International Journal of Applied Earth Observation and Geoinformation 16 (2012) 17–26

Contents lists available at SciVerse ScienceDirect

International Journal of Applied Earth Observation and

Geoinformation
journal homepage: www.elsevier.com/locate/jag

Landscape change and hydrologic alteration associated with dam construction

Qinghe Zhao, Shiliang Liu ∗ , Li Deng, Shikui Dong, Cong, Wang, Zhifeng Yang, Juejie Yang
State Key Laboratory of Water Environment Simulation, School of Environment, Beijing Normal University, Beijing 100875, China

a r t i c l e i n f o a b s t r a c t

Article history: Characterizing the landscape changes and hydrologic alterations associated with dam construction is very
Received 27 April 2011 important for watershed management. This paper presents a case study of the Lancang River in Yunnan
Accepted 2 November 2011 Province following dam construction. The landscape patterns and dynamics indicate the fragmentation,
shape, and diversity of the river in 1980, 1990, and 2000. The Range of Variability Approach (RVA) is used
Keywords: to evaluate the degree of hydrologic alteration (DHA) using 44 years (1957–2000) of hydrologic data.
Landscape change
The results indicate that the midstream and downstream landscapes were affected by dam construction,
Hydrologic alteration
becoming more complex and fragmented during the 1980–2000 period; the upstream area was not
Range of Variability Approach
Dam construction
influenced by dam construction and the reservoir impoundment exhibited less change. The variability
Lancang River in maximum runoff occurrence in the post-dam period was less than that in the pre-dam period. The
integrated DHAs of the Jiuzhou (upstream), Gajiu (midstream), and Yunjinghong (downstream) stations
were relatively low, reaching 26.28%, 33.40%, and 37.14%, respectively. However, the alteration became
obvious in the midstream area, and the situation worsened when the river was simultaneously influenced
by dam construction and other human activities (downstream). The results of the regression analysis
show strong relationships of landscape metric changes with DHA, and the forestland and water areas
with DHA. The DHA increased along with the aggravation of landscape fragmentation, the complexity of
the landscape shape, and the diversification of the landscape.
© 2011 Elsevier B.V. All rights reserved.

1. Introduction
Excessive anthropogenic activities such as agricultural land
reclamation, industrial development, road network expansion, and
dam construction play important roles in landscape change and
hydrologic alteration (Xu et al., 2007; Liu et al., 2008; Yang et al.,
2010; Bergerot et al., 2011). Dams, which are constructed for sea-
sonal flood control, navigation, and generation of hydroelectric
power, are often cited as the most significant impact on rivers
around the world, reducing the connectivity of rivers, fragmenting
watersheds, causing changes to hydrological processes, and result-
ing in downstream river channel erosion (Tiemann et al., 2004;
Hu et al., 2008). Therefore, concerns about the effects of dam con-
struction on the environment have increased with the increasing
number of dams in recent years (Jansson et al., 2000; Chovanec
et al., 2002; Tockner and Stanford, 2002; Dudgeon, 2005). Early
studies show that dam construction can affect a variety of processes
in both inner (Ellery et al., 2003; Hu et al., 2008; Walters et al., 2009;
Zeilhofer and De Moura, 2009) and outer river areas (Dudgeon,
2005; Zeilhofer and De Moura, 2009; Ouyang et al., 2010).
∗ Corresponding author. Tel.: +86 10 13522671206; fax: +86 10 58800397.
E-mail address: shiliangliu@bnu.edu.cn (S. Liu).
In the outer river area, landscape changes associated with land
use and land cover change are the most obvious impacts of dam
construction (Ouyang et al., 2010), and have a fundamental recipro-
cal relationship with ecological processes (Turner, 1989). Therefore,
the investigation and quantification of landscape changes caused
by dam construction in the outer area are the domain of landscape
ecology and the basis for sustainable environmental management.
A number of metrics have been developed to measure the influ-
ences of human activities on landscape structure (Theobald, 2010),
such as the total land area and individual land use type areas, patch
density, edge density, perimeter-to-area ratio, landscape diversity,
and so on (Palmer, 2004; Morgan et al., 2010). Usually, researchers
select metrics for quantification of landscape changes based on spe-
cific categories (e.g., fragmentation, shape and diversity) to avoid
linearity and redundancy between metrics.
Dam construction can affect the hydrologic regime in the inner
river area more immediately than the landscape in the outer river
area (Nilsson and Berggren, 2000; Walter and Merritts, 2008). Dam
construction can block the continuity of hydrology, disrupt sed-
iment transport and fish migration by modifying the seasonality
of flows, and alter surface and subsurface water levels, chang-
ing the magnitude, duration, frequency, timing, predictability, and
variability of flow events (Nilsson and Berggren, 2000; Ouyang
et al., 2011). These impacts may lead to a loss of biological diver-
sity and decrease the ecological functions in an aquatic ecosystem

