International Journal of Environment, Agriculture and Biotechnology

Vol-8, Issue-4; Jul-Aug, 2023

Peer-Reviewed International Journal
Journal Home Page Available: https://ijeab.com/
Journal DOI: 10.22161/ijeab

Spatial-Temporal Evolution Characteristics of Vegetation

Coverage and Urbanization Expansion in Dongguan Based
on Remote Sensing
Qiuhui Zhang, Ruei-Yuan Wang*

School of Sciences, Guangdong University of Petrochem Technology (GDUPT), Maoming 525000, China
*Corresponding author

Received: 24 Jun 2023; Received in revised form: 18 Jul 2023; Accepted: 27 Jul 2023; Available online: 01 Aug 2023
©2023 The Author(s). Published by Infogain Publication. This is an open access article under the CC BY license
(https://creativecommons.org/licenses/by/4.0/).

Abstract— The purpose of this study is to use the Landsat satellite image of Dongguan from 2005 to 2021,
combining the dimidiate pixel model and stochastic matrix method, to calculate the vegetation coverage
grade and land use type area, express the urbanization process with construction land, and analyze the
relationship between the temporal and spatial evolution characteristics of vegetation coverage and
urbanization expansion. The results indicate that: (1) The overall distribution of vegetation coverage is
consistent, showing a spatial distribution characteristic of "high in the east, low in the west, and lowest in
the north". (2) During this period, the overall coverage area of medium low, medium, and medium high
vegetation decreased, while both low and high vegetation coverage areas increased. (3) The area of
construction land has increased significantly, with the increased built-up area coming from bare land,
followed by vegetation and water bodies. (4) The spatio-temporal evolution characteristics of vegetation
coverage are closely related to urbanization expansion. This study can provide reasonable scientific data
support and an effective decision-making basis for evaluating the regional ecological environment,
construction, and urban development of Dongguan City.
Keywords— Dimidiate Pixel Model (DPM); Support Vector Machine (SVM); Land Use Type;
Urbanization; Stochastic Matrix; Remote Sensing

I. INTRODUCTION amount of rural labor is transferred to secondary and

With the continuous expansion of China's tertiary industries, resulting in idle rural land and the
urbanization, many problems have arisen. The process has restoration of forests.
negative and positive externalities for vegetation coverage. In addition, with economic growth and residents'
Many regions are occupying a large amount of ecological awareness of environmental protection rising, cities can
land during the rapid urbanization process, with some of it increase and restore vegetation resources through
being transformed into impermeable surfaces to support institutional and policy reform measures, making
human habitation, roads, and other infrastructure. urbanization and vegetation coverage develop in harmony.
Meanwhile, during the process of urbanization, a large Although urbanization brings burdens, it will also bring

ISSN: 2456-1878 (Int. J. Environ. Agric. Biotech.)

https://dx.doi.org/10.22161/ijeab.84.7 51
Zhang and Wang International Journal of Environment, Agriculture and Biotechnology, 8(4)-2023

positive externality to vegetation coverage [1] Its climate conditions are a subtropical monsoon
In this context, this paper takes Dongguan City, climate with a long summer without winter, an annual
Guangdong Province, as a confirmatory research area, and average temperature of 23.3 °C, abundant rainfall, and an
uses the Landsat data from 2005 to 2021 to carry out annual precipitation of 2042.6 mm. Topographic
research on the estimation of vegetation coverage in the conditions are mainly hilly tableland and Alluvial plains.
city, to quantitatively analyze its spatial and temporal The mountains in the southeast are large and concentrated,
distribution and change characteristics, and then to explore with an altitude of 200–600 meters, a gradient of about 30°,
the correlation between the change trend of vegetation and large fluctuations. The central and southern regions are
coverage and the level of urbanization. So as to clarify the mainly composed of low mountains and hills. The
spatial and temporal change law of vegetation and the northeast is dominated by Alluvial plains with an altitude
pattern of urbanization expansion, which is the regional of 30–80 meters. The slope is relatively small, and the
ecological environment construction and urban terrain undulates gently (Figure 1). The Forest cover of the
development of Dongguan City, provide reasonable and city reached 37.4%, and the greening coverage of the
scientific data support and an effective basis for central urban area reached 47.51%.
decision-making. Throughout the history of Dongguan's urban
development, in the past 40 years, it has developed from
II. STUDY AREA AND DATA SOURCES an agricultural county with a gross regional product of
2.1 Study Area only 600 million yuan to a global manufacturing base for
2
The total area of Dongguan City is about 2542.67 km , multinational companies with a total economic volume of
between 113° 31′ -114° 15′ E and 22° 39′ -23° 09′ N. more than 700 billion yuan. It is the first of the "Four
Located in the south-central part of Guangdong Province, Tigers of Guangdong", known as the "World Factory", and
on the east bank of the Pearl River Estuary and the Pearl an important transportation hub and foreign trade port in
River Delta at the lower reaches of the Dongjiang River, it Guangdong. This process has made the process of
borders Huizhou City in the east, Zengcheng District of urbanization and change dramatic. By the end of 2022,
Guangzhou City in the north, Panyu District of Guangzhou Dongguan will have a permanent resident population of
City across the sea in the west, and Bao'an District of 10.437 million, including 9.6281 million urban residents,
Shenzhen City in the south. It is an important waterway with an urbanization rate of 92.25%. The constructed area
between Guangzhou and Hong Kong. It currently governs will expand from less than 5 square kilometers to more
32 townships (4 streets, 28 towns), 546 village committees, than 900 square kilometers today, surpassing many
and 132 neighborhood committees.. provincial capitals and ranking on the new front line.

