Multi-Scale Resilience Assessment and Zonal Strategies for Storm Surge Adaptation in China’s Coastal Cities

Abstract

Storm surges are the leading marine disaster in China’s coastal cities, with their impacts exacerbated by climate change and rapid urbanization. Despite their significance, most existing studies focus on a single scale, neglecting the complex, multi-scale nature of urban resilience and the interrelated governance strategies needed to address storm surge risks. This study introduces a dual-scale resilience indicator system—macro (prefecture-level cities) and micro (coastal buffer grids)—within the “exposure–sensitivity–adaptation” framework, utilizing multi-source data for a comprehensive assessment. This research also explores the impact mechanisms of storm surges on urban areas and proposes zonal governance strategies. Findings indicate that resilience varies spatially in Chinese coastal cities, with a pattern of “high resilience in the north, low resilience in the south, and a mix in the center.” At the macro scale, key limitations include policy implementation, infrastructure capacity, and social vulnerability. At the micro scale, factors such as inadequate green space, increased impervious surfaces, limited shelter access, and low utility network density lead to the emergence of “low-resilience units” in ecologically sensitive and mixed coastal zones. The study further reveals the synergies between resilience drivers across scales, emphasizing the need for integrated cross-scale governance. This research advances resilience theory by expanding spatial scales and refining indicator systems, while proposing a zonal governance framework tailored to resilience gradation. It offers a quantitative basis and practical strategies for fostering “safe cities” and advancing “adaptive spatial planning” in the context of sustainable development.

1. Introduction

The intensification of global climate change, marked by rising sea levels and an increasing frequency of extreme weather events, has heightened the vulnerabilities of coastal cities, particularly in China’s coastal regions. Storm surges, the most prominent marine disasters, are abnormal rises in sea level caused by strong winds and low atmospheric pressure during storms, posing significant and complex risks to urban areas through coastal flooding and substantial damage to infrastructure and ecosystems [1]. Between 2001 and 2024, storm surge disasters in China led to direct economic losses exceeding 70 trillion yuan (approximately 10.8 trillion USD), as reported by the China Meteorological Administration (CMA), while also causing widespread flooding and significant infrastructure damage across coastal regions [2,3]. Although years of governance and development have reduced their impact on human life, considerable uncertainties regarding the spatial security, economic stability, and social functions of coastal urban agglomerations remain.

As coastlines face escalating exposure to climate risks, efforts to enhance resilience encounter significant challenges. “Coastal city resilience” is operationally defined as the capacity of urban systems and coastal spaces to sustain their core functions, absorb disturbances, and quickly recover and adapt to changes induced by marine disasters such as storm surges [4]. This includes the ability of social, economic, engineering, and urban systems to self-regulate and recover in the face of disruptions, ensuring the maintenance of essential services and infrastructure [5,6]. Unlike traditional studies on “vulnerability” or “disaster resilience capacity”, research on resilience emphasizes the dynamic response capacity of systems throughout the risk process, with particular focus on enhancing adaptability and fostering collaborative governance mechanisms in the context of spatial planning interventions [7,8]. Despite substantial progress, current studies on urban resilience in the context of storm surges remain limited in several key areas. First, most research has been confined to a single scale, focusing predominantly on prefecture-level cities or individual communities, while overlooking the importance of cross-scale collaboration within urban systems [9,10]. Second, methodological approaches are often constrained, with existing methods seldom integrating system-based, spatially explicit approaches linking data analytics with spatial planning—critical components for addressing practical needs related to storm surge disaster prevention, mitigation, and integration with territorial spatial planning.

To address these gaps, a “multiple scale—multiple factor—resilience enhancement path” research paradigm is proposed, focusing on 52 typical Chinese coastal cities. By adopting a dual-scale approach, from macro (prefecture-level cities) to micro (coastal buffer zone grids), the study constructs a resilience indicator system based on the “exposure–sensitivity–adaptation” framework. Remote sensing, statistical, and spatial data are integrated to identify the key factors influencing coastal cities’ responses to storm surges and their spatial heterogeneity. Based on these findings, targeted enhancement pathways are proposed to guide regional management and optimize spatial resilience. This research seeks to address the following key scientific questions:

(1)

How can a multi-scale, operational resilience assessment system for coastal cities be constructed?

(2)

What are the spatial patterns and dominant influencing factors of coastal city resilience, and how do they vary across scales?

(3)

What actionable, zoned, and implementable improvement strategies can be proposed for risk governance and planning practices?

2. Theoretical Background

2.1. Resilience Concepts and Measurement Methods in Coastal Storm Surge Contexts

The definition of urban resilience has evolved significantly [11,12,13]. Cutter et al. [14] emphasized that resilience includes three key capacities: resistance, recovery, and adaptation, which are essential for cities to withstand, recover from, and adjust to disasters. Meerow et al. built on this by highlighting the need for urban resilience to include not just functional recovery, but also adaptation in the context of socio-ecological systems [15]. The IPCC AR6 reinforced this idea by defining resilience as the ability of socio-ecological systems to maintain structure and function, self-organize, and adapt after disturbances [16]. More recent work has advanced these ideas further by proposing a dynamic framework where urban systems are linked to ecological resilience through multi-dimensional, coordinated responses to stressors [17,18,19]. This expanded understanding of resilience integrates ecological systems with urban planning, creating a more holistic view of resilience that incorporates both adaptation and recovery capacities [20]. Collectively, these frameworks highlight the interconnectedness of urban systems and the importance of integrating multiple resilience capacities to effectively address climate and disaster risks. In summary, resilience in urban systems is multi-dimensional, addressing the capacity to absorb shocks, recover, and adapt, and includes ecological, social, economic, and infrastructural components. This evolution in thinking aligns with recent frameworks such as those from the UNDRR (2023) and the IPCC AR6 reports, which stress the importance of adaptive capacities in complex socio-ecological systems.

Resilience measurement methods include indicator systems, simulation models, network analysis, and spatial statistical techniques [21,22,23,24]. The indicator system approach is most widely used, as it allows for the customization of evaluation dimensions based on different scales and facilitates the collection of relatively comprehensive input variables through multi-source data [25]. Although subject to challenges such as subjective weighting and multicollinearity, it remains the most practical for planning applications. Internationally, resilience indices have been integrated with GIS and machine learning for decision support. In China, research has advanced in parallel. Liu Lingna et al. analyzed resilience trends in 30 provincial capital cities in China, using an improved CRITIC model to explore urban resilience from 2000 to 2017 [26]. Research on coastal prefecture-level cities in eastern China highlights the complex, composite resilience structure of coastal cities, shaped by the interaction between natural resource constraints, ecosystem services, and development dynamics [21].

The concept of resilience is embedded within several United Nations Sustainable Development Goals (SDGs), particularly SDG 11 (Sustainable Cities and Communities), SDG 13 (Climate Action), and SDG 1 (End Poverty). These SDGs integrate key priorities of the Sendai Framework for Disaster Risk Reduction and provide a global framework for cities to align their local resilience strategies with international sustainability targets. Embedding resilience metrics within SDG 11 and SDG 13 enables measurable integration between local adaptation planning and international sustainability objectives [27,28]. By framing resilience within the SDG agenda, cities can not only align local adaptation efforts with global climate goals but also foster cross-scalar policy coordination, ensuring that local resilience strategies contribute to global resilience targets and are comparable across regions. This approach facilitates a more comprehensive and coordinated effort to address climate risks and disaster impacts, ensuring that resilience-building activities contribute to both local and global sustainable development.

In the context of storm surges, resilience assessment research has primarily focused on identifying risks, simulating inundation, and mapping vulnerability. Studies specifically addressing “urban resilience”, particularly the ability of urban systems to withstand, recover from, and adapt to storm surge impacts, remain relatively limited [29].

In the field of risk identification, Min and Long et al. used numerical hydrodynamic models to simulate the inundation extent, strengthening frequency, and changes in the frequency of storm surges, developing risk classification maps [30]. Shi Xianwu et al. conducted a storm surge risk zoning study based on data from 62 tidal gauge stations in China, identifying high-risk regions such as Bohai Bay, the Yangtze River Estuary, and the Pearl River Delta [31]. While these studies provide essential data for evaluating the “exposure” of coastal cities to storm surges, they offer limited insights into the mechanisms of “sensitivity” and “adaptive capacity.” Therefore, further research is needed to explore the overall resilience of cities to storm surges, encompassing dimensions such as “natural systems,” “infrastructure,” “social services,” and “governance mechanisms.” Most studies, however, tend to focus on isolated ecological buffer projects or specific city-level infrastructure responses, without forming integrated models that analyze collaborative mechanisms at both macro and micro scales [32].

Regarding the quantification of storm surge resilience, Meng et al. analyzed 15 coastal cities in China, including Tianjin, during the period 1990–2020 [33]. The study found a significant positive correlation between urban adaptability factors—such as fiscal expenditure, fixed asset investment, and healthcare infrastructure—and resilience, suggesting that strengthening adaptive capacity can effectively reduce urban vulnerability to disasters. Nevertheless, research integrating multi-scale resilience mechanisms within the context of storm surges remains scarce. In recent years, nature-based storm surge adaptation strategies have progressed to practical application, with notable examples including the pump station–wetland property rights synergy system in the North Sea region of the Netherlands and New York’s “Green Waterfront” resilience retrofit plan. These cases provide valuable insights for developing coastal resilience strategies in China, yet challenges remain in translating them into actionable operational planning pathways, particularly in terms of institutional coordination and spatial system integration. Current research on storm surge disaster resilience shows several key characteristics: diverse methodologies for risk identification, extensive analysis of storm surge impact ranges, and a growing focus on the multi-dimensional aspects of urban resilience. However, comprehensive, multi-dimensional evaluations of urban systems under storm surge conditions are still underdeveloped. Although increased investment in adaptive capacity has been shown to significantly enhance resilience, the recognition and integration of multi-scale spatial mechanisms into urban planning practices have not yet been effectively achieved. Therefore, this paper proposes a multi-scale indicator system based on the resilience framework and analyzes the underlying driving mechanisms, with the aim of providing scientific support and strategic guidance for urban–coastal spatial planning and zoned policymaking.