0303-2434/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.jag.2011.11.009
18 Q. Zhao et al. / International Journal of Applied Earth Observation and Geoinformation 16 (2012) 17–26
(Postel, 1998; Isik et al., 2008; El-Shafie et al., 2009). Accordingly,
it is essential to understand the impact of dam construction on
aquatic systems in the inner river area. Recently, the majority of
researchers in China and abroad have explored the impacts of build-
ing dams with a focus on hydrological changes in the inner river
area (Richter et al., 1998; Wang et al., 2005; Lajoie et al., 2007;
Ouyang et al., 2011). These studies demonstrate that the hydrologic
alterations and related environmental impacts of dam construction
have become a main concern in hydrological development plans
(Richter et al., 1998; Lajoie et al., 2007; Ouyang et al., 2011). Regard-
ing hydrologic alteration, many indicators have been developed
from the hydrological monitoring data to quantify flow character-
istics that are sensitive to human perturbations (Yang et al., 2008;
Chen et al., 2010). Although over 170 hydrologic indicators have
been developed to describe different components of flow regimes,
which hydrologic indicators to use to summarize flow properties
analogous to the use of the widely accepted metrics is still unclear
(Gao et al., 2009). Comparatively, the Indicators of Hydrologic Alter-
ation (IHA) comprising 32 parameters developed by Richter et al.
(1996) are more commonly used worldwide (Shiau and Wu, 2006;
Magilligan and Nislow, 2005; Hu et al., 2008; Chen et al., 2010).
Further, Richter et al. (1996) introduced a useful approach referred
to as the Range of Variability Approach (RVA) based on the 32
Indicators of Hydrologic Alteration (IHA) to quantitatively evalu-
ate the degree of hydrologic alteration (DHA) induced by human
disturbance. This method has been shown to be a practical, and
effective way to assess the DHA caused by dam construction and
other human activities (Maingi and Marsh, 2002; Magilligan and
Nislow, 2005; Kummu and Varis, 2007; Yang et al., 2008; Zeilhofer
and De Moura, 2009; Chen et al., 2010).
In general, dam construction has great impacts on landscape
and hydrology, so it is of scientific importance to investigate and
evaluate the landscape changes and hydrologic alterations induced
by dam construction. Nevertheless, many previous researchers
focused on quantifying land use change or hydrological alter-
ations without investigating the interrelation between landscape
changes and hydrological alterations under the influence of dam
construction. However, it is inevitable that land use changes such as
conversion of forest to cropland will directly influence the regional
hydrologic characteristics and processes (Zacharias et al., 2004;
Verbunt et al., 2005; Ouyang et al., 2011). Therefore, the objectives
of this study are to (1) appraise the magnitude of the influence
that dam construction imposes on local landscape and hydrology,
and (2) reveal the relationships between landscape change and
hydrological alteration in the Lancang River in Yunnan Province
(internationally known as the Mekong River) where large dams are
being constructed or have already been completed.
2. Materials and methods
2.1. Study area
As the largest international river in Asia, the Lancang River
passes through seven climatic zones, crosses different geographic
environments and connects different cultural, social, and economic
communities (Liu et al., 2008; Hu et al., 2009). The average rain-
fall in the Lancang River basin is about 500 mm in the southeast
and 250 mm in the northwest parts of the river. The Lancang River
exhibits strong seasonality of runoff, with 70% of the overall annual
water flow normally occurring in August and September (Jacobs,
2002). Additionally, the Lancang River has a main channel drop of
5000 m, and the portion of the river in Yunnan Province has a drop
of 1780 m and an average gradient of 0.15% (Jacobs, 2002; Liu et al.,
2008). Therefore, the unique natural and geographical features in
addition to plentiful hydraulic resources bring many advantages
Fig. 1. Location of the Manwan Dam and gauging stations in the Lancang Basin,
Yunnan Province.
for hydropower cascade development in this area (He et al., 2005,
2006).
The study area is a section of the Lancang River in Yunnan
Province, wherein the lower control point is the Yunjinghong gaug-
ing station and the upper control point is the entrance of the
Lancang River into Yunnan Province (Fig. 1). This area is a key part
of the upper stream of the Lancang River where four hydropower
dams have been constructed (Xiaowan, Manwan, Dachaoshan,
and Jinghong Dams) and several dams and storage reservoirs for
hydropower production are under construction or in planning (Wei
et al., 2009). The study area was divided into three parts accord-
ing to the hydrological and topography features (Gan et al., 2002).
The section from the inlet to the Jiuzhou station was defined as
the upstream area, and was not influenced by dam construction
or reservoir impoundment. The section from the Jiuzhou station
to the Gajiu station was defined as the midstream area, wherein
dam construction and reservoir impoundment were the dominat-
ing disturbances to the landscape and hydrology. The downstream
area includes the section from the Gajiu station to the Yunjinghong
station, and was simultaneously influenced by dam construction
and other human activities.
The Manwan Dam, which is the first dam of the Lancang main-
stream cascade and the first multimillion kilowatt hydropower
station in Yunnan Province, was completed in the middle reach
of the Lancang River in Yunnan Province in 1993. The dam is 132 m
high, has a crest length of 418 m, and forms a 337.1 m wide water
surface reservoir, which is two times more than that of the nat-
ural channel. The reservoir area is 23.6 km2 , which is 2.8 times
larger than the area existing prior to dam construction. The total
reservoir capacity is 1060 × 106 m3 , and the effective capacity is
257 × 106 m3 , which is dependent upon seasonal discharge regu-
lation (He et al., 2004, 2005, 2006; Fu and He, 2007; Zhao et al.,
2010). The associated hydrology and landscape mosaic of the areas
upstream and downstream from the dam have been affected by its
construction and operation (He et al., 2004).
2.2. Data and methods
2.2.1. Data
In this study, regional land use data from three periods (1980,
1990, and 2000) were directly extracted from the national land
use database acquired from the Data Center for Resources and
 Q. Zhao et al. / International Journal of Applied Earth Observation and Geoinformation 16 (2012) 17–26
Table 1
Records for gauging stations used in this analysis, and total number of years of records available for pre- and post-dam periods excluding the wet years (P ≥ Pmean+0.75stdv ,
1006 mm) and the dry years (P ≤ Pmean−0.75stdv , 883 mm).
Stations Locations
◦  ◦ 
Jiuzhou 25 17 N 99 13 E
Gajiu 24◦ 35 N 100◦ 27 E
Yunjinghong 22◦ 12 N 100◦ 28 E
Environmental Sciences, Chinese Academy of Sciences (RESDC).
These data were interpreted using Landsat MSS and TM series of
images and saved as a shape file format. The data were validated
using the ground truth data, and the results indicated that the accu-
racy of each land use map was at least 90% (Liu et al., 2006, 2008).
We then converted these data to raster format and imported them
into the FRAGSTATS software program to calculate the landscape
metrics. We analyzed the monthly average runoff data recorded
from 1957 to 2000 at the Jiuzhou, Gajiu, and Yunjinghong hydro-
logical monitoring stations (Fig. 1, Table 1). The Jiuzhou station is
located 269 km upstream from the Manwan Dam site, which not
affected by the backwater of the Manwan Reservoir. The Gajiu sta-
tion is located about 2 km downstream from the Manwan Dam site.
The Yunjinghong station is very close to the China–Burma border,
so the data collected there is always considered representative of
the Lancang River trans-boundary runoff. In this study, we divided
the data series from these three stations into pre- and post-dam
periods based on the time of dam completion. The precipitation
data (1957–2000) used to remove the potential impacts of pre-
cipitation variability on hydrological processes came from eight
surrounding National Reference Meteorological Stations.
2.2.2. Landscape metric selection and calculation
In order to quantify the landscape change, we computed land-
scape metrics indicating landscape fragmentation, shape, and
diversity information at the landscape level using FRAGSTATS soft-
ware (McGarigal et al., 2002). Patch density (PD), mean patch size
(MPS) and largest patch index (LPS) were used to quantify the frag-
mentation information. Landscape shape index (LSI), edge density
(ED), perimeter-area fractal dimension (PAFRAC), and the mean
fractal dimension (MFD) index were used to measure the shape and
size information. Contagion (CONTAG), Shannon’s diversity index
(SHDI) and Shannon’s evenness index (SHEI) were used to quantify
the compositional and structural components of the diversity infor-
mation (McGarigal et al., 2002; Ouyang et al., 2009; Huang et al.,
2010). The mathematical expressions and ecological meanings of
these metrics representing different landscape aspects are listed in
