Journal of Hydrology 551 (2017) 217–232
Contents lists available at ScienceDirect
Journal of Hydrology
journal homepage: www.elsevier.com/locate/jhydrol
Research papers
Observed changes in flow regimes in the Mekong River basin
Dongnan Li a, Di Long a, Jianshi Zhao a,⇑, Hui Lu b, Yang Hong a
a
b
State Key Laboratory of Hydro-Science and Engineering, Department of Hydraulic Engineering, Tsinghua University, Beijing, China
Center for Earth System Science, Tsinghua University, Beijing, China
a r t i c l e
i n f o
Article history:
Received 12 December 2016
Received in revised form 25 April 2017
Accepted 30 May 2017
Available online 1 June 2017
This manuscript was handled by Tim R.
McVicar, Editor-in-Chief, with the assistance
of Sergio M. Vicente-Serrano, Associate
Editor
Keywords:
IHA
Eco-flow metrics
Reservoir operation
Transboundary river
a b s t r a c t
Human activities, such as dam construction, significantly altered the flow regimes in the Mekong River,
particularly after the completion of two large dams, namely Xiaowan and Nuozhadu in 2010 and 2014,
respectively. Streamflow data from 1960 to 2014 obtained from five stations located along the Mekong
mainstream are divided into three periods, i.e., the pre-impact period (1960–1991), the transition period
(1992–2009), and the post-impact period (2010–2014). The flow regimes were investigated using ecoflow metrics and indicators of hydrologic alteration (IHA). The results show that the construction and initial filling of the upstream dams reduced the annual streamflow in the upstream Chiang Saen gauging
station, whereas no clear effect was observed in the downstream Stung Treng station. The operation of
dams reduces the streamflow in wet seasons and increases the streamflow in dry seasons, resulting in
a unique seasonal variation in the streamflow based on eco-flow metrics in the Chiang Saen gauging station, observed from 2010 to 2014. In addition, the maximum flow values decreased significantly in the
Chiang Saen gauging station during the year corresponding to the completion of the upstream dams.
The construction and operation of dams clearly have significant impacts on low pulse duration. It is
observed that climate change dictated the changes in the annual streamflow during the transition period
1992–2009 (82.28%), whereas human activities contributed more in the post-impact period 2010–2014
(61.88%). The results of this study could provide a reference for reservoir operation in the upstream
regions considering both ecological and economic benefits of such operations, as well as maximize the
interests of stakeholders in this region.
Ó 2017 Elsevier B.V. All rights reserved.
1. Introduction
In river ecosystems, the flow regime of runoff plays a significant
role in many fundamental ecological processes (Poff and
Zimmerman, 2010). Changes in flow regimes within the context
of climate change and human activities are significant to the
hydrological community, receiving considerable global attention.
River management currently focuses not only on the total volume
of runoff but also on its flow regime, which has been an important
objective in environmental systems for both ecological professionals and engineering managers since the 1990s (Richter et al., 1996;
Yin et al., 2011).
Climate change and human activities have been considered as
the two primary factors affecting flow regimes (Li et al., 2006;
Ma et al., 2014). In some basins, human activities are the main factors that alter flow regimes, particularly during the construction
and operation of large reservoirs (Poff et al., 1997; Fan et al.,
2015). Climate change can also be the dominant factor that alters
⇑ Corresponding author.
E-mail address: zhaojianshi@tsinghua.edu.cn (J. Zhao).
http://dx.doi.org/10.1016/j.jhydrol.2017.05.061
0022-1694/Ó 2017 Elsevier B.V. All rights reserved.
flow regimes (Li et al., 2006), which can change the pattern of precipitation and potential evaporation (Wang and Hejazi, 2011).
Human activities, such as dam construction and water withdrawal
activities (i.e., irrigation, industry, and municipal demands),
directly change flow regimes (Ma et al., 2014), thereby changing
river ecosystems (Gippel, 2001; Poff et al., 1997; Richter et al.,
2003, 2006). The impacts of the two factors are often analyzed separately. However, the effects of climate change and human activities are always combined, and their effects on some river basins are
difficult to identify. For example, climate change may cause
changes in precipitation, increasing the impacts caused by dams
as more water would be regulated by reservoirs during the long
dry seasons (Lu et al., 2014).
Many types of flow metrics and statistical methods have been
proposed to analyze the impacts of changes in flow regimes. However, the characteristics of such changes are not fully understood
from the perspectives of both ecological and human demands. Over
170 hydrologic metrics have been proposed to describe the variations and characteristics of flow regimes (Olden and Poff, 2003;
Gao et al., 2009). Indicators of hydrologic alteration (IHA) are
among the most popular metrics being widely used. The range of
218
D. Li et al. / Journal of Hydrology 551 (2017) 217–232
Table 1
Main conclusions on streamflow alteration in recent relevant research.
Study
Mean flow
Dry season flow
(Lu et al., 2008)
Post-dam period (1992–2003)
had lower water levels than the
pre-dam period (1962–1991).
1992 was the driest year since 1960, but water flow in dry
season was not the lowest.
(Kummu and Sarkkula,
2008)
Wet season flow
Small rises in the dry-season lake water level would
permanently inundate disproportionately large areas of
floodplain in Tonle Sap.
(Delgado et al., 2010)
(Piman et al., 2012)
(Räsänen et al., 2012)
(Cochrane et al., 2014)
At the 3S (Sekong, Sesan and Srepok) rivers of Mekong,
hydropower projects increased flows by 28% in the
current period and that number will be 63% after the
completed of all the dams.
A 90% increase in December–May
Flows in Chiang Saen after the construction of 6 Chinese
dams.
Compared with pre-1991 period, mean water levels for
Chiang Saen increased in excess of 30% for the dry season
months of March and April
(Vries et al., 2015)
(Räsänen et al., 2017)
Discharge in March-May 2014 increased by 121–187%,
41–74% compared to average discharges in Chiang Saen
and Kratie respectively.
variability approach (RVA) was proposed to measure hydrologic
changes (Richter et al., 1997), suggesting that the 25th and 75th
percentiles of IHA metrics should be the targets for maintaining
environmental flow (Yin et al., 2011). Based on the concept of flow
duration curves (FDCs), Vogel et al. (2007) proposed an eco-flow
metric comprising eco-deficit and eco-surplus. The two indices
are non-dimensional and directly show the deficit and surplus of
streamflow at different periods. Despite their simplicity, ecodeficit and eco-surplus indices are a promising overall representation of the degree of streamflow alteration (Gao et al., 2009). Ma
et al. (2014) proposed a hydrograph-based hydrologic alteration
assessment containing 25 indicators, which not only describe the
statistics for each year but also consider the characteristics of
extreme flow events. The integrated use of these methods could
measure changes in flow regimes comprehensively.
Ecological effects of streamflow alteration in some large international transboundary rivers, such as the Mekong River, are particularly significant. These effects cause significant problems
affecting collaborative management among riparian countries
(Ingram et al., 1994; Kirby et al., 2010). The Mekong River is the
ninth largest river globally, and the Mekong River Basin (MRB) harbors one of the most productive and diverse ecosystems in South
Asia (Kuenzer et al., 2013). Ecological issues in the MRB have been
a major concern to stakeholders, researchers, and other professionals globally. Over 70 dams spread across six riparian countries
make the management of this transboundary river quite complicated and sensitive. Several studies focused on the hydrological
effects of hydropower dams on the Mekong River, particularly
the dams in China. Following the literature review published by
Lu et al. (2008), Table 1 shows the literature review of some important studies conducted after 2008. A common perception is that
dams significantly altered the flow regimes at the basin scale
(Kuenzer et al., 2013). Some researchers have argued that construction of all the 78 tributary dams would produce less energy
and pose greater environmental risks as compared to having six
mainstream dams upstream of the Mekong River (Ziv et al.,
2012). Recently, Cochrane et al. (2014) analyzed the alteration in
water levels in six streamflow gauging stations along the mainstream of the Mekong River from 1960 to 2010 and discussed the
relationship between annual fluctuations and active reservoir stor-
During 20th century, although average
magnitude floods have a negative trend,
variability is increasing.
Wet season flows will decreased by 4% and
22% during current period and future period
at the outlet of 3S river respectively.
A 20–22% decrease in June–November flows,
in Chiang Saen after the construction of 6
Chinese dams.
Compared with pre-1991 period, monthly
increases betweenJune and December were
mostly less than 5%.
