Ocean & Coastal Management 135 (2017) 56e64

Contents lists available at ScienceDirect

Ocean & Coastal Management

journal homepage: www.elsevier.com/locate/ocecoaman

Using Landsat images to quantify different human threats to the

Shuangtai Estuary Ramsar site, China
Yanlin Tian a, 1, Ling Luo a, 1, Dehua Mao a, *, Zongming Wang a, **, Lin Li b, Jianping Liang a
a
Key Laboratory of Wetland Ecology and Environment, Northeast Institute of Geography and Agroecology, Chinese Academy of Sciences, No. 4888, Shengbei
Street, Changchun 130102, China
b
Department of Earth Sciences, Indiana University-Purdue University at Indianapolis, 420 University Blvd, Indianapolis, IN, USA

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

Article history: Multiple human activities can impose significant negative effects on wetland ecosystem. This study
Received 8 August 2016 aimed to identify and quantify the human threats on the Shuangtai Estuary Ramsar site (SERS) by means
Received in revised form of detecting wetland landscape changes using Landsat images from 1988 to 2014. Land-cover changes
6 November 2016
induced by agricultural reclamation, urban expansion, petroleum industry, aquaculture, and the in-
Accepted 7 November 2016
Available online 11 November 2016
creases in transportation land were identified using an object-oriented classification method and a
decision tree. Human-triggered wetland conversions and landscape metrics-characterized wetland
changes were documented in order to assess the protection effectiveness after this wetland was listed as
Keywords:
Remote sensing
a national nature reserve (1988) and a Ramsar site with international importance (2005). Results indicate
Human threats that wetlands area at this Ramsar site decreased from 127,526 ha in 1988e117,805 ha in 2005, 113,705 ha
Wetland loss in 2014 at an annual averaged loss rate of 531.6 ± 72 ha/yr. The only two vegetated wetland types,
Landscape metrics Phragmites australis and Suaeda heteroptera, have lost 4788 ha and 12,856 ha, respectively. Aquaculture
The Shuangtai Estuary Ramsar site ponds increased by 9715 ha, and almost all were transformed from natural wetlands. The landscape
metrics used in this study show notable fragmentation trend of wetland as a result of the expansion of
transportation lands. Multiple human activities occupied large areas of various wetland types. Compared
to agricultural reclamation and urbanization, wetland changes at the study site were affected primarily
by the development of aquaculture and petroleum industry. Obvious wetland loss and fragmentation
determined by the analysis of Landsat images suggests that the protection effect has been relatively low
in spite of the fact that the SERS is a national natural reserve and Ramsar site. As an internationally
critical rest habitat for migration waterfowl, ecological degradation of the SERS and severe human
threats create great challenges for ecosystem managements. This study also indicates that the protection
effectiveness of other costal wetland should also be assessed using remote sensing to objectively track
the conservation or restoration of these areas.
© 2016 Elsevier Ltd. All rights reserved.

1. Introduction examined worldwide, indicating these changes in wetlands affect

Wetland ecosystems are important due to their multiple
ecosystem services (Cao and Fox, 2009; Jafari, 2009; Keddy, 2010).
Climate change and human activities can drive changes in wetlands
(Junk et al., 2013; Wang et al., 2012; Wright and Wimberly, 2013).
Tremendous loss and fragmentation of wetlands have been
* Corresponding author.
** Corresponding author.
E-mail addresses: maodehua@iga.ac.cn (D. Mao), zongmingwang@iga.ac.cn
(Z. Wang).
1
Yanlin and Ling contributed equally to this work.
ecosystem services such as habitat supporting, water regulation,
biodiversity conservation etc. (Amler et al., 2015; Coleman et al.,
2008; Gong et al., 2010; Rebelo et al., 2009). Humans have great
difficulties to respond to the climate-resulted wetland changes, but
could adjust their behavior to reduce the direct effects on wetland
ecosystems. The quantification of direct threats from human ac-
tivities on wetlands is thus vital for scientific management and
conservation of wetlands.
Previous studies have documented shrinkage and fragmenta-
tion of wetlands on account of multiple human activities (An et al.,
2007; Murray et al., 2014). Agricultural reclamation has been
identified as the leading factor for vanished wetlands and their

