论文信息
论文解读
AI 辅助整理- 方法
- 土地利用转移矩阵、重心迁移模型与Getis-Ord Gi*热点分析,结合遥感与人
- 发现
- 人口增长热点与建设用地扩张轨迹基本吻合,三地呈现差异化的土地—人口联动机制。
- 机会点
- 可将该土地—人口耦合框架延伸至城乡结合部耕地占用与人口再分布协同监测研究。
论文摘要
摘要 香港、珠海、澳门大湾区城市群(HZM-GBAUC)共同构成世界上人口最密集、制度最多元的跨境都市群之一。本研究系统评估了2016至2025年间四个离散时间点上HZM-GBAUC的土地利用状况,并整合了四个相应时段的人口密度数据。本研究采用土地利用转移矩阵、空间重心迁移模型和Getis-Ord Gi*热点分析,通过ArcGIS 10.7平台对多源遥感土地利用数据和人口统计数据进行了定量分析。研究结果表明,珠海全域呈现高强度人口集聚;香港人口分布正从城市核心区向外围新界地区重新分布;澳门路氹填海区促进了人口向新开发建成区的定向转移。人口增长的热冷格局与HZM-GBAUC建成区扩张的空间轨迹大体吻合。本研究首次在统一的概念框架下考察了香港、珠海、澳门土地利用—人口关联的时空机制,为高密度跨境城市群在整合多元化住宅用地供应方面的全球治理提供了建议。
全文
Research SquareAssess land use changes and the impact on
population change across the Hong Kong-Zhuhai-
Macao Greater Bay Area Urban Clusters
Xiaotong Su
Macao Polytechnic University
Patrick Pang
Macao Polytechnic University
Ting Liu
Macao Polytechnic University
Yiming Taclis Luo
Macao Polytechnic University
Kin SunChan
University of Macau
Article
Keywords: Hong Kong-Zhuhai-Macao Greater Bay Area, Urban Clusters, Land use, Population dynamics,
Hot spot analysis
Posted Date: November 16th, 2026
DOI: https://doi.org/10.21203/rs.3.rs-11523017/v1
License: This work is licensed under a Creative Commons Attribution 4.0 International License.
Read Full License
Additional Declarations: No competing interests reported.
Assess land use changes and the impact on population change across
the Hong Kong-Zhuhai-Macao Greater Bay Area Urban Clusters
Xiaotong Sua., Patrick Panga.*,Ting Liua., Yiming Taclis Luoa. and Kin SunChanb.
a. Faculty of Applied Sciences, Macao Polytechnic University, Macao 999078, China;
b. Department of Government and Public Administration, Faculty of Social Sciences, University of
Macao
p2523603@mpu.edu.mo(X.S.);p2314982@mpu.edu.mo(T.L.);p2417889@mpu.edu.mo(Y .T.L.)
and kschan@um.edu.mo(K.S.C)
Correspondence: mail@patrickpang.net
Abstract
Hong Kong, Zhuhai, and Macao Greater Bay Area Urban Clusters (HZM-GBAUC )
collectively constitute one of the world's most densely populated and institutionally diverse cross-
boundary metropolitan clusters. This study systematically evaluates land use conditions across the
HZM-GBAUC at four discrete time points spanning 2016 to 2025, integrating four corresponding
periods of population density data. Employing a land use transfer matrix, a spatial center-of-gravity
migration model, and the Getis-Ord Gi hot spot analysis, this study conducts a quantitative analysis
of multi-source remote sensing land use data and demographic statistics through the ArcGIS 10.7
platform. The findings indicate that Zhuhai displays high-intensity population agglomeration across
its entire territory; Hong Kong's population distribution is redistributing outward from the urban
core toward the peripheral New Territories; and Macao's Cotai reclamation area has facilitated
directed population transfer into newly developed built zones. The hot-cold patterns of population
growth align broadly with the spatial trajectories of built-up land expansion across the HZM-
GBAUC. For the first time, under a unified conceptual framework, the spatio-temporal mechanisms
governing land–population linkages across Hong Kong, Zhuhai, and Macao are examined, offering
suggestions for the global governance of high-density cross-boundary urban agglomerations in
integrating diversified residential land supply.
Keywords: Hong Kong-Zhuhai-Macao Greater Bay Area; Urban Clusters; Land use; Population
dynamics; Hot spot analysis
- Introduction
In recent years, global research on land use change has shifted from quantitative expansion
towards examining spatial quality [1]. Under the dual pressures of the global energy transition and
food security, competition for land resources and land scarcity crises have intensified. The latest
research focus on the interplay between emissions reduction targets, population carrying capacity,
and food security [1].
Traditional models of population–land interaction have largely emphasised the linear
relationship in which population growth drives the expansion of urban construction land [3].
However, recent empirical studies demonstrate that in highly urbanised or intensively developed
areas, contemporary land use transitions are increasingly shaped by policy interventions [4], such
as regional spatial planning, ecological red-line demarcations, and green economy legislation, rather
than by demographic demand alone. Simultaneously, the negative feedback effects of land use
change-induced environmental degradation on population dynamics are receiving growing attention
[5]. Cross-disciplinary research integrating high-resolution remote sensing data with multi-
dimensional environmental indicators , including land surface temperature and vegetation indices ,
has confirmed that localised urban heat islands and air pollutant concentrations resulting from built-
up land expansion impose spatial environmental stress on elderly communities and vulnerable
populations within affected areas [6]. Such deterioration in environmental quality is triggering intra-
urban population mobility and spatial redistribution, rendering land use change an important
variable in reshaping the patterns of population distribution [7].
Land use is widely employed in environmental research to reflect the natural attributes of land,
referring to the long-term or cyclical productive activities that human societies undertake on land
resources, guided by their natural characteristics and in pursuit of defined objectives, using
biological and technical means. Land use change, in turn, denotes the evolutionary process by which
land cover types, landscape patterns, and utilisation intensities are transformed across defined
spatial and temporal scales, driven by both natural factors and human intervention [8]. This study
employs the land use transfer matrix , a core analytical tool derived from the Markov model in
systems analysis that characterises the direction and structural attributes of land use change [9], in
conjunction with the temporal sequence of mutual conversions between different land use types at
the beginning and end of the study period, to explain how land use not only quantitatively captures
the overall trend in the area of each land category but also precisely delineates the internal logic of
conversions between categories [10].
The completion of the Hong Kong-Zhuhai-Macao Bridge has triggered a cascade of changes
in land use configuration and population mobility across the HZM-GBAUC [11]. While the
empirical focus of this study is situated in the HZM-GBAUC, this region serves as far more than a
local case study; it represents a quintessential testbed for addressing global urban and environmental
challenges. Unlike conventional urban clusters, The HZM-GBAUC has become globally the most
institutionally diverse and geographically interconnected cross-boundary cluster. It boasts extremely
high population and building density, along with diverse institutional, economic, and urban planning
standards. The unique typology of the HZM-GBAUC becomes particularly evident when compared
to other world-class bay areas. For instance, the San Francisco Bay Area is primarily characterized
by tech-driven innovation distributed across a relatively low-density, sprawling spatial structure
under a unified national system. Conversely, the Tokyo Bay Area represents a mature, highly
industrialized, and high-density agglomeration, yet it benefits from the coordinated governance and
standardized data frameworks of a single jurisdiction.
Land use change exerts a degree of influence on population change , defined as the dynamic
evolution of population scale, spatial distribution, composition, and growth trends over time within
a given area [12], in the context of this study, population change encompasses both quantitative
shifts and spatial migration. An in-depth examination of the impacts of land use policy on population
change over the past decade reveals that the relationship between population dynamics and land use
is no longer simply one of land supply meeting demographic demand. For these developed and land-
constrained cities, the influence of land use change has transcended mere physical necessity to
become a pivotal factor in regulating the regional talent structure [13], balancing social welfare
distribution, and alleviating social class tensions.
The Hong Kong, Zhuhai and Macao government have proposed leveraging the most efficient
land-saving institutional arrangements to promote coordinated regional development and to
facilitate the orderly mobility and optimised allocation of population across the HZM-GBAUC area
[14][15]. Hong Kong's population management has undergone a transition over the past decade from
monitoring-oriented approaches to active intervention [16]. The Hong Kong SAR Government has
introduced a series of talent attraction measures to address the dual pressures of population outflow
and accelerating ageing [17]. Data released by the Census and Statistics Department indicate that
there remains a substantial inflow of people from mainland China and from across the globe.