Dams can only cause a small change in
water levels in the flood season until 2013.
Discharge in July-August 2014 decreased by
32–46%, 0–6% compared to average
discharges in Chiang Saen and Kratie
respectively.
age. However, the two largest reservoirs built after 2010, Xiaowan
and Nuozhadu, were not considered in the aforementioned study;
hence, the impacts of these reservoirs remain largely unknown.
In this study, we integrate the aforementioned methods, i.e.,
IHA for detailed metrics and eco-flow metrics for overall evaluation, to provide a complete analysis of the changes in the streamflow regime by taking the MRB as a case study. The following are
the objectives of this study: (1) conducting a comprehensive, systematic analysis to identify the alteration of the flow regimes
and their spatial and temporal patterns during the period 1960–
2014 in the five gauging stations located along the mainstream;
and (2) analyzing the causes and effects of the alteration of the
flow regimes in the MRB.
2. Study area and data
The Mekong River originates from the Tibet Plateau, flows
through six countries, namely China, Laos, Myanmar, Thailand,
Vietnam, and Cambodia, and then flows into the South China Sea
(Fig. 1). The MRB is generally divided into two sub-basins (Xi
et al., 2008). The upper Mekong Basin (UMB), which is known as
the Lancang River in China, covers an area of 195,000 km2 (24%
of the total drainage area) and flows through three provinces of
China, namely Qinghai, Tibet, and Yunnan. The lower Mekong
Basin (LMB), which covers an area of 600,000 km2 (76% of the
total drainage area) and flows through five countries, is generally
considered to exist in a near natural state because of less economic
development compared to the UMB (Hirsch, 2010; Kummu et al.,
2010; Piman et al., 2013). The Mekong River is more than
4180 km long with a drainage area of 795,000 km2 and a mean
annual streamflow of 14,500 m3/s (MRC, 2005). The streamflow
is dominated primarily by Southeast Asian monsoons, resulting
in a flood season and a dry season within a hydrologic year
(Cochrane et al., 2014).
Downstream of the LMB, several important ecological sites closely related to the flow regimes of the river exist; among them,
Lake Tonle Sap and the Mekong Delta are the most famous. Tonle
Sap, which is the largest lake in Southeast Asia (covering an area
of 8800 km2 on average) and connected with the mainstream of
the Mekong River at Phnom Penh, Cambodia, is regarded as the
D. Li et al. / Journal of Hydrology 551 (2017) 217–232
219
Fig. 1. Location of Mekong River Basin and the 6 completed dams and 19 planned dams along the mainstream.
‘‘natural reservoir” of the MRB. The water flowing into Tonle Sap
from the Mekong River during the wet season (June to October)
is nearly six times than that during the dry season (November to
May) (Cochrane et al., 2014). Tonle Sap is rich in aquatic products
and contains more than 300 species of freshwater fish. The down-
stream area of Phnom Penh is called the Mekong Delta, the area of
which is 55,000 km2 (4000 km2 in Vietnam; 11,000 km2 in Cambodia). The Mekong Delta is subjected to frequent drought and
flood events and influenced by tidal and seawater intrusion, particularly during dry seasons (Zhang et al., 2001).
220
D. Li et al. / Journal of Hydrology 551 (2017) 217–232
No dam was constructed on the mainstream of the Mekong
River until 1992. China has built six hydropower dams on the
mainstream of the Lancang River (i.e., the UMB) since 1992 (as
shown in Table 2). The two largest dams are Xiaowan (14.56 km3
in total storage, completed in 2010) and Nuozhadu (22.4 km3 in
total storage, completed in 2014), contributing to 36% and 55% of
the total storage capacity of all the existing reservoirs in the basin,
respectively. Although the streamflow regime has been altered
because of large-scale dams, the hydropower generation of the
Mekong River has kept increasing in recent years (MRC, 2010). In
the subsequent decades, 19 dams were planned to be built on
the mainstream of the Mekong River (8 in China, 10 in Laos, and
1 in Cambodia), and over 14 large dams (>10 MW) would be built
in the tributaries. The primary objective of these dams is hydropower generation, which is expected to yield considerable economic benefits for the riparian countries. Moreover, the
construction and operation of dams would improve navigation
conditions, increase irrigation opportunities, and enhance flood
control capacity (Lu et al., 2014). However, the construction and
operation of large dams would also affect the patterns of streamflow, resulting in multiple changes in streamflow regimes, thereby
causing a negative impact on ecosystems. The daily flow-time
series data obtained from the Mekong River Commission
(http://portal.mrcmekong.org/index) are used to explore these
comprehensive changes in the streamflow regimes in this study.
3. Methods
3.1. IHA metrics
The IHA metrics are the most widely used indicators to measure
changes in streamflow regimes. The IHA contains 33 hydrologic
indicators, which are generally categorized into the following six
groups (as listed in Table 3): 1) magnitude of monthly flows, 2) magnitude and duration of annual extreme flows, 3) timing of annual
extreme flows, 4) frequency and duration of high and low pulses,
5) rate and frequency of flow changes, and 6) zero flow events
(Richter et al., 1996). In this study, the indicators of groups (1) and
(2) are represented by the corresponding flow rates. The timing of
annual extreme flows is represented by a single number, which
implies the Xth day of a year (e.g., if minimum flow occurred on
February 10, the date of minimum flow is 41). The high and low
pulses are defined as the periods within a year wherein daily flow
exceeds the 75th percentile (high pulse) or drops below the 25th
percentile (low pulse) of the daily flow-time series under natural
conditions (i.e., 1960–1991 in MRB) (Gao et al., 2012). The base flow
index is calculated using the ratio of the seven-day minimum flow to
the annual mean flow. The rate and frequency of the flow changes
are represented using flow reversal from increasing to decreasing
or decreasing to increasing. Zero flow events do not exist in the MRB.
3.2. Eco-flow metrics
In addition to the IHA metrics, two other nondimensional metrics, namely, eco-deficit and eco-surplus, based on the FDC were
Table 3
Metrics of indicators of hydrologic alteration divided into six groups.
IHA statistics
Hydrologic parameters
Group 1
Mean flow in January
Mean flow in February
Mean flow in March
Mean flow in April
Mean flow in May
Mean flow in June
1-day maximum
3-day maximum
7-day maximum
30-day maximum
90-day maximum
Base flow index
Date of maximum
High pulse count
Low pulse count
Rise rate
Fall rate
Number of zero flow days
Group 2
Group 3
Group 4
Group 5
Group 6
Mean flow in July
Mean flow in August
Mean flow in September
Mean flow in October
Mean flow in November
Mean flow in December
1-day minimum
3-day minimum
7-day minimum
30-day minimum
90-day minimum
Date of minimum
High pulse duration
Low pulse duration
Number of reversals
introduced by Vogel et al. (2007). The FDC plots the streamflow
Qi as a function of its corresponding exceedance probability pi = i/
(n + 1), where i is the rank, and n is the total number of days.
The terms eco-surplus and eco-deficit were coined to represent
the overall loss or gain relative to a ‘‘reference scenario” or ‘‘post
scenario” for any period of interest (e.g., month, season, or year).
As shown in Fig. 2, the black line represents the unregulated FDC
corresponding to the ‘‘reference scenario,” while the red line represents the regulated FDC corresponding to the ‘‘post scenario.” The
area below the unregulated FDC and above the regulated FDC
(Zone 1) represents the amount of water deficiency in the riverine
ecosystems. The ratio of the area of Zone 1 to the total area under
the unregulated FDC (Zone 1 + Zone 3) is defined as the eco-deficit.
This ratio represents a fraction of the streamflow that is no longer
available to the river during that period. The area below the regulated FDC and above the unregulated FDC (Zone 2) represents the
amount of water surplus to the riverine ecosystems; accordingly,
the ratio of the area of Zone 2 to the areas of Zones 1 and 3 is
defined as the eco-surplus.
Ecodeficit ¼
Zone 1
Zone 1 þ Zone 3
Ecosurplus ¼
ð1Þ
Zone 2
Zone 1 þ Zone 3
ð2Þ
Fig. 2 Definitions of eco-surplus and eco-deficit in the flow
duration curve. Zones 1 and 2 represent eco-deficit and ecosurplus, respectively
In this study, the data of mean annual and seasonal FDCs of 2
flow gauges obtained from 1960 to 1991 are used to represent
the natural flow regime as the streamflow was in a near natural
state during this period. The red line represents the regulated
FDC. The annual and seasonal FDCs from 1960 to 2014 in two
streamflow stations are derived and the corresponding
Table 2
Characteristics of hydropower reservoirs in Upper Mekong Basin.
