Zhangye area alone. The grain output of Zhangye (Fig. 14(b)) fluctuant increased from 5.10 x 108 kg in 1950 to 109.23 x 108 kg in 2010. Agricultural development in the Zhangye area significantly reduced the runoff available for the lower HRB. The total actual evapotranspirationofthe farmland in the middle HRB was 11.13 x 108 m3,13.16 x 108 m3, and 14.91 x 108 m3 in 1967,1986, and 2000, respectively (Cheng, 2007).
Annual changes in the streamflow for the downstream stations reflect the impact of irrigation more clearly. Generally, yearly variation of streamflow at the upstream stations (e.g., the Yingluoxia station (YL) in Fig. 13) is a unimodal distribution. Streamflow begins to rise in March after the low flow periods of January and February, reaches a maximum in July or August, and then decreases continuously until December. If there are no human activities, yearly variation of streamflow for a downstream station would have a similar pattern. However, for the lower reaches of the HRB (e.g., the Zhengyixia station, ZY in Fig. 13), streamflow dries up from May to July, reemerges during the flood season (from July to October), and then decreases in discharge or dries up again in November. Spring and summer are the irrigation period for the middle reaches, the characteristics of streamflow changes are closely related to those of agricultural irrigation. The increasing intensity of the irrigation water diversion can also be seen from the dotted lines in Fig. 13. Water consumption increased more and more after 1980.
Increase in the water demand and consumption by the industrial and service sectors is another reason for streamflow reduction in the Heihe River although they account for only around 10% of the total water use. China's national economic reform began in 1978, and new industrial sectors have grown greatly since then, which drive up the GDP rapidly. The GDP of Zhangye (Fig. 14(a)) increased notably after 1985, reached 21.2 billion Yuan in 2010 (about 600 folds of the 1950). The industry gross output (Fig. 14(b)) also increased notably after 1985, reached 9.84 billion Yuan in 2010. According to the government statistical information, the proportion of outputs from the agriculture, industry and services, respectively, was 72%, 7%, and 21% in 1952,47%, 30%, and 23% in 1980, and 28%, 36%, and 36% in 2012. The rapid rise in industry and services sectors led to a tremendous increase in water demand (Wang et al., 2009).
Water consumption in the middle HRB increased from 5.13 x 108 to 8.71 x 108 m3 per year from 1985 to 2001 (Qi and Luo, 2005). The rapid development of agriculture and growth of economy began in the early 1980s. The downward abrupt change in the streamflow of the Zhengyixia station started in 1979, and significant upward abrupt change of the streamflow difference between Yingluoxia and Zhengyixia started in 1982. The consistency in the timing confirmed that the decrease in the streamflow of the Heihe River were mainly due to local agricultural and economic development.
5.2.2. Water-related policy preferences
For the middle and lower HRB where streamflow was greatly affected by human activities, the policy preference of the government was an important factor directly or indirectly contributing to streamflow changes. In the 1980s, the government desired to make the "Hexi Corridor" an important grain production base. The emphasis on grain production promoted the rapid advance of farming and irrigation projects. Unconstrained development resulted in streamflow being dried up in the lower HRB. The shortage of water for the lower HRB left people and the ecosystem in the downstream Gobi desert region to compete for limited water resources for survival. As a consequence, the fragile ecological system has been seriously damaged, and the conflict of water between the midstream and downstream became rampant.
To restore the severely degraded ecosystem in the lower HRB, the government relied on water transfer projects to cope with water shortage. The central government had invested 2.3 billion Yuan to implement the EWDP in 2000. The increase in the streamflow to the downstream of HRB is a direct outcome of the EWDP. From 2000 to 2005, there had been 16 times of intermittent watering to the lower Heihe River with the total volume of 5.28 billion cubic meters (Guo et al., 2009).
Table 3
Gray relational degrees between consumed water and impact factors.
Relational degree Impact factors
Grain output Gross industrial output value Population Rural Urban
1957-2010 0.77 0.32 0.83 0.81
1957-1980 0.80 0.58 0.81 0.79
1981-2000 0.87 0.69 0.91 0.90
2001-2010 0.85 0.77 0.85 0.82
5.2.3. Relational analysis between streamflow change and human activities
In the previous two sections the relationship between the streamflow change and human activities was evaluated qualitatively. In this section, a quantitative analysis was conducted to ascertain the correlation between the streamflow change and human activities in the middle HRB.