Appendix C of the FRAGSTATS software manual (McGarigal et al.,
2002). According to previous studies, the input pixel size of the
land use map principally influences the analytical metric results
(Liu et al., 2005; Uuemaa et al., 2005, 2008; Ouyang et al., 2009).
In this study, we selected 200 m as the input pixel size based on
previous studies (Uuemaa et al., 2008; Ouyang et al., 2009; Morgan
and Gergel, 2010).
2.2.3. Removal of potential precipitation variability impacts on
hydrological processes
The water years of the hydrological time-series data must be
separated in order to retain the same impact sources such as the
effect of dam construction on runoff (Yang et al., 2008; Chen et al.,
2010). Therefore, we identified the wet and dry years according
to the method developed by Yoo (2006), who recommended that
the proper threshold to confirm the independent occurrence of
dry or wet years based on the annual basin precipitation is the
level of mean ± 0.75stdv. A period with annual basin precipita-
tion less than mean − 0.75stdv (P ≤ Pmean−0.75stdv ) is considered a
19
Drainage area (km2 ) Series length (year)
Pre-dam (1957–1992) Post-dam (1993–2000)
34392.6 15 6
20058.4 15 6
10801.9 15 6
dry year, while a period with annual basin precipitation more
than mean + 0.75stdv (P ≥ Pmean+0.75stdv ) is considered a wet year.
The periods in which annual basin precipitation falls within the
threshold of mean ± 0.75stdv (Pmean−0.75stdv ≤ P ≤ Pmean+0.75stdv ) are
considered the normal years. In this study, only the runoff records
in normal years can be applied in the RVA hydrological alteration
assessment (Yang et al., 2008; Chen et al., 2010).
2.2.4. Range of Variability Approach and degree of hydrologic
alteration
IHA assesses alterations in terms of flow magnitude, timing,
frequency, duration and rate of change with 32 hydrological vari-
ables. However, these 32 variables are only defined when the daily
flow logs are available (Lajoie et al., 2007). Therefore, we defined
five parameters – magnitude, frequency, interannual variability of
magnitude, periods of occurrence of the monthly minimum and
maximum flows, and interannual variability of the periods of occur-
rence – based on monthly average flows, following the research of
Richter et al. (1996) and Lajoie et al. (2007). Next, we assessed the
hydrologic alterations using the five parameters with RVA. The RVA
target range for each parameter was determined from the pre-dam
flow presumed to be the ideal condition. We selected a simple mul-
tiple of the parameter standard deviations (mean ± stdv) for the
pre-dam flow as the RVA target range suggested as by Richter et al.
(1996). The degree to which the RVA target range is not attained
was regarded as a measure of hydrologic alteration (Hu et al., 2008;
Yang et al., 2008; Zeilhofer and De Moura, 2009). This measure of
hydrologic alteration, expressed as a percentage, can be calculated
as:
 Observed frequency − Expected frequency 
DHA = × 100% (1)
Expected frequency
where DHA is the degree of hydrologic alteration, “observed” is
the number of years in which the observed value of the hydro-
logic parameter falls within the target range, and “expected” is the
number of years for which the value is expected to fall within the
target range. DHA is equal to zero when the observed frequency of
the post-impact annual values within the RVA target range equals
the expected frequency. A positive deviation indicates that annual
parameter values fell inside the RVA target range more often than
expected, while negative values indicate that annual values fall
within the RVA target range less often than expected (Richter et al.,
1996; Hu et al., 2008; Yang et al., 2008; Zeilhofer and De Moura,
2009). DHA indicator values of hydrologic alteration will be used
to quantify the deviation of the post-dam flow regime from the
pre-dam flow. However, a single integrated index is needed to rep-
resent the overall hydrologic alteration (Shiau and Wu, 2006). Thus,
we used the average of the absolute DHA values of all parame-
ters (defined above) to provide an assessment of the overall impact
of dam construction (Richter et al., 1998). Moreover, to evaluate
whether a specific characteristic was severely altered, a simple
three-class system was used to evaluate the severity of hydrologic
alteration (Shiau and Wu, 2006). DHA values between 0 and 33
percent were classified as low alteration, values between 33 and
67 percent as moderate alteration, and 67 to 100 percent as high
alteration (Richter et al., 1998).
20 Q. Zhao et al. / International Journal of Applied Earth Observation and Geoinformation 16 (2012) 17–26
Fig. 2. Landscape fragmentation metrics during the 1980–2000 period.
2.2.5. Statistical data analysis
In this study, a simple linear regression method, which requires
a long time period and stable series (Phillips, 1986; Fu and He,
2007), was used to explore the relationship between runoff and
precipitation. This made it possible to determine the influence
of average monthly precipitation on average monthly runoff. The
simple linear regression method was also used to identify the
relationship between landscape change and hydrologic alteration
(Ouyang et al., 2009). Therefore, we analyzed the correlation
between runoff and precipitation within the study area using
monthly series data from 1957 to 2000, as well as the correlations
between rates of landscape change and DHA within the upstream,
midstream, downstream, and overall study areas. The data were
analyzed using SPSS 15 (Akin et al., 2010).
3. Results
3.1. Landscape metric change in upstream, midstream, and
downstream areas of the Lancang River
3.1.1. Landscape fragmentation metric change
The change in the landscape fragmentation metrics at the land-
scape level is shown in Fig. 2. In the overall study area, PD first
decreased from 1980 to 1990 and then increased from 1990 to 2000
with values of 0.222, 0.167, and 0.222, respectively. The MPS and
LPI show the opposite changes with values of 451.5 ha and 54.4% in
1980, and 600.2 ha and 63.8% in 1990, 451.2 ha and 53.7% in 2000,
Fig. 3. Landscape shape metrics during the1980–2000 period.
respectively. In the upstream, midstream, and downstream areas,
and landscape fragmentation metrics show the same changes as for
the overall study area. However, differences exist in the magnitude
of PD: the downstream area had the largest PD (largest fragmen-
tation degree) in all study periods, and the midstream area had
the lowest fragmentation degree. The upstream area was not influ-
enced by dam construction and reservoir impoundment, it showed
less change during the 1980–2000 period. Based on the change
in PD, the patch number increased and the landscape became
more fragmented in 2000 as compared with 1990, especially in
the midstream and downstream areas. The largest MPS indicates
the lowest fragmentation degree. The downstream area, which was
simultaneously influenced by dam construction and other human
activities, had the lowest MPS value in both the pre- and post-
dam periods with up–down variation over the study period. Fig. 2
shows that the original downstream MPS was 362.2 ha in 1980
increasing to 502.9 ha in 1990 and then decreasing to 357.2 ha
in 2000 after dam construction. The other three areas had simi-
lar change tendencies as the downstream area, revealing that the
patch size decreased and the patches became more fragmented
after dam construction with the exception of the upstream area,
where the MPS changed little during the 1980–2000 period. LPI
had the same change trend with MPS, and the downstream LPI
had the largest change, increasing from 50.7% to 67.1% during the
1980–1990 period, while decreasing to 49.7% in 2000 (Fig. 2). In
general, the midstream and downstream areas as well as the overall
study area showed larger variations from 1980 to 2000, and became
more fragmented after dam construction; however, the upstream
area showed little change during the study period.
3.1.2. Landscape shape metric change
Fig. 3 shows the changes in the landscape shape metrics during
the 1980–2000 period. The landscape shape metrics showed fewer
changes than the landscape fragmentation metrics, which were
relatively similar in that they decreased slightly in 1990. The down-
stream area had the largest ED, indicating that it had the longest
edge segments per unit area; on the other hand, the upstream area
had the lowest ED during the 1980–2000 period. All areas had
the same changing trend of LSI, decreasing before, and increasing
after, dam construction. Moreover, the LSI values increased from
the upstream to downstream areas and reached a maximum in the
overall study area. MFD changed little between 1980 and 2000,
meaning that the change in shape complexity for the area was
unobvious between the upstream and downstream areas before
and after dam construction. PAFRAC, the indicator of the com-
plexity of patch shapes, shows that the downstream area had the
most complex shape during the 1980–2000 period. The midstream
area, in which dam construction and reservoir impoundment were
the dominating disturbances to the landscape, had the most basic
 Q. Zhao et al. / International Journal of Applied Earth Observation and Geoinformation 16 (2012) 17–26 21