Dams
Year Completed
Total storage
(km3)
Hydropower capacity
(MW)
Dam height
Dam type
Manwan
Dachaoshan
Jinghong
Xiaowan
Gongguoqiao
Nuozhadu
1992
2003
2009
2010
2011
2014
0.92
0.93
1.23
14.56
3.16
22.4
1500
1350
1500
4200
900
5500
126
110
118
300
105
254
Gravity dam
Gravity dam
Gravity dam
Arch dam
Gravity dam
Gravity dam
D. Li et al. / Journal of Hydrology 551 (2017) 217–232
221
Based on the RC of Chiang Saen, an increasing trend was
observed on three occasions, which occurred between the completion of the first dam (Manwan) and the largest dam (Xiaowan).
Accordingly, the 55-year period is divided into the following three
periods based on the process of dam construction on the mainstream of the Mekong: (1) the pre-impact period (1960–1991, no
dam was constructed on the mainstream), (2) the transition period
(1992–2009, four dams with smaller capacities were constructed
on the mainstream), and (3) the post-impact period (2010–2014,
two of the largest dams were constructed on the mainstream).
3.5. Cubic spline smoothing
Fig. 2. Definitions of eco-surplus and eco-deficit in the flow duration curve. Zones 1
and 2 represent eco-deficit and eco-surplus, respectively.
eco-surplus and eco-deficit values are calculated to analyze the
changes in annual and seasonal streamflow.
Cubic spline smoothing is a method to demonstrate the longterm trend by fitting data using a piecewise cubic polynomial
curve. According to the length of series and need basis, divide
the sequence into m groups, applying least squares fitting to each
group, and then connect m to the piecewise curve. This method
reflects the trend of the sequence in a smooth way:
3.3 Mann–Kendall test
The Mann–Kendall test is a popular method used to test the
trend and intercept break point of long time series. UFk and UBk
are two key statistical parameters. Set mi as the cumulative number when the value of sample ‘‘i” is larger than sample ‘‘j”
(1 j i):
dk ¼
k
X
mi ;
FðtÞ ¼
8
y ðtÞ; a0 < t < a1
>
>
> 1
>
>
y ðtÞ; a1 < t < a2
>
>
> 2
>
>
>
<:
:
>
>
>
:
>
>
>
>
>
>
ym ðtÞ; am1 < t < am
>
>
:
ð8Þ
ð3Þ
ða0 < t 1 < a1 < a2 < . . . < am1 < t n < am Þ
i¼1
Assume that the original sequence is random and independent:
Eðdk Þ ¼
kðk 1Þ
;
4
varðdk Þ ¼
kðk 1Þð2k þ 5Þ
;
72
dk Eðdk Þ
UFk ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ;
varðdk Þ
ð4Þ
yk ðtÞ ¼ f k ðt; ak Þðk ¼ 1; 2; 3; . . . . . . mÞ
ð5Þ
where t is the order of data, ak is the interval point, F(t) is a piecewise function, and f k ðt; ak Þ is a cubic polynomial curve using the
least squares fitting method.
ð9Þ
ð6Þ
Reverse the time series and repeat the above process, obtaining
UBk. When the significant level a = 0.05 and UFk > 1.96, a significant ascend trend is observed, while a = 0.05 and UFk < 1.96
implies a significant descend trend. When UFk and UBk intersect
in the confidence interval, the intersection is a potential intercept
break point (Zhang and Lu, 2009; Moraes et al., 1998).
3.4. Regulating capacity of a basin
In this study, the regulating capacity (RC) is defined to characterize the storage capacity of the constructed dams by comparing
it to the drainage area. The RC represents the overall ability of
human activities to alter the streamflow.
RC ¼
a
;
A
ð7Þ
where A is the drainage area (103 km2), and a is the storage capacity, obtained by the summation of the active reservoir storage
within the drainage area (106 m3); hence, the RC (mm) represents
the regulated depth of the upstream area of a gauging station.
Fig. 3 shows the RC of the MRB. Increasing blocks can be observed
since the 1990s (Fig. 3), specifically after 2010, implying the rapid
development of damming in the MRB during those periods.
Fig. 3. Time series of regulating capacity in the MRB and mean annual flow at Stung
Treng and Chiang Saen gauging stations in the Mekong River.
222
D. Li et al. / Journal of Hydrology 551 (2017) 217–232
DQ m ¼ Q m2 Q m1 ;
3.6. The Budyko framework
Both climate change and human activities can result in the
alteration of flow regimes of the MRB. Modeling and understanding
the effects of climate change and human activities are crucial to
formulate adaption strategies for the basin (Kirby et al., 2010;
Räsänen et al., 2012; Lauri et al., 2012; Mainuddin et al., 2013).
The Budyko framework is a widely used tool to analyze this issue
(Wang et al., 2013; Liang et al., 2015). In this study, a decomposition method based on the Budyko hypothesis is used to address
this issue (Wang and Hejazi, 2011), as shown in Fig. 4.
Fig. 4 is a theoretical curve in which the evaporation ratio (E/P)
is the ratio of the mean annual evaporation to the mean annual
precipitation. The climate dryness index (PET/P) is the ratio of
the mean annual potential evaporation to the mean annual precipitation. The functional forms of the E/P and PET/P have been investigated by many researchers (Fu, 1981; Zhang et al., 2001; Yang
et al., 2008); Fu’s function is chosen for this study.
w 1=w
E
PET
PET
¼1þ
1þ
;
P
P
P
ð10Þ
where w is a parameter that varies with different basins. According
to Wang and Hejazi (2011), for a watershed without direct human
impact, if the climate (PET/P) becomes drier or wetter, the evaporation ratio (E/P) will change to a new status; however, it would still
follow the same Budyko-type curve. Moreover, direct human interferences could push the watershed to move only in the vertical
direction, i.e., change E/P by altering E. Hence, if a watershed moves
from A to B, as seen in Fig. 4, this change can be partitioned into two
phases, i.e., from A to B0 (induced by climate change) and from B0 to
B (induced by human activities). In this way, the two impact factors
for the changes in the flow regimes can be partitioned. For the longterm annual water balance, the soil-water storage is assumed to be
negligible, and the streamflow Q m can be expressed as
Qm ¼ P E
ð11Þ
Then, the climate-induced change in the streamflow DQmc and
the human-induced change of streamflow DQmh can be calculated
using
ð12Þ
1 1
w w
DQ mc ¼ P w1
ðPET w
ðP 2 P1 Þ
1
1 þ P1 Þ
1 1
w w
ðPET w
þ ½PET w1
1
1 þ P1 Þ
DQ mh ¼ DQ m DQ mc ;
1ðPET 2 PET 1 Þ
ð13Þ
ð14Þ
where Q m1 and Q m2 are the streamflow values for the periods before
and after the change, respectively.
The data for the precipitation, evaporation, and potential evaporation at the global scale from 1956 to 2015 were obtained from
the Global Land Data Assimilation System-1 (GLDAS-1) in which
PET was simulated by the NOAH model (ftp://hydro1.sci.gsfc.nasa.gov/data/s4pa/GLDAS/).
3.7. Linear regression model
A simple linear regression model was proposed by Beguería
et al. (2003) and was applied by López-Moreno et al. (2011) and
Vicente-Serrano et al. (2017) to simulate streamflow. Considering
that precipitation (P) and temperature (T) are the two major
climate-driven factors of streamflow, the effects of climatic factors
on streamflow change at a seasonal time-scale (both wet and dry
seasons) can be identified as:
Q sim ¼ a1 P þ a2 T þ a3
ð15Þ
RE ¼ Q sim Q obs
ð16Þ
RR ¼
RE
100%
Q obs
ð17Þ
where Qsim is the simulated seasonal streamflow according to the
simple linear regression model, Qobs is the observed seasonal
streamflow, the residual RE is defined as the difference between
simulated streamflow and observed streamflow, the relative residual (RR) is the ratio of residual to the mean seasonal streamflow,
and a1, a2, and a3 are coefficients. The effects of human activities
on streamflow can be estimated by RE and RR. RE represents the
changes of streamflow excluding the parts from the two climatic
driving factors (i.e., from human activities); thus RR stands for the
proportion of streamflow changes from human activities. This
regression model can be supplementary to the Budyko framework
to distinguish the effects of climate change and human activities
at the seasonal time-scale. Data from the pre-impact period
1960–1991 was used to regress the streamflow.