Fig.1 Location Map of Dongguan City

ISSN: 2456-1878 (Int. J. Environ. Agric. Biotech.)

https://dx.doi.org/10.22161/ijeab.84.7 52
Zhang and Wang International Journal of Environment, Agriculture and Biotechnology, 8(4)-2023

2.2 Data Sources spatial resolution of 30 m, come from the Chinese

In this study, remote sensing images, including Academy of Sciences geospatial data cloud
Landsat 5 TM remote sensing imagery from 2005 and (http://www.gscloud.cn/search), and the research data
2009 and Landsat 8 OLI from 2015 and 2021, with a interval is 5 years, as shown in Table 1.
Table 1 Landsat TM/OLI Satellite Imagery Information List
Landsat images Data Number Acquire date Resolution Cloud
cover/%

Landsat5 TM LT51220442005327BJC02 2005-11-23 30m 0.22

Landsat5 TM LT51220442009034BKT01 2009-02-03 30m 0.17
Landsat8 OLI LC81220442015291LGN01 2015-10-18 30m 0.58
Landsat8 OLI LC81220442021003LGN00 2021-01-03 30m 2.88

2.3 Data Preprocessing spatio-temporal evolution characteristics of land

For the remote sensing image data of the Dongguan use types.
area, ENVI 5.6 software was used for pre-processing, 2. Then, the image NDVI data was extracted, and
including band superposition, radiometric calibration, the Dimidiate Pixel Model (DPM) was used to
atmospheric correction, image mosaic, and mask, to obtain calculate the vegetation coverage, dividing their
the remote sensing image of the study area that meets the standard into five levels in this study area and
requirements. Based on this, Normalized Difference calculating the transfer matrix of the five levels,
Vegetation Index (NDVI), vegetation coverage, and which analyzed the spatio-temporal evolution
supervision classification were extracted. characteristics of vegetation coverage.
In atmospheric correction, the FLAASH processing 3. Finally, the urbanization process of Dongguan
function was used to perform radiometric calibration on city is represented by the change in construction
the four remote sensing images from 2005 to 2021, land area, and the spatio-temporal evolution
converting the brightness gray value of the images into characteristics of vegetation coverage and
absolute radiation brightness. Then the vector boundary of urbanization expansion are analyzed by
the study area was used to crop out the study area in the combining the urbanization process and the
image. spatio-temporal evolution characteristics of
vegetation coverage.
III. METHODOLOGY 3.2 Support Vector Machine (SVM)
3.1 Method SVM classification is a kind of binary classification
This study is based on Landsat-5 and Landsat-8 model that has many unique advantages in solving small
remote sensing images of Dongguan city, and its main sample, nonlinear, and high-dimensional pattern
analysis steps (Figure 2) are as follows: recognition problems. To date, SVM have been widely
1. First, the images were classified into four used to solve the problem of supervised classification of
categories: vegetation, construction land, bare high-dimensional data, especially for high-dimensional
land, and water bodies by Support Vector feature space and large data volume problems such as
Machine (SVM) classification. Then calculate hyper-spectral data.
the areas of these four categories and their
transfer matrix, which analyzes the

ISSN: 2456-1878 (Int. J. Environ. Agric. Biotech.)

https://dx.doi.org/10.22161/ijeab.84.7 53
Zhang and Wang International Journal of Environment, Agriculture and Biotechnology, 8(4)-2023