Fig. 5. Water year separation of the hydrological time-series data for the Lancang
River in Yunnan Province (the hydrological time-series corresponding to normal
years (shaded area) were used in the RVA analysis, the hydrological time-series
Fig. 4. Landscape diversity metrics during the1980–2000 period. corresponding to wet years and dry years were excluded).

3.2.1. Impacts of dam construction on magnitude and

shape, revealing the largest change in PAFRAC among all areas in interannual variability of average monthly runoff
the study. The ED, LSI, MFD, and PAFRAC indicate that the regional We compared the average monthly runoff and coefficient of
landscape shape became more complex and fragmented and less variation (CV) from the upstream (Jiuzhou), midstream (Gajiu),
aggregated in 2000, and this is particularly true for the downstream and downstream (Yunjinghong) areas of in the Manwan Dam on
area. the Lancang River for 12 months in the pre- and post-dam peri-
ods, respectively (Fig. 6). The upstream area was also divided into
the pre- and post-dam periods to analyze the effects of dam con-
3.1.3. Landscape diversity metric change struction on hydrologic alteration for comparison. The average
In the overall study area, the CONTAG value increased from monthly runoff at the Jiuzhou station (Fig. 6a) indicates that there
57.4 to 60.6 during the 1980–1990 period and decreased from was no significant change in runoff after dam construction during
60.6 to 57.2 during the 1990–2000 period (Fig. 4), indicating that January, February, March, April, October, or December. There was
the connectivity among the dominant patches increased in the a significant increase in runoff in May, July, August, September,
first decade whereas the patches were distributed more unevenly and November, but a significant decrease in runoffs in June. The
in the later decade. SHDI reflects the patch abundance and het- monthly runoff CV shows that the change in runoff increased after
erogeneity in the landscape and SHEI describes the distribution dam construction. The maximum runoff occurred in August dur-
evenness of landscape elements with CONTAG (Fig. 4). During the ing the pre-dam period and in July during the post-dam period.
1980–1990 period, SHDI and SHEI declined from 0.994 to 0.991
and from 0.553 to 0.527, respectively, suggesting that the richness
and evenness of the landscape types decreased and the composi-
tion of the ecosystem gradually simplified. During the 1990–2000
period, SHDI and SHEI increased from 0.991 to 0.997 and from
0.527 to 0.556, respectively. The midstream and downstream areas
showed similar trends with the overall study area, but the down-
stream area exhibited greater change during the 1980–2000 period.
The upstream area was almost unchanged during the 1980–2000
period. These results indicate that the landscape diversity increased
and the landscape elements were distributed more evenly in the
influenced area of dam construction. It also shows that human dis-
turbances such as dam construction intensified in the study area.

3.2. Impact of dam construction on monthly runoff characteristics

We analyzed the relationship between runoff and precipitation

in addition to the water years of the hydrological time-series before
analyzing the impact of dam construction on monthly runoff char-
acteristics. The results revealed a close relationship between runoff
and precipitation as shown by a correlation coefficient of 0.867,
which demonstrated that variation in precipitation may influence
the variation in natural runoff in the study period and area. The
location, name, and drainage area of stations, the time-series data
of the pre- and post-dam periods, and the series lengths of normal Fig. 6. Magnitude and interannual variability of average monthly runoff in the
years are shown in Fig. 5 and Table 1. upstream (a), midstream (b), and downstream (c) areas.
22 Q. Zhao et al. / International Journal of Applied Earth Observation and Geoinformation 16 (2012) 17–26

Fig. 8. Coefficients of variation of the monthly maximum (a) and minimum (b)
runoff in the pre-and post-dam periods.

from the Manwan Dam were affected by reservoir regulation

Fig. 7. Frequencies of monthly maximum and minimum runoff in the upstream (a),
midstream (b), and downstream (c) areas.
The largest CV was observed in August during both the pre- and
post-dam periods, amounting to 39.4% and 37.4%, respectively. The
change in average monthly runoff at the Gajiu station was similar
to that at the Jiuzhou and Yunjinghong stations (Fig. 6b). The mag-
nitude of the runoff during each month after dam construction was
very close to that before dam construction. However, the largest
CV was observed in October during the pre-dam period with a per-
centage of 32.1%, while in the post-dam period it was observed
in April with a percentage of 31.2%. The average monthly runoff
at the Yunjinghong station shows that the largest CV occurred in
November, reaching 34.1% after dam construction (Fig. 6c). The
largest CV was observed in June with a percentage of 33.8% prior to
dam construction.
activities.
3.2.3. The interannual variability of the monthly maximum and
minimum runoff
The interannual variability of the monthly maximum and min-
imum runoff expressed by the CV are presented in Fig. 8. The
monthly maximum runoff of the areas upstream, midstream, and
downstream from the Manwan Dam were characterized by high
variability (in terms of month of occurrence) in the pre-dam period
(Fig. 8a), reflected by a lower CV in the post-dam period than in the
pre-dam period. The trend for monthly minimum runoff was oppo-
site that of monthly maximum runoff (Fig. 8b). Indeed, the CV in the
post-dam period was larger than that in the pre-dam period in the
areas upstream, midstream, and downstream from the Manwan
Dam. These observations suggest that the variability in the months
in which maximum runoff occurred declined in the upstream and
downstream areas in the post-dam period as compared with the
pre-dam period. The opposite result was found for the months in
which minimum runoff occurred.

3.2.4. Evaluation of hydrologic alteration with monthly runoff

3.2.2. Impacts of dam construction on the maximum and
characteristics
minimum monthly runoff
The influence of dam construction on monthly runoff was esti-
Frequencies of monthly maximum and minimum runoff in the
mated using the monthly runoff data from 1957 to 2000 (Fig. 9). The
upstream (Jiuzhou), midstream (Gajiu), and downstream (Yun-
jinghong) areas (based on relation to the Manwan Dam) are shown
in Fig. 7. The monthly maximum runoff at the Jiuzhou station
(Fig. 7a) was frequently observed in July during the pre-dam period,
but occurred more frequently in July and August during the post-
dam period. The monthly minimum runoff frequently occurred in
February in both the pre- and post-dam periods. For the Gajiu sta-
tion (Fig. 7b), the monthly maximum runoff frequently occurred in
July in the post-dam period and in August in the pre-dam period.
The monthly minimum runoff frequently occurred in February and
March in the pre-dam period, but in January and February dur-
ing the post-dam period. The monthly maximum runoff at the
Yunjinghong station (Fig. 7c) frequently occurred in August, and
was scattered in July, September, and October in the pre-dam
period, while in the post-dam period it frequently occurred in
July and was scattered in August and September. The monthly
minimum runoff frequently occurred in February and March in
both the pre- and post-dam periods. The results indicate that the Fig. 9. Degrees of monthly average runoff alteration for three stream gauges on the
monthly maximum and minimum runoff frequencies downstream Lancang River in Yunnan Province.
 Q. Zhao et al. / International Journal of Applied Earth Observation and Geoinformation 16 (2012) 17–26 23

Table 3
Correlation between degree of hydrologic alteration and landscape type area.

Independent Dependent Correlation model R2

Forestland y = 0.622x − 25.731 0.930*

Grassland y = −2.4822x + 96.962 0.497
Degree of hydrologic Farmland y = −0.7323x + 36.937 0.208
alteration Water area y = 0.7047x − 24.482 0.975**
Construction land y = 0.8045x − 20.936 0.146
Unused land y = −2.7184x + 67.654 0.813*
*
Correlation is significant at the 0.05 level.
**
Correlation is significant at the 0.01 level.