4. Results
The long-term time series of the daily streamflow (1960–2014)
for five gauging stations are analyzed to determine the changes in
the streamflow regimes. Among them, the Chiang Saen station,
which is the most upstream station located at the boundary of
the UMB and LMB, and the Stung Treng station, which is the
most downstream station, are the two selected representative
gauging stations. The results of the changes in the annual
streamflow are presented first, and then five groups of the IHA
metrics are analyzed to compare the streamflow regimes in the
three periods.
4.1. Changes in annual streamflow
Fig. 4. Budyko curve, ‘‘A” and ‘‘B” represent the pre-impact and post-impact
periods, respectively.
The cubic spline smoothing method (Durrleman and Simon,
1989) is employed to illustrate the trends in the annual streamflow. Fig. 5 shows the trends in the mean annual streamflow in
the Chiang Saen station and the Stung Treng station. The annual
D. Li et al. / Journal of Hydrology 551 (2017) 217–232
streamflow in the Chiang Saen station (Fig. 5a) ranged from
1933 m3/s to 4026 m3/s during the 50-year period. The maximum
and minimum annual streamflow occurred in 1965 and 1992,
respectively. A notable downward trend was observed at the Chiang Saen station in the post-impact period, which could be closely
related to the initial filling of the upstream dams completed during
this period, with a total filling volume of nearly 41.35 km3. If this
amount of storage is filled in 12 months, it will need a flow rate
of 1311 m3/s. Compared with the streamflow rates in Fig. 5a, this
is a significant impact for the area downstream of the Chiang Saen
station. However, a similar downward trend during 2003–2005
should be attributed to the extreme drought event happening in
2003–2004 (Campbell and Manusthiparom, 2004), because the
storage capacity of the Dachaoshan Reservoir, which was completed in 2003, is only 0.93 km3. Four large dams were constructed
and completed after the year 2009, namely the Jinghong, Xiaowan,
Gongguoqiao, and Nuozhadu (largest) Dams, which were completed in the years 2009, 2010, 2011, and 2014, respectively, with
active storage capacities of 1.23 km3, 14.56 km3, 3.16 km3, and
22.4 km3, respectively. The completion of these dams significantly
increased the RC values of Chiang Saen from 30 mm to 230 mm,
whereas the initial filling of these dams reduced the annual
streamflow in Chiang Saen located near the downstream region
(Fig. 5a). However, gauging stations located farther downstream
show a different trend because their distances from these dams
are considerable (Fig. 5). The impacts of the dams gradually
decreased, reflected by the fact that no clear trend is observed in
Stung Treng station. This is because only approximately 16% of
the streamflow in this station comes from the upstream dammed
area in China, and the downstream hydrologic processes dominated the flow regime. Given that the effects of the reservoirs
decrease from the upstream region to the downstream region,
223
the focus of this study is to analyze the most influenced upstream
station (Chiang Saen) and the least influenced downstream station
(Stung Treng).
The findings aforementioned can be cross-checked using the
Mann–Kendall test, which is often applied to determine trends
and change points (Hamed and Rao, 1998) in a time series (as
shown in Fig. 6). UFk exhibits a downward trend, whereas UBk
shows an upward trend after the intersection point at Chiang Saen
station when approaching the year 2009 (Fig. 6a), thereby implying
a decreasing trend in the annual streamflow, though this change is
not significant (within the upper and lower boundaries of 0.05 significant level). However, in Fig. 6e, no clear trend is identified
based on the UFk and UBk curves during the same period, indicating
that no decreasing trend occurs at Stung Treng station. These
results confirm that damming in the upstream area reduced the
annual streamflow in the upstream Chiang Saen station, whereas
no clear effect in the downstream Stung Treng station is observed
because of different streamflow contributions of the damming area
in China to the two stations. It should be noted that the effects of
the dams should be re-checked when longer series of data are
available. Considering that streamflow change could be caused
by both climate change and human activities, more details about
the effects of these two factors are analyzed in Section 4.6.
Figs. 7 and 8 show the decadal statistics of the annual eco-flow
metrics. Extremely low eco-deficit values were recorded from 2010
to 2014 in Chiang Saen station because the construction and initial
filling of the four dams reduced the annual streamflow, resulting in
a low eco-deficit value during this period. The eco-deficit at Stung
Treng from 2010 to 2014 was low, but within a normal range,
implying that the construction and operation of the upstream
dam had no significant impact, with a runoff contribution of only
16% from China.
Fig. 5. Mean annual streamflow in the five gauging stations (Chiang Saen, Luang Prabang, Nong Khai, Pakse, and Stung Treng) along the Mekong River.
224
D. Li et al. / Journal of Hydrology 551 (2017) 217–232
Fig. 6. Mann–Kendall test for the annual streamflow in the five gauging stations along the mainstream of the Mekong River.
Fig. 7. Boxplots of the annual eco-surplus and eco-deficit values in Chiang Saen
gauging station.
Fig. 8. Boxplots of annual eco-surplus and eco-deficit values at the Stung Treng
gauging station.
D. Li et al. / Journal of Hydrology 551 (2017) 217–232
4.2. Changes in seasonal streamflow
Changes in the monthly streamflow can be represented by the
changes in the seasonal streamflow in the MRB case, because reservoirs store water in wet seasons and release water in dry seasons
to support hydropower generation and other objectives, resulting
in similar effects on monthly flow within a season. Dry seasons
(November to May) and wet seasons (June to October) are defined
to identify characteristics of the seasonal streamflow (Lu et al.,
2014). Fig. 9 shows the annual eco-surplus and eco-deficit values
of the seasonal streamflow from 1960 to 2014. The eco-flow metrics are strongly correlated with the mean value of the streamflow
at seasonal scales. The analysis is first conducted in the Chiang
Saen station. The variance of eco-surplus was small and remained
generally stable until 2011. For the dry seasons of 2013 and 2014,
after the completion of the Xiaowan Dam, the eco-surplus values
reached up to 0.44 and 0.66, respectively, considerably higher than
the mean value of 0.09. The mean value of the eco-deficit for the
dry seasons was 0.06, whereas the mean value in 2010–2014
was 0.05. The fluctuation range in the eco-deficit is smaller than
that in the eco-surplus (Fig. 8a). For the wet seasons, the ecosurplus and eco-deficit values range from 0 to 0.65 and 0 to
0.32, respectively (Fig. 8b). In the 1960s, the streamflow was relatively large during wet seasons, with extremely high eco-surplus
values in the wet seasons of 1966 and 1971. Extremely low ecodeficit values were observed in the wet season of 1992 (completion
of the Manwan Dam), 2003 (completion of the Dachaoshan Dam),
and 2012–2014 (construction and completion of the Nuozhadu
225
Dam). The eco-surplus values of the downstream Stung Treng
station for the dry seasons showed an increasing trend over the
recent years, whereas no clear trend was observed for the wet
seasons.
Overall, the abnormal results in the eco-flow metrics can be
mostly attributed to the operation of the constructed dams. Operations for hydropower generation, flood control, and navigation
would certainly reduce the streamflow during the wet seasons,
but increase the streamflow during the dry seasons, resulting in
unique changes in the seasonal streamflow based on the ecoflow metrics in Chiang Saen station from 2010 to 2014. For
instance, according to the operation rules of the Xiaowan Dam, it
stores water in the refill period (mainly in October), and its water
level is raised to the maximum water level 1240 m at the end of
October. The reservoir releases water for hydropower generation
and other objectives during the dry season (November to May),
and its water level is reduced to the limited water level for flood
control at the end of May. During the flood season (June to October), it is primarily used to regulate large floods, but allows small
floods to pass for a normal year (Guo and Zhou, 2011). Note that
the streamflow at the Chiang Saen station contributes only 21%
to the streamflow of the Stung Treng station during the wet seasons (ranging from 13% to 30%), whereas the contribution during the dry seasons increases up to approximately 34% (fluctuation
ranging from 24% to 50%) (Fig. 10). This finding indicates that
the fluctuations in the upstream streamflow tend to affect the
downstream streamflow in the dry seasons more than in wet seasons. Therefore, the eco-flow metrics of the two stations in the dry
Fig. 9. Eco-flow metrics at the Chiang Saen and Stung Treng gauging stations.
