From Conservation to Development: A Study of Land Use and Ecological Changes to Vegetation Around the Hainan Tropical Rainforest National Park

Abstract

Global ecosystems, particularly in biodiversity-rich tropical rainforests, are increasingly under pressure from human activities. As socio-economic development continues and populations steadily grow, the effective planning of areas surrounding national parks has become a global challenge. This study, based on remote sensing data and utilizing landscape ecology tools, such as ArcGIS 10.8, GeoDa 1.20, and Fragstats 4.2, combines spatial statistical methods, trend analysis, and the Hurst index to conduct a long-term analysis and forecast future trends in vegetation ecological quality indicators (VEQI) and landscape pattern changes within and around the Hainan Tropical Rainforest National Park. VEQI changes across various buffer zones were also assessed. Our results show that both arable and built-up land increased, especially from 2002 to 2022. Arable land decreased from 5566.8 km² to 4796.8 km², then increased to 5904.6 km²; built-up land expanded from 163.97 km² to 314.59 km², reflecting urbanization. Spatiotemporal analysis revealed that 42.54% of the study area experienced significant VEQI changes, with a 24.05% increase (mainly in the northwest) and an 18.49% decrease (mainly in the southeast). The VEQI improvements were consistent across all buffer zones, with the most significant growth in the 7.5 km zone. Landscape indices indicated high fragmentation in coastal areas, while inland areas remained stable, reflecting the tension between conservation and urbanization. These findings provide a theoretical basis for future ecological development and buffer zone policies in the park.

1. Introduction

As social development progresses and the demand for ecological conservation increases, national parks and nature reserves are facing substantial challenges, including habitat degradation, fragmentation, and isolation, exacerbated by surrounding developmental activities [1,2,3,4]. Previous studies have demonstrated that land use changes and the intensification of human activities in areas adjacent to national parks and nature reserves often result in the degradation of vegetation ecosystems in these regions [5,6,7]. Establishing buffer zones not only helps balance ecological protection with sustainable resource usage but also facilitates the integration of ecological conservation and economic development through the promotion of ecotourism. Globally, as awareness of ecological conservation and sustainable development grows, an increasing number of countries are expanding and prioritizing the protection of conservation areas [5,6,7]. By implementing these buffer zones, national parks can be integrated into broader ecological protection networks, aligning with global conservation trends and fostering international cooperation in ecological protection [8,9]. The Hainan Tropical Rainforest National Park, known for its rich biodiversity and unique vegetation ecosystems, plays a vital role in vegetation conservation in Hainan, China [10,11]. Thus, analyzing the spatiotemporal dynamics of landscape pattern changes and vegetation ecological quality in the areas surrounding the Hainan Tropical Rainforest National Park is of great significance.

With advancements in remote sensing technology, researchers can now monitor and study vegetation ecology across large spatial scales with greater precision. Earlier studies primarily relied on individual indicators, such as Net Primary Productivity (NPP) and Fractional Vegetation Cover (FVC), to explore the spatial distribution of vegetation’s ecological quality [12,13,14,15,16,17,18]. Deng et al. analyzed the impact of urban land expansion on cropland productivity in Northeast China using NPP data, finding that since 2000, urban expansion has significantly reduced cropland NPP [12]. He et al. examined the relationship between monthly NPP and the Standardized Precipitation-Evapotranspiration Index (SPEI) in the Qinling-Daba Mountains from 2001 to 2018, revealing that vegetation NPP in the region is highly sensitive to drought across multiple time scales [16]. He et al. also analyzed the spatiotemporal variation in FVC on the northern slopes of the Tianshan Mountains from 2001 to 2020, employing a pixel-binary model and Principal Component Analysis (PCA) to identify key drivers, including natural conditions, human activities, and economic growth [17]. Guo et al. investigated the spatiotemporal changes in vegetation cover and their driving factors on the Qinghai–Tibet Plateau from 2001 to 2020, finding an overall increase in vegetation cover, positively influenced by both climatic factors and human activities [18]. However, using single indicators such as FVC or NPP to assess vegetation may lead to incomplete and subjective evaluations, due to factors such as inconsistent NPP values under the same cover type [19]. To address this, Li et al. have proposed a new index that combines FVC and NPP to more comprehensively assess the role of vegetation in supporting environmental functions and biodiversity [20]. Peng et al. conducted a detailed assessment of the impact of major ecological restoration projects on vegetation across China based on the VEQI index [21]. The VEQI, noted for its simplicity and objectivity, has been adopted by China’s meteorological authorities for their annual national vegetation quality assessments.

Landscape indices serve as highly condensed representations of landscape pattern characteristics, providing essential tools for studying landscape patterns. These indices offer quantitative indicators that reflect various aspects of landscape structure, composition, and spatial configuration [22]. However, while landscape pattern optimization is an effective strategy for land use management and enhancing regional ecological protection, it is crucial to further investigate the relationship between landscape metrics and vegetation ecological quality, especially in tropical contexts [23,24]. In particular, understanding how landscape metrics such as fragmentation, connectivity, and patch composition influence vegetation ecological quality in tropical ecosystems remains underexplored. Kubacka et al. assessed changes in landscape patterns within the buffer zones of 159 national parks across 11 European countries from 1990 to 2018. Their analysis revealed that areas with high landscape value surrounding national parks significantly influenced the degree and rate of landscape fragmentation [25]. Zhang et al. developed a risk assessment framework to examine the impact of various human activities on landscape pattern changes within the Northeast China Tiger and Leopard National Park. Their findings indicated that factors such as land use intensity, population density, transportation accessibility, and industrial activities significantly affected the risk zones of landscape fragmentation [26]. Liao et al. analyzed the landscape patterns and geometric forms of 153 urban parks in Changsha, China, revealing the significant role of factors such as park size, tree and water body coverage, and spatial configuration in enhancing the intensity of the park’s cooling island effect. This analysis provided valuable landscape design insights for urban park planning and management [27]. Although these studies emphasize the importance of landscape patterns, they often overlook their direct relationships with vegetation’s ecological quality, particularly in tropical regions. Investigating the spatiotemporal dynamics of landscape patterns and the ecological quality of vegetation around the Hainan Tropical Rainforest National Park is therefore crucial for understanding how landscape metrics influence ecological health, for promoting green development, and for informing tourism resource management and policy formulation.

