Remote Observations in China's Ramsar Sites: Wetland Dynamics, Anthropogenic Threats, and Implications for Sustainable Development Goals
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
AAAS
Journal of Remote Sensing
Volume 2021, Article ID 9849343, 13 pages
https://doi.org/10.34133/2021/9849343
Research Article
Remote Observations in China’s Ramsar Sites: Wetland Dynamics,
Anthropogenic Threats, and Implications for Sustainable
Development Goals
D. Mao ,1 Z. Wang ,1,2 Y. Wang,3 C.-Y. Choi,4 M. Jia,1 M. V. Jackson ,5 and R. A. Fuller 5
1
Key Laboratory of Wetland Ecology and Environment, Northeast Institute of Geography and Agroecology, Chinese Academy
of Sciences, Changchun 130102, China
2
National Earth System Science Data Center, Beijing, China
3
Department of Natural Resources Sciences, University of Rhode Island, Kingston, RI 02881, USA
4
School of Environmental Science and Engineering, College of Engineering, Southern University of Science and Technology,
Shenzhen 518055, China
5
School of Biological Sciences, The University of Queensland, Brisbane, Australia
Correspondence should be addressed to Z. Wang; zongmingwang@neigae.ac.cn
Received 5 November 2020; Accepted 13 April 2021; Published 15 May 2021
Copyright © 2021 D. Mao et al. Exclusive Licensee Aerospace Information Research Institute, Chinese Academy of Sciences.
Distributed under a Creative Commons Attribution License (CC BY 4.0).
The Ramsar Convention on Wetlands is an international framework through which countries identify and protect important
wetlands. Yet Ramsar wetlands are under substantial anthropogenic pressure worldwide, and tracking ecological change relies
on multitemporal data sets. Here, we evaluated the spatial extent, temporal change, and anthropogenic threat to Ramsar
wetlands at a national scale across China to determine whether their management is currently sustainable. We analyzed Landsat
data to examine wetland dynamics and anthropogenic threats at the 57 Ramsar wetlands in China between 1980 and 2018.
Results reveal that Ramsar sites play important roles in preventing wetland loss compared to the dramatic decline of wetlands in
the surrounding areas. However, there are declines in wetland area at 18 Ramsar sites. Among those, six lost a wetland area
greater than 100 km2, primarily caused by agricultural activities. Consistent expansion of anthropogenic land covers occurred
within 43 (75%) Ramsar sites, and anthropogenic threats from land cover change were particularly notable in eastern China.
Aquaculture pond expansion and Spartina alterniflora invasion were prominent threats to coastal Ramsar wetlands. The
observations within China’s Ramsar sites, which in management regulations have higher levels of protection than other
wetlands, can help track progress towards achieving United Nations Sustainable Development Goals (SDGs). The study findings
suggest that further and timely actions are required to control the loss and degradation of wetland ecosystems.
1. Introduction able wetland ecosystem management and hinder policy
improvement.
Wetlands are among the most fragile ecosystems, vulnerable China has the fourth largest wetland coverage in the
to climate change and human disturbances, and have experi- world. Its wetlands account for 5% of the nation’s territorial
enced striking loss and degradation worldwide [1–3]. Under- area [7]. However, wetland loss has occurred rapidly in
standing the spatial extent, temporal changes, and ecosystem China despite the value of these ecosystems being recognized
threats of wetlands is important in achieving the United [8, 9]. Multiple studies have reported on wetland changes and
Nations Sustainable Development Goals (SDGs) which high- their driving factors [5, 10]. For example, it was estimated
light the protection of wetlands [4, 5]. However, due to the that the total wetland area in China declined by more than
difficulty in accurately delineating wetland boundaries 50,000 km2 between 1990 and 2000, and 33% of China’s wet-
caused by complex composition and spectral features, there lands were lost between 1978 and 2008 [11, 12]. Cropland
are few reliable assessments of wetland change at global or expansion contributed 60% to the loss of vegetated wetlands
national scales [6]. This deficiency may result in unsustain- in China between 1990 and 2010 [13]. Yet little is known
2 Journal of Remote Sensing
about the anthropogenic processes driving those losses, ham- Site No. 1,976) to the Dongzhaigang in Hainan in the south
pering sustainable management and long-term conservation (19° 59 ′ N, Site No. 553). According to the National Wetlands
of wetlands. China has established a large number of wetland Conservation Program (NWCP), the largest number of sites
protected areas [9], but it is unclear whether the actions have (16) occur in the coastal region (CRC), while the fewest (1)
stopped wetland losses within these protected areas. are found in the lower and middle reaches of the Yellow River
The Ramsar Convention on Wetlands, established in (YER; Figures 1(a) and 1(c)). The largest Ramsar site, the
1971, was the first globally coordinated institutional frame- Selincuo Wetlands in Tibet (Site No. 2,352), has an area of
work for wetland conservation and wise use [14]. Currently, 19,836 km2, while the Hangzhou Xixi Wetlands in Zhejiang
2,354 Ramsar sites are designated and distributed among (Site No. 1,867) has the smallest area (3 km2). China’s Ram-
170 countries, protecting 13–18% of global wetlands [3]. sar sites include multiple natural wetland types including
Since the adoption of the Ramsar Convention in 1992, China inland swamp, marsh, lake, river, coastal mangrove, bare
had designated 57 Ramsar wetland sites by 2019 with a total tidal flat, estuary, and shallow marine water. Detailed desig-
area of 69,486 km2 aimed at protecting the principal ecologi- nation number, name, and date for each Ramsar site is pro-
cal characteristics and biodiversity of those areas (http:// vided in supplementary Table S1.
www.ramsar.org). The protection of Ramsar wetlands has
strong governmental commitment through domestic legisla- 2.2. Landsat Satellite Images. Although the widely used
tion in China, and a systematic evaluation of Ramsar sites is national land use data product has advantages such as a
critical for informing their management and indicative for five-year interval and 30 m spatial resolution, we performed
wetland protection across the country [15]. our image classification for a finer wetland classification with
Ecological change in Ramsar wetlands can be assessed reliable data accuracy and consistency for each Ramsar site.