Proportion of dry and wet season
flow ˄%˅
226
D. Li et al. / Journal of Hydrology 551 (2017) 217–232
60
Wet season
50
Dry season
40
u ¼ 360 ði=365Þ
30
20
10
0
1960
of the minimum and maximum streamflow are transformed from
the normal date (i) to the circular date (u) (Gumbel, 1954;
Magilligan and Graber, 1996; Magiligan and Nislow, 2005) to
obtain Fig. 11.
1970
1980
1990
2000
2010
year
Fig. 10. Ratio of streamflow in Chiang Saen to Stung Treng in dry and wet seasons.
seasons (Fig. 9a and c) are more similar than those of the wet seasons (Figs. 9b and 8d).
4.3. Maximum and minimum streamflow
Fig. 11a and Table 4 give the dates of occurrences of the minimum and maximum streamflow. The mean dates of occurrences
ð18Þ
Most of the maximum flow generally occurred from July to
October, and more than 75% occurred in August or in the first half
of September, though two exceptions exist. In Chiang Saen station,
the maximum flow occurred in the second half of December in
2013 in addition to that of 1961. All the minimum flows occurred
from February to May, more than 68% of which were concentrated
in the second half of March or in the first half of April. The earliest
minimum flow occurred in 2003 (February 14, Chiang Saen), 2014
(February 12, Chiang Saen), and 2014 (February 13, Luang Prabang). Hence, we concluded that the dates of occurrences of
extreme flows in the reservoirs did not significantly change. Longer
time series of streamflow observations might be required to further analyze whether giant reservoirs (Xiaowan and Nuozhadu)
will alter the dates of occurrences of extreme flows in the future.
The water stored in the reservoirs reduces the maximum
streamflow. In the upstream Chiang Saen station, the maximum
flow values decreased significantly in the years corresponding to
Fig. 11. The rose diagrams of the dates of occurrences of minimum and maximum streamflow. The red lines represent the mean dates and red arcs represent the 90%
confidence interval. The dates of occurrences of maximum flow during the (A) pre-impact period (1960–1991) and (B) post-impact period (1992–2014). The dates of
occurrences of the minimum flow during the (C) pre-impact period (1960–1991); and (D) transition and post-impact periods (1992–2014).
227
D. Li et al. / Journal of Hydrology 551 (2017) 217–232
Table 4
Degree of changes in IHA (indicators of hydrologic alteration).
Chiang Saen
Jan flow
Feb flow
Mar flow
Apr flow
May flow
Jun flow
Jul flow
Aug flow
Sep flow
Oct flow
Nov flow
Dec flow
1-day max
1-day min
3-day max
3-day min
7-day max
7-day min
30-day max
30-day min
90-day max
90-day min
Base flow index
Date of max
Date of min
High pulse count
Low pulse count
High pulse duration
Low pulse duration
Number of
fluctuations
Rise rate
Fall rate
Mean flow
Stung Treng
Pre-impact period
(1960–1991)
Transition period
(1992–2009)
Post-impact period
(2010–2014
Pre-impact period
(1960–1991)
Transition period
(1992–2009)
Post-impact period
(2010–2014)
Mean value
Mean
value
Relative
change (%)
Mean
value
Relative
change (%)
Mean value
Mean
value
Relative
change (%)
Mean
value
Relative
change (%)
1159.0
936.1
820.0
901.7
1282.0
2511.9
4646.8
6665.9
5528.1
3890.1
2553.1
1611.7
10920.3
732.8
31547.2
2211.2
67641.2
5225.0
223085.5
23588.0
524813.2
78363.6
2.0
231.8
88.3
4.3
3.0
29.1
42.0
47.4
1146.6
935.7
862.6
936.5
1476.0
2436.6
4948.2
6111.4
5496.0
3838.3
2418.5
1520.4
10976.3
637.5
31127.7
1977.3
64870.0
4843.1
214027.5
23294.5
514246.3
80095.8
1.9
244.5
93.8
5.1
6.4
26.3
16.0
69.6
1.1
0.0
5.2
3.9
15.1
3.0
6.5
8.3
0.6
1.3
5.3
5.7
0.5
13.0
1.3
10.6
4.1
7.3
4.1
1.2
2.0
2.2
4.9
5.5
6.2
17.2
114.8
9.4
62.0
46.8
1499.6
1178.6
1272.9
1318.8
1625.2
1887.7
3119.6
4639.8
4169.2
3020.8
2297.5
1983.6
7189.0
863.8
19785.1
2617.6
40564.3
6190.1
146253.3
31567.7
375655.1
111600.3
1.4
253.8
79.8
5.2
2.4
15.3
11.3
73.6
29.4
25.9
55.2
46.3
26.8
24.9
32.9
30.4
24.6
22.3
10.0
23.1
34.2
17.9
37.3
18.4
40.0
18.5
34.4
33.8
28.4
42.4
27.0
9.5
9.6
20.6
20.0
47.3
73.2
55.4
3425.7
2541.7
2020.2
1877.5
3082.9
10561.2
20598.4
35029.2
35993.9
22453.1
10551.2
5447.1
50352.7
1655.9
149177.8
5013.3
337988.0
11843.1
1244186.8
53617.3
2952126.2
184988.8
0.9
240.0
104.8
2.3
1.6
48.1
74.5
44.8
3883.9
2907.2
2377.4
2525.0
4409.2
10646.1
22962.4
36325.3
36229.0
21158.4
10641.5
5856.3
51167.8
1919.5
151355.5
5809.4
341239.7
13891.6
1236426.5
65521.8
3042955.4
228292.1
1.0
243.3
97.0
2.1
2.4
53.3
26.6
46.8
13.4
14.4
17.7
34.5
43.0
0.8
11.5
3.7
0.7
5.8
0.9
7.5
1.6
15.9
1.5
15.9
1.0
17.3
0.6
22.2
3.1
23.4
11.6
1.4
7.4
9.9
52.9
10.7
64.3
4.4
3894.4
3161.2
2963.5
3188.2
4427.7
9168.1
19049.2
31490.6
31733.2
21202.3
10468.8
5937.9
45269.9
2380.3
134309.2
7323.5
303935.4
17501.6
1069891.7
83174.9
2675883.5
278207.5
1.4
249.0
78.2
2.8
0.8
35.3
18.9
55.8
13.7
24.4
46.7
69.8
43.6
13.2
7.5
10.1
11.8
5.6
0.8
9.0
10.1
43.7
10.0
46.1
10.1
47.8
14.0
55.1
9.4
50.4
52.5
3.8
25.3
19.5
48.8
26.7
74.6
24.4
245.4
136.6
2721.6
232.2
179.2
2690.0
5.4
31.3
1.2
199.9
146.3
2343.0
18.6
7.1
13.9
923.0
461.0
12861.5
847.2
513.1
13394.0
8.2
11.3
4.1
625.8
479.3
12282.7
32.2
4.0
4.5
the completion of the Manwan, Dachaoshan, and Jinghong Dams
(Fig. 12a). The same effect was observed in the downstream Stung
Treng station, but to a less significant degree (Fig. 12c).
The results given in Table 4 show that except for 1-day maximum, the maximum streamflow at the Chiang Saen station during
the transition period (1992–2009) decreased slightly compared to
that during the pre-impact period, i.e., 1-day maximum: 0.51%,
3-day maximum: 1.33%, 7-day maximum: 4.10%, 30-day
maximum: 4.06%, and 90-day maximum: 2.01%. During the
post-impact period (2010–2014), the maximum streamflow
significantly decreased, i.e., 1-day maximum: 34.17%, 3-day
maximum: 37.28%, 7-day maximum: 40.03%, 30-day
maximum: 34.44%, and 90-day maximum: 28.42%. Except for
the 90-day minimum, the minimum streamflow during the
transition period (1992–2009) decreased to varying degrees, i.e.,
1-day minimum: 13.04%, 3-day maximum: 10.58%, 7-day
maximum: 7.31%, 30-day maximum: 1.24%, and 90-day
maximum: 2.21%. During the post-impact period (2010–2014),
the minimum streamflow increased significantly, i.e., 1-day
minimum: 17.86%, 3-day maximum: 18.38%, 7-day maximum:
18.47%, 30-day maximum: 33.83%, and 90-day maximum:
42.41%. The value of the 1-day minimum streamflow in the Chiang
Saen station was 1350 m3/s, which is the highest value recorded in
history, occurring on February 12, 2014.
In the downstream Stung Treng station, the changes in the maximum and minimum flow events showed a pattern similar to that
of the upstream Chiang Saen station, though the changes are less
significant (Fig. 12c and d).