This study investigates the spatiotemporal dynamics of landscape pattern changes and the ecological quality of vegetation around the Hainan Tropical Rainforest National Park from 2002 to 2022, using the VEQI and landscape pattern indices. Long-term time series analysis is crucial for capturing the cumulative effects of urbanization, agricultural expansion, and policy changes on ecosystem services and ecological quality. Previous studies on landscape dynamics have often focused on short-term trends or isolated snapshots, which may have failed to capture gradual but significant ecological shifts that develop over longer periods. This research aims to bridge this gap by examining long-term trends, future projections, and spatial clustering patterns of VEQI and landscape patterns. The key aims of our research were the following: (1) to identify trends in VEQI and landscape pattern changes around the Hainan Tropical Rainforest National Park from 2002 to 2022; (2) to predict the spatial distribution of future VEQI trends in the park’s surrounding areas; (3) to explore the spatial clustering patterns of VEQI and landscape patterns in these areas; and (4) to provide scientific data to guide the future development of the park and its buffer zones, aligning with ongoing efforts in ecological conservation and sustainable land management.

2. Materials and Methods

2.1. Research Area

The Hainan Tropical Rainforest National Park (HTRNPNP), located on the southern island of Hainan, China, is renowned for its rich biodiversity and its role in preserving tropical rainforest ecosystems, which are among the most ecologically significant in the world. The park spans approximately 4500 square kilometers and plays a critical role in the conservation of tropical flora and fauna, especially endangered species endemic to this region [28]. Situated in the central and southern parts of Hainan, the park’s geographical coordinates range from 18°10′ N to 20°30′ N latitude and 108°50′ E to 110°10′ E longitude. The region experiences a tropical monsoon climate, characterized by high temperatures, abundant rainfall, and distinct wet and dry seasons. Annual precipitation typically ranges from 1500 mm to 2000 mm, and average annual temperatures range between 22 °C and 26 °C [29].

As a key component of China’s growing national park system, the HTRNPNP is embedded within broader efforts to safeguard the nation’s natural heritage. The development of national parks in China has gained prominence in recent years, especially due to the recognition of the urgent need to balance ecological protection with socio-economic development [3]. The establishment of buffer zones surrounding the park—ranging from 2.5 km to 10 km—is crucial for mitigating the impact of surrounding human activities, such as urbanization, agriculture, and infrastructure development. These buffer zones play a vital role in ecological restoration, land management, and maintaining biodiversity by providing a transitional area between human settlements and the core conservation area [30]. They also offer important opportunities for sustainable land use practices, such as eco-tourism and community-driven conservation initiatives.

This study examines not only the core area of the Hainan Tropical Rainforest National Park but also the surrounding buffer zones, which are integral to understanding the broader ecological dynamics, vegetation changes, and the long-term sustainability of the region’s ecosystems. The study area is illustrated in Figure 1.

2.2. Data Sources

In this study, we employed a land use dataset for the Hainan Tropical Rainforest National Park, covering the years 2007, 2012, 2017, and 2022, which was sourced from the ZENODO database [31]. The vector boundary data for the park were obtained from the Resource and Environmental Science Data Center (RESDC) [32]. Additionally, data from NASA’s Moderate Resolution Imaging Spectroradiometer (MODIS), including the Normalized Difference Vegetation Index (NDVI, MOD13A2) [33] and Net Primary Productivity (NPP, MOD17A3) [34], were incorporated into our analysis. These MODIS products were downloaded from the Google Earth Engine (GEE) platform, where they have undergone atmospheric correction and are available as pre-processed products, meaning additional quality assessment steps were not necessary. To ensure consistency and accuracy, all datasets were standardized to the Krasovsky 1940 Albers projection system and resampled to a spatial resolution of 1000 m × 1000 m.

2.3. Methods

This study employs a range of quantitative methods and statistical analyses to comprehensively investigate the VEQI and landscape patterns within and around the Hainan Tropical Rainforest National Park. The research framework is illustrated in Figure 2. First, a grid-based analysis was conducted to assess landscape patterns both within and beyond the park boundaries. Five widely used landscape indices—Patch Density (PD), Landscape Shape Index (LSI), Largest Patch Index (LPI), Contagion Index (CONTAG), and Shannon’s Evenness Index (SHEI)—were selected to quantify landscape evolution. Next, a comprehensive VEQI model was developed using a weighted method, integrating data from the Normalized Difference Vegetation Index (NDVI) and NPP. The weighting criteria for these components were determined based on prior studies, ensuring a balanced contribution from both vegetation productivity and coverage. For spatiotemporal analysis, the Theil–Sen estimator was applied to identify temporal trends in the VEQI and landscape indices. In spatial analysis, both global and local spatial autocorrelation methods were employed. The Global Moran’s I index was used to assess the overall spatial autocorrelation, while Local Indicators of Spatial Association (LISA) maps were applied to detect spatial clustering patterns, identifying high-value and low-value clusters of VEQI and landscape indices. Specifically, LISA cluster analysis was used to examine spatial clustering patterns within the study area, revealing the spatial autocorrelation characteristics of both high-value and low-value clusters. Additionally, the Hurst index was used to evaluate the long-term persistence of time series data. It ranges from 0 to 1, where H > 0.5 indicates a continued trend, while H < 0.5 suggests a potential reversal. This method provides insight into the stability of landscape changes and vegetation’s ecological quality over time.

2.3.1. Quantification of Landscape Patterns

The grid for the same area was constructed using ArcGIS, referencing the “Geographic Grid” [35] (GB12409-2009) and studies by other scholars. The grid size is recommended to be 2 to 5 times the average patch area [36]. In this study, the ecological risk unit was defined as a 2 km × 2 km ecological risk zone, with a total of 4963 units. Kriging interpolation was then applied to generate spatial distribution maps of landscape pattern indices for the regions surrounding the Hainan Tropical Rainforest National Park. Landscape pattern quantification primarily relies on landscape indices; however, due to the large number of available indices, some of which exhibit high correlations, the careful selection of appropriate indices is essential to ensure the scientific rigor of landscape evolution analysis [37]. To provide a comprehensive evaluation, this study employed five widely used landscape metrics—the Patch Density (PD), the Landscape Shape Index (LSI), the Largest Patch Index (LPI), the Contagion Index (CONTAG), and Shannon’s Evenness Index (SHEI)—to quantify landscape dynamics [38,39,40]. The PD indicates the degree of landscape fragmentation, which is critical for assessing habitat loss and the impacts of human activities. The LSI measures the complexity of patch shapes, which is important for understanding edge effects and ecosystem stability. The LPI evaluates the size and ecological integrity of the largest patches, which are vital for maintaining biodiversity. The CONTAG quantifies habitat connectivity, a key factor for species migration and gene flow. Finally, the SHEI reflects the distribution and diversity of vegetation types within the landscape, providing insight into the ecological stability and service functions of the ecosystem (

Table 1. Interpretation table for landscape metrics.