through multiple indicators including widely used spectral We selected 5 time periods at approximately decadal inter-
vegetation indices (e.g., normalized difference vegetation vals (1980, 1990, 2000, 2010, and 2018) to evaluate the effects
index, NDVI), landscape metrics, and directly measured eco- of the Ramsar Convention on Wetlands in China. Due to the
system parameters [16, 17]. Although wetland loss and deg- scarcity of available images in 1980, images in adjacent years
radation have been documented to some degree using those were supplemented to classify land cover in 1980. Sixty-four
methods in several Ramsar sites in China, specific anthropo- Landsat scenes were required to cover the entire extent of all
genic threats, such as agricultural cultivation, industrial foot- Ramsar sites (Figure 2). Multiseasonal images characterizing
print, aquaculture development, and biological invasions, phenological features could support the identification of dif-
have not been quantitatively investigated. A systematic ferent land cover types. Therefore, for obtaining the land
understanding of spatial extent, temporal changes, and major cover data sets for each the five years, we collected a total of
threats to Ramsar wetlands is required for wetland manage- 768 images (with cloud cover < 10%) from the open access
ment and policy improvement to achieve ecological source by the United States Geological Survey (USGS). Prior
sustainability. to image classification, data preprocessing including geomet-
In this study, we evaluated the effects that Ramsar desig- ric, topographic, and radiometric corrections was performed
nation has had on wetland conservation in China, with the for all images using the ENVI 5.1 software package.
aim of informing wetland management and protection mea-
sures at the national scale. To achieve this goal, we mapped 2.3. Land Cover Classification. To reduce seasonal variation
land cover changes within and adjacent to the 57 Ramsar in land cover, images in the months from June to September
sites between 1980 and 2018 at decadal intervals. We com- were primarily used, while images from other months were
pared wetland change within and adjacent to the Ramsar used as auxiliary data. A hybrid approach combining
sites before and after their designation and investigated the object-based image analysis and hierarchical decision-tree
coverage proportion and temporal changes in anthropogenic classification (HOHC) performed in the eCognition software
land covers including cropland and built-up land. Finally, we (version 9.2) was used due to its advantages in segmenting
examined the expansion of aquaculture ponds and damaging images into homogeneous objects and semiautomated distin-
invasive species, Spartina alterniflora (S. alterniflora), intro- guishing of these objects ensuring both classification accu-
duced into China’s coastal Ramsar wetlands. Our evaluation racy and efficiency [7, 18]. The classification accuracies of
of wetland protection efficacy across China’s Ramsar sites final data sets were evaluated using independent ground ref-
will provide valuable insights into management and policy erence samples, which were acquired from field surveys, pub-
priorities for wetlands to promote the protection of wetlands lic databases, and high-resolution images from Google Earth
and biodiversity worldwide. [19–21]. An outline of the process used for Landsat image
classification is adopted from Mao et al. [7] and illustrated
2. Materials and Methods in Figure 3.
We determined and adapted a classification by referen-
2.1. Study Area. Ramsar sites protect the habitat of a large cing the level of Class I of the National Land Cover Database
number of endangered migratory waterbirds, aquatic ani- of China (ChinaCover) [18] and considering the classifica-
mals, and natural plants or ecosystems and are an important tion potential of the Landsat images (Table 1). We were able
component of the wetland protection system in China. The to categorize land cover classes into natural (woodland,
57 Ramsar sites (Figure 1(a)) designated by 2019 occur from grassland, wetland, and barren land) and anthropogenic
the Grand Khingan Hanma Wetlands in the north (51° 35 ′ N, (cropland and built-up land) types according to Table 1. To
Journal of Remote Sensing 3
2018 8
2015 3
N 2013 5
2011 4
2009 1
Year
2008 6
2005 9
2002 14
1995 1
1992 6
0 3 6 9 12 15
Number
(b)
20
16
15 14
Number
10
10
7
5
5 4
1
0
0 300 600 km CRC NEC YAR QTP YGP IXP YER
Geographic region of NWCP
(a) (c)
Geographic region of NWCP Ramsar site
Northeast China (NEC) River
Lake
Qinghai-Tibet Plateau (QTP)
Coastline
Coastal region (CRC)
Vegetated wetland
Yunnan-Guizhou Plateau (YGP)
Country boundary
Southand Southeast China (SSC)
Inner Mongolia-Xingjiang Plateau (IXP)
Lower and middle reaches of the Yangtze River (YAR)
Lower and middle reaches of the Yellow River (YER)
Figure 1: Locations of China’s Ramsar sites: (a) distribution of Ramsar sites and division of the National Wetlands Conservation Program
(NWCP); (b) number of Ramsar sites designated in different years; (c) number of Ramsar sites located in each NWCP region.
permit a detailed assessment of anthropogenic land cover, high intensity of human management. In this study, classifi-
cropland was further divided into dry farmland and paddy cations were performed to the level of Class II for the inland
field, while built-up land was classified into residential land, Ramsar wetland sites and Class III for coastal Ramsar
industrial land, transportation land, and mining land. The wetland sites. To investigate conversions among wetland cat-
classification at the second level (Class II) for wetlands was egories, we separated human-made wetlands including reser-
derived from our proposed wetland classification system for voirs/artificial ponds and canals/channels from natural
national wetland mapping [7]. According the Ramsar wet- wetlands. Moreover, to examine the impacts of aquaculture
land definition, we classified all waterbodies in the Ramsar development on natural wetlands, the expansion of aquacul-
site as wetlands. Although a paddy field is an important ture ponds in the coastal Ramsar sites was assessed using the
human-made wetland type, we nonetheless classified it into methods described in Ren et al. [22]. Exotic plant invasion is
cropland to reflect its limited ecosystem function and the also impacting a number of Ramsar sites in China. S.
