Spatial–Temporal Change and Dominant Factors of Coastline in Zhuhai City from 1987 to 2022

Abstract

Understanding the spatiotemporal variations and driving mechanisms of coastlines is crucial for their adequate protection, utilization, and sustainable development. In this study, the changes in various coastline types in Zhuhai from 1987 to 2022 were analyzed by using long-term Landsat and GaoFen satellite imagery. The Index of Coastline Type Diversity (ICTD), Index of Coastline Utilization Degree (ICUD) and the Digital Shoreline Analysis System (DSAS) analysis indicators were employed to investigate coastline change. Both quantitative and qualitative analyses were integrated to comprehensively elucidate the impacts of various driving factors. The results indicate that the total length of Zhuhai coastline increased from 761.50 km in 1987 to 798.91 km in 2022, with natural coastlines decreasing by 89.82 km and artificial coastlines increasing by 153.40 km. The rapid expansion of artificial coastlines since 2007 led to a marked decline in the ICTD indicator, while the ICUD indicator increased from 146.42 in 1987 to 216.37 in 2022, reflecting the intensified and continuous influence of anthropogenic activities. Additionally, the end point rate (EPR) and Weighted Linear Regression Rate (WLR) changed by 8.09 m/yr and 6.62 m/yr, respectively. The Shoreline Change Envelope (SCE) and Net Shoreline Movement (NSM) exhibited average changes of 331.42 m and 224.32 m, respectively. Gray correlation and regression analyses further revealed that climate factors exhibited the strongest association with natural coastline changes, while economic development indicators showed the strongest correlation with artificial coastline dynamics. The relationship of Number of Berths in Main Ports (Nb) with coastline changes strongly suggests that human activities are the primary driver of these changes. These findings provide a robust scientific basis for coastal zone management in Zhuhai.

1. Introduction

Coastlines, as critical transitional zones between terrestrial and marine environments, function not only as ecologically sensitive corridors but also as pivotal hubs for anthropogenic activities and resource exploitation [1,2,3]. Globally, coastal zones accommodate approximately 50% of the world’s population [4,5]. In China, more than 50% of medium-sized and larger cities are located in coastal areas, contributing over 60% of the nation’s GDP [6], thereby intensifying the conflict between anthropogenic exploitation and ecological conservation. In recent decades, accelerated sea-level rise (SLR) and coastal erosion have further heightened the risks of coastline retreat. For example, projections suggest that by 2050, SLR-driven erosion may submerge up to 70% of Italy’s beaches [7]. Concurrently, anthropogenic interventions (e.g., land reclamation and port construction) have led to significant morphological fragmentation. In China, artificial coastlines expanded at an annual rate of 14.5% before 2010 [8]. These synergistic effects of natural and anthropogenic stressors necessitate a systematic investigation into the spatiotemporal evolution and multidimensional driving mechanisms of coastlines to reconcile development-conservation dilemmas, enhance coastal resilience, and achieve sustainable governance.

Accurate coastline mapping is fundamental to studying the spatial–temporal change and driving factors of the coastline. Traditional coastline mapping methods based on field surveys have gradually been supplanted by remote sensing technologies, owing to the high cost and labor intensity of ground-based approaches [9,10]. The change in coastline mapping has progressed from the use of single aerial or satellite images to the fusion of hyperspectral and LiDAR data, and from manual interpretation and band statistics to machine learning algorithms [11,12,13]. The research has also expanded from localized key areas to global applications [14,15]. It can be seen that the current research on the spatial–temporal change in the coastline has made significant progress in methodology and application scale. However, there remains a significant gap in the long-term analysis of morphological differentiation, spatial heterogeneity, and multidimensional driving mechanisms in Zhuhai—a core coastal city within China’s Special Economic Zone and the Guangdong-Hong Kong-Macao Greater Bay Area. Existing studies on the coastline of Zhuhai mostly focus on short periods [16,17,18,19], and although some studies have analyzed the long time-series changes in the Guangdong-Hong Kong-Macao Greater Bay Area, most of them are based on administrative divisions, which makes it difficult to reveal the changes in a refined way [20,21,22].

