Ecological adaptability assessment for adaptive management of island social–ecological systems: a potential–constraint framework applied to Dongtou District, China

Introduction: Artificial Island social–ecological systems face escalating, compound pressures from global environmental change and intensive human activities, while conventional post-hoc restoration approaches are often insufficient to ensure long-term sustainability. We propose an ecological adaptability assessment framework to support proactive, policy-responsive adaptive management.

Methods: We develop a potential–constraint framework that couples scenario-based spatial modeling with explicit policy feedback. Using Dongtou District (Zhejiang Province, China) as a representative case, we quantify trade-offs between ecological resilience and development potential under alternative policy trajectories. A multidimensional indicator system—covering habitat structure, ecosystem function, socio-economic conditions, marine utilization, and governance responses—is integrated into a gridded modeling platform to identify adaptive management pathways toward 2030.

Results: Ecological adaptability exhibits pronounced spatial heterogeneity: high-adaptability zones align with ecological reserves, whereas low-adaptability zones cluster on densely inhabited and industrialized islands. Scenario simulations indicate that policy orientation decisively reshapes adaptability patterns. Among the evaluated pathways, the eco–economic coordination scenario delivers the most balanced configuration, improving spatial connectivity and mitigating landscape fragmentation.

Discussion: The proposed framework provides a transferable, policy-linked approach that bridges ecological evaluation with spatial governance, offering actionable evidence to design and prioritize island adaptive management strategies.

1 Introduction

Under accelerating global environmental change and intensifying land–sea interactions, adaptive management has become a core paradigm for guiding island and coastal governance (Salafsky et al., 2001; Stankey et al., 2005; Rist et al., 2013; Ki-Moon, 2014). In this study, we focus on ecological adaptability assessment as the diagnostic core of adaptive management for island social–ecological systems. The foundational concept of “adaptation” originated in evolutionary ecology, describing how organisms or systems adjust to cope with environmental changes (Futuyama, 1979). Subsequently, Odum (1971) integrated adaptive principles into ecosystem theory by linking material, energy, and information flows with concepts of resilience and equilibrium. Building on this intellectual lineage, Holling (1978) proposed the concept of adaptive management, which Lee (1994) later formalized, emphasized its iterative cycle of monitoring, experimentation, evaluation, and adjustment as a necessary approach to dealing with ecological complexity and uncertainty. Within global change science, adaptation has become a central theme (Intergovernmental Panel on Climate Change (IPCC), 2001, Intergovernmental Panel on Climate Change (IPCC), 2007; Smit and Wandel, 2006), characterized by the overarching aim of reducing system vulnerability while simultaneously enhancing learning capacity and dynamic adjustment.

A rich body of work on ecological suitability and related assessment frameworks can be regarded as the intellectual precursors of ecological adaptability assessment. Early approaches such as McHarg’s classic “sieve mapping” and the FAO Land Evaluation system, largely focused on land suitability analysis at broad spatial scales (McHarg, 1969). Subsequent developments incorporated ecological carrying capacity models, multi-criteria suitability indices, and composite indicators that combine biophysical and socio-economic information to guide land-use planning and environmental management (Qian et al., 2017; Medland et al., 2020; Dang and Kawasaki, 2017; Yu, 1996; Wang et al., 2025). More recent contributions have begun to incorporate notions of ecosystem resilience, vulnerability, and adaptive capacity into spatial assessment frameworks (Boyd and Folke, 2011), particularly for sensitive coastal and island settings (Mcleod et al., 2019). Critically, however, most existing frameworks still emphasize static ecological conditions or isolated environmental indicators, often treating human development as an external pressure rather than an integral, co-evolving component of the system’s adaptive dynamics (Young, 2010). Consequently, they possess limited capacity to fully capture the essential feedbacks between ecological resilience and development dynamics, or to represent the fine-scale spatial trade-offs between ecological potential and critical environmental constraints.

These limitations are particularly acute in island social–ecological systems, which are simultaneously exposed to climate change and intensive development pressures. Island ecosystems are characterized by small area, isolation, limited resource endowment, and strong dependence on external inputs, making them among the most vulnerable socio-ecological systems (Nurse et al., 2014; Baldacchino, 2018). Their land–sea interfaces function as complex transition zones where terrestrial, coastal, and marine processes converge. Cross-boundary fluxes of sediment, nutrients, and pollutants, together with tidal and wave dynamics, generate strong biophysical coupling and nonlinear feedbacks that propagate across ecological and administrative boundaries (Ramesh et al., 2015; Barragán and De Andrés, 2015; Glavovic et al., 2014). Consequently, compounded pressures—including rising sea level, salinization, biodiversity loss, and socio-economic dependency—are exceedingly difficult to isolate and manage effectively within conventional sectoral or jurisdictional frameworks. This complexity poses persistent challenges for integrated island governance (Masselink et al., 2020; Duvat et al., 2021; Mycoo, 2018).

Despite the growing recognition of these challenges, most existing studies on island or coastal management remain predominantly focused on specific environmental elements—such as water quality, fisheries, or shoreline dynamics. Relatively fewer, however, have developed fully integrated, ecosystem-based frameworks that explicitly link ecological and socio-economic processes (Jiang et al., 2018; Gao et al., 2018; Semeoshenkova et al., 2016; McLeod and Leslie, 2009; Mumby et al., 2014). Marine ecosystems are inherently open, dynamic, and nonlinear, characterized by strong feedbacks between human activities and ecological processes—conditions that significantly amplify management uncertainty (Shen et al, 2009; Zhai et al., 2025; Sue et al., 2011). Island systems further intensify these challenges due to their bounded carrying capacity and the acute risk of crossing irreversible thresholds in land–sea interactions. Traditional post-hoc restoration approaches are therefore increasingly inadequate. An urgent shift is necessary toward forward-looking, system-oriented frameworks that can quantitatively assess ecological adaptability and explicitly reveal trade-offs between ecological resilience and development demands, thereby supporting adaptive management under accelerating land–sea coupling and climate-induced pressures. These forward-looking approaches are increasingly framed through ecosystem-services thinking and nature-based solutions to strengthen resilience while delivering co-benefits (Abernethy, 1999; Reyers et al., 2015; Cohen-Shacham et al., 2016; Cohen-Shacham et al., 2019; Seddon et al., 2020). They also commonly adopt scenario-based spatial planning and landscape modeling to explore development–conservation trade-offs across alternative trajectories (Steinitz, 1990; Decker et al., 1996; Hulse et al., 1997; Peter et al., 2002; Nichols and Williams, 2006; Verboom and Wamelink, 1999; Adams et al., 2016).

To address this critical gap, the present study introduces a potential–constraint framework for ecological adaptability assessment in island social–ecological systems. In this framework, ecological adaptability is conceptualized as emerging from the dynamic balance between two complementary dimensions: the “potential” dimension, which reflects intrinsic development capacity and opportunities, and the “constraint” dimension, which represents critical environmental limits and vulnerability. Specifically, resource potential (e.g., ecological quality, habitat connectivity, locational accessibility, and development opportunities) and environmental or anthropogenic constraints (e.g., pollution, ecological degradation, carrying-capacity thresholds, and policy-based ecological redlines) are rigorously quantified through a unified indicator system. These are then integrated within a spatial modeling platform to derive a comprehensive adaptability index. By jointly mapping potential and constraint, the framework makes explicit the inherent trade-off between development opportunities and ecological limitations, identifying zones where high potential and low constraint suggest adaptive development space, and areas where low potential and high constraint signal the need for strict ecological protection. In this way, the potential–constraint framework links ecological adaptability assessment directly to adaptive management by providing a spatially explicit, operational basis for zoning, phasing, and prioritizing island governance interventions.

In operational terms, we use ecological adaptability as the entry point to adaptive management. In this study, ecological adaptability is defined as the capacity of an island social–ecological system to maintain, or deliberately shift toward, desirable states under interacting environmental and development pressures (Walker et al., 2004). Conceptually, it extends classical notions of suitability and carrying capacity by embedding resilience and adaptive capacity: suitability focuses on how well current conditions support a given use, carrying capacity identifies limits that should not be exceeded, whereas ecological adaptability emphasizes whether and where the system can be adjusted through spatial reallocation and management intervention without crossing critical thresholds. Framed in this way, ecological adaptability assessment forms the diagnostic and feedback core of an adaptive management cycle (monitoring–spatial assessment–scenario comparison–zoning and staged policy adjustment) (Maor Cohen et al., 2020), providing spatially explicit evidence on which areas should be strictly protected, which can accommodate moderate development, and which require targeted restoration or risk mitigation (Zhang et al., 2023).In this context, Dongtou District in Wenzhou, Zhejiang Province, China, serves as a representative case study. The district encompasses a cluster of inhabited and uninhabited islands characterized by high ecological value but simultaneously facing intense development pressures, including port construction, tourism, aquaculture, and coastal engineering. Using Dongtou as a testbed, this study aims to:

1. Construct an indicator system and potential–constraint framework for ecological adaptability assessment in island social–ecological systems;

2. Map the baseline spatial pattern of ecological adaptability, identifying key spatial gradients and hotspots;

3. Simulate alternative 2030 scenarios that reflect different policy orientations and development intensities, and evaluate how ecological adaptability patterns reorganize under these trajectories; and

4. Derive management-relevant zones and actionable recommendations to support the adaptive management of island ecosystems.

By explicitly centering on ecological adaptability assessment and embedding it within an adaptive management context, this study seeks to provide an operational and transferable framework for integrated island governance in dynamic marine systems.

2 Study area and data sources

2.1 Study area profile

Dongtou District (27°41′19″ N–28°01′10″ N, 120°59′45″ E–121°15′58″ E), located in Zhejiang Province, China, is situated at the intersection of the Yangtze River Delta and the Western Taiwan Strait economic zones (Figure 1). The district encompasses 302 islands with a total land–sea area of 2,862 km², including approximately 2,700 km² of marine waters. The study area delineation includes two towns and five subdistricts, excluding Lingkun Subdistrict. Situated along the southern coast of Zhejiang, it is one of China’s few county-level administrative units consisting primarily of islands. Dongtou administers over 300 islands and approximately 351 km of coastline, forming two core port areas of Wenzhou Port and representing a highly characteristic land–sea coupled island region suitable for integrated management studies (Zhejiang Provincial Government, 2021).

Figure 1

Figure 1. Location and basic geographic features of the Dongtou district study area.