Based on the data collected in this study, the correlation between the total water consumption (i.e., the streamflow difference between Yingluoxia and Zhengyixia stations) and the factors of human activities (i.e., grain output, gross industrial output value, rural and urban populations) is quantified using a method referred to as "gray relational analysis", which calculates the geometric proximity between a reference sequence and comparative sequences within a system (Wong et al., 2006). The gray relational degree value (GRDV) indicates the degree of the relation between different sequences: the larger the gray relational degree value for a factor of human activities, the greater its effect on total water consumption. Table 3 shows gray relational degree results of four periods of different length, i.e., 1957-2010, 1957-1980, 1981-2000 and 2001-2010.
Overall, for the entire study period of 1957-2010, population is the most important impact factor that reduced the streamflow released to the downstream. The GRDV for both rural and urban populations is larger than 0.8. The rural population, which is related to combined water consumption by farming, forestry, animal husbandry and fishery, shows the greatest impact on total water consumption. The grain output, which represented the water consumption by crops, is the close second most important impact factor on total water consumption with a GRDV of 0.77. The gross industrial output value, which partially reflected industrial water use, has the smallest influence on total water consumption with a GRDV of 0.32.
From the results of three different periods, 1957-1980,1981-2000 and 2001-2010, it is noteworthy that the impact of industrial water use on the total water consumption increased with more recent periods. The impact of grain output and population on the total water consumption first increased and then decreased. This situation is related to the adjustment of industrial structure on one hand and the EWDP on the other hand.
The impact of human activities on water consumption is further evaluated based on the multiple linear regressive model (MLRM). The MLRM is first constructed between the total water consumption (Ywc) in the middle HRB and quantifiable human activities (i.e., X1: grain output, X2: gross industrial output value, X3: rural population and X4: urban populations) during the period of 1957-2000, and then used to forecast the water consumption for the period of 2001-2010. The equation for the MLRM is Ywc = 3.641 + 0.065X1 - 0.004X2 + 0.124X3 - 0.028X4. And the results of the MLRM (see Fig. 15) show that the actual and calculated water consumptions are in good agreement with the same changing trend before 2000. However, it has a poor predictive capability for the water consumption during 2001-2010. This suggests that the impact factors that dominated water consumption in the middle HRB prior to 2000 were not operating in the same manner, due to a different set of policy preferences by the government, such as the implementation of the EWDP in 2000.
5.3. Eco-environmental consequences of streamflow changes
Human activities in the midstream of the HRB have decreased the streamflow annually and altered its temporal and spatial distributions over the years. With the declination and temporal change in the streamflow of the Heihe River, serious environmental deterioration and ecosystem degradation
0 -1-'-1-1-1-1-'-1-1-1-
1957 1962 1967 1972 1977 1982 1987 1992 1997 2002 2007 Year
Fig. 15. The actual and calculated water consumptions in the middle HRB.
have occurred in the HRB in recent decades, especially in the lower reaches. The terminal lakes, West Juyan Lake (with a largest water surface of 560 km2) and East Juyan Lake, were completely dried up in 1961 and 1992, respectively. With desiccation of surface water, there was a decrease in the natural recharge to the groundwater which lowered the regional water table by 1.2-2.5 m and led to decline of groundwater-dependent vegetation and glowing desertification (Feng et al., 2005). The Ejina desert plain in the downstream was believed to be a source for the dust storms of North China (Wang et al., 2005a,b). Herbaceous plants in the HRB decreased from 200 to 80 species and the forage species decreased from 130 to 20 species from the 1950s to 1990s (Wang and Cheng, 2000). Populus euphratica in the riparian zone has been facing the danger of degrading or even collapsing. Vegetation degradation, in turn, has led to a decrease in wild animals. There used to be 26 species of rare wild animals in the HRB, however, nine species have disappeared and more than 10 species have migrated in the 1990s (Gong and Dong, 1998).
Implementation of the EWDP has improved the eco-environmental conditions in the lower HRB to some extent since 2000. Due to the increase of streamflow, shallow groundwater levels have been raised (Wang et al., 2011a,b), and vegetation of the oasis areas showed a recovery and expanding trend in the Ejina basin (Guo et al., 2009; Zhang et al., 2011). In 2002, the first water from the EDWP reached East Juyan Lake and formed a maximum water surface of 35.7 km2 in 2005 (Guo et al., 2009), which has improved the heath of the lake ecosystem. However, though the EWDP has achieved some success, it is far from enough. Since 2000 although the water released to the downstream has increased substantially, it has concentrated during July to October (Fig. 5). The plant growth is not only closely related to the water volume but also the duration of watering. Rational allocation and sustainable utilization of water resources remains a major challenge for the HRB.