4 Journal of Remote Sensing
70°E 80°E 90°E 100°E 110°E 120°E 130°E
50°N
50°N
Cropland
40°N
40°N
Oil footprint
30°N
30°N
Sp
ar
tin
aa
lte
rn
ifl
or
a
20°N
20°N
ond
re p
ultu
Elevation (m) uac
Aq
High: 8848
10°N
0 500 1,000 km
Low: –154 10°N
90°E 100°E 110°E 120°E
River Coastline
Ramsar sites Country boundary
Landsat coverage Provincial boundary
Figure 2: Landsat footprints necessary for mapping land covers and image examples for various anthropogenic threats in the Ramsar sites.
alterniflora, native to the Atlantic coastal America, has been ing the evident human disturbances from human activities
the most invasive species in coastal wetlands. In our study, on coastal wetlands, areal changes in aquaculture ponds were
data on the distribution of S. alterniflora were obtained from quantified during the investigated different periods. More-
the study of Liu et al. [23] and Mao et al. [24] to evaluate over, expansion of artificially introduced exotic species from
exotic plant invasion on natural wetlands. North America, S. alterniflora, was estimated in major
For evaluating the classification results, accuracies were coastal Ramsar sites from 1990. Therefore, the top four
assessed by field wetland samples which were provided by the anthropogenic threats to Ramsar wetlands considered in this
Wetland Science Data Centre of China (http://www.igadc.cn/). study were agricultural cultivation, built-up land expansion
These samples were investigated with the support of several including urbanization and industrial footprint, aquaculture
national research projects [7]. All classifications achieved development, and exotic species invasion.
consistency with validation samples characterized by overall For a Ramsar site with two different anthropogenic
accuracies for both Class II and Class III larger than 85%. threats (e.g., cropland and built-up land), the direct anthro-
pogenic threat was calculated as
2.4. Statistical Analysis. We compared areal changes in wet-
lands from 1980 to 2018 within and around (10 km buffer) A1 + A2
China’s Ramsar sites. We also used data sets in the four time Direct anthropogenic threat = , ð1Þ
Atotal
periods to investigate changes in wetland area and anthropo-
genic threats before and after their Ramsar designation. The
areal proportion of anthropogenic land covers within the where A1 represents the area of cropland in a Ramsar site, A2
Ramsar site was used to represent the direct anthropogenic represents the area of built-up land in a Ramsar site, and
threats in corresponding sites [25] (equation (1)). Consider- Atotal represents the total area of a Ramsar site.
Journal of Remote Sensing 5
Landsat image acquisition Auxiliary data acquisition Field investigation
and selection and processing Google Earth
ENVI
Multi-resolution Training and
Image preprocessing
eCognition segmentation OBIA validation database
Decision trees
Change
Hybrid image classification
detection
Product validation and
standardization
Multi-temporal data sets of land cover
Cropland expansion
Aquaculture pond expansion
Anthropogenic Internal type
Residential land expansion
threat identification conversion
Exotic species invasion
Industrial footprint
Existing problems and managing implications
Figure 3: General flowchart of Landsat data processing and data analysis.
Table 1: Land covers at different levels classified in this study.
Land covers Class I Class II Class III
Woodland
Grassland
Barren land
Natural
Human-made wetland (reservoir/artificial pond, canal/channel, and salt pan) Coastal aquaculture pond
Wetland Natural waterbody (lake, river, estuary, lagoon, and shallow marine water)
Other wetlands (swamp, marsh, and mudflat) S. alterniflora
Dry farmland
Cropland
Anthropogenic Paddy field
Built-up land
3. Results tze Estuarine Wetland Nature Reserve for Chinese Sturgeon
(Site No. 1,730) in Shanghai, the Nanpeng Archipelago Wet-
3.1. Spatial Extent and Temporal Changes of Wetlands within lands (Site No. 2,249) in Guangdong, and the Dalian
and Adjacent to Ramsar Sites. Landsat-based observations National Spotted Seal Nature Reserve (Site No. 1,147) in
revealed that there were 32,692 km2 of wetlands protected Liaoning, had wetland coverages of 100%. However, 11 sites
in the 57 Ramsar sites in 2018. Among these, 19 sites (33%) (19%) had wetland coverages of only 30%–50%, while
had wetland coverage larger than 90% and 14 sites had wet- another 13 sites (23%) had wetland coverage lower than
land coverage (including all water bodies) ranging between 30%. The Selincuo Wetlands in Tibet (Site No. 2,352) con-
50% and 90% (Figure 4(a)). Sites with high wetland coverage tained the largest total wetland area of 6,094 km2, dominated
mostly occurred in eastern China (Figure 4(b)), especially in by lakes. The Dajiu Lake Wetland (Site No. 2,186) in Hubei
the CRC. The Ramsar sites in Inner Mongolia-Xinjiang Pla- had the smallest wetland area (1.4 km2) and coverage (1%)
teau (MXP) and Yunnan-Guizhou Plateau (YGP) had lower among all the sites, being covered mainly by woodland. Spa-
wetland coverages than that those in other geographic tial patterns of land cover in each Ramsar site were provided
regions. Three sites in shallow marine waters, i.e., the Yang- in supplementary Figure S1.
6 Journal of Remote Sensing
90–100 19 50–100 4
Wetland coverage (%)
Direct anthropogenic
70–90 6 30–50 6
threats (%)
20–30 5
50–70 8
10–20 8
30–50 11
1–10 18
10–30 7 0–1 8
0–10 6 0 8
0 5 10 15 20 0 5 10 15 20
Number of Ramsar sites Number of Ramsar sites
(a) (c)
YER 97.5% YER 2.1%
Geographic region of
Geographic region of
IXP 9.1% IXP 12.1%
YGP 38.0% YGP 10.8%
NWCP
NWCP
QTP 56.1% QTP 3.9%
YAR 57.8% YAR 24.4%
NEC 47.3% NEC 20.8%
CRC 91.4% CRC 6.9%
0 20 40 60 80 100 0 10 20 30 40
Wetland coverage (%) Direct anthropogenic threats (%)
(b) (d)
Figure 4: Statistics of wetland coverage and direct anthropogenic threats in Ramsar sites in 2018: (a) number of Ramsar sites with various
wetland coverage classes; (b) mean value of Ramsar wetland coverages in different geographic regions of the National Wetlands
Conservation Program (NWCP); (c) number of Ramsar sites with various classes of direct anthropogenic threats; (d) mean value of direct
anthropogenic threats in different Ramsar sites over the geographic regions of NWCP: coastal region (CRC), Northeast China (NEC),
lower and middle reaches of the Yangtze River (YAR), Qinghai-Tibet Plateau (QTP), Yunnan-Guizhou Plateau (YGP), Inner Mongolia-
Xinjiang Plateau (IXP), and lower and middle reaches of the Yellow River (YER); error bars show standard deviations.
Total wetland area across the Ramsar estate decreased by nated as Ramsar sites. For example, the decreased wetland
an estimated 6.6% from 1980 to 2018, while wetland area in area in Xianghai (Site No. 548) designated as a Ramsar site
the 10 km buffer adjacent to the Ramsar sites declined by in 1992 has been reversed.