Ecologically, Dongtou hosts a rich a mosaic of ecosystems, including rocky and sandy coastlines, extensive tidal flats, coastal wetlands, and distinctive island vegetation, largely within the designated Dongtou National Marine Park system (Deng et al., 2024). These ecosystems provide essential functions notably shoreline protection, carbon sequestration, and critical nursery habitats for marine species (Cao et al., 2018). However, rapid coastal reclamation, uncontrolled aquaculture expansion, and intensive harbor construction have significantly intensified anthropogenic pressures on limited coastal resources, leading to habitat loss and ecosystem degradation (Wang, 2024). These dual characteristics- high ecological value and intense development pressure- establish Dongtou as a representative case study for examining the ecological adaptability of island social–ecological systems under coupled development and conservation demands.

2.2 Data assembly and baseline selection

To facilitate the support spatially explicit modeling and scenario analysis, we assembled and integrated a comprehensive set of multi-source datasets into a unified spatial framework. The datasets utilized in this study are grouped into five main categories:

1. Baseline Geographic and Spatial Data: Basic geographic information for Dongtou, including island and coastline distributions, reclaimed tidal flats, and the location of nature reserves and scenic areas. These datasets were utilized to characterize the locational context of the study area, patterns of land and sea uses, ecologically sensitive zones, and key conservation targets.

2. Ecological and Environmental Data: Indices of phytoplankton, zooplankton, and benthic diversity, a eutrophication index, and net primary productivity (NPP), along with supporting variables such as dissolved oxygen, nutrients, chlorophyll-a, and suspended sediment. These variables were derived from spring and autumn field surveys conducted in 2020 at 35 monitoring stations, covering intertidal, nearshore, and offshore zones (Supplementary Appendix). Point measurements were subsequently spatially interpolated to produce continuous ecological surfaces.

3. Socio-economic Data, Indicators such as the gross value added of marine industries, reclamation and coastal development intensity, and patterns of sea-use types. These were compiled primarily from the Wenzhou Statistical Yearbook, the Dongtou Statistical Yearbook, and official Marine Bulletins for the period around 2020.

4. Remote Sensing Data: Landsat 8 OLI and Sentinel-2 MSI imagery (10–30 m resolution) for 2020, which were radiometrically and atmospherically corrected to extract coastline positions, land-cover information, and vegetation indices; and DMSP/OLS nighttime-light imagery (provided by the National Geophysical Data Center, NGDC), which was used to represent the intensity and spatial distribution of human activities.

5. Policy and Planning Documents: Documents such as the Overall Development Plan of Dongtou County, the Master Plan for Scenic and Historic Interest Areas, and provincial-level coastal and marine spatial plans, provided by local government departments. These documents served as the basis for deriving, indicators such as the proportion of natural coastline retained and the spatial distribution of coastal risk.

All disparate datasets were spatially harmonized to a 30 m × 30 m grid and rigorously checked for consistency prior to indicator construction (Section 3.4) (Chen et al., 2023).

The year 2020 was selected as the baseline because it provides the most complete and high-quality spatial data and strategically aligns with the launch of Zhejiang’s “Ecological Civilization Demonstration Zone” and the 14th Five-Year Marine Development Plan (2021–2025) (Ma et al., 2022). This selection ensures the representativeness and comparability of the baseline data for the 2030 scenario simulations.

3 Methods

3.1 Operational framework for ecological adaptability assessment

Grounding the analysis in the conceptual definition of ecological adaptability and the potential–constraint idea introduced in Section 1, this study presents an operational framework that links data integration, indicator construction, potential–constraint modeling, and scenario analysis in a clear, stepwise manner (Figure 2). The overarching aim of this framework is to translate the abstract notion of “ecological adaptability” into a spatially explicit index that can effectively support adaptive planning for island social–ecological systems.

Figure 2

Figure 2. Operational framework for island ecological adaptability assessment and management.

First, multi-source spatial datasets describing habitat structure, ecosystem function, socio-economic conditions, sea-use patterns, and policy–management settings (Section 2) are harmonized to a common 30m × 30m grid. Based on this integration, we construct a comprehensive five-dimension, 13-indicator system (Section3.4; Table 1) that effectively captures both the structural attributes and functional processes of the island system.

Table 1

Table 1. Evaluation index system for island ecological adaptability.

Second, all indicators are categorized into potential-oriented, constraint-oriented, and policy-response components according to their respective functional roles in the system: potential-oriented indicators (e.g., accessibility, marine industrial development, human activity intensity) reflect the intrinsic capacity and extrinsic opportunities for socio-economic development; constraint-oriented indicators (e.g., ecological quality, biodiversity, environmental risk) delineate critical environmental limits and system vulnerability; policy-response indicators (e.g., ecological red-line coverage, protected-area ratio, shoreline protection) encode existing management interventions and mediate the relationship between development potential and ecological constraints.

Third, the composite potential and constraint indices are computed through a rigorous process of standardized scoring and weighting procedures. Indicator values are first normalized to comparable scales, then aggregated dimensionally, and finally integrated combined to obtain composite potential and constraint indices for every grid cell. These two indices are then mathematically combined into a single Ecological Adaptability Index (EAI), which reflects the relative ability of each location to maintain or strategically shift towards desirable states under coupled environmental and development pressures. This index forms the critical basis for identifying high-adaptability zones (high potential, low constraint) and ecologically fragile zones requiring strict protection (low potential, high constraint).

Finally, the adaptability framework is applied within a scenario-based evaluation context. Alternative 2030 scenarios—such as ecological conservation, economic development, and eco–economic coordination—are developed by modifying selected potential-oriented and constraint-oriented indicators based on predefined policy orientations and development intensities. The resulting changes in potential, constraint, and adaptability indices are mapped and compared to reveal ecological–economic trade-offs, potential spatial shifts in adaptability patterns, and priority areas for anticipatory adaptive management (Sections 3.2 and 4).

In this manner, the operational framework in Figure 2 connects the conceptual background presented in the introduction with a transparent, reproducible workflow that can be readily implemented in geographic information systems. It provides the methodological core of the evaluation–feedback–management model used to support adaptive decision-making for island ecosystems.

Within this structure, the potential–constraint framework occupies the assessment–feedback component of an adaptive management cycle. Monitoring programs and statistical reporting provide the input indicators; the potential–constraint model synthesizes them into baseline and scenario-based adaptability maps; and these maps are then translated into management zones, development–conservation trade-offs, and staged policy adjustments in Sections 4 and 5. As new monitoring data and policy objectives become available, the same workflow can be rerun, allowing ecological adaptability assessment to function as a recurring diagnostic step that continuously informs and updates adaptive island governance.

3.2 Integration of policy directives and multi-level constraints

This section details the integration of ecological constraints and multi-level policy directives from national, provincial, and local planning documents to systematically construct the development scenarios, using the baseline ecological adaptability pattern derived from Section 3.1. Within this study, ecological constraints are defined as the biophysical and environmental limits—including resource carrying capacity, habitat sensitivity, and ecological redlines—within which human activities and ecosystem processes can operate sustainably. These constraints are fundamentally combined with development potential to derive a spatially explicit Ecological Adaptability Index, which reflects the capacity of each location to maintain or strategically shift towards desirable states under multiple stressors.

3.2.1 Operationalizing policy integration

To operationalize the integration of ecological constraints and policy directives, we established a multi-level crosswalk between the spatial ecological baselines and officially released planning documents.

National level: The Marine Ecological Red Line Plan (2018) and the National Marine Functional Zoning Plan (2020–2030) were utilized to delineate mandatory ecological protection zones, development-restricted areas, and key functional zones.

Provincial level: The Zhejiang Provincial Coastal Zone and Marine Spatial Plan (2021–2035) and the Zhejiang Ecological Civilization Demonstration Zone Implementation Plan (2020) provided quantitative targets for ecological redline coverage, marine resource utilization intensity, and industrial layout.

Local level: The Dongtou District Marine Ecological Redline Delimitation (2020) and the Wenzhou Territorial Spatial Master Plan (2021–2035) were incorporated to refine boundary constraints and ensure alignment with local management regulations.

All pertinent policy layers were vectorized, rasterized to a 30m × 30m grid, and overlaid with the ecological constraint dataset (e.g., carrying capacity, community structure, environmental quality, shoreline protection zones) to form a composite policy–constraint layer. This layer serves two main methodological functions in the modeling workflow:

Anchoring High-Constraint Areas:. Ecological redline zones, core conservation areas, and highly sensitive habitats are locked as high-constraint cells across all scenarios. These zones are explicitly prevented from being downgraded by any development-oriented assumptions, thereby rigorously enforcing externally defined ecological bottom lines.

Defining policy-consistent constraint levels: The continuous constraint index is strategically classified into five ordinal levels (very low, low, medium, high, very high) using thresholds that are consistent with official planning targets (e.g., redline coverage proportion, seawater quality standards). These discrete levels are then combined with the similarly classified development potential index in a scenario-specific interaction matrix.

3.2.2 The potential–constraint interaction matrix

On this basis, we constructed a 5 × 5 interaction matrix that links five levels of development potential with five levels of ecological constraints under three policy orientations: Ecological Protection Priority (E), Ecological–Economic Coordination (S), and Development Priority (D). For each combination of potential level (1–5) and constraint level (1–5), the matrix assigns an adaptability-based management class (1–5) specific to each scenario (E/S/D). This integrative mechanism ensures that areas designated as strictly protected under national or provincial regulations remain in high-constraint, conservation-oriented management classes in all scenarios, whereas development-permitted or buffer zones are permitted to shift among different management classes depending on the specified policy scenario.

This method thus ensures that scenario construction reflects a coherent blend of top-down policy controls and bottom-up ecological thresholds, significantly improving the coherence between scientific modeling and regulatory practice.

3.3 Scenario simulation and adaptive management pathway

Accordingly, three distinct policy-oriented development scenarios were designed based on the integration of the potential–constraint framework and the official planning documents detailed in Section 3.2:

Ecological Protection Priority (Scenario E): This scenario emphasizes strict ecological conservation, enforcing minimal new reclamation and strong restrictions on high-intensity marine use. The core objective is to maximize ecosystem service provision and habitat integrity by prioritizing ecological protection over economic expansion.

Ecological–Economic Coordination (Scenario S): This pathway prioritizes ecological protection while simultaneously permitting selective and spatially optimized development. It actively seeks a balanced and sustainable configuration that harmonizes development potential and ecological constraints.

Development Priority (Scenario D): This scenario aims to maximize development intensity (e.g., land reclamation, port and industrial construction) within the absolute ecological and human safety bottom lines defined by policy. This allows for greater economic expansion, particularly in areas characterized by relatively lower ecological constraints.

3.3.1 Indicator aggregation and quantification

All 13 indicators described in Section 3.4 are initially aggregated into two composite metrics: a development potential index and an ecological constraint index:

Potential-oriented indicators include metrics such as reclamation intensity, marine industrial value added, human activity intensity, and locational accessibility, collectively reflecting the capacity and opportunities for socio-economic development.

