ArticlePDF Available

Detailed wetland-type classification using Landsat-8 time-series images: a pixel- and object-based algorithm with knowledge (POK)

Taylor & Francis
GIScience & Remote Sensing
Authors:

Abstract and Figures

Mapping detailed wetland types can offer useful information for wetland management and protection, which can strongly support the Global Biodiversity Framework. Many studies have conducted wetland classification at regional, national, and global scale, whereas fine-resolution wetland mapping with detailed wetland types is still challenging. To address this issue, we developed an integration of pixel- and object-based algorithms with knowledge (POK) by combining pixel-based random forest and an object-based hierarchical decision tree. Taking the Guangxi Beibu Gulf Economic Zone (GBGEZ) and Guangdong-Hong Kong-Macao Greater Bay Area (GBA) as our study areas, we produced wetland maps with 10 wetland types and 6 non-wetland types using Landsat-8 time series. In addition, to comprehensively evaluate the accuracy of our wetland classification, we implemented accuracy validation based on test samples and data inter-comparison based on existing datasets, respectively. The results indicate that the overall accuracy of our wetland map was 91.6% ±1.2%. For wetland types, agricultural pond, coastal shallow water, floodplain, mangrove, reservoir, river, and tidal flat achieved good accuracies, with both user accuracy and producer accuracy exceeding 88.0%. For non-wetland types, most accuracies were greater than 72.0%. By comparison with existing datasets, it was found that our wetland map had good consistencies with the China Ecosystem-type Classification Dataset (CECD) land use dataset, MC_LASAC mangrove dataset, and Tidal Wetlands in East Asia (TWEA) tidal flat dataset. In 2020, the wetland area was 4,198.8 km² in the GBGEZ and 10,932.2 km² in the GBA. The main wetland types in the two coastal urban agglomerations were agricultural ponds, coastal shallow waters, mangroves, reservoirs, rivers, and tidal flats. Our study successfully mapped detailed wetland types in the GBGEZ and GBA, serving the Global Biodiversity Framework of Convention on Biological Diversity.
This content is subject to copyright. Terms and conditions apply.
Detailed wetland-type classication using Landsat-8 time-series images: a pixel-
and object-based algorithm with knowledge (POK)
Kaifeng Peng
a
, Weiguo Jiang
b
, Peng Hou
c
, Zhifeng Wu
d
and Tiejun Cui
a
a
School of Geographic and Environmental Sciences, Tianjin Normal University, Tianjin, China;
b
State Key Laboratory of Remote Sensing
Science, Faculty of Geographical Science, Beijing Normal University, Beijing, China;
c
Satellite Environment Centre, Ministry of Ecology and
Environment, Beijing, China;
d
School of Geography and Remote Sensing, Guangzhou University, Guangzhou, China
ABSTRACT
Mapping detailed wetland types can oer useful information for wetland management and
protection, which can strongly support the Global Biodiversity Framework. Many studies have
conducted wetland classication at regional, national, and global scale, whereas ne-resolution
wetland mapping with detailed wetland types is still challenging. To address this issue, we
developed an integration of pixel- and object-based algorithms with knowledge (POK) by combin-
ing pixel-based random forest and an object-based hierarchical decision tree. Taking the Guangxi
Beibu Gulf Economic Zone (GBGEZ) and Guangdong-Hong Kong-Macao Greater Bay Area (GBA) as
our study areas, we produced wetland maps with 10 wetland types and 6 non-wetland types using
Landsat-8 time series. In addition, to comprehensively evaluate the accuracy of our wetland
classication, we implemented accuracy validation based on test samples and data inter-
comparison based on existing datasets, respectively. The results indicate that the overall accuracy
of our wetland map was 91.6% ±1.2%. For wetland types, agricultural pond, coastal shallow water,
oodplain, mangrove, reservoir, river, and tidal at achieved good accuracies, with both user
accuracy and producer accuracy exceeding 88.0%. For non-wetland types, most accuracies were
greater than 72.0%. By comparison with existing datasets, it was found that our wetland map had
good consistencies with the China Ecosystem-type Classication Dataset (CECD) land use dataset,
MC_LASAC mangrove dataset, and Tidal Wetlands in East Asia (TWEA) tidal at dataset. In 2020, the
wetland area was 4,198.8 km
2
in the GBGEZ and 10,932.2 km
2
in the GBA. The main wetland types
in the two coastal urban agglomerations were agricultural ponds, coastal shallow waters, man-
groves, reservoirs, rivers, and tidal ats. Our study successfully mapped detailed wetland types in
the GBGEZ and GBA, serving the Global Biodiversity Framework of Convention on Biological
Diversity.
ARTICLE HISTORY
Received 24 July 2023
Accepted 4 December 2023
KEYWORDS
Global biodiversity
framework;
Guangdong-Hong
Kong-Macao Greater Bay
Area; Guangxi beibu gulf
economic zone; remote
sensing classification;
Landsat-8 time series;
wetlands
1. Introduction
Wetlands are one of the three major ecosystems
formed by complex hydrological processes, and they
oer hydrological functions such as ood and
drought prevention, groundwater regulation, and
shoreline protection (Keith et al. 2023; Stovall et al.
2019; Xu et al. 2020). Due to their ecological impor-
tance, water-related ecosystems or wetlands are
valued by Convention on Biological Diversity and are
included in Global Biodiversity Framework, especially
in the target 1, 2, and 3 (Bon et al. 2022; Joly 2023).
However, due to climate change and human activ-
ities, global wetlands have experienced serious degra-
dation, as have wetlands in China (Mao et al. 2022).
From 1978 to 2008, approximately 33% of wetlands in
China disappeared or were degraded (Niu et al. 2012).
Therefore, it is necessary to monitor the extent of
wetlands to inform their protection and sustain-
able use.
Remote sensing is an eective technology for map-
ping the extent of wetlands. To date, many wetland
mapping datasets are available, such as water inunda-
tion (Pekel et al. 2016), mangroves (Jia et al. 2023),
tidal ats (Murray et al. 2022), other single wetland
types (e.g. salt marshes and agricultural ponds) (Hou
et al. 2022; Zhao et al. 2023), Land use/cover change
(LUCC) (Yang and Huang 2021), and multiple-wetland
type datasets (Liu et al. 2022). The single wetland
datasets just delineate the extent of the specic wet-
land, which cannot reect all wetland ecosystem
types’ distribution. The LUCC datasets are important
data source for delineating the extent of wetlands
CONTACT Weiguo Jiang jiangweiguo@bnu.edu.cn
GISCIENCE & REMOTE SENSING
2024, VOL. 61, NO. 1, 2293525
https://doi.org/10.1080/15481603.2023.2293525
© 2023 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use,
distribution, and reproduction in any medium, provided the original work is properly cited. The terms on which this article has been published allow the posting of the Accepted
Manuscript in a repository by the author(s) or with their consent.
(Elmahdy and Mohamed 2018; Owers et al. 2022).
However, LUCC datasets cannot reect the actual
status of detailed wetland types, and their wetland
type accuracy is low (Xu et al. 2022), for example,
Zhang et al. (2021) produced global land use maps
with wetland’s user accuracy and producer accuracy
of 43.4% and 61.8%, respectively. Gong et al. (2019)
mapped global land use patterns with wetland’s user
accuracy and producer accuracy of 34.48% and
10.14%, respectively.
For the multiple wetland dataset mapping, many
scholars have conducted studies on it. At the global
scale, the Global Lakes and Wetland Database (GLWD)
was the most widely used dataset that delineated 12
wetland types at a ~1 km spatial resolution (Lehner
and Doll 2004). However, this dataset delineated wet-
land extents in the 1980s by combining multiple exist-
ing wetland datasets, which is outdated for reecting
current wetland conditions in China. At the national
scale of China, there have been several detailed-type
wetland datasets, such as CAS_Wetlands (Mao et al.
2020) and CLUD (China Land-use/cover datasets)
(Kuang et al. 2022), which could used to support
Global Biodiversity Framework. For the
CAS_Wetlands, it mapped 14 wetland types at 30m
resolution with overall accuracy of 95.1%. However,
this dataset just delineated wetlands’ extent of 2015,
which cannot reect the wetland dynamics. For the
CLUD, it delineated 7 wetland types and 19 non-
wetland types at 30 m spatial resolution with overall
accuracy over 90%. But its produce cost was high,
because it was generated by visual interpretation.
Therefore, it is necessary to produce long-term wetland
maps at low cost for the detailed wetland categories.
Optical satellite images, such as Landsat, Sentinel-
2, and MODIS, are important data sources that can be
freely acquired for wetland mapping, which can oer
rich spectral information on dierent land cover types
(Chen et al. 2022; Hu et al. 2021; Jamali et al. 2021).
Among them, Landsat archives are widely used in
land cover or wetland classication because of their
long-term coverage and median spatial resolution
(Mao et al. 2020). However, as with other optical
images, Landsat images often suer from cloud
cover and shadow contamination, especially in
coastal regions. This phenomenon results in the
waste of a certain number of images and greatly limits
the use of the Landsat constellation. To overcome this
problem, some scholars have used Landsat time series
to classify wetlands, and the advent of Google Earth
Engine (GEE) has made it easy to realize (Li et al. 2022;
Navarro et al. 2021; Wang et al. 2023). Bad-quality
pixels of time-series images are masked out, and the
remaining good observations are utilized to map wet-
land types. To make full use of the dense time series of
Landsat images, scholars have attempted to explore
phenology-based features for wetland classication
(Murray et al. 2019; Wu et al. 2021). Compared with
simple features (e.g. single-date image, median fea-
tures), phenology-based features could oer more
useful information and reduce redundant information
of time-series images (Ni et al. 2021). Therefore, based
on Landsat time series, compositing suitable phenol-
ogy-based features of wetland types is promising for
wetland classication.
In recent decades, several classication algorithms
have been developed for wetland mapping.
Supervised classication methods, such as maximum
likelihood classication (MLC), support vector
machine (SVM), and random forest (RF), are com-
monly used for wetland classication due to their
good classication accuracy and high computation
eciency (Bhatt and Maclean 2023; Rodriguez-
Galiano and Chica-Rivas 2014). These methods are
primarily implemented at the pixel level and are sui-
table for broad wetland extractions. However, they
cannot well extract detailed wetland types based on
spectral features (Li and Niu 2022). Meanwhile, the
spectral indexes and index rules in combination with
x or dynamic thresholds were also commented used
for wetland mapping, such as waterbodies (Deng
et al. 2022), mangroves (Xia et al. 2022), and tidal at
(Zhang et al. 2022). However, these methods were
mainly used to map single wetland types, while it
was dicult to extract detailed wetland types. The
object-based hierarchical decision tree was another
common method for wetland mapping, which can
eectively extract detailed wetland types (Mao et al.
2020). Nevertheless, the construction of this method
is complex, and its thresholds and rule sets in decision
trees often change over time and regions (Fitoka et al.
2020). The complexity of this algorithm limits its wide-
spread use in large-scale wetland mapping. Thus,
detailed wetland type classication on a large scale
is challenging, and an accurate and robust wetland
mapping method is urgently needed.
Located in the southern coastal region of China,
the Guangxi Beibu Gulf Economic Zone (GBGEZ) and
2K. PENG ET AL.
the Guangdong-Hong Kong-Macao Greater Bay Area
(GBA) are two important coastal urban agglomera-
tions. The two regions are geographically adjacent
and have rich wetland resources (Cao et al. 2019; Wu
et al. 2020). In recent decades, these two urban
agglomerations have experienced rapid economic
development, especially in the GBA (Zhang et al.
2023). Undoubtedly, increasing human disturbance
has resulted in area losses of natural wetlands (Guo
et al. 2021). To clearly recognize the wetland extents
in the two regions, it is necessary to monitor the
wetland distribution of detailed wetland types.
Taking the GBGEZ and GBA as our study areas, this
study aimed to conduct a detailed wetland type clas-
sication in 2020 based on Landsat-8 time-series
images. The objectives of our study are as follows:
Develop an accurate and robust algorithm for
detailed wetland type classication by combin-
ing pixel-based RF algorithm and object-based
hierarchical decision tree.
Produce a wetland map with 10 wetland types
and 6 non-wetland types based on Landsat-8
image time series and evaluate its classication
accuracy based on test samples.
Implemented data inter-comparison with exist-
ing datasets to further check the accuracy of our
wetland classication.
2. Materials and methods
2.1. Study area
In our study, two coastal urban agglomerations were
selected: the Guangxi Beibu Gulf Economic Zone
(GBGEZ) and the Guangdong-Hong Kong-Macao
Greater Bay Area (GBA). The GBGEZ covers the latitude
20°40’and 24°2’ N, and the longitude 106°33’and 110°
53’ E. It includes six cities located in the Guangxi
Zhuang Autonomous Region (Figure 1). The GBA cov-
ers the latitude 21°25’and 24°23’ N, and the longitude
111°21’and 115°25’ E. It includes 11 cities, with 9 cities
in Guangdong Province, Hong Kong, and Macao. The
two urban agglomerations are geographically adja-
cent, and both are in the coastal area of southern
China. In addition, considering that some coasta
Figure 1. The location of our study area. In the figure (c-d), the black polygons denote the administrative boundary, and the blue
polygons denote the coastal expansion areas.