.
Fig.2 The Schema of This Study

The core idea of the SVM is based on the idea of which can be expressed as: 1/‖w‖, and the point is called
structural risk minimization as the induction principle. By the support vector. The sum of the minimum distance
nonlinear mapping, the sample is projected to the between two support vectors of different types and the
high-dimensional feature space, constructing the optimal hyperplane is 2/‖w‖, so finding the optimal hyperplane is to
classification surface with the lowest VC (Vapnik maximize the interval, which is minimized in the form, as
Chervonenkis) dimension as possible in the shown in (3):
high-dimensional space, which minimizes the upper bound 1
𝑚𝑖𝑛 ‖𝑤‖2 (3)
of classification risk so that the classifier has the optimal 2
generalization ability for unknown samples. Due to the Through the constraints of equation (2), the
fact that statistical learning theory has a complete expression of the optimization problem is obtained, as
theoretical system and solid mathematical foundation, and shown in (4):
the vector holding machine generated by it has strong 1
𝑚𝑖𝑛𝐿(𝑤, 𝑏, 𝛼) = ‖𝑤‖2
generalization ability, the theory and SVM have become a 2
new research hotspot in the field of machine learning 𝑁

around the world. The main principles are as follows [2]: − ∑ 𝛼𝑖 [𝑦𝑖 (𝑤 ∙ 𝑥𝑖 + 𝑏) − 1] (4)
If the vector x is linearly divisible in M-dimensional 𝑖=1

space, there is at least one hyperplane, dividing x into two After taking the partial derivative and setting it to 0,
types: -1 and 1, and then the hyperplane can be represented the constraint condition is obtained, as shown in (5):
as shown in (1): 𝑁

𝑓(𝑥) = 𝑤 ∙ 𝑥 + 𝑏 (1) ∑ 𝛼𝑖 𝑦𝑖 = 0 (5)

Let the mark of xi be yi, if the classification is correct, 𝑖=1

as shown i n (2): Obtaining, as shown in (6):

𝑦𝑖 (𝑤 ∙ 𝑥𝑖 + 𝑏) ≥ 1; 𝑖 = 1,2,3 ⋯ , 𝑁 (2) 𝑁

The minimum distance between the point closest to 𝑤 = ∑ 𝛼𝑖 𝑦𝑖 𝑥𝑖 (6)

the hyperplane and the hyperplane is called the interval, 𝑖=1

ISSN: 2456-1878 (Int. J. Environ. Agric. Biotech.)

https://dx.doi.org/10.22161/ijeab.84.7 54
Zhang and Wang International Journal of Environment, Agriculture and Biotechnology, 8(4)-2023

After equation s (5) and (6) are substituted into than or equal to", "le" is "less than or equal to", and "gt" is
equations (4), the dual optimization problem is obtained, "greater than". After outlier processing, the gray scale
as shown in (7): images of the NDVI in each year of Dongguan City are
𝑁 𝑁 𝑁 finally obtained.
1
𝑄(𝛼) = ∑ 𝛼𝑖 + ∑ ∑ 𝛼𝑖 𝛼𝑗 𝑦𝑖 𝑦𝑗 (𝑥𝑖 ∙ 𝑥𝑗 ) (7) Some studies indicate that the NDVI value of the
2
𝑖=1 𝑖=1 𝑗=1 vegetated area is greater than zero, and the higher the
Equation (7) can be solved by convex quadratic vegetation coverage, the greater the NDVI value. On rocks,
programming method, and the category decision function bare soil, and other surfaces, the reflectivity of visible light
is finally obtained, as shown in (8): and near infrared is similar, and the NDVI value is about
𝑁 zero. On water vapor, clouds, snow, and other surface
𝑓(𝑥) = ∑ 𝛼𝑖 𝑦𝑖 𝐾(𝑥𝑖 ∙ 𝑥) + 𝑏 (8) objects, the reflectivity of the visible light band is higher,
𝑖=1 and the NDVI value is usually less than zero. Therefore,
The above is a linear and nonlinear hard classification NDVI is suitable for large-scale vegetation monitoring and
problem, and the position of the hyperplane is strictly has a good indicator [3].
fixed. Considering that a few points cannot be classified 3.4 Dimidiate Pixel Model (DPM)
by hyperplane, by adding the relaxation variable ξ and the Dimidiate Pixel Model (DPM) is a commonly used
penalty factor C, equation (2) becomes: method to calculate the fraction of vegetation cover (FVC),
𝑦𝑖 (𝑤𝑥𝑖 + 𝑏) ≥ 1 − 𝜉𝑖 ; 𝑖 = 1,2, ⋯ , 𝑁 (9) which is one of the more widely used remote sensing
Equation (3) becomes, as shown in (10): estimation models. The model assumes that the ground
𝑁 area represented by each pixel is composed of two parts
1
𝑚𝑖𝑛 ‖𝑤‖2 + 𝐶 ∑ 𝜉𝑖 (10) covered by vegetation and no vegetation, and the reflection
2
𝑖=1 spectrum information of ground objects received by the
3.3 Normalized Difference Vegetation Index (NDVI) sensor is also formed by the superposition and
Based on the NDVI calculation tool in ENVI 5.6 combination of different reflection spectrum information
software, the NDVI was extracted from the pre-processed from these two parts. The weight of reflected spectral
Landsat images covering the study area in four phases in information of objects in different places represents the
2005, 2009, 2015, and 2021. The calculation equation is as proportion of objects in this pixel, so vegetation coverage
follows: can be used to represent the proportion of vegetation [4-6].
(𝑁𝐼𝑅 − 𝑅𝐸𝐷) NDVI data were used to calculate the vegetation
𝑁𝐷𝑉𝐼 = (11)
(𝑁𝐼𝑅 + 𝑅𝐸𝐷) coverage, and NDVI values with cumulative frequencies
In equation (11), NIR is the reflectivity of the near of 95% and 5% were used as the two extreme values of
infrared band (%), and RED is the reflectivity of the red NDVI to represent the NDVIveg value and NDVIsoil
band (%). value. When NDVI is less than NDVIsoil , the value of
Statistical analysis indicates that the calculated NDVI VFC is 0. When NDVI is greater than NDVIveg , the value
value is not within the range [-1, 1], which is because of VFC is 1. When the value of NDVI is greater than or
remote sensing images are affected by aerosol, shadow, equal to NDVIsoil and less than or equal to NDVIveg , that
and other factors, and excessive atmospheric correction is to say, when it is between the NDVIveg value and
leads to overflow phenomena. It is therefore necessary to NDVIsoil value, calculate equation (13). Using DPM to
normalize the NDVI value, that is to say, remove outliers. estimate the vegetation coverage equation, the calculation
The calculation equation is as follows: equation is as follows:
(𝑏1 𝑙𝑡 − 1) ∗ (−1) + (𝑏1 𝑔𝑡 1) ∗ 1 (𝑏1 𝑙𝑡 𝑁𝐷𝑉𝐼𝑠𝑜𝑖𝑙 ) ∗ 0 + (𝑏1 𝑔𝑡𝑁𝐷𝑉𝐼𝑣𝑒𝑔 ) ∗ 1 +