Constraint-oriented indicators include crucial ecological variables like community structure, environmental quality, and production supply, alongside policy-based limits such as shoreline protection.

Community structure is represented by the mean Shannon–Wiener diversity index of phytoplankton, zooplankton, and benthic macrofauna measured at each station.

Environmental quality is captured by a compound eutrophication index (EI) based on chemical oxygen demand (COD), dissolved inorganic nitrogen (DIN), and phosphate (PO₄^3-);

Production supply is represented by net primary productivity (NPP), which was estimated from chlorophyll-a and in situ primary productivity measurements using Cadee’s productivity index method.

These ecological variables were measured at 35 monitoring stations during spring and autumn 2020, and subsequently interpolated to continuous surfaces using ordinary kriging. The exact formulas, station locations, and variogram settings are provided in Section 2 and the Supplementary Appendix to ensure full reproducibility.

3.3.2 Scenario implementation via interaction matrix

The continuous potential and constraint indices are then classified into five ordinal levels (1 = very low, 5 = very high). Scenario simulation is implemented through the 5×5 interaction matrix (Table 2), which serves as the core mechanism for policy translation: for each 30m × 30m grid cell, we first identify its development potential level (1-5) and ecological constraint level (1-5) under the 2020 baseline, using the normalized and classified indices(Sections 3.4 and 3.5); using Table 2, we assign an adaptability-based management class (1-5) to that cell for each scenario (E, S, D).

Table 2

Table 2. Scenario-specific interaction matrix for adaptability-based management zones.

Crucially, the three scenarios share the same underlying spatial distribution of potential and constraint (i.e., the 2020 baseline surfaces). They differ only in how the interaction matrix translates each potential–constraint combination into final management classes based on the under different policy orientations.

In practical terms, Scenario E is more conservative: for cells with high constraints or those with both high potential and high constraints, the interaction matrix more restrictive, conservation-oriented management classes, prioritizing ecological protection. Scenario D is more permissive, allowing for greater development space classes (i.e., lower development space) for cells with high development potential and moderate constraints, while strictly adhering to redline-based high-constraint zones. Scenario S adopts an intermediate configuration, balancing ecological and economic objectives by strategically reallocating development toward lower-constraint areas and strengthening protection in high-constraint, high-potential zones.

3.3.3 Methodological assumptions and validation

It is important to note that the ecological input variables (such as community structure, environmental quality, and production supply) are not separately forecasted through process-based ecological models for 2030 horizon. Instead, their 2020 spatial patterns—captured in the constraint index derived from the 35-station survey and kriging interpolation—are assumed to remain within the same ordinal classes over the medium-term horizon. Scenario differences therefore arise primarily from the alternative management decisions encoded in the interaction matrix (e.g., stricter or more permissive treatment of high-potential/high-constraint areas) and from adjustments to socio-economic and marine-use indicators guided by planning targets. This design reflects the reality that, over a 10-year horizon, policy choices and spatial planning often exert a more immediate and controllable influence on island development trajectories than fully predictable ecological dynamics.

Finally, the scenario-specific adaptability maps are used to analyze spatial transitions in ecological adaptability under alternative policy orientations and development intensities. By comparing the outcomes of E, S, and D, we quantify ecological–economic trade-offs, identify priority zones for strict protection or adaptive development, and delineate adaptability-based management patterns aligned with sustainability targets. Baseline model performance was evaluated through consistency tests between observed (2020) spatial patterns and the classified adaptability map, yielding a Kappa coefficient of 0.82, which indicates satisfactory accuracy and provides a sound basis for the subsequent scenario analysis.

3.4 Indicator system design and quantification

The design of the indicator system is guided by international standards, such as the EU Water Framework Directive (WFD) and the Marine Strategy Framework Directive (MSFD), which mandate that ecosystem assessments integrate both structural and functional attributes and utilize policy-relevant and testable indicators (European Commission (EC), 2017; Van Leeuwen et al., 2014; Safi et al., 2019). Leveraging techniques like ecological-factor overlay, logical rule combination, and pattern–process coupling models (Lin et al., 2016), and recognizing the high uncertainty and spatial variability inherent in island marine ecosystems (Muniz et al., 2011; Wang et al., 2015), this study integrates multi-indicator evaluation with rule-based logical synthesis to construct a spatially explicit, feedback-oriented adaptive assessment framework.

3.4.1 Indicator system design

Our comprehensive evaluation–feedback–management system comprises five analytical dimensions and 13 measurable indicators (Table 1; see also Supplementary Table A1), guided by the principles of objectivity, integrity, hierarchy, land–sea coordination, and operability. This multi-dimensional framework effectively captures the current ecosystem state, its functional trajectory, and the coupled influences of policy and human activities (Huo, 2010; Liu et al., 2017a).

The five analytical dimensions describe the structural composition of the island social–ecological system, while the potential–constraint classification defines the functional roles of indicators:

Potential-oriented indicators (e.g., reclamation intensity, marine industrial value added) reflect the system’s development capacity and extrinsic opportunities (socio-economic conditions and marine utilization dimensions).

Constraint-oriented indicators (e.g., community structure, eutrophication level, net primary productivity) represent environmental limits and vulnerability (habitat structure and habitat function dimensions).

Policy-response indicators (e.g., ecological red-line coverage, protected-area ratio, shoreline protection) provide regulatory feedback, mediating the relationship between development potential and ecological constraints (policy response dimension).

Data derivation for these indicators was carried out as follows: habitat structure indicators were extracted from remote-sensing imagery and basic geographic data; ecosystem function indicators were calculated from ecological and environmental monitoring data (e.g., biodiversity, eutrophication, and primary productivity); socio-economic indicators were obtained from official statistical yearbooks and spatially allocated via population grids and land-use overlays; marine-utilization indicators were compiled from marine functional zoning, sea-use records, and port development statistics; and policy-response indicators were derived from coastal and marine spatial plans, ecological red-line delineations, and related planning documents. These diverse sources ensure consistent spatial and temporal coverage across the natural and human subsystems, guaranteeing that the five-dimension, 13-indicator framework is both comprehensive and reproducible.

3.4.2 Indicator quantification

Four complementary methodological approaches were applied to quantitatively derive the 13 indicators for the 30m × 30m grid:

Spatial operations: Metrics such as natural coastline retention rate, marine risk distribution index, reclamation intensity, and marine industrial value added were derived from vectorized baseline data and geospatial overlay analyses.

Semi-quantitative assignment: Indicators like ecological protection value and marine-use coordination were assessed through expert scoring based on the distribution of protected areas and sensitive zones.

Model-based calculations: Biodiversity indices, the eutrophication index, and net primary productivity (NPP) were estimated using standard mathematical models and ordinary kriging spatial interpolation based on field survey data from 35 monitoring stations.

Comprehensive measures: Locational accessibility and population density were derived from network accessibility analysis and DMSP/OLS nighttime-light data (Fang et al., 2004; Zou et al., 1983; Liu et al., 2003; Liu et al., 2017a).

This integrated framework provides a robust foundation for assessing the ecological adaptability of island social–ecological systems and forms the analytical core of the adaptive management system developed in this study.

3.5 Indicator normalization and weighting

3.5.1 Normalization and rescaling

All indicators were standardized based on either resource supply–demand balances or the thresholds of optimal ecosystem states, following established principles of ecological carrying capacity and system-threshold theory (Odum, 1971; Arrow et al., 1995). The primary goal was to normalize indicators to a 1–5 ordinal scale, where a higher score consistently indicates stronger ecological suitability or adaptability, and a lower score reflects environmental constraints or ecological pressures. All spatial analyses were performed in ArcGIS 10.8 at a fine 30m × 30 m resolution, ensuring a high-fidelity representation of the heterogeneous island landscapes.

For continuous indicators, classification thresholds were determined through a robust, three-step procedure: (1) Baseline Limits: deriving limits from empirical carrying-capacity values (e.g., maximum sustainable nutrient load, habitat support limits); (2) Regulatory Standards: referencing national and provincial standards, including the Marine Ecological Red Line Plan (2018) and GB 3097–2022 Seawater Quality Standard; and (3) Validation: validating threshold ranges using the interquartile range of observed values from comparable coastal ecosystems (Rockström et al., 2009).

Discrete indicators ((e.g., policy coordination or conservation value) were classified according to administrative zoning and expert evaluation. Statistical indicators were calibrated against historical optima or regional mean values to ensure ecological realism and consistency.

To minimize subjective bias from the initial data distribution, all indicators were first standardized using z-score normalization prior to the 1–5 re-scaling:

$$X_i^{\text{'}}=(X_i-{\mu }_i)/{\sigma }_i$$

where $X_i^{\text{'}}$ is the standardized value of indicator $i$, $X_i$ is the original value, and ${\mu }_i$ and ${\sigma }_i$ are the mean and standard deviation of indicator $i$, respectively. The standardized values were then linearly rescaled to the 1–5 range for direct comparability across all dimensions.

3.5.2 Weighting and index computation

To ensure the reliability and transparency of the potential–constraint assessment framework, a simple and robust equal-weighting scheme was adopted. All indicators in the evaluation system were assigned equal weights ($w_i=1/n$) to explicitly avoid subjective bias that can arise from expert-based or questionnaire derived weighting methods. The composite ecological suitability index (E) for the baseline year was computed as the arithmetic mean of all standardized indicator scores:

$$E=(1/n)\sum X_i$$

where $X_i$ is the standardized score of the i-th indicator (ranging from 1 to 5), and n is the total number of indicators (n=13 in this study).

Separate composite indices for development potential (P) and ecological constraints (C) were obtained in a similar manner by averaging only the potential-oriented and constraint-oriented indicators, respectively. These two indices are core inputs for the potential–constraint framework used to derive the final ecological adaptability index (EAI) in the subsequent scenario analysis.

3.5.3 Classification and comparability

The resulting index values for the baseline year (2020) were classified using the natural breaks (Jenks) method. This method was selected over equal-interval and quantile approaches due to its superior performance in minimizing intra-class variance while capturing the nonlinear distribution of ecological gradients in coastal systems (Gonçalves et al., 2025). The resulting baseline ecological suitability pattern (2020) serves as the spatial reference for subsequent scenario simulations.

To ensure comparability among scenarios, the classification thresholds defined for the baseline year were consistently applied to all simulated future scenarios (2030). This allows for a direct quantitative comparison of spatial transitions in ecological suitability and adaptability under alternative policy and management pathways.

Conceptually, ecological suitability represents the static baseline condition of the current ecosystem, whereas ecological adaptability extends this assessment to a dynamic dimension—reflecting the system’s capacity to adjust and reorganize under multiple disturbance and policy scenarios.