http://dx.doi.org/10.1016/j.ocecoaman.2016.11.011
0964-5691/© 2016 Elsevier Ltd. All rights reserved.
 Y. Tian et al. / Ocean & Coastal Management 135 (2017) 56e64
weakened ecosystem services (Rebelo et al., 2009; Finlayson and
Rea, 1999; Song et al., 2014; Wang et al., 2011a). Urban expansion
was determined to have significant effects on wetlands in different
regions of the world, especially in deltas or coastal regions (Bolca
et al., 2007; Lee et al., 2006; Kentula et al., 2004). The rapid
development of industrialized landscape has been reported to
impose evident impacts on wetland (Ren et al., 2010). Typical ex-
amples include oil exploitation causing the fragmentation of wet-
lands in the Yellow River delta of China (Zhang and Sun, 2005) and
the loss of ecosystem services in North America (Allred et al., 2015),
and construction of roads, ports, and dams resulting in noticeable
disappearance and fragmentation of wetlands (Bi et al., 2011;
Giosan et al., 2014; Li et al., 2014). In addition, Seto and Fragkias
(2007) revealed that aquaculture in Vietnam occupied substantial
mangrove wetlands for economic benefits. Although assessing
human threats on wetlands is beneficial for addressing wetland loss
and degradation as well as achieving effective protection and
management, the lack of information on wetland shrinkage and
their linkages to different human activities for a long period have
been a major barrier to such efforts.
Satellite remote sensing has the potential to provide infor-
mation for assessing human threats to wetland ecosystems
(Gallant, 2015; Seto and Fragkias, 2007), and thus has been
widely used in evaluating habitat conditions and monitoring
conservation effects (Barker and King, 2012; Nagendra et al.,
2013; Zheng et al., 2012). Satellite images were used to docu-
ment the status and distribution of mangrove forest over the
world (Giri et al., 2011), and to evaluate the protection efficacy of
national wetland reserves in China from 1978 to 2008 (Zheng
et al., 2012). Accessing a large area landscape is challenging
using the tradition field investigation method. Remote sensing
can provide both the latest images to characterize the current
status of wetlands at different scales and historical datasets to
determine the changes of wetland surface features. These ad-
vantages of remote sensing makes it the best way to identify
human-triggered wetland conversion. Furthermore, various
threats from human activities can be quantified to contribute to
the wetland management and conservation.
Wetland ecosystems have been largely protected in natural
reserves. Up to now, 2218 sites covering 214,131,110 ha wetlands in
the world were listed as the wetland with international impor-
tance (Ramsar site, https://ramsar.org), and 46 sites of these sites
covering 4,002,240 ha wetlands are in China. As one of the Ramsar
sites, Shuangtai Estuary wetland has the largest coastal Phragmites
australis (reed) wetland in Asia and the most famous Suaeda
heteroptera (saline seepweed) wetland named as “red beach”,
located in the south of the Liao River delta and the north of the
Bohai Sea Economic Rim. Although the present wetland was
approved to be national nature reserve in 1988 and has been a
Ramsar site since 2005, over the past decades, land expansions for
agriculture, urbanization, petroleum industry, aquaculture, and
tourism, as well as road and port construction have imposed
significant impacts on wetlands in Shuangtai Estuary Ramsar site
(SERS) as a result of local economic development (Mao et al.,
2014), and numerous striking features emerged at this site, for
example, enhanced oil industry and expansion of aquaculture
ponds. Therefore, it is important to assess potential threats from
these human activities to wetland in SERS and develop proactive
conservation strategies in order to maintain ecosystem structure
and services, and ensure the regional ecological security. This
objective motivated us to use remote sensing to investigate the
wetland types of this area from 1988 to 2005 and 2014, and to
quantify the spatiotemporal pattern of wetlands in response to
anthropogenic occupation.
57
2. Material and methods
2.1. Study area
The Shuangtai Estuary Ramsar site (No.1441) covering 1352 km2,
with latitude ranging from 40
450 to 41
100 N and longitude from
120
300 to 122
000 E, is located in the northern Liaodong gulf of
Northeast China (Fig. 1). A temperate semi-humid monsoon climate
prevails in the study area with the mean annual temperature being
8.5
C and the mean annual precipitation being 650 mm. This
wetland has more than 126 plant species with dominated plants
being Phragmites australis, Suaeda heteroptera, Typha, and
Hemarthria etc., and these plants grow from April to October. About
700 wild animals and bird species have been found in this wetland
(www.ramar.org). SERS is thus the important rest habitat for
migration waterfowl in the East Asia-Australasian Flyway (EAAF), in
which 45 bird species being protected.
Evident human disturbances to the wetlands of this area
occurred because of oil field prospection. Liaohe Oil Field was once
the third largest oil field in China, in which oil wells, factories,
roads, residences were extensively constructed. Meanwhile, the
enhanced aquaculture development and paddy cultivation were
promoted for their considerable economic benefits. In addition, the
red beach (Suaeda heteroptera marshland) attracts numerous
tourists to this region. All these human activities imply significant
disturbances to the SERS, and are responsible for the shrinkage and
fragmentation of natural wetlands in the SERS.
2.2. Remote sensing dataset
We select the 1988, 2005 and 2014 as the study nodes because
the study area was approved to be a national natural reserve since
1988 and a Ramsar site since 2005. Cloud-free Landsat image
scenes (path 120, row 31/32, Fig. 1) were acquired for the three
years from the American Earth Resources Observation and Science
Center (EROS). Both TM (Thematic Mapper) and OLI (Operational
Land Imager) images acquired from August were analyzed to obtain
land covers of the study area considering that land cover types can
be easily identified during this month (Lin et al., 2015; Ozesmi and
Bauer, 2002).
Before land cover classification can be done, the acquired images
were processed for atmospheric correction using the 6S radioactive
transfer model and geo-rectified with reference to 1:50,000 topo-
graphic maps, and at least 30 ground control points (GCPs) were
used for geo-rectifying each image. The root mean squared error of
geometric rectification was less than 1 pixel (or 30 m). The ENVI 5.0
image processing software was used in this geo-rectification step.
2.3. Extracting land covers information
For examining landscape dynamics, land cover of SERS in 1988,
2005, and 2014 was classified into 4 land cover categories (natural
wetland, human-made wetland, cropland, built-up land) and 12
land cover classes (Table 1) using the eCognition software.
Although paddy field is one of the important human-made wetland
types, it wasn't considered in this analysis since it is more appro-
priate classify them into cropland. The object-oriented classifica-
tion method, proven to be more accurate and robust than the
traditional pixel-based method (Dronova, 2015), was used for the
land cover classification consisting of three steps (Wang et al.,
2015). Segmentation is the first step of the object-oriented
method, which is the process of segmenting an image into groups
of homogeneous pixels so that the variability within the object was
minimized (Baatz and Scha €pe, 2000). Images segmentation pa-
rameters include scale, shape and compactness. The multi-
58 Y. Tian et al. / Ocean & Coastal Management 135 (2017) 56e64

Fig. 1. The geographical location and satellite images of the Shuangtai Estuary Ramsar site.

resolution segmentation algorithm was performed on Landsat TM matched the landscape characteristics of the study area. The shape
OLI images at five segmentation scales: 10, 30, 50, 80, and 100. parameter was set to 0.1 to maximize the influence of spectral in-
Through comparison of segmented results at different scales, the formation. The compactness was set to 0.8 to control the final shape
scale parameter 50 was found to be the most acceptable and of an image object. Second, the image objects were classified into

Table 1
The classified types of land cover in the Shuangtai Estuary Ramsar site (the images were composed by standard false color of Landsat bands).

