Qionghai and Wanning are located along the eastern coast of Hainan Island (Figure 1) and serve as the core cities within the eastern development cluster of the Hainan Free Trade Port. The permanent venue of the Boao Forum for Asia is situated in this region, granting it significant importance both within China and globally. The study area is characterized by a tropical monsoon climate, with strong monsoonal influence, abundant sunshine, high temperatures, and heavy rainfall. Typhoons occur frequently, and there is no distinct division of the four seasons, but rather a clear alternation between dry and wet seasons. The persistently warm and humid conditions create favorable circumstances for crop growth. However, uneven rainfall distribution often results in droughts and floods, while wind-related disasters (such as Qingming winds, dry hot winds, typhoons, and autumn cold winds) and cold damage caused by low-temperature and cloudy–rainy weather exert additional adverse effects on agricultural production. Topographically, the terrain of the study area slopes from west to east, comprising mainly medium- and low-elevation erosional mountains, denudation hills, undulating erosional plains, basalt platforms, marine depositional landforms, and fluvial erosion–accumulation features. Driven by government policies, population growth, and rapid economic development, the intensity of human activities in the region has increased markedly, placing growing pressure on the local ecological environment.

The study area currently confronts significant ecological and environmental challenges, primarily driven by intensified socioeconomic development. The principal issues include water eutrophication, shoreline erosion and sedimentation, ecological degradation from mining and landfill operations, soil erosion (notably riverbank collapse), land desertification, and heavy metal contamination in soils. These problems are particularly acute in the coastal zone, where unsustainable management of high-elevation aquaculture ponds has triggered seawater intrusion, soil and sediment contamination, and degradation of aquatic environments, ultimately leading to water quality deterioration and lagoon shrinkage. Seawater intrusion has elevated chloride concentrations beyond permissible limits, posing a tangible threat to freshwater resources in coastal areas. The rapid expansion of large-scale marine aquaculture has further exacerbated the problem, causing contamination of surface and groundwater, abandonment of domestic wells, salinization of farmland, and increasing soil desertification. As a result, the natural self-regulatory capacity of the coastal ecosystem has been substantially weakened - or even disrupted - and adjacent coral reef systems have suffered severe degradation, reflecting extensive ecological damage across the region.

The data used in this study primarily comprise spatial and statistical datasets. The spatial data include administrative boundary maps and base geographic maps of eastern Hainan Island. The statistical data encompass eleven key parameters: stratigraphy and lithology, slope, land use type, surface water quality, groundwater quality, soil quality, soil type, population density, and ecosystem type, among others. Specifically, topographic and geomorphological data were derived from a digital elevation model (DEM) with a spatial resolution of 2 meters. Stratigraphic and lithologic information was extracted from geological maps, while data on water and soil quality were obtained through laboratory testing and analytical measurements. The remaining datasets were acquired through field surveys and in situ observations. Details of the data sources are presented in Table 1.

Ecological and environmental quality is influenced by multiple interacting factors. Accordingly, this study developed an ecological environmental quality evaluation framework grounded in ecological–geological survey data. The framework integrates both the ecological and geological conditions of the region, as well as the natural geographical features and human activity characteristics of Qionghai and Wanning. Eleven evaluation indicators were selected to construct the assessment system, including topography and geomorphology, stratigraphy and lithology, soil type, parent material, land use, surface water quality, groundwater quality, soil quality, ecological–geological problems and geological hazards, human activity intensity, and ecosystem type. While “Land Use Type” and “Ecosystem Type” are derived from similar base data, they serve distinct purposes within the index system. “Land Use Type” is an indicator of anthropogenic pressure and management intensity, whereas “Ecosystem Type” serves as a proxy for intrinsic ecological structure, biodiversity potential, and natural resilience. This distinction ensures that both the human footprint on the landscape and the foundational ecological character are independently evaluated, avoiding a singular reliance on either socio-economic or purely natural classification.