Macao's demographic structure has long been constrained by the dual limitations of extreme land
scarcity and heavy dependence on the gaming and tourism industry [18]. Compared to Hong Kong,
Macao's talent attraction schemes are more industry-oriented, targeting four key sectors: health and
wellness, modern finance, high technology, and convention, trade, culture, and sports [19]. Zhuhai
is positioned as a major nodal city within the Guangdong-Hong Kong-Macao Greater Bay Area,
leveraging its land resources at scale to accommodate the spatial overflow of industry transferred
from Macao and the living and service functions of residents [20]Error! Reference source not
found.. With the continued improvement of port infrastructure, Zhuhai is undergoing a functional
transformation from a traditional industrial city towards a life-services hub, accompanied by deep-
seated adjustments in the residential structure of its permanent population and in its modes of land
utilisation.
Land utilisation in the HZM-GBAUC is approaching saturation, and the elevated land premium
has substantially raised the cost of living. For these developed and land-constrained cities, land use
planning has transcended mere physical necessity to become a critical factor in regulating the
regional talent composition, balancing social welfare distribution, and alleviating worker class
tensions [22]. Hong Kong faces the dual pressures of population outflow and demographic ageing,
relying heavily on external talent attraction without resolving the structural problem of endogenous
population growth [23]. Constrained by extreme land scarcity, Macao is pursuing an industry-
oriented talent attraction policy and urgently needs to expand its spatial capacity. Zhuhai, as a major
node of the Greater Bay Area, is exploring cross-administrative-boundary land supply sharing
models through platforms such as the Hengqin Guangdong-Macao In-Depth Cooperation Zone,
accommodating the spillover of industries and residential functions from Macao [23].
Most existing research on land use change focuses on the developmental trajectories of
individual cities. Few studies have delved deeply into the spatio-temporal linkages between land
use change and population dynamics across adjacent cross-boundary urban agglomerations, and the
relevant research is in its early stages and remains incomplete. Moreover, no prior study has
systematically revealed the spatio-temporal mechanisms governing land population linkages across
Hong Kong, Zhuhai, and Macao within a unified conceptual framework. The objectives of this study
are therefore as follows:
- To provide a scientific basis for establishing cross-administrative-boundary land supply
mechanisms, contributing to the integrated coordination of stock renewal in Hong Kong and Macao
with incremental development in Zhuhai, and to avoid regional resource misallocation resulting
from fragmented governance [24]. - To guide governments in formulating land policies that diversify residential land supply and
promote a balanced job–housing spatial structure, thereby enhancing the region's long-term capacity
to attract high-caliber labor [25]. - To reveal the land-population interrelationships that serve not only the high-quality
development of Hong Kong, Zhuhai, and Macao, but also offer experience for the efficient
management of land and the balanced distribution of population in high-density cross-boundary
urban agglomerations globally [26].
Two core questions structure this research:
RQ 1: Are population changes across the HZM-GBAUC related to land use change? How do
the direction and scale of built-up land expansion influence population dynamics in each city?
RQ 2: The spatio-temporal evolution of built-up land across the HZM-GBAUC exhibits phased
characteristics, what is the relationship between the spatial trajectory of center-of-gravity migration
and changes in regional population density? - Method
2.1 Overview of the study area
2.1.1 Geographical Overview
Figure 1 Geographical Overview of Hong Kong, Zhuhai, and Macao
As illustrated in Figure 1 , Hong Kong, Zhuhai, and Macao are all situated along the
southeastern coast of China, occupying the core zone of the Guangdong-Hong Kong-Macao Greater
Bay Area and sharing a distinctive and closely interconnected spatial locational advantage. Hong
Kong has a total land area of approximately 1,114 km²; its terrain is dominated by low hills, with
limited flat land. These constrained natural conditions have led to the formation of a high-density
[27], vertically structured urban spatial pattern. Zhuhai and Macao share a direct land boundary.
Zhuhai has a total land area of approximately 1,736 km² [28]; its coastline is intricate and its island
groups are numerous, with landforms primarily comprising plains and low hills, offering a
comparatively expansive land development hinterland Error! Reference source not found. . The
Macao is situated in the southeastern part of Zhuhai, composed of the Macao Peninsula, Taipa Island,
and Coloane Island [30]. Constrained by its historical development and extreme land scarcity,
Macao has long relied on land reclamation projects to extend its urban boundaries; its land area has
expanded from under 15 km² in the late nineteenth century to approximately 33.3 km² at present,
making it one of the most representative cases of land use change in the region. Since the opening
of the Hong Kong–Zhuhai–Macao Bridge, the HZM-GBAUC have achieved cross-sea direct
connection, breaking down their physical barriers and providing a geographical basis for optimising
the spatial structure of land use and enhancing connectivity within the urban agglomeration [32].
2.1.2 Population profile
The HZM-GBAUC exhibit significant differences in population size, density, and growth
dynamics, closely linked to their respective land carrying capacities and urban functional
orientations [32]. As a highly developed international metropolis, Hong Kong's current permanent
resident population stands at approximately 7.5 million. The urban cores of Kowloon City District
in particular maintain extremely high population densities. Macao is among the most densely
populated territories in the world [33]; as of early 2024, its total population is approximately 686,000,
with large numbers of people concentrated in the geographically small peninsula area. Zhuhai has
maintained robust population growth in recent years; by 2024, its permanent resident population has
surpassed 2.5 million, with an urbanisation rate exceeding 90% [34].
2.2 Dataset source
The primary data in this study are divided into land use data and population data.
Land use data were collected for the years 2016, 2019, 2022, and 2025. The 2016, 2019, and
2022 land use datasets are drawn from GLC_FCS30D: a global 30-m land-cover dynamic
monitoring product with a fine classification system covering 1985 to 2022. GLC_FCS30D is the
first global fine land cover dynamic product at 30-metre resolution to adopt continuous change
detection. It employs a refined classification system containing 35 land categories spanning 1985 to
2022; update cycles were every five years prior to 2000 and annual thereafter. The 2025 land use
data were sourced from the National Geomatics Centre of China [35][36][37].
Population data consist of population density datasets for Hong Kong, Zhuhai, and Macao for
2016, 2019, 2023, and 2024. These were obtained from the LandScan high-resolution global
population distribution dataset, developed by the U.S. Department of Energy's Oak Ridge National
Laboratory. As annual data are typically compiled in the following year, the 2024 population density
data represent the most recent available observation [38].
2.3 Software and tools
Spatial data processing, cartographic visualisation, and core spatial statistical analyses in this
study were conducted primarily using ESRI's ArcMap software platform (version ArcGIS 10.7) [39].
As a classical desktop GIS application, ArcMap offers powerful capabilities in spatial data
management, topological editing, overlay analysis, and spatial modelling, enabling efficient
processing of multi-source heterogeneous data [40], including raster data extracted from post-
classification remote sensing imagery and vector-format census data. Specifically, the data
management tools in ArcMap were used to perform unified projection transformation, clipping, and
mosaicking of the multi-temporal land use data and population spatial distribution data for the
HZM-GBAUC, establishing a consistent spatial reference datum. ArcToolbox's spatial statistics
tools and raster calculator were employed for land use reclassification, geometric area computation,
and subsequent extraction of deeper spatial characteristics such as center- of-gravity migration and
hot-cold spot patterns. The software provided both a robust computational environment for large-
volume geospatial data and, through its powerful symbology functions, an intuitive representation
of the spatio-temporal evolution maps of land and population in the study area [41].
2.4 Research methods
To reveal the spatio-temporal characteristics of land use change across the HZM-GBAUC and
their spatial correspondence with population dynamics, this study employs a suite of quantitative
geographical methods, including the land use transfer matrix, the center-of-gravity migration model,
and spatial hot-cold spot analysis [42].
