Abstract

The water shortage in the Huaihe River Basin (HRB), China, has been aggravated by population growth and climate change. To identify the characteristics of streamflow change and assess the impact of climate variability and human activities on hydrological processes, approximately 50 years of natural and observed streamflow data from 20 hydrological stations were examined. The Mann–Kendall test was employed to detect trends. The results showed the following. (i) Both the natural and the observed streamflow in the HRB present downward trends, and the decreasing rate of observed streamflow is generally faster than that of the natural streamflow. (ii) For the whole period, negative trends dominate in the four seasons in the basin. The highest decreasing trends for two kinds of streamflow both occurred in spring, and the lowest ones were in autumn and winter. (iii) Based on the above analysis and quantifying assessment for streamflow decrease, human activity was the main driving factor in the Xuanwu (80.78%), Zhuangqiao (79.92%), Yongcheng (74.80%), and Mengcheng (64.73%) stations which all belong to the Huaihe River System (HRS). On the other hand, climate variability was the major driving factor in the Daguanzhuang (68.89%) and Linyi (63.38%) stations which all belong to the Yishusi River System (YSR).

INTRODUCTION

Climate change and human activities have altered the hydrologic cycle (Huntington 2006; Wang & Hejazi 2011; Jiang et al. 2015). The accelerated hydrological cycle is altering the spatial and temporal distributions of regional water resources which could have extensive impacts on water resources availability, the frequency and intensity of floods and droughts, agriculture and aquatic ecosystems (Vörösmarty et al. 2000; Zhang et al. 2013a; Yan et al. 2016). Hydrology plays a central role in the development and management of water resources, and streamflow constitutes a major phase in the hydrological cycle (Miao et al. 2012). Therefore, it is important to analyze hydrological trends and assess the impacts of climate change and human activities on streamflow in order to understand the balance of water dynamics for water resources management, and to plan strategies for solving many problems such as the increasing imbalance of the water resources supply and demand around the world (e.g., Kahya & Kalayci 2004; Chen et al. 2012; Zhang et al. 2013b; Yang et al. 2016).

The hydrologic regime of a stream under specific geomorphic conditions represents the integrated basin response to various climatic inputs and human activities (Zhang et al. 2001; Jones 2011). However, different conclusions have been drawn which reflect the great diversity of the regional and global climates and hydrological regimes (Lettenmaier et al. 1994; Lins & Slack 1999; Zhang et al. 2001). The streamflow trends of various river systems across China were analyzed based on long-term historical records of climatic and hydrological data. It was found that streamflow had decreased in northern China, including the Yellow River Basin (Zhang et al. 2009), Haihe River Basin (Zhan et al. 2013), and Liaohe River Basin (Jiang et al. 2011a), while it had increased in southeastern China, e.g., the Yangtze River Basin (Ye et al. 2013; Omer et al. 2016), and Pearl River Basin (Zhang et al. 2008). In the west of China, streamflow in the headwater of the Tarim River exhibited significant increase, but a decreasing trend has been detected along the mainstream of the river (Tao et al. 2011). However, few studies on the long-term monotonic trend and abrupt change of streamflow in the Huaihe River Basin (HRB) have been carried out. Hao et al. (2011) found that annual streamflow of four hydrological stations in the HRB showed an insignificant decreasing trend, and the abrupt change in Linyi station occurred in 1965 and 1975. Shi et al. (2012) evaluated the runoff trends of five hydrological stations in the HRB and reported no significant change for annual runoff in the Huaihe River, while there was a significant decrease trend in some tributaries. However, no clear conclusions could be drawn as to whether there were significant trends at basin scale because of poor spatial coverage of the data. Xia et al. (2012) examined trends of extreme runoff events based on the daily runoff data from 20 hydrological stations above the Bengbu Sluice in the HRB and found 10 stations had a decreasing trend, while the other stations had an ambiguous increasing trend. However, none examined trend characteristics for different time intervals at the same time, such as seasonal and annual bases, to see whether or not a dramatic change occurs. Moreover, the above studies focused on the observed streamflow while there is a lack of sufficient study about the natural streamflow in the HRB. The difference between observed and natural streamflow can more effectively explain some real phenomena of streamflow change, and distinguish the impact of climate change and human activities on the streamflow.

The HRB, the sixth major river basin in China, is one of the most sensitive areas for annual runoff changes as a response to climate change in China (Gao et al. 2009) because of its transitional climate from sub-tropical zone to warm temperate zone and from wet zone to semi-arid and arid zone. This region, as one of the main cropping areas of China, has the highest population density (631 persons per km2). However, the amount of water resources per capita is only 438 m3, representing 1/5 of the average in China and 1/20 of the average in the world. Due to the rapid economic development and population growth, water withdrawal has increased substantially in this basin since the 1980s (Gao et al. 2009; Yang et al. 2010). This has resulted in frequent water shortages in the region and increasingly imbalance of the water resources supply and demand. All in all, water shortages in the HRB have become major challenges and have attracted much attention from both the relevant authorities and the general public (Shi et al. 2013). Therefore, with a background of climate change and human activities, it is essential to stress the importance of the detailed trend analysis in hydrologic variables (streamflow as the most noteworthy hydrological variable) in both space and time for the assessment of climate-induced risks and the pursuit of countermeasures in the HRB.

In this study, data of the natural and observed streamflow were collected at 20 hydrological stations to help us detect streamflow variations over the HRB, and discuss the influence of climate variability or human activities on streamflow change, with a particular emphasis on flow regime characteristics. This study can provide helpful information for policy-makers to manage water resources more effectively.

STUDY AREA

The HRB, encompassing nearly 12.7 million hectares of farmland and over 170 million people, is located in eastern China (111°55′–121°25′E, 30°55′–36°36′N), and spreads from the Tongbei-Dabie mountains in the west to the Yellow Sea in the east, and from the south levee of the Yellow River in the north to the Yangtze River Basin in the south (Figure 1). The area of the basin is 270,000 km2, which is divided into two sub-basins by the abandoned Yellow River: Huaihe River System (HRS, 190,000 km2) and Yishusi River System (YRS, 80,000 km2).

Figure 1

The geographical distribution of hydro stations and meteorological stations in the HRB.