Metric Name	Formulas and Explanations
Patch Density (PD)	$PD=\left(NP/A\right)$
	In the formula: $A$ represents the area of the landscape or a specific patch type; $PD$ denotes the number of patches per unit area, with higher $PD$ values indicating greater landscape fragmentation.
Landscape Shape Index (LSI)	$LSI={\displaystyle \frac{e_i}{min{e}_i}}{e}_i(NoneLSI\ge 1,withoutlimit)$
	In the formula: $e_i$ represents the total edge length (or perimeter) of type $i$; $min{e}_i$ denotes the minimum total edge length (or perimeter) for category $i$. A higher $LSI$ value indicates greater landscape fragmentation.
Largest Patch Index (LPI)	$LPI={\displaystyle \frac{a_{max}}{a_{total}}}\times 100\%(Percent0<LPI\le 100)$
	In the formula: $a_{max}$ represents the maximum patch area of the landscape or a specific patch type, and $a_{total}$ denotes the total area of the landscape or a specific patch type; $LPI$ reflects the dominance of a specific patch type within the landscape.
Contagion Index (CONTAG)	$CONTAG=\left[1+{\displaystyle \frac{\sum _{i=1}^m\sum _{k=1}^m\left[\left(p_i\right)\left(\frac{g_{ik}}{\sum _{k=1}^m{g}_{ik}}\right)\right]\left[in\left(p_i\right)\left(\frac{g_{ik}}{\sum _{k=1}^m{g}_{ik}}\right)\right]}{2in\left(m\right)}}\right]\times 100,(0<CONTAG\le 100)$
	$\mathrm{In}\mathrm{the}\mathrm{formula}:p_i$$\mathrm{represents}\mathrm{the}\mathrm{percentage}\mathrm{of}\mathrm{the}\mathrm{area}\mathrm{occupied}\mathrm{by}\mathrm{patch}\mathrm{type}i$$;g_{ik}$$\mathrm{denotes}\mathrm{the}\mathrm{number}\mathrm{of}\mathrm{adjacent}\mathrm{patches}\mathrm{between}\mathrm{type}i$$\mathrm{and}\mathrm{type}k$$;m$$\mathrm{represents}\mathrm{the}\mathrm{number}\mathrm{of}\mathrm{patch}\mathrm{types}\mathrm{in}\mathrm{the}\mathrm{landscape};CONTAG$$\mathrm{reflects}\mathrm{the}\mathrm{degree}\mathrm{of}\mathrm{patch}\mathrm{aggregation}\mathrm{or}\mathrm{the}\mathrm{tendency}\mathrm{for}\mathrm{extension}\mathrm{within}\mathrm{the}\mathrm{landscape}.\mathrm{A}\mathrm{higher}CONTAG$ value indicates that the landscape is composed of dominant patch types with good connectivity, while a lower value suggests that the landscape consists of multiple patch types with poor connectivity.
Shannon’s Evenness Index (SHEI)	$SHEI=-{\displaystyle \frac{\sum _{i=1}^m{p}_{i}in{p}_i}{inm}}$
	In the formula: pi refers to the proportion of patch type i within the landscape; m denotes the total number of patches. A small SHEI value indicates that the landscape is dominated by one or a few dominant patch types, whereas a larger SHEI value suggests a more even distribution of patch types across the landscape.

).

2.3.2. Quantification of Vegetation Ecological Quality

FVC is defined as the proportion of total land area covered by vegetation in a specific region. It serves as a key indicator for assessing the level and density of vegetation coverage and is widely used in ecological and environmental monitoring. Building on this concept, pixel decomposition techniques analyze the spectral information of individual pixels in satellite remote sensing images, allowing for effective identification and quantification of vegetation and non-vegetation components. This enables the derivation of FVC and other regional vegetation cover indices. In this study, we utilize MODIS satellite monthly composite data as the basis and apply the mixed pixel decomposition method to calculate FVC [41,42]. The specific formula is as follows:

$$\mathrm{F}\mathrm{V}\mathrm{C}={\displaystyle \frac{\mathrm{N}\mathrm{D}\mathrm{V}\mathrm{I}-{\mathrm{N}\mathrm{D}\mathrm{V}\mathrm{I}}_{\mathrm{s}\mathrm{o}\mathrm{i}\mathrm{l}}}{{\mathrm{N}\mathrm{D}\mathrm{V}\mathrm{I}}_{\mathrm{v}\mathrm{e}\mathrm{g}}-{\mathrm{N}\mathrm{D}\mathrm{V}\mathrm{I}}_{\mathrm{s}\mathrm{o}\mathrm{i}\mathrm{l}}}}\times 100\%$$

(1)

In this formula, the Normalized Difference Vegetation Index (NDVI) is an important indicator of vegetation growth status. It is calculated by comparing the reflectance differences between infrared and visible light for plants. Additionally, ${\mathrm{N}\mathrm{D}\mathrm{V}\mathrm{I}}_{\mathrm{s}\mathrm{o}\mathrm{i}\mathrm{l}}$, which represents the NDVI value under bare soil conditions with no vegetation cover, reflects the reflective properties of the soil surface. ${\mathrm{N}\mathrm{D}\mathrm{V}\mathrm{I}}_{\mathrm{v}\mathrm{e}\mathrm{g}}$, on the other hand, corresponds to the NDVI value for pixels under full vegetation cover. In this study, ${\mathrm{N}\mathrm{D}\mathrm{V}\mathrm{I}}_{\mathrm{s}\mathrm{o}\mathrm{i}\mathrm{l}}$ and ${\mathrm{N}\mathrm{D}\mathrm{V}\mathrm{I}}_{\mathrm{v}\mathrm{e}\mathrm{g}}$ are set to 0.05 and 0.95, respectively.

NPP and FVC are two key indicators for evaluating terrestrial ecosystem services and assessing vegetation ecological quality. NPP reflects the production potential of an ecosystem, while FVC represents its extent of coverage. However, relying solely on NPP or FVC for assessment only provides a partial view of vegetation ecological quality. To offer a more comprehensive analysis of the ecological quality of plant communities in the Hainan Tropical Rainforest National Park and their spatiotemporal distribution patterns, this study constructs the VEQI for the park, based on FVC and NPP data, using a weighted method. This model focuses on vegetation’s production and coverage capacities, offering a more holistic approach to evaluating vegetation ecological quality [41,42]. The specific calculation formula is as follows:

$$VEQI=\left(f_1\times FVC+f_2\times {\displaystyle \frac{NPP}{{NPP}_{max}}}\right)\times 100$$

(2)

In this formula, VEQI is a quantifiable measure of annual vegetation ecological quality, with values ranging from 0 to 100. The VEQI calculation is based on NPP and FVC. An increase in VEQI indicates a significant improvement in vegetation’s ecological quality. Specifically, a VEQI value of 0 means both NPP and FVC are at their lowest levels in the region, while a VEQI value of 100 indicates that NPP has reached its historical optimal level and FVC is at full coverage. The weight coefficients for NPP and FVC are denoted as $f_1$ and $f_2$, respectively. The determination of these coefficients depends on the ecological characteristics and research objectives of the study area. In this study, both weight coefficients are set to 0.5, following a methodology that has been successfully applied in analyzing the spatiotemporal distribution characteristics of vegetation VEQI across China [36].

