An Evaluation of Urban Resilience Using Structural Equation Modeling from Practitioners’ Perspective: An Empirical Investigation in Huangshi City, China

Abstract

As urbanization accelerates and climate change intensifies, cities are increasingly facing risks from natural disasters and human activities. