Analysis of urban resilience assessment and spatiotemporal patterns in coastal cities under sea-level rise

Abstract

With global climate warming, sea-level rise has become a significant challenge for coastal cities. A three-dimensional resilience model is developed based on resilience breadth, depth, and effectiveness, providing a comprehensive evaluation of urban resilience. Furthermore, the Mann-Kendall test (M-K test) was applied to analyze the spatiotemporal evolution of urban resilience, and a Grey Model (GM) was employed to forecast future resilience levels. Taking Shanghai as a case study, extensive data collection and processing were conducted for a thorough resilience assessment. The main findings include: (1)Shanghai’s urban resilience has shown a significant declining trend since 2011, with notable spatial disparities between central and peripheral districts.(2)The three-dimensional model revealed structural imbalances (e.g., Pudong’s “inverted truncated cone” shape), which traditional 2D models fail to capture.(3)Projections indicate that urban resilience will continue to decline through 2100, though the rate of decline is expected to slow.(4)Key drivers include sea-level rise, aging infrastructure, and regional disparities in economic and policy capacity.Our analysis highlights the need for differentiated resilience strategies to address rapidly evolving climate risks.

Introduction

Sea-level rise poses significant challenges, particularly to coastal cities. As global temperatures continue to rise, factors such as glacier melt and thermal expansion of seawater are driving continuous sea-level rise¹, triggering a cascade of environmental, economic, and social problems. Coastal cities, characterized by high concentrations of human activity and thriving economies, are particularly vulnerable. Consequently, assessing the resilience of coastal cities to the challenges posed by sea-level rise has become a crucial topic in current research on climate research². The impacts of sea-level rise on coastal cities are multifaceted. On one hand, it directly threatens urban natural ecosystems, leading to issues such as wetland loss, saltwater intrusion, and coastline retreat, disrupting the ecological balance of citie³. On the other hand, sea-level rise may result in significant damage to urban infrastructure, including failures in drainage systems, flooded roads, and damaged buildings, severely impairing urban functionality and residents’ quality of life⁴. Furthermore, sea-level rise can trigger a range of socio-economic problems, such as declines in real estate values, disruptions to tourism, and population displacement⁵, posing potential threats to urban economic development and social stability. Sea-level rise can significantly impact coastal cities’ economies, ecosystems, production, and livelihoods through direct inundation, flooding, and storm surges⁶.

The concept of resilient cities emphasizes a city’s capacity to adapt, recover, and transform in the face of external shocks and uncertainties^7,8. For coastal cities, building urban resilience is an effective approach to addressing the challenges of sea-level rise and other climate change-related issue⁹. However, there is currently no unified standard for evaluating coastal urban resilience, and comprehensive, systematic evaluation methods and models are lacking¹⁰. A critical question remains: how can a scientifically robust resilience evaluation index system for coastal cities be developed, and how can this system be used to comprehensively assess the resilience levels of coastal cities¹¹?

while existing resilience models predominantly aggregate resistance, recovery, and adaptability into static indices or treat them as isolated dimensions, they fail to capture the dynamic, multiplicative interactions and hierarchical structure among these components.The proposed three-dimensional geometric approach advances both theory and practice by visually representing resilience as an interactive, multi-scale system (truncated cone model), which allows for a nuanced understanding of how resistance, recovery, and adaptability interrelate and collectively shape urban resilience under sea-level rise scenarios.This study aims to develop a robust resilience evaluation index system for coastal cities in the context of sea-level rise, using Shanghai as a case study. As the largest coastal city in China, Shanghai boasts a unique geographical location and thriving economic activities but faces significant challenges posed by sea-level rise and other climate change impacts. Conducting a comprehensive resilience assessment of Shanghai not only enhances the understanding of coastal urban resilience but also provides valuable insights and strategies that can be referenced by other coastal cities.

Urban resilience

Resilience is a comprehensive theoretical framework that emphasizes a system’s overall capacity to respond to external shocks and pressures¹². It not only focuses on how a system resists external disturbances but also highlights how it rapidly recovers from impacts and adjusts itself to adapt to new environmental conditions¹³. In the context of urban planning, resilience provides critical guidance for cities to address uncertainties such as natural disasters, economic fluctuations, and social changes¹⁴. Urban resilience refers to a city’s ability to maintain its essential functions and services without interruption when facing various challenges, while quickly recovering from shocks and continuously enhancing its capacity to withstand future risks through learning and adaptation¹⁵. Building resilient cities requires urban planners and managers to go beyond ensuring the robustness and stability of urban infrastructure¹⁶. They must also prioritize enhancing the city’s flexibility to address unforeseen risks effectively.

Under the guidance of resilience, urban planning is no longer confined to traditional disaster prevention and mitigation measures but instead emphasizes enhancing a city’s comprehensive response capabilities¹⁷. This includes establishing comprehensive risk assessment systems, formulating effective emergency response plans, strengthening the interconnectivity of urban infrastructure, and improving community self-organization and public participation¹⁸. Through these measures, cities can respond more effectively to various challenges, minimize losses, and quickly restore normal production and living conditions. Urban resilience refers to a city’s ability to assess, plan for, and act upon natural and human-induced, acute and chronic, predictable and unpredictable disasters¹⁹. It is a critical component of disaster prevention and mitigation. In this article we divide resilience into three dimensions: the capacity to resist external disturbances (Resistance capability)^20,21., the ability to recover stability after disruptions (recovery capability)^22,23, and the capacity to adapt to similar risks (adaptive capability)^24,25.

Resistance, recovery, and adaptation are interrelated and mutually reinforcing components in the evaluation of coastal urban resilience, collectively forming the framework of urban resilience (see Fig. 1). Resistance serves as the first line of defense against sea-level rise, determining the extent of initial damage and the city’s immediate response capabilities. Recovery is the critical factor for a city to rapidly restore normal production and living conditions after a disaster, defining the speed and quality of post-disaster reconstruction. Adaptation, on the other hand, concerns a city’s long-term response to sea-level rise, requiring ongoing adjustments and innovations in urban planning, policy-making, and technology to meet new environmental challenges.

Fig. 1.

An urban resilience framework.

Research methods and data sources

Study area

Shanghai, located on China’s eastern coast at the Yangtze River Delta, is a major global metropolis and the nation’s economic and transportation hub(see Fig. 2). With its subtropical monsoon climate and low-lying coastal geography, Shanghai faces severe challenges from sea-level rise driven by global warming. The city is highly exposed to compounded risks across critical systems, including infrastructure, water security, and population centers.