2.3.3. Trend Analysis

The Theil–Sen estimator is a non-parametric statistical method used to determine the median slope of a trend line in a set of data points. This method, named after the independent work of Hendrik Theil and Sen Pranab Kumar, is widely used for analyzing trends in data [43,44,45]. In this study, we apply the Theil–Sen Median trend analysis method to calculate the trend changes of the VEQI and landscape pattern indices in the study area from 2002 to 2022. The specific formula is as follows:

$${\beta }_{slope}=Median\left({\displaystyle \frac{x_j-x_i}{j-i}}\right),\hspace{1em}2002\le i<j\le 2022$$

(3)

In the formula, $x_j$ and $x_i$ represent the VEQI and landscape pattern indices for the $i$-th and $i$-th years, respectively. If the Slope < 0, the VEQI and landscape pattern indices exhibit a downward trend over the time series; if the Slope > 0, the indices show an upward trend.

In this study, the Mann–Kendall test is employed to assess the statistical significance of the VEQI trend. The test statistics and are calculated as follows:

$$S={\sum }_{i=1}^{n-1}{\sum }_{j=i+1}^{n}sgn\left(x_j-x_i\right).$$

(4)

$$sgn\left(X_j-X_i\right)\left\{\begin{array}{c}1,\hspace{1em}{x}_j-x_i>0\\ 0,\hspace{1em}{x}_j-x_i=0\\ -1,\hspace{1em}{x}_j-x_i<0\end{array}\right.$$

(5)

$$Var\left(S\right)={\displaystyle \frac{n\left(n-1\right)\left(2n+5\right)}{18}}$$

(6)

$$Z_{Slope}=\left\{\begin{array}{c}{\displaystyle \frac{S-1}{\sqrt{Var\left(S\right)}}},\hspace{1em}S>0\\ 0,\hspace{1em}S=0\\ {\displaystyle \frac{S+1}{\sqrt{Var\left(S\right)}}},\hspace{1em}S<0\end{array}\right.$$

(7)

In the formula, $x_i$ and $x_j$ represent the VEQI and landscape pattern indices corresponding to years i and j in the time series, respectively, and $\mathrm{n}$ denotes the length of the time series, which is 22 years. The significance of the VEQI change trend in the study area from 2002 to 2022 is assessed at the 0.05 confidence level, when $Z_{Slope}>1.96$ indicates significant change. When $\left|Z_{Slope}\right|$ > Z_1-α/2, it suggests that the VEQI and landscape pattern indices in the region have undergone significant changes during the study period; otherwise, the changes are considered minor. In this paper, a change trend is defined as “highly significant” at the 0.01 confidence level and “significant” at the 0.05 confidence level, that is when $Z_{Slope}$ > 1.96, indicating the significance of the VEQI and landscape pattern index change trends in China from 2002 to 2022.

2.3.4. Spatial Autocorrelation Analysis

Global Spatial Autocorrelation Analysis: Global spatial autocorrelation provides a comprehensive description of the spatial relationships and differences across a region. The most commonly used statistic for this purpose is Global Moran’s I, which quantifies the average degree of spatial association between all spatial units and their surrounding areas across the entire region [46,47,48]. The Moran’s I index ranges from [−1, 1]. A positive value of Moran’s I indicates a spatially positive correlation of the variable, meaning that similar values tend to cluster in proximity. Conversely, a negative value suggests a spatially negative correlation, where dissimilar values are more likely to occur near each other. When Moran’s I equals zero, it indicates no spatial correlation between the variable and its neighbors. In this study, the Moran’s I index is used to explore the spatial correlations between the VEQI and landscape pattern indices in the areas surrounding the Hainan Tropical Rainforest National Park.

Local Spatial Autocorrelation Analysis: Local spatial autocorrelation analysis primarily examines the degree of aggregation of local spatial units with their neighboring units and their distribution patterns. It includes two main types of relationships: spatial coherence (high–high clustering, low–low clustering) and spatial trade-offs (high–low clustering, low–high clustering) [46,47,48]. High–high clustering refers to areas where both the study unit and its neighboring units exhibit strong functionality. Low–low clustering indicates areas where both the study unit and its neighbors show weak functionality. High–low clustering represents areas where the study unit has strong functionality, but its neighbors have weak functionality, while low–high clustering indicates the opposite. The Local Indicators of Spatial Association (LISA) map effectively identifies the spatial distribution of high and low value zones, facilitating the detection of local spatial anomalies. In this study, at the 2 km × 2 km grid scale, LISA maps are used to identify the local spatial clustering patterns of VEQI and landscape pattern indices in the vicinity of the Hainan Tropical Rainforest National Park. We used a contiguity-based spatial weights matrix, where neighboring grid cells sharing a border or vertex are considered spatially related. This approach is suitable for our analysis, as it reflects the geographic relationships between 2 km × 2 km grid cells in the study area. Other matrix types, such as distance-based matrices, could be explored in future research to assess how different spatial structures affect the results.

3. Results

3.1. Analysis of Land Use Dynamic Changes

As shown in

Table 2. Area of each land use type (10⁴ km²).

Land Use Type	2002	2012	2022
Cropland	5566.768	4796.811	5904.563
Forestland	11,248.508	11,971.644	10,877.705
Shrub land	1.519	0.962	0.761
Grassland	66.630	30.651	6.837
Water area	266.711	294.010	237.050
Bare land	16.799	3.105	1.432
Impervious Surface	163.971	233.722	314.594

, significant changes in land use occurred in the areas surrounding the Hainan Tropical Rainforest National Park between 2002, 2012, and 2022. The area of cultivated land decreased from 5566.768 km² in 2002 to 4796.811 km² in 2012, before rising again to 5904.563 km² in 2022, reflecting a trend of initial decline followed by an increase in the use of agricultural land. Forest area, on the other hand, first increased and then decreased over the same period. It rose from 11,248.508 km² in 2002 to 11,971.644 km² in 2012, only to decrease to 10,877.705 km² in 2022. Meanwhile, the area of impermeable surfaces (urbanized land) continued to grow, doubling from 163.971 km² in 2002 to 314.594 km² in 2022, reflecting rapid urbanization and infrastructure development in the region. These changes in land use reflect the impacts of regional development policies, resource management practices, and environmental conservation measures [49].