22.5% in the same period. Natural wetlands in the Ramsar
sites lost 12.3% of their total area, while natural wetlands in 3.2. Direct Anthropogenic Threats and Their Dynamics in the
the 10 km buffer adjacent to the Ramsar sites lost 26.1%. Ramsar Sites. Only 8 Ramsar sites had no sign of anthropo-
Wetland loss primarily occurred in the Ramsar sites of genic land covers in 2018, while wetlands in the other 49 sites
Northeast China (NEC) and CRC. However, human-made (86%) had been subject to various extents by anthropogenic
wetlands in the Ramsar sites expanded by 246% and in the threats over the past four decades (Figure 4(c)). Specifically,
10 km buffer adjacent to the Ramsar sites by 31%. The 10 sites (17%) had an areal proportion of anthropogenic land
expanded human-made wetlands were mostly identified in cover larger than 30%. Agricultural activities imposed the
Ramsar sites in the CRC (70%). Ramsar sites appear to have largest influences on wetlands, which were identified as the
played an important role in reducing wetland loss and pro- most severe in the San Jiang National Nature Reserve in Hei-
tecting wetlands. longjiang (Site No. 1,152). In this site, 64% of the whole area
From 1980 to 2018, wetland areas in 41 of the 57 Ramsar (2,232 km2) was covered by cropland in 2018. In addition to
sites (72%) showed changes (Figure 5). Specifically, wetland agricultural encroachment into wetlands, the largest propor-
area in 18 sites (32%) appeared to decline consistently despite tion (13%) of built-up land area dominated by residential
these wetlands having been designated as Ramsar sites. land was found for the Shengjin Lake National Nature
Another 6 sites showed an initial wetland increase followed Reserve in Anhui (Site No. 2,248), and significant industrial
by decrease. In particular, more than 100 km2 was lost from footprints occurred in the Liaohe Estuary in Liaoning (Site
each of 6 sites between 1980 and 2018. The San Jiang No. 1,441). Particularly acute anthropogenic threats were
National Nature Reserve in Heilongjiang (Site No. 1,152) observed in sites in the lower and middle reaches of the
had a most remarkable net loss of wetland (693 km2), mostly Yangtze River (YAR) and NEC compared to other geo-
resulting from agricultural cultivation. After the designation graphic regions (Figure 4(d)), with the former having 24.4%
of Ramsar site, 21 km2 of wetlands were converted into crop- of its area subject to anthropogenic threats.
land. In contrast, 7 sites (12%) showed consistent wetland From 1980 to 2018, increases in direct anthropogenic
increase, with the largest wetland increase (61 km2) being threats occurred in 40 sites, with an initial decrease followed
dominated by marsh and swamp expansion occurring in by an increase for 3 sites (supplementary Figure S2). The San
the Nanweng River National Nature Reserve in Heilongjiang Jiang National Nature Reserve in Heilongjiang (Site No.
(Site No. 1,976). Another 10 sites had an initial decrease in 1,152) had the largest areal increase of anthropogenic land
wetland area followed by increase after being formally desig- covers (692 km2), with 95% of the increase resulting from
Journal of Remote Sensing 7
100 No. 548 100 No. 549 100 100 No. 551 100 No. 552 100
75 75 75 No. 550 75 75 75 No. 553
50 50 50 50 50 50
25 25 25 25 25 25
0 0 0 0 0 0
100 100 100 100 100 100 No. 1,151
No. 1,146 No. 1,148
75 No. 1,144 75 No. 1,145 75 75 75 No. 1,149 75
50 50 50 50 50 50
25 25 25 25 25 25
0 0 0 0 0 0
100 No. 1,152 100 No. 1,154 100 No. 1,155 100 100 No. 1,157 100 No. 1,435
75 75 75 75 75 75
No. 1,156
50 50 50 50 50 50
25 25 25 25 25 25
0 0 0 0 0 0
Wetland coverage (%)
100 100 100 100 100 No. 1,727 100
75 No. 1,436 75 No. 1,439 75 75 No. 1,442 75 75 No. 1,728
No. 1,441
50 50 50 50 50 50
25 25 25 25 25 25
0 0 0 0 0 0
100 100 No. 1,731 100 100 No. 1,976 100 No. 1,977 100 No. 1,978
75 No. 1,729 75 75 75 75 75
50 50 50 No. 1,867 50 50 50
25 25 25 25 25 25
0 0 0 0 0 0
100 No. 2,184 100 No. 2,185 100 100 No. 2,188 100 No. 2,246 100 No. 2,347
75 75 75 No. 2,187 75 75 75
50 50 50 50 50 50
25 25 25 25 25 25
0 0 0 0 0 0
100 No. 2,349 100 No. 2,350 100 No. 2,351 100 No. 2,352 100 No. 2,353
75 75 75 75 75
50 50 50 50 50
25 25 25 25 25
0 0 0 0 0
Wetland coverage changes in different Ramsar sites (black dotted line denotes the designation year)
Consistent increase Decrease followed by increase
Consistent decline Increase followed by decrease
Figure 5: Various wetland changes from 1980 to 2018 over different Ramsar sites (time points are 1980, 1990, 2000, 2010, and 2018).
cropland expansion. However, the rate of increase of direct 3.3. Aquaculture Pond Expansion and Exotic S. alterniflora
anthropogenic threats slowed markedly after its designation Invasion in Coastal Ramsar Sites. A striking expansion of
as a Ramsar site (Figure 6). The greatest expansion area of aquaculture ponds at the expense of natural wetlands
built-up land was observed in the Dong Dongting Hu in occurred in 8 coastal sites, including 4 sites where aquacul-
Hunan (Site No. 551) with an area increase of 98.5 km2, ture ponds expanded by more than 100 km2, with expansion
with the increase occurring mostly after its designation as continuing beyond the date of Ramsar site designation
Ramsar site. (Table 2). The most obvious expansion occurred in the Yan-
Only the Haifeng Wetland site in Guangdong (Site No. cheng National Nature Reserve with an areal increase of
1,727) had a consistent decrease of anthropogenic land 752 km2. The Liaohe Estuary experienced a marked expan-
covers (Figure 6), while another 5 sites experienced an sion (106 km2) between 2010 and 2018 after its Ramsar des-
increase followed by a decrease of direct anthropogenic ignation in 2004. A rapid expansion of aquaculture pond in
threats. The Wang Lake site in Hubei (Site No. 2,349) experi- the Zhanjiang Mangrove National Nature Reserve was iden-
enced the most striking decrease in direct anthropogenic tified between 1990 and 2000 but slowed after Ramsar desig-
threats from 25% to 12%, while the Zhanjiang Mangrove nation in 2002.