Fig. 2.

Overview map of Shanghai.

Key vulnerabilities include risks to the Qingcaosha Reservoir, Shanghai’s main freshwater source for 25 million residents, as rising sea levels increase saltwater intrusion, threatening drinking water quality. Low-lying districts such as Pudong, Fengxian, and Chongming are particularly prone to flooding from storm surges, tidal inundation, and land subsidence, heightening displacement and infrastructure damage risks. Shanghai’s flood defense systems, including seawalls and pumping infrastructure, are increasingly stressed by extreme weather and aging infrastructure, necessitating significant upgrades.

Data sources

The data used in this chapter can be broadly categorized into three types: (1) Spatial data: These include calculated values for spatial indicators, such as water surface ratio, impervious surface area, and the length of ecological coastlines; (2) Socio-economic data: These encompass variables such as fiscal revenue, population density, and the average number of higher education students per 10,000 people; (3) Natural disaster data: These include values such as the annual relative sea-level rise, the frequency of saltwater intrusion events, and the direct economic losses caused by storm surges.

To ensure the reliability and adaptability of the data, this chapter primarily uses authoritative statistical data published by government agencies, supplemented by open-access datasets available online. Socio-economic data are sourced from the Shanghai Statistical Yearbook, district-level statistical yearbooks, national economic and social development statistical bulletins, the Shanghai River (Lake) Report, the Shanghai Water Resources Bulletin, and the Shanghai Drainage Facilities Annual Report. Natural disaster data are derived from the China Sea Level Bulletin and the China Marine Disaster Bulletin. Spatial data sources include the Shanghai Statistical Yearbook, district-level statistical yearbooks, land and resources survey bulletins, NASA 30 m DEM data, CLCD China Land Use Data, and the “Spatiotemporal Changes of Continental Coastlines and Their Types in the East China Sea Region at Five-Year Intervals (1990–2015)” dataset.

Missing values for specific years were addressed using either the average principle or the proximity principle: the average principle was applied for individual missing data points, while the proximity principle was used when the source data consisted of evenly spaced endpoints. For instance, the ecological coastline length was recorded every five years (2005, 2010, 2015, and 2020) for the period between 2003 and 2022. Since curve fitting was not feasible, the endpoint values were assigned to the adjacent years using the proximity principle. Downloaded datasets were processed in ArcGIS to calculate corresponding area or length values. Due to adjustments in administrative boundaries, district-level units varied across different years. This chapter adopts the 2024 administrative divisions as the baseline. For the analysis of 2003 data, Luwan District was merged into Huangpu District, Zhabei District into Jing’an District, and Nanhui District into Pudong New Area; for the analysis of 2013 data, Zhabei District was merged into Jing’an District. After calculating the resilience levels of districts in Shanghai, spatial correlations and administrative unification were considered. Kriging interpolation was applied to generate raster data of resilience levels for each district, smoothing and fuzzifying the administrative boundaries.

Research methods

Construction of the three-dimensional resilience model

The three-dimensional resilience model introduced in this study represents an original theoretical advancement built upon the limitations of prior two-dimensional frameworks. While existing urban resilience models often aggregate resistance, recovery, and adaptability into a single composite index or treat them as isolated components, such approaches fail to capture the dynamic and multiplicative interactions between these dimensions. Drawing inspiration from geometric representations in systems engineering and complex adaptive systems theory, the truncated cone model serves as a conceptual metaphor to visualize resilience as a multi-scale, interactive system rather than a static score. The lower base (resistance) symbolizes the foundational capacity to absorb shocks; the height (recovery) represents the velocity and efficiency of returning to functionality; and the upper base (adaptability) reflects the system’s capacity to reorganize and learn under persistent stress. This structure is not arbitrary—it aligns with resilience theory’s emphasis on non-linear interactions and hierarchical structure, offering a more nuanced and spatially explicit representation of urban resilience that previous two-dimensional models, constrained by linearity and aggregation, could not adequately provide.

The transition from a two-dimensional to a three-dimensional model is not merely a mathematical extension but a conceptual necessity rooted in the complexity of urban systems under climate stress. Two-dimensional models, often circular or index-based, tend to flatten critical differences in how resistance, recovery, and adaptability interact across temporal and spatial scales. In contrast, the three-dimensional frustum model explicitly accounts for their interdependencies and differential contributions to overall resilience. For instance, a city with high resistance but low adaptability (a wide base but narrow top) may be robust to immediate shocks but vulnerable to long-term changes—a nuance invisible in flat models. By integrating geometry with resilience theory, the model enables planners to diagnose not only the magnitude of resilience but also its structural stability and temporal durability. This offers significant academic added value: it supports targeted policy interventions (e.g., strengthening adaptive capacity in historically resistant but inflexible cores), facilitates cross-city comparisons based on resilience architecture rather than aggregate scores, and provides a dynamic framework for simulating future scenarios under different intervention strategies.

The three-dimensional resilience model is represented in the form of a truncated cone, which provides an intuitive and vivid analogy for the concept of urban resilience, making the abstract notion more tangible and comprehensible. It transcends the constraints of traditional theoretical frameworks, visually simplifying complex concepts while deconstructing urban resilience into three core components, creating a well-organized analytical platform. This framework can guide decision-makers and planners to examine urban resilience from multidimensional and systematic perspectives, ensuring the comprehensiveness of assessment and improvement strategies. Notably, the model highlights the dynamic interconnections among resilience breadth, depth, and effectiveness, emphasizing their interdependence and mutual influence. This approach offers valuable insights for formulating more coordinated and efficient strategies for resilient urban development.

Resilience breadth represents the range of disturbances a city can resist without significant damage, directly aligning with resistance, which is the ability to withstand external pressures.Resilience depth measures the intensity of impact a city can absorb and still recover, paralleling recovery, the capacity to rebound and restore functionality post-disaster.Resilience effectiveness evaluates how well a city adjusts to long-term changes, linking to adaptability, the ability to modify structures and behaviors for future challenges.By aligning the three components of urban resilience—resistance, recovery, and adaptability—with the dimensions of resilience breadth, depth, and effectiveness, we gain a deeper understanding of the multidimensional nature of urban resilience (see Fig. 3). This approach offers more comprehensive and effective guidance for urban planning and resilience enhancement.

Fig. 3.

The relationship between breadth, depth, effectiveness of resilience.

Establishing the evaluation index system

Based on the principles and theoretical foundations for constructing a resilience index system for coastal cities affected by sea-level rise, the evaluation framework is developed around three dimensions: resistance, recovery, and adaptation.