Understanding the spatial–temporal change in coastlines is essential for achieving sustainable development. Previous studies have identified both natural processes and anthropogenic activities as the principal factors influencing coastline changes, with natural factors traditionally regarded as the primary drivers [23]. However, in recent years, intensified anthropogenic activities have emerged as the predominant force behind these changes. Zhang et al. demonstrated through layer superposition and impact analysis that anthropogenic interventions, along with variations in material composition and landforms, play a critical role in shaping the coastlines of Southeast Asian islands [24]. Amrouni et al. found that the retreat of North African coastlines, coupled with rapid urban growth, has accelerated inland seawater intrusion by approximately 5 km, thereby negatively affecting crop production and food security [25]. Above results indicate that anthropogenic processes, rather than natural forces, primarily drive coastline retreat. In addition, Lansu et al. reported that coastal infrastructure development restricts the available space for coastal ecosystems, producing irreversible effects on coastline morphology [26]. Most existing research, however, tends to analyze driving factors using either qualitative or quantitative methods in isolation. Yang et al. combined tidal station data and reclamation project statistics to show that anthropogenic activities have dominated the evolution of mainland China’s coastline over the past 30 years [27]. Wang et al. qualitatively evaluated 50 years of coastline change in Huizhou city and concluded that anthropogenic factors play the dominant role, with natural factors having minimal influence [28]. Relying solely on a single analytical approach may underestimate the complexity and relative influence of various driving factors. Therefore, integrating both qualitative and quantitative analyses is necessary to comprehensively assess the factors influencing coastline changes.

To address the shortcomings in existing research on the spatiotemporal changes and driving factors of Zhuhai coastline, this study establishes a research framework comprising spatiotemporal change analysis, dynamic assessments of utilization and locational shifts, and driving factor evaluation. Utilizing methodologies such as utilization evaluation, the Digital Shoreline Analysis System (DSAS), gray correlation analysis, and regression analysis, we extract detailed spatiotemporal data on the coastline, assess changes in utilization and position, and quantify the relative influences of natural and socioeconomic factors to determine the dominant drivers. The objective of this study is to provide a robust scientific basis for coastal zone planning in Zhuhai and offer insights applicable to other estuarine cities.

2. Materials and Methods

2.1. Study Area

Zhuhai is located on the western margin of the Pearl River Estuary (21°48′–22°33′ N, 113°03′–114°19′ E) (Figure 1), which contains multiple islands and embayments, including Hengqin, Gaolan, Hebao and Qi’ao, parts of the Wanshan Archipelago, Xianglu Bay and Jiuzhou Bay. Coastal dynamics are governed by semi-diurnal tides and associated residual currents. Seasonal discharge from the Pearl River drives estuarine circulation and sediment plumes. Wave action and typhoon-driven storm surge reshape the shoreline during extreme events. Longshore sediment transport controls the morphology of beaches, tidal flats and shoals. Changes in river sediment supply and human activities have modified local hydrodynamics and sediment budgets. Human activities include large-scale land reclamation, port construction around Gaolan and Jiuzhou and the Hong Kong–Zhuhai–Macao Bridge with its artificial structures. The region has a subtropical climate with warm temperatures year-round and distinct wet and dry seasons [29]. This climate together with the geographic setting increases vulnerability to sea-level rise, coastal erosion and other climate-change stresses. Recent coastal urbanization has substantially altered the natural landscape and expanded artificial shorelines, particularly around Hengqin and Gaolan islands [30].

2.2. Data

In this study, a combination of Landsat series satellite imagery and GaoFen satellite data was employed as the primary data source (

Table 1. Remote sensing information used in the study area.

No.	Remote Sensing Data Type	Time
1	Landsat5 TM	8 December 1987
2	Landsat5 TM	1 November 1997
3	Landsat5 TM	29 January 2007
4	Landsat7 ETM	18 December 2017
5	GF-2, GF-7	3 February 2022

). Landsat imagery was acquired from the United States Geological Survey Earth Explorer platform (https://earthexplorer.usgs.gov (accessed on 12 December 2024)), and GaoFen data were obtained from ChinaGEOSS Data Sharing Network (https://noda.ac.cn (accessed on 13 December 2024)). Following standardized procedures, the imagery underwent radiometric calibration, mosaic merging and geometric correction to ensure spatial and spectral consistency. These preprocessed images formed the basis for coastline extraction. Coastline features were identified through the visual interpretation approach, in which remote sensing data were analyzed based on the spectral signatures of typical coastline elements. The final delineation of the coastline was conducted in accordance with established coastline classification criteria and definitional principles, ensuring both accuracy and consistency across the temporal dataset.