National Nature Reserve (Site No. 1,157) in Guangdong S. alterniflora invaded several coastal areas (Figure 7).
had the largest areal decline of anthropogenic land covers Notable S. alterniflora invasions of areas larger than 10 km2
(38 km2) after its Ramsar designation. However, built-up occurred in 4 Ramsar sites (Table 3). The Yancheng National
lands in all these 6 sites showed consistent expansion during Nature Reserve in Jiangsu had the most evident invasion of S.
the study period. alterniflora with the largest and consistent areal increase
8 Journal of Remote Sensing
N 1980 N 2000 N 2018 75 No. 1,152
Coverage proportion
50
25
0 15 30 km 0 15 30 km 0 15 30 km
0
N 1980 N 1990 N 2018 75 No. 551
Coverage proportion
50
25
0 5 10 km 0 5 10 km 0 5 10 km 0
N 1980 N 2010 N 2018 75 No. 1,727
Coverage proportion
50
25
0 5 10 km 0 5 10 km 0 5 10 km
0
N 1980 N 2000 N 2018 75 No. 2,349
Coverage proportion
50
25
0 3 6 km 0 3 6 km 0 3 6 km
0
N 1980 N 2000 N 2018 75 No. 1,157
Coverage proportion
50
25
0 15 30 km 0 15 30 km 0 15 30 km 0
Changes of ALC from 1980 to 2018
Dry farmland
Paddy field
Built-up land
Figure 6: Anthropogenic land cover (ALC) changes from 1980 to 2018 in Ramsar sites (the gray dotted line denotes its Ramsar designation
time; different colors in the line charts denote different change trends of direct anthropogenic threats).
from 3.2 km2 in 1990 to 111.2 km2 in 2018. The Dafeng The Chongming Dongtan Nature Reserve in Shanghai (Site
National Nature Reserve in Jiangsu (Site No. 1,145) had an No. 1,144) had S. alterniflora coverage of 15.3 km2 in 2018
increase in S. alterniflora coverage of 49.4 km2 between and S. alterniflora expansion of 10.4 km2 during 2000–2010.
1990 and 2000 but a slightly decreasing trend after 2000. It is noteworthy that S. alterniflora invasion from 2010 to
Journal of Remote Sensing 9
Table 2: Aquaculture pond expansion (km2) in selected Ramsar sites from 1980 to 2018.
Aquaculture pond expansion in different periods
Site number Site name
1980-1990 1990-2000 2000-2010 2010-2018
1,441 Liaohe Estuary 0 0.6 3 106.3
2,187 Yellow River Delta Wetland 13 0.1 0 -17.5
1,156 Yancheng National Nature Reserve 145.4 95.9 193.9 317.1
1,145 Dafeng National Nature Reserve 53 16.9 73.2 46.6
1,157 Zhanjiang Mangrove National Nature Reserve 25.3 198.4 30.8 21.4
1,726 Zhangjiangkou National Mangrove Nature Reserve 0 4.6 1.2 2.9
750 Mai Po Marshes and Inner Deep Bay 0 0.7 2.5 1.9
1,728 Beilun Estuary National Nature Reserve 0.4 0.4 0.1 2.5
1990 2000 2010 2018
N N N N
0 25 50 km 0 25 50 km 0 25 50 km 0 25 50 km
Yancheng National Nature Reserve Spartina alterniflora Image composition: Standard false color
Dafeng National Nature Reserve Chongming Dongtan Nature Reserve RGB: Bands 5, 4, 3
Figure 7: Expansion of S. alterniflora invaded into the coastal Ramsar sites.
Table 3: S. alterniflora changes in different coastal Ramsar sites (positive values denote areal increase and negative values denote areal
decline).
S. alterniflora areal changes in different periods (km2)
Site number Site name
1990-2000 2000-2010 2010-2018
2,187 Yellow River Delta Wetland 0 1.7 11.9
1,156 Yancheng National Nature Reserve 100.4 6.6 1.1
1,145 Dafeng National Nature Reserve 49.4 -1.7 -0.6
1,144 Chongming Dongtan Nature Reserve 3.9 10.4 1.0
1,726 Zhangjiangkou National Mangrove Nature Reserve 0.4 0.2 1.6
1,153 Shankou Mangrove Nature Reserve 0.5 1.8 0.6
10 Journal of Remote Sensing
2018 also occurred in a northern site, the Yellow River Delta wetland loss, wetland gain occurred in many Ramsar sites
Wetland in Shandong (Site No. 2,187), with an areal increase due to effective protection and wetland restoration efforts
of 11.9 km2. (Figure 5). On the one hand, improved hydrology caused
by increased precipitation or melting glaciers can facilitate
4. Discussion the formation of new wetlands. On the other hand, wetland
restoration from cropland was enhanced especially in the
In this study, we compared wetland areal changes within and protected areas from the National Wetland Conservation
adjacent to Ramsar sites in China through Landsat observa- Project [13]. To date, China has established 602 wetland pro-
tions and highlighted diverse anthropogenic threats to wet- tected areas and 1,699 wetland parks. Of these, 898 parks and
lands in these sites. China is to host the 15th UN more than 100 protected areas were designated at the
Conference of the Parities to the Convention on Biological national level (http://www.forestry.gov.cn). China’s pro-
Diversity and 14th Meeting of the Conference of the Con- tected areas for wetlands have been and will be increased
tracting Parties to the Ramsar Convention on Wetlands in continuously [9, 35]. Although large areas of wetlands have
2021. The findings will be beneficial to understanding the been protected, the effectiveness of these reserves or parks
achievements and challenges in conserving China’s wetlands. could be greatly improved through careful management
Considering the striking loss of natural wetlands across the and in particular through stopping expansion of anthropo-
whole country [10], the Ramsar sites have contributed sub- genic land covers and wetland conversion [17]. Anthropo-
stantially to protecting natural wetlands and halting the genic land covers, especially illegal agricultural cultivation,
downward trend of wetland area. However, as revealed in excessive tourism development, and industrial footprints,
our analysis, although wetland conservation has expanded should be rehabilitated into natural ecosystems as much as
significantly under the Ramsar Convention on Wetlands, possible to reduce the direct disturbance to wetlands [10,
anthropogenic impacts, including agricultural reclamation, 25]. Ecological migration, compensation mechanism, and
urbanization, industrialization, and aquaculture and tourism alternative livelihoods for native peoples in the Ramsar sites
development are pervasive in many Ramsar sites [26–28]. or national nature reserves should be enhanced for imple-
Anthropogenic threats to wetlands are also intensified in sev- menting returning cropland and aquaculture pond to natural
eral of China’s Ramsar sites (Figures 6 and 7). According to wetlands [36]. The utilization of wetland resources, e.g., tour-