6. Conclusions
Over the last several decades, although the streamflow coming from the upper reaches of the HRB has risen significantly, those flowing to the lower reaches have declined significantly. The direct cause is the increase of water consumption in the middle reaches arising from rapid population and economic growth. Since 2000, after the implementation of the Ecological Water Diversion Project (EWDP), the rising trend for the streamflow in the upper reaches has become more apparent while the declining trend for those in the lower reaches has weakened. This reflects the fact that the climate warming in the upstream headwater region has intensified, while the EWDP has allowed more flow in the midstream to be released to the downstream.
Climate changes have been shown to be partly responsible for the trends of streamflow variations detected in 13 gaging stations over the HRB. There are statistically significant increasing trends for the mean annual temperature and smaller increasing trends for the precipitation in the HRB. Rising precipitation and temperature in the upper HRB are the main reason for increased streamflow from
the upstream headwater region to the midstream oases. In the middle and lower HRB, higher air temperature may be attributed to streamflow decreases.
However, assessment of agricultural and socioeconomic development, based on both qualitative and quantitative evaluations, revealed that the human activity is the dominant driver for the decline of streamflow in the main stream of the HRB during the past several decades. The implementation of the EWDP is a determining factor that altered the hydrological regime of the downstream areas. Rational allocation and sustainable utilization of water resources for the HRB requires careful and systematic consideration of all relevant physical and socioeconomic conditions, which will be further explored and discussed in a future study.
Conflict of interest
None declared.
Acknowledgements
This work was supported by the National Natural Science Foundation of China (grants no. 91225301 and 91025019). We appreciate the data support from the Heihe Research Program (http://westdc.westgis.ac.cn/heihe) and also from the Cold and Arid Regions Environmental and Engineering Research Institute (CAREERI), Chinese Academy of Sciences. We also thank the editor Okke Batelaan and two anonymous reviewers for their invaluable comments.
References
Cao, Y., Nan, Z., Hu, X., 2012. Estimating groundwater storage changes in the Heihe river basin using grace. In: Geoscience and Remote Sensing Symposium (IGARSS), IEEE International, pp. 798-801.
Chen, Y., Xu, C., Chen, Y., Liu, Y., Li, W., 2013. Progress, challenges and prospects of ecohydrological studies in the Tarim River basin ofXinjiang, China. Environ. Manag. 51 (1), 138-153.
Chen, Y., Zhang, D.Q., Sun, Y.B., Liu, X.N., Wang, N.Z., Savenije, H.H., 2005. Water demand management: a case study ofthe Heihe river basin in China. Phys. Chem. Earth 30 (6), 508-519.
Cheng, Y.F., (Master thesis) 2007. Study on Temporal Land Spatial Distribution of Crop Evapotranspiration in Plain of the Main Heihe River Middle Reaches. Lanzhou University.
Cheng, G.D., Xiao, H.L.,Xu, Z.M., Li,J.X., Lu, M.F., 2006. Water issue and its countermeasure in the inland river basins of Northwest China - a case study in Heihe river basin. J. Glaciol. Geocryol. 28 (3), 506-513 (in Chinese).
Feng, Q., Wei, L., Su, Y.H., Zhang, Y.W., Si, J.H., 2005. Distribution and evolution of water chemistry in Heihe river basin. Environ. Geol. 55, 956-957.
Fu, G.B., Charles, S.P., Chiew, F.H.S., 2007. A two-parameter climate elasticity of streamflow index to assess climate change effects on annual streamflow. Water Resour. Res. 53, W11519, http://dx.doi.org/10.1029/2007WR005890.
Gong, J., Dong, G., 1998. Environment degradation of the Ejina Oasis and comprehensive rehabilitation in the lower reaches of the Heihe River. J. Desert Res. 18,55-59.
Guo, P., Chen, X.H., Tong, L., Li, J.B., Li, M., 2012. An optimization model for crop deficit irrigation system under uncertainty. Eng. Optim. 56(1), 1-15.
Guo, Q.L., Feng, Q., Li, J.L., 2009. Environmental changes after ecological water conveyance in the lower reaches of Heihe River, northwest China. Environ. Geol. 58 (7), 1387-1396.
Hamed, K.H., Rao, A.R., 1998. A modified Mann-Kendall trend test for autocorrelated data. J. Hydrol. 205,182-196.
Hamed, K.H., 2009. Enhancing the effectiveness of prewhitening in trend analysis of hydrological data. J. Hydrol. 368,153-155.
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.
Hu, L.T., Chen, C.X., Jiao, J.J., Wang, Z.J., 2007. Simulated groundwater interaction with rivers and springs in the Heihe river basin. Hydrol. Process. 21 (20), 2795-2806.