2.4.1 Land use transfer matrix
The land use transfer matrix is derived from the Markov model in systems analysis and serves
as a core method for characterising the direction and magnitude of mutual conversions between
different land use types over a given period [43]. The model comprehensively and specifically
reflects the flows and sources of each land use type between the initial and terminal time points of
the study period [44], thereby quantitatively depicting the competitive dynamics between ecological
and construction land during the urbanisation process in the HZM-GBAUC. In ArcMap, cross-
conversion area between land use raster or vector layers of different years is typically extracted
through spatial overlay analysis or the Tabulate Area tool [45]. The calculation formula is as follows:
𝑆 = [
𝑆11
𝑆21
𝑆12
𝑆22 ⋯⋯ 𝑆1
𝑆2
⋮ ⋮ ⋱ ⋮
𝑆
1 𝑆
2 ⋯ 𝑆
]
In this expression, 𝑆 denotes land area; n denotes the total number of land use types in the
study area; and denote the land use types at the beginning and end of the study period,
respectively; the diagonal elements 𝑆ᵢᵢ represent the area of land categories that remained
unchanged during the study period, while the off-diagonal elements 𝑆ᵢⱼ ( ≠ ) represent the area
converted from land type to land type Analysing this matrix allows precise tracking of, for
instance, the conversion of water bodies to built-up land through land reclamation, or the
encroachment of urban expansion onto forest and agricultural land [46][47].
2.4.2 The center-of-gravity migration model
The concept of the center of gravity originated in physics and was subsequently widely
introduced into human and physical geography to describe the overall tendency and migration
trajectory of a geographic phenomenon in regional spatial distribution [48]. To examine the spatial
expansion preferences of urban built-up land and the dynamic evolution of population
agglomeration centers across the HZM-GBAUC, this study employs a spatial center- of-gravity
migration model. Using the Mean Centre function within the Measuring Geographic Distributions
toolset in ArcMap, the geometric centroid coordinates of specific land use types or population
distributions are calculated for each year [49]. The calculation formula is as follows:
= ∑ ( × )
=1
∑
=1
, = ∑ ( × )
=1
∑
=1
In this expression, and denote respectively the longitude and latitude coordinates of
the center of gravity of a given geographic element (e.g., built-up land or population) in year ;
is the total number of spatial units in the study area [50]; and are the geometric centroid
coordinates of the i-th spatial unit; and is the attribute value (i.e., land patch area or population
count) of the -th spatial unit in year t. By connecting the centroid points for each period to generate
a migration trajectory, the macro-directional tendency of regional development can be intuitively
assessed [51].
2.4.3 Spatial hot-cold spot analysis
To further identify the significant local spatial agglomeration characteristics of land use change
intensity and population increase or decrease, this study employs the Getis-Ord Gi* index for local
spatial autocorrelation (hot-cold spot) analysis [52]. This method overcomes the limitation of global
spatial autocorrelation in being unable to specify the exact locations of agglomeration, and
effectively detects the specific spatial distribution of high-value clusters (hot spots) and low-value
clusters (cold spots) [53]. Using the Hot Spot Analysis tool in ArcMap, the statistic is computed for
gridded land change rates or population density increments as follows:
∗(") = ∑ #
=1 (") − ̅ ∑ #(")
=1
𝑆√ ∑ #2(")
=1 − (∑ #(")
=1 )
2
− 1
In this expression, is the observed value of spatial unit (e.g., population growth);
#(") is the spatial weight matrix based on distance ", taking a value of 1 when spatial units
and are adjacent and 0 otherwise; ̅ is the mean of all observed values; 𝑆 is the standard
deviation of observed values; and is the total number of spatial units. The resulting ∗ statistic
is the * -score: when the * -value is significantly and substantially positive, the area constitutes a
'hot spot' of significant high-value agglomeration (e.g., areas of rapid population growth or intensive
land development); conversely, when the * -value is significantly and substantially negative, the
area is a 'cold spot' of low-value agglomeration [53][55]. This analysis assists in revealing the
boundaries of spatial heterogeneity in the land–population interaction within the HZM-GBAUC.
3. Results
3.1 Land use types and areas in Hong Kong, Zhuhai, and Macao from 2016 to 2025
Using 2016, 2019, 2022, and 2025 as four temporal benchmarks, and drawing on land use
transfer matrices and population centroid data, this study systematically analyses the driving
mechanisms of land use change in the region and their implications for population dynamics.
The total study area covers approximately 3,300 km². This study primarily discusses seven
major land use categories: built-up land, cropland, forest, grassland, shrubland, water body, and
wetland.
Land type 2016 2019 2022 2025 Net change
(2016→2025)
Built-up Land 569.00 556.00 537.39 748.28 +179.28
Cropland 656.57 697.66 724.11 290.79 -365.78
Forest 809.96 791.45 787.95 1027.87 +217.91
Grassland 4.11 0.21 0.12 91.97 +87.86
Shrubland 188.57 191.36 194.59 21.81 -166.76
Water 857.32 831.39 832.31 1105.09 +247.77
Table 1 Summary of Land Use Areas at Each Temporal Benchmark (km²)
Based on monitoring and analysis of land use data for the HZM-GBAUC from 2016 to 2025,
land use types in the study area have evolved from relative stasis to structural transformation. Total
area remains stable at between 5,316.97 and 5,325.81 km², with overall growth attributable
primarily to human activities such as coastal land reclamation.
The area data for each land use type are shown in Table 1.The data indicate that between 2016
and 2022, built-up land declined marginally from 569.00 km² to 537.39 km². By 2025, built-up land
exhibited explosive growth, surging to 748.28 km². Between 2016 and 2022, cropland experienced
mild fluctuation and growth, rising from 656.57 km² to 724.11 km². By 2025, cropland fell sharply
to 290.79 km², a decline of more than 60%. Forest expanded markedly to 1,027.87 km² in 2025,
while grassland area leapt from near-zero to 91.97 km², compensating through afforestation and
green restoration for ecological losses caused by built-up expansion. Water body area increased
substantially in 2025, rising from 832.31 km² in 2022 to 1,115.09 km². Conversely, both wetland
and shrubland contracted to historic lows in 2025, at 5.01 km² and 21.81 km² respectively. The land
use type maps of Hong Kong, Zhuhai, and Macau are shown in Figure 2
Figure 2 Map of land use types in Hong Kong, Zhuhai, and Macao among different years (A: 2016; B: 2019; C:
2022; D: 2025)
3.2 2016-2025 Land use transfer across Hong Kong, Zhuhai, and Macao
3.2.1 2016-2019
The land use transfer matrix data for 2016-2019 is shown in Table S1. During this phase, built-
up land declined slightly from 569.00 km² to 556.00 km², a reduction of approximately 13 km²,
indicating localised land reversion and ecological restoration activity. Cropland area increased from
656.57 km² to 697.66 km², a net gain of approximately 41 km², while wetland also grew from 215.45
km² to 232.89 km².
Wetland 215.45 232.89 224.48 5.01 -210.44
The transfer matrix for this phase reveals that built-up land was converted to cropland at 45.83
km² and to wetland at 6.67 km², indicating that in cities such as Zhuhai, portions of development
zone land were reclaimed under policy influence. Cropland conversion to built-up land was only
35.17 km² demonstrating phased success for farmland protection policies. The land use transfer map
for 2016-2019 is shown in Figure 3, and the Sankey diagram of land use type transfer is shown in
Figure 4.
Figure 3 Land use type transfer map from 2016 to 2019
Figure 4 Land use type transfer Sankey map from 2016 to 2019
3.2.2 2019-2022
The land use transfer matrix data for 2019-2022 is shown in Table S2.The magnitude of change
in total land categories narrowed further in this phase, with the overall pattern approaching stability.
Built-up land fell to 537.39 km², a combined reduction of approximately 18.6 km² across the three
years; cropland steadily increased to 724.11 km², a cumulative gain of 10.2% relative to the 2016
baseline.
Water body area increased marginally from 831.39 km² to 832.31 km², while wetland declined
slightly to 224.48 km², reflecting dynamic water–wetland system interconversions: the Hong Kong–
Zhuhai–Macao Bridge and associated reclamation works had been substantially completed around
2018, causing surrounding marine ecological disturbance to gradually stabilise as natural wetlands
began recovering.
Built-up land contributed 50.25 km² to cropland and 7.43 km² to wetland conversions; cropland
contributed 35.72 km² to built-up land, indicating that urban sprawl continued but was effectively
offset by policy controls. The land use transfer map for 2019- 2022 is shown in Figure 5, and the
Sankey diagram of land use type transfer is shown in Figure 6.
Figure 5 Land use type transfer map from 2019 to 2022
Figure 6 Land use type transfer Sankey map from 2019 to 2022
3.2.3 2022-2025
The land use transfer matrix data for 2022-2025 is shown in
Table S3.This phase witnessed the most intense changes within the study period, with multiple
land categories undergoing large-scale restructuring. The most prominent feature is the dramatic
shrinkage of cropland (from 724.11 to 290.78 km², a reduction of approximately 433.3 km²), near-
disappearance of shrubland (from 194.59 to 21.81 km²), sharp decline of wetland to 5.00 km²,
substantial expansion of forest (from 787.95 to 1,027.86 km², an increase of approximately 239.91
km²), growth of grassland from near-zero to 91.97 km², increase of water body to 1,107.73 km², and
recovery of built-up land to 748.27 km².