+ (𝑏1 𝑔𝑒 − 1 𝑎𝑛𝑑 𝑏1 𝑙𝑒 1) ∗ 𝑏1 (12) (𝑏1 𝑔𝑒 𝑁𝐷𝑉𝐼𝑠𝑜𝑖𝑙 𝑎𝑛𝑑 𝑏1 𝑙𝑒 𝑁𝐷𝑉𝐼𝑣𝑒𝑔 ) ∗ (𝑏1 − 𝑁𝐷𝑉𝐼𝑠𝑜𝑖𝑙 )⁄( 𝑁𝐷𝑉𝐼𝑣𝑒𝑔 − 𝑁𝐷𝑉𝐼𝑠𝑜𝑖𝑙 ) (13)

In equation (12), "b1" is the NDVI value of the In equation (13), NDVIsoil is the NDVI value of bare
remote sensing image, "lt" is "less than", "ge" is "greater land or non-vegetated area. NDVIveg is the NDVI value

ISSN: 2456-1878 (Int. J. Environ. Agric. Biotech.)

https://dx.doi.org/10.22161/ijeab.84.7 55
Zhang and Wang International Journal of Environment, Agriculture and Biotechnology, 8(4)-2023

of the pixel in the area completely covered by vegetation. obtained according to the change in land cover status at
3.5 Vegetation Coverage Classification different times and phases in the same area, and the state
In order to better quantify the characteristics of and status of the system are described quantitatively
vegetation coverage evolution, this paper divides the [10-15]. The specific calculation equation is as follows:
vegetation coverage in this study area into 5 grades 𝑆𝑖𝑗
𝑃𝑖𝑗 = 𝑛 (13)
according to 《the Classification and Grading Standard of ∑𝑖=1 ∑𝑛𝑗=1 𝑆𝑖𝑗
Soil Erosion》 promulgated by the Ministry of Water In equation (13), Pij is the conversion probability of
Resources of the People's Republic of China in 2008. The land use type i to land use type j; Sij is the area of land
criteria are as follows: <30% (low coverage), 30%-45% use type i transformed into land use type j (km2).
(medium to low coverage), 45%-60% (medium coverage), The land use transfer matrix model [11] is:
60%-75% (medium to high coverage) and > 75% (high 𝑃11 ⋯ 𝑃1𝑗
coverage). [ ⋮ ⋱ ⋮ ] (14)
3.6 Land Use Transfer Matrix 𝑃𝑖1 ⋯ 𝑃𝑖𝑗