The extent and intensity of the influence of these evaluation factors on ecological and environmental quality vary across the study area. Considering the regional characteristics, the Delphi method was employed to construct a hierarchical decision-making model encompassing all selected indicators (Wang et al., 2024b; Bagheri et al., 2024). Experts were invited to evaluate and score the relative importance of each factor in terms of its impact on local ecological environmental quality. Based on these scores, an A–B judgment matrix (Table 2) was established, followed by consistency testing and hierarchical ranking. The finalized judgment matrix was then used to determine the weight of each indicator. Subsequently, using the spatial statistical tools in ArcGIS, an Ecological Environmental Quality Index (EEQI) was developed through a weighted summation model integrating the eleven suitability evaluation factors. Prior to the weighted overlay, all single-factor raster layers were normalized to a common dimensionless scale (0-1) to eliminate unit differences and ensure comparability. The min-max normalization method was applied using the following formula:

This method was chosen for its simplicity and effectiveness in preserving the original relative relationships between scores within each factor. The calculation formula is as follows:

Note:where A represents the comprehensive ecological environmental quality index of the study area; I_i denotes the value of the i-th evaluation indicator; and W_i is the corresponding weight assigned to indicator i.

To ensure the objectivity of the indicator weights, a Delphi method involving a panel of 15 experts was employed. The experts were selected from disciplines of ecology, geology, pedology, environmental science, and GIS. The consultation process consisted of two rounds. In the first round, experts independently scored the factors. The results were aggregated by calculating the arithmetic mean and standard deviation. These statistics, along with anonymous expert comments, were then fed back to the panel in the second round for re-scoring. The final weights were determined as the geometric mean of the second-round scores. This iterative and anonymous procedure helped reduce individual bias and build a robust consensus.

As shown in Table 2, the calculated values are λ = 11.1845, CI = 0.1440, RI = 1.49, and CR = 0.01 < 0.10. Since the consistency ratio (CR) is less than 0.10, the judgment matrix demonstrates good consistency, indicating that the constructed matrix is reasonable and the derived weights are reliable. The consistency ratio (CR = 0.01) confirms the logical coherence of the expert judgments. The derived weights (Table 2) reveal a clear hierarchy of influence among the factors. Ecological and Geological Problems and Hazards (C9, weight = 0.197) emerged as the most influential factor, underscoring that existing environmental degradation (e.g., water pollution, soil erosion) and geological vulnerability are the most direct constraints on ecological quality in this coastal region. This is followed closely by factors representing the current environmental state: Groundwater Quality (C7, 0.169) and Surface Water Quality (C6, 0.128), highlighting the critical role of water resources. Factors related to human pressure (Human Activity Intensity C10 and Land Use C5) and the ecological template (Ecosystem Type C11) also carry substantial weight (0.076–0.120), reflecting the significant impact of anthropogenic activities and the inherent resilience of ecosystems. In contrast, the fundamental geological and pedological background factors (Topography C1, Lithology C2, Soil Type C3, Parent Material C4) received the lowest weights (0.023–0.070), indicating that while they form the foundational matrix, their influence on current ecological quality is more indirect and subordinate to the pressures and states captured by the higher-ranked factors. This weight distribution aligns well with the study area’s context, where acute anthropogenic and biophysical pressures are driving ecological change.

Based on the ArcGIS software platform, the administrative boundaries of Qionghai and Wanning in eastern Hainan Island were used as the base map. Spatial overlay analysis and attribute spatialization were conducted using the collected datasets and the corresponding weights of each evaluation factor. A multi-factor comprehensive evaluation map was then generated to delineate the ecological environmental quality zones within the study area. Specifically, the eleven single-factor evaluation maps of the Qionghai–Wanning coastal zone were converted into raster layers, and spatial weighted overlay calculations were performed according to the weights presented in Table 2. Finally, the composite results were classified into different ecological environmental quality grades following the established classification criteria.