GISCIENCE & REMOTE SENSING 3
l wetlands are distributed outside the coastline, we
expanded the study area. According to Ramsar
Convention, inland and coastal wetlands are distrib-
uted in areas with water depth less than 6 m (Gell,
Finlayson, and Davidson 2023). However, due to tidal
variation and data errors of water depth, some coastal
wetlands (e.g. mangrove, tidal at) may occur in low
tidal level (Wang et al. 2022). Thus, some studies set
the threshold of 25 m by trial and error, which could
cover all possible wetland extent (Zhang et al. 2022;
Zhao and Qin 2020). In this study, we dened the area
outside the coastline with a water depth of less than
25 m as the coastal expansion area and implemented
wetland classication in both administrative and
coastal expansion areas (Figure 1).
2.2. Classication system of wetland ecosystem
Considering the separability of wetland ecosystem
types in satellite images, we designed a wetland clas-
sication system that is suitable for our study areas
based on previous studies (Keith et al. 2023; Liu et al.
2014; Mao et al. 2020). Our classication system
included ve inland wetlands (inland swamp, marsh,
river, lake, and oodplain), three coastal wetlands
(mangrove, tidal at, and coastal shallow water), two
human-made wetlands (reservoirs and agricultural
ponds). Meanwhile, for better conducting classica-
tion, we added six non-wetland types (forest, grass-
land, built-up land, paddy eld, dryland, and bare
land) in our classication system. The detailed
description of our classication was shown in Table 1.
2.3. Data sources
2.3.1. Landsat-8 image selection and processing
Landsat-8 images were available in the GEE platform.
The surface reectance of Landsat-8 satellite was used
in our study. All Landsat-8 images from June 2019 to
June 2021 were selected for wetland classication in
2020, with 660 images. To remove cloud contamina-
tion in each image, we implemented a mask opera-
tion based on the “QA_PIXEL” band of Landsat-8,
which could remove cloudy pixels. The remaining
pixels of image are good-quality observations. The
spatial distributions of good-quality observations are
shown in Figure 2(a,b). The pixel values in the two
images Figure 2(a,b) represent the number of good
observations of Landsat images in corresponding pix-
el’s location. It was found that good pixels without
cloud contaminations could completely covered the
study area. We also found that in both GBGEZ and
Table 1. Classification system of wetland ecosystem.
Category I Category II Description
sample
size
Coastal wetland Coastal shallow water Coastal water with depth less than 6 m, including coastal shallow water, bays, straits, etc. 1099
Mangrove Woody plant community that grow in coastal tidal flat, distributed in tropical and subtropical
climates
533
Tidal flat Intertidal zone between high-level tide and low-level tide in coastal areas, including intertidal
mudflats, rocks, sands, etc.
728
Human-made
wetland
Agricultural pond Small patchy waterbody with regular shape that used for aquaculture, mostly distributed in coastal
areas and nearby rivers.
1695
Reservoir Patchy waterbody that artificially constructed for flood regulation and power generation, often
equipped with dams.
1049
Inland wetland Floodplain Low vegetated areas inundated with seasonal water, including river flooded areas, lake beach,
alluvial fan, etc.
258
Inland swamp Vegetated wetlands dominated with woody vegetation in inland areas, including forest swamp and
shrub swamp
53
Marsh Vegetated wetland dominated with herbaceous vegetation 143
River Linear waterbody inundated with perennial water 993
Lake Nature patchy waterbody inundated with perennial water 88
Non-wetland Bare land Areas with low vegetation coverage, including soil, desert, rock, etc. 300
Built-up land Artificial construction facilities, including urban land, residential areas, roads, water conservancy
facilities, etc.
844
Dryland Cropland that relies on natural precipitation, or can be irrigated normally in general years, used for
corn, wheat, soybean, etc.
1868
Forest Woody vegetation with vegetation coverage greater than 20%, including forest and shrub 3903
Grassland Herbaceous vegetation with a coverage of more than 5%, including shrub grassland and sparse
forest grassland.
566
Paddy field Cropland used for rice, lotus root and other aquatic crops has irrigation facilities, and can be irrigated
normally in general years
834
The GBGEZ and GBA located in Southern China, with low altitude. As report by previous studies (Mao et al. 2020; Zhang et al. 2022), there are few coastal marsh
(i.e. salt marsh) in these two regions, which is also consistent with our study result. Thus, we do not divide the marsh into inland or coastal.
4K. PENG ET AL.
GBA, more than 99.9% of the pixels had more than
ve good-quality observations, and more than 89% of
the pixels had more than 10 good-quality observa-
tions Figure 2(c,d).
2.3.2. Auxiliary data
In our study, we collected a series of auxiliary datasets
that could be divided into six groups: water, tidal at,
mangrove, Land Use/cover Change (LUCC), topo-
graphic, and other auxiliary datasets. Detailed
descriptions of the datasets are provided in Table 2.
These datasets were used for sample production,
classication assistance, and data inter-comparison.
For the LUCC dataset, it includes China Ecosystem-
type Classication Dataset (CECD) and China Land
Cover Dataset (CLCD). The CECD includes 7 wetland
sub-types and 14 non-wetland broad types (Table A1),
which was produced by visual interpretation based
on multiple satellite images (e.g, Gaofen-1, Gaofen-2,
Ziyuan-3) in 2020, with overall accuracy over 90%. The
CLCD includes nine broad land use types, of which the
forest, water, and impervious have high accuracy. For
the POI, GOODD and OSM_Dam points, they were
used to assist in extracting reservoirs. In addition, it
should be noted that the four mangrove datasets for
all periods were merged into a potential mangrove
extent (PME) to assist mangrove extraction.
3. Methods
This study developed a new method for detailed
wetland type classication in two coastal urban
agglomerations, and the process can be divided
into three parts: (1) image selection and sample
production, (2) detailed classication of wetland
types using the POK algorithm, and (3) data inter-
0
4
8
12
16
20
24
28
32
36
5 10 15 20 25 30 35 40 45 >45
Number of pixel (%)
Good-quality observations
(c) GBGEZ-Histogram of
good-quality observations
0
4
8
12
16
20
24
28
32
36
5 10 15 20 25 30 35 40 45 >45
Number of pixel (%)
Good-quality observations
(d) GBA-Histogram of
good-quality observations
Figure 2. Image statistics of Landsat-8 time series from June 2019 to June 2021 in our study area. (a-b) are the numbers of spatial
distributions of good-quality observation (pixel without could cover) in the GBGEZ and GBA, respectively. (c-d) are histograms of
good-quality observations in the GBGEZ and GBA, respectively.
GISCIENCE & REMOTE SENSING 5
comparison based on our test samples and existing
datasets. The workow of this study is shown in
Figure 3.
3.1. Training sample generation
Based on auxiliary datasets, we produced training
samples by combining rule ltering and visual inter-
pretation. Our sample generation method was
divided into two parts (Table 3): automatic and semi-
automatic sample production.
First, we directly produced samples based on aux-
iliary datasets. For rivers, lakes, and reservoirs, we
used GSW, CECD, and HydroLAKES to produce train-
ing samples. Based on the GSW data from 1984 to
2020, water points were randomly generated within
a water inundation frequency (WIF) larger than 80%.
The river and lake samples were then selected using
the river and lake extent of CECD dataset. The reser-
voir samples and some of the lake samples were
selected using the reservoir and lake extent of
HydroLAKE. For mangroves and tidal ats, we
used the GMW and Global Intertidal Change data
from 2016 to produce their samples. For forest
and built-up land, we used the CLCD from 2015 to
2019 to produce random samples. The two land use
types in CLCD shared high map accuracy, with F-score
over 72%, which could be used to produce reliable
samples. To improve sample’s quality, we overlaid the
multiyear CLCD to generate stable forest and built-up
land regions and then randomly created samples
within the intersected regions.
In the second part, we rst produced potential
samples based on an auxiliary dataset and then
visually investigated and corrected them based on
very high-resolution (VHR) images of Google Earth
platform in 2020. Referring to (Peng et al. 2021,
2023), we produced sample points of oodplain,
inland swamp, and marsh by combining GSW,
MODIS NDVI time series, and visual interpretation.
For agricultural pond, paddy eld, dryland, grassland,
and bare land classes, we produced samples using the
CECD of 2020. We rstly extract their extents using the
CECD dataset, and then randomly generated sample
points in the corresponding type’s boundary in ArcGIS
10.4 software. Meanwhile, considering the inherent
errors of CECD, we manually inspected these samples
by visual interpretation, and corrected their attributes
if their sample labels were wrong. Finally, through the
above process, we produced 14,954 training samples.
The sample size of each type was shown in Table 1.
The spatial distributions of the training samples are
shown in Figure 4.
3.2. An algorithm of pixel- and object-based with
knowledge (POK)
By combining pixel-based random forest and an
object-based hierarchical decision tree, we devel-
oped a pixel- and object-based algorithm with
knowledge (POK). The “pixel” refers to the pixel-
based random forest classication (Deng et al.
2023), the “object” refers to polygons with geometric
properties segmented by chessboard segmentation,
and the “knowledge” refers to the rule set of the
Table 2. Auxiliary dataset list in our study.
Category Name Spatial resolution
Time
period Data source
Elevation dataset ETOPO1 1 arc-minute Amante and Eakins (2009)
Shuttle Radar Topography Mission (SRTM) 30 m 2000 Farr et al. (2007)
LUCC dataset China Ecosystem-type Classification Dataset (CECD) 5 m 2020 Ouyang et al. (2016)
China Land Cover Dataset (CLCD) 30 m 1985–2019 Yang and Huang (2021)
Mangrove dataset Global Mangrove Distribution 30 m 2000 Giri et al. (2011)
Global Mangrove Watch (GMW) 30 m 1984–2016 Bunting et al. (2018)
HSL_MangroveChina_LASAC_share 2 m/30 m 1978–2018 Zhang et al. (2020)
Mangrove Map of China for 2019 10 m 2019 Zhao and Qin (2020)
Open water dataset Global Surface Water (GSW) 30 m 1984–2020 Pekel et al. (2016)
HydroLakes 1:100 000 Messager et al. (2016)
Global georeferenced Database of Dams (GOODD) Mulligan, van Soesbergen,
and Saenz (2020)Open Street Map Dam (OSM_Dam)
Other auxiliary dataset Global River Widths from Landsat (GRWL) 30 m Allen and Pavelsky (2018)
Reservoir Point Of Interest (POI) Amap open platform
(https://lbs.amap.com/)
Tidal flat dataset Global Intertidal Change 30 m 1990–2017 Murray et al. (2019)
Tidal Wetlands in East Asia (TEWA) 10 m 2020 Zhang et al. (2022)
The LUCC denote Land Use/cover Change.
6K. PENG ET AL.
hierarchical decision tree. A concept graph of the
POK algorithm is shown in Figure 5.
3.2.1. Input features
To capture the seasonal characteristics of wetlands
and non-wetlands, we used Landsat-8 images from
June 2019 to June 2020 to construct phenology-
based features. In addition to six spectral bands
(green, blue, red, NIR [Near Infrared], SWIR1 [short-
wave Infrared1], and SWIR2 [Shortwave Infrared2]),
ve spectral indices were chosen, namely, the normal-
ized dierence vegetation index (NDVI) (Rouse et al.
1974), enhanced vegetation index (EVI) (Huete et al.
1997), normal dierence water index (NDWI)
(McFeeters 1996), modied normal dierence water
index (MNDWI) (Xu 2006), and automated water
extraction index (AWEI) (Feyisa et al. 2014). Based on
the GEE platform (Gorelick et al. 2017), we composited
all bands and indices of two-year images into phenol-
ogy-based features, including median, wettest, and
greenest images with spectral bands, and percentile,
mean interval, and statistical features of spectral
indexes (Figure 6).
The median image reect the average reectance
during the study period and was composited using
the median() function (Jia et al. 2020). The wettest
Figure 3. Workflow of detailed wetland type classification. SRTM denote Shuttle Radar Topography Mission, CECD denote China
Ecosystem-type classification dataset, MC_LASAC denote HSL_MangroveChina_LASAC_share, TWEA denote tidal wetlands in East
Asia.
GISCIENCE & REMOTE SENSING 7
image represents the maximum extent of perma-
nently inundated water bodies and was composited
using qualityMosaic (“MNDWI”) (Jia et al. 2021). The
greenest image represents pixels with high levels of
greenness across the two-year period and was com-
posited using qualityMosaic (“NDVI”) (Zhang et al.