As depicted in Figure 3, forests have demonstrated relatively high stability over the two decades, particularly in terms of area maintained as forested land. From 2002 to 2012, 10,682.3007 km² of forest remained unchanged, and from 2012 to 2022, 10,274.031 km² of forest remained intact. This indicates that the majority of forested areas have been well preserved. However, forest conversion to agricultural land was observed during both decades, with 553.9059 km² converted in the first decade and 1676.7162 km² in the second decade, signaling agricultural expansion in certain areas. The changes in cultivated land were also particularly notable, especially in terms of its conversion to impermeable surfaces (urbanized or construction land). Between 2002 and 2012, 60.2865 km² of cultivated land was converted to built-up areas, and this figure rose to 88.6005 km² in the subsequent decade, reflecting the ongoing impact of urbanization on agricultural land. The conversion of cultivated land to forests also highlights the trend of forest restoration, with 1279.0017 km² converted from cultivated land to forest between 2002 and 2012, while this figure decreased to 579.8196 km² between 2012 and 2022. Water bodies remained relatively stable across both periods, suggesting the potential success of water resource management strategies. From 2002 to 2012, 244.3221 km² of water remained unchanged, while between 2012 and 2022, 211.1283 km² of water bodies had maintained stability. Additionally, the conversion of grasslands and bare land demonstrated their sensitivity to urban expansion and other land-use pressures. The conversion of grasslands to impermeable surfaces during the second decade was 5.4018 km², relatively small but symbolically significant, indicating that even ecologically sensitive areas have been impacted by urbanization. Overall, the land use changes between 2002 and 2022 reflect a dynamic balance between the utilization and conservation of natural resources, while also revealing the interactions and conflicts between urbanization, agricultural activities, and the management of natural conservation areas [50,51].

3.2. Spatiotemporal Variation Analysis of the VEQI

As shown in Figure 4a, the VEQI of the areas surrounding the Hainan Tropical Rainforest National Park significantly exceeded that of the coastal regions between 2002 and 2022. Notably, the VEQI of Qiongzhong Li and Miao Autonomous County and Baoting Li and Miao Autonomous County was significantly higher than that of other areas, while the the ecological quality of vegetation in the coastal regions was notably lower. As illustrated in Figure 4b, the spatiotemporal trend of the VEQI in the surrounding areas of the Hainan Tropical Rainforest National Park from 2002 to 2022 exhibited a pattern of increasing values in the northwest and decreasing values in the southeast. A total of 42.54% of the study area experienced significant changes in the ecological quality of vegetation, while 57.46% remained stable. Specifically, approximately 24.05% (4111 km²) of the area showed a significant upward trend in the VEQI, primarily concentrated in the northwest region of the study area. In contrast, around 18.49% (3161 km²) of the area exhibited a significant decline in the VEQI, with the regions of decline concentrated in the eastern and southern parts of the study area.

Furthermore, we analyzed trends in the VEQI within the 2.5 km, 5 km, 7.5 km, and 10 km buffer zones surrounding the park from 2002 to 2022, as shown in Figure 5. The results indicate that while all the buffer zones experienced improvements in the VEQI, the magnitude of change varied. Among the buffer zones, the 7.5 km zone exhibited the highest percentage increase (10.190%), followed by the 10 km (8.934%), 5 km (8.775%), and 2.5 km (8.510%) zones. The highest average annual growth rate was recorded in the 7.5 km buffer zone (0.486%), while the lowest was observed in the 2.5 km zone (0.409%). According to our analysis, while all buffer zones showed improvements in the VEQI, the intensity of change differed across zones. The 7.5 km buffer zone experienced the greatest increase (10.190%), whereas the 2.5 km zone showed the smallest growth (8.510%). This variation may be attributed to several factors. First, the differences in conservation policies around the park’s core area could have played a significant role. The 2.5 km zone, being closest to the core, is subject to stricter conservation measures, which, while limiting human intervention, might also constrain positive ecological developments. In contrast, the 5 km and 7.5 km buffer zones likely benefited from more flexible land-use policies and eco-friendly land management practices, which contributed to more significant improvements in the VEQI. The strong recovery observed in the 10 km buffer zone could be attributed to large-scale ecological restoration projects aimed at enhancing vegetation cover beyond the immediate vicinity of the park, further driving significant improvements in VEQI in this zone.

Overall, despite experiencing declines in VEQI from 2012 to 2017 across all buffer zones, each region achieved positive growth from 2002 to 2022, demonstrating a continued improvement in vegetation quality. Among the buffer zones, the 2.5 km zone, closest to the core of the national park, exhibited the highest VEQI, while the 7.5 km zone experienced the most significant growth over the entire 20-year period.

3.3. Spatiotemporal Changes in Landscape Patterns

The landscape pattern indices of the study area exhibit distinct spatial differentiation characteristics. Spatially, as shown in Figure 6, the indices of PD, LSI, and SHEI generally exhibit a decreasing trend from the coastal areas to the inland, while the indices of the LPI and CONTAG show the opposite trend. Temporally, over the past two decades, the landscape pattern of the study area has fluctuated significantly, although the overall changes have been relatively minor. Fluctuations in the PD, LSI, and CONTAG were more pronounced, while LPI and SHEI remained relatively stable. As depicted in Figure 6, the SHEI of the areas surrounding the Hainan Tropical Rainforest National Park shows an overall upward trend, indicating that between 2002 and 2022, the fragmentation of landscape patches in the park’s vicinity increased, land use became more diversified, and the landscape types evolved toward a more balanced and equitable state. The landscape spatial structure has continued to exhibit diversity and uniformity. Between 2002 and 2022, the trend of LPI in the study area shows a spatial pattern of higher values in the north and lower values in the south, while the trend of LSI exhibits the opposite pattern. This suggests that landscape fragmentation in the northern part of the Hainan Tropical Rainforest National Park has eased, whereas fragmentation in the southern region has increased. The PD and CONTAG exhibited an overall downward trend from 2002 to 2022, indicating a decline in landscape connectivity around the national park.

To further investigate the landscape pattern trends around the Hainan Tropical Rainforest National Park, we analyzed the landscape pattern indices for the 2.5 km, 5 km, 7.5 km, and 10 km buffer zones around the park. As shown in Figure 7, the indices of the LSI, SHEI, and CONTAG are the lowest in the 2.5 km buffer zone compared to the other three buffer zones, and these values gradually increase with distance from the park. In contrast, the PD and LPI in the 2.5 km buffer zone are the highest among all the buffer zones, and LPI increases with increasing distance from the park. These observations suggest that landscape fragmentation is relatively low within the 2.5 km buffer zone, and fragmentation increases as the distance from the park increases.

However, it is important to note that while these findings suggest lower fragmentation near the park, they do not fully account for potential confounding factors, such as variations in human activity, land management practices, or natural environmental differences that could also influence landscape patterns. For example, the relative ecological integrity in the 2.5 km buffer zone may be influenced by stricter conservation policies, but other factors such as agricultural land use, urbanization pressures, and infrastructure development in surrounding areas could also contribute to the observed trends. Therefore, while the 2.5 km buffer zone shows relatively low levels of landscape fragmentation, these patterns should be interpreted with caution, as they may reflect a combination of both conservation efforts and other external influences.