2.3. Coastline Classification System

Based on the current development status of Zhuhai coast and considering features such as image scale, shape, shadow, color, and the surrounding environment derived from Landsat and high-fraction imagery, the continental coastline in the study area is classified into two primary categories: natural and artificial coastlines [8,31,32]. The natural coastline is further subdivided into bedrock, sandy and biological coastline. In contrast, the artificial coastline is categorized according to the marine functional area classification established by the State Oceanic Administration in China, which includes industrial, port terminal, road and breeding.

Table 2. Coastline classification system description in the study area.

Category	Class	Description	Image Example
Natural	Bedrock	Bedrock coastlines typically follow the water’s edge line as the shoreline reference, which is often irregular and jagged due to the resistance of exposed rock formations. On remote sensing imagery, the bedrock appears light-toned with striping patterns, and the shores are frequently dotted with reefs, boulders, sea cliffs, and other erosional landforms.
	Sandy	Sandy coastlines are typically delineated using the upper beach ridge line or the vegetation line as the shoreline reference. They exhibit a smoother water’s edge line; bright white, uniform hue in areas not reached by the tide, and a darker hue in areas wetted by the tide. Beaches often show striping patterns on remote sensing imagery.
	Biological	Biological coastline is a special kind of coastal space formed by the special development of a certain kind of organism. Among the types of biological coastlines visible in Zhuhai City, mangrove coastlines are generally mangrove forests, which are distributed in numerous independent patches in the intertidal zone and are dark green in true-color images and red in false-color images.
Artificial	Industrial	Industrial: Coastline used for waterfront industries for port construction development.
	Port Terminal	Coastline means coastline used for the construction of port terminals, including coastline used for piers, jetties, and other construction function uses.
	Road	Coastline used to connect land-linked islands, generally straight in shape.
	Breeding	Breeding generally refers to the use of aquaculture, fisheries production pond dike coastline, generally located in estuaries, shallow, fish ponds around the perimeter, mostly regular block target.
Island		The water–land boundary of atoll far from land or an island.

presents the complete classification system for the coastline in the study area.

2.4. Visual Interpretation

Given the characteristics of Zhuhai’s coastline, characterized by an intricate configuration, numerous embayments, extensive muddy shores, and ambiguous land-water boundaries in specific sectors, this study utilized visual interpretation for coastline delineation. This methodology integrated manual interpretation of remote sensing imagery with field verification. Initial interpretation leveraged key features within false-color composite imagery, including tone, texture, geomorphology, and surrounding features. Coastline positions were subsequently determined based on specific type, as detailed in Table 2. Uncertain shoreline segments were confirmed via field surveys to generate the final coastline dataset.

2.5. Methods

2.5.1. Coastline Utilization Index

In this study, the index of coastline type diversity (ICTD) and the index of coastline utilization degree (ICUD) were used to quantify the diversity and utilization intensity of coastline types, respectively. ICTD measure the diversity of coastline types based on their relative lengths [33]. The calculation formula is as follows:

$$\mathrm{ICTD} = 1-{\displaystyle \frac{\sum _{i=1}^n{L}_i^2}{{\left(\sum _{i=1}^n{L}_i\right)}^2}},$$

(1)

where ICTD represents the diversity of coastline types in the study area in a certain year, n denotes the total number of coastline types; i indexes each type; L_i represents the length of the ith coastline type. The ICTD value ranges from 0 to 1, with values closer to 0 indicating lower diversity and those nearer to 1 signifying a greater complexity and diversity in coastline types. This index is closely related to the proportionate lengths of the various coastline types. The ICUD quantifies the intensity of anthropogenic utilization along the coastline by incorporating both the proportion and the weight of different coastline types [34]. The calculation formula is as follows:

$$\mathrm{ICUD} = {\sum }_{i=1}^n\left(A_i \times { w}_i\right) \times 100,$$

(2)

where ICUD indicates the utilization intensity of the coastline in the study area in a certain year; i indexes each type; A_i stands for the proportion of ith length relative to the total coastline; W_i indicates the weight reflecting the degree of anthropogenic activity impact on that type. The higher ICUD value implies that anthropogenic activities exert a stronger influence on the coastline. The weighting factors (W_i) for each coastline type are determined by referring to previous studies (see

Table 3. Influence weight of anthropogenic activities on coastlines.

Class	Island	Bedrock	Sandy	Biological	Artificial	Breeding
Weight	1	1	2	3	4	4

for detailed information) [35,36].