Figure 1

The geographical distribution of hydro stations and meteorological stations in the HRB.

The climate of the basin is chiefly characterized by monsoonal weather conditions. The annual mean temperature is between 11 and 16 °C, and annual precipitation is about 898 mm, which generally decreases from about 1,000 mm in the southeast to less than 600 mm in the northwest. Meantime, precipitation varies monthly and annually. The precipitation in summer (June–August) generally accounts for 60–70% of the annual precipitation. Annual variability of precipitation is also very large due to the unstable characteristics of the duration, intensity, and impacting region of the subtropical high over the northern Pacific in summer. To provide water for agricultural, industrial, municipal, and hydroelectric power uses and flood control in the HRB, around 11,000 dams and sluices had been built by the year 2000. Thus, the basin has a highly regulated river system. Any significant change in the magnitude or timing of runoff induced by the climate changes or human activities would have significant implications for the economic prosperity and the eco-environmental security in this region (Jiang et al. 2011b).

DATA AND METHODOLOGY

Data

A dense network of hydrometric stations was established in the HRB in the 1950s, and complete data sets for streamflow have been recorded since. Simultaneously, considering the influence of human activities, such as water withdrawal from the river channel for irrigation, industrial, and domestic usage, and the role of dams in controlling the streamflow, the Huai River Conservancy Commission (HRCC) conducted a great deal of complex work to collect data and build the real, or so-called natural streamflow series. The differences between the natural and the observed streamflow include the water abstracted and consumed from the river, and the water transmission from outside the basin.

The natural streamflow at a given station can be calculated as:  
formula
(1)
where Qw,n is the calculated natural streamflow, Qw,m is the observed streamflow, and Qw,div is the net water quantity diverted from the river above the station. The formula for estimating natural runoff requires detailed information that is extremely difficult to collect. The differences between observed runoff and natural runoff are generally due to four factors: (i) the amount of water directly collected from the river channel for irrigation, industry, and domestic usage, and the amount returning to the downstream river channel after use; (ii) the amount of water controlled by dams, including extra water losses through evaporation and seepage; (iii) the amount of water transported into and out of the watershed; and (iv) the inter-basin water transfer projects such as the Yellow River and Yangzi River water transfer projects, causing observed runoff greater than the natural runoff in some periods. Although some hydrologists question the accuracy of natural runoff, it is generally believed that the natural runoff data published by the HRCC are the most authoritative and most accurate information. Hence, the data are widely used in water resource management and planning, and in hydrological engineering projects.

The observed and natural streamflow data from 20 hydrological stations (five stations situated along the mainstream, 15 on the major tributaries) in the HRB were selected for analysis of streamflow characteristics (Figure 1). It was assumed that the data from the selected stations were sufficient to characterize the hydrological trends over the entire basin. Detailed information concerning the hydrological records of these 20 hydrological stations is shown in Table 1. Natural and observed streamflow series were provided by the HRCC.

Table 1

Basic information of the 20 hydrological stations in the HRB

Sub-basin River Station Lon. (E°) Lat. (N°) Drainage area (104 × km2Data period
 
Observed streamflow Natural streamflow 
Huaihe River System (HRS) Huaihe R Changtaiguan 114.07 32.32 0.31 1956–2000 1956–2000 
Xixian 114.73 32.33 1.02 1956–2000 1956–2000 
Wangjiaba 115.6 32.43 3.06 1956–2000 1956–2000 
Lutaizi 116.63 32.57 8.86 1956–2000 1956–2000 
Wujiadu 117.37 32.95 12.13 1956–2000 1956–2000 
Honghe R Xincai 114.98 32.77 0.41 1956–2000 1956–2000 
Bantai 115.07 32.72 1.13 1956–2000 1956–2000 
Yinghe R Zhoukou 114.65 33.63 2.58 1956–2000 1956–2000 
Fuyang 115.83 32.9 3.52 1956–2000 1956–2000 
Qaunhe R Shenqiu 115.12 33.17 0.31 1956–2000 1956–2000 
Shihe R Jiangjiaji 115.73 32.3 0.59 1956–2000 1956–2000 
Pihe R Hengpaitou 116.37 31.6 0.44 1956–2000 1956–2000 
Huijihe R Zhuanqiao 115.35 34.02 0.34 1956–2000 1956–2000 
Wohe R Xuanwu 115.28 33.98 0.40 1959–2000 1959–2000 
Mengcheng 116.55 33.28 1.55 1956–2000 1956–2000 
Tuohe R Yongcheng 116.4 33.93 0.22 1956–2000 1956–2000 
Chihe R Mingguang 117.97 32.78 0.35 1956–2000 1956–2000 
Sanhe R Zhongdu 118.77 33.1 – 1956–2000 1956–2000 
Yishusi River System (YRS) Yihe R Linyi 118.4 35.02 1.03 1958–2000 1958–2000 
Shuhe R Daguanzhuan 118.55 34.8 0.45 1958–2000 1958–2000 
Sub-basin River Station Lon. (E°) Lat. (N°) Drainage area (104 × km2Data period
 
Observed streamflow Natural streamflow 
Huaihe River System (HRS) Huaihe R Changtaiguan 114.07 32.32 0.31 1956–2000 1956–2000 
Xixian 114.73 32.33 1.02 1956–2000 1956–2000 
Wangjiaba 115.6 32.43 3.06 1956–2000 1956–2000 
Lutaizi 116.63 32.57 8.86 1956–2000 1956–2000 
Wujiadu 117.37 32.95 12.13 1956–2000 1956–2000 
Honghe R Xincai 114.98 32.77 0.41 1956–2000 1956–2000 
Bantai 115.07 32.72 1.13 1956–2000 1956–2000 
Yinghe R Zhoukou 114.65 33.63 2.58 1956–2000 1956–2000 
Fuyang 115.83 32.9 3.52 1956–2000 1956–2000 
Qaunhe R Shenqiu 115.12 33.17 0.31 1956–2000 1956–2000 
Shihe R Jiangjiaji 115.73 32.3 0.59 1956–2000 1956–2000 
Pihe R Hengpaitou 116.37 31.6 0.44 1956–2000 1956–2000 
Huijihe R Zhuanqiao 115.35 34.02 0.34 1956–2000 1956–2000 
Wohe R Xuanwu 115.28 33.98 0.40 1959–2000 1959–2000 
Mengcheng 116.55 33.28 1.55 1956–2000 1956–2000 
Tuohe R Yongcheng 116.4 33.93 0.22 1956–2000 1956–2000 
Chihe R Mingguang 117.97 32.78 0.35 1956–2000 1956–2000 
Sanhe R Zhongdu 118.77 33.1 – 1956–2000 1956–2000 
Yishusi River System (YRS) Yihe R Linyi 118.4 35.02 1.03 1958–2000 1958–2000 
Shuhe R Daguanzhuan 118.55 34.8 0.45 1958–2000 1958–2000 

Monthly precipitation and temperature data from 14 evenly distributed meteorological stations (Figure 1) were collected from 1959 to 2008 by the China meteorological administration. The monthly and annual precipitation (and temperature) was then established from the collected data.