2.2. Resilience Assessment Method Based on the “Exposure-Sensitivity-Adaptation” Framework

The “Exposure-Sensitivity-Adaptation” (ESA) framework is a critical theoretical approach for assessing urban resilience, breaking down the vulnerability of a system into three dimensions [34]. Exposure represents the degree to which a system is subjected to natural or human-made risks. Sensitivity reflects the system’s response strength to impacts when affected by these risks [35]. Adaptation refers to the system’s ability to buffer, absorb shocks, and rapidly recover or even transform. In the context of storm surges, this framework offers high theoretical compatibility and strong operational utility, enabling the structured integration of multi-source heterogeneous data and facilitating the transition from qualitative understanding to quantitative evaluation.

Firstly, at the “Exposure” level, traditional studies have focused on the geographic spatial distribution of potential disaster impacts, such as storm surge inundation maps, tide gauge records, coastal topography, and coastline configurations, among other natural conditions. These fundamental factors determine the impact scope and the baseline risk status. For instance, Lu et al. quantified exposure by combining high-resolution remote sensing imagery with Digital Elevation Models (DEM), emphasizing the importance of factors such as “proximity to water bodies, topographic variations, and land types” in exposure assessments [36]. Data for the exposure dimension is primarily obtained through numerical simulations and remote sensing interpretation, which are suitable for multi-scale evaluations and provide a static risk baseline for planning interventions.

Secondly, the “Sensitivity” dimension focuses on how the internal resource attributes and structural characteristics of a system influence the intensity of its response to impacts [37]. This dimension reflects the degree of functional or structural damage caused under the same level of exposure. In urban resilience assessments, sensitivity indicators often include infrastructure density, population distribution, ecosystem characteristics, and transportation network disruptions. Research has shown that high-sensitivity areas often correlate with “social structure vulnerability + aging infrastructure,” leading to more significant functional failures in the urban system [38]. Therefore, accurate quantification of sensitivity at the urban scale is essential for identifying weak points in system functionality.

Finally, the “Adaptation” dimension centers on the resources and capabilities that support the system’s response and recovery strategies, including material assets, institutional resources, social capital, and knowledge mechanisms [39,40]. In urban resilience assessments, many scholars have quantified this dimension using indicators such as the density of public facilities, fiscal emergency capacity, per capita medical resources, and the planning response system [41,42]. Standardized adaptation scores are important indicators of policy intervention and resilience improvement trends. In practice, the “Exposure-Sensitivity-Adaptation” framework is widely used to construct indicator systems, quantify resilience levels, and analyze driving mechanisms. Although these methods are well-established in practice, they have key limitations. Sensitivity and adaptation assessments rely heavily on data availability and the representativeness of indicators, and assigning weights to indicators can lead to subjective biases. Therefore, it is necessary to optimize the indicator system within a multi-scale context and validate the effectiveness of the coupling mechanism across scales.

3. Research Framework and Study Area

3.1. Research Framework

This study proposes a comprehensive framework for assessing the resilience of coastal cities and zones in Eastern China in the context of storm surges. The framework is built around a “dual-scale analysis, multi-factor evaluation, and multi-path optimization” approach, with the primary goal of identifying spatial disparities, understanding the underlying mechanisms influencing resilience, and proposing actionable strategies for enhancement. The framework consists of three core components. First, the resilience evaluation system is developed based on the H-S-V (Hazard-Sensitivity-Vulnerability) theoretical model and encompasses natural, social, engineering, and institutional factors [43]. This system ensures a holistic evaluation, providing horizontal completeness across dimensions and vertical scalability for various spatial scales. Second, the dual-scale evaluation approach is employed. At the macro scale, the evaluation focuses on prefecture-level cities to assess their overall resilience to storm surges. At the micro scale, a 15 km coastal buffer zone is selected for case study cities, and fine-scale evaluations are conducted using a 1 km × 1 km grid. This approach specifically targets the identification of localized vulnerabilities and structural weaknesses. Finally, the study emphasizes mechanism identification and strategy formulation. By analyzing the dominance of influencing factors and exploring their spatial response mechanisms, the study formulates targeted resilience strategies. These strategies are designed to enhance resilience at different spatial levels, as illustrated in Figure 1.

3.2. Study Area

China’s coastal regions are home to 55 coastal prefecture-level cities (excluding Hong Kong, Macao, and Taiwan). This study focuses on these cities, selecting representative cases for a macro-micro linkage resilience evaluation and mechanism identification system. At the macro scale, 52 coastal prefecture-level cities from seven coastal provinces—Zhejiang, Fujian, Guangdong, Shandong, Jiangsu, Tianjin, and Shanghai—are selected for their spatial representativeness and comparability across natural geomorphology, coastline shape, land use intensity, and ecological background (see Figure 2). These cities span three major storm surge risk gradient zones: northern, central, and southern coastal China, ensuring broad spatial coverage. Local governance practices regarding urban resilience building, marine disaster response, and territorial spatial planning differ significantly across these cities, offering a robust foundation for identifying resilience mechanisms and classifying resilience types. To maintain consistency and operability, areas with special administrative status (such as Hong Kong, Macao, Taiwan, and Sansha) are excluded from the study. Geographically, the coastal boundary is defined by the Qiantang River and Hangzhou Bay, excluding the cities of Hangzhou and Shaoxing from the coastal category.

At the micro scale, Tianjin and Shanwei are selected as representative cities within the 15 km coastal buffer zone for a more detailed examination of spatial resilience patterns under varying marine and urban contexts. Tianjin, located in the northern part of Bohai Bay, is home to industrial clusters, such as petrochemical industries, port logistics, and high-end manufacturing, reflecting a “high-intensity development—high-risk pressure” pattern. The city faces compounded risks, including storm surges, sea-level rise, and saline-alkali erosion. In contrast, Shanwei, situated in the southern part of the Pearl River Estuary, is characterized by high agricultural exposure, significant ecological sensitivity, and underdeveloped infrastructure, embodying an “ecologically sensitive—lagging infrastructure—weak disaster prevention capacity” model. A comparison of these cities’ spatial resilience demonstrates that when enhancing coastal resilience to storm surges, tailored strategies are necessary. Tianjin requires robust infrastructure upgrades to withstand high-risk industrial pressures, while Shanwei demands a focus on ecological restoration and strengthening of disaster prevention capacity. This comparative study provides a practical and scalable model for spatial planning and risk prevention, offering insights that are applicable to other coastal cities facing similar challenges.

4. Methods and Data Sources

4.1. Research Methods

4.1.1. Macro Resilience Comprehensive Indicator System

The indicator system is organized into three logical tiers comprising objectives, criteria and indicators. It is grounded in the social, ecological and technical system resilience framework together with urban vulnerability assessment theory and the IPCC definitions of adaptive capacity and system response [44]. The system is based on the “exposure—sensitivity—adaptive capacity” framework proposed by Cutter et al. and subdivides the indicators into specific categories, including population exposure, spatial exposure, infrastructure sensitivity, socio-economic vulnerability, and adaptive governance capacity [45]. In the exposure dimension, indicators such as “population density” and “port and embankment length” represent the concentration of urban population and marine-related infrastructure, respectively, which are key risk factors for storm surge impacts. In the sensitivity dimension, indicators including “drainage network density” and “per capita road area” reflect the vulnerability of infrastructure to storm surge effects and the accessibility of emergency services. “Per capita GDP” and “proportion of the tertiary sector” capture the city’s economic flexibility and its capacity for resource mobilization [46]. In the adaptive capacity dimension, indicators such as “number of hospital beds” and “disaster risk reduction investment” assess the city’s emergency response resources and governance capacity [47]. “Insurance depth” and “mobile phone user numbers” serve as proxies for social protection coverage and information accessibility during crises [48]. The indicators were chosen based on their relevance to urban resilience in the context of storm surge adaptation, their ability to represent various dimensions of resilience, and the availability of reliable data. These indicators are grounded in clear theoretical principles and offer practical insights into the city’s ability to resist, absorb, mitigate, and recover from storm surge impacts. Additionally, the selection is grounded in prior research, as referenced in the relevant studies, ensuring that the indicators are well-established and applicable to this context [49,50,51]. A comprehensive list of indicators is provided in

Table A1. Macro-level Urban Resilience Comprehensive Indicator System.