Fig. 10. Degrees of monthly maximum and minimum runoff alteration for the three
stream gauges on the Lancang River in Yunnan Province. and the situation worsened when the river was simultane-
ously influenced by dam construction and other human activities
(downstream).

spatio-temporal hydrologic alterations differed among the three

stations. The hydrologic alterations in average monthly runoff in
February (downstream), April (up and downstream), and October
(upstream) were more obvious than those in other months. The
smallest changes in mean monthly flow among the 12 months in
the areas upstream, midstream, and downstream from the dam
site were 0% (January and February), 9.1% (December), and 9.1%
(May), respectively. However, the greatest effects of impoundment
in the midstream area were concentrated in three months: April,
June, and October (degrees of alteration were all above 50%). At
the Yunjinghong station, the greatest effect of dam construction
was concentrated in two months: February and April (with degree
of alteration reaching 77.3%). The hydrologic alteration of extreme
values based on monthly runoff shows that the degrees of monthly
maximum and minimum runoff alteration were relatively low at
the three gauging stations, suggesting insubstantial environmental
stress and disturbance of the Manwan Dam on the Lancang River
(Fig. 10).
Based on the above comparisons and analyses, different indica-
tors had diverse DHA, so we employed the average of the absolute
values of all parameters’ DHA to provide an assessment for the
overall impact of dam construction. The integrated DHA of the
Jiuzhou, Gajiu, and Yunjinghong stations reached 26.28%, 33.40%,
and 37.14%, respectively, indicating that the integrated DHA at the
three stations could be classified as low, moderate, and moderate
alteration, respectively. The construction of Manwan Dam altered
the natural runoff and spatial variability of the Lancang River.
The spatio-temporal hydrologic alterations differed among the
three stations. The lowest and highest integrated DHA values were
observed at the Jiuzhou and Yunjinghong stations, respectively.
This indicates that the degree of hydrologic alteration is relatively
low when there is no dam on the river (upstream); however,
more alteration took place after dam construction (midstream),
3.3. Relationship between landscape metric change and degree of
hydrologic alteration
The simple linear regression models and coefficients of corre-
lation between the landscape metric change and DHA are shown
in Table 2. The results indicate that the fragmentation of landscape
reflected by LPI had a significant positive relationship with DHA,
implying that an increase in LPI would augment DHA. However, the
relationships of PD and MPS with DHA were positive but insignif-
icant. Three out of the four metrics (with the exception of MFD)
had significant positive relationships with DHA, implying that more
complex patch shapes were associated with larger DHA. The results
regarding changes in DHA in response to landscape diversity metric
changes demonstrate that the landscape diversity metric change is
strongly related to DHA. CONTAG had a significant positive relation-
ship with DHA, and SHDI and SHEI had extremely strong positive
relationships with DHA (p < 0.01), with correlation coefficients of
0.997. A more diverse landscape is associated with larger DHA val-
ues.
The relationship between landscape area and DHA is presented
in Table 3. The simple linear regression models and coefficients
of correlation show that the forestland, water, and unused land
areas had significant relationships with DHA. The water areas had
the strongest (P < 0.01) positive relationship with DHA. The prob-
able reason for this pattern is that forestland and water were the
dominant landscapes in the inner and outer river areas and their
changes significantly influence each other. In conclusion, there
were strong relationships between landscape metric changes and
DHA: the more fragmented (LPI) and/or complex (ED, LSI, and
PAFRAC) and/or diverse (CONTAG, SHDI, and SHEI) the landscape
became, the larger the DHA value.

Table 2
Correlation between degree of hydrologic alteration and landscape metric change at the landscape level.