4.4. Duration of low pulses
Pulse frequency and duration of high and low water conditions together depict the pulsing behavior of environmental
variations within a year (Richter et al., 1996). The streamflow
data from the pre-impact period (1960–1991) are used to
identify the 75% and 25% thresholds of the pulses. No clear
change was observed for the duration of high pulses, while a
clear change for the duration of low pulses was observed, as
shown in Fig. 13.
Fig. 13 shows the low pulse duration in five gauging stations.
The low pulse duration in the five stations showed significant
decreasing trends during the transition and post-impact periods.
In Chiang Saen station, the mean values of the low-pulse duration
in the pre-impact, transition, and post-impact periods were 42,
16, and 11 days, respectively. When the Manwan Dam started
operations in 1992, an extremely low-pulse duration of 2.25 days
was recorded in the Chiang Saen station. The low-pulse duration
never reached 40 days after 1991. In 2014, the low pulse even
diminished to 0 in all the five stations. The decreasing treads
can be observed in the transition and post-impact periods in all
the five stations. The phenomenon could be associated with the
operation of the giant dams (such as Xiaowan and Nuozhadu),
which store water in wet seasons and release extra water in
dry seasons to support hydropower generation or other objectives
according to their operation rules (Guo and Zhou, 2011; Richter
and Thomas, 2007); hence, reducing the duration of low-flow
pulses.
228
D. Li et al. / Journal of Hydrology 551 (2017) 217–232
Fig. 12. Changes in the annual extreme streamflow in the Chiang Saen and Stung Treng stations from 1960 to 2014.
4.5. Reverse count and rate
The number of fluctuations represents the frequency of reserve
streamflow from rising to falling or falling to rising, and the reverse
rate (rise and fall rates) stands for the slope of the rising or falling
events. Table 4 lists the calculated values of the reverse, which
shows that the number of fluctuations increased slightly with time.
The number of fluctuations at the Chiang Saen station increased by
46% and 55% during the transition and post-impact periods,
respectively. The falling rates in the Chiang Saen station increased
229
D. Li et al. / Journal of Hydrology 551 (2017) 217–232
Fig. 13. Low pulse duration in five gauging stations.
by 31% and 7% during the transition and post-impact periods,
respectively. In addition, the rising rate decreased by 5% and
–19% during the transition and post-impact periods, respectively.
The changes in the flow reverse count and rate were similar to
those observed in the Stung Treng station, though the changes
are of lower magnitudes than those observed in the Chiang Saen
station.
4.6. Partitioning effects of climate change and reservoirs
The Budyko framework is used to distinguish the impact of climate change and human activities at a mean annual time-scale.
The parameter w, which in equation (13) was calculated as 1.76
for the Mekong Basin from 1952 to 1991Table 5 lists the results
of the transition period 1992–2009 and the post-impact period
2010–2015.
As shown in Table 5, compared with the baseline period 1952–
1991, no significant change in the annual streamflow (less than 3%)
occurred from 1992 to 2015 at the basin scale. The percentage of
climate-induced change was 82.29% from 1992 to 2009 wherein
the human-induced change corresponded to 17.71%, and both
parameters increased the streamflow. It is noted that human activities decreased the streamflow from 2010 to 2015, contributing
approximately 62% to the total streamflow change. Hence, it can
be concluded that the impacts of human activities on the streamflow increased at the decade scale.
The application of the Budyko framework requires a steady
state at an annual time-scale. As for the transition period from
1992 to 2009, only 3 reservoirs (Manwan, Dachaoshan, and Jinghong) were built. The total storage capacity of these 3 reservoirs
is 3.08 km3 (Table 2), while the mean annual runoff at Chiang Saen
station is 84 km3/yr. This means that reservoir construction and
operation during this period could not significantly impact the
steady state of the basin at an annual time-scale. During the
post-impact period 2010–2014, the total storage capacity
increased from 3.08 km3 to 43.2 km3, which became significant
compared to the annual runoff but still not enough for interannual
regulation. The operation of these reservoirs always follows steady
rules with an interannual time-scale (Guo and Zhou, 2011; Li et al.,
2015), implying a relevant steady status at an annual time-scale.
Thus, the steady-state assumptions of the Budyko framework is
valid.
In addition to the Budyko method, the simple regression model
is used to partition the effects of climate change and human activities at the seasonal time-scale. As shown in Table 6, the RR values
in the wet season are 59% and 24% at Chiang Saen and Stung Treng
Table 5
Human- and climate-induced change on annual streamflow.
1952–1991
1992–2009
2010–2015
Qm(mm/a)
PET(mm/a)
P(mm/a)
E(mm/a)
DQ(mm/a)
DQc(mm/a)
DQh(mm/a)
Qc%
Qh%
454.734
462.775
443.581
1646.567
1630.257
1974.449
1155.404
1161.445
1272.345
700.671
698.670
828.764
–
8.041
11.153
–
6.617
17.903
–
1.424
29.056
–
82.29
38.12
–
17.71
61.88
Note: percentages of change are calculated in absolute value.
230
D. Li et al. / Journal of Hydrology 551 (2017) 217–232
Table 6
Residuals and the relative variation of streamflow using linear regression models on seasonal scale.
Mean RE
Wet_Chiang Saen
Dry_Chiang Saen
Wet_Stuang Treng
Dry_Stung Treng
RR (%)
R_squre
Pre-impact period
Transition period
Post-Impact period
Transition period
Post-Impact period
0.53
0.22
0.39
0.04
0.02
0.01
0.01
0.05
127.59
15.72
358.38
549.23
1823.31
217.65
4629.39
489.32
3.15
1.21
1.63
11.69
59.26
13.67
23.73
9.91
in the post-impact period 2010–2014, respectively, which indicates that the streamflow in the wet season should be higher than
the observed value if human activities are excluded. Similarly, the
RR values in the dry season are 14% and 10% at Chiang Saen and
Stung Treng, respectively, in the post-impact period 2010–2014,
implying that human activities increased the streamflow in dry
seasons. The RR values in the transition period are very small,
meaning that human activities are not the dominant factor. This
conclusion is consistent with the storage capacity change during
this period.
In addition to reservoir construction and operation, other factors may also affect the flow regimes (Cook et al., 2012), which
need to be considered in this study. The data pertaining to land
use in 2010 and 2000 obtained from CCI-LC project version 2.3,
European Space Agency (ESA) were compared to detect possible
changes in land use in recent years (http://www.esa-landcovercci.org/?q=node/1). It was found that only 0.57% of the area in
the MRB changed during this 10-year period, implying that the
change in land use is not a significant factor affecting the streamflow among human activities. In addition, it has been shown that
snow melting contributed only a small proportion (0.1%) to the
mean annual discharge at Chiang Saen station (Eastham et al.,
2008).
5. Discussion
The flow regimes have changed to some degree after the construction and operation of dams on the mainstream, particularly
in the upstream, observed by several gauging stations during the
transition period (1992–2009) and post-impact period (2010–
2014). The mean flow and minimum flow increased in the dry seasons, while the mean flow and maximum flow decreased in the
wet seasons. The low pulse duration significantly declined, and
the number of fluctuations increased. The changes in the postimpact period are more severe than in the transition period.
The most important negative impact of these changes is the
destruction of fish habitats (Ziv et al., 2012). Similar to many tropical rivers globally, fish ecosystems in the MRB are extremely fragile. Many important and unique fishes in the MRB are seasonally
migratory fishes, and natural flow regimes and upstream shoals
and rapids are very important for their breeding. These changes
in flow regime also influence other riverine ecosystems.
However, increased flow in the dry seasons can significantly
benefit agricultural irrigation in the downstream areas. Water
resources in the upper reach of the Mekong River are mostly used
for power generation without water withdrawal. The five countries
along the LMB are agricultural countries; the Chi-Mun basin has
the largest irrigation system in the MRB. With growing population,
water demand for food and irrigation markedly increased. Barker
and Molle (2004) estimated that only 3% of the irrigated land in
the basin was exploited in 2002. MRC (2011) reported that irrigated areas in the LMB countries would increase from 6.6 million
ha in 2010 to 9.7 million ha in 2030.