Thus, the suitability index serves as the foundation al evaluation, ensuring a coherent analytical progression from present-state evaluation to scenario-based adaptive assessment.

3.6 Model validation and robustness verification

Because the proposed framework relies on composite indices and scenario-based mapping, we adopted a robust, three-fold validation strategy to ensure the reliability and generalizability of the results: (i) baseline consistency tests using independent ecological information, (ii) sensitivity analysis with respect to indicator weighting and classification thresholds, and (iii) cross-scenario consistency checks to ensure logical compatibility with the assumed policy orientations.

3.6.1 Baseline validation with independent data

For the baseline year (2020), we evaluated the internal consistency of the ecological suitability and adaptability maps by comparing them against independent ecological information. Specifically, we spatially overlaid the model-derived classification patterns with ecological red-line zones and designated protected areas derived from officially released planning documents (e.g., Zhejiang Provincial Government, 2021; Liu et al., 2017a). A confusion matrix was constructed and a Kappa coefficient was calculated to quantitatively assess the agreement between the model-derived classification and the independent, policy-based ecological-quality classification. The resulting Kappa value of 0.82 indicates a high level of statistical consistency, strongly supporting the reliability of the baseline ecological suitability and adaptability maps.

3.6.2 Sensitivity analysis

To evaluate the robustness of the potential–constraint framework, we conducted sensitivity tests along two dimensions, following common practice in spatial multi-criteria evaluation (Antle and Stoorvogel, 2006; Reyers et al., 2015).

First, sensitivity to Weighting Scheme: following Fang et al. (2024), we examined the stability of the composite indices against changes in the weighting scheme. Although the main analysis used equal weights, we tested the model’s resilience by perturbing the weight of each indicator in 1% increments within a ±30% range of its baseline value (while maintaining the sum of weights equal to 1). For each perturbed scheme, we recalculated and reclassified the indices, and then compared (i) the spatial distribution of suitability/adaptability classes and (ii) the area shares of each class. The results demonstrated that changes in both spatial patterns and class-area proportions were relatively small and remained within a narrow band, indicating that the model outputs are not highly sensitive to moderate perturbations in indicator weights.

Second, sensitivity to Classification Thresholds: we assessed the sensitivity of the results to the choice of classification method and thresholds. In addition to the natural breaks (Jenks) method used in the main analysis, we applied equal-interval and quantile-based schemes to the standardized indices. We then compared the resulting maps using class-area statistics and cross-tabulations. The analysis confirmed that the main high-suitability and low-suitability zones remained spatially stable across all classification methods, indicating that both the baseline and scenario patterns are robust to reasonable changes in classification schemes.

3.6.3 Cross-scenario consistency checks

Finally, we performed checks to verify the logical consistency of the three development scenarios (E: Ecological Protection Priority; S: Ecological–Economic Coordination; D: Development Priority). This ensures that the spatial outcomes of the scenario simulations genuinely reflect their intended policy orientations.

1. Bottom-Line Enforcement: We verified that areas designated as strictly protected by national or provincial planning documents (e.g., ecological red-line zones and core conservation areas) remain in high-constraint, strict-protection classes under all three scenarios. This confirms that the scenario rules successfully enforce externally imposed ecological bottom lines, preventing these zones from shifting into intensive-development classes even in Scenario D.

2. Expected Scenario Transitions: We examined how grid cells with high development potential and relatively low ecological constraints respond to different scenarios. As expected, Scenario D assigned more development-oriented classes to such cells than Scenario E, with Scenario S consistently yielding intermediate outcomes. These changes were summarized using transition matrices, which recorded the flows of cells among adaptability classes from E to S and from S to D. The observed transitions align perfectly with the conceptual design of the scenarios: scenario E is the most conservative, scenario D is the most permissive within ecological limits, and the scenario S successfully achieves the desired balance.

This comprehensive combination of baseline validation, parameter sensitivity tests, and cross-scenario logical checks is consistent with accepted common practice in spatial multi-criteria evaluation and scenario analysis (Antle and Stoorvogel, 2006; Reyers et al., 2015). The specific tailoring of the framework (e.g., the potential–constraint interaction matrix and three policy-oriented scenarios) to island social–ecological systems provides confidence that the model structure and outputs are both robust and suitable for supporting adaptive management decisions.

4 Results

4.1 Spatial pattern of island ecological adaptability (2020 baseline)

Using 2020 as the baseline year, we applied the multi-indicator potential–constraint framework (Table 3) and the procedures outlined in Section 3.2 to evaluate and map island ecological adaptability for Dongtou District. All input layers were standardized to a 30m × 30 m grid and cross-checked against official planning documents, including the Zhejiang Provincial Coastal Zone and Marine Spatial Plan (2021–2035), ensuring spatial consistency and alignment with current planning policies.

Table 3

Table 3. Classification of indicators for island ecological adaptability (2020 baseline).

The results reveal a clear conservation–development gradient across the district’s social–ecological landscape (Figure 3). The spatial distribution of the adaptability classes manifests the underlying tension between development potential and ecological constraint, and can be categorized as follows:

Figure 3

Figure 3. Baseline spatial pattern of island ecological adaptability in Dongtou (2020).

High-adaptability Zones (Priority Conservation): These zones constitute approximately 23% of the total area and are spatially concentrated in major ecological conservation regions. Key examples include the Nanbei Panshan Provincial Marine Special Protected Area, the southeastern marine biological resource conservation zone, and the Banping–Nance Island group in the south. These areas exhibit favorable ecological conditions (low constraint) but remain relatively remote, possessing limited infrastructure and low returns from current marine utilization(low potential). Within the adaptive management framework, they are therefore designated as priority conservation zones.

Very Low-Adaptability Zones (Priority Development): These zones account for around 10%of the total area and primarily cluster on Dongtou Island and its immediately neighboring inhabited islands, such as Daxiaomen in the north and the Daxiaoqu Islands in the south. These locations are characterized by strong socio-economic foundations, advanced infrastructure, and advantageous locations(high potential). Supported by ports, established aquaculture, and intensive marine industries, they are closely integrated with urban development and generate substantial economic benefits. Thus, they are identified as priority development zones.

Intermediate-Adaptability Zones (Transition Space): These zones account for approximately 65% of the total area and represent a critical transitional space, which is further subdivided:

Higher-intermediate zones (≈ 50%): These are mainly distributed in transitional marine areas, such as the eastern waters offshore of Dongtou Island. While lacking extensive urban infrastructure, they retain relatively favorable natural conditions (moderate constraint, moderate potential). They are consequently classified as general conservation zones.

Lower-intermediate zones (≈18%): These occur as fragmented patches surrounding inhabited islands, effectively functioning as buffers core between conservation and development spaces. Characterized by rich fishery and tourism resources and a foundation for marine industrial development, they are categorized as moderate development zones.

4.2 Scenario-based evolution of ecological adaptability (2030)

4.2.1 Overview of development scenarios

To explicitly explore how island ecological adaptability reorganizes under different policy orientations, we simulated three distinct development scenarios for 2030 within the adaptive management framework: E (Ecological Protection Priority), S (Eco–Economic Coordination), and D (Development Priority). These scenarios are systematically built upon the potential–constraint analytical framework and the multi-level planning directives in Sections 3.2 and 3.3, using 2020 as the established baseline year.

The scenario design was rigorously constrained by officially released national, provincial, and local planning documents, which jointly define mandatory ecological red-line zones, development-permitted areas, and marine utilization limits. Within these policy bounds, the potential–constraint interaction matrix (Table 2; Figure 4) translates various combinations of development potential and ecological constraints into five-level adaptability classes under each scenario. All scenarios are evaluated using the identical 13 indicator system and the same classification thresholds as the baseline assessment, ensuring robust comparability across time and policy options.

Figure 4

Figure 4. Scenario-specific interaction matrices linking development potential and ecological constraints to adaptability-based management zones: (a) Scenario E (Ecological Protection Priority), (b) Scenario S (Eco–Economic Coordination), and (c) Scenario D (Development Priority).

Conceptually, the three scenarios represent distinct trade-off structures between conservation and development:

Scenario E (Ecological Protection Priority): This scenario embodies the most conservation-oriented approach, emphasizing strict conservation by minimizing additional anthropogenic disturbance and maintaining baseline ecosystem integrity. Development is tightly constrained and largely limited to designated harbor and settlement zones.

Scenario S (Eco–Economic Coordination): This scenario pursues a balanced and adaptive pathway where moderate, spatially optimized development is permitted while simultaneously strengthening the protection of key ecological areas. Its aim is to maintain both ecological resilience and economic efficiency within predefined adaptive thresholds.

Scenario D (Development Priority): This scenario promotes accelerated economic expansion within the non-negotiable limits of ecological redlines. It concentrates industrial and infrastructural growth in areas with high development potential and relatively low ecological sensitivity, representing a configuration that embodies higher economic gain but potentially lower ecological stability.

Under this common framework, the three scenarios generate contrasting 2030 ecological adaptability patterns that can be directly and quantitatively compared in terms of spatial ecological performance and socio-economic implications.

4.2.2 Spatial patterns of ecological adaptability under alternative development scenarios

The simulated ecological adaptability patterns for 2030 reveal clear and contrasting spatial configurations among the three scenarios (Figure 5; Table 4).

Figure 5

Figure 5. (a) Spatial distribution of island ecological adaptability in Dongtou under scenario E (Ecological Protection Priority) for 2030. (b) Spatial distribution of island ecological adaptability in Dongtou under scenario S(Eco–Economic Coordination) for 2030. (c) Spatial distribution of island ecological adaptability in Dongtou under scenario D (Development Priority) for 2030.

Table 4

Table 4. Transition matrix of island ecological adaptability patterns under alternative scenarios.

4.2.2.1 Scenario E: ecological protection priority

Under Scenario E, the landscape is overwhelmingly dominated by conservation-oriented, high-adaptability zones. More than 60% of the area is classified as general or priority conservation classes, with development largely confined to a few major harbors and inhabited islands.

The transition of spatial matrix (Table 4; Figure 6) indicates a substantial upgrading of adaptability status: a significant proportion of baseline general conservation and moderate development zones shift into higher conservation-oriented classes, with transition ratios of roughly 8% and 24%, respectively. This leads to an expansion and consolidation of high-adaptability space, forming a more continuous, ecosystem-centered pattern that is highly consistent with long-term conservation objectives and ecological resilience goals.

Figure 6

Figure 6. Spatial transitions of island ecological adaptability under alternative development scenarios (2030), illustrated using Sankey diagrams: (a) Scenario E (Ecological Protection Priority), (b) Scenario S (Eco–Economic Coordination), and (c) Scenario D (Development Priority). ODZ, optimized development zone; MDZ, moderate development zone; GCZ, general conservation zone; PCZ, priority conservation zone.