Primary Dimension	Secondary Dimension	Indicator Name	Positive/Negative	ID	Indicator Explanation	Data Source
Exposure	Population Exposure	Population Density	-	A1	Proportion of urban population to land area	National Bureau of Statistics, local statistical yearbooks
		Aging Level	-	A2	Urban population aging rate
		Permanent Population Growth Rate	+	A3	Expansion of exposed areas due to urban expansion
	Spatial Exposure	Average Population Area in Built-up Zone	-	A4	Proxy for urban spatial development level	Landsat imagery, Earth Explorer platform
		Proportion of Built-up Land	-	A5	Proportion of built-up land to total urban land
		Length of Ports and Dikes	+	A6	Extent of coastal construction in the city	Coastal infrastructure data, GIS analysis
		Coastline Length	-	A7	Can be used for normalizing indicators like dikes
		Proportion of Fishery Output	-	A8	Represents coastal industry exposure	National Bureau of Statistics, local statistical yearbooks
Sensitivity	Traffic Infrastructure	Road Density	+	A9	More densely packed roads lead to slower recovery after a disaster.
		Per Capita Road Area	+	A10	Proportion of per capita road area, emergency response capacity
	Drainage System	Drainage Pipe Network Density	+	A11	Reflects urban drainage capacity	Local municipal data, water infrastructure
		Sewage Pipeline Density	+	A12	Ground impermeability, reflecting flood risks
		Runoff Coefficient	-	A13	Reflects urban surface permeability	Water runoff data, China Meteorological Administration, local hydrology data
	Medical Assistance	Medical and Health Beds per 10,000 People (Merged)	+	A14	Disaster medical treatment capacity	National Bureau of Statistics, local statistical yearbooks
		Proportion of Unemployment Insurance Coverage	+	A15	Disaster medical treatment capacity	Social insurance data, local labor department
	Environmental Ecology	Vegetation Coverage Rate	+	A16	Reflects natural water retention and buffering capacity	NDVI from Landsat imagery, Earth Explorer platform
	Economic Capacity	Per Capita GDP	+	A17	Foundation for resource mobilization post-disaster	National Bureau of Statistics, local statistical yearbooks
		Per Capita Disposable Income	+	A18	Household disaster resistance and recovery capacity
		Proportion of Tertiary Industry in GDP	+	A19	Higher resilience of light asset-based structures
		Economic Growth Rate	+	A20	Economic vitality indicator
Adaptation Capacity	Social Resources	Proportion of Young and Middle-aged Labor Force	+	A21	Human recovery resources
		Proportion of Educated Population	+	A22	Reflects knowledge acquisition and emergency response
		Depth of Comprehensive Insurance	+	A23	Total insurance expenditure/GDP	Insurance data, local finance department
		Mobile Phone Users per 10,000 People	+	A24	Information access and emergency communication capacity	Telecom data, mobile service providers
	Policy and Governance	Disaster Reduction and Prevention Investment	+	A25	Local financial emergency capacity	National Bureau of Statistics, local public financial data, disaster investment reports
		Proportion of Public Safety Expenditure	+	A26	Investment in law enforcement and rescue systems
		Proportion of Grain and Oil Reserves Expenditure	+	A27	Disaster recovery material guarantees

in Appendix A.

4.1.2. Micro Resilience Comprehensive Indicator System

This study focuses on the 15 km coastal buffer zones of Tianjin and Shanwei, selected to represent the urban and natural diversities of the Bohai Bay and South China Sea economic zones. The analysis maintains a consistent methodological approach, employing a unified indicator system to ensure spatial comparability and regional applicability. The system, based on the concept of “urban spatial resilience,” comprises four key dimensions: spatial morphology, adaptive capacity, a, and landscape pattern. It assesses the structural vulnerabilities, response capabilities, and ecological support systems of cities in the context of storm surges and other abrupt risks. Indicators are selected through a synthesis of prior literature, selection criteria, and data availability [41,52,53]. A total of 18 representative indicators is categorized into four aspects, as described below.

(1)

Spatial Morphology: Indicators such as “degree of mixed land use” and “road network density” reflect the urban spatial structure’s resilience and accessibility, which are essential for emergency response.

(2)

Adaptive Capacity: “Emergency shelter density” and “medical facility density” evaluate the city’s capacity to evacuate and provide medical assistance during emergencies.

(3)

Natural Background: Indicators such as elevation and annual precipitation are used to assess the city’s exposure and environmental carrying capacity. While annual precipitation may not directly correlate with storm surges, it can indicate potential flooding risks and contribute to understanding the city’s overall vulnerability to extreme weather events. Precipitation levels, when considered alongside other factors such as coastal proximity and infrastructure resilience, can provide additional insight into the region’s susceptibility to flooding and environmental stressors.

(4)

Landscape Pattern: Ecological connectivity is assessed using indicators like the “aggregation index” and “patch cohesion index,” which measure the stability and redundancy of the ecological system under extreme conditions. To ensure the indicator system’s adaptability across both cities, all indicators are normalized based on a unified calculation logic and standardized spatial scale (1 km × 1 km grid). The indicators are identified using a standardized nomenclature format, “T/S + serial number,” where “T” and “S” represent Tianjin and Shanwei, respectively. This standardized system facilitates cross-city comparisons within a shared framework, ensuring consistency in analysis and evaluation. A detailed list of indicators is shown in

Table A2. Micro-level Spatial Resilience Comprehensive Indicator System.

Primary Dimension	Indicator Name	Positive/Negative	ID	Indicator Explanation	Data Source
Spatial Functional Structure	Mixed Land Use Diversity Index	+	T/S1	The land use mix level reflects spatial resilience	Remote Sensing Data
	Road Network Density	+	T/S2	Reflects regional development intensity and traffic accessibility	Statistical Yearbooks and Government Public Data
	Surface Permeability Ratio	-	T/S3	Higher surface permeability helps alleviate waterlogging.	Remote Sensing Data
	NDVI Green Space Ratio	+	T/S4	An important component of ecological buffer zones can be extracted via remote sensing
Spatial Adaptation Capacity	Emergency Shelter Density (POI)	+	T/S5	Core element of evacuation and resettlement capacity	Points of Interest and Basic Spatial Data
	Emergency Shelter Distance	-	T/S6	Shorter distance to shelters improves response efficiency.
	Medical Facility Density	+	T/S7	Reflects the supply capacity of medical resources during a disaster
	Hospital Distance	-	T/S8	Shorter distance to hospitals enables faster rescue response.
	Firefighting Facility Density	+	T/S9	Reflects the density of emergency rescue facilities
	Firefighting Facility Distance	-	T/S10	Shorter distance improves emergency response efficiency.
Natural Environment	Elevation	+	T/S11	Higher elevation provides more safety, which can be obtained via remote sensing.	Remote Sensing Data
	Terrain Roughness	-	T/S12	More complex terrain makes drainage and construction more difficult.
	Annual Precipitation	-	T/S13	Higher precipitation increases waterlogging and secondary disaster risks.	Meteorological and Hydrological Data
	Distance to River	+	T/S14	Areas closer to rivers are more vulnerable to storm surge impact.	Remote Sensing Data
Landscape Pattern	LSI Landscape Index	-	T/S15	More complex boundaries make the system less stable.	Remote Sensing Data
	Patch Density (PD)	-	T/S16	More scattered patches indicate higher ecological fragmentation.
	AI Aggregation Index	+	T/S17	Higher aggregation indicates better ecological connectivity.
	Patch Cohesion Index (COHESION)	+	T/S18	Tighter connections between patches enhance overall resilience.

in Appendix A.

4.1.3. Indicator Selection and Multicollinearity Testing

To ensure the independence and statistical validity of each indicator in the macro-micro scale resilience evaluation, this study introduces a two-step indicator selection procedure before indicator weighting calculations. This procedure includes Pearson correlation analysis and variance inflation factor (VIF) testing to eliminate information redundancy and multicollinearity issues.

First, Pearson’s correlation coefficient (Pearson’s r) was used to perform pairwise correlation tests on all standardized indicators. If the absolute value of the correlation coefficient (|r|) between any two indicators is ≥0.8, and the two indicators are highly substitutable in theoretical significance or data attributes (e.g., “medical facility density” and “number of hospitals”), the indicator with stronger explanatory power or higher data quality was retained to avoid weight distortion or scoring bias caused by indicator redundancy. Subsequently, to detect potential multicollinearity issues, the VIF statistic was used. Its mathematical expression is as follows:

$${\mathrm{VIF}}_k={\displaystyle \frac{1}{1-R_k^2}}$$

(1)

The variance inflation factor (VIF) is defined for the i-th independent variable. R2k is the coefficient of determination (i.e., goodness of fit) when the i-th independent variable is used as the dependent variable and all other independent variables are used as independent variables in a regression analysis; 1 − $R_k^2$ is the degree of independence of the i-th variable in the regression model. A higher VIF indicates that the variable is strongly explained by other variables, suggesting strong multicollinearity. The general criteria for judgment are as follows: VIF < 5: multicollinearity is acceptable; 5 ≤ VIF < 10: some multicollinearity exists, requiring cautious handling; VIF ≥ 10: severe multicollinearity, variables should be removed or merged. After the above tests, indicators with good suitability are retained for subsequent weighted calculations to ensure that the scoring system has good discriminative power and explanatory validity.

4.1.4. Weight Determination

In terms of weighting the macro-micro indicator system, because urban resilience involves multi-system and cross-scale factors, relying solely on subjective judgment or objective algorithms is insufficient to fully reflect the actual weight distribution. Therefore, a combined weighting approach was adopted, integrating the Analytic Hierarchy Process (AHP) and the entropy method. The AHP was used to construct a weighting system that reflects subjective importance through expert scoring and judgment matrices, incorporating policy experience and practical knowledge. The entropy method assigned greater weights to indicators with higher variability, based on the information entropy of their value distributions, thereby capturing objective differences. The weights derived from both methods were then averaged to obtain a comprehensive weight value. The specific calculation formula is as follows.

$$W_i=\alpha \cdot W_i^{AHP}+(1-\alpha )·W_i^{Entropy}$$

(2)

where W_i is the comprehensive weight of the i-th indicator, $\alpha$ representing the composite coefficient of subjective and objective weights (set to 0.5 in this study).

4.1.5. Comprehensive Score

Based on the constructed indicator system and weight settings, this paper uses the weighted linear summation method to calculate the resilience scores of the macro-micro indicator system. Considering the positive and negative attributes of different indicators, all indicators are normalized before entering the calculation to ensure comparability between different units of measurement [54]. The formula for calculating the comprehensive macro resilience score is as follows:

$$R_j={\displaystyle \sum _{i=1}^n{W}_i}·X_{ij}$$

(3)

where R_j is the comprehensive resilience score of the j-th city, X_ij is the standardized value of the i-th indicator for the j-th city, and W_i is the corresponding comprehensive weight.