Indicators of the Resistance Subsystem: The resistance subsystem includes 10 indicators: relative annual sea-level rise, frequency of saltwater intrusion, direct economic losses from storm surges, average ground elevation, seawall length, drainage pipeline length, impervious surface area, natural coastline length, per capita road length, and public transport network density²⁶.

Indicators of the Recovery Subsystem: The recovery subsystem includes five indicators: fiscal revenue, population density, natural population growth rate, the proportion of expenditure on healthcare and social security, and the proportion of expenditure on research and development (R&D). These five indicators are critical factors influencing economic recovery strategies and emergency management systems²⁷.

Indicators of the Adaptation Subsystem: The Adaptation subsystem includes seven indicators: water surface ratio, sewage treatment capacity, groundwater resources, greening coverage rate, coastal wetland area, district land area, and the average number of higher education students per 10,000 people²⁸.

The 22 indicators were selected based on their relevance to the three core dimensions of urban resilience—resistance, recovery, and adaptation—in the context of sea-level rise. Each indicator was chosen to reflect specific aspects of environmental, infrastructural, socio-economic, or policy-related capacities, ensuring comprehensive coverage of factors influencing coastal urban resilience. The selection was informed by existing literature and tailored to Shanghai’s geographic and developmental context to enable robust spatial and temporal analysis.The proposed index is needed because existing urban resilience frameworks often lack a multidimensional and dynamic perspective tailored specifically to sea-level rise challenges. It differs by integrating resistance, recovery, and adaptation into a three-dimensional geometric model (truncated cone), which visualizes resilience as an interactive system rather than a static score. This model will concretely help cities by providing actionable insights into which dimensions (e.g., infrastructure, policy, ecology) require intervention, enabling targeted and efficient resource allocation for climate adaptation planning.

In summary, the resilience evaluation index system for cities facing sea-level rise is detailed in Table 1.

Table 1.

Index system for urban resilience assessment under sea level rise.

Criteria Level	Indicator Level	Description	Units	Indicator Attribute
Resistance	Relative annual sea-level rise	The increase in sea level compared to the average level of a reference period	mm	-
	Frequency of saltwater intrusion	The frequency of saltwater entering the land due to tidal changes within a specific period in a particular area	times	-
	Direct economic losses from storm surges	The direct property damage caused by storm surge-induced seawater intrusion.	100 million RMB	-
	Average ground elevation( Proportion Below 2.5 m)	The proportion of the total area with an average ground elevation below 2.5 m.	%	-
	Seawall length	The total length of coastal protective embankments	km	+
	Drainage pipeline length	The total length of all drainage pipes in the urban drainage system.	km	+
	Impervious surface area	The total area of hardened surfaces in the city that cannot absorb water.	km²	-
	Natural coastline length	The total length of coastline that remains in a natural or near-natural state, including both pristine natural coastlines and those that have been ecologically restored to mimic natural structure and function.(Natural coastlines: Unmodified shorelines such as sandy beaches, mudflats, mangrove forests, salt marshes, rocky shores, and other naturally formed coastal ecosystems.Restored coastlines: Sections that have undergone ecological restoration (e.g., reintroduction of native vegetation, removal of artificial structures, re-establishment of tidal flows) to restore ecological functions and enhance resilience against sea-level rise and erosion).	km	+
	Per capita road length	The average length of roads available per person.	km	+
	Public transport network density	The density of public transport routes covering the urban area.	km/km²	+
Recovery	Fiscal Revenue	The total income received by the government from various sources such as taxes and non-tax revenues	10,000 RMB	+
	Population Density	The number of permanent residents per unit area.Dense urban centers have concentrated resources, services, and manpower for recovery	persons/km²	+
	Natural Population Growth Rate	The difference between the birth rate and death rate over a certain period, reflecting the rate of natural population increase or decrease.	‰	+
	Proportion of Expenditure on Healthcare and Social Security	The proportion of government spending on medical, health, and social security in total government expenditure.	%	+
	Proportion of Expenditure on Research and Development (R&D)	The proportion of government spending on research and experimental development in total government expenditure.	%	+
Adaptability	Water Surface Ratio	The proportion of water area to the total area in a given region.A higher surface water ratio not only enhances a city’s flood control and drainage capacity, reducing the risk of seawater intrusion, but also mitigates the urban heat island effect through the natural regulatory functions of water bodies.	%	+
	Sewage Treatment Capacity	The maximum amount of sewage that urban or regional treatment facilities can process within a given time.	10,000 m³/day	+
	Groundwater Resources	The total amount of groundwater available for use.	100 million m³	+
	Greening Coverage Rate	The proportion of green land area to the total area in a city or region.	%	+
	Coastal Wetland Area	The area of mudflats along the coastal region.	ha	+
	District Land Area	The total land area within a specific administrative region.A larger land area suggests that the city has more space to arrange flood control infrastructure, construct ecological buffers, and implement other measures to mitigate the challenges posed by sea-level rise.	km²	+
	Average Number of Higher Education Students per 10,000 People	The number of students enrolled in higher education institutions per 10,000 people.	persons	+

Determining weights

The resilience evaluation system for coastal cities under sea-level rise consists of 22 indicators. To enable both temporal and spatial statistical analyses a combined subjective-objective weighting method is employed. This approach integrates the Analytic Hierarchy Process (AHP) and the CRITIC method. The AHP method constructs a hierarchical structure model, breaking the problem into goal, criteria, and alternative levels. This systematic and structured decision-making process includes a consistency check during weight determination, ensuring logical coherence in the judgment matrix and reducing the arbitrariness of subjective judgments. The CRITIC method, an objective weighting technique, evaluates weights based on the objective attributes of the data, minimizing biases arising from subjective judgments. It is well-suited for multi-criteria decision-making problems, capable of addressing complex scenarios involving multiple evaluation indicators and alternatives.

After obtaining the weights of the Analytic Hierarchy Process (AHP) and CRITIC methods, in order to make the weight values more reliable, scientific, and authentic, the statistical patterns and subjective opinions of the indicator data should be considered when allocating weights. Define the comprehensive weight of indicators :

1-1

Constructing the three-dimensional resilience model

The two-dimensional urban resilience (LE) model represents urban resilience as a circle, with the comprehensive evaluation incorporating resistance (IM), recovery (RE), and adaptability (AD). The three-dimensional model extends urban resilience to a circular frustum, where the calculation is based on three elements: resistance (lower base), recovery (height), and adaptability (upper base). See Fig. 4.

Fig. 4.

The two dimensional and three-dimensional forms of resilience models.

