Synergistic Development Mechanism Between Reservoir Resettlers’ Livelihoods and Host Regions

Abstract

The sustainability of reservoir resettlement depends on the synergistic development of resettlers’ livelihoods and host regions; however, existing studies lack an integrated analytical framework. Combining the Sustainable Livelihoods Framework with synergistic development theory, this study establishes a dual-system evaluation model comprising the Regional Development Support (RDS) and Resettlers’ Livelihood Development (RLD) indices. Using survey data from 289 households across 10 counties in Zhejiang’s QC Reservoir project, we apply composite weighting, coupling coordination modeling, and spatial analysis to evaluate the levels of synergistic development and examine spatial patterns. The findings reveal that (1) there is significant gradient differentiation in the Synergistic Development Index (SDI), with scores ranging from 0.134 to 0.738; (2) spatial autocorrelation is weak (Moran’s I = −0.089), reflecting industrial heterogeneity; and (3) four distinct coordination types are identified, with employment–industry mismatch and ecological constraints being the primary limiting factors. This study provides a diagnostic framework for assessing resettlement outcomes and offers guidance for formulating differentiated policy interventions.

1. Introduction

Water conservancy and hydropower projects play a crucial role in optimizing national water resources, mitigating flood risks, and ensuring energy security. However, their construction is often accompanied by large-scale, involuntary population displacement and resettlement. In China, such projects have relocated more than 25 million individuals in recent decades, making reservoir resettlement one of the world’s most complex and persistent development challenges [1]. Reservoir resettlers lose land, housing, and other livelihood foundations due to inundation and are collectively relocated to designated resettlement areas. These relocations entail both physical movement and a range of challenges, including livelihood reconstruction, social network reestablishment, and the restructuring of regional development models [2]. Studies in other national contexts have also emphasized the risks associated with resettlement; for instance, flood resettlers in Pakistan face multidimensional vulnerabilities after relocation, including heightened exposure to socioeconomic shocks, rupture of social ties, and compounded risks to physical and mental well-being [3]. These transformations position post-resettlement development as a representative coupled human–environment process characterized by substantial uncertainty and spatial heterogeneity.

Current research on involuntary resettlement primarily draws on Chambers and Conway’s Sustainable Livelihoods concept [4], the livelihood resilience framework [5], and the Impoverishment Risks and Reconstruction model [6]. These frameworks emphasize the sustainability of resettlers’ livelihoods as a key criterion for evaluating resettlement outcomes. The Chinese government has implemented a series of relocation and follow-up support policies that have successfully secured basic living conditions for resettlers. Nevertheless, in the current stage of China’s development, several critical challenges remain: how to facilitate a transition for resettlers from simply achieving “stable housing” to attaining “development capacity” and “prosperity,” and how to foster deeper coordination between the sustainability of resettlers’ livelihoods and broader regional development. Addressing these issues is vital for narrowing regional development disparities and maintaining social stability. These concerns extend beyond individual household welfare to influence social cohesion within resettlement areas and the long-term sustainability of regional economies.

Reservoir resettlement areas exemplify contexts in which micro-level livelihoods must be closely aligned with macro-level regional development. The enhancement and accumulation of resettlers’ livelihood capital depend on the industries, public services, and ecological resources available in resettlement regions, while the sustainable growth of these regions, in turn, relies on resettlers’ contributions through labor supply, consumption demand, and social participation. In practice, mismatches are frequently observed: some regions exhibit rapid economic development with limited benefits for resettlers, whereas others demonstrate high livelihood potential among resettlers but inadequate regional support. Evidence increasingly indicates that the relationship between regional development and resettlers’ livelihood recovery is neither linear nor deterministic. Instead, resettlement outcomes depend on how effectively local development pathways, industrial structures, social support systems, and environmental capacities align with resettlers’ adaptive capacities, resource endowments, and livelihood strategies. This dynamic interaction forms the conceptual foundation of what this study defines to as synergistic development.

Synergistic development refers to the mutually reinforcing and co-adaptive evolution between resettlers’ livelihood systems and the development trajectories of host regions. Rather than representing absolute levels of economic performance or livelihood recovery, synergy reflects the interactional compatibility between the two systems—whether regional structures offer appropriate opportunities for resettlers, whether resettlers success-fully integrate into regional development pathways, and whether both systems evolve in ways that mutually enhance rather than constrain one another. Under this framework, a region with moderate economic development may nonetheless exhibit high synergy if its industrial opportunities, policy environment, and sociocultural context align closely with resettlers’ livelihood strategies. Conversely, a region with strong economic performance may demonstrate low synergy if resettlers encounter institutional, structural, or cultural barriers to integration. Therefore, synergistic development is relational rather than hierarchical; it reflects systemic compatibility rather than absolute developmental scale.

Building on this framework, the study introduces the Synergistic Development Index (SDI) as an analytical tool to quantify the degree of co-evolution between the Regional Development Support (RDS) system and the Resettlers’ Livelihood Development (RLD) system. The RDS subsystem follows a three-dimensional “economic–social–ecological” coordination principle, emphasizing industrial foundations, public services, infrastructure, ecological environments, and governance policies as external supports for the accumulation and enhancement of livelihood capitals. The RLD subsystem, grounded in the five capitals of the Sustainable Livelihood Framework (SLF)—i.e., natural, financial, human, social, and physical capital—directly represents resettlers’ livelihood conditions. The SDI evaluates how these two systems interact, reinforce, or constrain each other, thereby providing an analytical foundation for identifying differentiated patterns and mechanisms of coordinated or discordant development across regions.

This study focuses on the QC Reservoir resettlement areas in Zhejiang Province as a representative case. The research contributes in three main respects. First, the QC Reservoir integrates economic benefits, ecological conservation, and resettlement needs, making it a pertinent context for analyzing synergistic development. Second, the cross-municipal nature of the resettlement encompasses 10 regions with considerable variation in agricultural, industrial, and commercial structures, allowing for comparative analysis of regional differentiation in coordination levels. Third, resettlers have entered a post-resettlement phase of livelihood stabilization, enabling current data on income, employment, and social adaptation to more accurately capture the state of synergistic development and improve the representativeness of the data.

1.1. Literature Review

Existing research largely falls into two relatively independent domains: “resettlers’ livelihoods” and “regional coordination,” with most studies employing a single theoretical framework. Integrated analytical approaches remain limited. This section is organized into three parts: micro-level livelihood studies, macro-level regional coordination studies, and limitations in cross-disciplinary integration. The objective of this study is to develop a research model that systematically combines these two dimensions to address the existing analytical gap.

The SLF serves as the primary theoretical foundation for analyzing resettlers’ livelihoods. Its core components include the vulnerability context, livelihood capitals, livelihood strategies, and livelihood outcomes [4]. International scholarship has primarily emphasized the roles of livelihood capitals, strategies, and associated risks. For instance, an empirical study of indigenous rural households in Mexico found that human capital (HC) and social capital (SC) significantly influence livelihood diversification [7]; a hydropower resettlement case in Nepal identified physical capital (PC) as a critical determinant of livelihood diversification [8]; research on a Pakistani hydropower project confirmed the positive contribution of PC to quality of life improvement [9]; and a study of refugees in Bangladesh revealed that deficits in natural capital (NC) and financial capital (FC) severely constrain long-term livelihood development [10]. These findings indicate that the mechanisms through which livelihood capitals function vary substantially depending on governance systems and regional contexts.

Complementing the SLF, the Community Capitals Framework (CCF), proposed by Cornelia Flora and Jan Flora, serves as another foundational theoretical model in rural sociology and community sustainability research. It categorizes community resources into seven dimensions: natural, human, social, cultural, political, financial, and built (physical) capital [11,12]. The framework underscores that community development relies on the dynamic equilibrium and cumulative enhancement of these capitals [13], thereby addressing the omission of non-economic dimensions in the five-capital model of the SLF. The CCF has been extensively applied in studies on livelihood improvement, disaster adaptation, and industrial development across diverse global contexts. For example, it has supported livelihood enhancement among cocoa producers in Colombia [14] and strengthened urban food security in Iran [15], illustrating both its explanatory power for complex community processes and its practical applicability.

In China, within the specific context of reservoir resettlement, political capital—primarily manifested through participation in decision-making—is largely embedded within resettlers’ social capital. Similarly, aspects of cultural capital are often expressed through social capital. Consequently, domestic research generally operates within the SLF and focuses on optimizing the composition and coordination of resettlers’ livelihood capitals. For example, Wang et al. [16] proposed a risk-analysis framework for differentiated livelihood strategies among reservoir resettlers from a resilience perspective, while Li et al. [17] introduced psychological capital to enrich the SLF. However, this body of work primarily focuses on static assessments of household-level livelihood capitals and often overlooks their interaction with broader regional development contexts. Consequently, it provides limited insight into how factors such as industrial layout and public services in resettlement areas influence the accumulation of livelihood capitals. Therefore, the first identified research gap concerns the lack of analytical integration between micro-level livelihoods and macro-level regional development.

Studies on macro-level regional coordination frequently adopt synergetics theory, which emphasizes the interaction mechanisms among subsystems. Originating from Haken’s theoretical framework, synergetics offers a systematic perspective for analyzing how localized interactions foster overall systemic order [18]. For example, Galderisi et al. [19], using case studies from peripheral regions in Italy, proposed a coordination model integrating “community participation and regional resource integration”. Wen et al. [20] developed a coordinated governance framework for watershed management encompassing water resources, hydraulic infrastructure, and socio-economic systems. These studies provide valuable methodological insights into cross-system coordination. Some studies align more closely with the Chinese policy context. For instance, Wu et al. [21] proposed the “enclave cluster resettlement” model emphasizing coordination between resettlers and local governments, while Chen et al. [22] highlighted that livelihood transformation requires dynamic alignment among economic, social, and ecological systems. Nonetheless, most of these studies emphasize macro-level conditions—such as infrastructure and industrial distribution—while neglecting detailed analyses of household-level decision-making and behavioral dynamics. Consequently, the second research gap involves the absence of micro-foundations in macro-level regional analysis, leaving unclear the specific mechanisms through which regional development enhances resettlers’ livelihoods.

A substantial research gap persists at the intersection of resettlers’ livelihood development and regional synergistic development. Few studies have examined how macro-regional factors shape livelihood systems within an integrated analytical framework. For instance, Liu et al. [23] investigated how relocation type and regional heterogeneity influence livelihood vulnerability; Liu et al. [24] incorporated market and financial macro-factors into evaluations of household livelihood sustainability; Wang et al. [25] demonstrated that both macro-regional conditions and micro-level household capitals jointly shape farmers’ behavioral decisions; and Su et al. [26] found that policy interventions such as the development of specialized industries strengthen sustainable livelihood capacities. However, these studies have not established a quantitative indicator system for evaluating coordination levels. The third major research gap therefore concerns the absence of robust methodological tools for quantitatively assessing the synergy between livelihood systems and regional development.

Currently, there is no unified academic standard for assessing the coupled or synergistic development between resettlers’ livelihoods and host region development; instead, researchers typically construct composite indices to integrate multidimensional data into a single, comparable metric [27]. For example, Wang et al. [28] employed the degree of coupling and coordination model (DCCM) to quantify coordination among subsystems within farmer livelihoods, offering a methodological foundation for this study. Accordingly, following comprehensive measurement of RDS and RLD, this research applies the DCCM to quantify the interaction intensity and overall development level between the two systems. This approach facilitates precise identification of the process of synergistic evolution—from subsystem interaction to ordered development—and provides a scientific basis for evaluating coordination across diverse resettlement regions.

1.2. Research Objectives and Innovations

To address the aforementioned research gaps, this study divides the analytical framework into two interrelated subsystems—RDS and RLD—and constructs the SDI to evaluate their level of coordination. The main innovations are as follows:

(1)

Integration of SLF and regional synergy theory to bridge the macro–micro divide. The coordination between RDS and RLD establishes a bidirectional mechanism that promotes mutual reinforcement between livelihood improvement and regional development.

(2)

Construction of the SDI as a diagnostic tool for evaluating the interactional compatibility within the dual-system (RDS–RLD) framework. The development levels of both subsystems are determined through a combined weighting method and the Technique for Order Preference by Similarity to an Ideal Solution (TOPSIS). The DCCM is then applied to quantify the degree of alignment between the two systems and derive the SDI, while Moran’s I is used to assess spatial autocorrelation. This integrated methodology provides an operational analytical framework for evaluating coordination dynamics.

(3)

Identification of differentiated synergy pathways using 10 cross-county case studies. The results offer empirical evidence to guide the formulation of more targeted and differentiated resettlement policies.

1.3. Technical Approach

Building upon the conceptualization of synergistic development and the dual-system framework described in the preceding sections, this study establishes an integrated analytical model linking RLD with RDS. The technical approach consists of four main components:

(1)

Field survey and dual-system indicator construction. Based on field investigations conducted at QC Reservoir resettlement sites, indicator systems are developed for both subsystems to comprehensively assess RDS and RLD. Indicator weights are determined using the AHP–EW method, and composite subsystem scores are calculated through TOPSIS.