The evaluation unit serves as the fundamental information carrier in ecological environmental quality assessment. Common methods for delineating evaluation units include the grid method, buffer zone method, and vector polygon method. In this study, the grid method was selected for spatial unit division. Comprehensive analysis of the study area’s topography, geomorphology, meteorological conditions, and land-use types revealed that the terrain is relatively gentle, with spatially uniform patterns of rainfall and temperature. Additionally, both stratigraphic–lithologic features and land-use distribution exhibit regular spatial continuity. To balance spatial detail with computational efficiency in this regional assessment, a grid size of 220 m × 220 m was selected. This resolution aligns with the effective scale of primary input datasets—such as the land use/cover classification—ensuring that each unit captures relevant spatial heterogeneity. Considering the spatial extent of the study area, the grid method was applied to divide it into evaluation units of 220 m × 220 m. Each grid cell was treated as an independent information carrier for assessing the ecological environmental vulnerability index, which was subsequently classified to obtain the overall ecological environmental quality assessment results of the study area.

Using a geographic information system, the eleven single-factor ecological environmental quality maps of the eastern coastal zone of Hainan Island were converted into raster format. Each raster layer was weighted according to the factor weights presented in Table 11 and spatially integrated using a weighted summation approach to calculate the Ecological Environmental Quality Index (EEQI) for each evaluation unit. The EEQI values were classified into four grades (Superior, Satisfactory, Moderate, Substandard) using the equal-interval method. This method was chosen for its objectivity, reproducibility, and the clear, fixed thresholds it provides, which are advantageous for environmental monitoring and communicating findings to stakeholders in management and planning contexts. The resulting comprehensive ecological environmental quality map of the study area is presented in Figure 8.

Beyond individual factor assessments, analyzing the spatial correlations among factors is crucial for understanding their synergistic effects on overall ecological quality. A visual comparison of the single-factor maps (Figures 2-7) reveals several pronounced spatial association patterns:

Strong Positive Association (Co-occurrence of Stressors): The spatial distributions of poor groundwater quality (Figure 5A), poor surface water quality (Figure 4B), and high human activity intensity (Figure 6B) exhibit remarkable overlap. These factors collectively form a “coastal degradation complex” concentrated in aquaculture zones and urban peripheries, indicating that anthropogenic pressure drives concurrent deterioration in both surface and subsurface water environments.

Strong Negative Association (Mutual Exclusivity): The high-quality zones for soil (Figure 5B) and vegetation (inferred from Land Use/Ecosystem Type, Figures 4A and 7) are predominantly located in the western forested uplands, which coincide with areas of low human activity intensity and favorable topography (gentle slopes, Figure 2A). This inverse pattern highlights that preserved natural ecosystems are spatially segregated from major anthropogenic disturbances.

Underlying Geological Control: The distribution of less resistant rock types (e.g., loose sediments, Figure 2B) and certain soil parent materials (Figure 3B) provides a foundational template that correlates with the locations of fertile soils and, subsequently, with historical and current agricultural land use. This suggests a cascading influence from geology to pedology to land use.

These observed spatial correlations confirm that the driving factors do not operate in isolation. Instead, they form interlinked clusters—where human pressure clusters with environmental degradation, and natural stability clusters with favorable geological conditions. This interconnectedness reinforces the necessity of the integrated multi-factor approach employed in this study and helps explain the emergence of the distinct EEQI spatial pattern shown in Figure 8.

We developed an integrated evaluation framework combining ArcGIS spatial analysis and the Analytic Hierarchy Process (AHP) to assess the ecological and environmental quality of the eastern coastal zone of Hainan Island. The model incorporates 11 key factors spanning topographic, geological, pedological, hydrological, ecological, and anthropogenic dimensions: topography and geomorphology, stratigraphic lithology, soil type, parent material, land use, surface water quality, groundwater quality, soil quality, ecological and geohazards, human activity intensity, and ecosystem type.