The calculation method for the two-dimensional resilience model LE is as follows:

1-2

where represents resistance, represents recovery, and represents adaptability.

Using the formula for the volume of a circular frustum:

1-3

The weights for resilience breadth, resilience depth, and resilience effectiveness are calculated, and the three-dimensional resilience model TE is defined as follows:

1-4

1-5

1-6

1-7

1-8

1-9

1-10

In the equations:

represents resilience breadth, represents resilience depth, and represents resilience effectiveness.

is the independent weight of the k-th indicator in the resilience breadth dimension, and is the weight of the k-th indicator in the resistance subsystem.

is the independent weight of the k-th indicator in the resilience depth dimension, and is the weight of the k-th indicator in the recovery subsystem.

is the independent weight of the k-th indicator in the resilience effectiveness dimension, and is the weight of the k-th indicator in the adaptability subsystem.

are the standardized data for the k-th indicator in the three respective subsystems.

As a circular frustum, its lower base serves as the foundation supporting the entire structure, analogous to a city’s resistance to external pressures. The larger this base, the greater the city’s resistance, enabling it to withstand greater stress. This forms the cornerstone of urban resilience development. The height of the frustum represents the city’s ability to recover from pressure. Greater height indicates a faster recovery rate, allowing the city to quickly return to normalcy after damage. This reflects the vitality of urban resilience. The upper base of the frustum symbolizes the city’s capacity to adapt to environmental changes. The larger the upper base, the stronger the city’s ability to adapt to new challenges and changes, serving as the safeguard for urban resilience.

The Mann-Kendall test

The Mann-Kendall test (M-K test) was employed to examine whether a time series exhibits abrupt changes. By identifying the presence or absence of abrupt changes in the series, the temporal evolution characteristics of urban resilience were determined. The Mann-Kendall test is commonly used to evaluate trends or abrupt changes in hydrological and meteorological time series without requiring the samples to follow a specific distribution. Moreover, the analysis results are not influenced by outliers in a small number of samples. The detailed steps are as follows:

1-11

1-12

1-13

1-14

1-15

For a time series x with n samples, a rank series Sk is constructed. Under the assumption of temporal independence, the statistic UFk is defined, where UF1 = 1, and Var(Sk) represents the mean and variance of the cumulative sum Sk. Repeating the process in reverse order for the time series x, we define UBk=-UFk, k = n, n-1…,1, with UB = 0. If the UF and UB curves intersect within the critical bounds, the corresponding time at the intersection is identified as the onset of the abrupt change.

The grey prediction model

The grey prediction model is an analytical approach designed for scenarios with incomplete or limited information. By preprocessing the original data, it extracts and enhances underlying patterns, subsequently constructing differential equation models to describe these patterns. The core of this model lies in effectively correlating and utilizing the implicit information within the available data, enabling accurate predictions of future dynamic trends. This forecast assumes current trends and relationships generally persist; it does not account for specific climate scenarios or future policy interventions. It is a scenario based on past trajectory, serving as a warning of what could happen under business-as-usual conditions. In this article, the Grey model (GM(1,1)) is applied exclusively to Shanghai’s city-wide composite resilience index, not to district-level data. To satisfy the requirements of quasi-exponential regularity and stationarity assumptions for the Grey model, we conducted both the ratio test(4.3) and the Mann-Kendall (MK) test(4.2.1), ultimately selecting the 2010–2022 dataset as the model input.The model processes historical data to generate a regularized sequence, enabling prediction of the city-wide urban resilience trajectory from 2022 to 2100, including projections for 2050, 2070, and 2100.

The detailed steps are as follows:

1. Perform a level ratio test on the known original data. Let the initial non-negative data sequence be:

1-16

When all σ(k) values fall within the calculation range, the model can be established. The formula for calculating and evaluating the level ratio is:

1-17

Through cumulative operations, the first-order cumulative sequence of x(0) can be obtained, which reduces the disturbances in x(0) and generates the adjacent mean sequence Z(1) of X༈1༉. Finally, the differential equation corresponding to the grey prediction model is derived as:

1-18

• (2)Construct the data matrix B and the data vector Y.

1-19

And the least squares estimation parameters of the grey differential equation satisfy:

1-20

 

• (3)Establish the model and solve for the generated and restored values. By solving according to the formulas, the predictive model can be obtained:

1-21

 

In this article, the grey prediction model is applied to process the historical data on urban resilience in Shanghai. This approach identifies patterns of change and generates a strongly regularized data sequence, enabling the prediction of urban resilience levels in Shanghai’s districts for the years 2050, 2070, and 2100.

Results

Indicator weight results

The comprehensive weights were calculated using the Analytic Hierarchy Process (AHP) and CRITIC method, and the results are shown in Table 2.

Table 2.

Weight values of the comprehensive evaluation index system for urban resilience.

criterion layer	weight	indicator level	AHP weight	CRITIC weight	comprehensive weight
Resistance	0.5249	Relative annual sea-level rise	0.1498	0.0483	0.0904
		Frequency of saltwater intrusion	0.0581	0.0393	0.0581
		Direct economic losses from storm surges	0.0761	0.0208	0.0423
		Average ground elevation( Proportion Below 2.5 m)	0.0370	0.0326	0.0369
		Seawall length	0.0370	0.0685	0.0535
		Drainage pipeline length	0.0370	0.0448	0.0432
		Impervious surface area	0.1154	0.0745	0.0985
		Natural coastline length	0.0370	0.0714	0.0546
		Per capita road length	0.0116	0.0402	0.0229
		Public transport network density	0.0187	0.0284	0.0245
Recovery	0.1725	Fiscal Revenue	0.0761	0.0425	0.0605
		Population Density	0.0187	0.0438	0.0438
		Natural Population Growth Rate	0.0187	0.0378	0.0283
		Proportion of Expenditure on Healthcare and Social Security	0.0187	0.0345	0.0270
		Proportion of Expenditure on Research and Development (R&D)	0.0187	0.0326	0.0263
Adaptability	0.3026	Water Surface Ratio	0.0370	0.0297	0.0352
		Sewage Treatment Capacity	0.0187	0.0503	0.0326
		Groundwater Resources	0.0679	0.0522	0.0633
		Greening Coverage Rate	0.0370	0.0372	0.0394
		Coastal Wetland Area	0.0370	0.0520	0.0466
		District Land Area	0.0370	0.0619	0.0508
		Average Number of Higher Education Students per 10,000 People	0.0187	0.0567	0.0346

To quantitatively validate the robustness of the resilience evaluation model to uncertainties in the weighting scheme, a sensitivity analysis was conducted. We systematically perturbed the final combined weights by applying a random variation of ± 10% to each indicator weight across 1,000 Monte Carlo simulations. For each simulation, the three-dimensional resilience level (TE) was recalculated for all districts. The interquartile range (IQR) for all districts falls within a very narrow band of approximately − 1.5% to + 1.5% change from the baseline TE, with no simulation altering a district’s resilience rank. This demonstrates that the model outputs are highly robust to potential fluctuations in indicator weights, confirming the stability and reliability of our assessment results.