the report of the Second National Wetland Inventory ism, should be controlled, even prohibited when necessary, in
(2009–2013) of China, the major threats to wetlands are pol- the core area of Ramsar sites by legislation [37]. Ecological
lution, agricultural encroachment, built-up land expansion, water supplement from reservoirs or rivers to wetlands with
overfishing and harvesting, and exotic species invasion obvious degradation, especially the wetlands in arid and
(http://www.forestry.gov.cn). Almost all these threat types semiarid zones, is necessary for maintaining vulnerable eco-
could be detected in the Chinese Ramsar estate. Although systems and biodiversity [38]. More scientific evidence is
pollution, overfishing, and harvesting were not directly mon- needed to evaluate methods for the prevention and eradica-
itored using remote sensing in this study, we can infer that tion of exotic S. alterniflora [24]. For example, the Yancheng
nonpoint source pollution related to agricultural and indus- National Nature Reserve in Jiangsu experienced marked nat-
trial activities and aquatic product acquisition from aquacul- ural wetland loss to agricultural cultivation and S. alterniflora
ture pond expansion also imposed considerable threats to invasion before its Ramsar designation in 2002. However,
wetland ecosystems in the Ramsar sites. after its Ramsar designation, the encroachment of cropland
In terms of evident wetland loss and diverse anthropo- and S. alterniflora into tidal flats was well controlled, but
genic threats (Figure 4), protection efficacy for most Ramsar the expansion of aquaculture ponds increased (Tables 2 and
sites in China could be improved. Even after the designation 3). Therefore, place-based policies and responsive decision-
of Ramsar sites, agricultural activities still encroached into making are required for the sustainable management of
large areas of Ramsar wetlands, especially in NEC [18, 29], China’s Ramsar sites, as well as other protected areas.
while infrastructure construction with urbanization and We found that boundaries of the Yancheng and Dafeng
industrialization occurred in areas such as the YAR [30, 31]. national nature reserves overlap (Figure 7). Some national
Moreover, aquaculture development and invasive S. alterni- nature reserve boundaries have been changed due to the
flora had encroached upon large areas of natural wetlands in revaluation of wetlands and biodiversity impacted by
coastal Ramsar sites that are known to be important for migra- regional socioeconomic development need [25, 39]. There-
tory waterbirds [32, 33]. While the preservation of natural fore, managers could reevaluate boundaries of the important
wetlands is of primary importance for migratory waterbirds, reserves and confirm the optimal boundaries using remote
enhanced management of existing human-made wetlands sensing and geographic information system technology. At
including aquaculture ponds and paddy fields can also be ben- present, a new protection system comprising national parks
eficial for biodiversity [34] and could be explored as a priority is being established for protecting China’s natural heritage
within Ramsar sites that already include aquaculture. [5]. Boundaries of wetland protected areas should be care-
Wetlands provide important ecosystem services includ- fully placed to represent biodiversity and include all relevant
ing aquatic food provision, water retention and purification, ecological processes [40, 41]. For example, a national park
climate and flood regulation, carbon sequestration, and bio- was suggested to be established in the Songnen Plain to com-
diversity conservation, and healthy wetlands thus contribute bine the 3 Ramsar sites and more than 10 wetland protected
towards achieving the SDGs [27, 28]. Besides the dramatic areas at different levels. In short, key priorities would appearJournal of Remote Sensing 11
to be (i) improving the effectiveness of existing protected Authors’ Contributions
areas, (ii) expanding protected areas, (iii) upgrading the pro-
tection level at some sites, (iv) optimizing the management D.M. and Z.W. designed the research; D.M. analyzed the data
efforts, and (v) communicating with the public on wetland and wrote the original manuscript; Y.W., C.C., M.J., and R.F.
conservation [17]. provided important comments and revised the manuscript;
The evaluation of China’s Ramsar wetlands is informative and D.M. and M.J. performed the image classification.
for other wetland protected areas and Ramsar wetlands in
other countries by providing methods based on an open data Acknowledgments
source (Landsat). Consistent data sets with long time series,
such as those used in this study, are beneficial to studying We are very grateful to the facility that made the Landsat
ecosystem processes, but accurate observations on diverse images accessible through the USGS (https://earthexplorer
anthropogenic threats to wetlands in Ramsar sites and other .usgs.gov/). We thank Prof. Weihua Xu for the important
protected areas are also important to deliver sustainable suggestions on our original manuscript. This study was
wetland management. High-resolution images could be more jointly supported by the National Key R&D Program of
easily acquired from satellite or airborne platforms [42]. China (grant numbers 2016YFC0500201 and
Therefore, mapping of land covers with high accuracy using 2016YFA0602301), the National Natural Science Foundation
finer-resolution images covering the whole Ramsar sites is of China (grant numbers 41771383 and 41730643), the
necessary. The current system of reporting and monitoring Science and Technology Development Program of Jilin Prov-
the Ramsar wetland status has room for improvement. Besides ince (grant number 20200301014RQ), and the funding from
land cover, remote sensing could characterize other ecosystem the Youth Innovation Promotion Association of Chinese
variables such as the fractional vegetation cover, leaf area Academy of Sciences (grant numbers 2017277 and
index, and vegetation productivity to assess ecosystem quality 2012178) and the National Earth System Science Data Center
[38, 43]. The assessments from perspectives of ecosystem (http://www.geodata.cn).