The main purpose of this paper is to study the

relationship between vegetation coverage and urbanization IV. ANALYSES AND RESULT
expansion. Thus, according to the image of Dongguan, 4.1 Spatiotemporal Evolution of Vegetation Coverage
land use types are defined as vegetation, construction land, The analysis data indicate that the overall distribution
bare land, and water substances. Vegetation includes of vegetation coverage in Dongguan is basically the same,
forestland, arable land, and grassland; construction land showing the spatial distribution characteristics of "high in
includes buildings, transportation facilities, and roads; bare the east-low in the west -the lowest in the north" (Figure 3).
land includes bare soil surface with less vegetation The eastern part of the mountain is huge, concentrated into
coverage, such as wasteland, abandoned arable, and quarry; pieces with relatively large ups and downs, forming a very
and water substance includes lakes, reservoirs, sea areas, obvious high vegetation belt, including Xiegang town,
and wetlands. The change in construction land area Qingxi town, Huangjiang Town and other areas. Among
indicates the urbanization process in Dongguan. them, Guanyin Mountain, Yinping Mountain, Dawang
In this study, the land use transfer matrix is used for Mountain, Baoshan, Nanmen Mountain and other areas of
dynamic monitoring and analysis of land use, which can the original forest are the main reasons for the formation of
not only quantitatively and accurately indicate the high vegetation areas. The distribution of middle, high and
transformation and change of different types of land use medium vegetation areas is relatively dispersed, mainly in
[7-9], but also quantitatively reveal the change rate of land the west and south, and the proportion of low vegetation
transfer among different types of comprehensive land use areas is relatively large.
transfer types. In other words, a two-dimensional matrix is

ISSN: 2456-1878 (Int. J. Environ. Agric. Biotech.)

https://dx.doi.org/10.22161/ijeab.84.7 56
Zhang and Wang International Journal of Environment, Agriculture and Biotechnology, 8(4)-2023

Fig.3 Vegetation Coverage Grade Map of Dongguan from 2005 to 2021

From 2005 to 2021, the area of low-vegetation area in increased areas of high vegetation are mainly located in
Dongguan increased significantly, with an increase of Huangjiang Town and Tangxia Town in the south in 2021,
2
about 92.14 km , and the area of high-vegetation area also while the increased areas of low vegetation are located in
increased, but the increase was relatively small, with an the west and central north, as shown in Table 2.
2
increase of about 13.36 km . Compared with 2005, the
Table 2 Grade of Vegetation Coverage in Dongguan from 2005 to 2021 (Unit: km2)
Year Low Rate Medium Rate Medium Rate Medium Rate High Rate
/% -low /% % -high /% /%
2005 993.98 40.53 311.28 12.69 261.20 10.65 249.60 10.13 637.70 26.00
2009 1152.70 47.00 352.58 14.35 249.62 10.13 210.31 8.58 489.60 19.94
2015 880.18 35.82 327.56 13.35 253.08 10.30 225.63 9.20 768.40 31.33
2021 1086.12 44.28 292.78 11.94 226.41 9.23 196.50 8.01 651.06 26.54

According to the distribution results of vegetation area of high vegetation conversion was 51.51 km2. In 2005,
coverage in 2005 and 2021, the transfer matrix can be the area of mid-high vegetation transferred out was 209.54
calculated, and the mutual transformation of vegetation km2, and in 2021, the area of low and high vegetation
coverage at all levels can be obtained by analyzing the transferred in was 56.95 km2, while the area of high
transfer matrix. vegetation transferred in was 88.05 km2. In 2005, the area
In Dongguan, the area of medium and low vegetation of low vegetation conversion was 250.11 km2. The area of
transferred out was 247.35 km2 in 2005, mainly to low and high vegetation transfer was 179.39 km2. In 2021, the area
medium vegetation; in 2021, the area of low vegetation transferred from low to medium vegetation is 228.86 km2,
transferred in was 140.87 km2, and the area of medium The transfer area of moderate vegetation is 218.03 km2,
2
vegetation transferred in was 46.63 km . In 2005, the area The transfer area of medium to high vegetation is 156.44
2
of medium vegetation conversion was 218.03 km , which km2, The area of low vegetation transfer in 2021 is 342.17
was converted to low and high vegetation in 2021. The km2, and the area transferred from high vegetation in 2021
area of low vegetation conversion was 83.88 km2, and the is 193.70 km2, as shown in Table 3.