2022). The percentile features, mean interval features,
and statistical features were composited based on the
time series of spectral indices (AWEI, EVI, NDVI, NDWI,
and MNDWI). Percentile features were calculated at
10%, 25%, 50%, 75%, and 90% using the percentile()
function (Murray et al. 2019). Specically, the histo-
gram of spectral index at a pixel location was calcu-
lated based on their two-years’ time series, and then
the percentile features were extracted at a certain
value (i.e. 10%, 25%, 50%, 75%, 90%) for this pixel
(Xie et al. 2019). The mean interval features were
calculated at intervals of 0–10%, 10–25%, 25–50%,
50–75%, 75–90%, 90–100%, 10–90%, and 25–75%
using the ee.Reducer.intervalMean() function (Wu
et al. 2021). Descriptive statistics were calculated,
including the maximum, minimum, median, and stan-
dard deviation of the time series of spectral indices
and were composited using the max(), min(), median
(), and stdDev() functions, respectively. Terrain fea-
tures, including elevation, slope, and aspect, were
calculated based on the SRTM dataset. Texture fea-
tures were calculated using a gray-level co-
occurrence matrix based on the median composite
NDVI image. Through these operations, we produced
114 feature layers in our phenology-based feature
collection.
In addition, to facilitate detailed wetland type clas-
sication, we chose four geometric features: compact-
ness, rectangular t, length/width, and area. The four
features were calculated based on chessboard seg-
mentation, which was implemented using eCognition
8.7 software (Zheng et al. 2016). It should be noted
that the geometric features were used in object-
Table 3. Illustration of training sample generation based on rule
filtering and visual interpretation.
Category Auxiliary dataset
Rule
filtering
Visual
interpretation
Automatic sample production
Built-up land CLCD
Forest CLCD
Lake CECD, GSW,
HydroLake
Mangrove GMW
Reservoir GSW, HydroLake
River GSW, CECD
Tidal flat Global Intertidal
Change
Semiautomatic sample production
Agricultural
pond
CECD
Bare land CECD
Dryland CECD
Floodplain GSW, MODIS
Grassland CECD
Inland swamp GSW, MODIS
Marsh GSW, MODIS
Paddy field CECD
In the GSW dataset, the occurrence layer that recorded the frequency with
which water was present from 1984 to 2022 was selected. Its pixel values
range from 0 to 100. The MODIS refer to MOD13Q1 product, with NDVI
band of 250 m spatial resolution. The MODIS NDVI from 2000 to 2020,
available in GEE platform, were selected to construct the NDVI time series.
Figure 4. Spatial distribution of training sample points in the GBGEZ (a) and GBA (b).
8K. PENG ET AL.
based hierarchical decision tree, rather than stack
them into phenology-based features for pixel-based
random forest classication.
3.2.2. Pixel-based random forest classification
Based on phenology-based features, we implemented
pixel-based random forest classication using the GEE
platform. Random forest (RF) is an ensemble algorithm
that can avoid overtting and is insensitive to feature
redundancy. Thus, RF is widely used in wetland and
land use mapping (Deng et al. 2023; Mahdianpari et al.
2020; Yang and Huang 2021). The main parameters of
the RF were the numbers of randomly selected features
(mtry) and decision trees (ntree). mtry is usually set as
the square of the total number of input feature layers
(Feng et al. 2022). Thus, the mtry value in our study was
11. Considering the mapping accuracy and computa-
tional eciency, we set ntree to 100. In our study, we
used the pixel-based RF algorithm to initially extract six
wetland types (water, swamp, marsh, oodplain, tidal
at, and agricultural pond) and six non-wetland types
(forest, grassland, built-up land, paddy eld, dryland,
and bare land). Considering the similar climate and
geographical environment, we used all the training
samples of GBGEZ and GBA to train the RF classier
and generate classication maps simultaneously.
Figure 5. Concept graph of the POK algorithm. GEE denotes the Google Earth Engine. PME denote the potential mangrove extent. POK
denotes the pixel- and object-based algorithm with knowledge.
Figure 6. Phenology-based features constructed by the GEE platform.
GISCIENCE & REMOTE SENSING 9
3.2.3. Object-based hierarchical decision tree
classification
To separate the six wetland types of the RF classica-
tion into detailed types, we designed an object-based
hierarchical decision tree by combining geometric fea-
tures and auxiliary datasets. Geometric features (com-
pactness, rectangular t, length/width, and area) of
water bodies were calculated using chessboard seg-
mentation in eCognition software. Firstly, the water-
body raster with binary values (i.e. water [1] and non-
water [0]) was extracted from the classication maps of
RF pixel-based algorithm. Secondly, the waterbody
vectors were derived from the water raster using the
ArcGIS 10.4 software. Thirdly, using the waterbody ras-
ter as the input raster and the waterbody vectors as the
input vector, the chessboard segmentation was used
to calculate compactness, rectangular t, length/width,
and area (Zheng et al. 2016).
Based on the training samples, we counted the geo-
metric features of the reservoirs, rivers, and lakes
(Figure 7). It was found that the river has the largest
compactness and length/width and the smallest rectan-
gular t among the three waterbody types. Based on the
above analysis and our trial and error, we found that
The rule of “Compactness >2.5” extracted 83.0% of riv-
ers, while the extraction results contained 69.1% of
reservoirs and 30.8% of lakes. The rule of
“Rectangular t <0.55” extracted 75.5% of rivers, while
its results contained 25.5% of reservoirs and 3.85% of
lakes. The rule of “Length/width >2.0” extracted
86.1% of rivers, while its results contained 32.7% of
reservoirs and 26.9% of lakes. Thus, we aggregated the
three individual rules to form a consolidated rule set for
river extraction, namely, “Compactness >2.5 &
Rectangular t < 0.55 & Length/width >2.0.” In addition,
we found that the above rule set may underestimate
large rivers, which could be well extracted by the rule of
“Compactness >6.0.” In summary, we used the rule set of
“Compactness >2.5 & Rectangular t <0.55 & Length/
width >2.0” or “Compactness >6.0” to extract the rivers.
Figure 7. Geometric features of reservoirs, rivers and lakes used in the development of the rule set of hierarchical decision. The three
geometric characteristics are shown for three size categories, between 0–2 km
2
, 2–10 km
2
and 10–50 km
2
for Compactness (a-c),
rectangular fit (d-f) and length/width (g-i), respectively.
10 K. PENG ET AL.
For the reservoirs, we used the reservoir POI,
GOODD, and OSM_Dam points to lter them from
the water objects. Reference to the study of Mao
et al. (2020), we extracted the coastal shallow water
using the shoreline and ETOPO1 datasets. Water
bodies located outside the shoreline with a depth of
less than 6 m were identied as coastal shallow
waters. We used the potential mangrove extent
(PME) to extract mangroves, and the remaining
swamp was labeled as inland swamp. The workow
of the hierarchical decision tree is shown in Figure 8.
3.3. Post-classication processing
In this study, additional steps were used in an attempt to
improve the wetland classications of our POK algo-
rithm. First, waterbodies with areas less than 0.01 km
2
were selected and then intersected with classied agri-
cultural ponds with a 30 m buer, which were then
labeled as agricultural ponds. Second, agricultural
ponds with areas less than 0.01 km
2
were selected and
intersected with classied rivers, which were then
labeled as rivers. Third, we used the GRWL dataset to
lter them from unclassied water bodies, and the
selected polygons were labeled rivers. Fourth, we imple-
mented visual interpretation to identify obvious unclas-
sied water bodies. Subsequently, water bodies with
a depth range of 6–25 m were set as the background,
and the remaining water bodies were labeled as unclas-
sied water. Finally, we used the majority of the algo-
rithms of the ENVI software to remove small patches of
marsh and inland swamps. In the majority algorithm, its
kernel size is 5 × 5, and its center pixel weight is 1.0.
3.4. Comparison analysis
3.4.1. Accuracy validation based on test samples
To validate our wetland classication accuracy, we
produced test samples by combining stratied ran-
dom sampling with visual interpretation (Olofsson
Figure 8. Illustration of hierarchical decision tree. The green boxes denote the initial input data, the yellow ovals denote process
operations, the light yellow boxes denote the intermediate output, and the light blue denote the final output. In addition, it should be
noted that the coastal deep water does not belong to wetlands, which was grouped into background.
GISCIENCE & REMOTE SENSING 11
et al. 2014). Specically, we rst generated random
points using stratied random sampling method in
ArcGIS 10.4 software, and then implemented visual
interpretation for these points by combining Google
Earth and Collect Earth platform (Bey et al. 2016). For
the visual interpretation, the Google Earth platform
oers VHR images, and the Collect Earth platform was
used to manage sample points. Finally, 2,034 points
were randomly produced. We then chose the overall
accuracy (OA), user accuracy (UA), and producer accu-
racy (PA) to evaluate the accuracy of our wetland map
(Stehman et al. 2021). The standard errors of the three
indices were also calculated to measure their uncer-
tainties at the 95% condence level. In addition, we
dened accuracy metrics greater than 80%, between
65% and 80%, and less than 65% as high accuracy,
medium accuracy, and low accuracy, respectively.
3.4.2. Intra-comparisons with existing datasets
To further check the accuracy of our wetland maps,
we compared them with existing datasets, including
the CECD, MC_LASAC, and TWEA datasets. We calcu-
lated the areas of individual types in dierent maps
and calculated the areas of spatial consistency by
overlay analysis.
The CECD dataset, acquired from the Ministry of
Ecology and Environment of China, was produced at
a spatial resolution of 5 m by visual interpretation, with
7 wetland types and 14 non-wetland types (Table A1).
This dataset was used to test the overall consistency of
the maps. Because the CECD merged the reservoirs and
agricultural ponds into one type (reservoir/pond), we
merge these two classes of our wetland maps into one
class, and compared it with CECD dataset. Meanwhile,
because the coastal shallow waters and tidal ats of the
CECD were not complete in the GBA but were relatively
complete in the GBGEZ, we only compared the tidal
ats in the GBGEZ. In addition, we aggregated the 19
non-wetland types into 6 types to match our classica-
tion system. The corresponding relationship between
the classes of CECD and our wetland classication was
shown in Table A1.
The MC_LASAC mangrove dataset of 2018, with a 2
m spatial resolution and an OA of 98%, was used to
check the accuracy of our predicted extent of man-
groves. The TWEA tidal at of 2020, at a spatial resolu-
tion of 10 m, with both UA and PA greater than 94%,
was used to test the accuracy of our predicted extent of
tidal ats. In our study, we implemented data inter-
comparison in ArcGIS 10.4 software. All datasets were
projected into WGS_1984_Albers coordinate system.
4. Results
4.1. Classication results of pixel-based RF
algorithm
Based on phenology-based features, we used a pixel-
based RF algorithm to extract six wetland types and
six non-wetland types. Their spatial distributions are
shown in Figure 9. The accuracy validation indicated
that the OA of the wetland map was 91.8%±1.2%
(Table 4). For wetland types, the water and tidal ats
had the highest accuracy, with UA and PA values over
92.2%. The swamp, oodplain, and agricultural ponds
also had high accuracy, with UA and PA values over
88.0%. The marsh had low accuracy with a low PA,
indicating that some omission errors existed. The rea-
son was that the small marshes were spectrally con-
fused with other types, such as dryland, swamp and
paddy eld. For non-wetland types, forest and built-
up land had the highest accuracy, with UA and PA
values over 91.2%. The dryland also achieved high
accuracy, with UA and PA values over 83.3%.
Grassland, paddy eld, and bare land had moderate
accuracy, with most UA and PA values over 72.0%.
4.2. Detailed wetland type classication results
We used an object-based hierarchical decision tree to
further divide the six wetland types into 10 wetland
types (Figure 10). Specically, we divided the water
into rivers, lakes, reservoirs and coastal shallow water,
and divided the swamps into inland swamp and man-
grove. Finally, there were 10 wetland types in our
study. The spatial comparison between before and
after ne classication in two typical regions was
shown in Figure 11.
The accuracy validation indicated that the OA of
our wetland map was 91.6%±1.2% (Table 5). For
inland wetlands, rivers and oodplains had high
accuracy, with UA and PA values over 88.0%. Lakes
had moderate accuracy, with a UA and PA of
>72.2%. The inland swamps and marshes had low
accuracy because they were dicult to extract
owing to their very small areas. For coastal wet-
lands, all three categories achieved good accuracy,
with UA and PA exceeding 90.7%. For human-made
12 K. PENG ET AL.
wetlands, reservoirs had the highest accuracy, with
UA and PA values over 90.7%. Agricultural ponds
also achieved good accuracy, with a UA and PA
of >88.1%.
For non-wetlands, forests and built-up lands had the
highest accuracy, with UA and PA values over 91.3%.
The drylands also achieved high accuracy, with a UA
and PA of over 83.2%. Grasslands, paddy elds, and
bare lands had moderate accuracy, with most accuracy
indices exceeding 0.72. In summary, we believe that
our wetland map achieved a high accuracy.
4.3. Intra-comparison with existing datasets
4.3.1. Comparison with the CECD land use dataset
A spatial comparison between the CECD and our
maps in two large typical regions is shown in
Figure 12. It was found that the land-use types in
the two datasets shared similar spatial distributions.
The agricultural ponds in our wetland map had
a distribution similar to that of the CECD Figure 12
(a-c). The rivers in our wetland map were accurately
extracted and consistent with the CECD dataset
Figure 12(d). For the paddy elds and drylands,
there were some inconsistencies, but the cropland
combined with paddy elds and drylands had good
spatial consistency in the two datasets Figure 12(b).