Land cover category Land cover class Image feature Land cover category Land cover class Image feature

Natural wetland Phragmites australis marshland Cropland Dry farmland

Suaeda heteroptera marshland Paddy field

River

Shallow sea Built-up land Residential land

Intertidal zone Industrial land

Human-made wetland Aquaculture pond Transportation land

Channel
 Y. Tian et al. / Ocean & Coastal Management 135 (2017) 56e64
certain categories using a number of decision rules. The classifica-
tion rules are related to different characteristics of segmented ob-
jects. For example, large normalized difference water index (NDWI)
contribute to extracting wetland, while normalized difference
vegetation index (NDVI) was applied to distinguish water body and
marshland. Color, texture, brightness, and adjacency largely assis-
ted in the identification of various wetland types, such as Phrag-
mites australis and Suaeda heteroptera. Normalized difference built-
up index (NDBI), brightness index (BI), and bare index (NIR þ R)
were used to identify built-up lands. Size, shape, and the length/
width index also were used to separate residential, industrial and
transportation lands. Furthermore, multi-season Landsat TM/OLI
data, ground filed check data, high resolution images from Google
earth, and other ancillary data (vegetation map, digital elevation
model, etc.) were used to improve the classification accuracy. Third,
the preliminary classification results were revised according to vi-
sual interpretation and ground survey data (Fig. 1). Our knowledge
to the region also contributed significantly to this process. The
classification accuracy for land covers was assessed through ground
surveys (reserved 200 verification points) and confusion matrix
method (Jia et al., 2014), as well as photos, and interviews with local
people. The overall classification accuracy and Kappa coefficient for
1988, 2005 and 2014 land cover maps was 91.8% and 0.87, 94.5% and
0.89, and 94.6% and 0.89, respectively.
2.4. Evaluating the direct threats from human activities on natural
wetland
As shown in Fig. 2, Landsat images document clear landscape
change of wetlands caused by multiple human activities from 1988
to 2005, and to 2014. Based on land cover classification result, the
spatial pattern and shrinkage of various wetlands were examined
for 1988, 2005, and 2014. For evaluating the direct threats from
multiple human activities and their linkages to landscape changes,
we mapped the wetland losses attributable farming reclamation,
urbanization, increasing oil industry footprint, development of
aquaculture, and enhanced transportation network. For investi-
gating the impacts of farming reclamation, the analysis was focused
on the expansion of dry farmland and paddy field, respectively. For
determining the influence of urbanization, the conversion of wet-
lands to residential lands was assessed. The impacts on wetlands
imposed by petroleum industry were addressed by examining oil
wells and factories. The areal increase of ponds for breeding crabs
or fishes were also quantified to demonstrate the effects of aqua-
culture. And also increased transportation lands and that caused
natural wetland changes were documented to discuss the impacts
of denser transportation network including roads and ports for
transporting crude oil or tourisms. Additionally, landscape metrics
in 1988, 2005, and 2014 were calculated by use of the software
Fragstats 4.2 for characterizing landscape fragmentation, including
the number of patches (NP), patch density (PD), Shannon's diversity
index (SHDI), contagion index (CONTAG), aggregation index (AI)
(Song et al., 2012). Furthermore, landscape dynamics were exam-
ined depending on the changes of landscape metrics. In this study,
the study periods of 1988~2005 and 2005~2014 was defined as
stage 1 and stage 2, respectively.
3. Results
3.1. Landscape features in the SERS
The 2014 classification results demonstrate that the area of
wetland, cropland, built-up land was 113,705 ha, 13,527 ha,
8005 ha, accounting for 84%, 10%, 6% of the total study region,
respectively. Marsh was the dominated wetland type including
59
43,951 ha of Phragmites australis and 3166 ha of Suaeda heteroptera,
which respectively accounted for 32.5% and 2.3% of the study area
and were mostly distributed in west side of the Shuangtaizi River.
Aquaculture pond was the primary type of human-made wetlands,
with an area of 11,732 ha, equivalent to 8.7% of the SERS. The river,
intertidal zone, and shallow sea presented the south of SERS, with
an area of 7381 ha, 1588 ha, 45,233 ha, respectively. Paddy field
held 87.3% of cropland with an area 11,810 ha and was largely
distributed in the southeast part of the SERS. In addition, trans-
portation land (4685 ha) had a larger area than industrial (950 ha)
and residential lands (2370 ha). The residential land was concen-
trated in five towns, while extensive traffic network and oil wells
for petroleum industry, tourism, and resident population observ-
ably fragmented wetland landscape.
3.2. Wetlands shrinkage in the SERS
Notable changes in the SERS during 1988~2014 were illustrated
in Fig. 3. Wetlands had decreased from 127,526 ha in
1988e117,805 ha in 2005, and to 113,705 ha in 2014, at an annual
averaged loss rate of 571.8 ± 73 ha/yr in stage 1 and 455.6 ± 51 ha/yr
in stage 2, and a great wetland loss occurred in the southern and
eastern parts. Although an obvious increase in human-made wet-
lands has been observed, the apparent loss of natural wetlands
drove the decreasing trend of the total wetlands in the SERS. For the
only two vegetated wetland types, Phragmites australis marshland
decreased by 4102 ha in stage 1 and 685 ha in stage 2, and Suaeda
heteroptera marshland lost by 11,372 ha in stage 1 and 685 in stage
2. The bigger disappearances of marsh patches mainly occurred in
the southern coastal region and the eastern side of Shuangtaizi
River. During the past 26 years, human-made wetland increased by
6764 ha in stage 1 and 3107 ha in stage 2, of which the expanded
aquaculture pond accounted for 98% of the increases. Meanwhile,
the intertidal zone had largely yielded to aquaculture pond in the
southwestern part and to paddy field in the southeastern part of the
SERS. Clear expansion of built-up lands also led to notable wetland
shrinkage. Not only did the human activities occupy massive nat-
ural wetlands, such as the expansion of agricultural land and
aquaculture pond, but also natural factors, such as rising sea levels,
led to remarkable wetland shrinkage (Fig. 3).
3.3. Occupation of multiple human activities on wetlands
3.3.1. Agriculture activities
Large area of wetlands were converted into cropland in the
study region as clearly illustrated in Fig. 4. Dry farmland increased
by 1121 ha in stage 1 and 104 ha in stage 2, and paddy field by
7056 ha in stage 1 and 1916 ha in stage 2. During the two stages,
676 ha (32.2%) of dry farmland and 9477 ha (80.2%) of paddy field
were reclaimed from wetland, respectively. The expansion of
cropland was mainly observed in the north of study region which is
largely covered by Phragmites australis (568 ha) and was mostly
observed in stage 1. In the eastern side of Shuangtaizi River and
close to ocean where there were 5656 ha Phragmites australis and
1311 ha Suaeda heteroptera marshes in 1988, as well as 2307 ha
intertidal zone, a large amount of natural wetlands were converted
to paddy fields.
3.3.2. Expansion of urban and rural residential lands
Accompanied by significantly increased population, an apparent
expansion of urban and rural residential lands was observed in
SERS (Fig. 5), as examples of the expansion of Huanxi town and
Xinghai village. The residential lands increased from 1768 ha in
1988e2085 ha in 2005, and to 2370 ha in 2014, with an increase
rate of 18.6 ha/yr in stage 1 and 31.7 ha/yr in stage 2, while 70% of
60 Y. Tian et al. / Ocean & Coastal Management 135 (2017) 56e64