The formula for calculating the micro-level resilience composite score is as follows:

$$R_{ij}={\displaystyle \sum _{k=1}^n{W}_k}·X_{ijk}$$

(4)

where R_ij is the comprehensive resilience score of the i-th spatial unit in city j; X_ijk is the standardized value of the i-th unit in city j for the kth indicator, W_k is the weight of the k-th indicator, i the grid spatial unit number, j the city code (T for Tianjin and S for Shanwei).

4.1.6. Hotspot Analysis

Based on the calculation of micro-scale resilience scores, hotspot analysis (Hotspot Analysis) is further employed to identify spatial distribution patterns. Specifically, the Getis-Ord Gi* statistic is used to conduct a significance test for the spatial clustering of cell resilience scores in a GIS platform, identifying regions with significantly high (hotspots) or low (cold spots) scores. This effectively reveals the spatial clustering characteristics of resilience distribution, analyses the spatial directionality of risk accumulation and capacity shortcomings in cities, and provides a quantitative basis for the classification of areas such as ‘priority intervention zones’, resilience enhancement zones’ and ‘ecological barrier zones’.

$$G_i^{*}={\displaystyle \frac{{\displaystyle \sum _{j=1}^n{w}_{ij}}{x}_j-\overline{X}{\displaystyle \sum _{j=1}^n{w}_{ij}}}{S\sqrt{\frac{n{\displaystyle \sum _{j=1}^n{w}_{ij}^2}-{\left({\displaystyle \sum _{j=1}^n{w}_{ij}}\right)}^2}{n-1}}}}$$

(5)

where the G_i* value is standardized to form a Z-score, which can be used for significance testing. If the Z-score is greater than 0 and significant, the cell is a hotspot (high resilience value area). If the Z-score is less than 0 and significant, it is a cold spot (low resilience value area). If the Z-score is not significant, it is a general area (non-aggregated area).

4.2. Data Sources

To ensure data completeness and accuracy, both macro- and micro-level indicator systems were developed using 2022 as the reference year. Data was obtained from the following sources:

(1)

Remote Sensing Data: High-resolution Landsat imagery (Landsat 8 OLI/TIRS) was used to map land use, land cover changes, and coastal distribution patterns. These images are essential for analyzing urban spatial structures and the impacts of storm surges. The Landsat images provide a spatial resolution of 30 m for most bands, which is suitable for identifying large-scale changes in land cover and infrastructure. The imagery was obtained from the Earth Explorer platform (https://earthexplorer.usgs.gov), accessed on 6 March 2025. Previous studies have validated the use of remote sensing data in assessing storm surge impacts and urban resilience, confirming the reliability of Landsat imagery in coastal impact analyses [55]. The advantages of Landsat imagery include its high spatial accuracy, consistent temporal resolution (available for multiple years), and its proven utility in environmental monitoring. However, limitations exist, including the lower spatial resolution compared to commercial satellite imagery (e.g., 10-m resolution from World View), which may affect the detection of finer details in urban environments. Additionally, cloud cover and atmospheric conditions can sometimes limit the clarity of images. Despite these limitations, Landsat imagery remains a reliable and widely used resource in coastal and urban resilience studies due to its long temporal history and global coverage.

(2)

Statistical Yearbooks and Government Public Data: Data from the National Bureau of Statistics and local statistical yearbooks provided socio-economic indicators such as population density, economic activity, industrial structure, and public facilities. These datasets are crucial for assessing urban carrying capacity and adaptive capacity. In addition, government-released information on disaster risk reduction investments, emergency management policies, and funding sources was collected to support the adaptability dimension (https://data.stats.gov.cn), accessed on 6 March 2025. These socio-economic datasets have been widely used in disaster vulnerability assessments, and their application in this study aligns with the standard practices in storm surge resilience research [56].

(3)

Meteorological and Hydrological Data: Historical data on storm surge events, tide monitoring, wind speed, and precipitation were obtained from the China Meteorological Administration and water resources departments (http://www.hydroinfo.gov.cn), accessed on 6 March 2025. These datasets are essential for understanding the frequency, intensity, and impacts of storm surges. As shown in Wang et al., such data has been extensively validated and utilized in storm surge impact studies, supporting its role in our framework for analyzing storm surge risks [57].

(4)

Points of Interest and Basic Spatial Data: Data on the spatial distribution of critical emergency facilities, such as medical services, shelters, and fire stations, were collected from open platforms like Gaode Maps (https://lbs.amap.com), accessed on 12 February 2025. This information indicates the city’s emergency response capabilities and social service infrastructure. Previous research has demonstrated the use of similar spatial data to assess the relationship between emergency infrastructure and disaster response efficiency, ensuring the robustness of this dataset in the current study [58].

(5)

Ecological Environment Data: Information on ecosystem services, vegetation cover, and wetland protection areas was integrated to assess the baseline ecological resilience of urban areas (https://www.nenv.org.cn), accessed on 12 February 2025. Chen and Zhang have shown that such ecological data is critical for assessing urban resilience to storm surges, particularly in coastal regions, and has been validated in similar storm surge assessments [59].

(6)

Survey and Field Data: Field surveys were conducted in key study areas to gather subjective indicators, including residents’ perceptions of storm surges, preparedness for emergencies, and satisfaction with government responses. These data complement the micro-level evaluation of social adaptive capacity. Previous studies have used similar survey data to assess social resilience and preparedness, ensuring the validity of these subjective indicators in our analysis [60].

5. Results

5.1. Macro-Scale Evaluation Results of 52 Coastal Cities

5.1.1. Correlation and VIF Analysis

Figure 3 presents the Pearson correlation and Variance Inflation Factor (VIF) analysis of the 27 selected indicators. The analysis reveals notable linear correlations among some variables. For instance, population density (A1) shows a strong positive correlation with built-up land ratio (A5) (r = 0.881, p < 0.01) and unemployment insurance participation ratio (A15) (r = 0.777, p < 0.01), indicating redundancy in A1 within the exposure dimension. Similarly, per capita road area (A10) demonstrates negative correlations with several variables (e.g., A5, A15, A21), suggesting potential cross-explanatory relationships between urban construction intensity and social structural factors. VIF analysis indicates that the majority of the indicators have VIF values below 5, suggesting that the linear correlations between most variables are within acceptable limits. However, a few indicators, such as population density (A1) and permanent population growth rate (A3), exhibit significantly high VIF values (9.57 and 9.63, respectively), approaching or exceeding the commonly accepted threshold of 10. Despite this, A1 is retained due to its critical role in representing urban exposure with strong explanatory power. Redundant variables, such as A5 and A15, were removed to avoid overlap, while A3 was kept for its effectiveness in representing urban development intensity and boundary expansion. Conversely, aging level (A2), despite exceeding the VIF threshold, was retained due to its low correlation with other social variables and significant contribution to evaluating social vulnerability.

By carefully selecting and removing highly correlated indicators, multicollinearity was effectively minimized, and the VIF for most variables was reduced to below 4. This ensures the robustness of subsequent regression and weighting analyses. The final model maintains a balanced distribution across the three primary dimensions: exposure, sensitivity, and adaptive capacity. Specifically, the exposure dimension includes seven indicators related to population structure (A1, A2, A3) and coastal involvement (A4, A6, A7, A8); the sensitivity dimension includes nine indicators covering transportation and drainage systems (A9–A13), healthcare infrastructure (A14), ecological regulation (A16), and economic support (A17, A19, A20); and the adaptive capacity dimension retains seven indicators reflecting social resources (A22–A24) and policy governance (A25–A27). In total, 23 indicators were selected for weight calculation and scoring.

5.1.2. Weight Comparison

The weight distribution of the 23 indicators is imbalanced but reflects the multi-dimensional nature of storm surge resilience, as shown in Figure 4. The NDVI green space rate (A16) (13.991%) and mobile phone users per 10,000 people (A24) (11.131%) have the highest weights, emphasizing the importance of natural ecosystems and communication infrastructure in urban disaster resilience. Green space, as a natural buffer, plays a crucial role in mitigating tidal impacts and slowing inland flooding, while mobile communication is essential for effective emergency response during extreme weather, highlighting the growing importance of “information resilience.”

Indicators such as port and dam length (A6), medical bed numbers (A14), and disaster reduction investments (A25) receive weights of 6.4–6.6%, reflecting the critical role of infrastructure and institutional mechanisms in disaster resistance. While coastal structures like ports and dams contribute to “hard protection,” their role in storm surge defense is more focused on erosion control than directly preventing flooding. Disaster reduction investments signify the fiscal resources for prevention and recovery. Medium-weight indicators, such as sewage pipeline density (A12, 4.793%) and public safety expenditure share (A26, 4.2995%), underscore the importance of urban recovery and emergency response. Low-weight indicators, like fisheries output proportion (A8, 1.0275%) and permanent population growth rate (A3, 1.182%), have limited structural impact on overall resilience, likely due to their lower data variability. Using the entropy weight method, NDVI green space rate (A16) and mobile phone user numbers (A24) have high entropy weights of 20.482% and 17.262%, respectively, showing strong differentiation across sample cities. In contrast, variables such as population growth rate (A3) and fisheries output proportion (A8) have low entropy weights (0.364% and 0.555%), indicating less influence on the model. Expert weighting methods assess the practical relevance of indicators within the resilience framework, emphasizing policy implications and theoretical significance. Despite the NDVI green space rate (A16) having the highest entropy weight, its expert weight is only 7.5%, lower than its entropy weight. Conversely, port and dam length (A6) and disaster reduction investments (A25) received expert weights of 6% and 6.5%, reflecting their central roles in storm surge protection.