Independent Dependent Correlation model R2

Landscape fragmentation metrics PD y = −2.9023x + 121.79 0.759

MPS y = 1.9169x − 82.939 0.711
LPI y = 2.0296x − 78.858 0.898*
Landscape shape metrics ED y = −1.9804x + 80.027 0.949*
LSI y = −1.9209x + 77.392 0.947*
Degree of hydrologic alteration
MFD y = −0.0177x + 0.4322 0.658
PAFRAC y = −0.373x + 13.554 0.843*
Landscape diversity metrics CONTAG y = 0.681x − 26.651 0.958*
SHDI y = −1.1197x + 41.371 0.997**
SHEI y = −1.1185x + 41.334 0.997**
*
Correlation is significant at the 0.05 level.
**
Correlation is significant at the 0.01 level.
24 Q. Zhao et al. / International Journal of Applied Earth Observation and Geoinformation 16 (2012) 17–26
4. Discussion and conclusions
4.1. Landscape metrics for assessing the impact of dam
construction on landscape patterns
In this study, we examined landscape pattern metrics using
regional land use data interpreted from Landsat MSS and TM
images. Similar to previous research, the landscape metrics selected
to describe variability in the landscape structure were based on cat-
egorical classifications (Ouyang et al., 2009; Morgan et al., 2010;
Theobald, 2010), implying that there are problems associated with
the classification scheme such as number of classes, arbitrary
classification schemes, classification accuracy, and spatial scale
(Wu, 2004; Corry, 2005; Diaz-Varela et al., 2009; Theobald, 2010).
Although some approaches were tested to eliminate the effects of
factors such as grain, extent, and classification scheme on land-
scape metrics, they are rarely found to have good applicability (Wu,
2004; Corry, 2005; Morgan and Gergel, 2010). Uuemaa et al. (2005,
2008) demonstrated that responses of landscape metrics to chang-
ing grain size vary among landscapes and metrics, such as patch
density and edge density decreasing sharply within the 100 m grain
size and then fluctuating slightly; the contagion decreasing signif-
icantly up to a grain size of 400 m then fluctuating with further
increases in grain size for all landscapes; Shannon’s diversity index
and evenness index decreasing up to one patch at the 400 m grain
size; mean shape index decreasing significantly up to a grain size
of 100 m and from 200 m showing erratic responses to increasing
grain size. Therefore, in this study, we selected 200 m as the input
pixel size based on the spatial resolution of Landsat MSS (80 m) and
TM (30 m) images and methods followed in other studies (Uuemaa
et al., 2008; Morgan and Gergel, 2010).
The landscape metrics we selected were related to landscape
composition or configuration based on previous research, which
were grouped into three major categories: fragmentation, shape,
and diversity. LSI is not really a measure of pattern shape in prac-
tice, but is more related to the degree of pattern fragmentation
(Saura and Castro, 2007). In this case, we analyzed LSI in the land-
scape shape metric changes that are similar to the findings of
other researchers (Gan et al., 2002; He et al., 2004; Ouyang et al.,
2009). The landscape fragmentation metrics demonstrate that the
midstream and downstream areas became more fragmented after
dam construction, and the landscape shape metrics show that the
regional landscape shape in the midstream and downstream areas
became more complex and fragmented as well as less aggregated
after dam construction. A higher level of landscape diversity in
the midstream area indicates that the landscape elements became
more diverse and more evenly distributed after dam construc-
tion. The spatio-temporal landscape pattern changes in the study
area are the basis for understanding the impacts of dam construc-
tion on landscape. However, Wu (2004) and Corry (2005) indicate
that the extent of the changes significantly affects landscape met-
rics, though the effects are generally less predictable than those
of changing pixel size, and the classification scheme can impact
the diversity and evenness of the regional landscape. These results
suggest that a more fragmented landscape pattern after dam con-
struction at the current study extent may exhibit connectedness at
fine-scales, and higher levels of landscape diversity and evenness
may exhibit a lower diverse or even level as the unique individ-
ual patches become one larger patch through dissolution when
shrinking the number of land cover types. Consequently, further
study on landscape responses to dam-induced influence is neces-
sary to use landscape metric scalograms in place of single-scale
values (Wu, 2004), explicitly recognize and address the selected
landscape matrices applied to fine-scale spatial data (Corry, 2005),
and present the optimized management approaches to mitigate the
negative ecological consequences of dam construction.
4.2. Range of Variability Approach for assessing the impact of
dam construction on hydrologic characteristic
The Range of Variability Approach (RVA) has been used world-
wide and is a practical, effective approach for assessing the DHA
caused by dam construction (Maingi and Marsh, 2002; Magilligan
and Nislow, 2005; Kummu and Varis, 2007; Yang et al., 2008;
Zeilhofer and De Moura, 2009; Chen et al., 2010). We analyzed DHA
based on monthly hydrologic data following the methods recom-
mended by Richter et al. (1998) and Lajoie et al. (2007) because
there was a lack of daily monitoring hydrologic data. The results
demonstrate that a combination of the two methods made it pos-
sible to assess the DHA caused by dam construction using the
monthly flow characteristics. The results indicate that operation
of the Manwan Dam brought forward the monthly maximum and
minimum runoff times in the downstream area. This is quite dif-
ferent than the findings of previous researchers (Yang et al., 2008;
Chen et al., 2010) studying in the middle and lower Yellow River
of China. Construction of the Manwan Dam resulted in a decrease
in the interannual variability of maximum runoff and an increase
in the interannual variability of minimum runoff. In spite of no sig-
nificant change in the magnitude of runoff, a significant increase of
CV in April and distinct decreases of CV in June, July, and October
indicate that the monthly runoff fluctuated abnormally in these
months in the post-dam period due to the dam operation. Gen-
erally, the results reveal that the impacts of the Manwan Dam on
hydrologic alteration are relatively complicated when combined
with magnitude and interannual variability of monthly runoff.
The degrees of monthly maximum and minimum runoff alter-
ation were relatively low in the three stream gauges, and the
integrated DHA of the Jiuzhou, Gajiu, and Yunjinghong stations
reached 26.28%, 33.40%, and 37.14%, respectively, suggesting no
substantial environmental stress or intense ecological disturbance
caused by the Manwan Dam on the monthly runoff of the Lan-
cang River. However, the alteration became remarkable after the
dam was built (midstream), and the situation worsened when the
river was simultaneously influenced by dam construction and other
human activities (downstream). Human activities could lead to
other disturbances (e.g., agricultural land reclamation, deforesta-
tion, urban sprawl, road networks expansion, and so on), as the
Yunjinghong station is located far downstream from the Manwan
Dam, which is close to the China–Burma border, and hydrologic
alteration occurring in the downstream area could be influenced by
other human activities. For example, the road construction stud-
ies conducted in this region indicate that road networks have a
significant effect on ecosystem changes, and changes in landscape
patterns in specific (Liu et al., 2006, 2008). The changed landscape
pattern impacts the regional hydrologic characteristics and pro-
cesses (Zacharias et al., 2004; Verbunt et al., 2005; Ouyang et al.,
2011), and the relationship between landscape change and degree
of hydrologic alteration was also demonstrated in this study. The
main objective of the present paper was to appraise the damming
effects on landscape and hydrology, so the individual impact of
other human activities on landscape and hydrology is not included.
The impacts of climate variability on initial hydrological time-
series have been demonstrated in different regions (Yang et al.,