2.5.2. DSAS Analysis

Before performing the DSAS analysis (developed by United States Geological Survey, Reston, VA, USA), transect generation was required. Considering the spatial characteristics of the study area and the resolution of the imagery, transects were generated at 100 m intervals and extended to a length of 2000 m. As a result, a total of 1736 transects intersecting the coastline were obtained; their spatial distribution is further detailed in the DSAS index analysis. The Net Shoreline Movement (NSM), End Point Rate (EPR), Weighted Linear Regression Rate (WLR) and Shoreline Change Envelope (SCE) were used to evaluate the shoreline dynamics in the study area. NSM represents the distance between the oldest and the most recent coastlines, thereby quantifying overall coastline change. EPR indicates the rate and direction of coastline movement [37]. WLR—an extension of the Linear Regression Rate (LRR)—determines the best-fit line that characterizes the coastline trend [38]. SCE is defined as the maximum distance between all coastlines intersecting a transect. For a detailed description of the DSAS metrics, see the DSAS Version 5.0 User Guide (https://pubs.usgs.gov/publication/ofr20181179 (accessed on 5 February 2025)).

2.5.3. Gray Correlation Analysis

Gray correlation analysis quantifies the degree of association between variables by examining the similarity of their trend curves (i.e., the more similar the trends, the higher the correlation). This method is particularly suitable for small sample sizes and data that do not conform to a normal distribution. The analysis proceeds through the following steps [39]:

Step 1: Mean normalization

$${\mathrm{x}}_i^{\prime }\left(k\right)={\displaystyle \frac{{\mathrm{x}}_i\left(k\right)}{\frac{1}{n}\sum _{k=1}^n{\mathrm{x}}_i\left(k\right)}},$$

(3)

where $x_i^{\prime }\left(k\right)$ represent the normalized value of the ith variable in year k. Each raw data series is normalized to remove dimensional effects.

Step 2: Calculation of absolute differences

$$\mathrm{a}=\begin{array}{c} min\\ i\end{array}\begin{array}{c}min\\ k\end{array}{\Delta }_{i}k,$$

(4)

$$\mathrm{b}=\begin{array}{c}max\\ i\end{array}\begin{array}{c}max\\ k\end{array}{\Delta }_{i}k,$$

(5)

where a and b are the minimum and maximum difference between the ith variable and the length of the study area coastline in year k, respectively. Δik is the absolute difference between the ith variable and the length of the study area coastline in year k.

Step 3: Calculation of correlation coefficient

$$y_{i}k={\displaystyle \frac{a+\mathsf{\rho }b}{{\Delta }_{i}k+\mathsf{\rho }b}},$$

(6)

where $y_{i}k$ denotes the correlation coefficient of the ith variable with the length of the coastline of the study area in year k, and $\mathsf{\rho }$ is the discrimination coefficient, which usually takes the value of 0.5.

Step 4: Gray correlation calculation

$$\mathrm{Y}={\displaystyle \frac{1}{n}}{\sum }_{k=1}^n{y}_{i}k,$$

(7)

where Y denotes the average correlation coefficient between the ith variable and the length of the coastline in the study area over k years.