To quantify the consistency and divergence
between our wetland classications and CECD data-
set, we calculated the confusion matrix between wet-
lands and non-wetland types in GBGEZ (Table 6) and
GBA (Table 7). The results indicated that the spatial
consistency of all land use types accounted for 84.9%
and 74.6% of GBGEZ and GBA, respectively.
For individual wetlands, the river had highest spa-
tial consistency, which accounted for 80.4% and
87.0% of our river in GBGEZ and GBA, respectively.
The tidal at, coastal shallow water, and reservoir/
pond have moderate spatial consistency. In GBGEZ,
the spatial consistency of four wetland types
accounted for 67.4%, 60.2%, and 64.6% of the corre-
sponding types of our wetlands, respectively. In GBA,
the reservoir/pond accounted for 68.7% of that of our
wetlands. For individual non-wetlands, the forest had
the highest spatial consistency, which accounted for
86.4% and 93.7% of our forest in GBGEZ and GBA,
respectively. The built-up land and cropland
Figure 9. Spatial distribution of six wetland types and six non-wetland types.
Table 4. Accuracy results of the pixel-based RF classifications.
OA 91.8%±1.2%
Individual accuracy metric UA PA
Agricultural pond 89.1% ± 8.3% 89.2% ± 7.8%
Bare land 92.3% ± 8.5% 72.0% ± 12.6%
Built-up land 91.2% ± 4.6% 97.8% ± 2.4%
Dryland 83.3% ± 3.9% 89.0% ± 3.4%
Floodplain 89.8% ± 8.6% 88.0% ± 9.1%
Forest 97.0% ± 1.1% 96.9% ± 0.9%
Grassland 75.0% ± 12.9% 66.0% ± 13.3%
Marsh 100.0% ± 0.0% 16.1% ± 12.0%
Paddy field 75.8% ± 10.4% 76.9% ± 10.3%
Swamp 89.1% ± 7.7% 95.0% ± 5.6%
Tidal flat 94.2% ± 6.4% 98.0% ± 3.9%
Water 92.2% ± 3.6% 96.7% ± 2.4%
GISCIENCE & REMOTE SENSING 13
(including paddy eld and dryland) had moderate
spatial consistency, of which most of their percentage
exceeded 65%. In summary, it was found that cate-
gories with large areas had high spatial consistency,
while rare land use types had low spatial consistency.
In addition, although our wetland classication has
high spatial consistency with CECD, there were still
divergences between these two datasets. The reasons
mainly involved three aspects. First, the production
method of datasets resulted in dierent regularities of
land use shapes. For the CECD, it was produced by
manual digitizing. If there were small land use patches
within a large land use patch, these small patches
were merged with the large patch. In contrast, the
small land use patches were accurately extracted from
our wetland maps, and they would not be merged
into the large land patches. Second, our wetland clas-
sications also had classication errors. For example,
the mapping accuracies of small land use types were
relative low. The paddy eld and dryland had some
misclassications due to their spectral similarity.
Third, for some land use types, their denitions were
dierent in the two classication systems. For exam-
ple, the CECD did not separate the agricultural ponds
and reservoir, and the reservoir/pond of CECD did not
completely include agricultural ponds.
4.3.2. Comparison with mangrove dataset
The statistical results indicated that the mangrove
areas of the two datasets were similar, with our
mangrove area being 122.5 km
2
and the MC_LASAC
mangrove area being 116.9 km
2
. Their spatial consis-
tency had an area of 90.8 km
2
, accounting for 74.1%
of our mangrove area. In addition, four typical regions
were chosen to further examine spatial details
(Figure 13). The two mangrove datasets shared an
almost identical spatial distribution and extent.
For the divergence between the two mangrove
datasets, it was mainly resulted from classication
methods and data sources. In our study, we extracted
the mangrove using the POK algorithm based on
Landsat images. In contrast, the MC_LASAC of 2018
was mapped by combining object-based image ana-
lysis and interpreter editing based on 2-m resolution
Gaofen-1 and Ziyuan-3 satellite imagery. Compared
with our mangrove, the patches of MC_LASAC man-
grove were more rened and regular.
4.3.3. Comparison with tidal flat dataset
The statistical results indicated that our tidal at
area was 349.2 km
2
, whereas the tidal at area of
the TWEA was 643.8 km
2
. Their spatial consistency
had an area of 278.8 km
2
, which accounted for
78.8% of our tidal at. Our tidal at area was
smaller than that of the TWEA dataset. These dif-
ferences mainly resulted from tidal variation. The
TWEA was produced by capturing the high tidal
images and low tidal images based on the
Sentinel-2 time series and delineated the maxi-
mum extent of the tidal at at a low level. Our
Figure 10. Spatial distribution of 10 wetland types in 2020.
14 K. PENG ET AL.
Figure 11. The spatial comparison between before and after fine classification in two typical regions. The (a) and (c) were
classifications with 6 wetlands using pixel-based random forest, and the (b) and (d) were classifications with 10 wetlands using object-
based hierarchical decision tree. The (a)-(b) were a typical region in GBGEZ, and the (c)-(d) are typical region in GBA.
Table 5. Accuracy results of hierarchical decision tree classification.
Category I Category II UA PA
Coastal wetland Coastal shallow water 97.7% ± 3.2% 95.5% ± 4.3%
Mangrove 90.7% ± 7.8% 98.0% ± 3.9%
Tidal flat 94.2% ± 6.4% 98.0% ± 3.9%
Human-made wetland Agriculture pond 88.1% ± 8.3% 94.5% ± 6.1%
Reservoir 90.7% ± 7.8% 98.0% ± 3.9%
Inland wetland Floodplain 89.8% ± 8.6% 88.0% ± 9.1%
Inland swamp 50.0% ± 43.8% 30.0% ± 29.9%
Lake 76.5% ± 20.8% 72.2% ± 21.3%
Marsh 100.0% ± 0.0% 16.2% ± 12.0%
River 90.6% ± 7.9% 96.0% ± 5.5%
Non-wetland Bare land 92.3% ± 8.5% 72.0% ± 12.6%
Built-up land 91.3% ± 4.6% 97.8% ± 2.4%
Dryland 83.2% ± 3.9% 89.6% ± 3.3%
Forest 96.9% ± 1.1% 97.2%±1.0%
Grassland 75.0% ± 12.9% 66.0% ± 13.3%
Paddy field 75.8% ± 10.4% 76.9% ± 10.3%
OA 91.6% ± 1.2%
GISCIENCE & REMOTE SENSING 15
phenology-based features characterized the aver-
age tidal level and could not capture the low tidal
level well. Thus, our wetland maps may underesti-
mate the tidal ats in the GBGZA and GBA.
However, the two tidal ats have similar spatial
distributions. As shown in Figure 14, the spatial
extent of the TWEA was larger than that of our
study, while their spatial distributions were similar.
Figure 12. Intra-comparison between our wetland maps and the CECD dataset. ourWetland refers to our wetland map.
Table 6. Confusion matrix between our wetland and CECD dataset in GBGEZ. The unit is percentage (%). In this table, the diagonal
values represent the percentage of consistency, while others represent the difference between two datasets.
CDCD dataset
Our Wetland Classifications La Riv FP TF CSW RP Ft GL BUL PF DL BL
La 0.1 10.1 0.2 0.0 0.0 72.7 5.0 0.0 2.1 4.4 5.4 0.0
Riv 0.1 80.4 2.1 0.1 0.1 3.4 6.2 0.1 1.1 2.6 3.9 0.0
FP 0.1 9.9 3.4 1.3 1.1 24.6 6.4 0.4 31.1 6.1 15.4 0.3
TF 0.0 4.0 0.4 67.4 24.6 0.9 1.3 0.1 1.1 0.2 0.0 0.0
CSW 0.0 12.5 0.1 26.1 60.2 0.6 0.2 0.0 0.3 0.0 0.0 0.0
RP 0.2 7.6 0.8 2.7 2.5 64.6 8.6 0.2 4.2 5.5 2.9 0.0
Ft 0.0 0.4 0.0 0.0 0.0 0.4 86.4 0.2 0.9 4.4 7.3 0.0
GL 0.0 0.5 0.1 0.0 0.0 0.4 64.3 0.5 7.9 11.1 15.1 0.1
BUL 0.0 1.2 0.2 0.3 0.2 3.9 6.4 0.4 77.1 5.5 4.5 0.2
PF 0.0 1.3 0.2 0.0 0.0 1.3 10.5 0.2 8.0 63.1 15.5 0.0
DL 0.0 1.2 0.1 0.0 0.0 1.4 22.7 0.2 10.5 24.8 39.1 0.0
BL 0.0 0.2 0.2 0.1 0.1 1.3 15.8 0.4 67.0 4.7 9.8 0.4
La->Lake, Riv->River, TF->Tidal flat, CSW->Coastal shallow water, RP->Reservoir/Pond, Ft->Forest, GL->Grassland, BUL->Built-up land, PF->Paddy field,
Dryland->DL, BL->Bare land.
16 K. PENG ET AL.
4.4. Area analysis of wetlands in two urban
agglomerations
Based on our POK algorithm and Landsat-8 time ser-
ies, we produced wetland maps with 10 wetland
types and 6 non-wetland types in the GBGEZ and
GBA. The area statistics of wetlands and non-
wetlands in the two coastal urban agglomerations
are shown in Table 8.
The wetland area of the GBGEZ in 2020 was 4198.8
km
2
, of which the areas of inland wetlands, coastal
wetlands and human-made wetlands were 666.4 km
2
,
2335.7 km
2
, and 1196.7 km
2
, respectively. The three
Table 7. Confusion matrix between our wetland and CECD dataset in GBA. The unit, meanings of elements, and abbreviations are
same as Table 6.
CDCD dataset
Our Wetland Classifications La Riv FP TF CSW RP Ft GL BUL PF DL BL
La 5.0 1.1 0.4 79.0 8.3 0.2 3.2 1.5 1.2 0.0
Riv 0.2 87.0 1.1 3.6 3.1 0.6 2.8 1.1 0.6 0.0
FP 8.6 9.8 6.9 31.5 7.3 1.7 21.6 8.7 3.0 0.8
TF
CSW
RP 1.2 5.0 0.3 68.7 5.3 0.7 9.1 7.5 2.2 0.0
Ft 0.0 0.3 0.0 0.9 93.7 0.1 1.2 2.0 1.8 0.0
GL 0.0 0.7 0.2 2.1 59.2 2.5 12.9 13.7 8.0 0.7
BUL 0.0 1.0 0.1 5.0 9.4 0.9 76.2 4.7 2.5 0.2
PF 0.0 1.6 0.4 4.3 17.1 0.9 8.8 54.2 12.6 0.1
DL 0.0 1.7 0.3 5.3 31.6 1.3 18.9 27.9 12.8 0.2
BL 0.2 0.3 0.3 3.0 15.2 1.3 68.7 4.1 4.5 2.4
Figure 13. Intra-comparison between the MC_LASAC dataset and our mangrove. The first row shows the mangrove distribution of our
wetland map in GBGEZ and GBA, respectively.
GISCIENCE & REMOTE SENSING 17
wetlands of categories I accounted for 15.9%, 47.5%,
and 37.5% of total wetland area of GBGEZ, respec-
tively. Among individual wetlands, the coastal shallow
waters had the largest area, followed by agricultural
ponds, rivers, reservoirs, tidal ats, mangroves,
oodplains, lakes, marshes, and inland swamps. For
inland wetlands, rivers had the largest area, with an
area of 505.7 km
2
. For coastal wetlands, areas of man-
grove, tidal at and coastal shallow water were 81.7
km
2
, 278.7 km
2
, and 1975.4 km
2
, respectively. For
Figure 14. Intra-comparison between the TWEA dataset and our tidal flat. The first row shows the spatial distribution of the tidal flat in
our wetland map.
Table 8. Areas of wetland types and non-wetland types in two urban agglomera-
tions (unit: km
2
).
Category I Category II GBGEZ GBA
Coastal wetland Coastal shallow water 1975.3 5078.8
Mangrove 81.7 40.9
Tidal flat 278.7 70.4
Human-made wetland Agricultural pond 697.8 3654.3
Reservoir 498.9 445.0
Inland wetland Inland swamp 18.9 110.2
Marsh 33.0 109.2
Lake 50.5 70.4
River 505.7 1234.1
Floodplain 58.3 118.8
Non-wetland Bare land 592.8 645.7
Built-up land 2010.8 8962.1
Dryland 19449.7 6466.1
Forest 44859.4 29928.5
Grassland 1275.0 725.1
Paddy field 2282.2 2916.6
18 K. PENG ET AL.
human-made wetlands, reservoirs, and agricultural
ponds had an area of 498.8 km
2
and 697.8 km
2
,
respectively.