(2)

Construction of the SDI. A unified n-system DCCM formulation is employed to quantify both internal subsystem coordination and cross-system coordination. The SDI functions as a diagnostic measure of the interactional compatibility and co-evolution between RDS and RLD.

(3)

Spatial correlation analysis. The spatial distribution patterns and interregional relationships of the SDI are examined using Moran’s I to evaluate spatial autocorrelation and clustering characteristics.

(4)

Policy recommendations. By integrating the comprehensive outcomes of both subsystems and the SDI, the study proposes actionable policy recommendations aimed at promoting coordinated and high-quality development of resettlers’ livelihoods and resettlement regions.

The overall research workflow is depicted in Figure 1.

The remainder of this paper is organized as follows: Section 2 describes the study area, data sources, and research methods; Section 3 presents and interprets the empirical findings; and Section 4 summarizes key conclusions and policy implications.

2. Research Model and Methods

2.1. Study Area and Data Sources

2.1.1. Study Area

This study employs the QC Reservoir resettlement area in Zhejiang Province, China, as a case study, based on the following considerations:

(1)

The QC Reservoir is a major hydraulic hub project that primarily provides water supply and flood control functions, with secondary roles in downstream irrigation and power generation. As it simultaneously addresses ecological protection, population resettlement, and regional development, it constitutes a representative context for examining synergistic development.

(2)

The relocation of QC Reservoir resettlers was completed in 2017. Fieldwork conducted in 2023 indicates that resettlers have largely restored their production and living standards. Indicators such as income, housing, and social adaptation have stabilized, ensuring that the collected data are temporally reliable and representative.

(3)

The QC Reservoir resettlement area encompasses 10 county-level administrative divisions. To maintain geographic confidentiality, these regions are anonymized and labeled R1–R10 according to their functional classification. Resettlers were relocated from R1 to ten regions, with R1 serving as the primary resettlement area and the remaining regions functioning as external dispersed resettlement zones. The allocation of resettlers across these regions was generally proportional to the water supply benefits that each region derived from the project. The distribution of resettled households is presented in Figure 2. Initial resettlement allocations were determined by random selection, after which resettlers were allowed to negotiate and adjust placements according to their development needs. The marked variation in economic development and industrial structures among these regions provides an ideal dataset for comparative analysis of coordination levels.

2.1.2. Data Sources

The data used in this study consist of two primary components.

(1) For RDS-related indicators, data were obtained from government statistical records and interviews with relevant agencies, including the Water Resources Bureau and the Statistics Bureau. To ensure reliability, five participating experts reviewed and refined the indicator system.

For indicators related to RDS, data were obtained through a mixed approach that combined official government statistics with expert-based standardized evaluations to ensure data reliability. Macroeconomic and industrial data were sourced from specialized reports issued by relevant government departments in Zhejiang Province and the ten study regions (including the Bureaus of Water Resources, Statistics, and Ecology and Environment). Key sources included the Zhejiang Statistical Yearbook 2024 (China Statistics Press, ISBN: 978-7-5230-0502-6) and publicly accessible datasets from official government websites (e.g., Zhejiang Provincial Bureau of Statistics: https://tjj.zj.gov.cn/; Zhejiang Provincial Department of Ecology and Environment: https://sthjt.zj.gov.cn/; and regional government information disclosure platforms).

For specialized indicators not covered by existing public sources—primarily resettlement-specific measures such as training coverage and industrial employment absorption—data were derived from annual work reports and semi-structured interviews with core resettlement management professionals and township leaders responsible for resettlement (conducted from 20 to 26 November 2023, and 28 to 30 January 2024).

For indicators requiring comprehensive evaluation, standardized scoring was performed by five experts: one professor specializing in resettlement studies, two senior engineers from design institutes, and two government officials with extensive practical experience in resettlement, each with over ten years of professional experience in reservoir resettlement. All experts provided independent assessments and were not involved in implementing or managing the QC Reservoir resettlement program, ensuring impartiality in the evaluation process. Expert scoring was strictly based on the objective data described above and conducted using a two-round Delphi process. After verifying the completeness and consistency of all objective data, experts independently assigned scores; indicators with a standard deviation greater than 0.8 underwent a second-round evaluation following group discussion. The final results passed Kendall’s coefficient of concordance test (Kendall’s W > 0.7 for all indicators, p < 0.05), confirming strong consistency and reliability while minimizing subjective bias. Detailed scoring criteria and expert profiles are available from the corresponding author upon request.

(2) Data for the RLD were obtained from a field survey conducted at the QC Reservoir resettlement sites in November 2023. The research team comprised one professor, one associate professor, five senior engineers with expertise in reservoir resettlement, and eight graduate students who received standardized training. The survey was carried out in three stages: (i) in-depth interviews with managers of local resettlement agencies; (ii) assessment of the livelihood development environment across the resettlement sites; and (iii) one-on-one household interviews using a structured questionnaire. The questionnaire was pre-tested and verified to have high reliability and validity. It collected data on household demographics, employment status, agricultural production, housing conditions, social adaptation, and the protection of resettlers’ rights and interests in 2023.

This study adopted a stratified random sampling method to select resettler households. First, the total sample size was determined based on the 2224 resettled households, following the minimum 5% sampling requirement specified in China’s Code for Resettlement Supervision and Evaluation of Water and Hydropower Projects (SL716-2015). Second, samples were proportionally allocated according to the number of resettled households in each region to reflect their spatial distribution. Because several regions (e.g., R2 and R5) have relatively small resettlement populations, their sampling ratios were intentionally increased to avoid under-representation. Although the absolute numbers of sampled households in these regions were small, their sampling proportions closely matched actual population shares, ensuring representativeness of the spatial distribution. Within each region, an additional stratification was performed based on current livelihood levels of resettlers (high/medium/low = 2:6:2), and sub-samples were drawn to capture intragroup heterogeneity and enhance overall representativeness. In total, 289 valid questionnaires were collected, yielding a sampling ratio of 13.0%. The distribution of sampled resettler households is shown in

Table 1. Distribution of resettlers and sample households.

Region	Actual Resettled Households	Sampled Households	Regions	Actual Resettled Households	Sampled Households
R1	1875	178	R6	70	20
R2	18	10	R7	76	21
R3	32	10	R8	40	10
R4	34	10	R9	30	10
R5	22	10	R10	27	10
Total	2224	289

.

2.2. Construction of the Evaluation Indicator System

2.2.1. Regional Development Support System

The sustainable development of resettlement areas provides the essential environmental foundation for stabilizing and improving resettlers’ livelihoods and forms the basis for coordinated regional–livelihood development. Although there is no universally recognized standard for constructing RDS indicators, the Economy–Social–Environment principle asserts that dynamic balance among these three dimensions promotes synergistic social development [29]. Drawing on relevant research and considering the developmental orientation of the QC Reservoir project and the industrial characteristics of its resettlement areas, this study emphasizes the capacity of resettlement regions to sustain resettlers’ livelihoods. Accordingly, the RDS subsystem is organized into three dimensions: Economic Support (ES), Social Support (SS), and Ecological Support (ECS). Eleven indicators are selected to develop a comprehensive evaluation framework (see

Table 2. Indicator system for the Regional Development Support System.

Dimension	Indicator	Indicator Description
Economic Support (ES)	Per capita GDP growth rate (ES1)	(Current year per capita GDP—previous year per capita GDP)/previous year per capita GDP × 100%.
	Share of tertiary industry value added (ES2)	Tertiary industry value added ÷ GDP × 100%.
	Performance of characteristic industries (ES3)	Based on industry scale, market position, technological innovation, development trend, and driving effect, composite expert score from 1 (very poor) to 5 (excellent).
	Employment absorption rate by resettler industries (ES4)	(Number of resettlers employed in primary characteristic industries/total employment in those industries) × 100%.
Social Support (SS)	Proportion of children attending local schools (SS1)	(Number of resettler children enrolled locally/total number of resettler children aged 6–15 in resettler households) × 100%.
	Community comprehensive service coverage (SS2)	(Number of resettler villages providing cultural integration, convenience services/total number of resettler villages) × 100%.
	Coverage of specialized resettler training (SS3)	(Number of resettler laborers who received industry-adapted skills training, employment guidance, and other targeted services/total number of resettler laborers) × 100%.
	Infrastructure convenience (SS4)	Based on transportation accessibility, completeness of living facilities, and industry supporting capacity, composite expert score from 1 (very poor) to 5 (excellent).
Ecological Support (ECS)	Per capita cultivated land area (ECS1)	Per capita cultivated land area for resettlers in the resettlement area.
	Ecological service supply (ECS2)	Based on water quality safety, forest coverage, ecological engineering, and pollution control, composite expert score from 1 (very poor) to 5 (excellent).
	Agricultural output value per unit sown area (ECS3)	(Total output value of agriculture, forestry, animal husbandry and fishery × agricultural output value extraction coefficient)/total sown area.

). The primary objective is to highlight the role of regionally driven development—under governmental coordination—in facilitating the sustainable improvement of resettlers’ livelihoods.

ES represents the material foundation for the coordinated development of resettlers’ livelihoods and resettlement regions, reflecting the industrial base and capacity to absorb resettler employment. The per capita GDP growth rate (ES1) reflects a region’s economic growth potential, while the proportion of tertiary industry value added (ES2) indicates the modernization level of the local economic structure. The performance of characteristic industries (ES3) and the employment absorption rate of resettlers in characteristic industries (ES4) directly measure industrial outcomes and the extent to which resettlers benefit from them. The structure, scale, and spatial distribution of characteristic industries determine the long-term viability of the resettlement area and serve as the primary channels for achieving “industry–resettler” coordinated development.

SS provides a structural safeguard for RLD by capturing the allocation and integration of social resources within resettlement areas. As compulsory education from primary to junior secondary school is universally accessible in China, the proportion of resettlers’ children enrolled in local schools (SS1) more accurately reflects the level of educational integration following relocation. Community comprehensive service coverage (SS2) assesses the availability of public and cultural services, such as cultural exchange programs and convenience services, which support diverse population needs. The coverage of specialized resettler training (SS3) indicates the effectiveness of government-supported vocational and employment training initiatives and the strength of the region–industry–resettler connection. Infrastructure convenience (SS4) evaluates the adequacy of basic infrastructure in supporting resettlers’ employment and daily life, thereby enriching the basis for assessing system coordination.

ECS, grounded in China’s commitment to green development, serves as the foundation for the long-term sustainability of resettlement regions. The per capita cultivated land area (ECS1) represents the local land resource endowment, while the ecological service supply score (ECS2)—constructed from a four-dimensional model encompassing water quality safety, forest coverage, ecological engineering, and pollution control—captures the region’s overall capacity to provide ecological services. Agricultural output value per unit sown area (ECS3) reflects the degree of agricultural modernization and productivity, serving as a critical indicator of ecological–economic synergy.

2.2.2. Resettlers’ Livelihood Development System

Drawing on the SLF, this study identifies 14 indicators across five types of livelihood capital—natural, financial, human, social, and physical—tailored to the actual circumstances of QC Reservoir resettlers to construct the livelihood development indicator system (see

Table 3. Indicator system for the Resettlers’ Livelihood Development System.

Dimension	Indicator	Indicator Description
Natural Capital (NC)	Per capita cultivated land area (NC1)	Cultivated land area obtained by resettler households in the resettlement area
	Cultivated land quality grade (NC2)	Overall quality level of cultivated land (soil fertility, irrigation conditions), scored from 1 (very poor) to 5 (excellent)
	Degree of match between cropping structure and characteristic agriculture (NC3)	1 = abandoned; 2 = self-cultivation focusing on staple food; 3 = land leased out or self-cultivation of local characteristic agricultural products; 4 = joined a cooperative for unified operation
Financial Capital (FC)	Household per capita annual income (FC1)	Total annual household income/household population
	Income diversity index (FC2)	Calculated using the Shannon–Wiener index (Equation (1))
	Number of types of government subsidies and preferential supports (FC3)	Types of government grants and loan programs resettlers can effectively access
Human Capital (HC)	Average household education level (HC1)	(1 × number with primary school or below + 2 × number with junior secondary + 3 × number with senior secondary/technical secondary + 4 × number with university or above)/total population
	Household labor ratio (HC2)	(Number of adult laborers in household/total household population) × 100%
	Labor employment rate (HC3)	(Number of employed adult laborers in household/total adult labor force) × 100%
	Proportion of labor with employment skills (HC4)	(Number of household laborers holding vocational skill certificates/total resettler labor force) × 100%
	Employment–industry matching degree (HC5)	Number of resettlers employed in skill-matched positions/total number of employed resettlers × 100%
Social Capital (SC)	Interpersonal interaction frequency (SC1)	Monthly frequency of visiting/exchanging with local friends: very rare (≤1 time) = 1; rare (2–3 times) = 2; moderate (4–6 times) = 3; medium-high (7–9 times) = 4; frequent (≥10 times) = 5
	Participation in local community activities (SC2)	1 = never participate, 2 = very rarely, 3 = occasionally, 4 = often, 5 = community organization member
	Satisfaction with neighborhood relations (SC3)	Very dissatisfied = 1; dissatisfied = 2; neutral = 3; fairly satisfied = 4; very satisfied = 5
Physical Capital (PC)	Per capita housing area (PC1)	Residential building area of the resettler household in the resettlement area/household population
	Renovation level of resettlement housing (PC2)	Unrenovated = 1; basic = 2; ordinary = 3; good = 4; deluxe = 5
	Number of household durable consumer goods (PC3)	Count items such as computer, motorcycle, e-bike (1 point each); cars counted per vehicle

for detailed indicator descriptions). From a coordinated development perspective, these indicators capture the capacity of resettler households to enhance their livelihoods through the accumulation and interaction of multiple forms of capital. Data for this subsystem were obtained through field questionnaire surveys of sampled households.