4.2.2.2 Scenario S: eco–economic coordination

Scenario S yields a more compact and spatially optimized configuration. Development activities are strategically concentrated on larger islands with existing infrastructure and moderate population density (e.g., Dongtou Island and the northern island cluster), while conservation is simultaneously strengthened in ecologically critical areas. For instance, the entire Dongtou National Marine Park is incorporated into strict protection classes, resulting in an increase of approximately 12% in total conservation area compared with the baseline.

The Sankey diagram (Figure 6) shows marked transitions from moderate development and general conservation categories toward higher adaptability states, accompanied by a desirable reduction in landscape fragmentation. Overall, Scenario S achieves the most balanced arrangement between conservation and development, effectively maintaining extensive ecological protection areas while enabling targeted, efficient use of limited development space.

4.2.2.3 Scenario D: development priority

By contrast, Scenario D is characterized by the widespread expansion of built-up and industrial zones within the permitted ecological and safety thresholds. Portions of former priority and general conservation areas are reclassified as moderate or optimized development zones. Table 4 indicates an overall decline of approximately 10% in ecological adaptability compared with the baseline, a consequence driven by the contraction of high-protection classes and the resultant growth of development-oriented classes. The Sankey diagram highlights dominant flows from conservation-oriented to development-oriented categories, reflecting increased anthropogenic pressure and reduced ecological resilience.

4.2.2.4 Synthesis of adaptive pathways

Taken together, the three scenarios outline a spectrum of adaptive pathways for Dongtou’s island system: Scenario E maximizes ecological integrity but at the opportunity cost of limiting development; Scenario D yields higher short-term economic gains but compresses high-adaptability zones and reduces the overall system resilience; Scenario S stands out as the most balanced and resilient pathway, maintaining a robust ecological foundation while enabling targeted, efficient development.

Quantitatively, conservation-oriented management zones (general + priority conservation) increase from about 72.1% of the district in the 2020 baseline to 96.1% in Scenario E and 93.1% in Scenario S, but decrease to 68.4% in Scenario D. High-adaptability zones (priority conservation) expand from 23.0% at baseline to 89.6% in Scenario E, remain roughly stable at 22.0% in Scenario S, and contract to 12.9% under Scenario D. Conversely, development-oriented zones (optimized + moderate development) shrink from 27.9% of the area in the baseline to 6.9% in Scenario S and 3.9% in Scenario E, yet expand to 31.6% in Scenario D (Table 4). These concise statistics make the contrasted spatial outcomes of the three scenarios more immediately comparable without requiring readers to repeatedly switch between figures and tables.

4.3 Indeterminancy analysis

Effective adaptive management of island ecosystems necessitates an explicit understanding of the multiple sources of indeterminacy and uncertainty that shape ecological adaptability. Based on the spatial differentiation and sensitivity analyses of the potential-oriented and constraint-oriented indicators in Dongtou District, three dominant dimensions of uncertainty indeterminacy were identified as critical to adaptive governance: land–sea boundary dynamics, natural environmental variability, and socio-economic development trajectories.

4.3.1 Land–sea boundary dynamics

The transitional nature of the island–sea interface represents a fundamental source of uncertainty in ecosystem management. Although Dongtou retains a relatively high proportion of natural coastline, progressive reclamation, bridge construction, and coastal engineering have continuously modified the shoreline morphology. Such anthropogenic alterations intensify the physical coupling between terrestrial and marine systems, producing non-linear feedbacks that fundamentally reshape hydrodynamic regimes, sediment transport, and critical habitat connectivity. These processes not only blur administrative boundaries but also pose significant management challenges, underscoring the necessity for integrated land–sea management strategies to sustain the continuity of ecological processes across island, intertidal, and nearshore zones.

4.3.2 Natural environmental variability

Environmental fluctuations exert a pronounced influence on the short- and medium-term trajectory of ecological adaptability. Constraint indicators—such as the ecological disturbance index and the marine risk distribution index—are particularly sensitive to episodic perturbations. Extreme events including harmful algal blooms (HABs), typhoons, storm surges, and shoreline erosion can rapidly alter ecosystem states, potentially inducing transient or even irreversible regime shifts. Such intrinsic variability significantly amplifies management uncertainty and highlights the crucial need for continuous environmental monitoring, adaptive threshold adjustment, and early-warning mechanisms to be fully embedded within the potential–constraint framework.

4.3.3 Socio-economic development trajectories

The dynamic balance between marine resource utilization and ecological protection introduces substantial uncertainty into future decision-making. Rapid industrial transformation in fisheries, ports, and coastal tourism strengthens human–ocean coupling, while simultaneous enforcement of ecological civilization policies imposes increasingly stricter environmental constraints. These interacting, often conflicting, forces generate evolving trade-offs and feedbacks between ecological carrying capacity and economic demand, which complicate the prediction of long-term adaptability. In this study, carrying capacity is defined as the environmental threshold beyond which an ecosystem’s ability to sustain human activities and economic development becomes impaired (Arrow et al., 1995). Within the potential–constraint framework, carrying capacity precisely delineates the ecological limitation zone and establishes the quantitative boundary for balancing ecological resilience with development potential.

5 Discussion

5.1 Comparative insights: potential–constraint framework versus established models

Recent advances in ecological adaptability assessment have largely evolved along four methodological strands: (i) ecological suitability and land evaluation, (ii) ecological carrying capacity (CC), (iii) resilience–vulnerability–adaptation (RVA) frameworks, and (iv) causal-chain environmental assessment models such as PSR (Pressure–State–Response) and DPSIR (Driving force–Pressure–State–Impact–Response) (OECD, 1993; Tscherning et al., 2012; Zhang et al., 2017). The potential–constraint (P-C) framework developed in this study is conceptually rooted in these traditions but is functionally distinct in several important ways. Below, we focus on the main aspects in which it adds value, rather than providing a full review.

5.1.1 Conceptual and structural advancement

5.1.1.1 Explicit binary structure for trade-off mapping

The P-C framework retains the multi-criteria structure of classical ecological suitability and land evaluation models (e.g., McHarg’s “sieve mapping,” FAO system), but reinterprets and restructures it around ecological adaptability. Instead of asking where conditions are simply “suitable” for a given use, the potential–constraint index evaluates where the island system can maintain or shift towards desirable states under ongoing and future pressures. Methodologically, this is achieved by an explicit binary partition of indicators into two semantically meaningful dimensions: resource potential (development capacity and opportunity) and environmental/anthropogenic constraints (ecological condition, sensitivity, and pressure). This dual structure makes trade-offs more transparent than in traditional, unstructured composite indices, where a high overall score can conflate high potential with high pressure.

5.1.2 Methodological advancement for adaptive management

5.1.2.1 Tight coupling with scenario analysis and management zoning

The P-C framework is designed from the outset to be tightly coupled with scenario analysis and management zoning, rather than remaining a one-off baseline diagnostic. The same indicator system and potential–constraint logic is consistently applied to the 2020 baseline and to the 2030 policy scenarios. A simple interaction matrix links discrete levels of potential and constraint to scenario-specific adaptability (management) classes, allowing alternative policy orientations—ecological protection priority, eco–economic coordination, and development priority—to be translated into spatially explicit, comparable maps. In practical terms, this enables planners to derive management zones (e.g., priority conservation, general conservation, moderate development, optimized development) directly from the adaptability index and to quantify how adaptability space is reallocated under different development trajectories, which goes beyond the static perspective typical of many suitability or CC studies.

5.1.2.2 Integrated, multi-dimensional land–sea perspective

The P-C framework integrates multiple dimensions across the land–sea interface at a common spatial resolution, while keeping data requirements moderate. Many island or coastal assessments remain sectoral, focusing on isolated themes such as water quality, fisheries, or shoreline dynamics, and do not explicitly link terrestrial land use, coastal infrastructures, and marine ecosystem conditions (Williams, 2011). By contrast, our indicator system jointly represents habitat structure, ecosystem function, socio-economic conditions, marine utilization, and policy responses on a 30m × 30m grid. This integrated, land–sea perspective is particularly important in island social–ecological systems, where limited land, intense coastal development, and strong cross-boundary flows make local trade-offs highly contingent on spatial configuration. Scenario simulations further show how locations of “high potential–high constraint” (e.g., high-quality habitats under strong development pressure) and “low potential–low constraint” (e.g., degraded but flexible zones) reconfigure under alternative development paths, providing planning insights that are difficult to obtain from single-subsystem analyses or generic RVA narratives.

5.1.3 Conceptual alignment and limitations

At the same time, the potential–constraint framework remains closely aligned with resilience and vulnerability thinking by operationalizing RVA concepts in a simple, map-based form. Potential broadly aligns with ecological quality, connectivity, and opportunities for adaptive responses, while constraints aggregate exposure, pressure, and reduced coping capacity. Management options can therefore be framed as either enhancing potential (e.g., safeguarding or restoring key habitats, maintaining connectivity) or alleviating constraints (e.g., reducing pollutant loads, retrofitting hard shorelines). In this sense, the framework translates the abstract language of resilience and vulnerability into directly interpretable spatial intervention strategies that can be iteratively updated as new monitoring data become available.

Finally, the potential–constraint framework should be viewed as a meso-scale decision-support tool that complements, rather than replaces, process-based or risk-specific models. Its outputs are unavoidably influenced by indicator selection, normalization choices, and scenario rules, and dynamic processes such as hydrodynamics or population dynamics are represented only indirectly through state indicators. These limitations are discussed in more detail in Section 5.4. When interpreted together with finer-scale modelling, stakeholder knowledge, and policy considerations, however, the framework provides a pragmatic bridge between complex theory and operational planning, helping island managers move from ad hoc, sectoral interventions towards more coherent, adaptive strategies that recognize both the opportunities and limits of fragile land–sea systems.

5.2 Ecological–economic trade-offs under alternative scenarios

The baseline ecological adaptability pattern for Dongtou already establishes a pronounced ecological–economic gradient: high-adaptability zones are consistently concentrated in relatively intact island ecosystems and designated conservation areas (low constraint, low potential), whereas low-adaptability zones coincide with dense urban clusters, major ports, and intensive aquaculture platforms (low constraint, high potential). This configuration reflects the historical accumulation of development pressures in nearshore hubs, where economic output has historically been prioritized over ecological performance.