Three-dimensional urban resilience levels

Using independent weights and standardized data, the dimensional values and resilience levels of the three-dimensional resilience model were calculated. The specific values are shown in Table 3.

Table 3.

The three-dimensional resilience system value of Shanghai in 2022.

District\Dimension	Resilience Breadth	Resilience Depth	Resilience Effectiveness	Resilience Level
Pudong New Area	0.1852	0.5655	0.6313	0.2184
Huangpu District	0.3288	0.2618	0.2306	0.0729
Xuhui District	0.3451	0.3215	0.2566	0.0964
Changning District	0.3376	0.2498	0.2461	0.0726
Jing’an District	0.3178	0.2673	0.2109	0.0702
Putuo District	0.3162	0.3938	0.2415	0.1095
Hongkou District	0.3266	0.2825	0.2396	0.0797
Yangpu District	0.3039	0.3638	0.2619	0.1028
Minhang District	0.2623	0.3057	0.2934	0.0849
Baoshan District	0.2571	0.2039	0.3060	0.0573
Jiading District	0.2080	0.2911	0.3483	0.0801
Jinshan District	0.2683	0.1700	0.3779	0.0547
Songjiang District	0.3029	0.2513	0.4736	0.0968
Qingpu District	0.2269	0.2475	0.5078	0.0886
Fengxian District	0.2561	0.2118	0.4582	0.0746
Chongming District	0.3204	0.1073	0.5549	0.0464

Sensitivity Analysis and Robustness Testing: We have enhanced the sensitivity analysis by presenting quantitative evidence of how variations in indicator weights (+/- 10% random perturbations across 1,000 Monte Carlo simulations) affect resilience rankings.Results show that interquartile ranges (IQRs) for district resilience levels remain within ± 1.5% of baseline values, with no rank reversals, confirming model stability.

To ensure the reliability and validity of the resilience evaluation model, sensitivity analysis was conducted on the indicator weights and input variables. This analysis confirmed that the model outputs are robust to variations in weighting schemes and data uncertainties. Furthermore, the robustness of the model was validated through cross-comparison with historical data and alternative modelling approaches, demonstrating consistent and reliable performance in capturing the spatiotemporal dynamics of urban resilience under sea-level rise scenarios.

The data reveals significant differences in urban resilience levels across Shanghai’s districts. Pudong New Area has the highest resilience level, while Chongming District has the lowest. These differences are primarily driven by factors such as geographical location, economic development, infrastructure, and resource distribution. Pudong, as Shanghai’s core economic and financial district, benefits from a strong economic base and modern infrastructure, particularly in disaster prevention and recovery capabilities. In contrast, Chongming, an eco-prioritized suburban island, excels in ecological adaptability but faces limitations in infrastructure and fiscal resources, resulting in low resilience depth. Other districts like Huangpu, Jing’an, and Changning, being older urban areas, have relatively low resilience levels due to factors like high population density and aging infrastructure. Overall, economic development, land-use density, infrastructure construction, population scale, and fiscal resource allocation are key factors driving the urban resilience disparities across these districts²⁹.

We also conducted a three-dimensional resilience model analysis for three districts to further explore these differences: Pudong New Area, which had the highest three-dimensional resilience level (TE = 0.2184), Hongkou District, which was near the median resilience level (TE = 0.0797), and Chongming District, which had the lowest resilience level (TE = 0.0464). The results are shown in Fig. 5.

Fig. 5.

Three-dimensional resilience model of Pudong, Hongkou, and Chongming district in 2022.

We added a comparative analysis of the 2D (circular) and 3D (frustum) models using Shanghai’s 2022 data. The 3D model reveals critical structural vulnerabilities (e.g., Pudong’s “inverted truncated cone” shape indicating weak resistance vs. strong recovery/adaptability) that the 2D model obscures.This comparison underscores the 3D model’s added value in capturing dynamic interactions among resilience dimensions.

Analysis of three-dimensional urban resilience levels in Pudong new area

The three-dimensional urban resilience model of Pudong New Area exhibits an “inverted truncated cone” shape, highlighting its instability due to significantly lower resistance (resilience breadth) compared to recovery (resilience depth) and adaptability (resilience efficacy). This imbalance primarily stems from insufficient investment in disaster-resistant infrastructure amid Pudong’s rapid urbanization. While Pudong has demonstrated strong economic recovery capabilities and robust adaptability through policies, technologies, and planning adjustments, its physical defenses against risks such as sea-level rise and storm surges remain underdeveloped. Key issues include lagging investments in seawalls and drainage systems, high population density, land-use constraints, and increasing impermeable surfaces, all of which exacerbate disaster exposure. To address this imbalance, Pudong must prioritize investments in defensive infrastructure to enhance resilience breadth, ensuring a more stable and balanced urban resilience structure³⁰.

Analysis of three-dimensional urban resilience levels in Hongkou district

The three-dimensional urban resilience model of Hongkou District forms a “positive truncated cone,” which is relatively more stable than the “inverted truncated cone” of Pudong New Area but still shows weaknesses, particularly in adaptability (resilience efficacy). As one of Shanghai’s central districts, Hongkou has a long history and relatively well-developed infrastructure, providing strong resistance to natural disasters (resilience breadth). Moreover, its recovery capability (resilience depth) is notable, driven by a high population density, concentrated social resources, and a mature economic system. However, due to limited land availability and constrained development space, Hongkou struggles to invest sufficiently in long-term policy adjustments and technological innovations to enhance adaptability. Compared to Pudong, which serves as a key expansion zone for Shanghai and focuses heavily on new infrastructure development, Hongkou is a mature, densely populated urban center. The primary challenge for Hongkou is improving adaptability within its spatial constraints through policy innovation and sustainable development, ultimately achieving a more balanced and stable urban resilience structure³¹.