5.1.3. Coastal City Resilience Scores and Spatial Patterns

As depicted in Figure 3 and Figure 4, storm surge resilience scores across coastal cities in eastern China exhibit significant variation. The overall resilience score ranges from 0.23 (Qinzhou) to 0.60 (Guangzhou), highlighting substantial spatial heterogeneity in infrastructure, ecosystem services, emergency response capabilities, and socio-economic structures (as shown in

Table 1. Resilience Scores for Coastal Cities in China Against Storm Surges.

Sorting	City	Resilience Score	Sorting	City	Resilience Score	Sorting	City	Resilience Score
1	Guangzhou	0.60	19	Rizhao	0.31	37	Tangshan	0.27
2	Shenzhen	0.49	20	Sanya	0.31	38	Qinhuangdao	0.26
3	Shanghai	0.48	21	Zhongshan	0.31	39	Jiaxing	0.26
4	Zhuhai	0.46	22	Jiangmen	0.31	40	Yancheng	0.26
5	Dongguan	0.40	23	Quanzhou	0.31	41	Shantou	0.26
6	Danzhou	0.39	24	Wenzhou	0.30	42	Jieyang	0.26
7	Dalian	0.38	25	Shantou	0.30	43	Beihai	0.26
8	Haikou	0.37	26	Putian	0.30	44	Maoming	0.25
9	Tianjin	0.36	27	Yangjiang	0.29	45	Dandong	0.25
10	Qingdao	0.35	28	Lianyungang	0.29	46	Cangzhou	0.25
11	Xiamen	0.35	29	Fangchenggang	0.29	47	Panjin	0.25
12	Ningbo	0.34	30	Nantong	0.28	48	Huludao	0.25
13	Chaozhou	0.34	31	Zhangzhou	0.28	49	Jinzhou	0.25
14	Yantai	0.34	32	Weihai	0.28	50	Ningde	0.24
15	Zhoushan	0.33	33	Binzhou	0.28	51	Yingkou	0.24
16	Fuzhou	0.32	34	Taizhou	0.27	52	Qinzhou	0.23
17	Huizhou	0.32	35	Weifang	0.27
18	Dongying	0.31	36	Zhanjiang	0.27

). High-resilience cities, such as Guangzhou (0.60), Shenzhen (0.49), Shanghai (0.48), Zhuhai (0.46), and Dongguan (0.40), are predominantly located in the core areas of the Pearl River Delta (PRD) and Yangtze River Delta (YRD). These cities share several common characteristics: robust economic strength, outstanding urban governance, and mature spatial functions. In contrast, cities with the lowest resilience scores, including Qinzhou (0.23), Yingkou (0.24), Ningde (0.24), and Jinzhou (0.25), are generally situated in peripheral coastal zones or resource-dependent regions, where resilience scores are notably lower. The low resilience of these cities is primarily attributed to three factors. First, limited socio-economic vitality, as indicated by the low proportion of tertiary industries and the limited capacity to mobilize resources effectively. Second, insufficient investment in essential infrastructure, particularly in medical services, drainage systems, and communication networks. Third, high ecological vulnerability, reflected in poor performance in indicators such as green space ratio and ecological connectivity. In general, high-resilience cities exhibit a combination of “economic support, spatial optimization, and governance capability,” while low-resilience cities face multiple challenges related to “resource vulnerability, infrastructure deficiencies, and ecological sensitivity”. This observation provides a solid foundation for developing targeted intervention strategies.

The north–south gradient in resilience across China’s coastal cities reveals notable differences in urban adaptability to storm surge risks. In the northern coastal regions, cities like Qinhuangdao, Tangshan, and Yantai show lower resilience, largely due to their reliance on traditional industries, limited infrastructure investment, and ecological vulnerabilities. These cities face greater challenges in disaster preparedness and recovery, with weak governance and inadequate emergency services exacerbating their vulnerability. In contrast, southern cities such as Guangzhou, Shenzhen, and Zhuhai exhibit higher resilience scores. Their economic strength, advanced infrastructure, and effective governance enable them to better mitigate and recover from storm surge events. The southern cities’ emphasis on ecological protection, disaster prevention systems, and high infrastructure connectivity contributes to their relatively stronger resilience.

Cities in the PRD and YRD regions form the core high-resilience zone. Cities such as Guangzhou, Shenzhen, Shanghai, Ningbo, and Dongguan benefit from strong industrial bases, ample financial resources, and advanced urban governance, which enable them to demonstrate high resilience to storm surge risks (As shown in Table 1). These cities boast well-established disaster prevention systems, high infrastructure connectivity, extensive ecological buffers, and comprehensive emergency evacuation systems. Cities in the Bohai Sea and Liaodong Bay areas, such as Yingkou, Jinzhou, Huludao, Qinhuangdao, and Cangzhou, generally score lower in resilience. These cities face challenges arising from their historical industrial bases, characterized by single economic structures and a lag in the development of the tertiary sector. This hampers their capacity for post-disaster recovery. Additionally, inadequate investment in infrastructure and emergency services, combined with fragile ecological environments, undermines their ability to create effective natural barriers against storm surges. Cities in the Beibu Gulf and Hainan regions show a degree of polarization in their resilience scores. Emerging tourist cities like Haikou (0.36) and Danzhou (0.40) demonstrate relatively better resilience, reflecting progress in ecological protection and infrastructure development. Conversely, cities such as Qinzhou (0.23) and Beihai (0.26) exhibit lower resilience, revealing significant deficiencies in ecological safety and governance response mechanisms.

This study compares the storm surge resilience scores of coastal cities in China with actual economic loss data to evaluate the model’s effectiveness. Guangzhou, with a resilience score of 0.60, faces coastal vulnerabilities, but according to the Guangzhou Municipal Government report (2014), the economic loss caused by Typhoon “Wilm” (Weimaxun) in 2014 amounted to RMB 1.2 billion. Given Guangzhou’s large economic scale, its storm surge losses constitute a relatively low proportion of its GDP, highlighting its strong disaster response and recovery capacity. In contrast, Xiamen, with a resilience score of 0.34, incurred economic losses of RMB 400 million from Typhoon “Meranti” in 2016, as reported by the Xiamen Meteorological Bureau (2016). As a smaller city, Xiamen’s storm surge losses represent a larger proportion of its GDP, reinforcing the trend of higher economic losses in cities with lower resilience scores. Similarly, Shanghai, with a resilience score of 0.484, faced substantial storm surge pressures during Typhoon “Meihua” in 2016, but the economic loss was relatively contained, amounting to RMB 2 billion, according to the Shanghai Municipal Government report (2016). As one of China’s largest economies, Shanghai’s storm surge losses constitute a smaller share of its GDP, demonstrating the city’s robust post-disaster recovery and emergency response capabilities. By contrast, cities with lower resilience scores, such as Qinhuangdao (0.26), Tangshan (0.27), and Yantai (0.27), experience significantly higher storm surge losses relative to their smaller economic bases. For example, in Qinhuangdao, storm surge losses represent a disproportionately high share of its GDP, highlighting the city’s infrastructure and emergency response weaknesses, which lead to more severe economic impacts. These comparisons illustrate a strong correlation between higher resilience scores and lower storm surge loss ratios, further validating the accuracy of our model in assessing urban resilience and post-disaster recovery capabilities.

5.2. Micro-Scale: Coastal Line Resilience Assessment at the Grid Level

5.2.1. Tianjin Coastal Line Resilience Assessment

(1)

Correlation and VIF of Tianjin’s Coastal Resilience Indicators

As shown in Figure 5a, the correlation structure of the micro-scale spatial resilience indicators for Tianjin generally demonstrates moderate to weak correlations, with some indicators exhibiting high coupling. Most correlation coefficients are below 0.4, indicating strong independence and complementarity among the indicators. For example, the correlation between “mixed land-use degree” (T1) and “medical facility density” (T7) is −0.096, highlighting the distinction between spatial structure and emergency facility functionality. Some negative correlations, such as between “surface infiltration capacity” (T3) and “NDVI green space ratio” (T4), show concentrated relationships, with coefficients exceeding 0.99. Notably, T15 (LSI Landscape Index) and T16 (Patch Density—PD) exhibit a near-perfect correlation (r = 0.9999), suggesting that as the number of patches increases, boundary complexity also rises, contributing to landscape fragmentation. This reinforces that areas with aggregated green land patches in Tianjin tend to have better spatial connectivity, enhancing ecological buffering.

Figure 5c shows that most indicators in Tianjin’s micro-scale resilience evaluation have reasonable Variance Inflation Factor (VIF) values, indicating good variable independence. Indicators such as T11 (Elevation) (VIF = 1.39), T12 (Terrain roughness) (VIF = 1.34), T13 (Precipitation) (VIF = 2.26), and T14 (Distance to river) (VIF = 1.67) exhibit significant independence, ensuring a stable data structure. Indicators like T1 (Mixed land-use) (VIF = 3.57), T5 (Shelter facility density) (VIF = 3.25), and T6 (Shelter distance) (VIF = 3.57) show moderate correlation but do not reach problematic collinearity levels. Indicators such as T2 (Road network density) (VIF = 6.84), T10 (Fire facilities distance) (VIF = 4.80), and T9 (Fire facilities density) (VIF = 4.32) exhibit reasonable collinearity with public service facility density and accessibility, maintaining analytical stability.