functions and services also are still necessary for sustainable
wetland conservation and management practice [44–46]. Supplementary Materials
In this study, Landsat series images from 1980 to 2018
were used to examine wetland extents, temporal changes, Table S1: detailed information for Ramsar sites in China. Fig-
and anthropogenic threats in China’s 57 Ramsar wetland ure S1: spatial pattern of land covers in 2018 over the Ramsar
sites. This study presents the first systematic assessment of sites (three Ramsar sites in shallow marine water were not
the effectiveness of conserving Ramsar sites in China on a shown here). Figure S2: various changes in direct anthropo-
national scale and provides a framework to inform sustain- genic threat from 1980 to 2018 over different Ramsar sites
able wetland management and policy improvement in other (time points are 1980, 1990, 2000, 2010, and 2018); red color
countries or even for global wetlands. Our results reveal that line denotes consistent increase trend, blue color line denotes
although Ramsar designation played important roles in halt- an initial decrease followed by an increase, and green color
ing wetland loss, natural wetland loss caused by anthropo- line denotes an initial increase followed by a decrease.
genic encroachment remain striking in many Ramsar sites. (Supplementary Materials)
Agricultural activities are the dominated anthropogenic
threats to Ramsar wetlands in China, while built-up land References
expansion, coastal aquaculture development, and exotic
species invasion are key threats in most coastal Ramsar [1] N. C. Davidson and C. M. Finlayson, “Extent, regional distri-
wetlands. Protective efficacy for most of Ramsar wetlands bution and changes in area of different classes of wetland,”
in China thus requires improvements given the evident wet- Marine and Freshwater Research, vol. 69, no. 10, article 1525,
land loss and diverse anthropogenic threats detected in the 2018.
Ramsar sites, despite these wetlands having been designated [2] Millennium Ecosystem Assessment (MEA), Ecosystems and
for their international importance. These observations Human Well-Being: Wetlands and Water Synthesis, World
Resources Institute, Washington, DC, 2005.
suggest that China needs to continue its efforts to control
the loss and degradation of wetland ecosystems to achieve [3] Ramsar Convention on Wetlands (RCW), Global Wetland
Outlook: State of the World’s Wetlands and Their Services to
sustainable development and ecological civilization.
People, Ramsar Convention Secretariat, Gland, Switzerland,
2018.
Data Availability [4] F. Jaramillo, A. Desormeaux, J. Hedlund et al., “Priorities and
interactions of sustainable development goals (SDGs) with
The multitemporal land cover data sets of Ramsar sites in focus on wetlands,” Water, vol. 11, no. 3, p. 619, 2019.
China generated in this paper are available at https://10 [5] W. Xu, S. L. Pimm, A. Du et al., “Transforming protected area
.5281/zenodo.4699319. management in China,” Trends in Ecology & Evolution, vol. 34,
no. 9, pp. 762–766, 2019.
Conflicts of Interest [6] C. Prigent, E. Matthews, F. Aires, and W. B. Rossow, “Remote
sensing of global wetland dynamics with multiple satellite data
The authors declare that there is no conflict of interest sets,” Geophysical Research Letters, vol. 28, no. 24, pp. 4631–
regarding the publication of this article. 4634, 2001.12 Journal of Remote Sensing
[7] D. Mao, Z. Wang, B. Du et al., “National wetland mapping in vations from Landsat OLI images,” Remote Sensing, vol. 10,
China: a new product resulting from object-based and hierar- no. 12, article 1933, 2018.
chical classification of Landsat 8 OLI images,” ISPRS Journal of [24] D. Mao, M. Liu, Z. Wang et al., “Rapid invasion of Spartina
Photogrammetry and Remote Sensing, vol. 164, pp. 11–25, alterniflora in the coastal zone of mainland China: spatiotem-
2020. poral patterns and human prevention,” Sensors, vol. 19, no. 10,
[8] Z. Niu, H. Zhang, and P. Gong, “More protection for China's article 2308, 2019.
wetlands,” Nature, vol. 471, no. 7338, p. 305, 2011. [25] W. Xu, X. Li, S. L. Pimm et al., “The effectiveness of the zoning
[9] Z. Wang, J. Wu, M. Madden, and D. Mao, “China’s wetlands: of China's protected areas,” Biological Conservation, vol. 204,
conservation plans and policy impacts,” Ambio, vol. 41, no. 7, pp. 231–236, 2016.
pp. 782–786, 2012. [26] C. M. Finlayson, “Forty years of wetland conservation and wise
[10] D. Mao, Z. Wang, J. Wu et al., “China’s wetlands loss to urban use,” Aquatic Conservation: Marine and Freshwater Ecosys-
expansion,” Land Degradation & Development, vol. 29, no. 8, tems, vol. 22, no. 2, pp. 139–143, 2012.
pp. 2644–2657, 2018. [27] M. Hettiarachchi, T. H. Morrison, and C. McAlpine, “Forty-
[11] P. Gong, Z. Niu, X. Cheng et al., “China’s wetland change three years of Ramsar and urban wetlands,” Global Environ-
(1990-2000) determined by remote sensing,” Science China: mental Change, vol. 32, pp. 57–66, 2015.
Earth Science, vol. 53, no. 7, pp. 1036–1042, 2010. [28] M. Jayanthi, S. Thirumurthy, M. Muralidhar, and
[12] Z. Niu, H. Zhang, X. Wang et al., “Mapping wetland changes in P. Ravichandran, “Impact of shrimp aquaculture development
China between 1978 and 2008,” Chinese Science Bulletin, on important ecosystems in India,” Global Environmental
vol. 57, no. 22, pp. 2813–2823, 2012. Change, vol. 52, pp. 10–21, 2018.
[13] D. Mao, L. Luo, Z. Wang et al., “Conversions between natural [29] C. Lu, Z. Wang, L. Li et al., “Assessing the conservation effec-
wetlands and farmland in China: a multiscale geospatial anal- tiveness of wetland protected areas in Northeast China,” Wet-
ysis,” Science of the Total Environment, vol. 634, pp. 550–560, lands Ecology and Management, vol. 24, no. 4, pp. 381–398,
2018. 2016.