NC focuses on the synergy between cultivated land use and regionally characteristic agriculture. The per capita cultivated land area (NC1) and cultivated land quality grade (NC2) represent the quantity and quality of natural resources accessible to resettler households. The degree of match between cropping structure and characteristic agriculture (NC3) reflects the extent to which household agricultural activities align with dominant regional agricultural practices.

FC highlights the coordination among policy support, resettler industries, and household income. Household per capita annual income (FC1) represents the income level of surveyed households, while the income diversity index (FC2), calculated using the Shannon–Wiener index [30], evaluates both the variety and structural balance of income sources. Its formula is:

$$S=-{\displaystyle {\sum }_{i=1}^n\left(p_i\times \ln \left(p_i\right)\right)},$$

(1)

where $p_i$ is the proportion of income source i in the total household income.

The number of types of government subsidies and preferential supports (FC3) reflects the degree of coordinated policy and financial assistance available to enhance resettlers’ livelihoods and industrial participation.

HC emphasizes the consistency between labor skills and regional industrial demand. The coordinated development of resettlers’ livelihoods relies on the alignment between the regional industrial structure and employment quality, consistent with empirical findings by Huang et al. [31] regarding social integration among migrant workers. The average household education level (HC1) represents the baseline capacity for policy comprehension and industrial participation. The household labor ratio (HC2), labor employment rate (HC3), and proportion of labor with employment skills (HC4) assess labor readiness, while the employment–industry matching degree (HC5) measures the alignment between existing labor skills and local industrial requirements.

SC captures the extent of social integration achieved by resettlers. Three indicators are selected: interpersonal interaction frequency (SC1), participation in local community activities (SC2), and satisfaction with neighborhood relations (SC3). Collectively, these indicators assess both the breadth and quality of post-resettlement social networks.

PC evaluates the interaction between basic living conditions and the urbanization level of resettlement areas. Per capita housing area (PC1) reflects baseline housing adequacy in urbanized resettlement contexts; renovation level of resettlement housing (PC2) represents improvements in housing quality; and the number of household durable consumer goods (PC3) serves as an indicator of consumer adaptation and integration into the resettlement environment.

2.3. Composite Measurement Method for Subsystems

To evaluate the development levels of RDS and RLD, this section first determines indicator weights using the AHP–EW method and then applies TOPSIS to calculate composite scores. This hybrid approach minimizes the subjective bias inherent in AHP while addressing the entropy method’s limitation in reflecting the practical relevance of indicators [32].

2.3.1. Determination of Indicator Weights

(1)

Determining subjective weights using AHP

The AHP is a subjective weighting technique grounded in expert judgment. It structures complex problems into hierarchical levels and constructs pairwise comparison matrices to quantify expert assessments of the relative importance of indicators. To determine the subjective weights, the five experts described in Section 2.1.2 independently performed pairwise comparisons of indicators within each hierarchical level using the Delphi-based Analytic Hierarchy Process (AHP) procedure. After the first round of scoring, all pairwise comparison matrices were evaluated using the consistency ratio (CR) test. Indicators that did not meet the consistency criterion in the first round were re-examined during a second Delphi round, in which experts discussed and revised their evaluations until all matrices achieved acceptable consistency. The CR values for all matrices in this study ranged from 0.008 to 0.059, fully satisfying the AHP requirement of CR < 0.1. The final subjective weights $W_{AHP}$ assigned to each indicator thus reflect expert consensus and conform to the internal consistency standards of AHP.

(2)

Determining objective weights using the entropy method

The entropy method is an objective weighting technique that evaluates indicator importance based on data dispersion. It determines the degree of influence of each indicator on the overall evaluation outcome by calculating its information entropy. Indicators with lower entropy values exhibit greater variability and therefore possess higher discriminative power, resulting in greater weights. This method has been widely applied in measuring regional development levels [33] and assessing household livelihood capitals [34].

First, indicator values are normalized using the min–max method to obtain standardized scores. Then, the proportion $f_{ij}$ of indicator j for sample i is calculated. The entropy $e_j$ of indicator j is derived using Equation (2), which quantifies the disorder of the data. Finally, the objective weight $W_{Entropy,j}$ is computed using Equation (3), where n is the number of samples and m is the total number of indicators.

$$e_j=-{\displaystyle \frac{1}{\ln n}}{\displaystyle {\sum }_{i=1}^n{f}_{ij}\ln f_{ij}}$$

(2)

$$W_{Entropy,j}={\displaystyle \frac{1-e_j}{{\displaystyle {\sum }_{j=1}^m\left(1-e_j\right)}}}$$

(3)

(3)

Weighted fusion to determine combined weights

AHP-derived subjective weights capture expert opinion and policy direction. For example, under the goal of coordinating resettlers’ livelihoods with regional development, the employment absorption rate of resettler industries (ES4) may be considered more important than per capita GDP growth rate (ES1). Conversely, entropy-derived objective weights reflect data variability; for example, a high variance in PC3 could result in a relatively large objective weight. To integrate these perspectives, combined weights $W_i$ are calculated via weighted fusion using Equation (4):

$$W_i=\alpha \cdot W_{AHP,i}+\beta \cdot W_{Entropy,i}$$

(4)

where α and β are fusion coefficients. In this study, (α = 0.6, β = 0.4) is set to emphasize the influence of government leadership and policy intervention, consistent with the practical characteristics of reservoir resettlement research.

Robustness checks employing alternative AHP–EW fusion coefficients (0.6–0.4, 0.5–0.5, 0.4–0.6) indicate that the SDI rankings remain highly stable, with Spearman rank correlation coefficients exceeding 0.90. Detailed results are presented in Appendix A,

Table A1. Robustness of the SDI to alternative AHP–EW fusion coefficients.

Region	SDI (α = 0.6, β = 0.4)	Rank (α = 0.6, β = 0.4)	SDI (α = 0.5, β = 0.5)	Rank (α = 0.5, β = 0.5)	SDI (α = 0.4, β = 0.6)	Rank (α = 0.4, β = 0.6)
R1	0.596	5	0.57	5	0.574	5
R2	0.315	9	0.314	9	0.311	9
R3	0.738	1	0.737	1	0.731	1
R4	0.725	2	0.687	4	0.694	4
R5	0.134	10	0.17	10	0.19	10
R6	0.428	8	0.515	6	0.531	6
R7	0.704	3	0.719	2	0.723	2
R8	0.685	4	0.698	3	0.701	3
R9	0.432	7	0.476	7	0.511	7
R10	0.441	6	0.459	8	0.456	8

and

Table A2. Spearman rank correlations among SDI rankings under alternative fusion coefficients.

Comparison	Spearman ρ
(α = 0.6, β = 0.4) vs. (α = 0.5, β = 0.5)	0.915
(α = 0.6, β = 0.4) vs. (α = 0.4, β = 0.6)	0.915
(α = 0.5, β = 0.5) vs. (α = 0.4, β = 0.6)	1.000

.

2.3.2. Calculation of Subsystem Composite Scores

TOPSIS is a widely applied method for comprehensive multi-indicator evaluation within systems. After adjusting the directionality of the original data and constructing a normalized decision matrix, TOPSIS calculates the Euclidean distance of each evaluation unit from both the positive ideal solution and the negative ideal solution. These distances are then used to determine the relative closeness of each sample to the ideal state, thereby establishing performance rankings [35]. The resulting values provide a clear representation of the extent to which each sample deviates from the optimal livelihood or regional development condition.

Assuming n samples and m indicators form an initial matrix $n\times m$, data are first normalized using the min–max method to produce the matrix $Z_{ij}$, where $z_{ij}$ is the normalized value of indicator j for sample i. Let $Z_j^{+}$ and $Z_j^{-}$ represent the positive (optimal) and negative (worst) ideal values of indicator j, respectively. Using the combined weights $W_j$ from Section 2.3.1, the distance of each sample iii from the ideal solutions—denoted $D_i^{+}$ and $D_i^{-}$—is calculated using Equation (5). The composite score $F_i$ for each sample is then derived using Equation (6):

$$\left\{\begin{array}{l}{D}_i^{+}=\sqrt{{\displaystyle {\sum }_{j=1}^m{W}_j\times {\left(Z_j^{+}-z_{ij}\right)}^2}}\\ D_i^{-}=\sqrt{{\displaystyle {\sum }_{j=1}^m{W}_j\times {\left(Z_j^{-}-z_{ij}\right)}^2}}\end{array}\right.$$

(5)

$$F_i={\displaystyle \frac{D_i^{-}}{D_i^{+}+D_i^{-}}}$$

(6)

This score reflects the relative performance of each sample in achieving subsystem objectives.

2.4. Measurement of the Synergistic Development Index

To quantitatively evaluate the synergistic development between the RDS and RLD, this study applies the DCCM, following the methodological framework proposed by Wang et al. [28] for assessing coupled systems. The DCCM is particularly well suited for analyzing nonlinear interactions among complex systems and enables a two-level evaluation: (i) internal coordination within each subsystem (RDS and RLD) and (ii) cross-system coordination, represented by the SDI. The former reflects the internal equilibrium of indicators within a subsystem, whereas the latter characterizes the interactional compatibility and co-evolution between the two subsystems.

Following Wang et al. [36], the coupling–coordination degree is computed within a unified n-system framework. Let $F_i$ denote the normalized development level of subsystem i, and let n be the total number of subsystems included. The coupling degree (C), comprehensive development index (T) and coupling coordination degree (D) are defined as:

$$C={\displaystyle \frac{{\left({\displaystyle \prod _{i=1}^n{F}_i}\right)}^{1/n}}{\frac{1}{n}{\displaystyle \sum _{i=1}^n{F}_i}}}$$

(7)

$$T={\displaystyle \sum _{i=1}^n{\beta }_i{F}_i}$$

(8)

$$D=\sqrt{C\times T}$$

(9)

This unified formulation allows for the evaluation of both internal subsystem coordination (where n equals the number of indicators within RDS or RLD) and cross-system coordination (where n = 2 for the RDS–RLD system). In the latter case, D is defined as the SDI, representing the extent to which the two systems evolve in a mutually adaptive and reinforcing manner.

The SDI should be understood as a diagnostic indicator of the interaction between the RDS and RLD systems, rather than as a measure of causal influence or absolute development performance. Because synergy represents system compatibility rather than magnitude, regions with moderate economic development may nonetheless exhibit high synergy when their development pathways align effectively with resettlers’ adaptive capacities and livelihood strategies.

${\beta }_i$ represents the corresponding weight, and ${\displaystyle \sum {\beta }_i}=1$. Following related research [30,37], this study applies equal weights (i.e., the arithmetic mean) to the indicators, thereby reflecting equal importance among all indicators within a subsystem. It is important to note that the arithmetic mean in Equation (8) functions as a structural parameter within the DCCM and does not imply any substantive weighting assumptions regarding the relative importance of the two subsystems. The substantive influence of each subsystem is instead represented by its corresponding development level ($F_i$).

Based on the value of D, coordination is classified into ten levels, with detailed classification criteria provided in

Table 4. Classification standards for coupling coordination degree.

Interval of D	Coordination Grade	Coupling Coordination Status	Interval of D	Coordination Grade	Coupling Coordination Status
[0.0–0.1)	1	Extremely disordered	[0.5–0.6)	6	Barely coordinated
[0.1–0.2)	2	Severely disordered	[0.6–0.7)	7	Primary coordination
[0.2–0.3)	3	Moderately disordered	[0.7–0.8)	8	Intermediate coordination
[0.3–0.4)	4	Mildly disordered	[0.8–0.9)	9	Good coordination
[0.4–0.5)	5	Near disordered	[0.9–1.0]	10	High-quality coordination

[36].

2.5. Spatial Autocorrelation Analysis

Moran’s I is a widely applied statistical measure used to determine whether the spatial distribution of multiple entities displays patterns of clustering, dispersion, or randomness [37]. The global Moran’s I assesses the overall spatial autocorrelation across all spatial units, whereas the local Moran’s I identifies specific regional clustering patterns within individual units [38]. In related studies on dual-system coupling, Moran’s I has been validated as a reliable metric for analyzing the spatial correlation between the two systems [39]. In this study, both global and local Moran’s I statistics are employed to examine whether the SDI values of the ten regions exhibit significant spatial autocorrelation.