ISSN: 2456-1878 (Int. J. Environ. Agric. Biotech.)

https://dx.doi.org/10.22161/ijeab.84.7 57
Zhang and Wang International Journal of Environment, Agriculture and Biotechnology, 8(4)-2023

The analysis shows that more vegetation has areas is increasing, which may be related to human
deteriorated from low-middle, middle, and middle-high activities. For example, with the increase in population and
vegetation to low vegetation, while less vegetation has the needs for economic and social development, cities
improved from low-middle, middle, and middle-high continue to expand, arable land is converted into
vegetation to high vegetation since 2005. During the construction land, and arable land and forest land are
period from 2005 to 2021, the area of middle-low, middle, reclaimed and developed, which results in the destruction
and middle-high vegetation decreased generally, while the and degradation of the original vegetation. On the other
area of low vegetation continued to increase with a hand, climate change may lead to changes in precipitation
relatively large increase and the area of high vegetation and temperature, affecting vegetation growth. Drought,
increased generally with a relatively small increase. reduced water resources, and high temperatures may lead
In general, the vegetation cover from 2005 to 2021 is to the apoptosis and degradation of vegetation.
decreasing, and the proportion of low vegetation cover
Table 3 Analysis of Transfer Matrix of Vegetation Coverage in Dongguan (Unit: km2)
Year Index 2021 year
Low High Medium Medium-low Medium-high
Low 743.60 25.85 64.67 123.95 35.64
High 60.47 457.18 37.53 31.64 49.76
2005 Medium 83.88 51.51 43.12 43.16 39.49
year Medium-low 140.87 28.29 46.63 63.88 31.56
Medium-high 56.95 88.05 34.43 30.11 40.01

4.2 Spatiotemporal Evolution of Land Use Types distributed in the northwest, including the Shima River and
In terms of the overall distribution of land use types Dongjiang River, and lakes and reservoirs are scattered
in Dongguan, vegetation dominates the south and east of and distributed in areas with high terrain and vegetation
the city (Figure 4), and water substance is divided into coverage. The sea areas are mainly in the Lion Ocean and
rivers, lakes, reservoirs, and some sea areas. Rivers are distributed in the west.

Fig.4 Distributions of Land Use Types in Dongguan from 2005 to 2021

ISSN: 2456-1878 (Int. J. Environ. Agric. Biotech.)

https://dx.doi.org/10.22161/ijeab.84.7 58
Zhang and Wang International Journal of Environment, Agriculture and Biotechnology, 8(4)-2023

According to the area changes of various types of During 2009-2015, the vegetation area recovered
land use (Figure 5), from 2005 to 2021, it can be seen that somewhat, the increase rate of urban construction area
the area of construction land in Dongguan increased from slowed down, and the bare land area decreased, indicating
2 2
1092.43 km to 1453.04 km , with a steady and large that the urban construction activities in this period mainly
increase in area. And the increased built-up area is mainly used the original bare land, and the vegetation covered
bare land; the second is vegetation, which is mainly from area recovered somewhat.
arable land and forest land; and the last is water, which is During the period from 2015 to 2021, the built-up
mainly from wetlands. area increased rapidly again, the bare land area continued
During the period from 2005 to 2009, the vegetation to decrease, and the area covered by medium and low
area decreased significantly, and the built-up area and bare vegetation decreased, indicating that the urban expansion
soil area continued to increase, indicating that sustained phenomenon of occupying arable and grassland appeared
and intense urban construction activities were carried out again.
during this period.

Fig.5 Area Changes of Land Use Types in Dongguan in Different Years

The land use status of Dongguan from 2005 to 2021 the water transfer area was 18.84 km 2, and the construction
includes the area of bare land transferred out, which is land transfer area was 519.21 km2, as shown in Table 4.
2
256.29 km , and the area converted into construction land, Among them, bare land has the largest net change,
2
which is 222.87 km . The area of vegetation transferred followed by vegetation and water bodies. The shift from
2
out is 249.49 km , and the area converted to construction bare land to construction land is the most obvious feature
2
land is 233.63 km . Simultaneously, the area of water of the urbanization process. The shift of vegetation
2
substance transferred out is 78.74 km , and the area towards construction land reflects the process of urban
2
converted into construction land is 62.71 km . The area of construction at the cost of occupying agricultural land and
2
construction land transferred out is 159.64 km . logging forest land. The diversion of water bodies towards
In addition, in 2021, the bare land transfer area was construction land mainly extends to wetlands, lakes, and
2 2
21.94 km , the vegetation transfer area was 183.18 km , other water bodies through land reclamation.