Methodology

Trend analysis

The Mann–Kendall (MK) trend test is a rank-based non-parametric method that has been widely applied for trend detection in hydro-climatic time series due to its robustness against the influence of abnormal data and especially its reliability for biased variables (Yue & Wang 2002; Partal & Kahya 2006; Zhan et al. 2013). Therefore, this method has been recommended for general use by the World Meteorological Organization (Mitchell et al. 1966). This study also used the MK test method to analyze trends in streamflow, precipitation, and temperature series. It should be noted here that the results of the MK test are heavily affected by serial correlation of the time series. Hence, before the trend analysis is performed, the MK test must be modified if the data are serially correlated (Zhang et al. 2001; Önöz & Bayazit 2012; Ye et al. 2013) in order to reduce the influence on trend detection; results revealed (not shown here) that no significant autocorrelation existed in the data. The MK test does not quantify the trend magnitude. The robust Kendall slope (β) was used to assess the slope of the monotone trend, which is insensitive to outliers and can be effectively used to quantify a trend in the hydrological time series (Li et al. 2012; Miao et al. 2012).

Change point analysis

To detect the existence of any step change points in the hydrological data, the sequential MK test and distribution-free CUSUM (cumulative sum) test were used in this study. The sequential MK test is widely used to detect change points of the hydro-climatic data (Partal & Kahya 2006; Zhang et al. 2013b), and has the following advantages (Bao et al. 2012): (1) it can handle non-normality and censoring data; (2) it has a high asymptotic efficiency. The distribution-free CUSUM test was used in order to test the hypothesis concerning the existence of the step change points in the hydro-climatic time series during the period of study. This test is also a rank-based non-parametric method whether the means in two parts of a record are different for an unknown time of change (Chiew & McMahon 1993; Kundzewicz & Robson 2004). In particular, successive observations are compared with the median of the series in order to detect the step change points in the mean of a time series after a number of observations (Potter et al. 2010; Paton et al. 2013).

Quantitative assessment of the attribution for streamflow variation

The driving factors for the observed streamflow change conclude climate variability and human activities. The change point of observed annual streamflow indicates the abrupt change of annual streamflow. Using the change point, the whole time series can be divided into two periods (Figure 2). The first one before the change point is called the ‘natural period’, in which there is not a statistically significant increasing or decreasing trend for streamflow. That means that the hydrological cycle and water resources system keep natural status and are insignificantly impacted by human activities. Spontaneously, the second one is after the change point, called the ‘impacted period’, in which there is a statistically significant change of streamflow, compared to in the ‘natural period’. That means the hydrological cycle and water resources system are extensively influenced by climate variability and/or human activities.

Figure 2

The framework for quantitative assessment of the contribution of climate variability and human activities for the streamflow.

Figure 2

The framework for quantitative assessment of the contribution of climate variability and human activities for the streamflow.

The illustration of the assessment of the attribution for streamflow decrease is shown in Figure 2. The observed annual streamflow in the ‘natural period’ and ‘impacted period’ are defined as Qon and Qoi, respectively. Similarly, the natural annual streamflow in the ‘natural period’ and ‘impacted period’ are defined as Qnn and Qni. Therefore, the changes of the observed and natural annual streamflow ( and ) can be calculated by (Bao et al. 2012):  
formula
(2)
 
formula
(3)
where contains two parts: the streamflow change caused by climate variability () and human activities (). In order to roughly estimate , this paper assumes that the change of the natural annual streamflow () is approximately equal to the streamflow change caused by climate variability (). So, the streamflow change caused by human activities () could be calculated by:  
formula
(4)
with the estimation of or , the contribution of climate variability and human activities to streamflow change, which are defined as ηC and ηH, respectively, and can be separated and estimated by:  
formula
(5)

RESULTS AND DISCUSSION

Trends for annual streamflow

The spatial distribution of trends for annual streamflow in the HRB over the whole available period is displayed in Figure 3. In addition, the magnitude of trend (Kendall slope, β) for annual streamflow was detected, as shown in Table 2.

Table 2

Results of the MK test

Sub-basin River Station MK test for observed streamflow
 
MK test for natural streamflow
 
Z Trend Β (m3·s−1·yr−1Z Trend Β (m3·s−1·yr−1
Huaihe River System (HRS) Huaihe R Changtaiguan −0.27 ↓ −0.07 −0.09 ↓ −0.02 
Xixian −0.10 ↓ −0.05 −0.22 ↓ −0.21 
Wangjiaba −0.41 ↓ −0.59 −0.40 ↓ −0.55 
Lutaizi −1.04 ↓ −4.18 −0.30 ↓ −1.76 
Wujiadu −1.02 ↓ −4.03 −0.50 ↓ −5.44 
Honghe R Xincai −0.57 ↓ −0.13 −0.46 ↓ −0.12 
Bantai −0.61 ↓ −0.35 −0.66 ↓ −0.51 
Yinghe R Zhoukou −1.92 ↓ −1.36 −0.99 ↓ −0.68 
Fuyang −1.68 ↓ −1.55 −0.73 ↓ −0.82 
Qaunhe R Shenqiu −1.30 ↓ −0.12 0.13 ↑ 0.01 
Shih eR Jiangjiaji −1.64 ↓ −0.76 −0.19 ↓ v0.09 
Pihe R Hengpaitou −0.45 ↓ −0.19 −0.79 ↓ −0.27 
Huijihe R Zhuanqiao − 2.80 ↓ − 0.22 −1.32 ↓ −0.05 
Wohe R Xuanwu − 3.59 ↓ − 0.15 0.98 ↑ 0.03 
Mengcheng − 4.24 ↓ − 1.32 −1.91 ↓ −0.39 
Tuohe R Yongcheng − 5.52 ↓ − 0.12 −0.64 ↓ −0.01 
Chihe R Mingguang −0.16 ↓ −0.03 0.26 ↑ 0.07 
Sanhe R Zhongdu −0.94 ↓ −6.09 −0.54 ↓ −2.84 
Yishusi River Yihe R Linyi − 2.71 ↓ − 1.25 −1.82 ↓ −1.22 
System (YRS) Shuhe R Daguanzhuan − 2.52 ↓ − 0.57 −0.98 ↓ −0.25 
Sub-basin River Station MK test for observed streamflow
 