Global Moran’s I is calculated using Equation (10), and local Moran’s I using Equation (11):

$$I_{Global}={\displaystyle \frac{{\displaystyle \sum _{i=1}^n{\displaystyle \sum _{j=1}^n{\omega }_{ij}\left(x_i-\overline{x}\right)\left(x_j-\overline{x}\right)}}}{s^2{\displaystyle \sum _{i=1}^n{\displaystyle \sum _{j=1}^n{\omega }_{ij}}}}},$$

(10)

$$I_{Local}={\displaystyle \frac{\left(x_i-\overline{x}\right){\displaystyle \sum _{j\ne 1}^n{{\omega }^{\prime }}_{ij}\left(x_j-\overline{x}\right)}}{s^2}},$$

(11)

where ${\omega }_{ij}$ denotes the spatial weight between county i and county j, and ${{\omega }^{\prime }}_{ij}$ is the row-standardized spatial weight matrix. $x_i$ represents the SDI value of county i, $\overline{x}$ is the mean SDI across all counties, $s^2$ is the variance, $i,j=\left(1,2,\dots ,n\right)$, and n = 10 is the sample size.

Moran’s I ranges from −1 to 1. A value I > 0 indicates spatial clustering (i.e., high–high or low–low adjacency), I < 0 indicates spatial dispersion (i.e., high–low adjacency), and I = 0 suggests a random spatial distribution. Statistical significance is assessed via the standardized Z-score (typically significant when $\left|Z\right|>1.96$, $\alpha =0.05$) and the p-value (p < 0.05 is considered to be statistically significant).

3. Results

3.1. Data Characteristics

Descriptive statistics for the 289 surveyed households are presented in

Table 5. Household characteristics of surveyed subjects.

Region	Average Household Population	Average Household Labor Population	Per Capita Cultivated Land Area (Mu)	Per Capita Housing Area (m2)	Per Capita Disposable Income (RMB)
R1	2.8	1.6	0.3	86.1	44,223
R2	3.8	2.5	0.9	31.6	48,746
R3	3.9	2.5	0.7	78.2	48,817
R4	2.8	1.5	1.3	87.8	48,975
R5	2.6	1.4	0.7	68.1	48,745
R6	3.9	2.7	0.5	84.0	47,054
R7	3.5	2.0	0.6	73.4	46,587
R8	3.4	2.7	0.6	99.2	43,285
R9	3.1	2.0	0.4	106.1	44,082
R10	3.6	2.2	0.6	70.8	43,502
Total	3.0	1.8	0.5	82.2	45,279

. On average, household size is 3.0 persons, and the per capita annual income is RMB 45,279. Following resettlement, the migrant population demonstrates characteristics of “adequate basic safeguards coupled with pronounced regional differentiation”. Considerable variation is observed across regions in both labor allocation and per capita cultivated land area. These baseline disparities directly affect resettlers’ ability to accumulate livelihood capitals and their alignment with regional industrial structures, thereby establishing the micro-level foundation for subsequent differences in coupling and coordination levels.

3.2. Regional Development Support System Analysis

3.2.1. Indicator Weight Determination for the Regional Development Support System

Using 2023 government statistical data for each region and information provided by resettlement management agencies, the AHP–EW method was applied to calculate indicator weights for the RDS subsystem based on evaluations from five domain experts (

Table 6. Calculation results of indicator weight for the RDS.

Dimension	AHP Weight	Entropy Weight	Combined Weight	Indicator	AHP Weight	Entropy Weight	Combined Weight
Economic Support (ES)	0.633	0.307	0.503	ES1	0.081	0.153	0.110
				ES2	0.154	0.222	0.182
				ES3	0.288	0.246	0.271
				ES4	0.476	0.379	0.437
Social Support (SS)	0.260	0.174	0.226	SS1	0.120	0.224	0.162
				SS2	0.260	0.238	0.251
				SS3	0.450	0.278	0.381
				SS4	0.171	0.259	0.206
Ecological Support (ECS)	0.106	0.519	0.271	ECS1	0.164	0.253	0.200
				ECS2	0.539	0.394	0.481
				ECS3	0.297	0.353	0.319

). All judgment matrices satisfied the consistency requirement (CR < 0.1). The resulting weight distribution reflects the coordinated logic characteristic of hydropower resettlement—economic priority, ecological constraint, and social foundation—highlighting the interdependent relationship between development, environmental protection, and livelihood support.

ES is the primary driver of coordinated development, with a combined weight of 0.503, functioning as the essential linkage between resettlers’ livelihoods and regional development. Industrial expansion and employment opportunities directly influence household income levels. The AHP-derived subjective weight for this dimension is 0.633, reflecting policy prioritization of economic growth, whereas the entropy-based objective weight is 0.308, indicating substantial data dispersion. Among the indicators, the employment absorption rate of resettler industries (ES4) holds the highest weight (0.437), signifying the capacity of industries to absorb labor. The performance of characteristic industries (ES3) also ranks highly (weight 0.271), representing the alignment between regional industrial structures and resettlers’ livelihood needs. These findings correspond to a synergistic development model characterized by “employment-led, industry-driven” growth.

ECS captures the distinctive ecological constraints and requirements of reservoir resettlement, with a combined weight of 0.271—higher than that of Social Support. The ecological service supply indicator (ECS2), encompassing water quality, forest coverage, ecological engineering, and pollution control, carries the greatest weight (0.462), serving as a core ecological constraint. The agricultural output value per unit sown area (ECS3) and per capita cultivated land area (ECS1) reflect the limited agricultural land resources in Zhejiang Province, underscoring the ecological specificity of the study area.

SS provides the foundational safeguards for livelihood development, with a combined weight of 0.226. SS indicators such as education, vocational training, and community services underpin the realization of synergistic development. The coverage of specialized resettler training (SS3) acts as a key linkage between industrial demand and skill enhancement (weight 0.381), while community service coverage (SS2) and infrastructure convenience (SS4) contribute to social integration and household livelihood stability.

3.2.2. System Performance Evaluation

Using the validated indicator weights and expert-reviewed data, the TOPSIS method was applied to calculate scores for each dimension and to derive composite measurements for the RDS subsystem (

Table 7. Comprehensive measurement results of the RDS.

Region	ES	SS	ECS	Composite Score	Rank
R1	0.778	0.826	0.747	0.852	1
R2	0.458	0.381	0.106	0.314	8
R3	0.452	0.305	0.036	0.289	10
R4	0.532	0.398	0.332	0.434	5
R5	0.545	0.509	0.421	0.501	4
R6	0.797	0.490	0.345	0.583	3
R7	0.257	0.570	0.379	0.323	7
R8	0.465	0.947	0.934	0.723	2
R9	0.102	0.338	0.539	0.302	9
R10	0.238	0.549	0.510	0.370	6

). The results reveal a distinct gradient pattern described as “high-level regions with balanced development and low-level regions with pronounced deficiencies”.

High-performing regions include R1 (0.852), R8 (0.728), and R6 (0.597), all exhibiting balanced and synergistic performance across the three dimensions. Radar charts for these regions are presented in Figure 3a. As the core water-source area, R1 demonstrates strong integration between ECS (score: 0.747) and ES (score: 0.778), while also maintaining high performance in SS. R8 benefits from a well-developed cultural and tourism industry, achieving the highest scores in both SS and ECS, reflecting a dual-advantage development model. R6, identified as an industrially advanced county, displays strong economic performance and high employment absorption capacity, resulting in the highest ES score among all regions.

Low-level regions include R3 (0.285), R9 (0.321), and R2 (0.323), each constrained by significant deficiencies in one or more key dimensions, as illustrated in Figure 3b. R3, a port-industrial hub dominated by heavy industry, performs poorly in SS (0.305) and extremely low in ECS (0.036). R9 exhibits weak economic development, with ES1, ES3, and ES4 identified as major limiting indicators. R2, characterized by a petrochemical-led economy, demonstrates a narrow industrial base and heightened ecological risk, resulting in low scores in both SS and ECS.

Medium-level regions—R5 (0.519), R4 (0.451), R10 (0.392), and R7 (0.349)—display mixed performance, with distinct strengths and weaknesses, as shown in Figure 3c. R5 has a well-developed commercial and service sector but limited employment alignment for resettlers, reflected in a low ES4 score. R10 performs relatively well in ECS, but its small industrial base constrains job creation, leading to a weak ES dimension.

Dimensional analysis confirms consistency between RDS rankings and the actual economic, social, and ecological profiles of the regions. For example, in the ES dimension, regions where tertiary industries contribute more than 50% of GDP (e.g., R4 and R5) exhibit strong employment potential but experience mismatches in training alignment. In the ECS dimension, water-source protection zones (e.g., R1) face restrictions on agricultural output due to ecological regulations, whereas non-protected industrial areas (e.g., R3) are constrained by pollution pressures. In the SS dimension, urban cores (e.g., R4 and R5) demonstrate high accessibility to public services, while peripheral regions (e.g., R9) require substantial infrastructure improvements.

This analytical framework enables precise identification of regions with robust support capacity (e.g., R1 and R8) and those with evident deficiencies (e.g., R3 and R9), providing actionable insights for improving system coordination. For instance, R3 requires targeted measures in vocational training and ecological restoration, whereas R9 needs strategies to convert ecological assets into economic advantages. Consequently, region-specific and complementary interventions can be formulated to strengthen overall system-level coordination.

3.2.3. Internal Coupling Coordination Assessment for the Regional Development Support System

The internal coordination degree of the RDS subsystem represents the extent to which its internal indicators evolve in a balanced and mutually reinforcing way. Using the DCCM (Equations (7)–(9)), the coupling coordination degree D was calculated among the three support dimensions, namely, ES, SS, and ECS, within the RDS system for the 10 regions. Coupling coordination grades follow the classification standards outlined in Table 4. Results are presented in

Table 8. Calculation results of internal coupling coordination for the RDS.

Region	Coupling Degree (C)	Comprehensive Development Index (T)	Coupling Coordination Degree (D)	Coordination Grade	Coupling Coordination Status
R1	0.996	0.851	0.921	10	High-quality coordination
R2	0.734	0.241	0.421	5	Near disordered
R3	0.212	0.175	0.192	2	Severely disordered
R4	0.857	0.367	0.561	6	Barely coordinated
R5	0.962	0.462	0.667	7	Primary coordination
R6	0.856	0.543	0.682	7	Primary coordination
R7	0.968	0.342	0.576	6	Barely coordinated
R8	0.959	0.834	0.894	9	Good coordination
R9	0.332	0.21	0.264	3	Moderately disordered
R10	0.928	0.371	0.586	6	Barely coordinated

and visualized in Figure 4.

The results reveal a distinct gradient of differentiation across regions. Each region’s coordination status closely corresponds to its industrial structure, policy orientation, and resource endowment—factors that collectively align with observed regional development patterns. Based on coordination levels, the regions are categorized into three groups: high coordination, medium coordination, and low coordination.

(1) High coordination areas include R1 (D = 0.921) and R8 (D = 0.894). These regions demonstrate well-balanced coordination across all three dimensions. Their development model can be characterized as a closed-loop synergy: ecological protection as the foundation, policy-driven support as the guarantee, industrial fit as the linkage, and livelihood security as the ultimate objective.

As the core resettlement area of the QC Reservoir, R1 demonstrates strong alignment between characteristic industries and its ecological base. In the social dimension, high scores in SS3 and SS2 indicate that both ecological and economic benefits are effectively transmitted to resettlers. Its leading ecological tourism industry successfully transforms natural resources such as water and cultivated land into economic value, exemplifying a typical “ecology–economy” synergy consistent with the SLF.

R8 leverages a mature eco-cultural tourism sector. Its high ES4 score, combined with leading performance in SS3 and SS4, reflects effective coordination among ecological assets, social service provision, and employment generation.

(2) Medium coordination areas include R6 (0.682), R5 (0.667), R10 (0.586), R7 (0.576), and R4 (0.561). These regions display partial coordination, wherein one dimension performs strongly while others show relative deficiencies.

R6 and R5 are primarily economy-driven regions that exhibit weaknesses in ecological coordination. In contrast, R10, R7, and R4 display relatively balanced development across the three dimensions but suffer from structural constraints such as limited industrial diversity. These regions require targeted improvements to address mismatches—particularly between economic and ecological systems, and between economic and social systems.

(3) Low coordination areas include R2 (0.421), R9 (0.264), and R3 (0.192). These regions are characterized by pronounced unidimensional weaknesses that hinder overall system interaction. R2’s petrochemical industry contributes to economic growth but imposes high ecological costs and weakens social linkages. R9, despite abundant ecological resources, faces difficulties in economic transformation and lacks adequate basic social support. R3 relies heavily on a single heavy-industry structure, with underdeveloped ecological and social support systems, resulting in a pronounced imbalance among the three dimensions. These low-coordination regions exemplify the challenges faced by economically unbalanced or geographically disadvantaged areas (e.g., mountainous regions) and call for precise, issue-oriented policy interventions.

(4) Spatial pattern of RDS coupling coordination indicates that high-coordination regions are primarily concentrated in ecological–economic transition zones (e.g., R1 and R8), whereas low-coordination regions are generally located in traditional industrial or rural agricultural areas (e.g., R3 and R9). This reflects the pattern: “ecological-priority areas coordinate strongly, traditional-industry areas coordinate weakly”. The driving mechanism underlying this pattern lies in the dynamic equilibrium among the three dimensions. In particular, ecological service supply (ECS2) and industrial employment adaptation (ES4) emerge as the key determinants of coordination levels, while SS performs a bridging function, connecting ecological and economic development with resettlers’ well-being.