Moreover, reduction in wet-season flow, but increase in dryseason flow due to the operation of the upstream dams are beneficial for downstream flood and drought management, navigation,
and even the ecosystem in the Mekong Delta. For instance, on
March 15 and April 10 of 2016, because of the extreme drought
event in the Mekong Delta in Vietnam, the Chinese government
released more than 2.7 billion m3 of stored water from the
upstream Jinghong Reservoir to the Mekong Delta for agriculture
and ecosystem demands. The discharge from the Jinghong Reservoir increased to 2000 m3/s, approximately twice the mean annual
discharge or more than three times the natural discharge for the
same period. This joint effort was a successful event for transboundary river management (MRC, 2016).
Both ecosystem and human demands in the MRB should be considered. Given that a large number of dams in the upper and lower
reaches of the Mekong River are to be constructed in the near
future, the impacts of human activities on streamflow would likely
intensify, posing great challenges for river management. Moreover,
given the uncertainty of climate change, collaborative and multiobjective management of these reservoirs are the keys to address
both ecosystem and human demands; hence, environmental flow
for ecosystem demands should be incorporated into reservoir
operation objectives in the MRB.
All the methods used in this study are statistical in nature (IHA,
eco-metrics, Mann–Kendall test, etc.). These methods can be used
in any river systems, particularly international transboundary river
basins wherein the cooperation among riparian countries and
management of water resources should be supported by illustrating trends and fluctuations of long-term flow regimes.
6. Conclusion
This study used long-term daily streamflow observations in five
gauging stations along the mainstream of the Mekong River to
examine changes and trends in the IHA and eco-flow metrics
(eco-surplus and eco-deficit). The results show that damming in
the upstream area led to a declining trend in the annual streamflow in the upstream Chiang Saen gauging station, whereas no
clear effect on the downstream Stung Treng station was observed.
The operation of the dams reduces streamflow in the wet seasons,
but increases the streamflow in the dry seasons, resulting in
unique seasonal variations in the streamflow based on the ecoflow metrics at Chiang Saen during the period 2010–2014. No
notable change in the date of extreme flows was observed, though
the maximum flow values decreased significantly at Chiang Saen
station in the years corresponding to the completion of the Manwan, Dachaoshan, and Jinghong Dams. The construction and operation of the dams had clear, significant impacts on the low-pulse
durations.
A decomposition method based on the Budyko hypothesis is
used to partition the effects of climate change and human activities
on the streamflow. It is shown that no significant change in the
annual streamflow (less than 3%) occurred from 1992 to 2015 at
the basin scale compared to the baseline period 1952–1991. During the transition period 1992–2009, climate change contributed
to 82.29% of the total streamflow change, whereas human activities contributed to 62% of the total change during the postimpact period 2010–2015. It is shown that only 0.57% of the land
use in the MRB changed during 2000–2010, implying that the
D. Li et al. / Journal of Hydrology 551 (2017) 217–232
land-use change was not a significant factor among human
activities.
Both positive and negative impacts of changes in the flow
regime on the management of water resources and river ecosystems are discussed. On one hand, changes in the flow regime
may influence fish habitats; on the other hand, reservoirs highly
benefit hydropower generation, irrigation, and other human needs.
The results of this study can serve as a basis for detecting changes
in the flow regime, which are valuable in transboundary cooperation and reservoir operation, particularly for better consideration
of environmental flow to address both ecosystem and human
needs in the future.
The statistical analysis performed in this study is based on the
historical streamflow records of five gauging stations on the mainstream of the Mekong River. Different numbers of years in each of
the three sub-periods analyzed may induce biases in the results of
the statistical analysis. Although this analysis yielded clear conclusions on the alteration of streamflow regimes, future observations
are still required to reduce the bias and further explore long-term
impacts of damming over this basin.
Acknowledgments
This research was funded by the National Key Research and
Development Program of China (2016YFC0402203) and the
National Natural Science Foundation of China (Grant No.
51579129 and 91547210). We are grateful to the editors and four
anonymous reviewers for their insightful and constructive comments that significantly helped improve the manuscript.
References
Barker, R., Molle, F., 2004. Evolution of irrigation in South and Southeast Asia.
Colombo, Sri Lanka: Comprehensive Assessment Secretariat. Comprehensive
Assessment Research Report 5.
Beguería, S., López-Moreno, J.I., Lorente, A., Seeger, M., García-Ruiz, J.M., 2003.
Assessing the effect of climate oscillations and land-use changes on streamflow
in the Central Spanish Pyrenees. AMBIO: J. Human Environ. 32 (4), 283–286.
Campbell, I.C., Manusthiparom, C., 2004. Technical report on rainfall and discharge
in the lower Mekong Basin in 2003–2004. Mekong River Commission
Secretariat, Vientiane.
Cochrane, T.A., Arias, M.E., Piman, T., 2014. Historical impact of water infrastructure
on water levels of the Mekong River and the Tonle Sap system. Hydrol. Earth
Syst. Sci. 18 (11), 4529–4541.
Cook, B.I., Bell, A.R., Anchukaitis, K.J., Buckley, B.M., 2012. Snow cover and
precipitation impacts on dry season streamflow in the Lower Mekong Basin. J.
Geophys. Res.: Atmos. 117 (D16).
Delgado, J.M., Apel, H., Merz, B., 2010. Flood trends and variability in the Mekong
river. Hydrol. Earth Syst. Sci. 14 (3), 407–418.
Durrleman, S., Simon, R., 1989. Flexible regression models with cubic splines. Stat.
Med. 8 (5), 551–561.
Eastham, J., Mpelasoka, F., Mainuddin, M., Ticehurst, C., Dyce, P., Hodgson, G., Ali, R.,
Kirby, M., 2008. Mekong River Basin Water Resources Assessment: Impacts of
Climate Change. CSIRO. Water for a Healthy Country National Research Flagship
report.
Fan, H., He, D., Wang, H., 2015. Environmental consequences of damming the
mainstream Lancang-Mekong River: A review. Earth Sci. Rev. 146, 77–91.
Fu, B.P., 1981. On the calculation of the evaporation from land surface [in Chinese].
Sci. Atmos. Sin. 5, 23–31.
Gao, Y., Vogel, R.M., Kroll, C.N., Poff, N.L., Olden, J.D., 2009. Development of
representative indicators of hydrologic alteration. J. Hydrol. 374 (1), 136–147.
Gao, B., Yang, D., Zhao, T., Yang, H., 2012. Changes in the eco-flow metrics of the
Upper Yangtze River from 1961 to 2008. J. Hydrol. 448, 30–38.
Guo, Y., Zhou, Y., 2011. Analyses on Xiaowan reservoir regulation and the operation
optimization of cascade hydropower stations in Lancang river. Hyropower 37
(12), 61–64 (in Chinese).
Gippel, C., 2001. Hydrological analyses for environmental flow assessment. Proc.
MODSIM 2001, 873–880.
Gumbel, E.J., 1954. Applications of the circular normal distribution. J. Am. Stat.
Assoc. 49 (266), 267–297.
Hamed, K.H., Rao, A.R., 1998. A modified Mann-Kendall trend test for autocorrelated
data. J. Hydrol. 204 (1), 182–196.
Hirsch, P., 2010. The changing political dynamics of dam building on the Mekong.
Water Altern. 3 (2), 312.
Ingram, H., Milich, L., Varady, R.G., 1994. Managing transboundary resources:
lessons from Ambos Nogales. Environ.: Sci. Policy Sustainable Dev. 36 (4), 6–38.
231
Kirby, M., Krittasudthacheewa, C., Mainuddin, M., Kemp-Benedict, E., Swartz, C., de
la Rosa, E., 2010. The Mekong: a diverse basin facing the tensions of
development. Water Int. 35 (5), 573–593.
Kuenzer, C., Campbell, I., Roch, M., Leinenkugel, P., Tuan, V.Q., Dech, S., 2013.
Understanding the impact of hydropower developments in the context of
upstream–downstream relations in the Mekong river basin. Sustain. Sci. 8 (4),
565–584.
Kummu, M., Sarkkula, J., 2008. Impact of the Mekong River flow alteration on the
Tonle Sap flood pulse. AMBIO: J. Human Environ. 37 (3), 185–192.
Kummu, M., Lu, X.X., Wang, J.J., Varis, O., 2010. Basin-wide sediment trapping
efficiency of emerging reservoirs along the Mekong. Geomorphology 119 (3),
181–197.
Lauri, H., Moel, H.D., Ward, P.J., Räsänen, T.A., Keskinen, M., Kummu, M., 2012.
Future changes in Mekong River hydrology: impact of climate change and
reservoir operation on discharge. Hydrol. Earth Syst. Sci. 16 (12), 4603–4619.