MK test for natural streamflow
 
Z Trend Β (m3·s−1·yr−1Z Trend Β (m3·s−1·yr−1
Huaihe River System (HRS) Huaihe R Changtaiguan −0.27 ↓ −0.07 −0.09 ↓ −0.02 
Xixian −0.10 ↓ −0.05 −0.22 ↓ −0.21 
Wangjiaba −0.41 ↓ −0.59 −0.40 ↓ −0.55 
Lutaizi −1.04 ↓ −4.18 −0.30 ↓ −1.76 
Wujiadu −1.02 ↓ −4.03 −0.50 ↓ −5.44 
Honghe R Xincai −0.57 ↓ −0.13 −0.46 ↓ −0.12 
Bantai −0.61 ↓ −0.35 −0.66 ↓ −0.51 
Yinghe R Zhoukou −1.92 ↓ −1.36 −0.99 ↓ −0.68 
Fuyang −1.68 ↓ −1.55 −0.73 ↓ −0.82 
Qaunhe R Shenqiu −1.30 ↓ −0.12 0.13 ↑ 0.01 
Shih eR Jiangjiaji −1.64 ↓ −0.76 −0.19 ↓ v0.09 
Pihe R Hengpaitou −0.45 ↓ −0.19 −0.79 ↓ −0.27 
Huijihe R Zhuanqiao − 2.80 ↓ − 0.22 −1.32 ↓ −0.05 
Wohe R Xuanwu − 3.59 ↓ − 0.15 0.98 ↑ 0.03 
Mengcheng − 4.24 ↓ − 1.32 −1.91 ↓ −0.39 
Tuohe R Yongcheng − 5.52 ↓ − 0.12 −0.64 ↓ −0.01 
Chihe R Mingguang −0.16 ↓ −0.03 0.26 ↑ 0.07 
Sanhe R Zhongdu −0.94 ↓ −6.09 −0.54 ↓ −2.84 
Yishusi River Yihe R Linyi − 2.71 ↓ − 1.25 −1.82 ↓ −1.22 
System (YRS) Shuhe R Daguanzhuan − 2.52 ↓ − 0.57 −0.98 ↓ −0.25 

Data series with significant trends at the 0.05 significance level are shown in bold.

Figure 3

Spatial distribution of trends for annual streamflow over the HRB: (a) natural streamflow and (b) observed streamflow. denotes significant decrease; denotes no significant decrease; denotes significant increase.

Figure 3

Spatial distribution of trends for annual streamflow over the HRB: (a) natural streamflow and (b) observed streamflow. denotes significant decrease; denotes no significant decrease; denotes significant increase.

Overall, the results of the MK test showed that decreasing trends for both the observed and natural streamflow were observed in the HRB from 1956 to 2000 (Figure 3 and Table 2). For the natural streamflow, except for the Shenqiu station, Xuanwu station, and Mingguang station, a decreasing trend for annual streamflow of the other 17 stations was observed. However, the observed streamflow of all stations have more significant decreasing trends than the natural streamflow. Meanwhile, the observed streamflow of six stations passes the 5% significance decreasing test, among which were four in the north of the HRS and two in the YRS, but none for the natural streamflow. For both the observed and natural streamflows, it can also be found that compared with the stations in the south of the Huaihe River, the downward trend of annual streamflow in the north appeared more significant, and streamflow along the Huaihe River displayed an aggravated decreasing trend from the upper to the lower reach. The possible reasons may be the weakening northward propagation of water vapor, intensively decreasing precipitation, and the influence of anthropogenic activities such as irrigation and increased population in the north of the HRB (Zhang et al. 2011b).

The value of slope during the MK test reflects the change rate being analyzed (Table 2). Except for Xixian, Wujiadu, Bantai, and Hengpaitou stations, the decreasing rate of the observed streamflow was greater than that of the natural streamflow. For both the observed and natural streamflow, the absolute values of slope were of spatially significant differences in the 20 stations; the observed streamflow increases from about 0.03 m3·s−1·yr−1 at Mingguang station to about 6.09 m3·s−1·yr−1 at Zhongdu station, and that of the natural streamflow increases from 0.01 m3·s−1·yr−1 at Yongcheng station to 5.44 m3·s−1·yr−1 at Wujiadu station. The absolute value of slope in the Huaihe River increases gradually downstream from 0.07–4.03 m3·s−1·yr−1 for the observed streamflow and 0.02–5.44 m3·s−1·yr−1 for the natural streamflow, respectively. Furthermore, the change rate of streamflow correlates significantly (p < 0.01) with the drainage area, and it is shown that the change rate declines with increasing drainage area (Figure 4). This may be due to the capacity of resisting disturbance decrease with the drainage area, that under the same impact of climate changes or human activities, the slope of MK test becomes higher along with the smaller drainage area.

Figure 4

Correlation between drainage area and the slope value of the MK test: (a) natural streamflow and (b) observed streamflow.

Figure 4

Correlation between drainage area and the slope value of the MK test: (a) natural streamflow and (b) observed streamflow.

The natural streamflow data are mainly obtained by the sub-investigation method, which mainly considered agricultural water consumption, industrial water consumption, loss of the reservoir evaporation leakage and other reductions. The data have been validated with historical data, and the restored streamflow data could reflect the state of runoff without the influence of human activities. However, the statistical deviation of agricultural water consumption, industrial water consumption, etc., may bring uncertainty to the natural streamflow data restoration.