Large areas of cropland and shrubland were converted to forest and grassland through
ecological restoration projects; portions of reclaimed areas were formally classified as water bodies;
multiple major construction projects in Zhuhai, Hong Kong's Northern Metropolis, and elsewhere
entered peak land-use phases, driving a net increase of approximately 210 km² in built-up land. The
transfer matrix indicates combined outflows from cropland to forest, water body, and built-up land
of approximately 490.4 km², and from shrubland to forest of 163.9 km², all are the largest single-
phase transfers. The land use transfer map for 20 22-2025 is shown in Figure 7, and the Sankey
diagram of land use type transfer is shown in Figure 8.
Figure 7 Land use type transfer map from 2022 to 2025
Figure 8 Land use type transfer Sankey map from 2012 to 2025
3.3 Population changes in Hong Kong, Zhuhai, and Macao from 2016 to 2024
Based on analysis of population statistics for 2016, 2019, 2022, and 2024, the HZM-GBAUC
has experienced significant population expansion and internal spatial restructuring over the past two
decades, each following distinctly different evolutionary logics. The population density distribution
map of Hong Kong, Zhuhai, and Macao in 2016-2024 is shown in Figure 9.
Figure 9 Population density distribution map of Hong Kong, Zhuhai, and Macao (A: 2016; B: 2019; C: 2022; D:
2024)
Population density in Hong Kong's core urban areas showed marked differentiation during the
study period (See Table 2). Yau Tsim Mong District recorded the highest density in the territory,
approximately 55,309 persons pre km² in 2016, rising to 57,613 persons pre km² in 2022 and
remaining extremely high at 58,270 persons pre km² in 2024. Wan Chai District declined
continuously from 34,547 persons/km² in 2016 to 29,753 persons pre km² in 2024, a decrease of
13.9%. The Southern District on Hong Kong Island fell from 5,346 to 4,739 persons pre km²
(−11.4%), consistent with trends of commercialisation pressure inflating residential costs and
residents relocating outward. Tai Po District density rose from 825 to 1,156 persons pre km²
(+40.1%), recording the largest increase in the territory; Sai Kung District rose by 23.6%, Yuen
Long District by 18.2%, and Tuen Mun District by 13.9%. The North District though at a low base
density, also registered positive growth (+4.7%). Kwun Tong District's 2024 population density rose
to 26,309 persons pre km², a 17.4% increase from 2016, the most pronounced rise in Kowloon City
District. In contrast, the peri-urban Tsuen Wan District and Kwai Tsing Districts fell sharply by 33.2%
and 19.5% respectively.
Administrative district 2016 2019 2022 2024 Net change
Central and Western District 14069 14201 13190 15721 ▲+11.7%
Wanchai District 34547 34872 32389 29753 ▼ -13.9%
Eastern District 6517 6307 5942 6298 ▼ -3.4%
Southern District 5346 5397 5011 4739 ▼ -11.4%
Yau Tsim Mong District 55309 55812 57613 58270 ▲ +5.4%
Sham Shui Po District 39606 39972 411147 39331 ▼ -0.7%
Kowloon City District 54567 55064 56841 53795 ▼ -1.4%
Wong Tai Sin District 23239 23450 24206 22870 ▼ -1.6%
Kwun Tong District 22403 22607 23337 26309 ▲+17.4%
Tsuen Wan District 6171 6216 6275 4126 ▼ -33.2%
Tuen Mun District 3154 3179 3262 3590 ▲+13.9%
Yuen Long District 4992 5018 5007 5898 ▲+18.2%
North District 2435 2453 2467 2549 ▲ +4.7%
Tai Po District 825 844 845 1156 ▲ 40.1%
Sai Kung District 1214 1222 1231 1500 ▲ 23.6%
Sha Tin District 5828 5878 5932 5708 ▼ -2.1%
Kwai Tsing District 19029 19180 19357 15317 ▼ -19.5%
Islands District 660 559 701 695 ▲ +5.3%
Note: Population density unit is persons pre km². ▲ denotes increase; ▼ denotes decrease. Net
change represents the relative rate of change from 2016 to 2024.
Table 2 Population density data of Hong Kong from 2016 to 2024
Population density data for Macau are shown in Table 3, St. Anthony Parish recorded a
population density of 123,636 persons/km² in 2016, the highest in the territory, which plummeted
to 67,000 persons pre km² in 2022 before recovering marginally to 68,273 persons pre km² in 2024,
representing an overall decline of 44.8%. St. Lawrence Parish density decreased slightly by 2.6%,
remaining broadly stable. St. Francis Parish is Macao's most dynamic emerging district in terms of
population density growth, rising from 4,450 persons pre km² in 2016 to 6,656 persons pre km² in
2024, an increase of 49.6%. St. Lazarus Parish density surged from 55,167 persons pre km² in 2016
to 87,000 persons/km² in 2024, a remarkable increase of 57.7%. Our Lady of Fatima Parish grew
by 18.3% and Cathedral Parish by 29.1%, both maintaining positive growth as principal areas on
the Macao Peninsula.
Parish 2016 2019 2022 2024 Net change
Our Lady of Fatima Parish 76,250 81,688 88,688 90,219 ▲+18.3%
St. Anthony Parish 123,636 125,000 67,000 68,273 ▼ -44.8%
St. Lazarus Parish 55,167 55,167 86,667 87,000 ▲ +57.7%
Cathedral Parish 14,471 15,294 18,029 18,676 ▲ +29.1%
St. Lawrence Parish 53,700 54,500 50,700 52,300 ▼ -2.6%
Our Lady of Carmel Parish 13,316 1,3797 14,316 14,671 ▲ +10.2%
St. Francis Parish 4,450 5,033 6,262 6,656 ▲ +49.6%
Note: Population density unit is persons pre km². ▲ denotes increase; ▼ denotes decrease. Net
change represents the relative rate of change from 2016 to 2024.
Table 3 Population density data of Macao from 2016 to 2024
Population density data for Zhuhai are shown in Table 4. Xiangzhou District, as the urban core
of Zhuhai, recorded a population density of 1,736 persons pre km² in 2016, declining to 1,530
persons pre km² in 2022 and remaining broadly flat at 1,536 persons pre km² in 2024. Doumen
District population density fell from 730 to 574 persons pre km², a decline of 21.4%, the sharpest
decrease among Zhuhai's three districts. Jinwan District density rose from 383 to 433 persons pre
km² (+13.1%), briefly peaking at 578 persons pre km² in 2019.
Administrative district 2016 2019 202 2024 Net change
Xiangzhou District 1,736 1,765 1,530 1,536 ▼ -11.5%
Doumen District 730 632 577 574 ▼-21.4%
Jinwan District 383 578 427 433 ▲ 13.1%
Note: Population density unit is persons pre km². ▲ denotes increase; ▼ denotes decrease. Net
change represents the relative rate of change from 2016 to 2024.
Table 4 Population density data of Zhuhai City from 2016 to 2024
3.4 Spatial distribution pattern of building land in Hong Kong, Zhuhai, and
Macao
The standard deviational ellipses and center-of-gravity migration maps (see Figure 10) indicate
that the spatial distribution of built-up land in the HZM-GBAUC exhibits marked directionality. All
three standard deviational ellipses display a clear northeast–southwest orientation, demonstrating
that built-up land in this region does not expand uniformly but rather clusters and extends along a
specific axis. In terms of ellipse area, the range differences across the three standard deviation levels
are considerable; the third standard deviational ellipse (red) covers a significantly larger area than
the first (blue) and second (yellow), indicating that the overall distribution of built-up land is
relatively dispersed with a small number of edge developments located far from the core, while the
core agglomeration zone (within the first standard deviational ellipse) remains highly compact.
The center- of-gravity migration trajectory shows that the centroid of built-up land shifted
notably from 2016 (red dot) through 2019 (yellow dot) and 2022 (blue dot) towards the southwest,
before partially reverting in a southeasterly direction by 2025 (green dot). This trajectory indicates
that built-up land expansion was not uniformly distributed across the study period but shifted in
emphasis across phases: during the earlier phase (2016-2022), the construction centroid moved
towards the southwest, reflecting the relative intensification of construction activities in the Zhuhai
and Macao directions; during the later phase (2022-2025), the centroid partially recovered, possibly
associated with a phased resurgence of construction activity in Hong Kong and its environs. The
overall magnitude of centroid migration is relatively limited, indicating that the spatial pattern of
built-up land in the HZM-GBAUC broadly stable at the macro level, with construction activity
concentrated within established urbanised core areas.