Li, Jiang, Chen, Cheng, Zhou, Yi, 2015. Study on joint operational strategy for
Xiaowan and Nuozhadu on Lancang river. Yunnan Electric Power 43 (3), 39–41
(In Chinese).
Li, S., He, D., Fu, K., 2006. The correlations of multi-timescale characteristics of water
level processes in Lancang-Mekong River. Chin. Sci. Bull. 51, 50–58.
Liang, W., Bai, D., Wang, F., Fu, B., Yan, J., Wang, S., Yang, Y., Long, D., Feng, M., 2015.
Quantifying the impacts of climate change and ecological restoration on
streamflow changes based on a Budyko hydrological model in China’s Loess
Plateau. Water Resour. Res. 51 (8), 6500–6519.
López-Moreno, J.I., Vicente-Serrano, S.M., Moran-Tejeda, E., Zabalza, J., LorenzoLacruz, J., García-Ruiz, J.M., 2011. Impact of climate evolution and land use
changes on water yield in the Ebro basin. Hydrol. Earth Syst. Sci. 15 (1),
311.
Lu, X.X., Jian-Jun, W., Grundy-Warr, C., 2008. Are the Chinese dams to be blamed for
the lower water levels in the Lower Mekong? Modern Myths of the Mekong,
39–51.
Lu, X.X., Li, S., Kummu, M., Padawangi, R., Wang, J.J., 2014. Observed changes in the
water flow at Chiang Saen in the lower Mekong: Impacts of Chinese dams?
Quatern. Int. 336, 145–157.
Ma, Z.Z., Wang, Z.J., Xia, T., Gippel, C.J., Speed, R., 2014. Hydrograph-based
hydrologic alteration assessment and its application to the yellow river. J.
Environ. Inf. 23 (1).
Magilligan, F.J., Graber, B.E., 1996. Hydroclimatological and geomorphic controls on
the timing and spatial variability of floods in New England, USA. J. Hydrol. 178
(1), 159–180.
Magilligan, F.J., Nislow, K.H., 2005. Changes in hydrologic regime by dams.
Geomorphology 71 (1), 61–78.
Mainuddin, M., Kirby, M., Hoanh, C.T., 2013. Impact of climate change on rainfed
rice and options for adaptation in the lower Mekong Basin. Nat. Hazards 66 (2),
905–938.
Mekong River Commission (MRC), 2005. Overview of the hydrology of the Mekong
Basin, Vietiane, 82, Mekong River Commission.
Mekong River Commission (MRC), 2010. State of the Basin Report 2010. Mekong
River Commission, Vientiane.
Mekong River Commission (MRC), 2011. Assessment of Basin-Wide Development
Scenarios-Main Report. Mekong River Commission, Vientiane.
Mekong River Commission (MRC). (2016). Mekong River Commission, China discuss
joint study. http://www.mrcmekong.org/news-and-events/events/mekongriver-commission-china-discuss-joint-study/.
Moraes, J.M., Pellegrino, G.Q., Ballester, M.V., Martinelli, L.A., Victoria, R.L., Krusche,
A.V., 1998. Trends in hydrological parameters of a southern Brazilian watershed
and its relation to human induced changes. Water Resour. Manage. 12 (4), 295–
311.
Olden, J.D., Poff, N.L., 2003. Redundancy and the choice of hydrologic indices for
characterizing streamflow regimes. River Res. Appl. 19 (2), 101–121.
Piman, T., Cochrane, T.A., Arias, M.E., Green, A., Dat, N.D., 2012. Assessment of flow
changes from hydropower development and operations in sekong, sesan, and
Srepok Rivers of the Mekong Basin. J. Water Resour. Plann. Manage. 139 (6),
723–732.
Piman, T., Lennaerts, T., Southalack, P., 2013. Assessment of hydrological changes in
the lower Mekong Basin from Basin-Wide development scenarios. Hydrol.
Process. 27 (15), 2115–2125.
Poff, N.L., Allan, J.D., Bain, M.B., Karr, J.R., Prestegaard, K.L., Richter, B.D., Stromberg, J.
C., 1997. The natural flow regime. Bioscience 47 (11), 769–784.
Poff, N.L., Zimmerman, J.K., 2010. Ecological responses to altered flow regimes: a
literature review to inform the science and management of environmental
flows. Freshw. Biol. 55 (1), 194–205.
Räsänen, T.A., Koponen, J., Lauri, H., Kummu, M., 2012. Downstream hydrological
impacts of hydropower development in the Upper Mekong Basin. Water Resour.
Manage 26 (12), 3495–3513.
Räsänen, T.A., Someth, P., Lauri, H., Koponen, J., Sarkkula, J., Kummu, M., 2017.
Observed river discharge changes due to hydropower operations in the Upper
Mekong Basin. J. Hydrol. 545, 28–41.
Richter, B.D., Baumgartner, J.V., Powell, J., Braun, D.P., 1996. A method for
assessing hydrologic alteration within ecosystems. Conserv. Biol. 10 (4),
1163–1174.
Richter, B., Baumgartner, J., Wigington, R., Braun, D., 1997. How much water does a
river need? Freshw. Biol. 37 (1), 231–249.
Richter, B.D., Mathews, R., Harrison, D.L., Wigington, R., 2003. Ecologically
sustainable water management: managing river flows for ecological integrity.
Ecol. Appl. 13 (1), 206–224.
232
D. Li et al. / Journal of Hydrology 551 (2017) 217–232
Richter, B.D., Warner, A.T., Meyer, J.L., Lutz, K., 2006. A collaborative and adaptive
process for developing environmental flow recommendations. River Res. Appl.
22 (3), 297–318.
Richter, B.D., Thomas, G.A., 2007. Restoring environmental flows by modifying dam
operations. Ecol. Soc. 12 (1).
Vicente-Serrano, S.M., Zabalza-Martínez, J., Borràs, G., López-Moreno, J.I., Pla, E.,
Pascual, D., Peña-Gallardo, M., 2017. Effect of reservoirs on streamflow and river
regimes in a heavily regulated river basin of Northeast Spain. Catena 149, 727–
741. Ellipsis.
Vogel, R.M., Sieber, J., Archfield, S.A., Smith, M.P., Apse, C.D., Huber-Lee, A., 2007.
Relations among storage, yield, and instream flow. Water Resour. Res. 43 (5).
Vries, T.T., Arias, M.E., Dang, D.T., Cochrane, T.A., Van, P.D.T., 2015. Analysis of water
level changes in the Mekong Floodplain impacted by flood prevention systems
and upstream dams.
Wang, D., Hejazi, M., 2011. Quantifying the relative contribution of the climate and
direct human impacts on mean annual streamflow in the contiguous United
States. Water Resour. Res. 47 (10).
Wang, S.P., Zhang, Z.Q., McVicar, T.R., Guo, J.T., Tang, Y., Yao, A.K., 2013. Isolating the
impacts of climate variability and land use change on decadal streamflow
variation: Assessing three complementary approaches. J. Hydrol. 507, 63–74.
http://dx.doi.org/10.1016/j.jhydrol.2013.10.018.
Xi, L.X., Jian-Jun, W., Grundy-Warr, C., 2008. Are the Chinese dams to be blamed for
the lower water levels in the Lower Mekong? Modern Myths of the Mekong,
39–51.
Yang, H., Yang, D., Lei, Z., Sun, F., 2008. New analytical derivation of the mean
annual water-energy balance equation. Water Resour. Res. 44, W03410. http://
dx.doi.org/10.1029/2007WR006135.
Yin, X.A., Yang, Z.F., Petts, G.E., 2011. Reservoir operating rules to sustain
environmental flows in regulated rivers. Water Resour. Res. 47 (8).
Zhang, L., Dawes, W.R., Walker, G.R., 2001. Response of mean annual
evapotranspiration to vegetation changes at catchment scale. Water Resour.
Res. 37 (3), 701–708. http://dx.doi.org/10.1029/2000WR900325.
Zhang, S., Lu, X.X., 2009. Hydrological responses to precipitation variation and
diverse human activities in a mountainous tributary of the lower Xijiang, China.
Catena 77 (2), 130–142.
Ziv, G., Baran, E., Nam, S., Rodríguez-Iturbe, I., Levin, S.A., 2012. Trading-off fish
biodiversity, food security, and hydropower in the Mekong River Basin. Proc.
Natl. Acad. Sci. 109 (15), 5609–5614.