3. Results

3.1. Spatial–Temporal Change in Coastline

Figure 2 illustrates the dynamic evolution of the coastline in the study area from 1987 to 2022. Significant spatiotemporal changes were observed over the study period. In 1987, the total coastline length was 761.50 km, comprising 184.63 km (24.24%) of natural coastlines and 103.91 km (13.64%) of artificial coastlines. Among the natural segments, bedrock coastlines dominated, accounting for 83.49% (154.32 km) of the natural coastline, while aquaculture dikes constituted the major component (67.13%, 69.71 km) of the artificial coastline. By 1997, the overall coastline length had increased to 768.62 km; however, the natural coastline decreased to 118.62 km (15.43%), whereas the artificial coastline expanded to 202.27 km (26.31%). Within the natural category, bedrock coastlines continued to predominate (88.67%, 105.23 km). In contrast, the composition of the artificial coastline shifted, with artificial construction becoming the dominant component (62.20%, 125.82 km). In 2007, the total coastline slightly decreased to 750.50 km. Natural coastlines further declined to 107.80 km (14.36%), maintaining bedrock as the primary type (84.00%, 90.55 km), while the artificial coastline exhibited substantial growth, reaching 270.52 km (36.05%), of which artificial construction accounted for 54.54% (147.57 km). By 2017, the total coastline had expanded to 785.15 km—an overall increase of approximately 5.00% since 1987. During this period, natural coastlines experienced a marked reduction to 91.85 km (11.70%), with bedrock coastlines notably declining by 69.39%. Conversely, the artificial coastline surged to 277.89 km (35.39%), with artificial construction comprising 66.53% (184.72 km) of this total. In 2022, the total coastline reached 798.91 km. Natural coastlines were reduced further to 94.81 km (11.87%), reflecting an overall decline of 48.64% since 1987; bedrock coastlines suffered a particularly severe reduction (−69.39%), while biological coastlines paradoxically increased by 158.82%. In contrast, artificial coastlines grew markedly (147.62%), with artificial construction exhibiting an extraordinary rise of 466.73%. Notably, the aquaculture dikes showed an initial increase followed by a subsequent decline over the study period. Spatial analysis revealed a pattern of natural coastline retreat concurrent with artificial coastline expansion, predominantly concentrated in the central and southern regions of the study area (Figure 3). The spatial overlap strongly shows that intensive human activities are the primary driver of coastal transformation, with significant ecological implications arising from the accelerated loss of natural habitats and the proliferation of anthropogenic infrastructure.

3.2. Dynamic Changes in Coastline Utilization and Location

The dynamic evolution of coastline utilization within the study area reveals distinct temporal patterns (Figure 4). ICTD exhibited a notable increase between 1987 and 2007, followed by a significant decline from 2007 to 2022. This temporal trend reflects the trans-formation in the proportional distribution of coastline types over time. During the 1987–2007 period, the proportions of major coastline categories—such as bedrock coastlines (ranging from 83.49% to 88.67% of natural coastlines), artificial construction coastlines (rising from 54.54% to 67.13% of artificial coastlines), and aquaculture dikes (from 62.20% to 75.22%)—exhibited only modest fluctuations. These relatively stable proportions contributed to a higher diversity of coastline types. However, post-2007, substantial structural changes occurred. Bedrock coastlines experienced a sharp decline, decreasing by 69.39% by 2022, while aquaculture dike coastlines also significantly contracted. In contrast, artificial construction coastlines expanded dramatically, with an increase of 466.73%. This rapid expansion of a single artificial type led to a reduction in the overall diversity of coastline types, resulting in the observed decline in ICTD. Simultaneously, ICUD exhibited a steady upward trend, increasing at an average rate of 1.97 units per year from 1987 to 2022. This continuous rise indicates a growing intensity of anthropogenic intervention and utilization along the coastline. The opposite trend of ICTD and ICUD reflects that the coastal development in the study area has shifted from relative balance to increasingly intensive utilization dominated by anthropogenic in the past 35 years.

The DSAS analysis results for the study area from 1987 to 2022, as illustrated in Figure 5, revealed pronounced spatiotemporal changes in the coastline. EPR analysis indicated an average coastline change rate of 8.09 ± 0.01 m/yr, with 26.27% of transects exhibiting statistically significant erosion (mean rate: −7.32 m/yr) and 72.69% showing statistically significant accretion (mean rate: 13.64 m/yr). These results reflect a prevailing trend of coastline progradation over the study period. WLR analysis indicated a slightly lower average rate of 6.62 ± 3.34 m/yr. Among all transects, 26.04% were classified as erosional (only 5.23% statistically significant, with an average rate of −3.29 m/yr) and 73.96% as accretional (of which 13.83% were statistically significant, averaging 10.10 m/yr). These findings further support the general trend of coastline advancement, though localized erosional processes persist. SCE represents the maximum displacement among all observed coastlines at each transect, averaged 331.42 m, ranging from 0.65 m to 1001.95 m, reflecting the spatial heterogeneity of coastline responses across the study area. NSM represents the linear distance between the earliest and most recent coastlines, recording an average change of 224.32 m. Notably, 26.48% of transects showed negative displacement, with a maximum retreat of −859.40 m and an average of −111.68 m, whereas 73.52% of transects experienced positive displacement, with a maximum advancement of 1001.54 m and an average of 345.33 m. Spatially, the most pronounced EPR reductions were concentrated in the eastern region, while localized areas in the south exhibited the highest accretion rates. The spatial patterns of WLR, SCE, and NSM closely aligned with those observed in the EPR analysis, reinforcing the consistency of regional coastline dynamics across multiple indicators.