Figure 10 Map of the center of gravity and standard deviation ellipse of the building land in the study area
3.5 Factors that lead to changes in land use
The establishment in 2019 of the HZM-GBAUC's core positioning within the national strategy
directly propelled the implementation of large-scale land development plans including Zhuhai's
Hengqin New Area and Hong Kong's Northern Metropolis. The rapid recovery of built-up land after
2022 closely coincides with the commencement of substantive development phases for these
strategic projects. The sustained decline in built-up land and continuous growth of cropland between
2016 and 2022, contrary to the conventional pattern of net cropland loss driven by urbanization,
reflects the strong enforcement of cropland balance policies and the designation of permanent basic
farmland. During this phase, Zhuhai reclaimed portions of urban fringe construction land for
cultivation while maintaining a dynamic balance in the compensation for cropland newly occupied
by construction. Between 2022 and 2025, forest land across the HZM-GBAUC registered a net
increase of approximately 240 km² and grassland of approximately 91 km², closely correlated with
the concurrent implementation of ecological corridor restoration in Zhuhai and forest zone
expansion in Hong Kong. The large-scale conversion of shrubland to forest is the result of
afforestation programmes, reflecting a policy orientation of replacing shrubland with forest to
enhance ecosystem quality. Water body area, beyond the effects of reclamation activities, was also
influenced by the increased frequency of extreme precipitation events. Growth from 857 km² in
2016 to 1,115 km² in 2025, an increase of approximately 30%, reflects the fact that the HZM-
GBAUC has experienced frequent super typhoons and prolonged rainstorms in recent years, causing
the extent of seasonal water bodies to be captured in remote sensing imagery statistics. Concurrently,
the abrupt decline of wetland to 5.01 km² reflects adjustments in the land classification system:
coastal tidal flat areas previously classified as wetland, due to sea-level rise and land reclamation
activities, have partially converted to permanent water bodies.
The Hong Kong–Zhuhai–Macao Bridge and associated infrastructure consumed large areas of
construction land and water bodies during 2016-2019, and generated substantial reclamation activity,
causing localised fluctuations in wetland and water body statistics during that phase. Between 2019
and 2022, project completion and ecological restoration proceeded concurrently, stabilising water
body and wetland areas. The Zhuhai Jinwan Airport expansion and auxiliary works for the
Shenzhen–Zhongshan Link also accelerated during 2022-2025, directly correlating with the
recovery of built-up land.
Based on population centroid migration data, the population center of gravity in the HZM-
GBAUC shifted overall in a southwesterly direction during the study period, reflecting the sustained
enhancement of population attractiveness on the mainland side of the Greater Bay Area. Population
growth driving demand for residential, commercial, and infrastructure uses propelled the significant
expansion of built-up land after 2022; while the deceleration or net outflow of Hong Kong's
population growth partially relieved land expansion pressure in Hong Kong, contributing to the
characteristic pattern of built-up land statistics at the regional level showing a decline-then-rise
trajectory. Macao, with its extremely small area and near-saturated land development, has limited
influence on regional aggregate change, though its high-density built environment imposes long-
term pressure on local ecosystems, primarily manifested in the continued decline of wetlands.
3.6 Analysis of the impact of land use change
The net increase of approximately 210 km² in built-up land from 2022 to 2025 was spatially
concentrated in areas such as Hengqin in Zhuhai, the Northern Metropolis development area in
Hong Kong, and the Zhuhai Aeronautics and Astronautics Industry Park. This spatial expansion
provided critical spatial carriers for emerging industries including biomedicine, integrated circuits,
aviation manufacturing, and fintech. The net reduction of approximately 433 km² in cropland from
2022 to 2025 has severely compressed Zhuhai's agricultural production space. Although some losses
have been partially offset by urban agriculture and vertical farming within built-up areas, traditional
grain and vegetable production functions have been substantively impacted, further reducing
regional food self-sufficiency and intensifying dependence on external supply chains. The
significant expansion of forest and grassland has enhanced the regional stock of green ecological
assets and provided resource foundations for ecological tourism and forest wellness.
The HZM-GBAUC constitutes an important migratory bird corridor globally and provides
overwintering habitats for multiple rare and endangered species. The precipitous contraction of
wetland area, at a rate of 97.8%, will lead to severe biodiversity decline and habitat fragmentation
for migratory birds; reduced coastal protection functions and elevated risk of typhoon-driven storm
surges; and diminished offshore water purification capacity with increased probability of harmful
algal blooms and eutrophication. The net increase of 218 km² in forest has effectively enhanced
regional carbon sequestration capacity, water conservation functions, and biodiversity corridor
connectivity, making a positive contribution to offsetting the carbon emission increment generated
by built-up land expansion. However, the concurrent significant increase in built-up land intensifies
urban heat island effects, simultaneously creating pressure on the region's carbon reduction targets.
The expansion of water body area, on the one hand, enhances regional flood regulation capacity and
enriches waterfront landscape resources; on the other hand, the extensive expansion of permanent
water bodies alters local hydrodynamic conditions, with certain impacts on fish spawning habitats
and the distribution of benthic biological communities.
The study's land use typology encompasses seven categories: built-up land, cropland, forest,
grassland, shrubland, water and wetland. Among these, built-up land exhibits the most pronounced
change during the study period; the following analysis therefore focuses primarily on the impact of
built-up land change on population dynamics. According to the land use analysis data, built-up land
in the HZM-GBAUC underwent a phased trajectory of initial decline followed by recovery between
2016 and 2024: from 2016 to 2022, built-up land contracted modestly from 569 km² to 537 km²,
then surged sharply from 2022 to 2025 to 748 km², a net increase of approximately 210 km². From
the population density distribution maps, high-density agglomeration zones (darker areas) are
increasingly concentrated in the Kowloon–Hong Kong Island corridor, while low-density dispersal
zones extend towards the southwest, corroborating the northeast- to-southwest migration trajectory
of the built-up land centroid.
Population density data indicate that density growth in the northern districts (Tai Po District,
Sai Kung District and Yuen Long District) aligns closely with land development in this direction,
while the declining density in older districts such as Wan Chai District, Tsuen Wan District, and
Kwai Tsing District reflects the initial manifestation of population deconcentration effects. In Macao,
the incremental built-up land is primarily derived from continued development of St. Francis Parish
reclamation area, which not only provides new residential and commercial space but has also
become the fastest-growing area in terms of population density. The establishment of the Hengqin
Guangdong–Macao In-Depth Cooperation Zone has further extended Macao's population carrying
hinterland, objectively relieving population pressure in the Macao Peninsula, enabling older areas
such as St. Anthony Parish to achieve some decompression despite extremely high population
densities.
The opening of the Hong Kong-Zhuhai-Macao Bridge in 2018 broke the physical isolation
among the HZM-GBAUC. The rapid expansion of built-up land in areas adjacent to the bridge's
landing points in Zhuhai stimulated population growth in Jinwan District; the construction of
reclaimed new zones and port facilities on the Macao side also accelerated population agglomeration
in St. Francis Parish area; and Hong Kong's Lantau Islands District, as the bridge's Hong Kong
terminal, has seen its land development gradually intensify, with population density increasing from
660 to 695 persons pre km², laying the foundation for subsequent Northern Metropolis development.