3.4. Correlation Analysis Between the VEQI and Landscape Patterns

Table 3. Bivariate global Moran’s I values for vegetation ecological quality and landscape pattern indices in the area surrounding the Hainan Tropical Rainforest from 2000 to 2020.

Years	PD	LPI	LSI	CONTAG	SHEI
2002	−0.136	0.358	−0.396	0.088	−0.36
2007	−0.136	0.365	−0.375	0.078	−0.371
2012	−0.141	0.353	−0.351	0.099	−0.35
2017	−0.139	0.309	−0.31	0.125	−0.282
2022	−0.174	0.288	−0.236	0.097	−0.229

presents the bivariate global Moran’s I values between vegetation ecological quality and landscape pattern indices in the surrounding areas of the Hainan Tropical Rainforest from 2000 to 2020. The indices analyzed include the PD, LPI, LSI, CONTAG, and SHEI. Moran’s I is a statistical measure used to assess the spatial correlation between these indices, where positive values indicate a positive spatial correlation and negative values indicate a negative spatial correlation.

From 2002 to 2022, the Moran’s I value for the PD remained negative, decreasing from −0.136 to −0.174, reflecting a 27.9% increase in spatial negative correlation with other landscape indices, suggesting a growing spatial dispersion in patch distribution. In contrast, the Moran’s I value for the LPI consistently remained positive, but declined from 0.358 to 0.288, a reduction of approximately 19.6%. This indicates that while the positive spatial correlation between the largest patch and other landscape elements still exists, its intensity has weakened. The Moran’s I value for the LSI increased from −0.396 to −0.236, showing a 40.4% rise, which suggests a reduction in the negative spatial correlation with other landscape attributes, indicating a trend towards more regularized landscape shapes.

Similarly, the Moran’s I values for the SHEI and VEQI were negative but increased from −0.36 to −0.229, reflecting a 36.4% improvement. This suggests that the distribution of patch types within the landscape has become more evenly distributed. The Moran’s I value for the CONTAG rose from 0.088 to 0.125, an increase of 42.0%, indicating a growing aggregation of landscape elements.

In summary, these data reflect the complex and dynamic changes in the landscape pattern of the Hainan Tropical Rainforest region over the past two decades. This includes increasing landscape fragmentation, gradual ecological balance, regularization of landscape shapes, and the close relationship between these changes and the region’s natural development as well as human interventions.

From a temporal perspective (Figure 8), the spatial clustering patterns between the VEQI and various landscape pattern indices in the surrounding areas of the Hainan Tropical Rainforest National Park remained relatively stable from 2002 to 2022. Throughout this period, the spatial clustering patterns of the VEQI with LSI, PD, and SHEI in the eastern and southern regions of the park were predominantly characterized by “high–low” clustering. This indicates that these areas exhibited high ecological quality but also had relatively high landscape fragmentation and low biodiversity evenness. Specifically, “high–low” clustering suggests that while ecological quality was high, the fragmentation of the landscape resulted in a less continuous ecosystem, leading to challenges in maintaining biodiversity and ecological stability. In contrast, the northern region of the park showed “low–low” clustering between the VEQI and LSI, PD, and SHEI, reflecting poorer ecological quality, greater landscape fragmentation, and uneven distribution of biodiversity. The “low–low” clustering in this region suggests a more degraded ecosystem with significant disturbances and fragmentation.

Notably, the spatial clustering patterns of the VEQI with the LPI and CONTAG in the eastern and southern regions primarily exhibited “high–high” clustering. This suggests that these areas possess vegetation of a high ecological quality, with large, continuous landscape patches, which are often associated with healthier and more intact ecosystems. The “high–high” clustering indicates that both ecological quality and landscape connectivity are high, promoting better ecosystem functioning and resilience. Conversely, in the coastal areas, the spatial clustering patterns between the VEQI and LPI were predominantly “low–low”, indicating lower ecological quality and a lack of large, contiguous ecological patches. This pattern likely reflects the high developmental pressures in coastal regions, which have led to habitat destruction and a significant reduction in ecological connectivity. The “low–low” clustering here highlights the negative effects of urbanization and development on the landscape, leading to fragmentation and diminished ecological quality.

From a temporal perspective, the spatial autocorrelation of the trend changes in the VEQI with landscape indices highlights significant spatial dependencies in the evolving landscape patterns. For instance, the “high–high” clustering of the VEQI with LPI and CONTAG in the eastern and southern regions indicates areas where both ecological quality and landscape connectivity have improved over time. This suggests that these regions have experienced positive trends in both vegetation quality and landscape configuration, which are characteristic of healthy and well-connected ecosystems. On the other hand, the “low–low” clustering observed in coastal areas indicates negative trends in both the VEQI and landscape connectivity, reflecting ecological degradation and fragmentation likely driven by development pressures. Further analysis of the relationship between the VEQI and LSI, PD, and SHEI reveals weak spatial autocorrelation of the trends, with “high–low” and “low–low” clustering patterns observed. This suggests that while some regions with higher ecological quality maintain stable landscape patterns, others with high fragmentation and low ecological quality show negative trends, further complicating the landscape dynamics.

4. Discussion

4.1. Prediction of Future VEQI Trends

Studying the future trend of the VEQI in the surrounding areas of the national park is crucial for the implementation of future conservation policies and the establishment of buffer zones. To this end, we simulated the future trend of the VEQI by calculating the Hurst exponent for the 10 km buffer zone surrounding the Hainan Tropical Rainforest. The Hurst exponent is commonly used to describe the sustainability of time series data, and in this study, it was employed to model future changes in the VEQI for the regions surrounding the Hainan Tropical Rainforest [52,53,54]. In our analysis, we ensured that the time series data did not contain any missing values or outliers. No data gaps were present, and potential outliers were identified and addressed through standard data cleaning procedures, ensuring the robustness and accuracy of the Hurst exponent calculation.