Trends for seasonal streamflow

Analysis of trends in seasonal streamflow provides much greater temporal detail, and can help to reveal and understand the cause of trends in annual streamflow. The spatial distribution of trends for seasonal streamflow in the HRB over the whole available period is displayed in Figure 5. The MK testing statistics values (Z) for seasonal streamflow throughout the whole period of the 20 stations are presented in Figure 6.

Figure 5

Spatial distribution of trends for seasonal streamflow over the HRB: (a) natural streamflow and (b) observed streamflow. denotes significant decrease; denotes no significant decrease; denotes significant increase; denotes no significant increase.

Figure 5

Spatial distribution of trends for seasonal streamflow over the HRB: (a) natural streamflow and (b) observed streamflow. denotes significant decrease; denotes no significant decrease; denotes significant increase; denotes no significant increase.

Figure 6

The MK testing statistics values (Z) for seasonal streamflow over the HRB: (a) natural streamflow and (b) observed streamflow. (The boxes indicate the 25th and 75th percentiles; the whiskers indicate the lowest and highest data value; and strigula indicate the 50th percentiles value.).

Figure 6

The MK testing statistics values (Z) for seasonal streamflow over the HRB: (a) natural streamflow and (b) observed streamflow. (The boxes indicate the 25th and 75th percentiles; the whiskers indicate the lowest and highest data value; and strigula indicate the 50th percentiles value.).

For the whole available period, the natural and observed streamflow of the entire basin had a statistically decreasing trend in each season (Figures 5 and 6). For the natural streamflow, the stations with decreasing streamflow accounted for 65%, 85%, 55%, and 50% of the total stations in spring, summer, autumn, and winter in the HRB, respectively. However, that of the observed streamflow accounted for 95%, 95%, 80%, and 80%, respectively (Figure 5). The highest decreasing trends for two kinds of streamflow occurred in spring (lower than −1.96, α= 0.05 level), and the lowest ones were in autumn and winter, respectively (also negative). It can obviously be found that the observed streamflow has more significant decreasing trends than the natural streamflow (Figures 5 and 6). Meanwhile, more stations for both the observed and natural streamflow were associated with the decreasing streamflow in spring and summer compared to autumn and winter. There is an important reason for this in that corn and winter wheat are common crops in the HRB and require massive amounts of water during their growing periods in spring and summer. Besides, scarce precipitation and stronger evaporation during spring in the HRB also contribute significantly to the decrease in streamflow (Zhang et al. 2011a).

In addition, for the observed streamflow, the stations with statistically significant decreasing streamflow accounted for 35%, 30%, 30%, and 40% of the total stations in spring, summer, autumn, and winter, respectively. However, there was less than 10% of the total stations with significant decreasing trends for the natural streamflow at the same time (Figure 5). This result further indicates that the observed streamflow has more significant decreasing trends than the natural streamflow. There are four stations with statistically significant decreasing trends being the same in four seasons and these are mainly located in the north of the middle catchment of the HRS and include Zhuanqiao (Huijihe River), Xuanwu (Wohe River), Mengcheng (Wohe River), and Yongcheng station (Tuohe River) (Figure 5).

Change point for annual streamflow

The sequential MK and distribution-free CUSUM tests were performed to detect the change points of the observed annual streamflow in the six stations with the statistically decreasing trend at significance level 0.05. The identified change points are given in Table 3 and the corresponding graphs are shown in Figure 7.

Table 3

Years with abrupt change detected using the sequential MK and distribution-free CUSUM tests

Station Direction of trend MK CUSUM Station Direction of trend MK CUSUM 
Zhuanqiao Downward 1988 1980 Xuanwu Downward 1988 1987 
Mengcheng Downward 1981 1980 Yongcheng Downward 1981 1985 
Linyi Downward 1965 1965, 1975 Daguanzhuang Downward 1965 1975 
Station Direction of trend MK CUSUM Station Direction of trend MK CUSUM 
Zhuanqiao Downward 1988 1980 Xuanwu Downward 1988 1987 
Mengcheng Downward 1981 1980 Yongcheng Downward 1981 1985 
Linyi Downward 1965 1965, 1975 Daguanzhuang Downward 1965 1975 
Figure 7

Change points of the observed annual streamflow in the six stations with statistically decreasing trend at significance level 0.05 detected by the sequential MK test (UB-UF) and the distribution-free CUSUM test (Vk).

Figure 7

Change points of the observed annual streamflow in the six stations with statistically decreasing trend at significance level 0.05 detected by the sequential MK test (UB-UF) and the distribution-free CUSUM test (Vk).

The results show that the change points detected using the two tests were similar. Among the six stations, the earliest ones occurred in 1965 in Linyi and Daguanzhua stations of the YRS by the sequential MK test, while in Daguanzhua station another change point was detected in 1975 by the CUSUM test. This was consistent with the conclusion of Xue & Tan (2011). They found that due to the construction of mega-reservoirs, largely done from the early 1960s to mid-1970s in the YRS, the runoff of the upper reaches of Linyi station has two change points that appeared in 1965 and 1975, respectively. In addition, change points in four other stations in the HRS mostly happened in the 1980s, which was consistent with rapid economic development, population growth, and substantially increasing water withdrawal in this basin since the 1980s (Gao et al. 2009; Yang et al. 2010). Therefore, human activity may be considered the important factor driving runoff abrupt change in the HRB. As a whole, it can be concluded that the change points of the observed annual streamflow in the HRS happened in the 1980s, and in the YRS happened in 1965 and 1975.

Trends for climate

The MK test was applied to analyze the trends of precipitation based on the data from 14 meteorological stations in the HRS and the YRS from 1956 to 2008 (Figures 8 and 9). The result of the trend analysis shows there are different trends of annual precipitation in different areas in the HRB (Table 4). In the HRS, the annual precipitation showed no significant increase, while in the YRS the annual precipitation showed no significant decrease.