Our assessment reveals that the overall ecological condition of the region is generally favorable. Notably, areas classified as Superior account for 78.86% of the total, predominantly located in forested uplands with minimal human disturbance. The seemingly high proportion of ‘excellent or good’ area alongside the identification of human activity as the dominant driver is explained by the landscape’s structure: acute human pressures are spatially confined to coastal and urban hotspots, while the favorable ratings are sustained by the extensive, less-disturbed forested and agricultural lands that comprise most of the study area. This result is consistent with findings by Peng et al. (2018) in central Hainan, underscoring the robust natural ecological baseline shared across the island’s eastern and central regions. While Peng et al. (2018) established a robust evaluation system for the island’s predominantly natural, mountainous interior, our framework is explicitly designed for the human-nature interface of the coastal zone. By integrating direct anthropogenic pressure indicators (e.g., human activity intensity) with underlying geological vulnerabilities, we not only corroborate the location of high-quality refugia but also diagnose the processes creating sharp ecological gradients at the coast. Thus, we advance from a regional-scale state assessment to a management-scale process diagnostic, providing actionable insights for the island’s most dynamically stressed ecological frontier. Satisfactory (12.52%) are mainly distributed throughout agricultural landscapes, where moderate human activity coexists with largely preserved ecosystem structures. In contrast, zones rated as Moderate (2.35%) and Unsatisfactory (6.27%), though limited in extent, exhibit pronounced spatial clustering that strongly correlates with intense anthropogenic pressure-highlighting human disturbance as a principal factor driving localized ecological degradation.

Spatial analysis reveals a distinct concentration of ecologically substandard areas within the coastal aquaculture and urban zones of Wancheng Town, Dong’ao Town, and Wanning Xiao Hai. This clustered distribution reflects systematic environmental pressures driven by three interconnected factors: Geologically, these regions are characterized by poorly consolidated sediments and interbedded soft-hard rock strata with low weathering resistance, establishing an inherently vulnerable ecological foundation. Remotely sensed and land-use data further demonstrate concentrated built-up areas, diminished vegetation cover, and pronounced landscape fragmentation-factors that collectively compromise ecosystem integrity and resilience. Furthermore, population densities exceeding 1,200 persons/km² sustain intensive anthropogenic activities, notably expansive agriculture and large-scale mariculture, which have persistently degraded soil structure and fertility while amplifying eutrophication, soil salinization, and sediment contamination. This finding aligns with Ma et al. (2025), who documented intensifying human pressure across Hainan Island, and further refines the conclusion by Yu et al. (2024) that anthropogenic disturbance is a critical driver of ecological quality-specifically pinpointing coastal urban and aquacultural zones as priority areas for ecological risk management.

The integrated evaluation framework developed in this study, combining the Analytic Hierarchy Process (AHP) with GIS-based spatial overlay analysis, successfully synthesizes multi-source geospatial data at a 1:50,000 scale to systematically assess the complex coastal ecosystem. This methodology proves both applicable and operationally viable for regional environmental diagnosis. Nevertheless, several limitations should be acknowledged. Certain input data—particularly geological hazard susceptibility—were derived from historical surveys, and their limited temporal resolution may affect the accurate capture of dynamic environmental processes. Furthermore, while surface water quality interpolation was applied across the study area to serve as a continuous influence layer within the overlay model, future studies could employ more hydologically explicit methods, such as confining interpolation to water bodies and modeling land-based influence through defined buffer zones or dispersion models. Additionally, while the 220 m × 220 m evaluation unit was selected based on data compatibility and regional assessment needs, a systematic scale-sensitivity analysis was not conducted. Exploring the impact of different grid resolutions on the assessment outcomes would strengthen the methodological robustness in future applications. Moreover, although the Delphi method was used to consolidate expert knowledge in factor scoring and weight determination, some inherent subjectivity remains in the valuation process. The ecosystem type classification, while informative, is a broad categorization. Future studies would benefit from incorporating more granular metrics such as landscape pattern indices (e.g., patch density, connectivity) or direct biodiversity assessments. To advance this research direction, future work should prioritize the integration of high-resolution remote sensing (e.g., Sentinel-2) to dynamically monitor indicators like NDVI and chlorophyll-a concentration and in-situ sensor networks to facilitate real-time monitoring and updating of critical parameters. The incorporation of machine learning approaches (e.g., Random Forest for feature importance ranking) also holds promise for refining the weight determination process, offering a pathway toward more adaptive, objective, and precise ecological assessment models.