2008; Chen et al., 2010), and the significant impact of climate vari-
ability can be regarded as precipitation (Labat et al., 2004; Menzel
and Burger, 2002; Wang and Zhou, 2005). Precipitation not only
affects the magnitude of annual, seasonal, and monthly runoff,
but is also likely to affect the frequency of extreme flow events.
Inevitably, the variation of precipitation between the pre- and post-
dam periods will substantially influence the value of DHA. So, in
order to clarify the influence of dam construction on hydrologic
characteristics, we tried to remove the possible impacts of pre-
cipitation on runoff using the method developed by Yoo (2006).
 Q. Zhao et al. / International Journal of Applied Earth Observation and Geoinformation 16 (2012) 17–26
However, it is quite difficult to distinguish all potential effects of
natural and artificial processes on hydrological regimes (Yang et al.,
2008). So, more uncertainties should be identified and assessed.
Furthermore, the RVA method used in this paper is sensitive
to the length of the hydrologic series (Richter et al., 1996; Yang
et al., 2008), so the validity of the DHA values should be taken with
caution when the available hydrologic series are short for either
the pre- or post-dam period (Yang et al., 2008). In this study, the
lengths of the observed hydrologic series were 36 and 8 years before
the removal of precipitation variability impacts for the pre- and
post-dam periods, respectively, reducing to 15 and 6 years after
removing the precipitation impacts, respectively. The short length
of the hydrologic series during the post-dam period could influ-
ence the DHA results. Therefore, further study of feasible methods
to extend the length of the hydrologic series and investigate the
response of DHA to series length is necessary. Nevertheless, the
results from the DHA analyses presented important insights on the
hydrologic alterations induced by dam construction that can be
used to provide guidance for river development and management
in the future.
4.3. Correlation between rate of landscape change and degree of
hydrologic alteration
Regression analysis was used to identify the factors influ-
encing hydrologic changes in the river areas downstream from
dams (Matteau et al., 2009) and the accumulated effects on land-
scape patterns from hydroelectric cascade exploitation (Ouyang
et al., 2009). In our analyses, we attempted to correlate land-
scape change and hydrologic alteration using a regression analysis
method. We found significant relationships between DHA and land-
scape metric changes, especially for the forestland and water areas,
including the outer and inner river area. This further supports
other researchers’ findings on the indirect relationship between
the watershed area and hydrologic alteration (Lajoie et al., 2007;
Matteau et al., 2009). The LPI of the landscape fragmentation met-
rics and the ED, LSI, and PAFRAC of the landscape shape metrics in
addition to the three metrics of landscape diversity had close rela-
tionships with DHA, indicating that a change in landscape patterns
would influence hydrologic characteristics within the watershed.
The DHA increased simultaneously with the aggravation of land-
scape fragmentation, the complication of landscape shape, and
the diversification of landscape. However, the regression results
indicate that the statistical relevance between variables and the
corresponding cause–effect relationship was uncertain. Although
many researchers have modeled the impact of land use and land
cover changes on hydrology (Guo et al., 2008; Verstraeten and
Prosser, 2008; Oeurng et al., 2011; Xie and Cui, 2011), the influence
of land use change induced by dam construction on hydrology has
rarely been explored. Therefore, further investigations and compre-
hensive simulation of the impact of dam construction on landscape
change and hydrologic alteration as well as the internal relationship
between landscape and hydrology are needed.
Acknowledgements
This research was funded by the National Natural Sciences
Foundation of China (Nos. 50939001, 40871237), the Nonprofit
Environment Protection Specific Project of China (No. 201209029),
and the Fundamental Research Funds for the Central Universities.
References
Akin, B.S., Atıcı, T., Katircioglu, H., Keskin, F., 2010. Investigation of water quality on
Gökçekaya dam lake using multivariate statistical analysis, in Eskişehir, Turkey.
Environ. Earth Sci., doi:10.1007/s12665-010-0798-6.
25
Bergerot, B., Fontaine, B., Julliard, R., Baguette, M., 2011. Landscape variables impact
the structure and composition of butterfly assemblages along an urbanization
gradient. Landscape Ecol. 26, 83–94.
Chen, Y.Q.D., Yang, T., Xu, C.Y., Zhang, Q., Chen, X., Hao, Z.C., 2010. Hydrologic alter-
ation along the Middle and Upper East River (Dongjiang) basin, South China:
a visually enhanced mining on the results of RVA method. Stoch. Environ. Res.
Risk A 24, 9–18.
Chovanec, A., Schiemer, F., Waidbacher, H., Spolwind, R., 2002. Rehabilitation of
a heavily modified river section of the Danube in Vienna (Austria): biological
assessment of landscape linkages on different scales. Int. Rev. Hydrobiol. 87,
183–195.
Corry, R.C., 2005. Characterizing fine-scale patterns of alternative agricultural land-
scapes with landscape pattern indices. Landscape Ecol. 20, 591–608.
Diaz-Varela, E.R., Marey-Pérez, M.F., Rigueiro-Rodriguez, A., Álvarez-Álvarez, P.,
2009. Landscape metrics for characterization of forest landscapes in a sustain-
able management framework: Potential application and prevention of misuse.
Ann. For. Sci. 66, 301.
Dudgeon, D., 2005. River rehabilitation for conservation of fish biodiver-
sity in monsoonal Asia. Ecol. Soc. 10, 15. http://www.ecologyandsociety.
org/vol10/iss2/art15/.
El-Shafie, A., Abdin, A.E., Noureldin, A., Taha, M.R., 2009. Enhancing inflow forecast-
ing model at Aswan high dam utilizing radial basis neural network and upstream
monitoring stations measurements. Water Resour. Manage. 23, 2289–2315.
Ellery, W.N., Dahlberg, A.C., Strydom, R., Neal, M.J., Jackson, J., 2003. Diversion of
water flow from a floodplain wetland stream: an analysis of geomorphological
setting and hydrological and ecological consequences. J. Environ. Manage. 68,
51–71.
Fu, K.D., He, D.M., 2007. Analysis and prediction of sediment trapping efficiencies
of the reservoirs in the mainstream of the Lancang River. Chin. Sci. Bull. 52,
134–140.
Gan, S., He, D.M., Dang, C.L., 2002. The comparison of three case study area on land-
scape structure of Lancang River basin in Yunnan Province. J. Mount Res. 20,
564–569.
Gao, Y., Vogel, R., Kroll, C., Poff, N., Olden, J., 2009. Development of representative
indicators of hydrologic alteration. J. Hydrol. 374, 136–147.
Guo, H., Hu, Q., Jiang, T., 2008. Annual and seasonal streamflow responses to climate
and land-cover changes in the Poyang Lake basin, China. J. Hydrol. 355, 106–122.
He, D.M., Feng, Y., Gan, S., Magee, D., You, W.H., 2006. Transboundary hydrological
effects of hydropower dam construction on the Lancang River. Chin. Sci. Bull. 51,
16–24.
He, D.M., Wu, S.H., Peng, H., Yang, Z.F., Ou, X.K., Cui, B.S., 2005. A study of ecosystem
changes in longitudinal Range-Gorge region and transboundary eco-security in
southwest China. Adv. Earth Sci. 20, 338–344.
He, D.M., Zhao, W.J., Chen, L.H., 2004. The ecological changes in Manwan reservoir
area and its causes. J. Yunnan Univ. (Nat.Sci.) 26, 220–226.
Hu, B., Cui, B.S., Dong, S.K., Zhai, H.J., Liu, H.J., 2009. Ecological water requirement
(EWR) analysis of high mountain and steep gorge (HMSG) river-application to
Upper Lancang-Mekong River. Water Resour. Manage. 23, 341–366.
Hu, W.W., Wang, G.X., Deng, W., Li, S.N., 2008. The influence of dams on ecohydro-
logical conditions in the Huaihe River basin, China. Ecol. Eng. 33, 233–241.
Huang, J.L., Lin, J., Tu, Z.S., 2010. Detecting spatiotemporal change of land use and
landscape pattern in a coastal gulf region, southeast of China. Environ. Dev.
Sustain. 12, 35–48.
Isik, S., Dogan, E., Kalin, L., Sasal, M., Agiralioglu, N., 2008. Effects of anthropogenic
activities on the Lower Sakarya River. Catena 75, 172–181.
Jacobs, J.W., 2002. The Mekong River Commission: transboundary water resources