Table 4

Results of MK and Pettitt tests for regional annual precipitation

Sub-basin MK test
 
Sequential MK test CUSUM test 
Trend T T 
Huaihe River Basin (HRB) −1.15 ↓ – – 
Huaihe River System (HRS) −0.63 ↓ – – 
Yishusi River System (YRS) −2.52* ↓ 1967 1971, 1975 
Sub-basin MK test
 
Sequential MK test CUSUM test 
Trend T T 
Huaihe River Basin (HRB) −1.15 ↓ – – 
Huaihe River System (HRS) −0.63 ↓ – – 
Yishusi River System (YRS) −2.52* ↓ 1967 1971, 1975 

*Delineates significance at 0.05 significance level.

Figure 8

Spatial distribution of trends for annual precipitation over the HRB: denotes no significant decrease; denotes no significant increase.

Figure 8

Spatial distribution of trends for annual precipitation over the HRB: denotes no significant decrease; denotes no significant increase.

Figure 9

The Z values of MK test of annual precipitation.

Figure 9

The Z values of MK test of annual precipitation.

The spatial distribution of seasonal precipitation trends in the HRB is displayed in Figure 10. The trend of precipitation in spring and autumn showed a decrease, and the meteorological stations with decreasing precipitation accounted for 71.4% and 100% of the total meteorological stations in spring and autumn, while there were two stations (Luoshan and Fuyang) presenting a significant decrease in the autumn. The precipitation had an increasing trend in summer and winter in the HRB, and the meteorological stations with increasing precipitation accounted for 85.7% and 100% of the total meteorological stations in summer and winter. Among them, there was a significant increase in Bengbu in summer and there were three stations (Yongcheng, Mengcheng, and Huoshan) presenting a significant increase in winter. Through the runoff trend analysis, the natural and observed streamflow of the entire basin had a statistically decreasing trend in each season. Therefore, in summer and winter, the changes of runoff were greatly influenced by human activities in the HRB.

Figure 10

Spatial distribution of trends for seasonal precipitation over the HRB: denotes significant decrease; denotes no significant decrease; denotes significant increase; denotes no significant increase.

Figure 10

Spatial distribution of trends for seasonal precipitation over the HRB: denotes significant decrease; denotes no significant decrease; denotes significant increase; denotes no significant increase.

Furthermore, the sequential MK test was used to analyze the trends of annual average temperature during 1956–2008 in the HRB (Figures 11 and 12). The result of trend analysis shows that increasing trend of annual average temperature was detected in the HRB.

Figure 11

Spatial distribution of trends for annual temperature over the HRB: denotes significant increase; denotes no significant increase.

Figure 11

Spatial distribution of trends for annual temperature over the HRB: denotes significant increase; denotes no significant increase.

Figure 12

The Z values of MK test annual temperature.

Figure 12

The Z values of MK test annual temperature.

The sequential MK test was used to analyze the change points of 13 meteorological stations where the annual average temperatures presented a significant increase, as seen as Figure 13. The results show that the change points of most meteorological stations occurred in the 1990s. Among the 13 meteorological stations, the earliest one change point occurred in 1987 in Yongcheng, while the latest occurred in 2004 in Yuzhou. Overall, the change points for temperature occurred later than those for streamflow. Thereafter, except for the influence of temperature, the streamflow is influenced by many other factors in the HRB. For example, in the early 1980s, the surface water consumption rate in the HRB was 51% and increased to 87% in 1997 (Xu 2005), which could contribute to the decrease of the observed streamflow of the HRB.

Figure 13

Change points of the annual temperature in 13 stations (except for Juxian station) with statistically decreasing trend at significance level 0.05 detected by the sequential MK test (UB-UF).

Figure 13

Change points of the annual temperature in 13 stations (except for Juxian station) with statistically decreasing trend at significance level 0.05 detected by the sequential MK test (UB-UF).

Contribution of climate variability and human activities to streamflow change

The impacts of climate variability and human activities on annual streamflow of the six stations with the statistically decreasing trend at significance level 0.05 were analyzed, as shown in Table 5 and Figure 14.

Table 5

Changes of annual streamflow by the change points (by the sequential MK test) of the six stations with the statistically decreasing trend at significance level 0.05

Station Period Year QOa (m3/s) Qna (m3/s) 
Zhuangqiao Natural period 1956–1987 15.70 7.55 
Impacted period 1988–2000 5.15 5.43 
Change  10.55 2.12 
Xuanwu Natural period 1959–1986 7.53 7.50 
Impacted period 1987–2000 1.69 6.37 
Change  5.84 1.12 
Mengcheng Natural period 1956–1980 54.97 47.72 
Impacted period 1981–2000 18.23 34.03 
Change  36.74 13.69 
Yongcheng Natural period 1956–1980 5.26 5.36 
Impacted period 1981–2000 1.13 4.32 
Change  4.13 1.04 
Linyi Natural period 1958–1966 115.38 123.90 
Impacted period 1967–200 46.23 76.26 
Change  69.15 47.64 
Daguanzhuang Natural period 1958–1966 46.79 49.21 
Impacted period 1967–2000 25.99 36.03 
Change  20.80 13.19 
Station Period Year QOa (m3/s) Qna (m3/s) 
Zhuangqiao Natural period 1956–1987 15.70 7.55 
Impacted period 1988–2000 5.15 5.43 
Change  10.55 2.12 
Xuanwu Natural period 1959–1986 7.53 7.50 
Impacted period 1987–2000 1.69 6.37 
Change  5.84 1.12 
Mengcheng Natural period 1956–1980 54.97 47.72 
Impacted period 1981–2000 18.23 34.03 
Change  36.74 13.69 
Yongcheng Natural period 1956–1980 5.26 5.36 
Impacted period 1981–2000 1.13 4.32 
Change  4.13 1.04 
Linyi Natural period 1958–1966 115.38 123.90 
Impacted period 1967–200 46.23 76.26 
Change  69.15 47.64 
Daguanzhuang Natural period 1958–1966 46.79 49.21 
Impacted period 1967–2000 25.99 36.03 
Change  20.80 13.19 
Figure 14

Contribution of climate variability and human activities for annual streamflow by the change point (by the sequential MK test) of the six stations with the statistically decreasing trend at significance level 0.05.

Figure 14

Contribution of climate variability and human activities for annual streamflow by the change point (by the sequential MK test) of the six stations with the statistically decreasing trend at significance level 0.05.