Compared with the GBGEZ, the GBA had larger
wetland areas with a value of 10,932.2 km
2
, of which
the inland wetland area, coastal wetland area, and
human-made wetland area were 1642.7 km
2
, 5190.1
km
2
and 4099.3 km
2
, respectively. The three wetlands
of categories I accounted for 15.9%, 47.5%, and 37.5%
of total wetland area of GBA, respectively. For inland
wetlands, rivers had the largest area with a value of
1234.1 km
2
. For coastal wetlands, areas of mangroves,
tidal at and coastal shallow water were 40.9 km
2
,
70.4 km
2
, and 5078.8 km
2
, respectively. Human-made
wetlands, reservoirs, and agricultural ponds had an
area of 445.0 km
2
and 3654.4 km
2
, respectively.
5. Discussion
5.1. Eectiveness of our POK algorithm
Wetland mapping can oer useful information for wet-
land management and protection, which can well serve
in the Convention on Biological Diversity and support
the Global Biodiversity Framework (Fitoka et al. 2020;
van Rees et al. 2021). Wetland classication methods can
be grouped into two categories: supervised classica-
tion and hierarchical decision trees. Supervised classi-
cation methods can eectively extract rough land use
types (Piaser and Villa 2023; Xu et al. 2022) but are not
suitable for detailed wetland-type classication. For
example, supervised classication can extract water
bodies well but cannot identify water cover types (e.g.
rivers, lakes, reservoirs) (Li and Niu 2022). The hierarch-
ical decision tree method can eectively realize detailed
wetland-type classication by combining spectral and
geometric features. However, its rules and thresholds
always change over time and regions (Fitoka et al.
2020). The complexity of the algorithm limits its wide
application in wetland classication.
In our study, we developed a novel algorithm for
detailed wetland classication. Our POK algorithm
was accurate, eective, and robust for detailed wet-
land classication. In our POK algorithm, the pixel-
based RF was used to extract 6 wetland types
(Table 4), and the object-based hierarchical decision
tree was used to divide the above wetlands into 10
wetland types (Table 5). Compared with traditional
supervised classication, the POK algorithm can well
distinguish wetland types with similar spectra, such as
lakes, rivers, and reservoirs. Compared with the
object-based decision tree algorithm, the POK algo-
rithm exhibited better generalization. Its classication
rules and thresholds will not change over time and
regions. Meanwhile, the POK algorithm utilized the
phenology-based features that could capture the
spectral and seasonal characteristics of wetlands
well (Wu et al. 2021), which well assisted in the classi-
cation of complex wetland types.
The accuracy validation results indicated that the OA
of our wetland map was 91.8%±1.2%. For wetland types,
most had UA and PA values greater than 88.0%. For the
non-wetland types, the UA and PA values were over
72.0%. The above accuracy indices show that our wet-
land map had high accuracy. To further test the accuracy
of our map, we implemented a cross comparison with
existing datasets. It was found that the spatial consis-
tencies between our map and CECD accounted for
84.9% and 74.6% in GBGEZ and GBA, respectively. The
spatial consistency between our mangrove and
MC_LASAC accounted for 74.1% of our mangroves,
and the spatial consistency between our tidal at and
TWEA accounted for 78.8% of our tidal ats. Based on
above analysis, we believe that our wetland map could
reliably reect the wetland extents in our study areas.
5.2. Limitations, uncertainties, and outlooks
In this study, we successfully mapped 10 wetland
types and 6 non-wetland types with high accuracy.
However, there are uncertainties and limitations. First,
tidal ats were underestimated in the wetland map of
our POK algorithm. Our phenology-based features
represent the overall status of wetland conditions
but do not capture high and low tides (Jia et al.
2021). Second, the accuracy of inland swamps and
marshes was low because of their very small areas.
The eectiveness of our algorithm for large inland
swamps and marsh extraction needs to be further
explored in other regions. Third, there were misclassi-
cations between paddy elds and drylands due to
sample uncertainties caused by interpretation confu-
sion between paddy elds and drylands and their
spectral similarity. Forth, we just determined the
value of ntree based on literature review. However,
for this parameter, it needed conduct experiments on
relationship between ntree and out of bag (OBB)
errors. Fifth, we use the ETOPO1 to assist the
GISCIENCE & REMOTE SENSING 19
extraction of coastal shallow water, which may lead to
classication errors, because the ETOPO1 was static
water depth data with rough spatial resolution.
For our POK algorithm, there are some addition
limitations and potential factors aecting our wetland
classication. As is known, wetland extents often
changes with seasons. Our study did not consider
the seasonal variations on wetland changes, which
may resulted in errors in mapping wetland extents
(Xu, Niu, and Tang 2018). Due to the spectral com-
plexity of wetlands, spectral confusion of wetland
types was another factor inuencing our classication
accuracy. For example, the marsh and inland swamp
may misclassify with forest and paddy eld
(Mahdianpari et al. 2020). In addition, the quality of
training samples also aected our POK algorithm. For
our training samples, the error of wetland datasets
and roughness of manual interpretation may lead to
sample errors, which further aected the classication
accuracy (Peng et al. 2023).
In future studies, satellite images with high spatial
and temporal resolutions should be used to improve
wetland classications, such as Sentinel-2, Gaofen 1–
2, and IKONOS. The applicability and accuracy of our
POK algorithm when applied to these images
require further investigation. Previous studies have
indicated that SAR images are benecial to wetland
classication (Feng et al. 2022; Fu et al. 2017)
because they can capture valuable information
about the water and ground conditions under vege-
tation canopies. Compared to using optical images
alone, combining optical images with SAR will
improve the wetland classication (Lu and Wang
2021). However, using SAR images alone for wetland
classication is inferior in accuracy compared to
utilizing optical image time series (Vanderhoof
et al. 2023), because the optical images, with multi-
spectral bands, spectral bands, and phenology-
based features, can provide more valuable informa-
tion to the classication algorithm. In addition, to
further enhance the application of our study, it is
necessary to implement long-term wetland classi-
cation and monitor wetland dynamics using our POK
algorithm, which could reveal the spatiotemporal
evolution of detailed wetland types and oer useful
information and knowledge for wetland resource
protection and future policy-making.
6. Conclusion
In this study, we developed a POK algorithm for
detailed wetland classication by combining a pixel-
based random forest algorithm and an object-based
hierarchical decision tree. The results indicated that
the OA of our wetland map with 10 wetland types
was 91.6%±1.2%. Seven wetland types resulted in
high accuracies in our POK classication, including
rivers, oodplains, mangroves, tidal ats, coastal
shallow waters, reservoirs, and agricultural ponds,
with UA and PA values over 88.0%. For non-
wetland types, most UA and PA values were over
72.0%. By comparing with existing datasets, it was
found that the spatial consistency between our wet-
lands and CECD accounted for 84.9% and 74.6% of
GBGEZ and GBA, respectively. For the MC_LASAC
and TWEA dataset, our mangroves and tidal ats
had 74.1% and 78.8% spatial consistency with cor-
responding datasets.
In the GBGEZ in 2020, the wetland area was
4,198.8 km
2
, of which the inland wetland area
was 666.4 km
2
(15.9%), the coastal wetland area
was 2,335.7 km
2
(47.5%), and the human-made
wetland area was 1,196.7 km
2
(37.5%). In the
GBA in 2020, the wetland area was 10,932.2
km
2
, of which the inland wetland area was
1,642.7 km
2
(15.9%), the coastal wetland area
was 5,190.1 km
2
(47.5%), and the human-made
wetland area was 4,099.3 km
2
(37.5%). Our
study successfully mapped detailed wetland
types at a 30 m resolution in two urban agglom-
erations, which could potentially inform deci-
sion-making and planning.
Disclosure statement
No potential conict of interest was reported by the author(s).
Funding
This work was supported by the National Natural Science
Foundation of China under grant numbers U1901219,
U21A2022, 42301413, 42071393, and 41971306 and the
Natural Science Foundation of Tianjin under grant number
22JCQNJC01420.
20 K. PENG ET AL.
Data availability statement
China Ecosystem-type Classification
Dataset (CECD)
The data are not publicly available
because its production cost is
high, which just available to
specific users and departments
China Land Cover Dataset (CLCD) https://doi.org/10.5281/zenodo.
4417810
ETOPO1 https://earthengine.google.com/
Global Mangrove Distribution https://earthengine.google.com/
Global Mangrove Watch (GMW) https://data.unep-wcmc.org/data
sets/45
Global Intertidal Change https://earthengine.google.com
Global River Widths from Landsat
(GRWL)
https://zenodo.org/record/
1297434#.ZCboFHZByUk
Global Surface Water (GSW) https://earthengine.google.com
GOODD https://www.globaldamwatch.org/
directory
Tidal Wetlands in East Asia (TEWA) https://figshare.com/articles/data
set/Fujian_zip/14331785
HSL_MangroveChina_LASAC_share http://www.sasclouds.com/chinese/
home/
HydroLakes https://www.hydrosheds.org/pro
ducts/hydrolakes
Mangrove Map of China for 2019 https://doi.org/10.11922/sciencedb.
00245
Open Street Map Dam (OSM_Dam) https://www.globaldamwatch.org/
directory
Reservoir Point Of Interest (POI) https://lbs.amap.com/
Shuttle Radar Topography Mission
(SRTM)
https://earthengine.google.com/
References
Allen, G. H., and T. M. Pavelsky. 2018. “Global Extent of Rivers
and Streams.” Science 361 (6402): 585–587. https://doi.org/
10.1126/science.aat0636.
Amante, C., and B. W. Eakins. 2009. “ETOPO1 arc-minute global
relief model: procedures, data sources and analysis.” NOAA
Technical Memorandum NESDIS NGDC-24 19 (3): 2009.
https://repository.library.noaa.gov/view/noaa/1163
Bey, A., A. Sánchez-Paus Díaz, D. Maniatis, G. Marchi,
D. Mollicone, S. Ricci, J. F. Bastin, et al. 2016. “Collect Earth:
Land Use and Land Cover Assessment Through Augmented
Visual Interpretation.” Remote Sensing 8 (10): 807.
Bhatt, P., and A. L. Maclean. 2023. “Comparison of
High-Resolution NAIP and Unmanned Aerial Vehicle (UAV)
Imagery for Natural Vegetation Communities Classication
Using Machine Learning Approaches.” GIScience & Remote
Sensing 60:1. https://doi.org/10.1080/15481603.2023.
2177448.
Bon, G., E. Turak, D. Dudgeon, B. Bendandi, M. Thieme, J. Lento,
M. Simpson, and FWBON. 2022. Inland Waters in the Post-
2020 Global Biodiversity Framework. https://geobon.org/
science-briefs/ .
Bunting, P., A. Rosenqvist, R. M. Lucas, L. M. Rebelo, L. Hilarides,
N. Thomas, A. Hardy, T. Itoh, M. Shimada, and
C. M. Finlayson. 2018. “The Global Mangrove WatchA New
2010 Global Baseline of Mangrove Extent.” Remote Sensing
10 (10). https://doi.org/10.3390/rs10101669.
Cao, X. S., S. S. Ouyang, W. Y. Yang, Y. Luo, B. C. Li, and D. Liu.
2019. “Transport Accessibility and Spatial Connections of
Cities in the Guangdong-Hong Kong-Macao Greater Bay
Area.” Chinese Geographical Science 29 (5): 820–833.
https://doi.org/10.1007/s11769-019-1034-2.
Chen, K. X., P. F. Cong, L. M. Qu, S. X. Liang, and Z. C. Sun. 2022.
“Annual Variation of the Landscape Pattern in the Liao River
Delta Wetland from 1976 to 2020.” Ocean & Coastal
Management 224. https://doi.org/10.1016/j.ocecoaman.
2022.106175.
Deng, Y. W., W. G. Jiang, Z. F. Wu, Z. Y. Ling, K. F. Peng, and
Y. Deng. 2022. “Assessing Surface Water Losses and Gains
Under Rapid Urbanization for SDG 6.6. 1 Using Long-Term
Landsat Imagery in the Guangdong-Hong Kong-Macao
Greater Bay Area, China.” Remote Sensing 14 (4): 881.
Deng, Y., Z. F. Shao, C. Y. Dang, X. Huang, W. F. Wu,
Q. W. Zhuang, and Q. Ding. 2023. “Assessing Urban
Wetlands Dynamics in Wuhan and Nanchang, China.”
Science of the Total Environment 901:165777. https://doi.
org/10.1016/j.scitotenv.2023.165777 .
Elmahdy, S. I., and M. M. Mohamed. 2018. “Monitoring and
Analysing the Emirate of Dubai’s Land Use/Land Cover
Changes: An Integrated, Low-Cost Remote Sensing
Approach.” International Journal of Digital Earth 11 (11):
1132–1150. https://doi.org/10.1080/17538947.2017.
1379563.
Farr, T. G., P. A. Rosen, E. Caro, R. Crippen, R. Duren, S. Hensley,
M. Kobrick, et al. 2007. “The Shuttle Radar Topography
Mission.” Reviews of Geophysics 45 (2). https://doi.org/10.
1029/2005RG000183.
Feng, K. D., D. H. Mao, Z. Q. Qiu, Y. X. Zhao, and Z. M. Wang.
2022. “Can Time-Series Sentinel Images Be Used to Properly
Identify Wetland Plant Communities?” GIScience & Remote
Sensing 59 (1): 2202–2216.