3.3. Influencing Factors of Spatial–Temporal Change in Coastline

The results of the gray relational analysis revealed that the highest gray relational coefficients with natural coastline were associated with MF (0.85), MAT (0.83), and AP (0.81), followed by PRP (0.65), AA (0.65), Nb (0.65), QL (0.61), and APO (0.61). The lowest coefficients were observed for GDP (0.49) and Ex (0.47), both below the critical threshold of 0.5, indicating weak correlations (Figure 6). In contrast, the highest gray relational coefficients with artificial coastline were observed for Ex (0.64), PT (0.63), TFT (0.61), and GDP (0.60). Coefficients below 0.5 were recorded for PRP (0.49), MF (0.49), MAT (0.49), AA (0.47), and Nb (0.46). Regarding the total coastline, the strongest associations were observed with MAT (0.98) and AP (0.85), while Ex (0.46) was the only factor with a coefficient below 0.5. Overall, climatic variables (specifically MAT and AP) demonstrated the strongest associations with natural coastline changes, whereas economic development indicators such as GDP, TFT, and Ex showed the strongest correlations with artificial coastline dynamics. Further regression analysis revealed that MF and Nb were significantly associated with natural coastlines (Figure 7). Specifically, MF exhibited a significant positive correlation (slope = 1.05, p = 0.01), whereas Nb showed a significant negative correlation (slope = −1.03, p = 0.002). Regarding the artificial coastline, MF, TFT, AA, and Nb all displayed statistically significant correlations. Among these, MF had a significant negative correlation (slope = −1.03, p = 0.03), while TFT, AA, and Nb showed significant positive correlations. No variables were found to be significantly correlated with the total coastline length. However, Nb exhibited an especially strong relationship with coastline change, demonstrating an inverse association with natural coastline and a direct association with artificial coastline. This suggests that anthropogenic activities was the most influential factor driving changes in coastline length.

Similar conclusions can be drawn from typical regional coastline changes. In the southwestern part of the study area, Gaolan Island exhibited a clear trend of seaward expansion from 1987 to 1997, during which road infrastructure connecting the island to the mainland was completed (Figure 8). This expansion trend weakened between 1997 and 2007, although land reclamation began in the northern part of Gaolan Island in 2003, accompanied by the initiation of engineering construction on its western side. By 2022, land enclosure in the northern region had been completed, and construction activities had extended to the western and southern sections connected to the mainland, eventually forming a port. As of 2022, the northern part of the island was fully reclaimed, with ongoing infrastructure development in the west and south, contributing to continued westward expansion of the harbor. Hengqin Island, located in the southern portion of the study area, experienced particularly significant seaward expansion during the 1987–1997 period (Figure 9). The eastern islands underwent widespread outward growth due to polder farming and the development of coastline engineering projects. These efforts resulted in the enclosure of the original natural coastline and the creation of new artificial coastline, leading to a substantial increase in coastal zone area. Consequently, the proportion of artificial coastline steadily approached that of natural coastline. Between 1997 and 2007, this expansion trend diminished, with reclamation and aquaculture mainly concentrated in the Nanwan and Jinwan areas. After 2007, the rate of coastline expansion further declined compared to the previous periods. Expansion became primarily driven by engineering construction projects, while the contribution from reclamation and aquaculture decreased.

4. Discussion

The coastline of Zhuhai has undergone marked spatiotemporal changes from 1987 to 2022, characterized by a dominant trend of artificial coastline expansion and natural coastline retreat. DSAS analysis revealed a prevailing accretional trend, particularly in the southern parts of Zhuhai, such as Gaolan and Hengqin Islands, where extensive land reclamation and coastal infrastructure construction have been concentrated. The average EPR reached 8.09 ± 0.01 m/yr, suggesting strong anthropogenic influence over natural geomorphic processes. In the early period (1987–1997), infrastructure development such as road construction and initial reclamation projects marked the beginning of rapid coastline advancement. Between 1997 and 2007, land enclosure expanded northward and westward in Gaolan Island, while large-scale polder farming and coastline engineering in Hengqin significantly altered the coastline configuration. After 2007, the pace of expansion slowed, replaced by more localized engineering projects. This change aligns with observations in other rapidly urbanizing Chinese coastal cities. Han et al. (2023) found similar patterns of morphological change driven by anthropogenic activity in estuarine environments [40]. Li and Dong (2023) documented phased and spatially uneven ecological transformations in the Pearl River Delta, closely tied to urban development intensity [41]. While the morphological gain from artificial coastline construction supports urban growth and port development, it has also led to a marked reduction in ICTD and ICUD, signaling a shift from balanced land–sea interaction to anthropogenic coastal systems.