As can be seen from Table 5 and Figure 7, human activities (ηH) played a dominant role in the HRS. On the contrary, climate variability (ηC) was the major driving factor in the YRS. This conclusion was consistent with the results of annual streamflow and annual precipitation in the HRS and the YRS. Taking Zhuangqiao station as an example, the observed average annual streamflow decreased from 15.70 m3/s to 5.15 m3/s in the ‘natural period’ and ‘impacted period’, respectively. Meanwhile, the natural average annual streamflow decreased from 7.55 m3/s to 5.43 m3/s in the ‘natural period’ and ‘impacted period’, respectively. That meant climate variability and human activities accounted for 20.08% (2.12 m3/s out of 10.55 m3/s) and 79.92%, respectively (Figure 7). Similarly, the contribution of climate variability accounted for 19.22%, 37.27%, 25.20%, 68.89%, and 63.38%, and human activities accounted for 80.78%, 62.73%, 74.80%, 31.11%, and 36.62% to the streamflow decrease in the Xuanwu, Mengcheng, Yongcheng, Linyi and Daguanzhuang stations, respectively. The above results show that the stations driven by human activities are Xuanwu, Zhuangqiao, Yongcheng, and Mengcheng located in the HRS, and the other two in the YSR are driven by climate variability.

CONCLUSIONS

In this study, the variations of observed streamflow and natural streamflow from 20 hydrological stations in the HRB were analyzed. The following conclusions were obtained:

  1. Both observed and natural streamflow in the whole basin show a downward trend from 1956 to 2000. However, the observed streamflow of all stations have decreasing trends which were more significant than those of the natural streamflow. Meanwhile, the decreasing rate of observed streamflow is higher than that of the natural streamflow. Furthermore, negative trends dominate the four seasons in the HRB. The highest decreasing trends for two kinds of streamflow occurred in spring (lower than −1.96, α = 0.05 level), and the lowest ones were in autumn and winter, respectively (also negative).

  2. The observed streamflow of six stations, including four stations in the north of the HRS and two stations in the YRS, showed significant downward trend (α= 0.05 level), but none for the natural streamflow. The change points of the observed annual streamflow in the HRS happened in the 1980s, and in the YRS occurred in 1965 and 1975. For the annual precipitation, a slightly decreasing trend was detected in the HRS, but a statistical significant decreasing trend at α = 0.05 level occurred in the YRS. The change points of annual precipitation occurred in 1965 and the 1970s, which was basically consistent with that of the observed annual streamflow in the YRS.

  3. According to the change point of the annual streamflow, the whole period was divided into two periods: ‘natural period’ (before the change point) and ‘impacted period’ (after the change point). By comparing the difference of average annual streamflow in these two periods, quantitative assessment of streamflow variation under the attribution climate change and human activities was conducted. The results indicated that human activities played a dominant role in the HRS, including at Xuanwu (80.78%), Zhuangqiao (79.92%), Yongcheng (74.80%), and Mengcheng stations (64.73%). On the contrary, climate variability was the main driving factor for streamflow decrease in the YRS, including Daguanzhuang (63.28%) and Linyi stations (68.89%).

This study can provide an important scientific basis for water resources allocation, regional water transfer, water ecosystem security, and so on in the HRB. Meanwhile, it has a certain scientific significance on the assessment of climate change and human activities on hydrological processes in the HRB.

ACKNOWLEDGEMENTS

This research was supported by the National Natural Science Foundation of China (Grant No. 41230640) and the National Key Project of Science and Technology, ‘Pollution Control on the Huaihe River basin (No. 2012ZX07204-003)’.