Fig. 2. Images for typical landscape changes of wetlands caused by multiple human activities from 1988 to 2014 (A, conversion from wetland to paddy field; B, conversion from
wetland to urban land; C, conversion from wetland to industrial land; D, conversion from natural wetland to aquaculture pond; E, conversion from wetland to transportation land; F,
wetland fragmentation).

Fig. 3. Land cover maps of the Shuangtai Estuary Ramsar site.

the increase was attributable to transformation of Phragmites aus-
tralis marshlands. Urbanization was concentrated in each of the five
towns, which significantly encroached wetlands. And the urbani-
zation processes were accelerated during stage 2.
3.3.3. Oil industrial footprint
Massive oil industrial footprints could be observed widely in the
overall study region (Fig. 6). Great demands of oil made the number
of oil wells increase by 87% in the past 26 years, while lands for
refining factory increased by 577 ha. The industrial lands expanded
from 172 ha in 1988 to 514 ha in 2005, and 950 ha in 2014, when
faster expansion occurred in stage 2 than 1 and a large number of
wetlands were occupied.
Phragmites australis marsh was the largest wetland type lost to
 Y. Tian et al. / Ocean & Coastal Management 135 (2017) 56e64 61

Fig. 4. Conversion of wetlands to croplands (bar graph shows the area changes of dry Fig. 6. Industrial lands converted from wetlands (bar graph shows the area changes of
farmland and paddy field). industrial lands).

marshlands (Fig. 7). Aquaculture ponds increased from 2019 ha in

1988e8627 ha in 2005, to 11,732 ha in 2014, and had an annual
averaged increase rate of 182.6 ha/yr in stage 1 and 345 ha/yr in
stage 2. Aquaculture was especially promoted in the southeastern
part and the southern part adjacent to sea and during the last
decade. In addition, the information from the satellite images
showed that the Phragmites australis marshland was fragmented
into a large number of patches because small patches are suited to
breeding crabs and fishes.

Fig. 5. Residential lands converted from wetlands (bar graph show the area changes of
residential lands).

the industrial land expansion, mainly occurring in the western side

of the Shuangtaizi River with the loss area being 589 ha, followed
by the disappearance of 144 ha intertidal zone in the study period
caused by increased industrial lands largely for oil industry.