Kang, E., Lu, L., Xu, Z., 2007. Vegetation and carbon sequestration and their relation to water resources in an inland river basin of Northwest China. J. Environ. Manag. 85 (3), 702-710.
Kendall, M.G. (Ed.), 1975. Rank Correlation Methods., 2nd ed. C. Griffin, London.
Li, B.F., Chen, Y.N., Chen, Z.S., Li, W.H., 2012. Trends in runoffversus climate change in typical rivers in the arid region of northwest China. Quat. Int. 282, 87-95.
Li, X., Cheng, G.D., Liu, S.M., et al., 2013. Heihe watershed allied telemetry experimental research (HiWATER): scientific objectives and experimental design. Bull. Am. Meteorol. Soc. 95 (8), 1155-1160.
Ma, M.G., Hu, X.L., Song, Y., Liu, X.J., Xu, G.J., 2009. Development and application on the information system of water conservancy project statues for Zhangye city. Remote Sens. Technol. Appl. 25 (5), 559-566 (in Chinese).
Mann, H., 1945. Non-parametric tests against trend. Econometrica 13, 245-259.
Min, L., Yu,J., Liu, C., Zhu,J., Wang, P., 2013. The spatial variability of streambed vertical hydraulic conductivity in an intermittent river, northwestern China. Environ. Earth Sci. 69 (3), 873-883.
Nakawo, M., 2009. Shrinkage of summer-accumulation - glaciers in Asia under consideration of downstream population. In: Assessment of Snow, Glacier and Water Resources in Central Asia., pp. 19-25.
Pettitt, A.N., 1979. A non-parametric approach to the change point problem. Appl. Stat. 28 (2), 126-135.
Qi, S.Z., Luo, F., 2005. Water environmental degradation of the Heihe river basin in arid northwestern China. Environ. Monit. Assess. 108 (1-3), 205-215.
Qin, D., Qian, Y., Han, L., Wang, Z.M., Li, C., Zhao, Z.F., 2011. Assessing impact of irrigation water on groundwater recharge and quality in arid environment using CFCs, tritium and stable isotopes, in the Zhangye Basin, northwest China. J. Hydrol. 505
(1),195-208.
Qin, J., Ding, Y.J., Wu, J.K., Gao, M.J., Yi, S.H., Zhao, C.C., Ye, B.S., Wang, L.M.S.X., 2013. Understanding the impact of mountain landscapes on water balance in the upper Heihe River watershed in northwestern China. J. Arid Land 5 (3), 366-383.
Rouge, C., Ge, Y., Cai, X.M., 2012. Detecting gradual and abrupt changes in hydrological records. Adv. Water Resour. 53,33-55.
Sakai, A., Fujita, K., Duan, K., Pu, J., Nakawo, M., Yao, T., 2006. Five decades of shrinkage of July 1st glacier, Qilian Shan, China. J. Glaciol. 52(176), 11-16.
Sen, P.K., 1968. Estimates of the regression coefficient based on Kendall's tau. J. Am. Stat. Assoc. 63,1379-1389.
Wang, G.X., Cheng, G.D., 2000. The characteristics of water resources and the changes of the hydrological process and environment in the arid zone of northwest China. Environ. Geol. 39 (7), 783-790.
Wang, G.X., Yang, L.Y., Chen, L., Kubota, J., 2005a. Impacts of land use changes on groundwater resources in the Heihe river basin. J. Geogr. Sci. 15,505-515.
Wang, J., Li, H., Hao, X., 2010. Responses of snowmelt runoff to climatic change in an inland river basin, northwestern China, overthe past 50 years. Hydrol. Earth Syst. Sci. 15 (10), 1979-1987.
Wang, P., Yu, J.J., Zhang, Y.C., Fu, G.B., Min, L.L., Ao, F., 2011a. Impacts of environmental flow controls on the water table and groundwater chemistry in the Ejina Delta, northwestern China. Environ. Earth Sci. 65 (1), 15-25.
Wang, X.M., Dong, Z.B., Zhang, J.W., Liu, L.C., 2005b. Modern dust storms in China: an overview. J. Arid Environ. 58, 559-575.
Wang, Y., Feng, Q., Chen, L., Yu, T., 2013. Significance and effect of ecological rehabilitation project in inland river basins in northwest China. Environ. Manag. 52 (1), 209-220.
Wang, P.Y., Li, Z.Q., Gao, W.Y., 2011b. Rapid shrinking of glaciers in the Middle Qilian Mountains region of Northwest China during the last 50 years. J. Earth Sci. 22, 539-558.
Wang, Y., Xiao, H.L., Lu, M.F., 2009. Analysis of water consumption using a regional input-output model: model development and application to Zhangye city, northwestern China. J. Arid Environ. 73 (10), 895-900.