However, indicators like T15 (LSI), T16 (Patch Density—PD), T17 (Aggregation Index—AI), and T18 (Cohesion—COHESION) show high collinearity, indicating redundancy. Despite T15 slightly exceeding the VIF threshold, it remains integral to the landscape morphology dimension and was retained in the final model. Following VIF analysis and correlation review, T16, T17, and T18 were removed due to high collinearity. The final indicator system integrates multiple dimensions—urban spatial structure, emergency adaptability, ecological exposure, and natural vulnerability—while maintaining strong explanatory power and spatial perceptibility.

(2)

Tianjin Coastal Resilience Indicator Weighting System

As shown in Figure 6, the entropy-weighted values exhibit a strong polarization trend. For instance, T7 (medical facility density) has an entropy weight of 0.370, indicating a high degree of variability across different units and making it a key variable in the heterogeneous resilience space. On the other hand, natural attribute indicators such as T11 (elevation), T12 (terrain roughness), and T13 (annual precipitation) have very low entropy weights (all < 0.01), suggesting that their spatial distribution is relatively homogeneous, with insufficient variability to significantly distinguish resilience differences across regions. In contrast, the expert-based weighting method results in a more balanced distribution, reflecting the systematic judgments of the expert group regarding the intrinsic value of different indicators. T3 (surface infiltration capacity, 0.11), T1 (mixed land-use degree, 0.09), and T4 (green space ratio, 0.09) received higher weights, indicating that experts place significant emphasis on the role of urban green infrastructure and spatial layout in enhancing the city’s disaster response capacity. Meanwhile, indicators such as T6, T8, and T10 received lower weights (all ≤ 0.05), as they are seen as relatively auxiliary or interchangeable in practical operations and planning interventions. In the overall weighting ranking, the top three indicators are: T7 (medical facility density, 0.207), T5 (emergency shelter density, 0.130), and T3 (surface infiltration capacity, 0.082). These three indicators, respectively, reflect the core support for urban resilience in the “rapid response—passive defense—natural buffering” mechanisms. The high weight of T7 highlights the spatial accessibility of medical resources, which has become a prominent weakness in the resilience system and a key governance focus on today’s high-density urban contexts.

The next highest weighted indicators include T1 (mixed land-use degree, 0.093), T14 (distance to river, 0.073), and T4 (NDVI green space ratio, 0.062), which emphasize the importance of land-use elasticity, disaster risk exposure, and green space buffering in overall resilience. Notably, T1 and T4, as built environment variables, not only have high weights but also demonstrate strong independence (low VIF), ensuring stable contributions in the model.

(3)

Tianjin Coastal Resilience Spatial Distribution Characteristics and Hotspot Analysis

As shown in Figure 5 and Figure 6, significant spatial variations in resilience scores across Tianjin’s coastal areas highlight urban resilience differentiation. Hujia Garden Street has the highest resilience score (0.6378), significantly exceeding the average, with strong scores in key social adaptability indicators such as emergency shelter density (T5 = 0.9138), medical facility density (T7 = 1.0), and accessibility to hospitals and fire facilities (T8 = 1.0, T10 = 0.959). Located in the northern part of Jinnan District near the city center, this area benefits from a well-developed built environment, abundant green space (T4 = 0.4789), and high ecological foundation scores (T11 = 0.8327, T12 = 0.9126). It represents a “spatial complexity + functional integration” model with high urban resilience.

The Huanggang Ecological Leisure Residential Area, located at the southern edge of the city center, shows resilience through its strong ecological foundations, including green space ratio (T4 = 0.5948), elevation (T11 = 0.8680), terrain roughness (T12 = 0.6754), and annual precipitation (T13 = 0.7785), along with moderate shelter and medical resources. This area follows a “ecological buffering + moderate human activity” model, characterized by good ecological conditions, developed infrastructure, moderate population density, and compact spatial layout.

In contrast, Lingang Industrial Park, a traditional heavy industrial zone, scores low (below 0.55). Its low emergency shelter (T5) and medical facility density (T7), combined with poor accessibility (T8 and T10), reflect its monofunctional, industrial nature with weak residential support. The area’s extensive hard pavement and terrain modifications lead to low resilience factors like NDVI green space ratio (T4) and surface infiltration capacity (T3). Its low mixed land-use degree (T1) further hampers functional integration and self-regulation. Historically a key industrial area, it is highly exposed and poorly adaptable, making it vulnerable to storm surges and extreme weather. Overall, Figure 4 shows that central and northwestern ecological residential areas exhibit high resilience, while coastal industrial zones and outer suburban areas generally have lower resilience.

5.2.2. Shanwei City Coastal Resilience Evaluation

(1)

Shanwei City Coastal Resilience Indicator System Correlation and VIF

As presented in Figure 5b, the correlation analysis of Shanwei City’s micro-scale resilience indicator system reveals several significant associations, reflecting inherent structural couplings among certain indicators. Notably, Service Facility Indicators: There is a high degree of correlation between indicators related to emergency facilities, such as T5 (emergency shelter density), T6 (hospital density), and T7 (medical facility density). Specifically, the correlation between T5 and T7 reaches 0.76, indicating a strong spatial and functional synergy within the city’s emergency services. This reflects the urban planning philosophy of “multi-functional complexity,” where emergency services are integrated with conventional infrastructure to optimize land use and enhance efficiency. T2 (population density) exhibits a positive correlation with several facility-related indicators (T5–T8), suggesting that densely populated areas tend to attract a higher concentration of resources and infrastructure. This creates a “population-facility-resilience” coupling, where increased population density drives the concentration of essential services and thereby contributes to enhanced urban resilience. However, T2 also shows a negative correlation with T3 (land infiltration rate), illustrating the trade-off between urbanization and ecological systems, as the development of hard infrastructure reduces the space available for ecological functions such as water infiltration. Terrain-related indicators such as T11 (elevation), T12 (terrain roughness), and T13 (annual precipitation) also demonstrate strong correlations. Notably, T11 and T12 show a significant negative correlation of −0.87, which reflects the inherent spatial gradient structure of the terrain. These indicators, representing physical geographical features, help assess the city’s vulnerability to natural disasters and its ecological sensitivity. Indicators T15 to T18 (e.g., Landscape Shape Index (LSI), Patch Density (PD), Aggregation Index (AI), and Cohesion (COHESION)) exhibit nearly perfect collinearity (|r| > 0.99). This redundancy suggests that these indicators largely capture the same underlying landscape complexity characteristics, such as patch boundary complexity and aggregation levels. Retaining all these indicators would likely compromise the model’s robustness due to multicollinearity. Therefore, to ensure model stability and interpretability, the most representative indicator, T15 (LSI), was retained, while the other landscape pattern indicators (T16–T18) were excluded.

(2)

Shanwei City Coastal Resilience Indicator System Weighting

As shown in Figure 5d, T9 (firefighting facility coverage) holds the highest weight at 0.222, significantly surpassing other indicators. This reflects that it exhibits the greatest variability between samples and holds the most information. Additionally, T15 (landscape pattern complexity, LSI) and T11 (elevation) show relatively high entropy values (0.172 and 0.154, respectively), indicating that natural terrain and ecological patterns play a significant role in spatial resilience in cities like Shanwei, which are ecologically sensitive and rich in coastal areas. These variables not only have critical significance in the physical environment but also stand out statistically due to their high internal variability and uneven spatial distribution. Conversely, indicators like T3 (land infiltration rate) and T4 (street connectivity) exhibit very low entropy, indicating minimal variation between samples and high stability, which limits their explanatory capacity for distinguishing spatial units, possibly serving as background support. From the expert scoring dimension, T3 (land infiltration rate) received the highest score (0.11), followed by T4 (street connectivity) and T5 (emergency shelter density). This suggests that from the perspective of urban planners, ecological permeability and infrastructure accessibility are indispensable key factors for building resilience. Despite their low variability, they are considered highly valuable from a strategic planning perspective. In terms of combined weight ranking, T9, T15, and T11 occupy the top three positions, indicating that the resilience construction in Shanwei is more influenced by spatial physical structure and emergency response capacity. The second tier includes public service facilities indicators such as T7 (hospital density) and T5 (emergency shelter density), emphasizing the synergistic support of the core “safety-service-environment” elements in building resilience.

(3)

Shanwei City Coastal Resilience Spatial Distribution and Hotspot Analysis

As seen in Figure 5f and Figure 7, the weighted analysis of Shanwei City’s micro-scale spatial resilience indicator system reveals that areas such as Jiesheng Town (Ciyun Temple region) and the southwestern region of Haifeng County score significantly higher, reflecting superior performance across multiple resilience factors. These areas typically feature moderate terrain, balanced infrastructure, and well-developed service support systems. In the highest-scoring samples, several key indicators scored highly, such as T5 (emergency shelter density), T6 (medical service accessibility), T9 (firefighting facility coverage), and T10 (drainage system completeness), all nearing 1. This indicates strong emergency response capabilities and public service support systems in these regions. Additionally, T7 (hospital density) and T8 (school density) also score highly, while T3 (land infiltration rate) and T4 (street connectivity) perform well, further illustrating that these regions have formed a high degree of synergy in spatial function connectivity and ecological resilience. It is especially noteworthy that some natural environmental indicators, such as T11 (elevation), have slightly lower scores but do not significantly reduce the overall score. This suggests that the infrastructure and service systems have a compensatory effect, which helps maintain resilience despite weaker natural factors. In contrast, areas like Xiangkeng Village and Jiaxi Town score relatively low, with more complex underlying reasons. These areas generally score poorly on indicators like T5, T6, T7, and T13 (green space accessibility), pointing to issues with incomplete public service facilities and poor accessibility. While some regions perform well in ecological resilience indicators (such as T3, T11, and T15), these advantages cannot fully offset the deficits in critical “supporting resilience factors” like shelter, medical services, and firefighting. For instance, despite Xiangkeng Village having certain natural green space advantages and ecological infrastructure, the low scores in key resilience factors like shelter and medical facilities result in a decline in the overall score. Furthermore, these areas may be located in remote or hilly regions, where infrastructure investment and planning are lagging, preventing the full transformation of natural advantages into spatial resilience. Shanwei City’s micro-scale resilience pattern shows a clear “infrastructure-driven difference”: areas with dense infrastructure and public services significantly outperform areas with good natural conditions but weak human settlement environments.