ISSN: 2456-1878 (Int. J. Environ. Agric. Biotech.)

https://dx.doi.org/10.22161/ijeab.84.7 59
Zhang and Wang International Journal of Environment, Agriculture and Biotechnology, 8(4)-2023

Table 4 Transfer Matrix of Land Use Types in Dongguan (Unit: km2)

Year Index 2021 year

Construction Bare land Water substance Vegetation

land
Construction land 933.68 8.13 12.85 137.66
2005
Bare land 222.87 4.12 1.07 32.35
year
Water substance 62.71 2.7 147.42 13.17

Vegetation 233.63 10.93 4.93 623.79

4.3 Correlation Analysis of Vegetation Coverage and of middle and high vegetation transferred was 119.55 km2,
Urbanization and the total area of high vegetation transferred was
Objectively, there is a close relationship between 112.67 km2. At this stage, with the destruction of the
urbanization and the ecological environment. Due to the ecological environment and the deterioration of the city's
transfer of rural populations to cities and the continuous internal environment, Dongguan's economic development
expansion of urban areas, land resources in natural was affected. In the meantime, human beings realized the
ecological space must be occupied, resulting in changes in impact of the environmental destruction and began to
vegetation and the ecological environment, which puts restore the ecological environment, so that the vegetation
forward higher requirements for urban environmental area recovered somewhat, the increase rate of urban
protection. Simultaneously, the increase in economic construction area slowed down, and the urbanization
aggregate makes cities more capable of environmental process slowed down. In this period, urban construction
protection investment and protects vegetation resources to activities mainly used the original bare land; the bare land
a certain extent. Therefore, urbanization has a dual effect area was greatly reduced, and the vegetation coverage area
on vegetation resources. was restored.
According to the vegetation coverage transfer matrix During the period from 2015 to 2021, the area of low
analysis, during the period from 2005 to 2009, the area of vegetation and high vegetation continued to increase, due
low vegetation and middle and low vegetation continued to the transformation of middle and low vegetation into
to increase, mainly from medium vegetation to low low vegetation and the area of low vegetation into 162.48
vegetation and high vegetation. The area transferred by km2. During this period, the built-up area increased rapidly
low vegetation was 65.85 km2, and the area transferred by again, the bare land area continued to decrease, the middle
high vegetation was 75.70 km2. In this stage, the area of and low vegetation areas decreased, but the high
construction land and bare land continued to increase, and vegetation areas increased, indicating that population
the area covered by vegetation continued to decrease, growth required greater space and resources, and the urban
indicating that human beings carried out continuous and expansion phenomenon of occupying arable and grassland
intense urban construction activities at the cost of again appeared. However, during this period, Dongguan
destroying vegetation during this period, and the implemented many policies, such as returning arable to
urbanization process increased significantly, resulting in an forest, protecting natural forests, adjusting industrial
increase in the area of low-vegetation areas, that is to say, structure, and developing tourism such as ecological parks,
construction land and bare land. which should also pay attention to the protection of the
During the period from 2009 to 2015, the area of high ecological environment and reduce the destruction of
vegetation and middle and high vegetation continued to forest land in some natural ecological areas.
increase, from low vegetation and middle and low In summary, the urbanization level of Dongguan can
vegetation to middle and high vegetation, and the total area be used as one of the reference factors for the change of

ISSN: 2456-1878 (Int. J. Environ. Agric. Biotech.)

https://dx.doi.org/10.22161/ijeab.84.7 60
Zhang and Wang International Journal of Environment, Agriculture and Biotechnology, 8(4)-2023

vegetation coverage, but the level of urbanization cannot coverage in Dongguan are closely related to urbanization
be used as an absolute influencing factor to measure the expansion. On the one hand, urbanization will destroy
change of urban vegetation coverage, and comprehensive vegetation in various ways. For example, urbanization
analysis should be carried out in combination with other increases the area of construction land, leading to the loss
influencing factors. of arable land and forest land. On the other hand, with the
acceleration of urbanization, the state will also implement
V. CONCLUSION various measures to protect the ecological environment,
In this study, the DPM and transfer matrix were used such as returning arable to forest, adjusting industrial
to estimate the vegetation coverage of Dongguan, structure, and developing some tourism to protect forest
analyzing the area proportion and change of vegetation land.
coverage area at all levels. Meanwhile, the transfer matrix Finally, this study only selected urbanization and
method of vegetation coverage can effectively analyze the vegetation coverage at all levels for comparative analysis,
mutual transformation of vegetation coverage area in each and did not consider other natural and social factors, such
year. Moreover, by analyzing the change trend of land type as natural disasters, population density, urban economic
area and the transfer matrix of each land type, the change development and other factors, which would also have an
index of construction land area was compared with the impact on vegetation coverage. Therefore, the subsequent
change degree of vegetation coverage, analyzing research will also incorporate these factors affecting the
correlation between the change degree of vegetation vegetation coverage into the decision-making of the
coverage and urbanization expansion. analysis.
This study provides a new way of thinking about the
spatio-temporal evolution characteristics of vegetation ACKNOWLEDGEMENTS
coverage in the past 15 years and the interannual variation The author is grateful for the research grants given to
law of vegetation coverage, which can also provide a Ruei-Yuan Wang from GDUPT Talents Recruitment
decision-making basis for ecological environment (No.2019rc098), and ZY Chen from Talents Recruitment
construction in Dongguan. The main conclusions are as of GDUPT (No. 2021rc002), in Guangdong Province,
follows: Peoples R China, and Academic Affairs in GDUPT for
(1) The overall distribution of vegetation coverage Goal Problem-Oriented Teaching Innovation and Practice
in Dongguan is basically the same, showing the spatial Project Grant No.701-234660.
distribution characteristics of "high in the east-low in the
west-the lowest in the north". REFERENCES
(2) During the period from 2005 to 2021, the [1] Shi, Y. Temporal and Spatial Evolution and Influence of
coverage area of middle-low, middle and middle-high Vegetation Cover in the Process of Urbanization in China,
vegetation will decrease overall, while the coverage area Northeast Forestry University, 2022.
of low vegetation will continue to increase with a [2] Sun, L. Research on Classification of Hyperspectral Images
relatively large increase, and the coverage area of high Based on Support Vector Machine, Harbin Engineering
vegetation will increase with a relatively small increase. University, 2011.
(3) During the period from 2005 to 2021, the area of [3] Liu, H. The Research on the Vegetation Coverage Dynamic
construction land in Dongguan will increase steadily and Change of Guangdong Province Based on RS and GIS,
greatly, and the increased built-up area is mainly bare land, South China Agricultural University, 2015.
followed by vegetation and water area. [4] Jiang, Q. Analysis of vegetation coverage change in Shanbei
(4) Through the combined analysis of the change area of Huainan City based on RS. Agriculture and
trend of vegetation coverage and land type area at all technology, 2021, 41(14):130-134.
levels from 2005 to 2021, which indicates that the [5] Wu, J., Li, H., and Lei, J. Remote sensing monitoring of
spatial-temporal evolution characteristics of vegetation spatiotemporal variation of vegetation cover in Dongjiang