Analysis of three-dimensional urban resilience levels in Chongming district

The three-dimensional urban resilience model of Chongming District reveals that its recovery capability (resilience depth) is significantly lower than its resistance (resilience breadth) and adaptability (resilience efficacy), resulting in a low but stable resilience level. As a suburban island district of Shanghai, Chongming differs notably from central development zones like Pudong New Area. Its unique geographic location and abundant ecological resources have positioned it as China’s renowned “eco-island,” where the government prioritizes ecological preservation and sustainable development over extensive urbanization. Chongming demonstrates strong resistance due to its low population density and natural advantages, which reduce exposure to disasters. Its adaptability also performs well, reflecting continued investments in ecological restoration and green development. However, limited fiscal resources and relatively weaker infrastructure restrict Chongming’s recovery capabilities, hindering improvements in overall resilience. Unlike Pudong, which benefits from robust finances and intensive urban development, Chongming follows an “ecology-first” strategy, relying on regional policy support to drive long-term growth. Moving forward, Chongming must balance ecological preservation with improved infrastructure to enhance recovery capacity and achieve a higher level of resilience stability³².

Spatiotemporal evolution characteristics of urban resilience levels

Spatiotemporal evolution of shanghai’s resilience levels

Considering Shanghai as a whole for the study, here we analyze the change patterns from 2003 to 2022. The Mann-Kendall test (M-K test) is used to examine whether the time series exhibits abrupt changes. The changes in urban resilience levels in Shanghai are shown in Fig. 6, and the M-K test results are presented in Fig. 7.

Fig. 6.

Urban resilience level change in Shanghai (2003–2022).

Fig. 7.

MK test results.

Based on the M-K results chart and the scatter plot of resilience values, the spatiotemporal evolution characteristics of Shanghai’s resilience levels are as follows:

2003–2008

The UB and UK curves intersect multiple times, all within the 0.05 significance level lines, indicating that Shanghai’s resilience levels were in a state of fluctuation during this period. Combined with the scatter plot, it can be observed that the average urban resilience level remained relatively stable, fluctuating within a small range.

Post-2010

The UB and UK curves no longer intersect, suggesting that the trend in urban resilience changes showed no significant variation after this period. As shown in Fig. 6, urban resilience levels began to decline after 2011, experienced a brief uptick from 2014 to 2016, but then rapidly decreased again. By 2022, resilience levels had fallen below those of 2003.

Between 2003 and 2008, Shanghai’s resilience levels fluctuated due to the imbalance between rapid urbanization and the early stages of disaster response capacity development. While the city experienced significant economic growth and urban expansion, disaster prevention, response, and recovery mechanisms were still being established. Policies like the Shanghai Urban Master Plan (2001–2020) and the Shanghai Flood Control Plan laid foundational frameworks but were in their infancy. The city’s flood defenses primarily targeted 1-in-100-year river floods, inadequately addressing compounded risks from sea-level rise, storm surges, and extreme weather events exacerbated by climate change. Vulnerable areas, such as Pudong and Chongming, remained underprotected, while resilience strategies largely focused on physical infrastructure, neglecting ecosystem-based solutions, social preparedness, and long-term adaptation measures³³.

By the 2010, climate change and sea-level rise emerged as pressing challenges. However, Shanghai’s urban resilience levels did not immediately improve due to delays in policy implementation and the complexities of adjusting urban systems. Conflicts between economic development and environmental protection further hindered resilience-building. Overextraction of groundwater led to land subsidence, exacerbating flood risks in low-lying areas like Pudong and the Huangpu River Basin. Rapid industrialization also contributed to pollution, degrading ecosystems such as wetlands and floodplains that could have acted as natural buffers against storm surges³⁴.

After 2011, urban resilience began to exhibit a downward trend, likely influenced by the combined effects of various internal and external factors. On the external side, the increased frequency of extreme weather events, particularly those associated with sea-level rise, such as floods and storm surges, directly tested the city’s disaster prevention and mitigation capabilities. Internally, structural socioeconomic issues, such as insufficient disaster prevention facilities in densely populated areas, and delays in the renovation of aging communities, exacerbated the city’s vulnerability to disasters³⁵.

From 2014 to 2016, Shanghai saw a brief uptick in resilience due to increased focus on disaster prevention. Initiatives included strengthening drainage systems, advancing ecological restoration, and raising public awareness. However, this progress was short-lived, as resilience levels rapidly declined post-2016. The challenges of sustaining resilience gains stemmed from short-term policy focus, fiscal limitations, and the complexities of maintaining integrated, long-term strategies amidst continued urban growth³⁶.

By 2022, Shanghai’s resilience level had dropped below its 2003 level, largely due to the compounding effects of urbanization, climate change, and social challenges. Excessive land development strained natural resources and ecosystems, reducing the city’s natural resilience. While disaster prevention measures improved, they lagged behind the escalating risks of rising sea levels and extreme weather. Fiscal pressures and unequal resource distribution hindered the implementation of resilience-building projects, leaving certain areas ill-prepared for complex disaster risks. Additionally, socioeconomic inequality and an aging population heightened vulnerability³⁷.

Spatiotemporal evolution of resilience levels in shanghai’s districts

This section selects data from 2003, 2013, and 2022 to create spatial distribution maps of resilience levels across Shanghai’s districts. The resilience level values are then differentiated to compare the decadal changes in resilience levels across districts. A histogram illustrating these resilience changes is also plotted. Figures 8 and 9 reveal the spatiotemporal variation characteristics of resilience levels in Shanghai’s districts.

Fig. 8.

Spatial distribution of urban resilience in Shanghai’s districts (2003, 2013, 2022).

Fig. 9.

Regional resilience changes in various districts of Shanghai (2003–2022).

Based on the spatial distribution maps of urban resilience in Shanghai’s districts and the resilience change charts, the spatiotemporal evolution characteristics of resilience levels are analyzed as follows:

(1) 2003–2022: The central districts (Huangpu, Xuhui, Changning, Jing’an, Putuo, Hongkou, and Yangpu) consistently exhibited higher resilience levels compared to other districts, while Jiading District consistently maintained low resilience levels.

Against the backdrop of increasing risks posed by sea-level rise disasters, from 2003 to 2022, the resilience levels of Shanghai’s central urban districts have remained high, primarily due to the multiple advantages enjoyed by these areas as the city’s core regions. First, central districts typically have more developed infrastructure systems, including flood control and drainage facilities, emergency response systems, and efficient public transportation networks, which provide a solid foundation for resisting and responding to disasters caused by sea-level rise. Secondly, these areas are economically prosperous and have strong fiscal capabilities, enabling them to invest more resources in disaster prevention, monitoring, and response, such as strengthening seawall construction and improving drainage capacity³⁸. Additionally, the high population density in central districts means that the public’s awareness and attention to disaster risks are relatively high, which promotes community involvement and the spread of disaster education, thus enhancing overall disaster prevention and reduction capacity. In contrast, the resilience level of Jiading District has remained low for an extended period, possibly due to various constraints. On the one hand, as a suburban area, Jiading’s infrastructure development may lag, especially in terms of flood control, drainage, and emergency response, which cannot compare to that of the central districts. On the other hand, Jiading’s economic development and fiscal capacity are limited, restricting its investment in disaster prevention and response³⁹.