5.3. Impact Mechanisms

Storm surges, triggered by external forces such as high tide superposition, strong winds, sea-level rise, and rapid pressure drops, pose significant risks to urban areas, leading to multi-level, multi-dimensional secondary disaster effects. These impacts can be categorized into two main pathways: macro-systemic shock chains and micro-physical vulnerability chains. According to the World Meteorological Organization (WMO) and the United Nations Office for Disaster Risk Reduction (UNDRR), storm surges exceeding 1.5 m are considered critical in causing substantial coastal flooding. Furthermore, waves above 5 m, as seen during events like Typhoon Haiyan (2013) and Hurricane Sandy (2012), are associated with severe damage to coastal infrastructure and ecosystems.

At the macro level, storm surges cause extensive flooding, disrupting land use, regional economies, population movement, and public services. In coastal industrial zones and densely populated residential areas, infrastructure failures can trigger cascading systemic risks. The concentration of dominant industries and the vulnerability of essential services exacerbate these risks during storm surge events. At the micro level, urban physical environment and infrastructure conditions drive significant differences in storm surge resilience. Factors such as building density, elevation, green space ratio, and impervious surface ratio influence water accumulation and flood retention. Underground pipe networks and drainage systems provide rapid response capabilities, while shelter accessibility and coastal infrastructure protection determine the safety and efficiency of evacuation and emergency responses. Low-lying areas, coastal zones, and regions with informal construction, fragmented green spaces, and aging infrastructure are particularly vulnerable to increased disaster exposure, as shown in Figure 8.

To enhance storm surge resilience, cities can be classified into high, medium, and low resilience zones based on spatial characteristics and comprehensive scores, with tailored response strategies. High-resilience areas, with strong industries and infrastructure, should focus on proactive measures like early warning systems, digital twin technology, ecological coastal buffers, and regional collaboration mechanisms. Medium-resilience areas should strengthen drainage systems, rebuild shelters, and reorganize functional zones. Low-resilience areas, facing severe vulnerabilities, require fundamental improvements in environmental capacity and social governance. Key actions include optimizing green infrastructure, updating aging pipelines, and reclassifying land use based on simulations. Storm surges are spatially and sequentially triggered disasters, shaped by the complex interaction of terrain, facilities, and population. This framework offers both theoretical support and practical guidance for developing spatially tailored resilience strategies for coastal cities.

6. Discussion

6.1. Application of a Macro-Micro Resilience Evaluation Framework for Storm Surge Response

As depicted in Figure 3 and Figure 4, in summary, the resilience of coastal cities in China to storm surges exhibits clear spatial clustering and differentiation. Core cities are highly clustered, secondary cities are more dispersed, and peripheral cities exhibit lower resilience scores. This spatial structure follows a typical “center-periphery” hierarchical pattern, which not only reflects differences in urban development stages and resource endowments but also highlights disparities in adaptive capacity in the face of climate-related disasters. These findings provide practical insights for regional collaborative disaster prevention efforts and targeted resilience improvement strategies.

On the macro scale, factors such as economic strength, infrastructure coverage, and population density are key drivers of urban resilience, highlighting the need to enhance economic and infrastructural capacity to mitigate storm surge risks. On the micro scale, resilience is more intricately shaped by finer spatial elements such as green space ratio, impermeable surface ratio, pipeline network density, and shelter accessibility. These smaller-scale indicators provide a more detailed understanding of resilience at the local level, particularly in how urban spaces function during extreme events. Multivariate linear diagnostics and correlation analyses indicated the presence of strong multicollinearity among certain indicators, suggesting the need for careful consideration in constructing the resilience indicator system to avoid redundancy that could complicate policymaking and effective decision-making.

This study integrates macro and micro analyses to assess urban resilience to storm surges, offering a comprehensive, multi-scale framework for planning and governance. At the macro level, resilience is evaluated based on socio-economic conditions, policy frameworks, infrastructure, and governance capacity, identifying gaps in policy execution and social vulnerabilities across cities. This provides a basis for strategic interventions at the city level. In contrast, the micro analysis focuses on spatial units, highlighting vulnerabilities in areas such as coastal buffer zones, low-lying regions, and ecologically sensitive areas, driven by factors like insufficient green space and increased impervious surfaces. The integration of these two analyses allows for a dual-pathway approach: macro-level interventions aim to improve governance and infrastructure, while micro-level strategies focus on enhancing local resilience through targeted measures like green infrastructure and optimized emergency shelter distribution. This dual approach ensures a comprehensive solution to storm surge challenges, combining city-wide policy insights with localized spatial interventions, thus enhancing resilience and promoting sustainable urban development.

6.2. Coastal Resilience Spatial Characteristics and Differences

(1)

Analysis of Coastal Resilience Spatial Characteristics in Tianjin and Shanwei

The spatial distribution of coastal resilience in Tianjin and Shanwei reveals distinct yet logically consistent structural patterns (As shown in Figure 9). These differences reflect variations in urban development, resource availability, and governance strategies, each shaping the resilience capacity of these cities in response to coastal risks. The resilience space in Tianjin demonstrates a clear “center-periphery” gradient structure. High-resilience areas are primarily concentrated in the urban periphery and ecological corridor intersections, such as Hujia Garden and Huanggang Ecological Leisure Residential Area. These regions represent a core resilience zone, supported by a combination of advanced public service infrastructure and ecological connectivity. In contrast, Shanwei’s coastal resilience spatial pattern places greater emphasis on ecological embedding and natural topography. High-resilience areas, such as Ciyun Temple and the Jiesheng Town coastal region, rely more on the natural environment, with topographical advantages and ecological features playing a fundamental role in resilience. The resilience of these regions is further supported by well-coordinated infrastructure, particularly emergency services.

The resilience disparity between Tianjin and Shanwei is shaped by a complex interplay of socio-economic and ecological factors. While Tianjin’s high industrialization and economic strength should theoretically enhance resilience, its aging infrastructure, heavy industrial concentration, and insufficient green spaces exacerbate vulnerability to storm surges. The city’s dense urban environment and high impervious surface ratios further amplify flood risks, particularly in lower-income areas. In contrast, Shanwei’s resilience is primarily driven by ecological factors, such as coastal protection measures and mangrove restoration, which provide natural defenses against storm surges. However, Shanwei’s socio-economic limitations—marked by low economic diversification and limited financial resources—restrict its capacity to invest in critical infrastructure and disaster management systems. This comparison illustrates that Tianjin’s resilience is constrained by infrastructural and industrial pressures, while Shanwei’s strength lies in ecological strategies, though it is hindered by socio-economic vulnerabilities, highlighting the need for integrated approaches that address both ecological and socio-economic drivers in urban resilience planning.

(2)

Differences in Coastal Resilience Space between Tianjin and Shanwei

Tianjin and Shanwei display substantial differences in their coastal resilience paths and spatial structures, shaped by factors such as urbanization stages, resource endowments, governance frameworks, and planning priorities.

Tianjin’s high-resilience areas are largely driven by the completeness of infrastructure systems and functional aggregation. For example, Hujia Garden, a high-resilience zone, features a high concentration of essential facilities such as medical centers, shelters, and drainage systems. This clustering of infrastructure results in strong system integration and spatial coupling, enhancing resilience. The resilience of Tianjin is thus more “built environment-driven,” with high-density urban development playing a central role. In these areas, technological interventions, such as advanced urban planning and infrastructure management, primarily drive resilience restoration and regeneration. Shanwei’s Resilience Path: Conversely, Shanwei’s coastal resilience is more rooted in ecological factors and natural topography. The city’s resilience is underpinned by its geographical advantages and the coordination between ecological elements and infrastructure. However, the resilience of certain regions is limited by the uneven urban-rural development, with facility-related indicators (such as T6 (medical accessibility) and T7 (hospital density)) scoring lower in more remote or rural areas. This limits the effectiveness of spatial emergency response systems in those regions. Furthermore, Shanwei’s resilience is constrained by its dependence on ecological redundancy, particularly in areas with challenging topographical features. These areas face a “good natural resource but weak service support” structural shortcoming, where natural advantages are not always matched by adequate emergency infrastructure.

Tianjin faces the challenge of reconciling the high-density built environment with the need to preserve ecological space. As urban expansion encroaches on natural areas, the city must increasingly rely on green space systems and landscape reconstruction to enhance its ecological resilience. Tianjin’s resilience strategy is characterized by a high-density, high-function aggregation approach, where technological factors and high facility coverage drive the city’s overall disaster preparedness and recovery capacity. Shanwei, on the other hand, faces challenges stemming from its more ecologically driven resilience strategy, where natural advantages like topography and ecological connectivity are key. However, the city’s infrastructure is less developed compared to Tianjin, especially in more remote areas. Shanwei’s reliance on ecological redundancy and local coordination for resilience enhancement highlights a need for more comprehensive service support, particularly in the emergency response domain. The coastal resilience patterns of Tianjin and Shanwei illustrate clear structural differences that reflect their unique urbanization processes and resource distributions. Tianjin’s resilience is more reliant on infrastructure aggregation and built environment systems, while Shanwei’s resilience is deeply embedded in its natural topography and ecological foundations. Understanding these differences provides a theoretical framework for designing differentiated resilience strategies, guiding urban planning in both cities toward more targeted interventions that address their specific vulnerabilities and resource constraints. This comparison also underscores the importance of adopting context-sensitive resilience strategies that consider local conditions, infrastructure needs, and ecological assets to build adaptive capacity against coastal risks.

6.3. Resilience Enhancement Strategies

(1)

“Macro-Micro” Systemic Enhancement Strategies

To enhance resilience against storm surges, a synergistic approach that integrates both macro-level governance and micro-level physical interventions is essential. This two-way strategy combines top-down policy development with bottom-up community-level improvements, ensuring a comprehensive response to storm surge risks.

The macro-level focus centers on cross-level governance and spatial planning, which are vital for creating a holistic resilience framework. Establishing a comprehensive governance system at the regional level that integrates multiple stakeholders, including emergency response teams, urban planners, resource managers, and policymakers. This system should focus on storm surge early warning, resource allocation, and functional recovery. Redesigning coastal land-use through re-planning and industrial risk restructuring. This includes re-evaluating the population carrying capacity and enhancing the resilience and redundancy within regional spatial layouts. Developing multiple layers of resilience at the regional level, integrating both ecological buffers and urban infrastructure systems, to enhance the capacity to withstand storm surges.

At the micro-level, physical vulnerability mitigation is the primary focus. This includes improving the built environment and infrastructure resilience within smaller units, such as neighborhoods or grids, to provide localized solutions. The strategies comprise functional overlap and redundancy, which involve introducing overlapping functions within infrastructure (e.g., multifunctional green spaces) to enhance resilience against storm surge impacts. They also include intelligent regulation, whereby smart technologies are deployed to enable real-time monitoring and adaptive responses during storm surge events. Micro-Renovation of Resilience involves localized interventions at the 1 km grid level, including upgrading low-lying areas (e.g., creating green sunken plazas), building underground retention pools to manage waterlogging, and optimizing evacuation routes in coastal residential zones. These measures aim to improve evacuation efficiency and stormwater management.

(2)

Differentiated Resilience Governance Strategies for Coastal Spaces

Based on comprehensive resilience scores, coastal areas can be categorized into three types, and a gradient-based strategy can be adopted for progressive governance.

High Resilience Areas (e.g., Hujiayuan): The strategy should focus on “early warning + ecological protection + system redundancy”. Key actions include strengthening ecological coastal belt construction, deploying multi-source monitoring systems, establishing regional emergency linkage mechanisms, and promoting the development of “smart resilience coasts” as demonstration zones.

Medium Resilience Areas (e.g., Huanggang Ecological Leisure Area): This strategy should emphasize “function supplementation + infrastructure densification + spatial decoupling”. Optimizing drainage systems, improving green space connectivity, adding shelter nodes, and diversifying regional functions are crucial steps. The goal is to build a “composite resilience corridor”.

Low Resilience Areas (e.g., parts of the coastal industrial parks): The priority should be “risk control + infrastructure upgrading”. Key actions include renovating aging pipeline systems, conducting risk assessments and relocation plans for key enterprises, and establishing redundant evacuation systems. This should reinforce a “baseline protection” system, as shown in Figure 8.

(3)

Precise Governance Strategies for Coastal Cities with Different Characteristics

Coastal cities exhibit distinct geomorphological, land-use, and functional characteristics. Strategies should be tailored to these characteristics, as shown in Figure 8:

Coastal Industrial Cluster Zones (e.g., Tianjin Port Free Trade Zone, Shanwei Port Industrial Park), Strengthen industrial disaster collaborative assessments and key enterprise relocation plans. Establish a multi-functional coastal protection system, integrating industrial, infrastructure, and ecological coastal elements, forming an “industrial resilience closed loop”. Ecological Mudflats and Wetland Areas, Focus on the restoration of natural buffer zone functions and the construction of ecological dikes. Promote the integration of ecological and disaster reduction objectives, shaping a “natural coastal buffer zone”. Low-lying Coastal Residential Areas or Urban-Rural Junctions, Prioritize addressing issues such as green space fragmentation, aging pipelines, and difficulties in evacuation. Implement “community resilience renewal projects”, such as networking shelter nodes, upgrading old drainage systems, and redeveloping inefficient land plots.

(4)

Zone-based Strategies for “0–5 km, 5–10 km, and 10–15 km” Spatial Impact Zones

Based on the physical impact distance and attenuation effects of storm surges, the coastline can be divided into core impact zones, radiation influence zones, and ecological buffer zones. The following spatial resilience enhancement paths should be proposed, as shown in Figure 10:

0–5 km Core Impact Zone: As the area most directly impacted by storm surges, it should build a “resilience shield” system, including high-standard flood barriers, ecological disaster reduction facilities (e.g., mangrove belts), dense evacuation facilities, and disaster rapid-response command points.

5–10 km Radiation Influence Zone: The focus here is on addressing tidal backflow, secondary flooding, and traffic disruptions. The recommended strategies include sponge city micro-unit renovations, construction of underground comprehensive utility corridors, and optimization of integrated storage and drainage systems. This will strengthen the region’s pressure-bearing and recovery capabilities.

10–15 km Ecological Buffer Zone: This zone should leverage the “green barrier” and “ecological redundancy” for resilience support. Key actions include wetland protection, the development of green disaster-reduction agricultural belts, and integrating ecological conservation with land-use control to form a regional-scale “multi-layered defense line”.

The storm surge resilience strategies in the Netherlands are deeply rooted in a long history of water management, featuring advanced flood protection infrastructure such as dikes, storm surge barriers, and the “Room for the River” program [61]. The Dutch approach emphasizes an integrated, multi-tier governance framework, characterized by close coordination between national, regional, and local authorities, as well as public–private partnerships. In New York, the storm surge resilience strategy has been significantly shaped by the impact of Hurricane Sandy, with a strong focus on adaptive urban planning, incorporating green infrastructure such as coastal wetlands and the development of resilient waterfronts [62]. Notably, New York places a special emphasis on social equity, ensuring that marginalized communities affected by storm surges benefit from resilience-building initiatives [63]. In Southeast Asia, cities like Jakarta and Manila face similar challenges, where urbanization and population growth have exacerbated coastal vulnerabilities to storm surges. Jakarta’s resilience strategy is primarily focused on flood control through the construction of seawalls, canals, and drainage systems, while also integrating ecological solutions such as mangrove restoration along the coast [64]. In contrast, Manila has embraced the “sponge city” concept to enhance rainwater management, combining hard infrastructure with green spaces designed to absorb rainwater and mitigate flood risks [65].

A comparison of the resilience strategies in these regions with those in China’s coastal cities reveals both similarities and differences. While all regions emphasize the improvement of infrastructure and the protection of vulnerable communities, the approaches in the Netherlands and New York place greater emphasis on advanced technological solutions and social equity—areas that could be further emphasized in the strategies of Chinese cities. For instance, in cities such as Guangzhou and Shenzhen, the focus has been on enhancing macro-level infrastructure and policy execution. However, there has been relatively less attention paid to community-driven strategies and ecological solutions, such as mangrove restoration or the development of green spaces. This difference suggests that Chinese cities could benefit from the experiences of the Netherlands and New York by incorporating more community-centered and ecologically integrated resilience strategies into future urban planning and governance frameworks. This study’s findings align with policy frameworks such as China’s “National Climate Adaptation Strategy 2035” and the UNDRR’s “Global Assessment Report” (2023), which emphasize the importance of multi-scale resilience planning. The study’s dual-pathway approach provides a useful framework for operationalizing “adaptive spatial planning” in urban settings, offering clear guidance for policy interventions and spatial governance. By linking resilience assessments to urban planning and disaster mitigation strategies, this research provides actionable insights for enhancing storm surge resilience at both the policy and planning levels.

6.4. Limitations and Uncertainties

While this study provides a comprehensive resilience framework, there are certain limitations that should be addressed. The resolution of data used in this study, especially related to infrastructure and ecological factors, may impact the accuracy of resilience assessments. Additionally, the subjectivity of weight assignments in the resilience indicator system could introduce bias in the evaluation. Furthermore, the study is based on a single-year analysis, which may not fully capture the variability of storm surge impacts over time. Future research should address these limitations by incorporating multi-year data and refining indicator weights to reduce subjectivity. Moreover, uncertainty in the data, such as the variability of storm surge events and socio-economic changes, could affect the robustness of the model. These uncertainties should be acknowledged and further analyzed in future studies.

7. Conclusions

This study develops a dual-scale coupling framework for assessing coastal urban resilience to storm surges, integrating both prefecture-level and 1 km grid analyses. The framework reveals key spatial gradients in resilience, highlighting significant variations between coastal and inland areas.

(1)

Key Findings

Macro-scale resilience tends to decline from coastal belts to inland areas. Coastal zones typically exhibit higher resilience due to better infrastructure and proximity to defense systems. Micro-scale vulnerabilities are particularly noticeable in highly localized areas, such as green space fragmentation, inadequate shelter accessibility, and insufficient tidal defenses. Case studies of Tianjin and Shanwei reveal a divergence between macro-level policy execution and micro-level structural response, emphasizing the need for tailored strategies at different spatial scales.

(2)

Drivers of Resilience

Macro-scale determinants include economic capacity, infrastructure coverage, and population density, which influence broader resilience patterns. Micro-scale determinants involve the green space ratio, impermeable surface ratio, pipeline network density, and shelter accessibility, which shape resilience at more localized levels.

(3)

Governance Strategies

Based on these findings, a set of gradient-based and zone-specific governance strategies is proposed: Regional collaborative governance to enhance coordination between coastal and inland regions. Targeted infrastructure reinforcement in high-risk areas to bolster defenses. Ecological restoration and the implementation of multi-layered defense systems to mitigate storm surge impacts.