Feyisa, G. L., H. Meilby, R. Fensholt, and S. R. Proud. 2014.
“Automated Water Extraction Index: A New Technique for
Surface Water Mapping Using Landsat Imagery.” Remote
Sensing of Environment 140:23–35. https://doi.org/10.1016/
j.rse.2013.08.029.
Fitoka, E., M. Tompoulidou, L. Hatziiordanou, A. Apostolakis,
R. Hofer, K. Weise, and C. Ververis. 2020. “Water-Related
ecosystems’ Mapping and Assessment Based on Remote
Sensing Techniques and Geospatial Analysis: The SWOS
National Service Case of the Greek Ramsar Sites and Their
Catchments.” Remote Sensing of Environment 245. https://
doi.org/10.1016/j.rse.2020.111795.
Fu, B. L., Y. Q. Wang, A. Campbell, Y. Li, B. Zhang, S. B. Yin,
Z. F. Xing, and X. M. Jin. 2017. “Comparison of Object-Based
and Pixel-Based Random Forest Algorithm for Wetland
Vegetation Mapping Using High Spatial Resolution GF-1
and SAR Data.” Ecological Indicators 73:105–117. https://
doi.org/10.1016/j.ecolind.2016.09.029 .
Gell, P. A., C. M. Finlayson, N. C. Davidson. 2023. “An
Introduction to the Ramsar Convention on Wetlands.” In
Ramsar Wetlands, 1–36. Switzerland: Elsevier.
Giri, C., E. Ochieng, L. L. Tieszen, Z. Zhu, A. Singh, T. Loveland,
J. Masek, and N. Duke. 2011. “Status and Distribution of
GISCIENCE & REMOTE SENSING 21
Mangrove Forests of the World Using Earth Observation
Satellite Data.” Global Ecology & Biogeography 20 (1):
154–159. https://doi.org/10.1111/j.1466-8238.2010.00584.x.
Gong, P., H. Liu, M. N. Zhang, C. C. Li, J. Wang, H. B. Huang,
N. Clinton, et al. 2019. “Stable Classication with Limited
Sample: Transferring a 30-M Resolution Sample Set
Collected in 2015 to Mapping 10-M Resolution Global Land
Cover in 2017.” Science Bulletin 64 (6): 370–373. https://doi.
org/10.1016/j.scib.2019.03.002.
Gorelick, N., M. Hancher, M. Dixon, S. Ilyushchenko, D. Thau,
and R. Moore. 2017. “Google Earth Engine: Planetary-Scale
Geospatial Analysis for Everyone.” Remote Sensing of
Environment 202:18–27. https://doi.org/10.1016/j.rse.2017.
06.031 .
Guo, H. J., Y. P. Cai, Z. F. Yang, Z. C. Zhu, and Y. R. Ouyang. 2021.
“Dynamic Simulation of Coastal Wetlands for
Guangdong-Hong Kong-Macao Greater Bay Area Based on
Multi-Temporal Landsat Images and FLUS Model.” Ecological
Indicators 125. https://doi.org/10.1016/j.ecolind.2021.
107559
Hou, T. T., W. W. Sun, C. Chen, G. Yang, X. C. Meng, and
J. T. Peng. 2022. “Marine Floating Raft Aquaculture
Extraction of Hyperspectral Remote Sensing Images Based
Decision Tree Algorithm.” International Journal of Applied
Earth Observation and Geoinformation 111:102846. https://
doi.org/10.1016/j.jag.2022.102846 .
Huete, A., H. Liu, K. Batchily, and W. Van Leeuwen. 1997.
“A Comparison of Vegetation Indices Over a Global Set of
TM Images for EOS-MODIS.” Remote Sensing of Environment
59 (3): 440–451.
Hu, X. D., P. L. Zhang, Q. Zhang, and J. Q. Wang. 2021.
“Improving Wetland Cover Classication Using Articial
Neural Networks with Ensemble Techniques.” GIScience &
Remote Sensing 58 (4): 603–623.
Jamali, A., M. Mahdianpari, B. Brisco, J. Granger,
F. Mohammadimanesh, and B. Salehi. 2021. “Deep Forest
Classier for Wetland Mapping Using the Combination of
Sentinel-1 and Sentinel-2 Data.” GIScience & Remote Sensing
58 (7): 1072–1089. https://doi.org/10.1080/15481603.2021.
1965399.
Jia, M. M., D. H. Mao, Z. M. Wang, C. Y. Ren, Q. D. Zhu, X. C. Li,
and Y. Z. Zhang. 2020. “Tracking Long-Term Floodplain
Wetland Changes: A Case Study in the China Side of the
Amur River Basin.” International Journal of Applied Earth
Observation and Geoinformation 92:102185. https://doi.org/
10.1016/j.jag.2020.102185 .
Jia, M. M., Z. M. Wang, D. H. Mao, C. Y. Ren, K. S. Song, C. P. Zhao,
C. Wang, X. M. Xiao, and Y. Q. Wang. 2023. “Mapping Global
Distribution of Mangrove Forests at 10-M Resolution.”
Science Bulletin 68:1306–1316. https://doi.org/10.1016/j.
scib.2023.05.004.
Jia, M. M., Z. M. Wang, D. H. Mao, C. Y. Ren, C. Wang, and
Y. Q. Wang. 2021. “Rapid, Robust, and Automated Mapping
of Tidal Flats in China Using Time Series Sentinel-2 Images
and Google Earth Engine.” Remote Sensing of Environment
255. https://doi.org/10.1016/j.rse.2021.112285.
Joly, C. A. 2023. The Kunming-Montreal Global Biodiversity
Framework. Kunming: SciELO Brasil.
Keith, D. A., D. H. Benson, I. R. C. Baird, L. Watts, C. C. Simpson,
M. Krogh, S. Gorissen, J. R. Ferrer-Paris, and T. J. Mason. 2023.
“Eects of Interactions Between Anthropogenic Stressors
and Recurring Perturbations on Ecosystem Resilience and
Collapse.” Conservation Biology: The Journal of the Society for
Conservation Biology 37 (1): e13995. https://doi.org/10.1111/
cobi.13995.
Kuang, W. H., S. W. Zhang, G. M. Du, C. Z. Yan, S. X. Wu, R. D. Li,
D. S. Lu, et al. 2022. “Monitoring Periodically National Land
Use Changes and Analyzing Their Spatiotemporal Patterns
in China During 2015-2020.” Journal of Geographical Sciences
32 (9): 1705–1723. https://doi.org/10.1007/s11442-022-
2019-0.
Lehner, B., and P. Doll. 2004. “Development and Validation of
a Global Database of Lakes, Reservoirs and Wetlands.”
Canadian Journal of Fisheries and Aquatic Sciences 296 (1–
4): 1–22. https://doi.org/10.1016/j.jhydrol.2004.03.028.
Li, Y., and Z. G. Niu. 2022. “Systematic Method for Mapping
Fine-Resolution Water Cover Types in China Based on Time
Series Sentinel-1 and 2 Images.” International Journal of
Applied Earth Observation and Geoinformation 106. https://
doi.org/10.1016/j.jag.2021.102656.
Li, A. Z., K. S. Song, S. B. Chen, Y. L. Mu, Z. Y. Xu, and Q. H. Zeng.
2022. “Mapping African Wetlands for 2020 Using Multiple
Spectral, Geo-Ecological Features and Google Earth Engine.”
Isprs Journal of Photogrammetry & Remote Sensing 193:252–
268. https://doi.org/10.1016/j.isprsjprs.2022.09.009 .
Liu, J. Y., W. H. Kuang, Z. X. Zhang, X. L. Xu, Y. W. Qin, J. Ning,
W. C. Zhou, et al. 2014. “Spatiotemporal Characteristics,
Patterns, and Causes of Land-Use Changes in China Since
the Late 1980s.” Journal of Geographical Sciences 24 (2):
195–210. https://doi.org/10.1007/s11442-014-1082-6.
Liu, Y., H. Q. Zhang, M. Zhang, Z. Y. Cui, K. X. Lei, J. Zhang,
T. D. Yang, and P. Ji. 2022. “Vietnam Wetland Cover Map:
Using Hydro-Periods Sentinel-2 Images and Google Earth
Engine to Explore the Mapping Method of Tropical
Wetland.” International Journal of Applied Earth Observation
and Geoinformation 115:103122. https://doi.org/10.1016/j.
jag.2022.103122 .
Lu, Y., and L. Wang. 2021. “How to Automate Timely
Large-Scale Mangrove Mapping with Remote Sensing.”
Remote Sensing of Environment 264. https://doi.org/10.
1016/j.rse.2021.112584.
Mahdianpari, M., H. Jafarzadeh, J. E. Granger,
F. Mohammadimanesh, B. Brisco, B. Salehi, S. Homayouni,
and Q. H. Weng. 2020. “A Large-Scale Change Monitoring of
Wetlands Using Time Series Landsat Imagery on Google
Earth Engine: A Case Study in Newfoundland.” GIScience &
Remote Sensing 57 (8): 1102–1124.
Mao, D. H., Z. M. Wang, B. J. Du, L. Li, Y. L. Tian, M. M. Jia, Y. Zeng,
K. S. Song, M. Jiang, and Y. Q. Wang. 2020. “National Wetland
Mapping in China: A New Product Resulting from
Object-Based and Hierarchical Classication of Landsat 8
OLI Images.” Isprs Journal of Photogrammetry & Remote
22 K. PENG ET AL.
Sensing 164:11–25. https://doi.org/10.1016/j.isprsjprs.2020.
03.020.
Mao, D. H., H. Yang, Z. M. Wang, K. Song, J. R. Thompson, and
R. J. Flower. 2022. “Reverse the Hidden Loss of China’s
Wetlands.” Science 376 (6597): 1061–1061. https://doi.org/
10.1126/science.adc8833.
McFeeters, S. K. 1996. “The Use of the Normalized Dierence
Water Index (NDWI) in the Delineation of Open Water
Features.” International Journal of Remote Sensing 17 (7):
1425–1432.
Messager, M. L., B. Lehner, G. Grill, I. Nedeva, and O. Schmitt.
2016. “Estimating the Volume and Age of Water Stored in
Global Lakes Using a Geo-Statistical Approach.” Nature
Communications 7. https://doi.org/10.1038/ncomms13603.
Mulligan, M., A. van Soesbergen, and L. Saenz. 2020. “GOODD,
a Global Dataset of More Than 38,000 Georeferenced Dams.”
Scientic Data 7 (1). https://doi.org/10.1038/s41597-020-
0362-5.
Murray, N. J., S. R. Phinn, M. DeWitt, R. Ferrari, R. Johnston,
M. B. Lyons, N. Clinton, D. Thau, and R. A. Fuller. 2019. “The
Global Distribution and Trajectory of Tidal Flats.” Nature
565 (7738): 222±. https://doi.org/10.1038/s41586-018-0805-
8.
Murray, N. J., T. A. Worthington, P. Bunting, S. Duce, V. Hagger,
C. E. Lovelock, R. Lucas, et al. 2022. “High-Resolution
Mapping of Losses and Gains of Earth’s Tidal Wetlands.”
Science 376 (6594): 744±. https://doi.org/10.1126/science.
abm9583.
Navarro, A., M. Young, P. I. Macreadie, E. Nicholson, and
D. Ierodiaconou. 2021. “Mangrove and Saltmarsh
Distribution Mapping and Land Cover Change Assessment
for South-Eastern Australia from 1991 to 2015.” Remote
Sensing 13 (8): 1450.
Ni, R. G., J. Y. Tian, X. J. Li, D. M. Yin, J. W. Li, H. L. Gong, J. Zhang,
L. Zhu, and D. L. Wu. 2021. “An Enhanced Pixel-Based
Phenological Feature for Accurate Paddy Rice Mapping
with Sentinel-2 Imagery in Google Earth Engine.” Isprs
Journal of Photogrammetry & Remote Sensing 178:282–296.
https://doi.org/10.1016/j.isprsjprs.2021.06.018.
Niu, Z. G., H. Y. Zhang, X. W. Wang, W. B. Yao, D. M. Zhou,
K. Y. Zhao, H. Zhao, et al. 2012. “Mapping Wetland Changes
in China Between 1978 and 2008.” Chinese Science Bulletin
57 (22): 2813–2823. https://doi.org/10.1007/s11434-012-
5093-3.
Olofsson, P., G. M. Foody, M. Herold, S. V. Stehman,
C. E. Woodcock, and M. A. Wulder. 2014. “Good Practices
for Estimating Area and Assessing Accuracy of Land
Change.” Remote Sensing of Environment 148:42–57.
https://doi.org/10.1016/j.rse.2014.02.015.
Ouyang, Z. Y., H. Zheng, Y. Xiao, S. Polasky, J. G. Liu, W. H. Xu,
Q. Wang, L. Zhang, Y. Xiao, and E. M. Rao. 2016.
“Improvements in Ecosystem Services from Investments in
Natural Capital.” Science 352 (6292): 1455–1459.