3.3.4. Aquaculture development

Results suggested notable transforms from natural wetlands to
aquaculture ponds, including 5468 ha intertidal zones, 2699 ha Fig. 7. Natural wetlands converted to aquaculture ponds (bar graph shows the area
Suaeda heteroptera marshlands, and 1627 ha Phragmites australis changes of aquaculture ponds).
62 Y. Tian et al. / Ocean & Coastal Management 135 (2017) 56e64
Fig. 8. Expansion of transportation lands at the expense of wetlands (bar graph shows
the area changes of transportation lands).
3.3.5. Enhancement of transportation network
Apparently, dense transportation networks, resulting from the
development of oil industry, aquaculture, tourism, contributed to
the fragmentation of wetlands from 1988 to 2014 (Fig. 8). Moreover,
the expansion of transportation lands (907 ha in stage 1 and
1336 ha in stage 2) induced the largest wetland loss considering all
of the built-up land types. In the past 26 years, increased trans-
portation lands for roads and ports have resulted in a decrease of
Phragmites australis marshland by 1712 ha, intertidal zone by
368 ha, and Suaeda heteroptera marshland by 234 ha. The man-
agement of oil wells and transportation of crude oil call for
adequate roads. Consumption of wetlands by constructing roads
can be observed widely in the SERS, especially in the second stage,
while the increased lands for three ports were found in the
southern part.
3.4. Wetlands fragmentation in the SERS
Human activities not only lead to area loss of wetlands, but also
result in wetland landscape changes. For example, denser trans-
portation network fragmented the natural wetlands. The landscape
metrics for the SERS in 1988, 2005, and 2014 summarized in Table 2
indicate the fragmentation trend of wetlands in the SERS during the
past 26 years. Changes in each of the five landscape metrics
represent the fragmentation of wetlands in the study area. The
number of patches (NP) increased substantially from 1146 in
1988e2076 in 2005, to 2495 in 2014, while patch density (PD)
increased dramatically from 0.81 to 1.35 and 1.63 per 100 ha.
Noticeable raised SHDI from 1.71 to 1.78 and to 1.81 reflected the
increment of landscape heterogeneity. In addition, the evidently
decreased contagion index values (CONTAG) from 63.49% to 62.15%
and to 61.48% and aggregation index (AI) from 97.17 to 95.67
demonstrated a degradation trend in landscape connectivity and
landscape aggregation, respectively.
4. Discussion
In this study, Landsat images were applied to differentiate
Table 2
Changes of landscape metrics of wetlands.
Year NP PD SHDI CONTAG AI
1988 1146 0.81 1.71 63.49 97.16
2005 2076 1.35 1.78 62.15 96.18
2014 2495 1.63 1.81 61.48 95.67
NP: number of patches; PD: patch density (/100ha); SHDI: Shannon's diversity in-
dex; CONTAG: unit, %, 0 < CONTAG  100; AI: Aggregation Index (0e100).
different land cover types in the SERS. Based on the obvious spec-
tral distinction and phenology, the SERS was classified into crop-
lands, built-up lands, and major wetland vegetation types, e.g.,
Phragmites australis and Suaeda heteroptera. In the process of our
classification, a decision tree containing a number of decision rules
was designed to classify different types of land cover. On the one
hand, object-oriented method effectively avoided the Salt and
Pepper noise (Dronova et al., 2011) and characterized the features
of various land cover types, which is significantly superior to pixel-
based classification method (Juel et al., 2015). On the other hand,
multiresolution segmentation conducted on the images with a
resolution of 30 m contributed more or less in the classification
errors. Although NDWI was applied to distinguish the wetland and
dry farmland, similar color, texture, and vicinity between these two
types led to some errors. However, the distinguishing between
Phragmites australis marshland and paddy field can be easily
accomplished based on color difference in the composition of
bands 5, 4, 3 of TM and bands 6, 5, 4 of OLI (paddy field has bright
green color and Phragmites australis has olivine color). A clear
change pattern of the wetlands at this site as well as human dis-
turbances can be observed using Landsat images. Notable shrinkage
and fragmentation demonstrated by the analysis of Landsat images
suggested severe disturbances from multiple human activities to
the wetland in SERS. More importantly, because of the 30 m reso-
lution of Landsate images analyzed here, it is inevitable that oil
wells as well as some narrow roads smaller than this resolution
could be omitted. Therefore actual human threats, especially the
petroleum industry, were more severe than those detected by
Landsat images. Additional assessment using higher resolution
images is a desirable.
Significant shrinkage and fragmentation of wetlands have been
observed in various regions of Northeast China (Wang et al., 2011a,
2011b) and other regions of the world (Kentula et al., 2004; Murray
et al., 2014). The influence of multiple human activities on wetlands
is particularly evident in the SERS. Our results indicate that the
obvious expansion of agricultural land and built-up land resulted in
both the shrinkage and fragmentation of the wetland in SERS