[14] E. Maltby, “Wetland management goals: wise use and conser- [30] Y. Chen, J. Dong, X. Xiao et al., “Effects of reclamation and
vation,” Landscape and Urban Planning, vol. 20, no. 1-3, pp. 9– natural changes on coastal wetlands bordering China’s Yellow
18, 1991. Sea from 1984 to 2015,” Land Degradation & Development,
[15] E. Fitoka, M. Tompoulidou, L. Hatziiordanou et al., “Water- vol. 30, no. 13, pp. 1533–1544, 2019.
related ecosystems' mapping and assessment based on remote [31] Y. Li, Y. Shi, X. Zhu, H. Cao, and T. Yu, “Coastal wetland loss
sensing techniques and geospatial analysis: the SWOS and environmental change due to rapid urban expansion in
national service case of the Greek Ramsar sites and their Lianyungang, Jiangsu, China,” Regional Environmental
catchments,” Remote Sensing of Environment, vol. 245, arti- Change, vol. 14, no. 3, pp. 1175–1188, 2014.
cle 111795, 2020. [32] X. Gan, Y. Cai, Z. Choi, Z. Ma, J. Chen, and B. Li, “Potential
[16] K. C. Seto and M. Fragkias, “Mangrove conversion and aqua- impacts of invasive Spartina alterniflora on spring bird com-
culture development in Vietnam: a remote sensing- based munities at Chongming Dongtan, a Chinese wetland of inter-
approach for evaluating the Ramsar Convention on Wet- national importance,” Estuarine Coastal and Shelf Science,
lands,” Global Environmental Change, vol. 17, no. 3-4, vol. 83, no. 2, pp. 211–218, 2009.
pp. 486–500, 2007. [33] C. E. Studds, B. E. Kendall, N. J. Murray et al., “Rapid popula-
[17] Y. Zheng, H. Zhang, Z. Niu, and P. Gong, “Protection efficacy tion decline in migratory shorebirds relying on Yellow Sea
of national wetland reserves in China,” Chinese Science Bulle- tidal mudflats as stopover sites,” Nature Communications,
tin, vol. 57, no. 10, pp. 1116–1134, 2012. vol. 8, no. 1, article 14895, 2017.
[18] D. Mao, X. He, Z. Wang et al., “Diverse policies leading to con- [34] M. V. Jackson, L. R. Carrasco, C. Y. Choi et al., “Multiple hab-
trasting impacts on land cover and ecosystem services in itat use by declining migratory birds necessitates joined-up
Northeast China,” Journal of Cleaner Production, vol. 240, arti- conservation,” Ecology & Evolution, vol. 9, no. 5, pp. 2505–
cle 117961, 2019. 2515, 2019.
[19] M. Jia, Z. Wang, Y. Zhang, D. Mao, and C. Wang, “Monitoring [35] Z. Ma, Y. Chen, D. S. Melville et al., “Changes in area and
loss and recovery of mangrove forests during 42 years: the number of nature reserves in China,” Conservation Biology,
achievements of mangrove conservation in China,” Interna- vol. 33, no. 5, pp. 1066–1075, 2019.
tional Journal of Applied Earth Observation and Geoinforma- [36] H. Xiang, Z. Wang, D. Mao et al., “What did China's national
tion, vol. 73, pp. 535–545, 2018. wetland conservation program achieve? Observations of
[20] R. Ma, H. Duan, C. Hu et al., “A half-century of changes in changes in land cover and ecosystem services in the Sanjiang
China's lakes: global warming or human influence?,” Geophys- Plain,” Journal of Environmental Management, vol. 267, article
ical Research Letters, vol. 37, no. 24, pp. 283–289, 2010. 110623, 2020.
[21] K. Song, Y. Shang, Z. Wen et al., “Characterization of CDOM [37] Y. Dai, L. Feng, X. Hou et al., “Policy-driven changes in enclo-
in saline and freshwater lakes across China using spectroscopic sure fisheries of large lakes in the Yangtze Plain: evidence from
analysis,” Water Research, vol. 150, pp. 403–417, 2019. satellite imagery,” Science of the Total Environment, vol. 688,
[22] C. Ren, Z. Wang, Y. Zhang et al., “Rapid expansion of coastal pp. 1286–1297, 2019.
aquaculture ponds in China from Landsat observations during [38] D. Mao, Z. Wang, B. Wu, Y. Zeng, L. Luo, and B. Zhang, “Land
1984-2016,” International Journal of Applied Earth Observa- degradation and restoration in the arid and semiarid zones of
tion and Geoinformation, vol. 82, article 101902, 2019. China: quantified evidence and implications from satellites,”
[23] M. Liu, D. Mao, Z. Wang et al., “Rapid invasion of Spartina Land Degradation & Development, vol. 29, no. 11, pp. 3841–
alterniflora in the coastal zone of mainland China: new obser- 3851, 2018.Journal of Remote Sensing 13
[39] R. Wu, S. Zhang, D. Yu et al., “Effectiveness of China’s nature
reserves in representing ecological diversity,” Frontiers in Ecol-
ogy and the Environment, vol. 9, no. 7, pp. 383–389, 2011.
[40] C. Y. Choi, H. Peng, P. He et al., “Where to draw the line?
Using movement data to inform protected area design and
conserve mobile species,” Biological Conservation, vol. 234,
pp. 64–71, 2019.
[41] R. A. Fuller, E. McDonald-Madden, K. A. Wilson et al.,
“Replacing underperforming protected areas achieves better
conservation outcomes,” Nature, vol. 466, no. 7304, pp. 365–
367, 2010.
[42] M. Jia, Z. Wang, C. Wang, D. Mao, and Y. Zhang, “A new veg-
etation index to detect periodically submerged mangrove for-
est using single-tide sentinel-2 imagery,” Remote Sensing,
vol. 11, no. 17, article 2043, 2019.
[43] H. Nagendra, R. Lucas, J. P. Honrado et al., “Remote sensing
for conservation monitoring: assessing protected areas, habitat
extent, habitat condition, species diversity, and threats,” Eco-
logical Indicators, vol. 33, pp. 45–59, 2013.
[44] R. J. McInnes, M. Simpson, B. Lopez, R. Hawkins, and
R. Shore, “Wetland ecosystem services and the Ramsar Con-
vention: an assessment of needs,” Wetlands, vol. 37, no. 1,
pp. 123–134, 2017.
[45] L. F. Ricaurte, M. H. Olaya-Rodríguez, J. Cepeda-Valencia
et al., “Future impacts of drivers of change on wetland ecosys-
tem services in Colombia,” Global Environmental Change,
vol. 44, pp. 158–169, 2017.
[46] W. Xu, Y. Xiao, J. Zhang et al., “Strengthening protected areas
for biodiversity and ecosystem services in China,” Proceedings
of the National Academy of Sciences, vol. 114, no. 7, pp. 1601–
1606, 2017.You can also read