The three simulated 2030 scenarios explicitly frame the non-fixed nature of these trade-offs, demonstrating how different policy choices lead to contrasting spatial outcomes:

5.2.1 Scenario E: ecological protection priority (high ecological integrity)

Under scenario E, Dongtou shifts toward a strongly conservation-oriented configuration. More than 60% of the territory falls into general or priority conservation zones(Figure 5a). Large, contiguous patches of high adaptability emerge around existing protected areas and less disturbed island groups, indicating spatial consolidation of conservation space. The transition matrix (Table 4) shows substantial upward transitions from general conservation (≈8%) and moderate development (≈24%) categories into higher-protection classes. While ecological resilience is maximized, this comes at the cost of markedly constraining future socio-economic development, particularly in land-scarce harbor-adjacent and nearshore segments. Scenario E occupies one extreme of the trade-off spectrum: high ecological integrity and strong adaptive capacity, but limited room for additional development.

5.2.2 Scenario D: development priority (high short-term gain)

By contrast, scenario D pushes the system toward the opposite end of this spectrum. Built-up and industrial zones expand within the limits of ecological red lines, and portions of former priority conservation areas are converted into moderate or optimized development zones (Figure 5c). The transition matrix (Table 4) shows dominant flows from conservation-oriented to development-oriented classes, and the total area of high-adaptability categories declines by roughly 10% compared with the baseline. While this supports short- to medium-term economic gains and infrastructure expansion, it simultaneously compresses ecological buffers and increases exposure to cumulative risks along heavily reclaimed shorelines. Scenario D thus illustrates the upper bound of development-centered choices that respect formal ecological bottom lines, but entail a noticeable erosion of system resilience.

5.2.3 Scenario S: ecological–economic coordination (optimal balance)

The Ecological–Economic Coordination scenario (S) occupies a crucial intermediate and non-trivial position. This scenario does not simply average conservation and development measures. Instead, it redistributes land–sea uses to align high-intensity activities with zones of relatively lower ecological constraints/potential, while prioritizing strict protection and restoration in high-potential, high-constraint areas. Spatially, development is concentrated on larger islands with established infrastructure, while key habitats are incorporated into strict protection zones, increasing total conservation area by about 12% (Figure 5b). The Sankey diagram (Figure 6) reveals that many cells in moderate development and general conservation classes shift toward higher-adaptability states without the extensive conversion of conservation space seen in Scenario D, and overall landscape fragmentation is reduced. Scenario S demonstrates that it is possible to achieve non-trivial gains in ecological resilience while still accommodating a reasonable level of socio-economic development, especially when spatial allocation is optimized to manage trade-offs effectively.

The three scenarios collectively show that ecological–economic trade-offs in island social–ecological systems are highly sensitive to the spatial allocation of development. Some island and coastal segments exhibit steep ecological costs for relatively small additional development, suggesting that strict protection or active restoration should be prioritized there. The potential–constraint framework makes these trade-offs explicit by decomposing changes into shifts in potential and shifts in constraints. This provides a clear spatial decision basis for balancing conservation and development—identifying locations where the marginal ecological costs of additional development are unacceptably high, and others where carefully managed growth could be compatible with maintaining or even improving ecological adaptability.

5.3 Management implications

The analysis of spatial adaptability patterns across the three scenarios (Figures 5, 6; Table 4) reveals not only the inherent robustness or fragility of island ecosystems, but also the crucial role of policy orientation in reorganizing ecological–economic trade-offs. These results offer several direct and actionable implications for adaptive island governance under accelerating land–sea coupling and climate change pressures (Schultz et al., 2015; Rölfer et al., 2022).

5.3.1 From reactive restoration to anticipatory and learning-based governance

The consistent concentration of low-adaptability zones in harbor–urban interfaces and intensively used aquaculture embayment— areas characterized by high development potential and strong ecological constraints demands a shift in management philosophy.

In Scenario D, many of these areas remain low-adaptability hotspots, indicating minimal capacity to absorb further disturbance. Relying on reactive, post-hoc restoration after functional collapse is costly and uncertain. The potential–constraint maps support a shift to anticipatory governance. Zones that repeatedly appear as low-adaptability under multiple scenarios should be treated as “early-warning regions.” Managers should prioritize stricter environmental regulation, ecological shoreline retrofitting, or phased retreat strategies before critical thresholds are crossed. Because the adaptability index can be updated iteratively with new monitoring data, it serves as a learning device. Managers can observe how adaptability responds spatially to specific measures, refine parameter assumptions, and continuously adjust zoning rules and spatial priorities.

5.3.2 Mainstreaming uncertainty and spatial risk into management priorities

The pronounced spatial heterogeneity of adaptability, especially along narrow land–sea transition belts, highlights locations where small changes in use intensity can trigger disproportionate shifts in system state. For example, in Scenario S, several medium-adaptability “hinge regions” can swing towards either higher or lower adaptability depending on localized allocation decisions.

The framework allows managers to explicitly mainstream spatial risk into prioritization. Zones where adaptability is low and projected to decline across scenarios (e.g., the transition areas of Scenario D) should be assigned higher precautionary weights, tighter pollution caps, or stricter constraints on new reclamation. Areas where results are more uncertain or where scenario outcomes diverge strongly are suitable for flexible and reversible uses (e.g., low-impact tourism, seasonal activities). These zones should be targeted for enhanced monitoring. In practice, this means implementing differentiated monitoring intensity, establishing contingency buffers in use quotas, and applying adaptive zoning rules that permit revision based on new ecological or socio-economic information.

5.3.3 Optimizing decision-making across multiple policy scenarios

The scenario simulations demonstrate that while policy orientation decisively shapes the spatial distribution of ecological adaptability, some ecological–economic conflicts can be mitigated through strategic reallocation rather than simply through overall restriction.

Scenario S successfully maintains large high-adaptability blocks around key island ecosystems (e.g., ~12% expansion of strict protection areas around the Dongtou National Marine Park) while concentrating new development in already transformed or lower-potential segments. This contrasts sharply with Scenario D, which achieves only a 10% reduction in high-adaptability area.

For planners, this implies that scenario analysis must be embedded in spatially explicit adaptability evaluation, moving beyond abstract narratives. Comparing baseline, E, S, and D within the same potential–constraint framework enables decision-makers to identify:

Robust “No-Regret” Areas: areas where all scenarios agree on the need for strict protection (e.g., high-constraint zones);

Flexible zones: areas where multiple development options lead to similar adaptability outcomes, making negotiated solutions feasible;

Conflict hotspots: areas where certain development choices produce sharp declines in adaptability, requiring stringent safeguards, compensatory measures, or alternative locations (Du et al., 2022).

This integrated view helps shift stakeholder negotiations from abstract debates about “development versus protection” to concrete maps and quantifiable trade-offs, improving the transparency and accountability of spatial decision-making.

Overall, grounding management strategies in spatially explicit ecological adaptability assessment encourages a shift away from fragmented, sector-by-sector interventions towards coherent, island-scale planning. Used iteratively with monitoring data and stakeholder input, the potential–constraint framework supports a gradual transition from crisis-driven responses to proactive, adaptive governance of fragile land–sea systems.

5.4 Limitations and future directions

Despite the conceptual rigor and practical contributions of the potential–constraint framework, several inherent limitations of this study must be acknowledged to guide future methodological advancements.

5.4.1 Data fidelity, uncertainty quantification, and system complexity

First, the design of the indicator system and subsequent scenario simulations were fundamentally constrained by data availability and resolution. While a multi-dimensional evaluation framework was established, several key facets of island social–ecological systems remain only partially represented. Specifically, the current adaptability index relies on static, snapshot surveys of biodiversity and productivity, which do not yet capture interannual variability, the intricacies of fine-scale habitat processes, or critical socio-cultural values (e.g., local knowledge regarding ecosystem services).

Future research must transition from static state assessment to dynamic representation. This requires strengthening the data foundation through the integration of high-resolution remote sensing, establishing continuous in situ observation platforms, and conducting participatory social surveys to fully capture the complexity, heterogeneity, and functional dynamics of land–sea systems.

5.4.2 Model structure, simplification of feedbacks, and uncertainty sources

Second, the potential–constraint model, in its current form, inevitably simplifies the endogenous feedbacks within coupled human–environment subsystems.

5.4.2.1 Sources of uncertainty

The current adaptability classification is subject to spatial uncertainty (e.g., inherent errors from interpolating scattered socio-economic or pollution point data) and temporal uncertainty (mismatches among datasets collected in different years). Generally, these uncertainties tend to dampen the true extremes of high and low adaptability, resulting in a slightly more conservative (less spatially heterogeneous) classification than reality. Future work must incorporate explicit uncertainty propagation analysis to quantify the likely error bounds of the adaptability index.

5.4.2.2 Simplification of feedbacks

The present implementation, based on static indicator combinations and scenario rules, only partially captures crucial phenomena such as non-linear interactions, cumulative anthropogenic effects, and threshold dynamics. Future work should aim for a more mechanistic representation of system behavior. This could involve integrating system dynamics models, Bayesian hierarchical frameworks, or machine-learning-based causal inference into the assessment. Such integration is essential to more accurately model feedback loops and uncertainty propagation, particularly to explore how major shocks (e.g., extreme pollution events, critical infrastructure failures) cascade through interconnected land–sea networks.

5.4.3 Temporal horizon and response to external shocks

Third, the scenario simulations are limited to a medium-term horizon (2030) and assume relatively stable policy trajectories aligned with current planning documents. In reality, island governance is highly susceptible to sudden, exogenous shocks—such as extreme climate events, pandemics, or global market fluctuations—which can rapidly disrupt both ecological conditions and development priorities. Future studies should extend the temporal horizon and incorporate resilience-oriented or shock-response scenarios (e.g., simulating post-disaster recovery or climate-induced sea-level rise). This advancement would improve the model outputs’ robustness and policy relevance by explicitly exploring the conditions under which system adaptability can be maintained or restored following major disturbances.

5.4.4 Cross-context validation and transferability requirements

Finally, this study focuses on a single representative case (Dongtou District). While the results offer robust methodological and empirical insights, broader cross-context validation is necessary to test the generality and transferability of the framework.

5.4.4.1 Recalibration requirements

When transferring the Potential–Constraint (P-C) framework to a new island or coastal region, the following components would require recalibration to ensure local relevance and accuracy:

Normalization and Constraint Thresholds: The normalization thresholds and specific constraint thresholds must be reset based on local environmental standards and empirically determined carrying capacity limits.

Weights: The relative importance of indicators would require expert elicitation or Bayesian updating based on local knowledge and specific management priorities, reflecting the regional goals for balancing development versus conservation.

5.4.4.2 Future validation approach

Future applications should prioritize comparative and multi-scalar analysis across diverse island and coastal regions—especially in highly vulnerable zones like tropical and subtropical islands and in Small Island Developing States (SIDS).

Implementing nested local–regional assessments or multi-case comparative analysis would further enhance the transferability and scalability of the potential–constraint approach for evidence-based sustainable management globally.