ISSN: 2456-1878 (Int. J. Environ. Agric. Biotech.)

https://dx.doi.org/10.22161/ijeab.84.7 61
Zhang and Wang International Journal of Environment, Agriculture and Biotechnology, 8(4)-2023

source area, Journal of Jiangxi University of Science and

Technology, 2017, 38(01):29-36.
[6] Chen, F., and Wu, F. Analysis of temporal and spatial
evolution of vegetation coverage in Nanchang based on
NDVI, Changjiang Information & Communications,
2022,35(11):120-122.
[7] Liang, B., Lu, R., Li, G., and Li, Q. The Effects of Land Use
Carbon Emission Based on the Carbon Transfer Matrix —A
Case of Guangxi Beibu Gulf Economic Zone, Hubei
Agricultural Sciences, 2016,55(11):2779-2784.
[8] Zhang, Y., Zeng, J., Hu, X., Sun, D., and Song, Z. Study on
land－use change in midstream of Shule River based on
land use transfer matrix, Yangtze River, 2017,48(21):34-39.
[9] Qiao, W., Sheng, Y., Fang, B., and Wang, Y. Land use
change information mining in highly urbanized area based
on transfer matrix: A case study of Suzhou, Jiangsu Province,
GEOGRAPHICAL RESEARCH, 2013,(08):1497-1507.
[10] Chang, D., Cheng, A., Bai, T., and Jiao, Y. Characteristics
analysis of land use change in Tangjiahui Coal mine based
on transfer matrix. Journal of Green Science and Technology,
2021, 23(22):222-224.
[11] Zhu, H., and Li, X. Discussion on the Index Method of
Regional Land Use Change, ACTA EOPGRAPHIC SINICA,
2003,(05):643-650.
[12] Zhong, J., Huang, Y. and Guo, M. Spatio-temporal Change
Characteristics Analysis of Vegetation Coverage in Western
Hubei Province in Recent 20 Years. GEOSPATIAL
INFORMATION, 2023,21(02):50-55. doi:10.3969/j.issn.
1672-4623.2023.02.010
[13] Xiao, X., Li, J., Han, B., Lu, C., and Xue, B.
Spatial-temporal characteristics of vegetation coverage and
its correlation with urbanization in traditional industrial area
of Northeastern China, Ecological Science, 2017, 36(6):
71-77.
[14] Wang, Y., Zhang, X., and Shang, G. Analysis of Spatial and
Temporal Changes of Vegetation Fraction Coverage in
Beijing from 2013 to 2020, Journal of Hebei GEO
University,2022,(01):99-105.
[15] Li, Y., Lin, F., and Yan, Z. Analysis of Urbanization
Evolution Based on Landsat Remote Sensing Image,
Urbanism and Architecture,2019,16(25):138-142. DOI ：
10.19892/j.cnki.csjz.2019.25.030

ISSN: 2456-1878 (Int. J. Environ. Agric. Biotech.)

https://dx.doi.org/10.22161/ijeab.84.7 62