- Discussion
4.1 Policy-Guided Spatial Interactions among Land Use, Population, and Ecology
This study reveals the dominant mechanism of policy intervention in reshaping land-
population feedback relationships. Land use change and its spatial configuration across the HZM-
GBAUC between 2016 and 2025 represent the core factors shaping population redistribution in this
high-density cross-boundary urban agglomeration. The implementation of the Outline Development
Plan for the Guangdong-Hong Kong-Macao Greater Bay Area in 2019 shifted the regional
governance paradigm from administrative isolation to cross-boundary functional integration. This
policy mandated the convergence of infrastructure and relaxed institutional barriers for cross-border
businesses, which directly catalyzed large-scale infrastructure investment [57]. Particularly between
2022 and 2025, Zhuhai’s policy designation as a key West-Bank node with relaxed talent-entry
thresholds and subsidized industrial land quotas created a policy pull effect. This institutional
incentive converted built-up land expansion directly into a structured population vacuum, drawing
over 2.5 million permanent residents [58]. The built-up land expansion and population growth
exhibit a clearly positive correlation. In 2021, Hong Kong’s Northern Urban Development Strategy
initiated a structural shift in land supply, aiming to direct residential and industrial development
northward, closer to the periphery. Hong Kong's built-up land development strategy is oriented
primarily towards northern New Territories areas such as Tai Po District, Yuen Long District and
the North District; through large-scale residential land supply, it guides residents from high-density
Kowloon and Hong Kong Island to relocate outward [59]. This was not merely a passive land supply
shift, but a deliberate policy-driven de-concentration strategy designed to alleviate the socio-spatial
pressure of the southern core and leverage the economic spillover of the mainland-border economy,
thereby directing the population gravity southwestward.
Throughout the study period, the spatial competition between built-up land and ecological land
was intense. Afforestation and green restoration projects drove net increases in forest and notable
growth in grassland, partially enhancing regional carbon sink capacity and ecological connectivity
[60]. The simultaneous sharp decline of wetland area represents a land cover conversion that not
only exacerbates habitat fragmentation for migratory birds and increases the risk of typhoon storm
surges, but also imposes long-term spatial environmental stress on coastal biodiversity and
vulnerable populations. Zhuhai, drawing on the ecological endowment of the Wanshan Islands and
Hengqin wetlands, has already formed an embryonic coastal ecotourism product portfolio; the
expansion of Hong Kong's country parks similarly strengthens its global green city profile, attracting
high-end tourism and convention-economy activity.
4.2 Institutional Intervention Strategies for Global Cross-Boundary Urban
Agglomerations
The HZM-GBAUC, as one of the most institutionally diverse cross-boundary urban
agglomerations in the world, faces a land–population trade-off challenge that is not unique but rather
a typical bottleneck encountered by mega-scale trans-boundary urban regions globally during deep
integration processes [61]. High institutional and economic gradients, mediated through differences
in land policy, can readily lead to regional spatial resource misallocation and destructive competition.
The Singapore–Johor model employs a core–hinterland collaborative development approach.
Given the extreme scarcity of land in Hong Kong and Macao contrasted with the spatial potential
of Zhuhai's hinterland, it is recommended to establish a land quota coordination mechanism that
transcends administrative and customs boundaries [62]. In population and industry mobility
planning, moving beyond a unidirectional land development model and coordinating the stock
renewal of high-density core areas with the incremental development of hinterland cities would be
advisable. Through the construction of cross-boundary joint parks, the orderly movement of
population, especially high-caliber talent, towards newly developed areas could be facilitated,
substantively alleviating spatial saturation and social compression effects in older districts. Th e
substantial contraction of wetland in the HZM-GBAUC and degradation of coastal habitats has
reached a scale where localised administrative intervention can no longer halt the overall decline of
the ecosystem. Drawing on the successful experience of the Oresund region in transnational
ecological co-governance and joint maritime spatial planning [63], it is recommended that local
governments establish cross-boundary ecological compensation frameworks with genuine binding
force and joint environmental monitoring networks. While advancing cross-sea infrastructure and
reclamation activities, ecological r ed-lines and cross-regional wetland restoration plans should be
incorporated into the joint statutory planning of the HZM-GBAUC, ensuring that urban expansion
does not come at the cost of regional climate resilience. The San Diego-Tijuana metropolitan area
on the US-Mexico border has experienced spatial misallocation and tidal cross-border congestion
due to institutional disparities and income differentials [64]; the HZM-GBAUC, when leveraging
land expansion to guide cross-boundary population redistribution, should go beyond providing mere
physical housing supply. Governments should strive to advance equitable cross-boundary public
service reforms. The research findings indicate that only when the supply of social public services
in new expansion areas can smoothly interface with, or even match, core area standards will policy-
led population transfer across administrative boundaries successfully transform into voluntary
integration driven by the pursuit of a higher quality of life.
In sum, this study has, within a unified conceptual framework, comprehensively assessed the
spatio-temporal transformation of land use across the HZM-GBAUC and the mechanisms through
which it influences population dynamics. The directed expansion of built-up land and the
coordinated migration of the population center of gravity demonstrate that systematic land planning
that transcends the traditional supply-and-demand framework is fundamental to achieving balanced
population distribution in high-density urban agglomerations. As regional integration continues to
deepen [64], incorporating diversified residential land supply, cross-boundary population mobility
guidance, and rigorous coastal ecosystem protection into a comprehensive decision-making
framework is essential both for safeguarding regional ecological security and for advancing the
sustainable development of high-density cross-boundary urban agglomerations globally.
5. Limitations
The land use data employed in this study span multiple sources, including the GLC_FCS30D
global 30-metre resolution land cover dynamic product and the 2025 data released by the National
Geomatics Centre of China. Cloud cover interference in remote sensing imagery and seasonal
variations in surface conditions may affect the extraction accuracy of water bodies and wetlands in
certain periods.
Furthermore, the spatial statistical units for population data in this study vary in granularity,
making rigorous grid-based comparison within a unified spatial analysis framework difficult, a
limitation that constrains the comparability and precision of hot-cold spot analysis across cities to
some degree. Given the close spillover and linkage effects between population flows and land
development in the HZM-GBAUC and surrounding cities, research conclusions focusing only on
these three locations have limited extrapolation capacity to the broader and more complex Bay Areas.
Future research could reconstruct analytical units at a larger spatial scale that transcends
administrative boundaries.
6. Conclusion
The land use pattern of the HZM-GBAUC underwent an evolution from relative stasis to
structural transformation during the study period. Between 2016 and 2022, constrained by policy
and stock-planning orientations, the total volume of built-up land exhibited a gentle contracting
trend, cropland area steadily recovered, and land-intensification management policies achieved
phased outcomes. Between 2022 and 2025, as the Guangdong-Hong Kong-Macao Greater Bay Area
strategy advanced in depth, built-up land registered a net increase of approximately 210 km², while
forest and grassland expanded markedly in the same period, forming a composite pattern in which
development and ecological restoration advanced in parallel, yet the precipitous decline of wetland
to historic lows reveals the severe pressure confronting coastal ecosystem protection. Population
space in the HZM-GBAUC displays pronounced differentiation: Zhuhai progressively achieved
region-wide population agglomeration anchored to the Greater Bay Area strategy; Hong Kong's
population space continued to consolidate from the core towards peripheral new development zones
in the northern New Territories, closely aligned with the trajectory of built-up land expansion;
Macao's population distribution, facilitated by large-scale land development in St. Francis Parish
reclamation area, has undergone directed transfer towards newly completed built zones. The
direction, scale, and pace of built-up land expansion are positively correlated in space with the hot-
cold patterns of population density across the cities, corroborating the fundamental principle that
land development activity serves as the core driver of population agglomeration.
Declarations
Ethics approval and consent to participate
Not applicable.
Consent for publication
Not applicable.
Availability of data and materials
The author confirms that all data generated or analyzed during this study are included in this
published article. Furthermore, primary and secondary sources and data supporting the findings of
this study were all publicly available at the time of submission.
Competing interests
The authors declare no competing interests.
Funding
This research was not funded.
7.References
[1] Long, H., Zhang, Y ., Ma, L., & Tu, S. (2021). Land use transitions: Progress, challenges and
prospects. Land, 10(9), 903.
[2] Scheidel, A., & Sorman, A. H. (2012). Energy transitions and the global land rush: Ultimate
drivers and persistent consequences. Global environmental change, 22(3), 588-595.
[3] Marshall, J. D. (2007). Urban land area and population growth: A new scaling relationship for
metropolitan expansion. Urban studies, 44(10), 1889-1904.
[4] Qu, Y ., & Long, H. (2018). The economic and environmental effects of land use transitions under
rapid urbanization and the implications for land use management. Habitat International, 82, 113-
121.
[5] Chu, X., & Xu, L. (2025). The research progress and future frontiers of agricultural geography:
A comprehensive review of theoretical evolution, technological transformation, and practice-driven
development. Geographical Research Bulletin, 4, 680-723.
[6] Singh, N., Singh, S., & Mall, R. K. (2020). Urban ecology and human health: implications of
urban heat island, air pollution and climate change nexus. In Urban ecology (pp. 317-334). Elsevier.