(Figs. 4e8). Petroleum industry and aquaculture brought about
more serious shrinkage and fragmentation of wetland in the SERS
than agriculture, which is different from the wetland conversion
reported for the Sanjiang Plain (Wang et al., 2011a), the whole
China (Gong et al., 2010), and other regions in the world (Rebelo
et al., 2009) where shrinkage and fragmentation of natural wet-
lands were mainly caused by the expansion of agricultural lands.
The ecological disturbances studying on the SERS conducted by
Chen et al. (2010) revealed likewise that human threats, especially
the agricultural reclamation, resulted in dramatic wetland loss and
fragmentation before 2005. Additionally, the expansion of cropland
was slowed down after the study area was listed as one of the
Ramsar sites, however, urbanization, petroleum industry, aqua-
culture development, and transportation network were markedly
promoted after 2005 than before. Aiming to an effective conser-
vation of wetlands in the SERS, two works should be carried out. On
the one hand, the management organization of SERS must pay
 Y. Tian et al. / Ocean & Coastal Management 135 (2017) 56e64
more attentions on these threats, especially the development of
petroleum industry and aquaculture, to reduce a further wetland
loss and fragmentation. On the other hand, scientific wetland
restoration is urgently needed including reversion from cultivated
land and abandoned oil well industrial land to wetland.
Obvious human disturbances and the corresponding changes in
wetland landscape structure further resulted in negative ecological
effects. As examples, the water and soil quality of wetland estuaries
has been observably polluted by chemical fertilizers and pesticides
from cropland (Chen et al., 2015), oil industry (Ji et al., 2007; Zhang
and Sun, 2005), and domestic sewage (Zhao et al., 2005). The hy-
drological process can be largely affected by irrigation, construc-
tions of roads, ports, drainage ditches, and levees (Kingsford, 2000;
Liu et al., 2004). Massive intertidal zones have been converted into
aquaculture ponds and ports, resulting in the invasion of sea water
to the wetland, intensified natural disasters, weakened self-
purification capacities, and impaired interception of sediment and
contaminant (Kirwan and Megonigal, 2013; Murray et al., 2014;
Seto and Fragkias, 2007). In addition, degraded wetlands could
affect regional climate (Allred et al., 2015; Zhao et al., 2005). Even
worse is the biological invasion due to degradation of native
marshes (An et al., 2007), which well further affect the habitat of
rare waterfowls. All these human-induced changes contribute to
degradation of habitat condition and influence biological diversity
(He et al., 2014; Lu et al., 2015). As one of the critical rest habitat for
migration waterfowl in the East Asia-Australasian Flyway, the SERS
provides habitats for the waterfowl, which is one of the most
important service provided by wetland ecosystem. However,
degraded habitat conditions and more severe human disturbances
restrict the supporting capacity. The famous landscape of “red
beach” dominated with Suaeda heteroptera may not be as attractive
to tourists as it used to be because of its dramatic shrinkage and
fragmentation.
According to our results, natural wetlands in the study region
disappeared by 16,485 ha during 1988~2005 and 7207 ha
2005~2014, indicating that the protection efforts has not been so
successful despite the present wetland was approved to be national
natural reserve in 1988 and has been a Ramsar site since 2005. In
fact, pursuing economic benefits promoted the expansion of urban
and agricultural lands and the development of petroleum industry
and aquaculture, when providing management and conservation
do not slow down. Based on our analysis and field investigation,
agricultural reclamation, road construction, and oil exploitation
still continue in the core area of the SERS. In the background of
revitalizing the old industrial base of Northeast China and the
strategy of the Belt and Road (“B & R”), great challenges still exist to
effectively protect natural wetlands in the SERS. Any feasible pro-
posal on protecting the natural wetland in SERS should come up
with long-term solutions and take into account financial diffi-
culties. Therefore effective policies and regulations which can bal-
ance the economic value and ecological services must be
implemented and enforced to sustain local conservation and
management of natural wetlands (Valiela and Fox, 2008).
5. Conclusions
In this study, Landsat images were applied to examine the
shrinkage and fragmentation of the Shuangtai Estuary wetland, as
well as direct threats from multiple human activities. Quantitative
analysis of Landsat images was carried out to differentiate land
cover types of this area. Our results indicate that expansion of
cropland and built-up land in the SERS happened at cost of massive
wetlands. The development of petroleum industry, aquaculture,
and tourism resulted in dramatic fragmentation of the wetland
landscape. Annual averaged loss rate of 571.8 ± 73 ha/yr from 1988
63
to 2005 and 455.6 ± 51 ha/yr from 2005 to 2014 and obvious
fragmentation trend of wetlands suggest that the current protec-