The coastal transformation in Zhuhai is shaped by both climatic variables and socio-economic development. Gray relational analysis revealed that natural coastline change was most strongly associated with climate factors, particularly mean annual temperature (MAT, 0.83) and annual precipitation (AP, 0.81). These results suggest that hydroclimatic dynamics affect natural coastline morphology through influencing runoff, erosion, and sedimentation, although their influence is increasingly overshadowed by anthropogenic activity. In contrast, artificial coastlines showed strong correlations with economic indicators such as Ex, PT and TFT, which are closely tied to infrastructure expansion. Regression analysis also highlighted the number of berths (Nb) as a key indicator: negatively correlated with natural coastlines (p = 0.002) and positively with artificial coastlines (p = 0.01), consistent with the narrative of land use transformation. These findings reinforce prior studies on the role of economic growth in coastal change. Han et al. (2023) emphasized how anthropogenic activities have restructured estuarine geomorphology [40]. Li and Dong (2023) identified GDP-driven expansion as a primary contributor to ecological decline in urban coastal zones [41]. In Zhuhai, the shift from natural to engineered coastlines reflects broader national trends, but also suggests an urgent need to balance development with ecological resilience.

However, this study operates within medium-resolution remote sensing data. Firstly, Landsat imagery was established as the robust tool for regional coastline monitoring, imposes inherent limitations on the detection of sub-pixel features such as narrow beach berms or minor scarps [14,42,43]. In addition, the observed reduction in rocky coastline and the variability of biological coasts (including aquaculture areas) likely reflect both human and natural drivers. Reduced riverine sediment supply caused by upstream water diversion and intensive agriculture, together with port construction and other engineering works, can prevent estuarine replenishment of adjacent beaches and alter local hydrodynamics. Enhanced wave attack on exposed promontories and typhoon-driven storm surge also contribute to shoreline change. Disentangling these influences requires higher-resolution imagery, in situ sediment and hydrodynamic measurements, and explicit mapping of aquaculture and reclamation activities; future work should integrate these data with tidal and wave records to better separate anthropogenic from natural controls.

To ensure sustainable coastal development in Zhuhai, it is essential to address the growing tension between economic expansion and environmental conservation. Based on the identified trends and driving mechanisms, several recommendations are proposed. Firstly, coastal ecological zoning should be strengthened to protect remaining natural coastline types, particularly bedrock and biological coastlines that serve vital ecological functions [44]. Designating conservation zones can help prevent irreversible habitat loss caused by unchecked urban encroachment. Secondly, the city should adopt an Integrated Coastal Zone Management (ICZM) framework to coordinate land use planning, engineering development, and ecological protection. The integration of nature-based solutions such as mangrove restoration and green buffer zones alongside necessary gray infrastructure has proven effective in enhancing coastal resilience [45]. Thirdly, a long-term dynamic monitoring system should be implemented using remote sensing and coastline analysis tools (e.g., DSAS) to detect early warning signs of erosion, coastline degradation, or ecological decline. Recent studies recommend coupling remote sensing data with hydrodynamic models for improved decision-making. Finally, stakeholder participation and policy coordination between municipal departments, developers, and local communities are crucial. Public awareness campaigns and transparent coastline usage regulations can increase compliance and foster a shared responsibility for coastal stewardship.

5. Conclusions

This study reveals that anthropogenic drivers dominate coastline evolution in rapidly urbanizing coastal zones, with implications for global deltaic and bay cities. Analysis of Zhuhai (1987–2022) demonstrates that land reclamation and infrastructure development drove extensive coastline accretion, particularly in southern areas like Hengqin and Gaolan Islands. Crucially, economic forces (GDP, Ex, TFT) outweighed climatic factors (MAT, AP) in steering artificial coastline changes, while Nb correlations confirm port development as a primary morphological driver. These findings highlight a global-scale paradigm: where coastal urbanization intensifies, economic priorities systematically override natural dynamics. The irreversible transition from natural to anthropogenic coastlines underscores an urgent need for strategic balancing of development and ecological resilience—a critical challenge for coastal cities worldwide.