[7] Yue, X., Wang, Y ., Li, W., Wu, Y ., Wang, Y ., Zhang, H. O., & Ma, Z. (2024). Research progress
and trends in urban residential segregation. Buildings, 14(7), 1962.
[8] Lambin, E. F., & Geist, H. J. (Eds.). (2008). Land-use and land-cover change: local processes
and global impacts. Springer Science & Business Media.
[9] Rimal, B., Zhang, L., Keshtkar, H., Haack, B. N., Rijal, S., & Zhang, P. (2018). Land use/land
cover dynamics and modeling of urban land expansion by the integration of cellular automata and
markov chain. ISPRS International Journal of Geo-Information, 7(4), 154.
[10] Li, Y ., Ma, X., Zhao, J., Zhang, S., & Liu, C. (2025). Spatio-Temporal Evolution and Conflict
Diagnosis of Territorial Space in Mountainous–Flatland Areas from a Multi-Scale Perspective: A
Case Study of the Central Yunnan Urban Agglomeration. Land, 14(4), 703.
[11] Scanlon, B. (2024). The geopolitics of infrastructure and the unmaking of an island: The Hong
Kong-Zhuhai-Macau Bridge. Geopolitics, 29(2), 444-470.
[12] Chhabra, A., Geist, H., Houghton, R. A., Haberl, H., Braimoh, A. K., Vlek, P. L., ... & Lambin,
E. F. (2006). Multiple impacts of land-use/cover change. In Land-use and land-cover change: local
processes and global impacts (pp. 71-116). Berlin, Heidelberg: Springer Berlin Heidelberg..
[13] Basu, D. N., & Kashyap, S. P. (1996). way. Instead of hierarchical mapping of priorities and
programmes, ACRP begins with space specific resources (land, water, climatic features). Agro-
climatic Regional Planning in India: Themes and case studies, 2, 27.
[14] Zuo, Y ., Yang, C., Xin, G., Wu, Y ., & Chen, R. (2023). Driving mechanism of comprehensive
land consolidation on urban–rural development elements integration. Land, 12(11), 2037.
[15] Li, Z., Lin, K., & Tian, H. (2019). The development of a conservation-oriented and
environmentally friendly ecological economy. In Chinese Dream and Practice in Zhejiang–
Ecology (pp. 29-94). Singapore: Springer Singapore.
[16] So, B. K., Lui, I. D., Wat, K. P., Bai, Y ., Lam, K. F., & Yip, P. S. (2025). Assessment of the
impact of migration on Hong Kong’s population and its development. Asian Population
Studies, 21(3), 243-256.
[17] Lam, G., & Tu, E. J. C. (2015). Hong Kong’s population policies on immigration: challenges
and feasibility. Asian Education and Development Studies, 4(2), 180-189.
[18] Sheng, M., & Gu, C. (2018). Economic growth and development in Macau (1999–2016): The
role of the booming gaming industry. Cities, 75, 72-80.
[19] Yang, W. (2024). Strategies for Talent Acquisition and Development: Policies for Enhancing
University Talent Growth in the Guangdong-Hong Kong-Macao Greater Bay Area. Cultura:
International Journal of Philosophy of Culture and Axiology, 21(5).
[20] Tang, X., Xu, J., Wang, R., Li, J. V ., Jiang, L., & Li, C. Z. (2026). Drivers of Cross-Boundary
Land Use and Cover Change in a Megacity Region: Evidence from the Guangdong–Hong Kong–
Macao Greater Bay Area. Sustainability, 18(1), 470.
[21] Yu, H. (2024). Regional development in China under the “dual-circulation” strategy: The case
of Guangdong-Hong Kong-Macau greater bay area. Journal of Contemporary China, 33(148), 683-
703.
[22] Dunford, M., & Perrons, D. (1994). Regional inequality, regimes of accumulation and
economic development in contemporary Europe. Transactions of the Institute of British
geographers, 163-182.
[23] Lam, K. W. K., & Fong, E. (2025). The impact of post-2019 migration on Hong Kong
population dynamics. In Global Hong Kong (pp. 145-160). Routledge.
[24] Liu, Y ., & He, S. (2024). From the Pearl River Delta to the Greater Bay Area: State spatial
selectivity, contingent socio-spatial processes, and variegated geographies of China’s city-
regionalism. Transactions in Planning and Urban Research, 3(3), 216-240.
[25] Weng, M., Qiu, Y ., Wu, J., & Li, X. Spatial correlation network characteristics and driving
mechanisms of the construction industry in southeastern coastal urban agglomerations of
China. Available at SSRN 5646372.
[26] Zheng, L., Zheng, Y ., & Fu, Z. (2024). The Impact of Urban Renewal on Spatial–Temporal
Changes in the Human Settlement Environment in the Yangtze River Delta, China. Land, 13(6), 841.
[27] Zhu, J., Jiang, Y ., & Wang, C. (2026). From spatial mismatch to urban equity: a governance-
oriented supply–demand coupling framework for evaluating urban park green space equity in high-
density Tianjin. International Journal of Urban Sciences, 1-36.
[28] Leung, R. C. T. (1978). Hong Kong, a case of urban settlement expansion: an investigation of
implications of waterborne and hillside development (Doctoral dissertation, Massachusetts Institute
of Technology).
[29] Bosselmann, P., & Moos, S. (2014). The metropolitan landscapes of the pearl river delta and
the San Francisco Bay Area. Built Environment, 40(2), 244-264.
[30] Löffler, E. (1982). Landforms and landform development. In Biogeography and Ecology of
New Guinea: Part One-Seven (pp. 57-72). Dordrecht: Springer Netherlands.
[31] Tang, U. W., & Sheng, N. (2009). Macao. Cities, 26(4), 220-231.
[32] Wang, Z., Xu, G., Bao, C., Xu, J., & Sun, F. (2017). Spatial and economic effects of the Bohai
Strait Cross-Sea Channel on the transportation accessibility in China. Applied Geography, 83, 86-
99.
[33] Shi, Y ., Shi, S., & Wang, H. (2019). Reconsideration of the methodology for estimation of land
population carrying capacity in Shanghai metropolis. Science of the Total Environment, 652, 367-
381.
[34] Mak, H. W. L., & Ng, D. C. Y . (2021). Spatial and socio-classification of traffic pollutant
emissions and associated mortality rates in high-density Hong Kong via improved data analytic
approaches. International Journal of Environmental Research and Public Health, 18(12), 6532.
[35] Zou, Z., Zhao, X., Liu, S., & Zhou, C. (2025). Driving forces and future scenario simulation of
urban agglomeration expansion in China: a case study of the Pearl River Delta urban
agglomeration. Remote Sensing, 17(14), 2455.
[36] Zhang, X., Liu, L., Zhao, T., Chen, X., Lin, S., Wang, J., ... & Liu, W. (2023). GWL_FCS30: a
global 30 m wetland map with a fine classification system using multi-sourced and time-series
remote sensing imagery in 2020. Earth System Science Data, 15(1), 265-293.
[37] Zhang, X., Liu, L., Zhao, T., Gao, Y ., Chen, X., & Mi, J. (2022). GISD30: Global 30 m
impervious-surface dynamic dataset from 1985 to 2020 using time-series Landsat imagery on the
Google Earth Engine platform. Earth System Science Data, 14(4), 1831-1856.
[38] Zhang, X., Liu, L., Chen, X., Gao, Y ., Xie, S., & Mi, J. (2021). GLC_FCS30: Global land-
cover product with fine classification system at 30 m using time-series Landsat imagery. Earth
System Science Data, 13(6), 2753-2776.
[39] Dobson, J. E., Bright, E. A., Coleman, P. R., Durfee, R. C., & Worley, B. A. (2000). LandScan:
a global population database for estimating populations at risk. Photogrammetric engineering and
remote sensing, 66(7), 849-857.
[40] Liu, T., Luo, Y., Pang, P. C. I., & Xia, Y. (2024). Leveraging digital technologies for public health
education in High-Density community spaces: A Geospatial analysis. Systems, 12(11), 504.
[41] Daghlavi, M., Ghasemi-Nejad-Raeini, M., Marzban, A., Taki, M., Moazami, M., & Nabavi-
Pelesaraei, A. (2026). High-precision crop rotation mapping in arid agroecosystems: integrating
sentinel-2 time series, SVM, and GIS. Scientific Reports.
[42] Su, X., Liu, T., Pang, P., Luo, Y . T., & Wong, D. (2026). How Can Large Language Models
Drive Environmental Sustainability? A Systematic Scoping Review. Sustainability, 18(9), 4327.