REFERENCES

REFERENCES
Bao
,
Z. X.
,
Zhang
,
J. Y.
,
Wang
,
G. Q.
,
Fu
,
G. B.
,
He
,
R. M.
,
Yan
,
X. L.
,
Jin
,
J. L.
,
Liu
,
Y. L.
&
Zhang
,
A. J.
2012
Attribution for decreasing streamflow of the Haihe River basin, northern China: climate variability or human activities?
J. Hydrol.
460–461
,
117
129
.
Chen
,
H.
,
Xiang
,
T.
,
Zhou
,
X.
&
Xu
,
C. Y.
2012
Impacts of climate change on the Qingjiang Watershed's runoff change trend in China
.
Stoch. Env. Res. Risk. A.
26
(
6
),
847
858
.
Chiew
,
F. H. S.
&
McMahon
,
T. A.
1993
Detection of trend or change in annual flow of Australian rivers
.
Int. J. Climatol.
13
,
643
653
.
Gao
,
C.
,
Gemmer
,
M.
,
Zeng
,
X.
,
Liu
,
B.
,
Su
,
B.
&
Wen
,
Y.
2009
Projected streamflow in the Huaihe River Basin (2010–2100) using artificial neural network
.
Stoch. Env. Res. Risk. A.
24
,
685
697
.
Hao
,
T. T.
,
Zhong
,
P. A.
&
Wei
,
P.
2011
Variation characteristics analysis of natural runoff series of Huaihe River Basin over past 50 years
.
Water Resources and Power
29
(
9
),
4
7
(in Chinese).
Jiang
,
Y.
,
Peng
,
Q. D.
,
Luo
,
H. H.
&
Ma
,
W.
2011b
Analysis of spatial and temporal variation of water quality in Huaihe River Basin
.
Journal of Hydraulic Engineering
42
(
11
),
1283
1288
(in Chinese).
Kahya
,
E.
&
Kalayci
,
S.
2004
Trend analysis of streamflow in Turkey
.
J. Hydrol.
289
(
1–4
),
128
144
.
Kundzewicz
,
Z. W.
&
Robson
,
A. J.
2004
Change detection in hydrological records – a review of the methodology
.
Hydrolog. Sci. J.
49
,
1
19
.
Lettenmaier
,
D. P.
,
Wood
,
E. F.
&
Wallis
,
J. R.
1994
Hydro-climatological trends in the continental United States, 1948–88
.
J. Climate.
7
,
586
607
.
Li
,
D. L.
,
Wang
,
W. S.
,
Hu
,
S. X.
&
Li
,
Y. Q.
2012
Characteristics of annual runoff variation in major rivers of China
.
Hydrol. Process.
26
,
2866
2877
.
Lins
,
H. F.
&
Slack
,
J. R.
1999
Streamflow trends in the United States
.
J Geophys. Res.
26
(
2
),
227
230
.
Miao
,
C. Y.
,
Shi
,
W.
,
Chen
,
X. H.
&
Yang
,
L.
2012
Spatio-temporal variability of streamflow in the Yellow River: possible causes and implications
.
Hydrolog. Sci. J.
57
(
7
),
1355
1367
.
Mitchell, J. M., Dzerdzeevskii, B., Flohn, H., Hofmeyr, W. L., Lamb, H. H., Rao, K. N. & Wallen, C. C
.
1966
Climate Change, WMO Technical Note no. 79. World Meteorological Organization, Geneva
.
Partal
,
T.
&
Kahya
,
E.
2006
Trend analysis in Turkish precipitation data
.
Hydrol. Process.
20
,
2011
2026
.
Shi
,
H. L.
,
Hu
,
C. H.
,
Wang
,
Y. G.
&
Tian
,
Q. Q.
2012
Variation trend and cause of runoff and sediment load variations in Huaihe River
.
J. Hydraul. Eng.
43
(
5
),
571
579
(in Chinese).
Shi
,
P.
,
Ma
,
X.
,
Hou
,
Y.
,
Li
,
Q.
,
Zhang
,
Z.
,
Qu
,
S.
,
Chen
,
C.
,
Cai
,
T.
&
Fang
,
X.
2013
Effects of land-use and climate change on hydrological processes in the upstream of Huai River, China
.
Water Resour. Manage.
27
,
1263
1278
.
Tao
,
H.
,
Gemmer
,
M.
,
Bai
,
Y. G.
,
Su
,
B. D.
&
Mao
,
W. Y.
2011
Trends of streamflow in the Tarim River Basin during the past 50 years: human impact or climate change?
J. Hydrol.
400
(
1–2
),
1
9
.
Vörösmarty
,
C. J.
,
Green
,
P.
,
Salisbury
,
J.
&
Lammers
,
R. B.
2000
Global water resources: vulnerability from climate change and population growth
.
Science
289
(
5477
),
284
288
.
Xia
,
J.
,
Du
,
H.
,
Zeng
,
S. D.
,
She
,
D. X.
,
Zhang
,
Y. Y.
,
Yan
,
Z. Q.
&
Ye
,
Y.
2012
Temporal and spatial variations and statistical models of extreme runoff in Huaihe River Basin during 1956–2010
.
J. Geogr. Sci.
22
(
6
),
1045
1060
.
Xu
,
B. B.
2005
Status quo, problems and measures of water resources development and utilization in Huai river basin
.
China Water Resource.
22
,
26
27
.
Xue
,
L. F.
&
Tan
,
H. Q.
2011
Variations of the hydrological characteristics and driving factors in the Yihe River Basin
.
Progress in Geography
30
(
11
),
1354
1360
(in Chinese).
Yan
,
R. H.
,
Huang
,
J. C.
,
Wang
,
Y.
,
Gao
,
J. F.
&
Qi
,
L. Y.
2016
Modeling the combined impact of future climate and land use changes on streamflow of Xinjiang Basin, China
.
Hydrol. Res.
Doi: 10.2166/nh.2015.206
.
Yang
,
C. G.
,
Lin
,
Z. H.
,
Yu
,
Z. B.
,
Hao
,
Z. C.
&
Liu
,
S. F.
2010
Analysis and simulation of human activity impact on streamflow in the Huaihe River Basin with a large-scale hydrologic model
.
J. Hydrometeorol.
11
(
3
),
810
821
.
Yang
,
X. L.
,
Chen
,
H. L.
,
Wang
,
Y. L.
&
Xu
,
C.-Y.
2016
Evaluation of the effect of land use/cover change on flood characteristics using an integrated approach coupling land and flood analysis
.
Hydrol. Res.
DOI: 10.2166/nh.2016.108
.
Yue
,
S.
&
Wang
,
C. T.
2002
Applicability of prewhitening to eliminate the influence of serial correlation on the Mann-Kendall test
.
Water Resour. Res.
38
(
6
),
4-1
4-7
.
DOI: 10.1029/2001WR000861
.
Zhang
,
X. B.
,
Harvey
,
K. B.
,
Hogg
,
W. D.
&
Yuzyk
,
T. R.
2001
Trends in Canadian streamflow
.
Water Resour. Res.
37
,
987
998
.
Zhang
,
S. R.
,
Lu
,
X. X.
,
Higgitt
,
D. L.
,
Chen
,
C. T.
,
Han
,
J.
&
Sun
,
H.
2008
Recent changes of water discharge and sediment load in the Zhujiang (Pearl River) Basin, China
.
Global Planet. Change
60
,
365
380
.
Zhang
,
Q.
,
Xu
,
C. Y.
&
Yang
,
T.
2009
Variability of water resource in the Yellow River Basin of past 50 years, China
.
Water Resour. Manage.
23
,
1157
1170
.
Zhang
,
Q.
,
Singh
,
V. P.
,
Sun
,
P.
,
Chen
,
X.
,
Zhang
,
Z.
&
Li
,
J.
2011a
Precipitation and streamflow changes in China: changing patterns, causes and implications
.
J. Hydrol.
410
,
204
216
.
Zhang
,
Q.
,
Xu
,
C. Y.
,
Chen
,
X. H.
&
Zhang
,
Z. X.
2011b
Statistical behaviors of precipitation regimes in China and their links with atmospheric circulation 1960–2005
.
Int. J. Climatol.
31
,
1665
1678
.
Zhang
,
L. L.
,
Podlasly
,
C.
,
Ren
,
Y.
,
Feger
,
K. H.
,
Wang
,
Y. H.
&
Schwärzel
,
K.
2013a
Separating the effects of changes in land management and climatic conditions on long-term streamflow trends analyzed for a small catchment in the Loess Plateau region, NW China
.
Hydrol. Process.
28
(
3
),
1284
1293
.
DOI: 10.1002/hyp.9663
.
Zhang
,
Q.
,
Singh
,
V. P.
,
Xu
,
C. Y.
&
Chen
,
X.
2013b
Abrupt behaviours of streamflow and sediment load variations of the Yangtze River basin, China
.
Hydrol. Process.
27
,
444
452
.
DOI: 10.1002/hyp.9278
.