3.3. Resettlers’ Livelihood Development System Analysis

3.3.1. Indicator Weight Determination for the Resettlers’ Livelihood Development System

Using field survey data from 289 sample households and expert evaluations from five domain specialists, the AHP–EW method was employed to calculate indicator weights for the RLD subsystem (

Table 9. Calculation results of indicator weight for the RLD.

Dimension	AHP Weight	Entropy Weight	Combined Weight	Indicator	AHP Weight	Entropy Weight	Combined Weight
Natural Capital (NC)	0.049	0.407	0.192	NC1	0.164	0.268	0.205
				NC2	0.297	0.521	0.387
				NC3	0.539	0.211	0.408
Financial Capital (FC)	0.462	0.186	0.351	FC1	0.557	0.153	0.395
				FC2	0.320	0.402	0.353
				FC3	0.123	0.445	0.252
Human Capital (HC)	0.275	0.282	0.278	HC1	0.064	0.092	0.075
				HC2	0.109	0.166	0.132
				HC3	0.215	0.128	0.180
				HC4	0.239	0.392	0.300
				HC5	0.373	0.222	0.312
Social Capital (SC)	0.089	0.059	0.077	SC1	0.164	0.411	0.263
				SC2	0.539	0.355	0.465
				SC3	0.297	0.234	0.272
Physical Capital (PC)	0.126	0.066	0.102	PC1	0.539	0.317	0.450
				PC2	0.164	0.166	0.165
				PC3	0.297	0.517	0.385

). All judgment matrices satisfied the consistency test criteria. The subjective weight distribution reflects resettlers’ primary concerns—such as income stability, employment security, and skill alignment, while the entropy-derived objective weights capture regional disparities in livelihood conditions. The resulting structure is consistent with the core logic of the SLF: FC serves as the foundation, HC functions as the driving force, NC provides supplementary support, and PC together with SC act as fundamental guarantees for sustainable livelihood development.

FC serves as the cornerstone of resettlers’ livelihood systems, with a combined weight of 0.351. Among its components, per capita annual income (FC1) holds the highest weight and demonstrates strong policy relevance. The income diversification index (FC2), weighted at 0.353, underscores the essential contribution of diversified income sources to enhancing livelihood resilience. The number of government subsidy types (FC3), with a weight of 0.271, reflects the importance of policy and financial support mechanisms in sustaining resettlers’ livelihoods. The relatively narrow income disparity among resettlers across regions indicates both Zhejiang’s advanced stage of economic development and its well-established social security network. This finding suggests that resettlement policies have successfully secured basic income levels for vulnerable groups, thereby mitigating livelihood risks associated with

HC functions as the primary driver of livelihood transformation, with a combined weight of 0.278. Within this category, the employment–industry matching degree (HC5) carries the highest weight (0.312), directly representing the degree of alignment between household labor skills and regional industrial demands. Survey respondents frequently emphasized that “job fit influences income more than formal education,” highlighting the significance of the proportion of labor with employment skills (HC4). The labor employment rate (HC3) and household labor ratio (HC2) reflect overall labor participation levels and household workforce structure. The average household education level (HC1) holds the lowest weight, as the widespread attainment of basic education in Zhejiang reduces variability. Moreover, education alone does not guarantee improved livelihoods unless it is effectively converted into employable skills aligned with local labor market requirements.

NC, with a combined weight of 0.192, emphasizes alignment with regionally characteristic agriculture. Cultivated land resources in resettlement areas are generally limited, and some households have opted for pension-insurance–based resettlement instead of land redistribution, thereby reducing dependence on cropland. The degree of match between cropping structure and characteristic agriculture (NC3) holds the highest weight (0.408), reflecting Zhejiang’s capacity to sustain specialized agricultural production despite its constrained land availability. The cultivated land quality grade (NC2) carries a greater weight than per capita cultivated land area (NC1), underscoring that land quality contributes more significantly than quantity to livelihood sustainability and agricultural productivity.

PC and SC have relatively lower combined weights—0.102 and 0.077, respectively—indicating their roles as foundational supports rather than principal drivers of livelihood development. During the resettlement process, local governments implemented multiple initiatives to facilitate social integration. Survey respondents generally expressed satisfaction with interpersonal relationships and community participation; however, they consistently emphasized that economic stability remains central to their livelihood strategies. Within these categories, per capita housing area (PC1) and community participation (SC2) hold the highest indicator weights (0.450 and 0.465, respectively), functioning as baseline guarantees that underpin broader livelihood development and social adaptation.

3.3.2. Livelihood Capital Evaluation

Individual indicator scores were computed for all 289 surveyed households based on the field data (Table 3). The TOPSIS method was subsequently applied, using these indicator scores and the previously determined weights, to calculate both dimension-level scores and overall composite measurements for the RLD subsystem (

Table 10. Comprehensive measurement results of the RLD.

Region	NC	FC	HC	SC	PC	Composite Score	Rank
R1	0.337	0.286	0.539	0.679	0.336	0.496	9
R2	0.730	0.320	0.554	0.539	0.347	0.605	3
R3	0.763	0.374	0.621	0.535	0.401	0.658	1
R4	0.781	0.403	0.540	0.554	0.339	0.648	2
R5	0.695	0.311	0.545	0.462	0.372	0.581	4
R6	0.645	0.283	0.592	0.490	0.374	0.579	5
R7	0.699	0.267	0.546	0.410	0.373	0.567	6
R8	0.398	0.346	0.547	0.686	0.389	0.540	7
R9	0.281	0.379	0.462	0.587	0.371	0.474	10
R10	0.323	0.208	0.703	0.513	0.355	0.514	8

). The results reveal a distinct gradient pattern: regions with higher scores exhibit well-balanced livelihood capital structures, whereas low-scoring areas display pronounced deficiencies in one or more forms of capital accumulation.

High-score regions include R3 (0.658), R4 (0.648), and R2 (0.605). These regions exhibit well-balanced livelihood capital portfolios, each displaying notable strengths in one or more dimensions (Figure 5a). R3 resettlers benefit from robust HC, supported by high-skilled employment opportunities within the port industry. Correspondingly, HC and PC scores are the highest, reflecting the advantages of an industrialized urban environment that features stable employment and well-developed infrastructure. In R4, the modern service economy contributes to income stability, with resettlers achieving strong performance across NC, FC, and SC—demonstrating the advantages of diversified urban settings. R2’s petrochemical industry provides steady employment and consistent support across all capital types, with no significant deficiencies observed in its livelihood structure.

Low-score regions include R9 (0.474), R1 (0.496), and R10 (0.514), all of which exhibit multidimensional constraints (Figure 5b). R9 records the lowest scores in HC and NC due to slow agricultural transformation, limited industrial capacity, weak employment opportunities, and poor skill alignment. In R1, resettlers face restrictions imposed by water-source protection regulations and are confined to cultivating “suitable” crops such as rice and legumes, thereby reducing returns on NC. Many resettlers retained their previous employment modes because of intra-county relocation, resulting in path dependency that suppresses FC scores. R10 reveals a mismatch between HC and FC—low income levels and limited income diversification stem from the volatility of local fisheries and the absence of robust alternative industries.

Medium-score regions, including R5 (0.581), R6 (0.579), R7 (0.567), and R8 (0.540), show moderate performance without pronounced strengths or critical weaknesses (Figure 5c). R5’s commercial-urban context yields balanced yet undifferentiated livelihood capitals. R6 resettlers exhibit sufficient industrial skills but limited income diversification. R7 possesses strong agricultural potential yet weak community integration. R8 demonstrates strong social support, but its agricultural adaptation capacity for resettlers remains underdeveloped.

Analysis by capital dimension confirms consistency with observed regional realities. In FC, urban centers (e.g., R4) attain higher scores, whereas agricultural regions (e.g., R7) face greater income risks due to industrial concentration. In HC, industrialized areas (e.g., R3) display superior skill alignment, while rural zones (e.g., R9) experience lower employment rates. In NC, water-source protection regions (e.g., R1) are notably constrained by environmental regulations. For PC and SC, urban localities (e.g., R4) benefit from improved housing conditions and greater social participation, whereas peripheral regions (e.g., R10) face limitations in basic support infrastructure.

Overall, these results highlight significant spatial heterogeneity in resettlers’ livelihood capitals and provide an empirical foundation for differentiated policy interventions. For instance, R9 would benefit from enhanced vocational training to strengthen skill adaptability, while R10 requires expanded non-farm employment opportunities to stabilize household income and mitigate livelihood vulnerability.

3.3.3. Internal Coupling Coordination Assessment for the Resettlers’ Livelihood Development System

The internal coordination degree of the RLD subsystem reflects the extent to which its livelihood-related indicators evolve coherently and structurally reinforce one another. Using the DCCM, the coupling and coordination relationships among the five livelihood capitals within the RLD subsystem were computed to evaluate each region’s efficiency in resource allocation and overall system order. The classification of coupling coordination grades follows the criteria outlined in Table 4, while the corresponding results are summarized in

Table 11. Calculation results of internal coupling coordination for the RLD.

Region	Coupling Degree (C)	Comprehensive Development Index (T)	Coupling Coordination Degree (D)	Coordination Grade	Coupling Coordination Status
R1	0.472	0.364	0.415	5	Near disordered
R2	0.879	0.498	0.662	7	Primary coordination
R3	0.962	0.780	0.866	9	Good coordination
R4	0.679	0.577	0.626	7	Primary coordination
R5	0.897	0.489	0.662	7	Primary coordination
R6	0.953	0.505	0.694	7	Primary coordination
R7	0.531	0.413	0.468	5	Near disordered
R8	0.879	0.620	0.738	8	Intermediate coordination
R9	0.301	0.413	0.353	4	Mildly disordered
R10	0.451	0.353	0.399	4	Mildly disordered

and illustrated in Figure 6.

The results indicate substantial variation in coupling coordination levels across regions, reflecting how effectively resettlers’ livelihood capital structures align with regional industrial demands. Overall, the observed patterns correspond closely with the distinctive livelihood characteristics of local resettlers. For analytical clarity, the regions are classified into three coordination categories—high, medium, and low—based on their coupling coordination degrees.

(1) High-coordination area: Only R3 qualifies as a high-coordination region (D = 0.866). Its defining feature is the strong interaction and synergy among all five forms of livelihood capital. Resettlers in R3 actively utilize local economic opportunities and have successfully adapted to new production and living environments. The high score for employment–industry matching degree (HC5) drives the accumulation of FC and reinforces PC. This dynamic creates a virtuous cycle characterized by deep coupling between human and financial capital, strong physical capital, and well-developed SC. These mutually reinforcing relationships result in a highly efficient and well-integrated livelihood system.

(2) Medium-coordination areas: R8 (D = 0.738), R6 (D = 0.694), R2 (D = 0.662), R5 (D = 0.662), and R4 (D = 0.626) fall within the medium-coordination category. These regions maintain relatively balanced livelihood capital structures that correspond to local development contexts, though none exhibit dominance in a single capital dimension.

In R8, R5, and R4, SC plays a pivotal role. Their tertiary-sector–based economies foster balanced interactions among the various capitals, producing moderate synergies between NC and FC. However, limited agricultural adaptability constrains further capital enhancement. R6 and R2 demonstrate strong integration between regional industrial development and livelihood capital formation. In both regions, HC and FC align effectively with local employment structures, though income diversification remains modest. These areas would benefit from policies that enhance cross-capital complementarities and mitigate reliance on single-dimensional growth.

(3) Low-coordination areas: R7 (D = 0.468), R1 (D = 0.415), R10 (D = 0.399), and R9 (D = 0.353) are identified as low-coordination regions. These areas experience significant deficits in one or more livelihood capitals, which hinder capital interaction and impede the formation of a self-reinforcing livelihood system.

In R7, ongoing industrial upgrading in mold manufacturing has resulted in skill mismatches and shortages in financial ad social capital. In R1, watershed protection policies constrain agricultural income, while path dependency in employment limits the compensatory roles of social and human capital in offsetting deficiencies in FC. R10, influenced by volatility in the fisheries sector, experiences unstable income and difficulty converting human capital into financial returns. R9, a mountainous agricultural county, suffers from limited arable land and underdeveloped human capital, leading to concurrent weaknesses in both natural and human capital dimensions. These regions require tailored policy interventions—such as programs for mitigating skill mismatches, promoting non-farm employment, and implementing ecological compensation mechanisms—to address localized vulnerabilities.

(4) Spatial pattern and mechanism: The gradient pattern of coupling coordination within the RLD system exhibits a clear spatial hierarchy: urban and industrially advanced regions > traditional agricultural regions > mountainous regions. The primary mechanism driving this differentiation lies in the dynamic balancing capacity among the five livelihood capitals. HC and FC—particularly HC5 and FC2—are the dominant determinants of coupling strength, while NC and SC serve complementary roles. Nevertheless, weak industrial alignment or stringent ecological constraints can create structural bottlenecks, disrupting overall system integration and limiting sustainable livelihood transformation.