6 Conclusion

This study has established and applied a potential–constraint framework to evaluate the ecological adaptability of island social–ecological systems, utilizing Dongtou District, China, as a representative case. By integrating multi-dimensional indicators with scenario-based spatial modeling, the framework successfully translates the abstract notion of ecological adaptability into operational, spatially explicit information for planning. The results reveal pronounced spatial heterogeneity in adaptability and demonstrate that policy orientation decisively reshapes the spatial pattern. Specifically, the eco–economic coordination scenario yields the most balanced configuration, significantly enhancing landscape connectivity and achieving a more favorable trade-off between conservation and development than either strict-protection or development-priority strategies.

Collectively, these findings advocate for a tangible shift in island governance from reactive, post-hoc restoration toward anticipatory, learning-based management. By explicitly embedding uncertainty and spatial risk into adaptive planning, the proposed framework provides a scientifically grounded, transferable tool for fostering sustainable and resilient development in vulnerable island and coastal regions under accelerating global environmental change.

Data availability statement

Publicly available datasets were analyzed in this study. DMSP/OLS nighttime-light data are available from NOAA/NGDC (https://eogdata.mines.edu/products/vnl/). Socio-economic statistics are from the Wenzhou Statistical Yearbook, the Dongtou Statistical Yearbook, and official Marine Bulletins. Field survey data and derived/interpolated indicator layers supporting the findings are available from the corresponding author upon reasonable request.

Author contributions

YX: Methodology, Conceptualization, Investigation, Writing – review & editing, Writing – original draft, Visualization, Software. CW: Resources, Methodology, Validation, Writing – review & editing.

Funding

The author(s) declared financial support was received for this work and/or its publication. This work was supported by the National Natural Science Foundation of China under Grant Agreements 41506140.

Acknowledgments

Sincere thanks are given for the comments and suggestions of reviewers and members of the editorial team.

Conflict of interest

The authors declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declare that Generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fmars.2025.1722774/full#supplementary-material

References

Abernethy V. D. (1999). Nature’s services: societal dependence on natural ecosystems. Popul. Environ. 20, 277. doi: 10.1023/A:1023307309124

Crossref Full Text | Google Scholar

Adams V. M., Pressey R. L., and Álvarez-Romero J. G. (2016). Using optimal land-use scenarios to assess trade-offs between conservation, development, and social values. PloS One 11, e0158350. doi: 10.1371/journal.pone.0158350

PubMed Abstract | Crossref Full Text | Google Scholar

Antle J. M. and Stoorvogel J. J. (2006). Predicting the supply of ecosystem services from agriculture. Am. J. Agric. Econ 88, 1174–1180. doi: 10.1111/j.1467-8276.2006.00929.x

Crossref Full Text | Google Scholar

Arrow K., Bolin B., Costanza R., Dasgupta P., Folke C., Holling C. S., et al. (1995). Economic growth, carrying capacity, and the environment. Ecol. Econ 15, 91–95. doi: 10.1016/0921-8009(95)00059-3

Crossref Full Text | Google Scholar

Baldacchino G. (2018). The Routledge international handbook of island studies (London & New York: Routledge).

Google Scholar

Barragán J. M. and De Andrés M. (2015). Analysis and trends of the world’s coastal cities and agglomerations. Ocean. Coast. Manage. 114, 11–20.

Google Scholar

Boyd E. and Folke C. (Eds.) (2011). Adapting institutions: governance, complexity and social-ecological resilience (Cambridge University Press).

Google Scholar

Cao L., Li J., Ye M., Pu R., Liu Y., Guo Q., et al. (2018). Changes of ecosystem service value in a coastal zone of Zhejiang Province, China, during rapid urbanization. Int. J. Environ. Res. Public Health 15, 1301. doi: 10.3390/ijerph15071301

PubMed Abstract | Crossref Full Text | Google Scholar

Chen M., Zeng C., Zeng X., Liu Y., Wang Z., Shi X., et al. (2023). Assessment of marine protected areas in the East China Sea using a management effectiveness tracking tool. Front. Mar. Sci. 10, 1081036. doi: 10.3389/fmars.2023.1081036

Crossref Full Text | Google Scholar

Cohen-Shacham E., Andrade A., Dalton J., Dudley N., Jones M., Kumar C., et al. (2019). Core principles for successfully implementing and upscaling nature-based solutions. Environ. Sci. Policy 98, 20–29. doi: 10.1016/j.envsci.2019.04.014

Crossref Full Text | Google Scholar

Cohen-Shacham E., Walters G., Janzen C., and Maginnis S. (Eds.) (2016). Nature-based solutions to address global societal challenges (Gland, Switzerland: IUCN).

Google Scholar

Dang A. N. and Kawasaki A. (2017). Integrating biophysical and socio-economic factors for land-use and land-cover change projection in agricultural economic regions. Ecol. Model. 344, 29–37. doi: 10.1016/j.ecolmodel.2016.11.004

Crossref Full Text | Google Scholar

Decker D. J., Chase L. C., and Forman R. T. T. (1996). Landscape ecology principles in landscape architecture and land-use planning (Washington, DC: Harvard University Graduate School of Design, Island Press and ASLA).

Google Scholar

Deng W. J., Wu H. Z., Li T. X., Zhou L. N., Hu R. Y., Jin X. J., et al. (2024). Main vegetation types and characteristics in Dongtou National Marine Park, Zhejiang, China (in Chinese). Chin. J. Plant Ecol. 48 (2), 254–268. doi: 10.17521/cjpe.2023.0126

Crossref Full Text | Google Scholar

Du Y., Li X., He X., Li X., Yang G., Li D., et al. (2022). Multi-scenario simulation and trade-off analysis of ecological service value in the Manas River Basin based on land use optimization in China. Int. J. Environ. Res. Public Health 19, 6216. doi: 10.3390/ijerph19106216

PubMed Abstract | Crossref Full Text | Google Scholar

Duvat V. K., Magnan A. K., Perry C. T., Spencer T., Bell J. D., Wabnitz C. C., et al. (2021). Risks to future atoll habitability from climate-driven environmental changes. Wiley. Interdiscip. Rev.: Climate Change 12, e700.

Google Scholar

European Commission (EC) (2017). Directive 2017/845/EC of 17 may 2017 amending directive 2008/56/EC of the european parliament and of the council. Off. J. Eur. Comm., L125/27.

Google Scholar

Fang J. Y., Shen Z. H., Tang Z. Y., and Wang Z. H. (2004). The protocol for the survey plan for plant species diversity of China’s mountains (in Chinese). Biodivers. Sci. 12, 5–9.

Google Scholar

Fang X., Wu Y., Xia L., Wang Z., Ni W., Zhang Y., et al. (2024). Vulnerability analysis of socio-ecological systems on small islands from the perspective of fisheries and tourism: A case study of Zhoushan Gouqi Island. J. Clean. Prod. 473, 143550. doi: 10.1016/j.jclepro.2024.143550

Crossref Full Text | Google Scholar

Futuyama D. J. (1979). Evolutionary biology (Sunderland: Sinauer).

Google Scholar

Gao S., Sun H. H., Cao G. X., Lin H., and Hong T. (2018). Dynamic assessment of ecological security of Pingtan Island in southeast China (in Chinese). J. Appl. Oceanogr. 37, 94–101.

Google Scholar

Glavovic B., Kelly M., Kay R., and Travers A. (Eds.) (2014). Climate change and the coast: building resilient communities (London: CRC Press), 596.

Google Scholar

Gonçalves L. A., De Souza E. G., Nóbrega L. H., et al. (2025). Spatial and temporal variability of soil apparent electrical conductivity. Precis. Agric. 26, 10. doi: 10.1007/s11119-024-10209-x

Crossref Full Text | Google Scholar

Holling C. S. (Ed.) (1978). Adaptive environmental assessment and management (London: John Wiley & Sons).

Google Scholar

Hulse D., Goorjian L., and Richey D. (1997). Possible futures for the muddy Creek Watershed, Benton County, Oregon (Oregon, USA: University of Oregon).

Google Scholar

Huo J. (2010). Influencing factors and evaluation index system of marine carrying capacity (Qingdao: Ocean University of China).

Google Scholar

Intergovernmental Panel on Climate Change (IPCC) (2001). Climate change 2001: The science of climate change (Cambridge: Cambridge University Press).

Google Scholar

Intergovernmental Panel on Climate Change (IPCC) (2007). Climate change 2007: Impacts, adaptation and vulnerability (Cambridge: Cambridge University Press).

Google Scholar

Jiang Y., Tian Y., and Zheng S. (2018). A study on urban eco-security pattern of an island city: A case study of Pingtan Island (in Chinese). Acta Ecol. Sin. 38, 769–777.

Google Scholar

Ki-Moon B. (2014). The road to dignity by 2030: Ending poverty, transforming all lives and protecting the planet (New York: United Nations).

Google Scholar

Lee K. N. (1994). Compass and gyroscope: Integrating science and politics for the environment (Washington, DC: Island Press).

Google Scholar

Lin Y., Fan J., and Wen Q. (2016). Primary exploration of ecological theories and technologies for delineation of ecological redline zones (in Chinese). Acta Ecol. Sin. 36, 1244–1252.

Google Scholar

Liu J., Liu M., Zhuang D., Zhang Z., and Deng X. (2003). Study on spatial pattern of land-use change in China during 1995–2000. Science in China (Series D) 46, 373–384. doi: 10.1360/03yd9033

Crossref Full Text | Google Scholar

Liu S., Wang Y., and Wang R. (2017b). Study on the delimitation method of the overloading area of marine resources and environment in the red line area (in Chinese). Mar. Sci. Bull. 36, 497–503.

Google Scholar

Liu F., Wu T., and Jiang G. (2017a). Dynamic response of coastline and coastal landscape patterns to Hemeroby: A case study along the south coast of Yingkou. Acta Ecol. Sin. 37, 7427–7437.

Google Scholar

Ma R., Chen J., Pan Q., Cheng Y., Wang W., Zhu B., et al. (2022). How should the effectiveness of marine functional zoning be evaluated? Taking Wenzhou Marine Functional Zoning as an example. Land 11, 267. doi: 10.3390/land11020267

Crossref Full Text | Google Scholar

Maor Cohen M., Shamon H., Dolev A., Reichman A., David S. B., and Saltz D. (2020). Long-term reevaluation of spatially explicit models as a means for adaptive wildlife management. Ecol. Appl. 30, e02088. doi: 10.1002/eap.2088

PubMed Abstract | Crossref Full Text | Google Scholar

Masselink G., Beetham E., and Kench P. (2020). Coral reef islands can accrete vertically in response to sea level rise. Sci. Adv. 6, eaay3656. doi: 10.1126/sciadv.aay3656

PubMed Abstract | Crossref Full Text | Google Scholar

McHarg I. (1969). Design with nature (New York: Natural History Press).