[43] Peng, H., Lou, H., Zhang, Y ., Wang, W., He, Q., Liu, P., & Yang, Y . (2025). Study of
identification and simulation of ecological zoning through integration of landscape ecological risk
and ecosystem health. Frontiers in Ecology and Evolution, 13, 1662739.
[44] Luo, Y . T., Liu, T., Pang, P., & Kan, H. Y . (2026). Decoding dynamic emotional valence in
GenAI interactions: insights from covariate-dependent Markov chains. Frontiers in Psychology, 17,
1783364.
[45] Hurtt, G. C., Chini, L. P., Frolking, S., Betts, R. A., Feddema, J., Fischer, G., ... & Wang, Y . P.
(2011). Harmonization of land-use scenarios for the period 1500–2100: 600 years of global gridded
annual land-use transitions, wood harvest, and resulting secondary lands. Climatic change, 109(1),
117.
[46] Sun, Q., Wu, Z., & Tan, J. (2012). The relationship between land surface temperature and land
use/land cover in Guangzhou, China. Environmental Earth Sciences, 65(6), 1687-1694.
[47] Liu, J., Liu, B., Wu, L., Miao, H., Liu, J., Jiang, K., ... & Liu, T. (2024). Prediction of land use
for the next 30 years using the PLUS model's multi-scenario simulation in Guizhou Province,
China. Scientific Reports, 14(1), 13143.
[48] Alshayeb, M. J. (2025). Evaluating ecosystem service trade-offs and recovery dynamics in
response to urban expansion: implications for sustainable management
strategies. Sustainability, 17(5), 2194.
[49] Capoani, L. (2023). Review of the gravity model: origins and critical analysis of its theoretical
development. SN Business & Economics, 3(5), 95.
[50] Ma, M., Liu, Y ., Wang, B., Yan, X., & Li, H. (2023). Spatial-temporal evolution and driving
mechanism of urban land use efficiency based on T-DEA model: a case study of anhui province,
China. Sustainability, 15(13), 10087.
[51] Muyibul, Z., & Abulizi, M. (2026). Evolutionary characteristics of production, living, and
ecological spaces and related environmental effects in Urumqi, Northwest China. Environment,
Development and Sustainability, 28(5), 12533-12555.
[52] Zhou, H., Liu, Y ., & Liu, Y . (2019). An approach to tracking meteorological drought
migration. Water Resources Research, 55(4), 3266-3284.
[53] Songchitruksa, P., & Zeng, X. (2010). Getis–Ord spatial statistics to identify hot spots by using
incident management data. Transportation research record, 2165(1), 42-51.
[54] Wang, G., Wang, Y ., Li, Y ., & Chen, T. (2023). Identification of urban clusters based on
multisource data—an example of three major urban agglomerations in China. Land, 12(5), 1058.
[55] Anselin, L. (1992). Spatial data analysis with GIS: an introduction to application in the social
sciences.
[56] Yang, M., Ma, J., Jia, P., Pu, Y ., & Chen, G. (2011, June). The use of spatial autocorrelation to
analyze changes in spatial distribution patterns of population density in Jiangsu province, China.
In 2011 19th International Conference on Geoinformatics (pp. 1-6). IEEE.
[57] Xiao, Y . (2025). The Guangdong-Hong Kong-Macao Greater Bay Area: The Challenge of
Economic and Political Integration (Doctoral dissertation, University of Liverpool).
[58] Jahid, M. S. R. (2022). Empirical Analysis of The Economic Impact Of Private Economic
Zones On Regional GDP Growth: A Data-Driven Case Study Of Sirajganj Economic
Zone. American Journal of Scholarly Research and Innovation, 1(02), 01-29.
[59] Wu, Y ., Wang, J., & Gou, A. (2023). The evolution characteristics and driving mechanism of
urban construction land in Hong Kong since 1972. Applied Sciences, 13(4), 2565.
[60] Huang, T., Ma, L., Wang, X., Sheng, X., Huang, M., & Wang, Y . (2026). Carbon Neutrality
Pathways Through Ecological Restoration: Synergies Between Climate Change Mitigation and
Degradation Control. Land Degradation & Development.
[61] Lee, S., & Lee, S. (2021). China's water resources management . Springer International
Publishing.
[62] Hutchinson, F. E. (2021). In the gateway's shadow: Interactions between Singapore's
Hinterlands. Growth and Change, 52(1), 71-87.
[63] MORF, A., SANDSTEN, J., PRUTZER, M., & LINDROTH, N. (2021). Blue Economy and
Coastal Development–Sharing Swedish Experiences (No. 2021, p. 5). Report.
[64] Herzog, L. A. (2018). The US–Mexico transfrontier metropolis: Theoretical and empirical
explorations. In Handbook of emerging 21st-century cities (pp. 222-243). Edward Elgar Publishing.
[65] Wu, L., Lang, W., & Chen, T. (2024). Deciphering Urban Land Use Patterns in the Shenzhen–
Dongguan Cross-Boundary Region Based on Multisource Data. Land, 13(2), 161.
Table S1 Land Use Transfer Matrix in Hong Kong, Zhuhai, and Macao from 2016 to 2019(km²):
Summation term: area 2019
2016 Built-up Land Cropland Forest Grassland Shrubland Water Wetland Total
Built-up Land 512.7687 45.8298 0.5103 0.0369 0.099 3.0861 6.6735 569.0043
Cropland 35.1684 597.6423 5.1795 0.0855 2.502 6.9462 9.0414 656.5653
Forest 0.7587 20.5713 782.73 0.0018 4.8186 0.297 0.7848 809.9622
Grassland 0.009 0.0756 0.1611 0.0000 3.861 0.0000 0.0000 4.1067
Shrubland 0.3168 5.6979 2.2905 0.0702 180.0405 0.0261 0.1251 188.5671
Water 1.656 12.2355 0.1872 0.0063 0.0135 777.0897 66.1284 857.3166
Wetland 5.3244 15.6069 0.3888 0.0135 0.0297 43.9452 150.1371 215.4456
Total 556.002 697.6593 791.4474 0.2142 191.3643 831.3903 232.8903 3300.9678
Table S2 Land Use Transfer Matrix in Hong Kong, Zhuhai, and Macao from 2019 to 2022(km²):
Summation term: area 2022
2019 Built-up land Cropland Forest Grassland Shrubland Water Wetland Total
Built-up land 494.4321 50.2452 0.6588 0.0324 0.2763 2.925 7.4322 556.002
Cropland 35.7192 632.5182 5.7915 0.0342 3.1149 8.2935 12.1878 697.6593
Forest 0.729 7.4529 779.2092 0.0072 3.2742 0.1809 0.594 791.4474
Grassland 0.018 0.1341 0.0009 0.0477 0.0018 0.0009 0.0108 0.2142
Shrubland 0.0999 2.1375 1.2717 0.0000 187.7607 0.0189 0.0756 191.3643
Water 2.1825 9.72 0.2889 0.0009 0.0288 768.1734 50.9958 831.3903
Wetland 4.2093 21.9069 0.7299 0.0009 0.1359 52.722 153.1854 232.8903
Total 537.39 724.1148 787.9509 0.1233 194.5926 832.3146 224.4816 3300.9678
Table S3 Land Use Transfer Matrix in Hong Kong, Zhuhai, and Macao from 2022 to 2025(km²):
Summation term: area 2025
2022 Built-up land Cropland Forest Grassland Shrubland Water Wetland Total
Built-up land 454.4278 20.5034 18.3762 13.1771 0.9183 29.0527 0.2613 536.7168
Cropland 207.2589 192.8800 152.1007 37.2413 3.0127 130.6661 0.2166 723.3763
Forest 27.2661 41.5226 671.9100 23.7545 15.1131 6.5993 0.3743 786.5398
Grassland 0.0773 0.0060 0.0176 0.0178 0.0000 0.0046 0.0000 0.1233
Shrubland 6.8498 16.7161 163.9054 4.8538 1.1312 0.7798 0.0792 194.3152
Water 16.7104 7.7219 6.1198 4.1584 0.8317 789.2763 0.5114 825.3298
Wetland 34.9980 11.2254 13.8236 8.4530 0.6132 151.3484 3.5289 223.9906
Total 747.5882 290.5753 1026.2533 91.6558 21.6202 1107.7272 4.9717 3290.3918
论文原文与出版信息
Research Square