Owers, C. J., R. M. Lucas, D. Clewley, B. Tissott, S. M. T. Chua,
G. Hunt, N. Mueller, et al. 2022. “Operational
Continental-Scale Land Cover Mapping of Australia Using
the Open Data Cube.” International Journal of Digital Earth
15 (1): 1715–1737. https://doi.org/10.1080/17538947.2022.
2130461.
Pekel, J. F., A. Cottam, N. Gorelick, and A. S. Belward. 2016.
“High-Resolution Mapping of Global Surface Water and Its
Long-Term Changes.” Nature 540 (7633): 418±. https://doi.
org/10.1038/nature20584.
Peng, K. F., W. G. Jiang, P. Hou, Z. Y. Ling, D. H. Mao, and
Z. H. Huang. 2021. “Dense Wetland Sample Production at
Large Scale by Combining Multi-Source Thematic Datasets
and Visual Interpretation.” National Remote Sensing Bulletin
XX (XX): 13. https://doi.org/10.11834/jrs.20211152.
Peng, K. F., W. G. Jiang, P. Hou, Z. F. Wu, Z. Y. Ling, X. Y. Wang,
Z. G. Niu, and D. H. Mao. 2023. “Continental-Scale Wetland
Mapping: A Novel Algorithm for Detailed Wetland Types
Classication Based on Time Series Sentinel-1/2 Images.”
Ecological Indicators 148:110113. https://doi.org/10.1016/j.
ecolind.2023.110113.
Piaser, E., and P. Villa. 2023. “Evaluating Capabilities of Machine
Learning Algorithms for Aquatic Vegetation Classication in
Temperate Wetlands Using Multi-Temporal Sentinel-2 Data.”
International Journal of Applied Earth Observation and
Geoinformation 117:103202. https://doi.org/10.1016/j.jag.
2023.103202 .
Rodriguez-Galiano, V. F., and M. Chica-Rivas. 2014. “Evaluation
of Dierent Machine Learning Methods for Land Cover
Mapping of a Mediterranean Area Using Multi-Seasonal
Landsat Images and Digital Terrain Models.” International
Journal of Digital Earth 7 (6): 492–509. https://doi.org/10.
1080/17538947.2012.748848.
Rouse, J. W., R. C. Haas, J. A. Schell, and D. W. Deering. 1974.
“Monitoring Vegetation Systems in the Great Plains with
ERTS.” NASA Special Publications 351 (1): 309.
Stehman, S. V., B. W. Pengra, J. A. Horton, and D. F. Wellington.
2021. “Validation of the US Geological Survey’s Land Change
Monitoring, Assessment and Projection (LCMAP) Collection
1.0 Annual Land Cover Products 1985–2017.” Remote
Sensing of Environment 265:112646. https://doi.org/10.
1016/j.rse.2021.112646.
Stovall, A. E. L., J. S. Diamond, R. A. Slesak, D. L. McLaughlin, and
H. Shugart. 2019. “Quantifying Wetland Microtopography
with Terrestrial Laser Scanning.” Remote Sensing of
Environment 232. https://doi.org/10.1016/j.rse.2019.111271.
Vanderhoof, M. K., L. Alexander, J. Christensen, K. Solvik,
P. Nieuwlandt, and M. Sagehorn. 2023. “High-Frequency
Time Series Comparison of Sentinel-1 and Sentinel-2
Satellites for Mapping Open and Vegetated Water Across
the United States (2017-2021.” Remote Sensing of
Environment 288. https://doi.org/10.1016/j.rse.2023.113498.
van Rees, C. B., K. A. Waylen, A. Schmidt-Kloiber, S. J. Thackeray,
G. Kalinkat, K. Martens, S. Domisch, et al. 2021.
“Safeguarding Freshwater Life Beyond 2020:
Recommendations for the New Global Biodiversity
Framework from the European Experience.” Conservation
Letters 14 (1). https://doi.org/10.1111/conl.12771.
Wang, X. Y., W. G. Jiang, K. F. Peng, Z. Li, and P. Z. Rao. 2022.
“A Framework for Fine Classication of Urban Wetlands
Based on Random Forest and Knowledge Rules: Taking the
GISCIENCE & REMOTE SENSING 23
Wetland Cities of Haikou and Yinchuan as Examples.”
GIScience & Remote Sensing 59 (1): 2144–2163. https://doi.
org/10.1080/15481603.2022.2152926.
Wang, M., D. H. Mao, Y. Q. Wang, X. M. Xiao, H. X. Xiang,
K. D. Feng, L. Luo, M. M. Jia, K. S. Song, and Z. M. Wang.
2023. “Wetland Mapping in East Asia by Two-Stage
Object-Based Random Forest and Hierarchical Decision
Tree Algorithms on Sentinel-1/2 Images.” Remote Sensing
of Environment 297 (August): 113793–113793. https://doi.
org/10.1016/j.rse.2023.113793.
Wu, Z., Z. Cao, S. Song, W. Jiang, G. Guo, and Y. Wu. 2020.
“Wetland Remote Sensing Monitoring and Assessment in
Guangdong-Hong Kong-Macau Greater Bay Area: Current
Status, Challenges and Future Perspectives.” Acta Ecologica
Sinica 40 (23): 11.
Wu, N., R. H. Shi, W. Zhuo, C. Zhang, B. C. Zhou, Z. L. Xia, Z. Tao,
W. Gao, and B. Tian. 2021. “A Classication of Tidal Flat
Wetland Vegetation Combining Phenological Features with
Google Earth Engine.” Remote Sensing 13:3. https://doi.org/
10.3390/rs13030443.
Xia, Q., T. T. He, C. Z. Qin, X. M. Xing, and W. Xiao. 2022. “An
Improved Submerged Mangrove Recognition Index-Based
Method for Mapping Mangrove Forests by Removing the
Disturbance of Tidal Dynamics and S. Alterniora.” Remote
Sensing 14 (13): 3112.
Xie, S., L. Y. Liu, X. Zhang, J. N. Yang, X. D. Chen, and Y. Gao.
2019. “Automatic Land-Cover Mapping Using Landsat
Time-Series Data Based on Google Earth Engine.” Remote
Sensing 11 (24): 3023.
Xu, H. Q. 2006. “Modication of Normalised Dierence Water
Index (NDWI) to Enhance Open Water Features in Remotely
Sensed Imagery.” International Journal of Remote Sensing
27 (14): 3025–3033.
Xu, P. P., M. Herold, N. E. Tsendbazar, and J. Clevers. 2020.
“Towards a Comprehensive and Consistent Global Aquatic
Land Cover Characterization Framework Addressing
Multiple User Needs.” Remote Sensing of Environment 250.
https://doi.org/10.1016/j.rse.2020.112034.
Xu, P. P., Z. G. Niu, and P. Tang. 2018. “Comparison and
Assessment of NDVI Time Series for Seasonal Wetland
Classication.” International Journal of Digital Earth 11 (11):
1103–1131.
Xu, P. P., N. E. Tsendbazar, M. Herold, J. Clevers, and L. L. Li.
2022. “Improving the Characterization of Global Aquatic
Land Cover Types Using Multi-Source Earth Observation
Data.” Remote Sensing of Environment 278. https://doi.org/
10.1016/j.rse.2022.113103.
Yang, J., and X. Huang. 2021. “The 30 M Annual Land Cover
Dataset and Its Dynamics in China from 1990 to 2019.” Earth
System Science Data 13 (8): 3907–3925. https://doi.org/10.
5194/essd-13-3907-2021.
Zhang, Z., W. G. Jiang, K. F. Peng, Z. F. Wu, Z. Y. Ling, and Z. Li.
2023. “Assessment of the Impact of Wetland Changes on
Carbon Storage in Coastal Urban Agglomerations from 1990
to 2035 in Support of SDG15. 1.” Science of the Total
Environment 877:162824. https://doi.org/10.1016/j.scito
tenv.2023.162824 .
Zhang, X., L. Y. Liu, X. D. Chen, Y. Gao, S. Xie, and J. Mi. 2021.
“GLC_FCS30: Global Land-Cover Product with Fine
Classication System at 30 M Using Time-Series Landsat
Imagery.” Earth System Science Data 13 (6): 2753–2776.
Zhang, Z., N. Xu, Y. F. Li, and Y. Li. 2022. “Sub-Continental-Scale
Mapping of Tidal Wetland Composition for East Asia:
A Novel Algorithm Integrating Satellite Tide-Level and
Phenological Features.” Remote Sensing of Environment
269. https://doi.org/10.1016/j.rse.2021.112799.
Zhang, T., S. You, X. Yang, and S. Hu. 2020. “Mangroves Map of
China 2018 (MC2018) Derived from 2-Meter Resolution
Satellite Observations and Field Data.” Science Data Bank:
Beijing, China 1. https://doi.org/10.11922/sciencedb.00449.
Zhao, C. P., M. M. Jia, Z. M. Wang, D. H. Mao, and Y. Q. Wang.
2023. “Toward a Better Understanding of Coastal Salt Marsh
Mapping: A Case from China Using Dual-Temporal Images.”
Remote Sensing of Environment 295:113664. https://doi.org/
10.1016/j.rse.2023.113664 .
Zhao, C. P., and C. Z. Qin. 2020. “10-M-Resolution Mangrove
Maps of China Derived from Multi-Source and
Multi-Temporal Satellite Observations.” Isprs Journal of
Photogrammetry & Remote Sensing 169:389–405. https://
doi.org/10.1016/j.isprsjprs.2020.10.001.
Zheng, X. Y., Y. Wang, M. Y. Gan, J. Zhang, L. M. Teng, K. Wang,
Z. Q. Shen, and L. Zhang. 2016. “Discrimination of
Settlement and Industrial Area Using Landscape Metrics in
Rural Region.” Remote Sensing 8 (10): 845.
24 K. PENG ET AL.
Appendix
Table A1. The corresponding relationship between the classes of CECD dataset and our wetland classification.
Classification system of CECD Classification system of our study
Category I Category II Category
Wetland River River
Lake Lake
Reservoir/pond Reservoir, Agricultural pond
Tidal flat Tidal flat
Floodplain Floodplain
Coastal shallow water Coastal shallow water
Vegetated wetland Inland swamp, marsh, mangrove
Forest Forestland Forest
Shrub land
Sparse woodland
Other forest land
Grassland High vegetation cover grassland Grassland
Median vegetation cover grassland
Low vegetation cover grassland
Built-up land Urban land Built-up land
Rural Settlement
Other construction land
Cropland Paddy field Paddy field
Dryland Dryland
Unused land Sand land Bare land
Saline-Alkali Land
Bare soil
Bare rock
Other
GISCIENCE & REMOTE SENSING 25
... Regarding sample point selection, discrepancies between the collection time of samples and the acquisition time of remote sensing images also impacted the stability and reliability of wetland monitoring. Wetland type distribution changes over time [68], primarily due to dynamic water fluctuations, which alter wetland boundaries [69]. These changes are more pronounced during the rainy summer months. ...
Article
Full-text available
Wetlands play a vital role in regulating the global carbon cycle, providing biodiversity, and reducing flood risks. These functions maintain ecological balance and ensure human well-being. Timely, accurate monitoring of wetlands is essential, not only for conservation efforts, but also for achieving Sustainable Development Goals (SDGs). In this study, we combined Sentinel-1/2 images, terrain data, and field observation data collected in 2020 to better understand wetland distribution. A total of 22 feature variables were extracted from multi-source data, including spectral bands, spectral indices (especially red edge indices), terrain features, and radar features. To avoid high correlations between variables and reduce data redundancy, we selected a subset of features based on recursive feature elimination (RFE) and Pearson correlation analysis methods. We adopted the random forest (RF) method to construct six wetland delineation schemes and incorporated multiple types of characteristic variables. These variables were based on remote sensing image pixels and objects. Combining red-edge features, terrain data, and radar data significantly improved the accuracy of land cover information extracted in low-mountain and hilly areas. Moreover, the accuracy of object-oriented schemes surpassed that of pixel-level methods when applied to wetland classification. Among the three pixel-based schemes, the addition of terrain and radar data increased the overall classification accuracy by 7.26%. In the object-based schemes, the inclusion of radar and terrain data improved classification accuracy by 4.34%. The object-based classification method achieved the best results for swamps, water bodies, and built-up land, with relative accuracies of 96.00%, 90.91%, and 96.67%, respectively. Even higher accuracies were observed in the pixel-based schemes for marshes, forests, and bare land, with relative accuracies of 98.67%, 97.53%, and 80.00%, respectively. This study’s methodology can provide valuable reference information for wetland data extraction research and can be applied to a wide range of future research studies.
... However, for the study of the nature of alpine wetland ecosystems, it is equally necessary to discuss their water-related surface changes such as the reduction of snowmelt, river recharge, and the decreased wetland soil water content. Synthetic aperture radar and optical images from Sentinel missions are commonly used in wetland mapping [ 17 ], including detailed pixel-based and object-based wetland-type classification [ 18 ] and water cover mapping based on time series images, shape features, and flooding frequency [ 19 -21 ]. ...