The basic principle of the Hurst exponent is as follows: For a given time series variable $\left\{\mathsf{\xi }\right(\mathrm{t}\left)\right\},\mathrm{t}=1,2,\dots ,$ for any given moment $\mathsf{\tau }\ge 1$, define its mean sequence as:

$${〈\xi 〉}_{\tau }={\displaystyle \frac{1}{\tau }}{\sum }_{t=1}^{\tau }\xi \left(t\right),\hspace{1em}\tau =1,\hspace{1em}2,\hspace{1em}\dots$$

(8)

Cumulative deviation sequence:

$$X\left(t,\tau \right)={\sum }_{u=1}^t\left(\xi \left(u\right)-{〈\xi 〉}_{\tau }\right),\hspace{1em}1\le t\le \tau$$

(9)

Range sequence:

$$R\left(\tau \right)={}_{1\le t\le \tau }{}^{max}X\left(t,\tau \right)-{}_{1\le t\le \tau }{}^{min}X\left(t,\tau \right),\hspace{1em}\tau =1,\hspace{1em}2,\hspace{1em}\dots$$

(10)

Standard deviation sequence:

$$S\left(\tau \right)={\left[{\displaystyle \frac{1}{\tau }}{\sum }_{t=1}^{\tau }{\left(\xi \left(t\right)-{〈\xi 〉}_{\tau }\right)}^2\right]}^{{\displaystyle \frac{1}{2}}},\hspace{1em}\tau =1,\hspace{1em}2,\hspace{1em}\dots$$

(11)

If $\mathrm{R}/\mathrm{S}\propto {\mathsf{\tau }}^{\mathrm{H}}$, it indicates the presence of Hurst phenomenon in the time series. The Hurst exponent, denoted as H, can be derived using least squares fitting based on ${\mathrm{l}\mathrm{o}\mathrm{g}\left(\mathrm{R}/\mathrm{S}\right)}_{\mathrm{n}}=\mathrm{a}+\mathrm{H}\times \mathrm{l}\mathrm{o}\mathrm{g}\left(\mathrm{n}\right)$. The Hurst exponent ranges from 0 to 1. When 0.5 < H < 1, it indicates positive persistence in the time series, meaning that the future trend is consistent with the past, and the data exhibit strong long-term autocorrelation. The closer H is to 1, the stronger the continuity of the data. When H = 0.5, it indicates that the time series is random with no long-term correlation, meaning the future trend is independent of the past, and the data exhibit no long-term memory. When 0 < H < 0.5, it indicates negative persistence, meaning that the future trend is contrary to the past, and the data show strong long-term anti-correlation. The closer H is to 0, the stronger the negative persistence.

Figure 9 illustrates the projected future changes in the VEQI across different buffer zones. In all buffer zones, the most frequent trend observed is “Up → Down”, particularly in the 2.5 km buffer zone, where 1508 km² is predicted to experience a reversal from an increasing to a decreasing trend. This suggests that these areas may face a potential decline in ecological quality. In contrast, the “Up → Up” trend occurs less frequently, with the majority concentrated in the western regions of the National Park and the 10 km buffer zone, indicating that areas farther from the park center may sustain a continuous improvement in vegetation ecological quality. The frequency of the “Down → Down” trend is relatively low across all buffer zones, suggesting that a persistent decline is not widespread. Notably, only 65 km² in the 10 km buffer zone is predicted to continue declining, which indicates that most areas will not experience long-term ecological degradation. The frequency of the “Down → Up” trend is fairly balanced across all buffer zones, with a concentration in the eastern part of the National Park, suggesting that these areas may recover and shift from a declining trend to an upward trajectory. Overall, regions closer to the center of the Hainan Tropical Rainforest National Park exhibit greater ecological dynamics, particularly within the 2.5 km and 5 km buffer zones. This indicates that these areas may be more strongly influenced by human activities or ecological pressures. As the distance from the park center increases, especially in the 7.5 km and 10 km buffer zones, the negative trends decrease, suggesting that these more distant areas may possess greater ecological stability and potential for recovery.

Therefore, when formulating strategies for the development of the Hainan Tropical Rainforest National Park and the establishment of its buffer zones, it is essential to account for the spatial variations and ecological risks revealed by the VEQI trend analysis. Specifically, the 2.5 km and 5 km buffer zones exhibit a marked “Up → Down” trend, indicating that areas closer to the park’s core are facing higher risks of ecological degradation, likely due to intensified human activities or insufficient environmental management measures [54]. To mitigate these risks, it is recommended to strengthen conservation efforts in these regions, such as implementing stricter land-use regulations and increasing ecological restoration projects, to alleviate pressures from human activities and promote the stability and recovery of ecological quality [55]. In contrast, the 10 km buffer zone, particularly its western regions, shows a more pronounced “Up → Up” trend, suggesting that these areas have significant potential for sustained improvement in ecological quality. In these regions, it is advisable to promote sustainable development practices and the construction of eco-friendly infrastructure. Additionally, these areas could be prioritized for ecological tourism or educational initiatives to leverage their ecological assets and foster sustainable regional economic development. For the eastern regions, which exhibit a “Down → Up” recovery trend, focus should be placed on ecological monitoring and management. In these areas, ecological interventions such as reforestation and habitat restoration could be implemented to accelerate the recovery of ecosystems from their degraded states [56].

4.2. Policy Recommendations

The study of vegetation’s ecological quality in the Hainan Tropical Rainforest National Park and its surrounding buffer zones reveals a significant improvement in overall vegetation quality from 2002 to 2022. Notably, the 2.5 km zone closest to the park center experienced the most pronounced increase in VEQI, suggesting that conservation measures implemented in these areas have been highly effective. In contrast, the declining trend in vegetation quality in the eastern and southern regions highlights the need for reinforced ecological protection and management in these areas. The spatial and trend analysis, along with the Hurst index predictions, suggests that areas showing declining VEQI trends may face long-term ecological degradation if not addressed promptly. Therefore, it is crucial to prioritize more stringent environmental regulations, including the restriction of high-impact human activities, such as large-scale land development and unsustainable agricultural practices, to protect and enhance ecological quality in these areas [57]. Moreover, while previous studies have primarily focused on vegetation dynamics in specific regions, our research places particular emphasis on the long-term spatiotemporal analysis and future predictions for vegetation’s ecological quality in the buffer zones surrounding national parks [58,59]. The 7.5 km buffer zone, which exhibited the most significant growth over the 20-year period, demonstrates the effectiveness of ecological restoration and conservation strategies. This zone’s improvement, coupled with the long-term sustainability indicated by the Hurst index analysis, suggests that continued and expanded conservation efforts in this area are essential. It is recommended to review and enhance existing conservation measures to ensure their adaptability to evolving environmental pressures. This may include expanding conservation strategies to neighboring buffer zones, as well as addressing specific challenges such as fragmentation and biodiversity loss. Given the promising outcomes within the 7.5 km zone, extending restoration efforts and strengthening ecological corridors between zones could promote further integration of the landscape, thus enhancing resilience to ecological pressures.

Additionally, considering the generally lower ecological quality of vegetation in the coastal regions, there is a need for stronger management and establishment of coastal ecological reserves. The Hurst index predictions indicate potential long-term trends of degradation, emphasizing the importance of sustainable coastal protection. Promoting eco-tourism and environmental education initiatives in these regions can also play a critical role in raising public awareness, while simultaneously contributing to conservation funding and community engagement [60,61]. Furthermore, urbanization and its associated impacts on buffer zone effectiveness require a more detailed exploration. As urban development in surrounding areas increases, it is essential to account for the pressures that urbanization places on buffer zones. The expansion of infrastructure and residential areas often leads to habitat fragmentation, which could undermine the effectiveness of current buffer zones. More targeted studies on urbanization’s effects on buffer zone integrity are necessary to refine management strategies. Collaboration with urban planners and local governments is essential to ensure that buffer zones are properly integrated into urban development plans, allowing for sustainable coexistence between urban growth and conservation efforts.