planning and regional security. Geogr. J. 168, 354–364.
Jansson, R., Nilsson, C., Renofalt, B., 2000. Fragmentation of riparian floras in rivers
with multiple dams. Ecology 81, 899–903.
Kummu, M., Varis, O., 2007. Sediment-related impacts due to upstream reservoir
trapping, the Lower Mekong River. Geomorphology 85, 275–293.
Labat, D., Godderis, Y., Probst, J.L., Guyot, J.L., 2004. Evidence for global runoff increase
related to climate warming. Adv. Water Resour. 27, 631–642.
Lajoie, F., Assani, A.A., Roy, A.G., Mesfioui, M., 2007. Impacts of dams on monthly
flow characteristics. The influence of watershed size and seasons. J. Hydrol. 334,
423–439.
Liu, J.Y., Liu, M.L., Tian, H.Q., Zhuang, D.F., Zhang, Z.X., Zhang, W., Tang, X.M.,
Deng, X.Z., 2005. Spatial and temporal patterns of China’s cropland during
1990–2000: an analysis based on Landsat TM data. Remote Sens. Environ. 98,
442–456.
Liu, S.L., Cui, B.S., Dong, S.K., Yang, Z.F., Yang, M., Holt, K., 2008. Evaluating the influ-
ence of road networks on landscape and regional ecological risk – a case study
in Lancang River Valley of Southwest China. Ecol. Eng. 34, 91–99.
Liu, S.L., Wen, M.X., Cui, B.S., Dong, S.K., 2006. Effects of road networks on regional
ecosystems in southwest mountain area: a case study in Jinghong of Longitudi-
nal Range-Gorge Region. Acta Ecol. Sin. 26, 3018–3024.
Magilligan, F.J., Nislow, K.H., 2005. Changes in hydrologic regime by dams. Geomor-
phology 71, 61–78.
Maingi, J.K., Marsh, S.E., 2002. Quantifying hydrologic impacts following dam con-
struction along the Tana River, Kenya. J. Arid Environ. 50, 53–79.
Matteau, M., Assani, A.A., Mesfioui, M., 2009. Application of multivariate statistical
analysis methods to the dam hydrologic impact studies. J. Hydrol. 371, 120–128.
Mcgarigal, K., Cushman, S.A., Neel, M.C., Ene, E., 2002. Fragstats: Spatial Pattern
Analysis Software for Categorical Maps. University of Massachusetts, Amherst.
Menzel, L., Burger, G., 2002. Climate change scenarios and runoff response in the
Mulde catchment (Southern Elbe, Germany). J. Hydrol. 267, 53–64.
26 Q. Zhao et al. / International Journal of Applied Earth Observation and Geoinformation 16 (2012) 17–26
Morgan, J.L., Gergel, S.E., 2010. Quantifying historic landscape heterogeneity
from aerial photographs using object-based analysis. Landscape Ecol. 25,
985–998.
Morgan, J.L., Gergel, S.E., Coops, N.C., 2010. Aerial photography: a rapidly evolving
tool for ecological management. BioScience 60, 47–59.
Nilsson, C., Berggren, K., 2000. Alterations of riparian ecosystems caused by river
regulation. BioScience 50, 783–792.
Oeurng, C., Sauvage, S., Sánchez-Pérez, J.M., 2011. Assessment of hydrology, sed-
iment and particulate organic carbon yield in a large agricultural catchment
using the SWAT model. J. Hydrol. 401, 145–153.
Ouyang, W., Hao, F.H., Song, K.Y., Zhang, X., 2011. Cascade dam-induced hydrological
disturbance and environmental impact in the upper stream of the Yellow River.
Water Resour. Manage. 25, 913–927.
Ouyang, W., Hao, F.H., Zhao, C., Lin, C., 2010. Vegetation response to 30 years
hydropower cascade exploitation in upper stream of Yellow River. Commun.
Nonlinear Sci. Numer. Sim. 15, 1928–1941.
Ouyang, W., Skidmore, A.K., Hao, F.H., Toxopeus, A.G., d Abkar, A., 2009. Accumulated
effects on landscape pattern by hydroelectric cascade exploitation in the Yellow
River basin from 1977 to 2006. Landscape Urban Plann., 163–171.
Palmer, J.F., 2004. Using spatial metrics to predict scenic perception in a changing
landscape: Dennis, Massachusetts. Landscape Urban Plann. 69, 201–218.
Phillips, P.C.B., 1986. Understanding spurious regressions in econometrics. J. Econ.
33, 311–340.
Postel, S.L., 1998. Water for food production: will there be enough in 2025? Bio-
Science 48, 629–637.
Richter, B.D., Baumgartner, J.V., Braun, D.P., Powell, J., 1998. A spatial assessment
of hydrologic alteration within a river network. Regul. Rivers: Res. Manage. 14,
329–340.
Richter, B.D., Baumgartner, J.V., Powell, J., Braun, D.P., 1996. A method for assessing
hydrologic alteration within ecosystems. Conserv. Biol. 10, 1163–1174.
Saura, S., Castro, S., 2007. Scaling functions for landscape pattern metrics derived
from remotely sensed data: are their subpixel estimates really accurate? ISPRS
Photogramm. Eng. Remote Sens. 62, 201–216.
Shiau, J.T., Wu, F.C., 2006. Compromise programming methodology for determining
instream flow under multiobjective water allocation criteria. J. Am. Water Works
Assoc. 42, 1179–1191.
Theobald, D.M., 2010. Estimating natural landscape changes from 1992 to 2030 in
the conterminous US. Landscape Ecol. 25, 999–1011.
Tiemann, J.S., Gillette, D.P., Wildhaber, M.L., Edds, D.R., 2004. Effects of lowhead dams
on riffle-dwelling fishes and macroinvertebrates in a midwestern river. Trans.
Am. Fish. Soc. 133, 705–717.
Tockner, K., Stanford, J.A., 2002. Riverine flood plains: present state and future trends.
Environ. Conserv. 29, 308–330.
Turner, M.G., 1989. Landscape ecology: the effect of pattern on process. Annu. Rev.
Ecol. Syst. 20, 171–197.
Uuemaa, E., Roosaare, J., Kanal, A., Mander, Ü., 2008. Spatial correlograms of soil
cover as an indicator of landscape heterogeneity. Ecol. Indic. 8, 783–794.
Uuemaa, E., Roosaare, J., Mander, Ü., 2005. Scale dependence of landscape metrics
and their indicatory value for nutrient and organic matter losses from catch-
ments. Ecol. Indic. 5, 350–369.
Verbunt, M., Zwaaftink, M., Gurtz, J., 2005. The hydrologic impact of land cover
changes and hydropower stations in the Alpine Rhine basin. Ecol. Model 187,
71–84.
Verstraeten, G., Prosser, I.P., 2008. Modelling the impact of land-use change and farm
dam construction on hillslope sediment delivery to rivers at the regional scale.
Geomorphology 98, 199–212.
Walter, R.C., Merritts, D.J., 2008. Natural streams and the legacy of water-powered
mills. Science 319, 299.
Walters, A.W., Barnes, R.T., Post, D.M., 2009. Anadromous alewives (Alosa pseu-
doharengus) contribute marine-derived nutrients to coastal stream food webs.
Can. J. Fish. Aquat. Sci. 66, 439–448.
Wang, G.Q., Wu, B.S., Wang, Z.Y., 2005. Sedimentation problems and management
strategies of Sanmenxia Reservoir, Yellow River, China. Water Resour. Res. 1,
W09417, doi:10.1029/2004WR0039190.
Wang, Y.Q., Zhou, L., 2005. Observed trends in extreme precipitation events in China
during 1961–2001 and the associated changes in large-scale circulation. Geo-
phys. Res. Lett. 2, L09707, doi:10.1029/2005GL0225740.
Wei, G.L., Yang, Z.F., Cui, B.S., Li, B., Chen, H., Bai, J.H., Dong, S.K., 2009. Impact of
dam construction on water quality and water self-purification capacity of the
Lancang River, China. Water Resour. Manage. 23, 1763–1780.
Wu, J., 2004. Effects of changing scale on landscape pattern analysis: scaling rela-
tions. Landscape Ecol. 19, 125–138.
Xie, X.H., Cui, Y.L., 2011. Development and test of SWAT for modeling hydrological
processes in irrigation districts with paddy rice. J. Hydrol. 396, 61–71.
Xu, C., Liu, M.S., Zhang, C., An, S.Q., Yu, W., Chen, J., 2007. The spatiotemporal dynam-
ics of rapid urban growth in the Nanjing metropolitan region of China. Landscape
Ecol. 22, 925–937.
Yang, J.J., Liu, S.L., Dong, S.K., Zhao, Q.H., Zhang, Z.L., 2010. Spatial analysis of three
vegetation types in Xishuangbanna on a road network using the network K-
function. Proc. Environ. Sci. 2, 1534–1539.
Yang, T., Zhang, Q., Chen, Y.Q.D., Tao, X., Xu, C.Y., Chen, X., 2008. A spatial assessment
of hydrologic alteration caused by dam construction in the middle and lower
Yellow River, China. Hydrol. Process. 22, 3829–3843.
Yoo, C., 2006. Long term analysis of wet and dry years in Seoul, Korea. J. Hydrol. 318,
24–36.
Zacharias, I., Dimitriou, E., Koussouris, T., 2004. Quantifying land-use alterations and
associated hydrologic impacts at a wetland area by using remote sensing and
modeling techniques. Environ. Model. Assess. 9, 23–32.
Zeilhofer, P., De Moura, R.M., 2009. Hydrological changes in the northern Pantanal
caused by the Manso dam: impact analysis and suggestions for mitigation. Ecol.
Eng. 35, 105–117.
Zhao, Q.H., Liu, S.L., Dong, S.K., 2010. Effect of dam construction on spatial-temporal
change of land use: a case study of Manwan, Lancang River, Yunnan, China. Proc.
Environ. Sci. 2, 852–858.