3.4. Analysis of the Synergistic Development Index

Building on the development levels of RDS and RLD discussed in Section 3.2, along with their internal coordination analyzed in Section 3.3, this section employs the DCCM to calculate the SDI. The SDI quantitatively measures the degree of synergy between resettlers’ livelihood development and regional development. A higher SDI value indicates a more balanced and mutually reinforcing relationship between the two subsystems, reflecting a healthier bidirectional dynamic of “livelihood improvement supported by regional development and regional progress driven by livelihood enhancement”. Classification grades are consistent with those presented in Table 4, while the calculated results are summarized in

Table 12. Calculation results of the Synergistic Development Index.

Region	Coupling Degree (C)	Comprehensive Development Index (T)	Synergistic Development Index	Synergistic Grade	Synergistic Development Status
R1	0.635	0.559	0.596	6	Barely coordinated
R2	0.511	0.381	0.441	5	Near disordered
R3	0.199	0.500	0.315	4	Mildly disordered
R4	0.827	0.600	0.704	8	Intermediate coordination
R5	0.978	0.479	0.685	7	Primary coordination
R6	0.999	0.545	0.738	8	Intermediate coordination
R7	0.651	0.287	0.432	5	Near disordered
R8	0.934	0.563	0.725	8	Intermediate coordination
R9	0.847	0.021	0.134	2	Severely disordered
R10	0.981	0.187	0.428	5	Near disordered

and visualized in Figure 7.

The SDI reveals a distinct gradient across the ten regions, illustrating the varying degrees of alignment between resettlers’ livelihood needs and the developmental capacities of their host regions. Fundamentally, the SDI captures the efficiency of bidirectional interactions—that is, the extent to which the regional support system can effectively provide resources to the livelihood system, and whether the livelihood system can, in turn, generate feedback that strengthens regional development. Considering each region’s industrial structure, resource endowment, and resettler characteristics, the regions are categorized into four synergy types: Benign Synergy, Region-driven, Unipolar Constraint, and Dual-weak Lock-in. The distribution of the SDI is shown in Figure 8.

(1) Benign Synergy type includes R6 (SDI = 0.738), R8 (SDI = 0.725), and R4 (SDI = 0.704). These regions demonstrate strong interaction between the two subsystems (high C) and relatively high overall development levels (high T). The defining feature of this type is reciprocal empowerment between the regional support system and resettlers’ livelihood system, forming a positive feedback loop in which regional development generates livelihood opportunities, and livelihood advancement, in turn, reinforces local economic growth. Coordination mechanisms in these regions are both mature and effective, promoting sustainable, high-level development.

R6 records the highest SDI, primarily due to elevated scores in ES4 and HC5, representing strong employment absorption by characteristic industries and a high degree of labor–industry matching. These two key indicators establish a virtuous cycle of industrial drive → skill supply → income growth → consumption feedback, reinforcing long-term system stability. R8 successfully transforms ecological assets into economic momentum through its mature eco-cultural tourism sector. High employment absorption among resettlers strengthens FC accumulation, achieving a balanced outcome that combines ecological conservation with livelihood improvement—a clear demonstration of “green synergy” between environment and development.

(2) Region-driven type includes R5 (SDI = 0.685) and R1 (SDI = 0.596). These regions feature a strong RDS system that drives the RLD subsystem forward. However, internal weaknesses within the livelihood system constrain overall synergy. To achieve higher coordination, these regions must enhance the transmission efficiency through which regional advantages translate into direct livelihood improvements for resettlers.

R5 is a traditional commercial hub, and benefits from well-developed regional infrastructure and service systems. Nevertheless, persistent skill mismatches among resettlers weaken employment alignment, thereby reducing the effectiveness of converting economic advantages into tangible livelihood outcomes.

R1 serving as the core watershed protection area, attains the highest RDS composite score (0.852) due to strong policy support in ecological services and public infrastructure. However, RLD performance remains limited. Strict environmental regulations—such as prohibitions on polluting industries within ecological redline zones—curtail productive activities and disrupt the balance between NC and local economic demand. This results in a “high support–low livelihood” configuration, where resettlers’ developmental potential and self-driven initiative are insufficiently activated.

(3) Unipolar Constraint type includes R2 (SDI = 0.441), R7 (SDI = 0.432), and R10 (SDI = 0.428). In these regions, substantial deficiencies in one subsystem suppress overall coordination, necessitating targeted interventions to alleviate single-dimension bottlenecks.

In R2, the petrochemical industry drives regional economic development but simultaneously exerts severe ecological pressure, forming an “ecological constraint synergy” pattern where industrial growth conflicts with environmental sustainability. R7 is currently undergoing industrial upgrading toward research and high-tech sectors; however, the resettlers’ livelihood system lacks the requisite HC to adapt. Skill mismatches between existing labor capacity and emerging industrial demands significantly hinder coordinated progress.

(4) Dual-weak Lock-in type includes R3 (SDI = 0.315) and R9 (SDI = 0.134). These regions experience concurrent weaknesses in both the RDS and RLD subsystems, resulting in minimal interaction and entrapment in a low-level development loop. Achieving coordination in such contexts requires comprehensive structural transformation.

R3’s dependence on heavy industry has led to ecological degradation and underdeveloped social infrastructure. Although the RLD subsystem performs moderately, resettlers’ labor participation and consumption contribute little to regional development, producing a fragmented structure characterized by an “independent support system and self-contained livelihood system”. This disconnection prevents genuine synergy between the two subsystems. R9, as a mountainous agricultural county, lacks both economic and social infrastructure, and its RLD score is the lowest among all regions, resulting in a pronounced dual-weak lock-in condition.

(5) From the typological analysis, Benign Synergy regions should focus on consolidating their multidimensional advantages to sustain high-level coordination. Region-driven areas must strengthen mechanisms that effectively translate regional development potential into tangible livelihood improvements. Unipolar Constraint regions require targeted policy interventions to alleviate subsystem bottlenecks and restore balance. Dual-weak Lock-in regions, by contrast, demand fundamental structural reform and the simultaneous empowerment of both subsystems to break out of low-level equilibrium traps.

The core determinant of SDI performance lies in the strength of supply–demand alignment between the two systems. Specifically, HC5 (employment–industry matching degree) and ES4 (employment absorption by characteristic industries) function as the principal drivers of synergy, fostering mutual reinforcement between labor capacity and industrial demand. ECS2 (ecological service supply) and FC2 (income diversity index) act as modulators that can either constrain or amplify synergy, depending on local environmental and economic contexts. Collectively, these findings offer robust empirical evidence to guide differentiated and precision-targeted policy strategies aimed at improving the sustainability and effectiveness of resettlement outcomes.

3.5. Spatial Correlation Analysis of the SDI

To examine the spatial association of the SDI across the ten regions, Moran’s I was calculated to evaluate both global and local spatial correlations (using Equations (10) and (11)). This analysis identifies patterns of geographic clustering or dispersion in county-level coordination performance. The results reveal a spatial configuration characterized as “weakly dispersed distribution with heterogeneous local associations”. While global spatial dependence is not statistically significant, localized clustering and dispersion patterns are evident. These findings indicate that variations in industrial structure and resource flows play a critical role in shaping the spatial differentiation of coordination levels among regions.

3.5.1. Global Spatial Correlation Analysis

Global Moran’s I measures the overall spatial dependence of a regional variable. For the SDI, the global Moran’s I is $I_{Global}=-0.089$, with a theoretical expectation E(I) = −0.111, Z = 0.135, and p = 0.446. These values indicate that the overall SDI distribution across the 10 regions follows a “weakly dispersed” pattern, which is not statistically significant.

This weak spatial dispersion primarily results from the pronounced heterogeneity in industrial structures among regions. Central urban areas (e.g., R4 and R5) are dominated by service-oriented industries, coastal industrial zones (e.g., R3 and R2) display heavy industrial activity, and ecological regions (e.g., R1 and R10) emphasize agriculture and eco-tourism. Such industrial differentiation disrupts the conventional expectation that geographically adjacent regions should display similar levels of coordination. In addition, the randomized resettlement process—where resettlers are allocated to relocation sites through a lottery mechanism—introduces heterogeneity in skill composition and livelihood capacity, thereby weakening patterns of global spatial clustering.

3.5.2. Local Spatial Correlation Analysis

Local Moran’s I identifies spatial association patterns between each region and its immediate neighbors. The sign of $I_{Local}$ indicates clustering or dispersion, and significance is determined by the Z-score and p-value. Results are presented in

Table 13. Local spatial correlation results for the Synergistic Development Index.

Location	Local Moran’s I	Z-Score	p-Value
R1	−0.166	−0.744	0.114
R2	−0.114	−0.572	0.142
R3	−0.265	−0.483	0.157
R4	0.561	2.116	0.009
R5	0.156	1.209	0.057
R6	−0.023	0.389	0.174
R7	−0.365	−2.127	0.008
R8	−0.116	−0.162	0.218
R9	−0.593	−0.237	0.203
R10	0.123	0.889	0.093

. Based on significance and direction, local spatial associations are categorized into Significant-association, Marginal-association, and Weak-association types.

(1) Significant-association type (p < 0.05) includes R4 and R7, which represent opposite spatial patterns—clustering and dispersion—and serve as illustrative typological examples.

R4 shows significant clustering ($I_{Local}=0.561>0$). It forms a “high–high” cluster with neighboring R5 and R8. This is driven by strong industrial complementarity: R4’s modern service economy complements R5’s commercial sector and R8’s eco-tourism, while interconnected public services enhance regional cohesion.

R7 exhibits significant dispersion ($I_{Local}=-0.365<0$), forming a “low–high” contrast with higher-performing neighbors like R6 and R4. R7 is in industrial transition (from mold manufacturing to tech-intensive sectors), but resettler skills and social adaptability lag, creating a coordination gap.

(2) Marginal-association type (p < 0.1) includes R5 and R10, where spatial clustering is evident but not statistically strong due to weaker underlying foundations.

R5 marginally clusters with R4 and R8 in a “high–high” configuration, enabled by commercial complementarities and frequent regional interaction. However, resettler–industry mismatch weakens overall livelihood linkages.

R10 forms a marginal “low–low” cluster with adjacent R9. Both regions lack industrial depth and have limited resource circulation, reinforcing their low-coordination status.

(3) Weak-association type (p ≥ 0.1) includes the remaining six regions, namely, R6, R8, R1, R2, R3, and R9, none of which show statistically significant spatial correlation. This reflects industrial heterogeneity and isolated development dynamics.

The root causes of the weak spatial association observed among these regions stem from several structural factors: industrial mismatches (e.g., R3’s heavy-industrial orientation contrasted with neighboring service-based economies), ecological constraints (e.g., R1’s stringent water-source protection policies that restrict cross-regional coordination), and geographic or administrative isolation (e.g., R9’s mountainous terrain limiting interregional connectivity). In such contexts, the interaction between livelihood and support systems tends to be self-contained or structurally fragmented, thereby reducing spatial regularity and impeding the diffusion of synergistic development effects across adjacent regions.

(4) Implications: The spatial differentiation of SDI is not solely determined by geographic proximity but emerges from a complex interaction among industrial specialization, policy alignment, and livelihood–system compatibility. Regions with homogeneous or complementary industrial structures (e.g., R4–R5–R8) tend to form spatial clusters, while heterogeneous regions (e.g., R7 versus R6) exhibit spatial dispersion due to mismatches in industrial orientation or disparities in resettler capacity. Public service interconnectivity strengthens spatial cohesion and regional “stickiness” (e.g., R4 and its neighboring areas), whereas administrative fragmentation or geophysical isolation (e.g., R10) weakens spatial linkages.

To promote spatially balanced development, it is essential to dismantle administrative barriers, enhance industrial integration, improve cross-regional resource circulation, and prevent the emergence of “islanded” or fragmented development zones. These findings offer a solid empirical basis for formulating cross-regional coordination and support mechanisms in future reservoir resettlement planning and implementation.

4. Discussion

4.1. Core Driving Mechanisms Within Subsystems

(1)

Internal driving mechanism of the RDS System

Composite RDS scores differ markedly across regions, driven primarily by variations in the ES dimension (combined weight: 0.503). Among its indicators, the employment absorption rate of resettlers by characteristic industries (ES4) plays a decisive role, carrying the highest weight (0.437) and reflecting the labor-pulling capacity of regional industrial structures. This observation aligns with Yu et al. [40], who concluded that “global knowledge drives structural differentiation within clusters”. The ECS dimension, with a combined weight of 0.271, highlights the distinctive environmental context of reservoir resettlement areas. Contrasting ecological (e.g., R8) and heavy-industry (e.g., R3) orientations result in pronounced developmental disparities—consistent with Wen et al. [20], who emphasized that “basin coordination must prioritize ecosystem protection”.