Wei, Z., Liu, Y., Xu, D., Cai, J.B., Zhang, B.Z., 2013. Application and comparison of winter wheat canopy resistance estimation models based on the scaling-up of leaf stomatal conductance. Chin. Sci. Bull. 58 (23), 2909-2916.
Wong, H., Hu, B.Q., Ip, W.C., Xia, J., 2006. Change-point analysis of hydrological time series using grey relational method. J. Hydrol. 324(1), 323-338.
Wu, B., Zheng, Y., Tian, Y., Wu, X., Yao, Y., Han, F., Liu, J., Zheng, C., 2014. Systematic assessment of the uncertainty in integrated surface water-groundwater modeling based on the probabilistic collocation method. Water Resour. Res. 50, 5848-5865, http://dx.doi.org/10.1002/2014WR015366.
Xu, Z.X., Liu, Z.F., Fu, G.B., Chen, Y.N., 2010. Trends of major hydroclimatic variables in the Tarim River basin during the past 50 years. J. Arid Environ. 75 (2), 256-267.
Yao, Y., Zheng, C., Liu, J., Cao, G., Xiao, H., Li, H., Li, W., 2014. Conceptual and numerical models for groundwater flow in an arid inland river basin. Hydrol. Process., http://dx.doi.org/10.1002/hyp.10276.
Yin, Z.L., Xiao, H.L., Zou, S.B., Lu, Z.X., Wang, W.H., Liu, Z.X., Xu, B.R., 2012. Analysis on water balance in different land cover types at upper reaches of Heihe river basin in northwestern China using SWAT model. In: 2012 International Symposium on Geomatics for Integrated Water Resources Management (GIWRM), pp. 1-6.
Yue, S., Pilon, P., Phinney, B., Cavadias, G., 2002. The influence of autocorrelation on the ability to detect trend in hydrological series. Hydrol. Process. 16,1807-1829.
Zang, C., Liu, J., Jiang, L., Gerten, D., 2013. Impacts of human activities and climate variability on green and blue water flows in the Heihe river basin in Northwest China. Hydrol. Earth Syst. Sci. 10 (7), 9505-9577.
Zhang, Y., Ma, J., Chang, X., van Wonderen, J., Yan, L., Han, J., 2012a. Water resources assessment in the Minqin Basin: an arid inland river basin under intensive irrigation in northwest China. Environ. Earth Sci. 65 (6), 1831-1839.
Zhang, Y.C., Yu,J.J., Wang, P., Fu, G.B., 2011. Vegetation responses to integrated water management in the Ejina basin, northwest China. Hydrol. Process. 25 (22), 3558-3561.
Zhang, Y.S., Liu, S.Y., Shangguan, D.H., Li, J., Zhao, J.D., 2012b. Thinning and shrinkage of Laohugou No. 12 glacier in the Western Qilian Mountains, China, from 1957 to 2007. J. Mt. Sci-Engl. 9 (3), 350-353.
Zhao, W.Z., Chang, X.L., He, Z.B., Zhang, Z.H., 2007. Study on vegetation ecological water requirement in Ejina Oasis. Sci. China Ser. D 50 (1), 121-129.
Zheng, C., Cheng, G., Fu, B., Song, C., Xiao, H., 2012. An integrated study of ecological and hydrological processes i the Heihe river basin, northwest China (invited), Abstract GC23D-05 Presente at 2012 Fall Meeting, AGU, San Francisco, Calif., 3-7 December.
Zheng, J.H., Huang, G.H., Jia, D.D., Wang, J., Mota, M., Pereira, L.S., Huang, Q.Z., Xu, X., Liu, H.J., 2013. Responses of drip irrigated tomato (Solanum lycopersicum L.) yield, quality and water productivity to various soil matric potential thresholds in an arid region of northwest China. Agric. Water Manag. 129 (11), 181-193.
Zhou, J., Hu, B., Cheng, X., Wang, G.G., Li, X., 2011. Development of a three-dimensional watershed modelling system for water cycle in the middle part of the Heihe river shed, in the west of China. Hydrol. Process. 25 (12), 1965-1978.
Zhu, Y.H., Wu, Y.Q., Drake, S., 2005. A survey: obstacles and strategies for the development of ground-water resources in arid inland river basins of Western China. J. Arid Environ. 59 (2), 351-367.
Zou, Y., Zhang, B., 2012. Identification and analysis the change points of hydrological process ofYingluoxia station in the Heihe river basin. In: World Automation Congress (WAC), IEEE, pp. 1-5.