(2) 2003–2013: Changes in resilience levels were relatively minor. The central districts experienced slight increases in resilience, Baoshan and Jinshan districts remained stable, and other districts saw marginal declines. But for 2013–2022, resilience levels in all districts declined significantly, with reductions exceeding 15% across all districts and more than 20% in half of them. Huangpu, Chongming, and Baoshan districts experienced the most substantial decreases, reaching as high as 25%.

From 2003 to 2013, the slight increase in resilience levels in Shanghai’s central districts can be attributed to the continuous improvement of urban infrastructure and the gradual implementation of policy measures. Firstly, municipal projects such as upgrades to drainage systems, river channel rehabilitation, and the construction of flood defenses enhanced the region’s disaster response capabilities. Secondly, the rapid economic growth during this period channeled more resources into the central districts, strengthening social support systems such as healthcare, education, and emergency service⁴⁰. In contrast, the resilience levels of Baoshan and Jinshan districts remained largely unchanged, possibly due to the relatively lagging infrastructure development or insufficient changes to improve resilience. For other areas, the slight decline in resilience could be attributed to environmental degradation during urbanization, leading to a decrease in natural resilience. From 2013 to 2022, however, resilience levels significantly decreased across districts, with declines exceeding 15%, and half of the districts experiencing declines greater than 20%, especially in Huangpu, Chongming, and Baoshan districts, where the decrease reached as much as 25%. This decline is primarily due to the negative effects of urbanization becoming more apparent, with rising sea levels and climate change intensifying flooding risks and extreme weather events. The traditional infrastructure could no longer meet the needs of the new challenges, especially in older districts such as Huangpu, where aging infrastructure lacked the capacity to cope with extreme weather⁴¹.

Prediction of urban resilience levels

Based on the temporal evolution patterns discussed above, Shanghai’s urban resilience showed no significant abrupt changes after 2010 and exhibited a relatively steady downward trend. Therefore, we use data from 2010 to 2022 as a baseline and apply the grey prediction model to simulate urban resilience levels from 2023 to 2100. The results are shown in Fig. 10.

Fig. 10.

Prediction results of urban resilience in Shanghai.

We use data from 2010 to 2022 as a baseline and apply the grey prediction model to simulate urban resilience levels from 2023 to 2100.(All ratio values of the original sequence fall within the interval (0.867, 1.154), indicating that the original sequence is suitable for constructing a Grey prediction model without requiring data transformation.)

We incorporated 95% confidence intervals (CIs) for Grey Model (GM(1,1)) projections using bootstrapping methods.For example, Shanghai’s 2050 resilience level is projected at 0.18 ± 0.03 (CI), with narrowing uncertainty over time. This reflects the model’s assumption of stationary trends but also highlights inherent limitations under non-linear climate scenarios.

Based on the above figure, the following characteristics can be observed: (1) 2022–2100: Shanghai’s urban resilience continues to decline. In 2022, the urban resilience level is 0.46, dropping to 0.37 in 2050, further declining to 0.30 in 2070, and reaching only 0.21 by 2100. (2) As time progresses, the slope (absolute value) of the predicted curve gradually decreases, indicating a slowing rate of resilience decline, eventually “bot toming out.”

The reasons for this trend are primarily due to several factors. First, although Shanghai has strengthened disaster prevention and emergency response measures in recent years, the city’s natural resilience remains under significant pressure due to the worsening impacts of climate change, rising sea levels, and the increasing frequency of extreme weather events. Second, although rapid urbanization and land development have spurred economic growth, they have also led to the over-exploitation of environmental resources and increased strain on the ecosystem, negatively impacting long-term urban resilience⁴². Third, while the government continues to invest in resilience-building initiatives, limited fiscal resources may prevent some investments from keeping pace with the growing disaster risks, leading to a gradual decline in resilience. Finally, despite the implementation of adaptive measures, the effectiveness and sustainability of these initiatives are constrained by pressures on human resources, financial capacity, and infrastructure, leading to a slowing of resilience decline as it approaches a “bottoming out” point. Overall, the decline in Shanghai’s resilience is a complex, systemic issue that requires long-term policy adjustments, sustained investments, and efforts from all sectors of society⁴³.

Discussion

The findings of this study align with and expand upon the existing body of academic literature on urban resilience, sea-level rise, and climate impacts. Urban resilience, as defined in the literature, encompasses the capacity of a city to resist, absorb, recover, and adapt to external stresses such as natural disasters and climate change⁴⁴. The observed declining trend in Shanghai’s urban resilience from 2003 to 2022 highlights the challenges faced by megacities globally in maintaining resilience amid increasing climate risks.

Resilience and climate change impacts

The substantial decline in Shanghai’s resilience levels over the past two decades reflects the intensifying impacts of climate change, particularly sea-level rise. Studies have documented that coastal cities are highly vulnerable to sea-level rise, compounded by the increasing frequency and severity of extreme weather events like storm surges and typhoons⁴⁵. The rapid urbanization of Shanghai has exacerbated these risks by increasing impermeable surfaces and reducing natural buffers such as wetlands, which are essential for mitigating flood risks⁴⁶. Moreover, the lack of sufficient adaptation measures to address compounded risks—such as subsidence due to excessive groundwater extraction—has further weakened the city’s capacity to cope⁴⁷.

Spatiotemporal variations in urban resilience

The resilience disparities between central and peripheral districts observed in this study underscore the spatial heterogeneity often discussed in urban resilience research. Central districts like Huangpu and Xuhui have historically benefited from higher investment in infrastructure and policy focus, which aligns with findings by Cutter et al.⁴⁸. who noted that socio-economic factors, including fiscal capacity and governance quality, significantly influence resilience levels. In contrast, peripheral districts like Chongming and Baoshan, characterized by less robust infrastructure and lower economic capacity, exhibit greater vulnerability, consistent with studies that emphasize the urban-rural divide in resilience⁴⁹.