The SS dimension (weight: 0.226) centers on training coverage (SS3), which acts as a critical bridge between economic development and livelihood adaptation. High-performing regions (e.g., R1 and R8) exhibit strong internal coordination across economic, ecological, and social dimensions, forming virtuous development loops. In contrast, low-performing regions (e.g., R3 and R9) demonstrate unidimensional weaknesses, where ecological restrictions or insufficient social services constrain the achievement of coordinated, sustainable development.

(2)

Internal driving mechanism of the RLD System

RLD scores exhibit substantial regional variation, primarily driven by FC (weight: 0.351) and reinforced by HC (weight: 0.278). Among the indicators, employment–industry matching (HC5) and income diversity (FC2) emerge as the dominant influencing factors. High-performing regions (e.g., R3) display well-balanced livelihood capital portfolios, characterized by strong integration between HC and FC, while low-performing regions (e.g., R9) show concurrent deficits in NC and HC, leading to heightened livelihood vulnerability. This finding supports Zhang et al.’s [41] assertion that “insufficient non-farm employment exacerbates household poverty vulnerability”.

The influence of NC manifests primarily through the alignment of cropping structures with specialty agriculture (NC3) rather than through the absolute quantity of cultivated land (NC1). This contrasts with the findings of Khanal et al. [8] in their hydropower resettlement study in Nepal, which identified PC—including farmland and housing—as the core determinant of livelihood outcomes. The divergence reflects differences in regional development trajectories: Zhejiang’s ongoing industrial modernization (e.g., R6’s intelligent manufacturing sector) has elevated the marginal utility of HC, whereas Nepal’s predominantly agrarian economy places greater emphasis on the security of physical resources.

4.2. Coupling and Interaction Mechanisms Between the Two Systems

The interaction between resettlers’ livelihoods and regional development operates as a fundamentally bidirectional dynamic—in which regional systems provide resources, infrastructure, and institutional support, while livelihood systems generate feedback through labor participation, consumption demand, and contributions to social cohesion. Synergistic development depends not on the absolute performance of either subsystem but on their structural compatibility and co-evolution. Based on the SDI results (Section 3.4) and indicator weight analysis (Section 3.2.1 and Section 3.3.1), five distinct synergistic development pathways are identified, each representing a different mechanism through which regional support and livelihood strategies reinforce one another.

4.2.1. Specific Synergistic Development Paths

(1)

Industry–Livelihood Matching Path (ES4→HC5→FC2)

Regions with strong industrial employment absorption capacity (ES4, weight = 0.437; core RDS indicator) provide diverse employment opportunities. When resettlers possess or acquire adaptive skills (HC5, weight = 0.312; core RLD indicator), the alignment between labor supply and industrial demand improves, while financial accessibility (FC2, weight = 0.353) supports costs related to training, commuting, or entrepreneurial transitions. This forms a reinforcing chain—industrial absorption → skill matching → financial empowerment → livelihood upgrading—that substantially enhances synergy. This mechanism explains why R6 exhibits the highest SDI (0.738; Benign Synergy type), as its smart-manufacturing sector and resettlers’ skill adaptation operate in a mutually reinforcing cycle.

(2)

Public Service–Social Integration Path (PS3→SC4→HC4/HC5)

Improved public service provision (PS3) reduces adaptation costs and promotes social integration (SC4). Enhanced integration subsequently facilitates human capital accumulation (HC4, HC5) through skill upgrading and active labor participation, forming a positive feedback loop between livelihood improvement and regional development. This mechanism is characteristic of urban and peri-urban areas such as R4 (SDI = 0.704) and R5 (SDI = 0.685), where high public service accessibility offsets moderate economic disparities and fosters strong synergy.

(3)

Ecology–Agriculture Compatibility Path (ES2→PC2→FC3)

In regions with ecological advantages but weaker economies (e.g., R1, SDI = 0.596; R8, SDI = 0.725), ecological suitability (ES2) and agricultural production capacity (PC2) align well with the livelihood strategies of agriculture-dependent resettlers. Agricultural subsidies and ecological compensation (FC3) stabilize income, mitigating the limitations imposed by reduced industrial opportunities.

(4)

Transportation–Employment Accessibility Path (ES1→HC3→SC2)

Enhanced transportation infrastructure (ES1) broadens the employment radius for resettlers (HC3) and facilitates the reconstruction of social networks (SC2). This promotes livelihood diversification and strengthens cross-system feedback, enabling regions with limited industrial capacity (e.g., R10, SDI = 0.428) to leverage transportation connectivity to offset industrial constraints.

(5)

Institution–Behavior Adaptation Path (PS4→SC2→HC5)

Effective implementation of resettlement policies (PS4) strengthens institutional trust (SC2), motivating resettlers to participate in training and entrepreneurial activities (HC5). This pathway is particularly vital in regions where economic capacity alone does not ensure adaptation, such as R9 (SDI = 0.134; dual-weak lock-in type) and R7 (SDI = 0.432; Unipolar Constraint type), as it activates endogenous motivation and enables resettlers to align more effectively with regional development trajectories.

4.2.2. Core Characteristics of Synergistic Mechanisms

These pathways reveal two key principles underlying synergistic development:

(1)

Regions may achieve similar SDI levels through different mechanisms. For example, R4 follows a public service–social integration pathway, while R8 relies on an ecology–agriculture compatibility pathway; yet both achieve SDI values above 0.7 and fall within the “Intermediate Coordination” category despite fundamentally different development foundations.

(2)

Similar regional conditions can yield divergent synergy levels depending on resettlers’ adaptive capacity. For instance, although both R2 and R3 are industrial regions, their SDI levels differ markedly—0.441 and 0.315, respectively—because the petrochemical sector in R2 offers stronger skill-matching opportunities than the heavy-industry base in R3.

Across all pathways, the dominant drivers are ES4 and HC5, which directly link regional industrial structures with resettlers’ labor supply. Modulating influences include FC2 and ECS2, which enhance or constrain synergy depending on local economic and ecological contexts. Social support indicators (PS3, SC4) and effective policy implementation (PS4) act as enabling mechanisms, ensuring that regional advantages are effectively translated into sustainable livelihood improvements.

4.3. Spatial Differentiation Patterns of Coupled Development

(1)

Gradient differentiation pattern

The spatial distribution of the coupling development index displays a weakly dispersed pattern, with a global Moran’s I value of −0.089. Unlike rural natural settlements, whose spatial patterns are primarily influenced by topography and urbanization processes [42], the SDI in resettlement regions is shaped by more than just regional economic scale or locational advantages. The observed gradient differentiation corresponds to the degree of alignment between regional industrial structures and resettler skill profiles. A clear typological gradient emerges—Benignly Coupled → Region-Driven → Unipolar-Constrained → Dual-Weak-Locked.

Regions such as R6, where local industries are closely aligned with resettlers’ skill sets, demonstrate the highest coupling levels. Central service-oriented areas such as R4 and R5 exhibit strong regional development capacity but have yet to fully convert this advantage into substantial livelihood improvements. In contrast, R1 depends heavily on policy compensation to balance ecological restrictions with livelihood needs. Dual-weak regions (e.g., R3 and R9) remain constrained by industrial monocultures and weak foundational capacities, limiting the potential for dynamic interaction between the livelihood and regional systems.

This gradient pattern corresponds with the “core–periphery” spatial differentiation identified by Yin et al. [43] in Gansu Province. However, the present study extends this understanding by revealing the internal driving mechanism: regional resource endowments translate into higher coupling levels only when effectively mediated through “industry–skill” matching. The diversity, adaptability, and inclusiveness of industrial structures—and their compatibility with resettlers’ HC—constitute the fundamental determinants of improved coupling and coordination outcomes.

(2)

Local association pattern

Results from the local Moran’s I analysis indicate that spatial association patterns are strongly influenced by industrial similarity and interregional connectivity. Regions R4, R5, and R8 constitute a significant “high–high” clustering hotspot owing to their high degree of industrial complementarity and shared public service infrastructure. In contrast, R7 displays a “low–high” outlier pattern, as its resettler skill base is poorly aligned with the labor demands of adjacent industrial regions, resulting in localized decoupling between livelihood and regional systems.

These results emphasize that geographic proximity alone does not ensure coordination. Instead, industrial structure similarity and resource flow interoperability are the principal determinants of local coupling. Similarly, in studies on reclaimed-water coupling coordination, Chen et al. [44] attributed observed spatial differentiation to differences in regional industrial structures and resource endowments. Regions with homogeneous or complementary industrial systems are more likely to form spatial clusters, whereas industrial heterogeneity tends to produce dispersion. Furthermore, interconnected public services enhance spatial cohesion and coupling intensity, while geographic isolation weakens interregional linkages and reduces the overall strength of coordination.

(3)

Spatial commonality of constraining factors

Spatially, two principal categories of constraints—skills gaps and ecological shortfalls—exhibit distinct regional clustering patterns. The primary driving indicators are the employment–industry matching degree (HC5) and the industry employment absorption rate (ES4), while ecological service supply (ECS2) and income diversity (FC2) serve as moderating variables influencing the overall coordination effect.

Skills gaps are concentrated in agriculturally dominated regions (e.g., R9) and in areas undergoing industrial upgrading (e.g., R7), where resettlers’ labor capacities fail to meet the changing skill demands of emerging industries. Ecological shortfalls, by contrast, cluster in heavy-industry regions (e.g., R3 and R2), where environmental degradation undermines the long-term sustainability of livelihood systems. The spatial overlap between these two categories of constraint—particularly in regions facing both industrial mismatch and ecological stress—intensifies developmental challenges and exacerbates disparities in coupling levels.

This spatially explicit identification of constraint clusters provides clear, actionable insights for formulating differentiated and precision-targeted regional intervention policies. Such strategies can enhance the adaptive capacity of resettlers while promoting more balanced and sustainable coordination between livelihoods and regional development.

4.4. Research Limitations and Future Directions

Despite the model’s innovation and broad applicability, several limitations remain, highlighting key directions for future research.

(1) Cross-sectional data are inherently limited in capturing dynamic co-evolutionary trends. This study is designed to diagnose the relative coordination between the two systems rather than to infer causal relationships or temporal dynamics. Because the analysis is based on cross-sectional data collected in 2023, it captures only the relatively stable post-resettlement period and does not reflect the long-term co-evolution of resettlers’ livelihoods and regional development. Consequently, the study cannot establish causal link-ages or fully account for potential bidirectional interactions between the two systems. Second, although the overall sample size (n = 289, representing 13.0% of all resettled households) is sufficient for examining regional-level patterns, the absolute number of sampled households in several small receiving regions remains limited due to the small scale of their resettler populations. This constrains the capacity to conduct independent and detailed subgroup analyses for these regions. Future research will address these limitations by conducting longitudinal follow-up surveys to develop multi-wave panel datasets and by expanding the sample to support more comprehensive analyses of smaller subgroups.

(2) The case study has limited regional generalizability. The research site—Zhejiang Province—possesses a strong economic foundation, well-coordinated inter-city linkages, and substantial policy support. These favorable conditions contrast sharply with those of underdeveloped central and western regions, where resettlers frequently encounter resource constraints and coordination challenges [23]. Future research should therefore undertake comparative analyses in low-capacity regions (e.g., Gansu and Shaanxi) to further evaluate the model’s robustness and generalizability across diverse socio-economic and geographic contexts.

(3) Certain NC indicators (e.g., NC1 and NC3) within the current indicator system implicitly assume the existence of arable land, thereby constraining the model’s applicability in non-agricultural resettlement contexts. This structural limitation reduces the model’s relevance for urban or land-scarce scenarios. Compared with the CCF, which conceptualizes the seven capitals as distinct analytical dimensions, the present model emphasizes system-level interaction and co-evolution rather than the classification of individual capitals. Future research will expand upon frameworks such as those of Huang et al. [31] and Cornelia Flora et al. [11] to integrate indicators reflecting non-farm employment, urban adaptation, and other pertinent dimensions of livelihood sustainability, thereby enhancing the model’s comprehensiveness and applicability across diverse resettlement settings.

(4) The livelihood-to-region feedback pathways warrant further investigation. This study primarily examines feedback effects from HC and NC on regional development, relying on indirect inferences regarding labor supply and consumption demand derived from income data. Future research should explicitly analyze how various livelihood capitals contribute to regional industrial dynamics, particularly through mechanisms such as local consumption, labor market participation, and entrepreneurial activity. Such analysis would deepen understanding of the bidirectional interactions between livelihood systems and regional development processes.

5. Conclusions

5.1. Research Conclusions

Using the QC Reservoir as a case study, this research integrates the micro-level SLF with macro-level regional synergistic development theory to construct a dual-system analytical model of “resettler livelihoods–regional development”. By quantifying the SDI, the model evaluates the coupling and coordination pathways between resettlers’ livelihood systems and host region development.

Quantitative analysis of the SDI reveals a distinct gradient among the ten regions, ranging from 0.134 (R9, severely disordered) to 0.738 (R6, intermediate coordination). Overall, the SDI demonstrates a weakly dispersed spatial pattern (global Moran’s I = −0.089, p = 0.446). These results align closely with the actual conditions observed at both regional and household levels within the sample, confirming the model’s empirical validity and offering a transferable analytical framework for assessing synergistic development in other reservoir resettlement contexts.