Google Scholar

Mcleod E., Bruton-Adams M., Förster J., Franco C., Gaines G., Gorong B., et al. (2019). Lessons from the Pacific Islands–adapting to climate change by supporting social and ecological resilience. Front. Mar. Sci. 6, 289. doi: 10.3389/fmars.2019.00289

Crossref Full Text | Google Scholar

McLeod K. L. and Leslie H. M. (2009). “Why ecosystem-based management?,” in Ecosystem-based management for the oceans. Eds. McLeod K. L. and Leslie H. M. (Island Press, Washington, DC), 3–12.

Google Scholar

Medland S. J., Shaker R. R., Forsythe K. W., Mackay B. R., and Rybarczyk G. (2020). A multi-criteria wetland suitability index for restoration across Ontario’s mixedwood plains. Sustainability 12, 9953. doi: 10.3390/su12239953

Crossref Full Text | Google Scholar

Mumby P. J., Chollett I., Wolff N. W., and Bozec Y. M. (2014). Ecological resilience, robustness, and vulnerability: How do these concepts benefit ecosystem management? Curr. Opin. Environ. Sustain. 7, 22–27. doi: 10.1016/j.cosust.2013.11.021

Crossref Full Text | Google Scholar

Muniz P., Venturini N., and Hutton M. (2011). Ecosystem health of Montevideo coastal zone: A multi-approach using benthic indicators to improve a ten-year-ago assessment. J. Sea. Res. 65, 38–50. doi: 10.1016/j.seares.2010.07.001

Crossref Full Text | Google Scholar

Mycoo M. A. (2018). Beyond 1.5 C: vulnerabilities and adaptation strategies for Caribbean Small Island developing states. Reg. Environ. Change 18, 2341–2353. doi: 10.1007/s10113-017-1248-8

Crossref Full Text | Google Scholar

Nichols J. D. and Williams B. K. (2006). Monitoring for conservation. Trends Ecol. Evol. 21, 668–673. doi: 10.1016/j.tree.2006.08.007

PubMed Abstract | Crossref Full Text | Google Scholar

Nurse L. A., McLean R. F., Agard J., Briguglio L. P., Duvat-Magnan V., Pelesikoti N., et al. (2014). “Small islands,” in Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part B: Regional Aspects (Cambridge University Press, Cambridge), 1613–1654.

Google Scholar

Odum H. T. (1971). Environment, power, and society (New York: John Wiley & Sons).

Google Scholar

OECD (1993). “Core set of indicators for environmental performance reviews,” in OECD Environment Monographs No. 83 (OECD Publishing, Paris).

Google Scholar

Peter H., Verburg W., Soepboer W., and Veldkamp A. (2002). Modeling the spatial dynamics of regional land use: The CLUE-S model. Environ. Manage. 20, 391–405.

PubMed Abstract | Google Scholar

Qian X. U. E., Wei S., Yili Z., and Fengyun M. (2017). Research progress in ecological carrying capacity: Implications, assessment methods and current focus. J. Resour. Ecol. 8, 514–525. doi: 10.5814/j.issn.1674-764x.2017.05.009

Crossref Full Text | Google Scholar

Ramesh R., Chen Z., Cummins V., Day J., D’elia C., Dennison B., et al. (2015). Land–ocean interactions in the coastal zone: Past, present & future. Anthropocene 12, 85–98. doi: 10.1016/j.ancene.2016.01.005

Crossref Full Text | Google Scholar

Reyers B., Nel J. L., O’Farrell P. J., Sitas N., and Nel D. C. (2015). Navigating complexity through knowledge co-production: Mainstreaming ecosystem services into disaster risk reduction. Proc. Natl. Acad. Sci. 117 (24), 7362–7368. doi: 10.1073/pnas.1414374112

PubMed Abstract | Crossref Full Text | Google Scholar

Rist L., Felton A., and Samuelsson L. (2013). A new paradigm for adaptive management. Ecol. Soc. 18, 63. doi: 10.5751/ES-06183-180463

Crossref Full Text | Google Scholar

Rockström J., Steffen W., Noone K., Persson Å., Chapin F. S., Lambin E. F., et al. (2009). A safe operating space for humanity. Nature 461, 472–475. doi: 10.1038/461472a

PubMed Abstract | Crossref Full Text | Google Scholar

Rölfer L., Celliers L., and Abson D. (2022). Resilience and coastal governance: Knowledge and navigation between stability and transformation. Ecol. Soc. 27. doi: 10.5751/ES-13244-270240

Crossref Full Text | Google Scholar

Safi G., Giebels D., Arroyo N. L., Heymans J. J., Preciado I., Raoux A., et al. (2019). Vitamine ENA: A framework for the development of ecosystem-based indicators for decision makers. Ocean. Coast. Manage. 174, 116–130. doi: 10.1016/j.ocecoaman.2019.03.005

Crossref Full Text | Google Scholar

Salafsky N., Margoluis R., and Redford K. (2001). Adaptive management: A tool for conservation practitioners (World Wildlife Fund, Washington, DC: Biodiversity Support Program).

Google Scholar

Schultz L., Folke C., Österblom H., and Olsson P. (2015). Adaptive governance, ecosystem management, and natural capital. Proc. Natl. Acad. Sci. 112, 7369–7374. doi: 10.1073/pnas.1406493112

PubMed Abstract | Crossref Full Text | Google Scholar

Seddon N., Chausson A., and Berry P. (2020). Understanding the value and limits of nature-based solutions to climate change and other global challenges. Philos. Trans. R. Soc. B. 375, 20190120. doi: 10.1098/rstb.2019.0120

PubMed Abstract | Crossref Full Text | Google Scholar

Semeoshenkova V., Newton A., and Rojas M. (2016). A combined DPSIR and SAF approach for adaptive management of beach erosion in Monte Hermoso and Pehuen Co (Argentina). Ocean. Coast. Manage. 130, 179–189.

Google Scholar

Shen G., Huang L., and Guo F. (2009). Marine ecology (in Chinese) (Beijing: Science Press).

Google Scholar

Smit B. and Wandel J. (2006). Adaptation, adaptive capacity and vulnerability. Global Environ. Change 16, 282–292. doi: 10.1016/j.gloenvcha.2006.03.008

Crossref Full Text | Google Scholar

Stankey G. H., Clark R. N., and Bormann B. T. (2005). Adaptive management of natural resources: Theory, concepts, and management institutions (Portland, OR: USDA Forest Service, Pacific Northwest Research Station).

Google Scholar

Steinitz C. (1990). A framework for theory applicable to the education of landscape architects. J. Landscape 9, 136–143. doi: 10.3368/lj.9.2.136

Crossref Full Text | Google Scholar

Sue K., Andy P., and Chris F. (2011). The ecosystem approach to marine planning and management (London: Earthscan Ltd.).

Google Scholar

Tscherning K., Helming K., Krippner B., Sieber S., and Paloma S. G. (2012). Does research applying the DPSIR framework support decision making? Land. Use Policy 29, 102–110. doi: 10.1016/j.landusepol.2011.05.009

Crossref Full Text | Google Scholar

van Leeuwen J., Raakjaer J., Van Hoof L., van Tatenhove J., Long R., and Ounanian K. (2014). Implementing the Marine Strategy Framework Directive: A policy perspective. Mar. Policy 50, 325–330. doi: 10.1016/j.marpol.2014.03.004

Crossref Full Text | Google Scholar

Verboom J. and Wamelink W. (1999). “Spatial modeling in landscape ecology,” in Issues in landscape ecology: Proceedings of the Fifth World Congress. Eds. Wiens J. A. and Moss M. R. (International Association for Landscape Ecology, Snowmass Village, CO, USA), 38–44.

Google Scholar

Walker B., Holling C. S., Carpenter S. R., and Kinzig A. (2004). Resilience, adaptability and transformability in social–ecological systems. Ecol. Soc. 9. doi: 10.5751/ES-00650-090205

Crossref Full Text | Google Scholar

Wang W. (2024). The functioning mechanism of a collaborative environmental governance network in a coastal zone: A case study of the Wenzhou Dongtou coastal zone. Sustainability 16, 10159. doi: 10.3390/su162310159

Crossref Full Text | Google Scholar

Wang Q., Meng J., Qi Y., Peng F., and Mao X. (2015). Ecosystem adaptive management based on dynamically changed ecological carrying capacity: A case study of Uxin Banner in Ordos (in Chinese). Areal. Res. Dev. 34 (4), 154–159. doi: 10.3969/j.issn.1003-2363.2015.04.029

Crossref Full Text | Google Scholar

Williams B. K. (2011). Adaptive management of natural resources—framework and issues. J. Environ. Manage. 92, 1346–1353. doi: 10.1016/j.jenvman.2010.10.041

PubMed Abstract | Crossref Full Text | Google Scholar

Young O. R. (2010). Institutional dynamics: Resilience, vulnerability and adaptation in environmental and resource regimes. Global Environ. Change 20, 378–385. doi: 10.1016/j.gloenvcha.2009.10.001

Crossref Full Text | Google Scholar

Yu K. J. (1996). Security patterns and surface model in landscape planning. Landscape Urban. Plann. 36, 1–17. doi: 10.1016/S0169-2046(96)00331-3

Crossref Full Text | Google Scholar

Zhai J., Pu L., and Qie L. (2025). Changes of land use and landscape pattern along sea–land gradient in developed coastal region: A case study of Jiangsu Province, China. Ecol. Indic. 176, 113635. doi: 10.1016/j.ecolind.2025.113635

Crossref Full Text | Google Scholar

Zhang D., Ren H., Sun P., Jing P., and Guo B. (2023). Construction of multi-level ecological security network in fragmented nature landscape using the three-dimensional framework of ecological adaptability. Ecol. Indic. 157, 111229. doi: 10.1016/j.ecolind.2023.111229

Crossref Full Text | Google Scholar

Zhang X., Wang L., Fu X., Li H., and Xu C. (2017). Ecological vulnerability assessment based on PSSR in Yellow River Delta. J. Clean. Prod. 167, 1106–1111. doi: 10.1016/j.jclepro.2017.04.106

Crossref Full Text | Google Scholar

Zhejiang Provincial Government (2021). Dongtou District – Zhejiang Province overview. Available online at: http://www.ezhejiang.gov.cn/wenzhou/2021-08/19/c_653179.htm.

Google Scholar

Zou J., Dong L., and Qin B. (1983). Preliminary study on eutrophication and red tide problems in Bohai Bay (in Chinese). Mar. Environ. Sci. 2, 41–54.

Google Scholar