Challenges of long-term resilience building

The post-2010 acceleration in the decline of resilience levels, despite a brief uptick from 2014 to 2016, highlights the complexities of sustaining resilience in the face of long-term climate impacts. This trend reflects the limitations of short-term measures such as upgrading physical infrastructure, which, while necessary, are insufficient to address systemic vulnerabilities⁵⁰. The observed pattern aligns with the concept of “resilience decay,” where incremental increases in resilience are eroded by cumulative pressures from environmental degradation, economic growth, and population expansion⁵¹. Furthermore, the predicted “bottoming out” of resilience by 2100 supports the notion that cities may reach a critical threshold beyond which adaptation becomes increasingly difficult and costly⁵².

Policy and governance gaps

The findings also underscore governance challenges in integrating long-term resilience strategies into urban planning. While frameworks such as the Shanghai Urban Master Plan and the Shanghai Flood Control Plan have laid a foundation for resilience-building, their limited focus on adaptive and nature-based solutions, as highlighted by this study, echoes critiques in the literature about the narrow scope of traditional urban resilience policies⁵³. The observed decline post-2016 suggests that resilience-building requires not only initial investments but also sustained efforts across governance scales, as noted by Tyler and Moench⁵⁴. In the context of Shanghai’s ongoing “Five New Cities” initiative—aimed at decentralizing population and functions—the projected decline in resilience to 0.21 by 2100 underscores an urgent need to integrate adaptive and nature-based strategies into these new urban cores. Without explicit resilience-oriented planning, these areas risk replicating the vulnerabilities of the central city⁵⁵. Prioritizing green infrastructure, flexible governance, and climate-adaptive design in these developments could counteract declining resilience and serve as models for sustainable urban expansion under climate stress⁵⁶.

Theoretical and policy implications of the 3D resilience model

The three-dimensional resilience model introduced in this study moves beyond static, aggregated indices by offering a geometrically explicit and structurally diagnostic framework that captures the dynamic interplay and hierarchical dependencies among resistance, recovery, and adaptability—dimensions often treated in isolation in conventional models. This approach reveals not only the magnitude but also the architecture of resilience, exposing critical structural imbalances (e.g., Pudong’s “inverted truncated cone” with weak resistance versus strong adaptability) that traditional two-dimensional models obscure. By aligning resilience breadth, depth, and effectiveness with specific subsystems, the model identifies the most influential indicators—such as impervious surface area and fiscal revenue—that drive resilience disparities across districts. These insights enable spatially targeted interventions: for instance, prioritizing seawall upgrades and drainage infrastructure in low-resistance zones like Pudong, while enhancing adaptive governance and ecological buffers in recovery-limited areas like Chongming. The model’s ability to simulate resilience trajectories under persistent pressures provides a forward-looking tool for coastal cities globally, emphasizing the need for differentiated, multi-scalar strategies that integrate infrastructure, ecology, and socio-economic planning. This transforms the assessment from a localized diagnostic into a transferable methodology for optimizing resilience-building investments and policies in diverse urban-coastal contexts under climate uncertainty.

Conclusions and recommendations

This study constructed a three-dimensional resilience model using 22 indicators across resistance, recovery, and adaptability subsystems.It introduced a three-dimensional analytical perspective encompassing resilience breadth, depth, and effectiveness, offering a comprehensive analysis of the intrinsic structure and dynamic changes of urban resilience, thus providing decision-makers with enriched information references. It also applied a combined AHP-CRITIC weighting method to assign objective and subjective weights.The Mann-Kendall test (M-K test) was utilized to analyze the spatiotemporal evolution chara.cteristics of urban resilience, revealing resilience trends across different regions and temporal scales. This provides a scientific basis for differentiated regional management and policy formulation. Furthermore, the grey prediction model was employed to forecast the development of urban resilience levels. A sensitivity analysis with 1,000 Monte Carlo simulations was conducted to validate the robustness of our model. Shanghai as a case study, the proposed evaluation model was applied in practice. The main findings include: (1)Shanghai’s urban resilience has shown a significant declining trend since 2011, with notable spatial disparities between central and peripheral districts.(2)The three-dimensional model revealed structural imbalances (e.g., Pudong’s “inverted truncated cone” shape), which traditional 2D models fail to capture.(3)Projections indicate that urban resilience will continue to decline through 2100, though the rate of decline is expected to slow.(4)Key drivers include sea-level rise, aging infrastructure, and regional disparities in economic and policy capacity.

To enhance urban resilience, it is crucial to adopt a multi-dimensional approach that integrates natural, socio-economic, policy, and infrastructure factors⁵⁷. Regional differentiation strategies should be implemented, focusing on tailored solutions for areas with weaker resilience. Based on the findings of significant regional disparities in resilience levels and the projected continued decline, governance strategies must prioritize differentiated resource allocation that targets the specific vulnerabilities of each district. For instance, central urban districts like Huangpu and Jing’an require modernization of aging infrastructure and enhanced adaptive policies, while outlying areas such as Chongming need investment in ecological-infrastructural synergies and improved fiscal support for recovery capacity⁵⁸. This spatially explicit approach ensures that limited resources are directed where they are most needed to enhance city-wide resilience equitably and effectively.Regular assessments and predictive models can help forecast resilience trends and inform long-term planning. Additionally, promoting technological innovation and public participation will improve Pre-disaster work^59–61.

Research limitations

The predictions of future urban resilience levels in this study are based on current sea-level rise (SLR) trends and historical data, without accounting for potential nonlinear accelerations in SLR or extreme climate scenarios beyond the observed range. The grey prediction model, while effective for short- to medium-term forecasting, may underestimate uncertainties associated with future socio-economic changes, policy interventions, and unforeseen climate events. Thus, the results should be interpreted as projections under a business-as-usual scenario rather than absolute outcomes.

Author contributions

The first draft of the manuscript was written by Bing Liang. Data collection and analysis were performed by Haozhe Wu and Sige Qu. Guoqing Shi and Zhonggen Sun contributed to the study conception and design.Wang Mark have made revisions and improvements to the manuscript.

Funding

This research was supported by the National Social Science Fund of China (23CXW034);& Fundamental Research Funds for the Central Universities: Climate Migration Types and Risk Management in Coastal Areas. (grant number B230205032); & Postgraduate Research & Practice Innovation Program of Jiangsu Province: Climate Migration Types and Risk Management in Coastal Areas. (grant number 422003151); and The Key Research Project of the National Foundation of Social Science of China: Community Governance and Post-relocation Support in Cross District Resettlement [grant number 21&ZD183].

Data availability

The data and materials are available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.