tion practice has very low effectiveness in spite of the fact that the
present wetland is a national natural reserve and a Ramsar site. As
the critical rest habitat for migration waterfowl, the degradation of
wetlands in the SERS and severe human threats gave rise to great
challenges for its effective conservation and management. These
threads largely come from aquaculture and petroleum industry
rather than agricultural reclamation and urbanization.
This study has demonstrated that moderate resolution remote
sensing images is effective in assessing wetland landscape dy-
namics and human disturbances, and the results are useful to
promote regional ecosystem managements and services. Mean-
while, similar assessment for the protection effectiveness for other
wetland nature reserves and Ramsar sites should be undertaken to
track the management or conservation effectiveness objectively.
Funding
This work was jointly supported by the National Natural Science
Foundation of China (Nos. 41401502, 41371403, and 41671219), the
National Key Research Program of China (2016YFC0500201-03),
and the Outstanding Young Scientist Foundation of IGA, CAS
(Y5H1061001).
References
Allred, B.W., Smith, W.K., Twidwell, D., Haggerty, J.H., Running, S.W., Naugle, D.E.,
Fuhlendorf, S.D., 2015. Ecosystem services lost to oil and gas in North America.
Science 348 (6233), 401e402.
Amler, E., Schmidt, M., Menz, G., 2015. Definitions and mapping of east African
wetlands: a review. Remote Sens. 7, 5256e5282.
An, S.Q., Li, H.B., Guan, B.H., Zhou, C.F., Wang, Z.S., Deng, Z.F., Zhi, Y.B., Liu, Y.H.,
Xu, C., Fang, S.B., Jiang, J.H., Li, H.L., 2007. China's natural wetlands: past
problems, current status, and future challenges. AMBIO 36 (4), 335e342.
Baatz, M., Scha €pe, A., 2000. Multiresolution segmentation: an optimization
approach for high quality multi-scale image segmentation. Angew. Geogr. Inf.
XII, 12e23.
Barker, R., King, D.J., 2012. Blanding's Turtle (Emydoidea blandingii) potential
habitat mapping using aerial orthophotographic imagery and object based
classification. Remote Sens. 4, 194e219.
Bi, X.L., Wang, B., Lu, Q.S., 2011. Fragmentation effects of oil wells and roads on the
Yellow River delta, North China. Ocean Coast. Manag. 54 (3), 256e264.
Bolca, M., Turkyilmaz, B., Kurucu, Y., Altinbas, U., Esetlili, M.T., Gulgun, B., 2007.
Determination of impact of urbanization on agricultural land and wetland land
use in Balcovas's delta by remote sensing and GIS technique. Environ. Monit.
Assess. 131, 409e419.
Cao, L., Fox, A.D., 2009. Birds and people both depend on China's wetlands. Nature
460, 173e173.
Chen, A.L., Zhu, B.Q., Chen, L.D., Wu, Y.H., Sun, R.H., 2010. Dynamic changes of
landscape pattern and eco-disturbance degree in Shuangtai estuary wetland of
Liaoning Province, China. Chin. J. Appl. Ecol. 21 (5), 1120e1128.
Chen, H.J., Wang, G.P., Lu, X.G., Jiang, M., Mendelssohn, I.A., 2015. Balancing the
needs of China's wetland conservation and rice production. Environ. Sci.
Technol. 49 (11), 6385e6393.
Coleman, J.M., Huh, O.K., Braud Jr., D., 2008. Wetland loss in world deltas. J. Coast.
Res. 24 (1), 1e14.
Dronova, I., 2015. Object-based image analysis in wetland research: a review.
Remote Sens. 7, 6380e6413.
Dronova, I., Gong, P., Wang, L., 2011. Object-based analysis and change detection of
major wetland cover types and their classification uncertainty during the low
water period at Poyang Lake, China. Remote Sens. Environ. 115 (12),
3220e3236.
Finlayson, C.M., Rea, N., 1999. Reasons for the loss and degradation of Australian
wetlands. Wetl. Ecol. Manag. 7, 1e11.
Gallant, A.L., 2015. The challenges of remote monitoring of wetlands. Remote Sens.
7, 10938e10950.
Giosan, L., Syvitski, J., Constantinescu, S., Day, J., 2014. Protect the world's deltas.
Nature 516, 31e33.
Giri, C., Ochieng, E., Tieszen, L.L., Zhu, Z., Singh, A., Loveland, T., Masek, J., Duke, N.,
2011. Status and distribution of mangrove forests of the world using earth
observation satellite data. Glob. Ecol. Biogeogr. 20 (1), 154e159.
Gong, P., Niu, J.G., Cheng, X., Zhao, K.Y., Zhou, D.M., Guo, J.H., Liang, L., Wang, X.F.,
Li, D.D., Huang, H.B., Wang, Y., Wang, K., Li, W.N., Wang, X.W., Ying, Q., Yang, Z.Z.,
Ye, Y.F., Li, Z., Zhuang, D.F., Chi, Y.B., Zhou, H.Z., Yan, J., 2010. China's wetland
change (1990e2000) determined by remote sensing. Sci. China earth Sci. 53 (1),
64 Y. Tian et al. / Ocean & Coastal Management 135 (2017) 56e64
1e7.
He, C.Y., Liu, Z.F., Tian, J., Ma, Q., 2014. Urban expansion dynamics and natural
habitat loss in China: a multiscale landscape perspective. Glob. Change Biol. 20,
2886e2902.
Jafari, N., 2009. Ecological integrity of wetland, their functions and sustainable use.
J. Ecol. Nat. Environ. 1 (3), 045e054.
Ji, G.D., Sun, T.H., Ni, J.R., 2007. Impact of heavy oil-polluted soils on reed wetlands.
Ecol. Eng. 29, 272e279.
Jia, M.M., Zhang, Y.Z., Wang, Z.M., Song, K.S., Ren, C.Y., 2014. Mapping the distri-
bution of mangrove species in the core zone of Mai Po marshes nature reserve,
Hong Kong, using hyperspectral data and high-resolution data. Int. J. Appl. Earth
Obs. Geoinform. 33, 226e231.
Juel, A., Groom, G.B., Svenning, J.C., Ejrnaes, R., 2015. Spatial application of random