Article
Full-text available
Monitoring alpine wetland degradation on the Qinghai–Tibet Plateau is crucial for understanding the responses to and resilience against climate change but has been challenging due to limited images in cloudy high-mountain areas. Based on 3 elements, spectral–temporal characterization, classification, and degradation detection for wetland covers, this study proposes a continuous classification and degradation detection algorithm for alpine wetlands (AW-CCD). This algorithm relates to water-related landscape change processes, including multiscenario detection of snowmelt, lake, and river shrinkage and the transition of a swampy meadow to an alpine meadow with decreased soil wetness. AW-CCD uses the spectral–temporal index features to classify wetlands on an annual basis and then capture wetland degradation processes to combine long-time-series inter-annual parameters and seasonal soil wetness. This study detected snow cover from clouds based on the Landsat Quality Assessment band and spectral changes during snow–bare rock transition. Through the meadow spectral ratio vegetation index and seasonal soil wetness frequency across years, swampy and alpine meadow dynamics are tracked by wetness loss and increasing grass signal. By effectively characterizing multiple surface changes through spectral–temporal analysis, AW-CCD provides annual wetland mapping and monitoring metrics for multiscenario degradation. Results show an improvement in snow and meadow mapping accuracy by 5% and 3%, respectively, with a mapping accuracy of 94.9% in the Maidika Wetland in 2022. Spatial–temporal patterns demonstrated multiscenario degradation during 2 decades, with snow and river areas decreasing by 5.04% and 16.74%, respectively, and 3.23% of swampy meadows transitioning to alpine meadows. Degradation was most pronounced before 2009, followed by stability until 2015 and renewed degradation thereafter. This study highlights the effectiveness of AW-CCD in capturing the multiscenario responses of alpine wetlands to climatic changes on the Qinghai–Tibet Plateau.
... Rectangular Fit [71,72] is a measure of rectangularity, where a perfect rectangle has a value of 1, and non-rectangular shapes have values less than 1. The formula for Rectangular Fit is defined as: ...
Article
Full-text available
At present, the extraction of irrigation canal network distribution information is of great significance for developing a digital twin irrigation district. However, due to the low resolution of remote sensing images, it is difficult to effectively identify the canal networks, especially for channels with a width of less than 1 m, where recognition is insufficient. Therefore, the purpose of this study is to extract canal networks of different widths in an irrigation district in Shaanxi Province as the research area. A rule-based object-oriented classification method was employed, utilizing image data collected by the DJI Mavic 3 multispectral UAV (Unmanned Aerial Vehicle) to explore the accuracy of this method in extracting canal distribution information. Based on UAV multispectral remote sensing imagery, the segmentation parameters for the remote sensing imagery were determined using ENVI 5.6 software, with the segmentation threshold set at 60 and the merging threshold set at 80. By combining the spectral and spatial differences between the canals and other ground objects, rules for extracting canal network distribution information were established, and the information on the distribution of channels in this irrigation area was finally obtained. The experimental results showed a maximum recall rate of 91.88% and a maximum precision rate of 57.59%. The overall recall precision rates for the irrigation district were 85.74% and 55.08%, respectively. This method provides a new solution for identifying and extracting canal systems in irrigation districts, offering valuable insights for acquiring canal distribution information and providing a scientific basis for precision irrigation.
... Wang et al. 2020). In contrast, remote sensing technologies offer various advantages such as wide coverage, repeatable observations, and comprehensive informational content, making them highly suitable for wetland mapping (Amani et al. 2021;Peng et al. 2024;. The rapid development of computer technology has facilitated the evolution of wetland mapping from manual visual interpretation towards semi-automatic and automatic methods based on machine learning (ML) . ...
... The mangrove forest data from 1990-2035 adopted in this study were contributed by Peng (Peng 2022;Peng et al. 2024). In Peng's study, the initial mangrove sample points were generated by Global Mangrove Watch (https://data.unep-wcmc.org/datasets/45), and the Global Mangrove Watch data from 1996 to 2016 with 7 periods was spatially overlaid to generate random mangrove points in the common intersection area. ...
Article
Full-text available
It is crucial to mangrove management by understanding its conservation efforts. This study proposed a spatiotemporal analysis model to evaluate mangrove changes in the Guangxi Beibu Gulf (GBG) and the Guangdong-Hong Kong-Macao Greater Bay Area (GBA), focusing on past, present, and future dynamics. The analysis covered mangrove loss, stability, and gain from 1990 to 2020 and provided projections for 2020 to 2035 across multiple scenarios. Then further assessed the efficacy of past mangrove conservation with the Sustainable Development Goals (SDGs) and the prospective conservation potential with the Global Mangrove Alliance Goal (GMAG) indicators. Findings indicate that by 2020, the GBG and GBA had 81.65 km² and 40.92 km² of mangroves, respectively, with increases of 31.51% and 164.09% since 1990. Projected expansions are anticipated under all three scenarios during 2020- 2035, with the highest growth projected under the trend continuation scenario, followed by ecological restoration and protection, and economic development. While mangrove protection in these regions met SDG targets 6.6.1 and 14.5.1 from 2015 to 2020, the 80% conservation milestone set by GMAG for 2030 was not achieved, necessitating the establishment of nature reserves. The study offers novel insights for sustainable governance of mangrove forests at local, regional, and global levels.
Article
Images derived from an individual sensor fail to simultaneously satisfy the demands of high spatial, spectral, and temporal resolutions. Multisource remote sensing image fusion provides efficient access to high-spatial-resolution multispectral images (spatial spectral fusion, SSF) and high-temporal-spatial-resolution images (spatiotemporal fusion, STF). Whereas existing deep learning-based models can mainly implement either SSF or STF, there is an urgent need for models that can simultaneously implement both SSF and STF. A unified generative adversarial network with convolution and Transformer (CTUGAN) for SSF and STF is proposed. CTUGAN contains an adaptive-resolution generator (ACTG) and multiresolution discriminators (MCTDs), both with the convolution and Transformer. First, a bidirectional local-global feature encoder is devised in the ACTG to extract local-global features via a high-to-low resolution and a low-to-high resolution, respectively. Then, a multihead cross-attention fusion decoder is devised to aggregate and fuse complementary local-global features of various levels and resolutions hierarchically to restore valuable information. Moreover, MCTDs adversely learn multiresolution local-global features to identify the relative reality of products, and a generalized loss function is built to accomplish full supervision. Finally, numerous experiments on the SSF data (Gaofen-2 and QuikBird) and STF data (CIA and LGC) demonstrate that the proposed CTUGAN model outperforms both subjective and objective evaluations.
Article
Full-text available
The quantitative assessment and spatial representation of wetland carbon storage, which play a critical role in the global carbon cycle and human production, can provide useful data and knowledge for decision-making in achieving sustainable development goals (SDGs). Currently, human activities and climate change impacts pose a challenge for the assessment of wetland carbon storage in coastal urban clusters. We proposed a "past-present-future" long time series refined wetland carbon storage assessment model using Guangxi Beibu Gulf (GBG) and Guangdong, Hong Kong, Macao and the Greater Bay Area (GBA) as the study area. The CLUE-S and InVEST models were coupled to conduct a comparative analysis of the spatial and temporal changes in wetland carbon storage and the spatial identification of damages from 1990 to 2035 and finally explore the sensitivity of wetland changes to carbon storage and quantitatively assess the SDG15.1 target. The results showed that (1) both urban clusters are characterized by many reservoirs/farming ponds, large river areas and few lakes. 1990-2035 rivers, shallow waters and mudflats have a decreasing trend to be distributed in the middle of their respective regions, mangroves are on an increasing trend, GBG is mainly distributed in the Maowei Sea and GBA is mainly distributed in Shenzhen Bay. (2) Wetland carbon storage of the two urban clusters show an overall fluctuating downward trend, with rivers, lakes and beaches all showing a downward trend. The multiyear average carbon storage of the GBG are 3.2 times higher than those of the GBA. In ecological protection scenario (EPS) policy planning, it is reasonable to help wetland carbon sequestration in coastal urban clusters. (3) The trend of wetland change from 1990 to 2020 was positive for carbon storage. The rate of recovery of wetland carbon stocks is lower in GBA than in GBG under the natural increase scenario (NIS) and the ecological protection scenario (EPS). The economic development scenario (EDS) contributes least to the realisation of SDG15.1 for the coastal urban agglomeration. The ecological protection scenario (EPS) contributes the most to the realisation of SDG15.1 for the coastal urban agglomeration.
Article
Full-text available
To map and manage forest vegetation including wetland communities, remote sensing technology has been shown to be a valid and widely employed technology. In this paper, two ecologically different study areas were evaluated using free and widely available high-resolution multispectral National Agriculture Imagery Program (NAIP) and ultra-high-resolution multispectral unmanned aerial vehicle (UAV) imagery located in the Upper Great Lakes Laurentian Mixed Forest. Three different machine learning algorithms, random forest (RF), support vector machine (SVM), and averaged neural network (avNNet), were evaluated to classify complex natural habitat communities as defined by the Michigan Natural Features Inventory. Accurate training sets were developed using both spectral enhancement and transformation techniques, field collected data, soil data, texture, spectral indices, and expert knowledge. The utility of the various ancillary datasets significantly improved classification results. Using the RF classifier, overall accuracies (OA) between 83.8% and 87.7% with kappa (k) values between 0.79 and 0.85 for the NAIP imagery and between 87.3% and 93.7% OA with k values between 0.83 and 0.92 for the UAV dataset were achieved. Based on the results, we concluded RF to be a robust choice for classifying complex forest vegetation including surrounding wetland communities.
Article
Full-text available
Frequent observations of surface water at fine spatial scales will provide critical data to support the management of aquatic habitat, flood risk and water quality. Sentinel-1 and Sentinel-2 satellites can provide such observations, but algorithms are still needed that perform well across diverse climate and vegetation conditions. We developed surface inundation algorithms for Sentinel-1 and Sentinel-2, respectively, at 12 sites across the conterminous United States (CONUS), covering a total of >536,000 km2 and representing diverse hydrologic and vegetation landscapes. Each scene in the 5-year (2017-2021) time series was classified into open water, vegetated water, and non-water at 20 m resolution using variables from Sentinel-1 and Sentinel-2, as well as variables derived from topographic and weather datasets. The Sentinel-1 algorithm was developed distinct from the Sentinel-2 model to explore if and where the two time series could potentially be integrated into a single high-frequency time series. Within each model, open water and vegetated water (vegetated palustrine, lacustrine, and riverine wetlands) classes were mapped. The models were validated using imagery from WorldView and PlanetScope. Classification accuracy for open water was high across the 5-year period, with an omission and commission error of only 3.1% and 0.9% for the Sentinel-1 algorithm and 3.1% and 0.5% for the Sentinel-2 algorithm, respectively. Vegetated water accuracy was lower, as expected given that the class represents mixed pixels. The Sentinel-2 algorithm showed higher accuracy (10.7% omission and 7.9% commission error) relative to the Sentinel-1 algorithm (28.4% omission and 16.0% commission error). Patterns over time in the proportion of area mapped as open or vegetated water by the Sentinel-1 and Sentinel-2 algorithms were charted and correlated for a subset of all 12 sites. Our results showed that the Sentinel-1 and Sentinel-2 algorithm open water time series can be integrated at all 12 sites to improve the temporal resolution, but sensor-specific differences, such as sensitivity to vegetation structure versus pixel color, complicate the data integration for mixed-pixel, vegetated water. The methods developed here provide inundation at 5-day (Sentinel-2 algorithm) and 12-day (Sentinel-1 algorithm) time steps to improve our understanding of the short- and long-term response of surface water to climate and land use drivers in different ecoregions.
Article
Coastal salt marshes suffering from anthropogenic coastal development and sea level rise have attracted much attention because of their capacity for carbon sequestration and global climate change mitigation. Accurate mapping of coastal salt marshes is always the first step for their protection, management, and restoration. The inherent complexities of vegetation, dynamics of tides, and anthropogenic disturbances pose challenges for remote sensing-based approaches. Existing studies have utilized phenology information and various black-box algorithms to reduce misclassifications. However, the approaches with dual-temporal images containing phenology information have suffered from inefficiency; the misclassifications have been objectively post-processed rather than considered in the method design, and the tacit knowledge of the trained black-box models has not been revealed. To address the above issues, we proposed a new approach with solid improvements in dual-temporal image construction, misclassification processing, and tacit knowledge analysis, including an efficient method to synthesize dual-temporal images based on the common divisor of the green-up season or senescence season resulting from latitudinal gradients in narrow coastal areas of China, a detailed classification scheme by interpretation of iteratively collected samples, and a key decision rule approximating the trained model. We applied the approach to Sentinel-1/2 images and DEM data, thus deriving a 10-m resolution coastal salt marsh map of China with an overall accuracy of 92.5%. A qualitative comparison showed that the map produced in this study was fitted well with actual salt marsh distributions, resulting in improved accuracy when compared to recently generated maps. The most important contribution is that the overall nature of the trained model observed from the training samples was approximated by a four-feature decision rule following the principle of explainable artificial intelligence, i.e., B8 senescence /B4 senescence < 2.06 & B4 green /B8 green < 0.78 & B12 green /B11 green < 0.72 & elevation < 2.13, and thus establi