To effectively implement these recommendations, it is crucial to address the potential barriers and challenges. One primary obstacle may be the resistance from local communities that rely on land-use practices such as agriculture and development for their livelihoods. Collaboration with these communities is essential to achieving sustainable conservation goals. Establishing partnerships with local stakeholders—including farmers, landowners, and businesses—could help create a shared understanding of the importance of preserving biodiversity while also supporting local economic growth. For example, offering incentives such as eco-friendly farming subsidies or promoting sustainable land management practices can encourage community involvement in conservation initiatives. Another barrier may arise from limited financial resources and insufficient infrastructure to support large-scale ecological restoration and conservation projects. To address this, a more detailed funding strategy should be developed, involving collaboration between government agencies, non-governmental organizations (NGOs), and private sector stakeholders. Additionally, leveraging international financial support and green development funds could help address these financial gaps. Educational initiatives are key to overcoming long-term ecological challenges. In addition to eco-tourism, incorporating environmental education into school curricula and community-based workshops could raise awareness about the importance of maintaining ecological balance and ecosystem services. These initiatives should focus on engaging the younger generation to build a culture of ecological stewardship, which will help ensure that conservation practices become an integral part of local values and everyday life. Furthermore, community-led monitoring programs could foster a sense of ownership and responsibility among local residents, encouraging them to take an active role in monitoring the ecological health of the park and its buffer zones.

Finally, ensuring that conservation measures are adaptable and flexible to evolving environmental challenges is critical. A robust monitoring system should be implemented to track vegetation quality, landscape fragmentation, and biodiversity trends, allowing for timely adjustments to management strategies. Regular assessments and stakeholder consultations will be necessary to refine conservation policies and ensure that they remain effective as both environmental and socio-economic conditions change. By integrating these comprehensive measures—community collaboration, education, financial support, and adaptive management—the unique biodiversity of the Hainan Tropical Rainforest National Park can be more effectively protected, while promoting sustainable development and ecological stability in the surrounding areas [62,63,64].

4.3. Limitations and Future Research Directions

This study offers valuable insights into the spatiotemporal dynamics of vegetation ecological quality and landscape patterns in the buffer zones of the Hainan Tropical Rainforest National Park. However, several limitations must be addressed in future research. First, the spatial resolution of the VEQI data was limited to 1000 m, which may impact the precision of the analysis. Future studies should use higher resolution data for more detailed and accurate insights. Additionally, our study focused on a single national park and its surrounding buffer zones, which limits our ability to generalize our findings to other national parks. Future research should investigate buffer zones in different tropical protected areas worldwide to assess the applicability of these findings across varied ecological and socio-economic contexts. The identification of the 7.5 km buffer zone as an optimal area for conservation efforts requires further ecological and socioeconomic analysis. While the VEQI showed significant improvement in this zone, further studies on ecological connectivity, species migration, and human activity impacts are necessary to confirm the effectiveness of this distance for conservation. Furthermore, the time series analysis did not specifically account for potential temporal autocorrelation, which could influence the accuracy of trend predictions. Future research should incorporate advanced statistical methods, such as autoregressive moving average (ARMA) models or generalized least squares (GLS), to address this limitation. Finally, this study relies primarily on the VEQI as a key indicator of vegetation quality, but it does not account for other essential ecological factors such as soil quality, hydrology, and species diversity. Future research should incorporate these additional indicators for a more comprehensive assessment of ecosystem health. Moreover, the impact of urbanization on buffer zone effectiveness warrants more detailed exploration in future studies to better understand the challenges posed by development in protected areas.

5. Conclusions

This study examined land use changes at five time points between 2002 and 2022 in the Hainan Tropical Rainforest National Park and its surrounding areas. A variety of software tools, including ArcGIS, GeoDa, and Fragstats, were employed to conduct landscape pattern analysis, spatial statistical analysis, trend analysis, and Hurst index analysis. The primary objective was to investigate the spatiotemporal evolution of land use changes and their effects on landscape patterns and ecosystem service values within the study area. The key findings are as follows:

(1) From 2002 to 2022, the vegetation ecological quality in the Hainan Tropical Rainforest National Park and its surrounding buffer zones showed significant changes both spatially and temporally. The 2.5 km buffer zone, which is closest to the park, experienced the most pronounced improvement in VEQI. However, this zone is at risk of a potential decline in vegetation quality in the future due to increasing pressures or challenges, highlighting the need for ongoing management. In contrast, the 7.5 km and 10 km buffer zones demonstrated sustained improvements in VEQI, with the 7.5 km zone showing especially notable progress. This suggests that ecological restoration strategies implemented in these areas have been highly effective.

(2) Landscape pattern indices in the surrounding areas of the Hainan Tropical Rainforest National Park showed clear spatial differentiation. The indices of PD, LSI, and SHEI generally decreased from coastal to inland areas, while the Largest Patch Index (LPI) and CONTAG exhibited an increasing trend. Despite some fluctuations in these indices, the overall changes were relatively modest. The 2.5 km buffer zone had the lowest level of landscape fragmentation, with fragmentation increasing progressively further from the park’s center. This suggests that areas close to the park maintain higher ecological integrity, while more distant areas face growing fragmentation pressures.

(3) Significant regional differences were observed in the VEQI, LSI, PD, and SHEI in the surrounding areas of the Hainan Tropical Rainforest National Park. The eastern and southern regions exhibited a “high–low” clustering pattern for the VEQI, LSI, PD, and SHEI, indicating high ecological quality but considerable landscape fragmentation. Conversely, the same regions showed a “high–high” clustering pattern between the VEQI and LPI/CONTAG, reflecting not only high ecological quality but also larger, continuous landscape patches. In contrast, coastal areas exhibited a “low–low” clustering pattern between the VEQI and LPI, indicating lower ecological quality and reduced landscape connectivity due to the impacts of human development.

To further improve ecological quality in the buffer zones, it is crucial to implement targeted restoration measures, particularly in areas experiencing degradation (such as the 2.5 km zone). Promoting eco-friendly land-use practices, increasing afforestation efforts, and managing human interventions will be essential for balancing conservation and development. Moreover, continued monitoring and adaptive management should be prioritized to mitigate potential negative trends and enhance landscape connectivity, particularly in fragmented coastal zones.

This study fills a critical gap in understanding land use and ecological changes around the Hainan Tropical Rainforest National Park, providing a foundation for future ecological management strategies to preserve biodiversity and ecosystem services.