forest models for fine-scale coastal vegetation classification using object based
analysis of aerial orthophoto and DEM data. Int. J. Appl. Earth Obs. Geoinform.
42, 106e114.
Junk, W.J., An, S.Q., Finlayson, C.M., Gopal, B., Kvet, J., Mitchell, S.A., Mitsch, W.J.,
Robarts, R.D., 2013. Current state of knowledge regarding the world's wetlands
and their future under global climate change: a synthesis. Aquat. Sci. 75,
151e167.
Keddy, P.A., 2010. Wetland Ecology: Principles and Conservation. Cambridge Uni-
versity Press, Cambridge.
Kentula, M.E., Gwin, S.E., Pierson, S.M., 2004. Tracking changes in wetlands with
urbanization: sixteen years of experience in Portland, Oregon, USA. Wetlands
24 (4), 734e743.
Kingsford, 2000. Ecological impacts of dams, water diversions and river manage-
ment on floodplain wetlands in Australia. Aust. Ecol. 25, 109e127.
Kirwan, M.L., Megonigal, J.P., 2013. Tidal wetland stability in the face of human
impacts and sea-level rise. Nature 504, 53e60.
Lee, S.Y., Dunn, R.J.K., Young, R.A., Connolly, R.M., Dale, P.E.R., Dehayr, R.,
Lemckert, C.J., Mckinnon, S., Powell, B., Teasdale, P.R., Welsh, D.T., 2006. Impact
of urbanization on coastal wetland structure and function. Austral Ecol. 31,
149e163.
Li, Y.F., Shi, Y.L., Zhu, X.D., Cao, H.H., Yu, T., 2014. Coastal wetland loss and envi-
ronmental change due to rapid urban expansion in Lianyungang, Jiangsu, China.
Reg. Environ. Change 14, 1175e1188.
Lin, W.P., Chen, G.S., Guo, P.P., Zhu, W.Q., Zhang, D.H., 2015. Remote-sensing
monitoring of dominated plant species distribution and dynamics at jiuduansha
wetland in Shanghai, China. Remote Sens. 7, 10227e10241.
Liu, H.Y., Zhang, S.K., Li, Z.F., Lu, X.G., Yang, Q., 2004. Impacts on wetlands of large-
scale land-use changes by agricultural development: the small Sanjiang plain,
China. AMBIO 33 (6), 306e310.
Lu, C.Y., Wang, Z.M., Li, L., Wu, P.Z., Mao, D.H., Jia, M.M., Dong, Z.Y., 2015. Assessing
the conservation effectiveness of wetland protected areas in Northeast China.
Wetl. Ecol. Manag. http://dx.doi.org/10.1007/s11273-015-9462-y.
Mao, D.H., Wang, Z.M., Li, L., Song, K.S., Jia, M.M., 2014. Quantitative assessment of
human-induced impacts on marshes in Northeast China from 2000 to 2011.
Ecol. Eng. 68, 97e104.
Murray, N.J., Clemens, R.S., Phinn, S.R., Possingham, H.P., Fuller, R.A., 2014. Tracking
the rapid loss of tidal wetlands in the Yellow Sea. Front. Ecol. Environ. 12 (5),
267e272.
Nagendra, H., Lucas, R., Honrado, J.P., Jongman, R.H.G., Tarantino, C., Adamo, M.,
Mairota, P., 2013. Remote sensing for conservation monitoring: assessing pro-
tected areas, habitat extent, habitat condition, species diversity, and threats.
Ecol. Indic. 33, 45e59.
Ozesmi, S.L., Bauer, M.E., 2002. Satellite remote sensing of wetlands. Wetl. Ecol.
Manag. 10, 381e402.
Rebelo, L.M., Finlayson, C.M., Nagabhatla, N., 2009. Remote sensing and GIS for
wetland inventory, mapping and change analysis. J. Environ. Manag. 90,
2144e2153.
Ren, L.Y., Yue, W.Z., Li, J.L., Wu, C.F., 2010. Impact of economic development on
wetlands in Hangzhou Bay industrial Belt. J. Geogr. Sci. 20 (3), 406e416.
Seto, K.C., Fragkias, M., 2007. Mangrove conversion and aquaculture development in
Vietnam: a remote sensing-based approach for evaluating the Ramsar
Convention on Wetlands. Glob. Environ. Change 17 (3), 486e500.
Song, K.S., Wang, Z.M., Li, L., Tedesco, L., Li, F., Jin, C., Du, J., 2012. Wetland shrinkage,
fragmentation and their links to agriculture in the Muleng-Xingkai, Plain,
China. J. Environ. Manag. 111, 120e132.
Song, K.S., Wang, Z.M., Du, J., Liu, L., Zeng, L.H., Ren, C.Y., 2014. Wetland degradation:
it's driving forces and environmental impacts in the Sanjiang Plain, China.
Environ. Manag. 54, 255e271.
Valiela, I., Fox, S.E., 2008. Managing coastal wetlands. Science 319, 290e291.
Wang, Z.M., Song, K.S., Ma, W.H., Ren, C.Y., Zhang, B., Liu, D.W., Chen, J.M., Song, C.C.,
2011a. Loss and fragmentation of marshes in the Sanjiang Plain, Northeast
China, 1954e2005. Wetlands 31, 945e954.
Wang, Z.M., Huang, N., Luo, L., Ren, C.Y., Song, K.S., Chen, J.M., 2011b. Shrinkage and
fragmentation of marshes in the west Songnen plain, China, from 1954 to 2008
and its possible causes. Int. J. Appl. Earth Obs. Geoinform. 13 (3), 477e486.
Wang, Z.M., Wu, J.G., Madden, M., Mao, D.H., 2012. China's wetlands: conservation
plans and policy impacts. AMBIO 41, 782e786.
Wang, Z.M., Mao, D.H., Li, L., Jia, M.M., Dong, Z.Y., Miao, Z.H., 2015. Quantifying
changes in multiple ecosystem services during 1992-2012 in the Sanjiang Plain
of China. Sci. Total Environ. 514, 119e130.
Wright, C.K., Wimberly, M.C., 2013. Recent land use change in the Western Corn Belt
threatens grasslands and wetlands. PNAS 110 (10), 4134e4139.
Zhang, J.F., Sun, Q.X., 2005. Causes of wetland degradation and ecological restora-
tion in the Yellow River delta region. For. Stud. China 7 (2), 15e18.
Zhao, S.Q., Da, L.J., Tang, Z.Y., Fang, H.J., Song, K., Fang, J.Y., 2005. Ecological conse-
quences of rapid urban expansion: Shanghai, China. Front. Ecol. Environ. 4 (7),
341e346.
Zheng, Y.M., Zhang, H.Y., Niu, Z.G., Gong, P., 2012. Protection efficacy of national
wetland reserves in China. Chin. Sci. Bull. 57 (10), 1116e1134.