(1) The proposed model addresses the shortcomings of single-theory approaches by integrating the five livelihood capitals of the SLF with the regional economic–social–ecological system, thereby establishing a mutually reinforcing RDS–RLD framework. The eight dimensions and 25 selected indicators effectively capture both the core livelihood development needs of resettlers and the macro-level coordination objectives of host regions.

Within the RDS, three key indicators reflect the fundamental challenges of resettlement: the employment absorption rate by resettler industries (ES4, weight = 0.437) under the ES dimension represents industrial adaptation; the coverage of specialized resettler training (SS3, weight = 0.381) under the SS dimension corresponds to skills enhancement; and the ecological service supply (ECS2, weight = 0.481) under the ECS dimension embodies ecological constraints.

In the RLD, the employment–industry matching degree (HC5, weight = 0.312) and the degree of match between cropping structure and characteristic agriculture (NC3, weight = 0.408) demonstrate close alignment with regional support indicators, collectively reflecting the compatibility and feedback linkages between the livelihood and regional subsystems. This dual-system integration bridges the micro–macro divide, overcoming the explanatory limitations inherent in single-framework models. The conceptual logic parallels Wang et al.’s [28] “economic–social–environmental” ternary evaluation system for ecological resettlement, yet places greater emphasis on the interactive coupling between regional industries and resettler livelihoods as a dynamic mechanism for sustainable development.

(2) The model employs a hybrid AHP–EW approach to balance expert judgment with data-driven objectivity in determining indicator weights. The TOPSIS is applied for composite scoring, effectively revealing regional differentiation. The DCCM is utilized to overcome the compensatory limitations of linear aggregation methods, capturing nonlinear interactions between subsystems. Finally, Moran’s I is used to detect spatial correlation, enabling identification of clustering or dispersion patterns across regions.

The composite score highlights clear contrasts in the development levels of the two subsystems, effectively distinguishing between regions characterized by “high support–low livelihood” and those exhibiting “high livelihood–low support”. Within the RDS, high-level regions (R1, R8, R6) display a well-balanced structure across the economic, social, and ecological dimensions, whereas low-level regions (R3, R9, R2) reveal weaknesses concentrated in a single dimension—ecological fragility in R3 and economic underdevelopment in R9. In the RLD, high-scoring regions (R3, R4, R2) exhibit strong coupling between HC and FC, reflecting effective labor–income synergy. Conversely, low-scoring regions (R9, R1, R10) are hindered by deficiencies in critical capital forms or by ecological constraints, which limit the overall capacity for livelihood improvement and system-level coordination.

The construction of the SDI directly addresses calls for improved robustness and interpretability in composite indices (e.g., Greco et al. [32]) by integrating both the interaction intensity and development level of the two subsystems into a single, quantifiable metric. Based on SDI values, the ten regions can be classified into four distinct categories: Benign Synergy (characterized by bidirectional reinforcement between livelihood and regional systems), Region-Driven (marked by strong regional support but limited livelihood transmission), Unipolar Constraint (where progress is restricted by a single-system bottleneck), and Dual-Weak Lock-In (defined by low-level disconnection in both systems). Overall, SDI values across the ten regions exhibit a clear gradient differentiation pattern and weak spatial dispersion, reflecting variations in the depth and balance of livelihood–region coordination.

(3) The model was applied to 10 regions within Zhejiang’s QC Reservoir area, utilizing data collected from 289 resettler households. The resulting composite rankings correspond closely with each region’s industrial characteristics and resettler capital profiles, demonstrating the model’s strong reliability and validity. The analysis effectively distinguishes four coordination types and identifies critical developmental bottlenecks, notably skill–industry mismatches (e.g., HC5 and ES4) and ecological constraints, which serve as primary limiting factors in achieving balanced, synergistic development.

For instance, the model identifies the core constraint in R3 (SDI = 0.315) as a pronounced mismatch between heavy-industrial structures and resettlers’ skill composition, revealed through cross-analysis of ecological weakness in the RDS and skill misalignment in the RLD. The comparison between ecosystem service provision and NC in R1 (SDI = 0.596) highlights a synergy gap between ecological conservation imperatives and livelihood transformation capacity. In contrast, R6 (SDI = 0.738) exhibits a virtuous “industry-driven–livelihood-enhancing” feedback cycle, arising from the high compatibility between employment absorption in the RDS and skill matching in the RLD, which jointly reinforce sustainable regional–livelihood coordination.

This analysis provides targeted empirical evidence to inform differentiated policy formulation. For highly synergistic regions, policy efforts should prioritize the consolidation and expansion of multidimensional advantages to sustain balanced growth. For moderately synergistic regions, the focus should be on strengthening endogenous development momentum by enhancing the internal drivers of coordination. For low-synergy regions, priority should be given to addressing capital deficiencies and restoring fundamental interactions and functional linkages between the livelihood and regional systems to re-establish a stable foundation for coordinated development.

5.2. Practical Proposals

Empirical measurement of the SDI demonstrates that achieving high-level coordination between resettlers’ livelihoods and regional development depends on two-way empowerment: both the reciprocal interaction between regional support systems and livelihood capitals, and the empowerment of resettlers as active agents capable of driving their own adaptive development and integration within the regional economy.

5.2.1. Differentiated Strategies to Optimize Coordination

Policy interventions should precisely target the core bottlenecks and strengths associated with each coupling type and provide classified guidance accordingly.

(1)

Benign Synergy regions: consolidate strengths and build demonstration benchmarks

Benign Synergy regions (R6, R8, and R4) demonstrate robust system interactions with no major deficits across subsystems. Policy efforts in these regions should prioritize reinforcing multidimensional advantages in industry–skill–ecology integration, fostering innovation capacity, and establishing demonstration zones for coordinated development. For industry-fitted region R6, local human resource and social security bureaus, in collaboration with characteristic enterprises in benign synergy regions (e.g., R6’s smart manufacturing enterprises), should develop customized vocational training systems through a cooperative model of “enterprise orders + government subsidies + training providers” to strengthen HC5 (employment–industry matching) and sustain industrial adaptability. For ecology-adapted region R8, county-level resettlement bureaus and cultural and tourism bureaus should jointly support policies to facilitate resettlers’ entrepreneurship in eco-cultural tourism (e.g., homestays and ecological experience projects) to convert ecological resources into sustainable livelihood income. This approach will enhance NC3 (cropping–industry alignment) and promote sustainable ecological utilization. For service-adapted region R4, public service management departments at all levels should jointly implement measures to improve the interoperability of public service across administrative districts—particularly in the fields of education and healthcare. This initiative aims to reinforce SC2 (community participation and integration) and consolidate the social foundation for regional cohesion.

(2)

Region-driven areas: address livelihood gaps and improve support transmission

These regions (R5 and R1) possess robust regional support systems yet exhibit underperforming livelihood systems. Policy efforts should therefore focus on enhancing the conversion efficiency of regional resources into tangible livelihood outcomes. For commerce-led region R5, local human resource and social security bureaus, in collaboration with bureaus of commerce and digital economy development, should jointly implement digital upskilling programs to strengthen ES4 (employment absorption) and improve access to higher-value employment opportunities within the service and digital economy sectors. For water-ecology-sensitive region R1, local water conservancy bureaus and agriculture and rural bureaus should develop water-compatible industries such as organic agriculture and reservoir-based tourism, and create ecological protection and tourism service positions to achieve a sustainable balance between environmental conservation and livelihood enhancement.

(3)

Unipolar-Constrained regions: target single-system deficits to activate coordination

These regions (R2, R7, and R10) face developmental constraints arising from weaknesses in a single subsystem. Targeted interventions should therefore focus on removing these bottlenecks and restoring system balance between regional support and livelihood development. For heavy industry region R2: Bureaus of industry, ecology and tourism should foster complementary service sectors—such as industrial logistics and industrial heritage tourism—while advancing the green upgrading of core industries. These measures will mitigate ecological stress and enhance industrial sustainability. For industrial-transition region R7: Human resources and social security bureaus, together with local industry associations, should establish component-processing and auxiliary industry linkages with neighboring R6 to leverage industrial spillover effects. Simultaneously, vocational training and industrial alignment initiatives should be strengthened to enhance resettlers’ skill adaptability. For ecology-led region R10: Local cultural and tourism bureaus, agriculture and rural bureaus, and commerce departments should expand non-farm livelihood opportunities through initiatives such as fishing-village homestays, seafood e-commerce platforms, and ecological experience projects. This strategy will diversify resettlers’ income sources (enhancing FC2) and strengthen household resilience.

(4)

Dual-Weak-Lock-in regions: break low-level equilibrium and reconstruct coordination foundations

Regions such as R3 and R9 exhibit systemic deficiencies across both the RDS and RLD subsystems, resulting in a persistent lack of coordination and functional disconnection. Effective interventions must therefore adopt a dual-focus strategy that simultaneously strengthens regional support capacity and enhances resettlers’ livelihood foundations.

The government of R3, in collaboration with development and reform commissions, should establish interregional cooperation mechanisms with higher-performing regions (e.g., R4 and R6) to jointly develop resettler entrepreneurship parks. Such partnerships would promote industrial diversification, reduce over-reliance on heavy industry, and facilitate technology transfer, market integration, and employment expansion—thereby strengthening ES3 (performance of characteristic industries) and rebuilding the interaction foundation between RDS and RLD.

The government of R9 should encourage relevant departments—such as agriculture and rural bureaus, human resources and social security bureaus, and rural revitalization offices—to develop regional ecological agriculture brands through contract farming models, standardized production systems, and inclusive vocational training programs. This strategy commercializes local ecological resources, strengthen resettlers’ human capital, and compensates for existing skill and capacity deficits.

5.2.2. Cross-Regional Coordination and Safeguard Mechanisms

To effectively implement differentiated development strategies and advance territory-wide coordination, it is essential to dismantle institutional barriers, optimize resource allocation, and strengthen integration mechanisms between the livelihood and regional systems [45]. These measures will facilitate more equitable resource flows, improve cross-sector collaboration, and enhance the overall efficiency of coordinated resettlement governance.

(1)

Establish factor-flow mechanisms

Higher-level governments should lead the construction of a cross-regional industry–skills information-sharing platform to align industrial labor demands with resettlers’ skill profiles. Building on this foundation, this platform will align cross-district vocational training and certification programs to enhance workforce adaptability, and develop ecological product trading platforms to promote resource complementarity and interregional benefit sharing. Especially for regions exhibiting low SDI scores (e.g., R3, R9), county resettlement bureaus and human resources and social security bureaus should reinforce cross-regional value chains and employment linkages to better align livelihood opportunities with broader regional development trajectories. These measures would strengthen functional linkages across regions and foster integrated, sustainable development.

(2)

Strengthen policy coordination mechanisms

Under the leadership of higher-level governments, establish cross-city and cross-county coordination groups to guide and oversee integrated resettlement development. Standardize training criteria and related technical protocols to prevent interregional policy divergence, ensuring consistency in implementation and evaluation. Furthermore, strengthen grassroots governance capacity by creating incentive-compatible mechanisms [46] and integrating administrative and financial resources to better align resettlers’ needs with regional development opportunities.

5.2.3. Paths to Activate Resettlers’ Agency

Long-term coordination sustainability fundamentally relies on enhancing resettlers’ internal development capacity and initiative, thereby reinforcing their endogenous motivation to participate actively in and contribute to regional development processes. These strategies represent the fundamental pathway for enhancing synergy across all regions. In particular, low-synergy regions experience more pronounced structural mismatches between the RLD and RDS systems and therefore require deeper and more sustained policy implementation. Specifically, county-level resettlement bureaus, rural revitalization offices, human resource and social security bureaus, and township governments must intensify efforts to align resettlers’ skills with local industrial demands, expand participatory governance mechanisms, and strengthen policy communication, with the objective of narrowing the coordination gap more effectively.

(1)

Implement demand-driven, targeted skills training

Drawing on industry forecasts and resettlers’ willingness surveys, develop “menu-style” training systems co-designed with enterprises to enhance program relevance and practical impact. Complement these efforts by providing entrepreneurship support, low-interest financing, and mentorship programs to reduce start-up risks and strengthen HC5 (employment–industry matching), thereby improving the alignment between workforce skills and regional labor market demands.

(2)

Expand channels for resettler participation in governance

Establish participatory platforms enabling resettlers to engage directly in decision-making related to resettlement planning, industrial development, and community governance. Drawing on the Living Lab model proposed by Galderisi et al. [19], create multi-stakeholder governance forums that convene regularly to address key issues such as training programs, industry alignment, and ecological policy coordination. These initiatives will strengthen resettlers’ sense of ownership, agency, and belonging, fostering more inclusive and sustainable governance of resettlement processes.

(3)

Strengthen policy outreach and interpretation

Enhance policy transparency through community briefings, dedicated hotlines, and digital communication tools (e.g., short-form videos). As demonstrated by Shi et al. [47], transparent and accessible communication significantly improves policy comprehension and resettler engagement. Furthermore, optimize the implementation of follow-up policies to strengthen resettlers’ sense of acquisition and elevate their development expectations, thereby fostering proactive participation and sustained motivation in livelihood advancement.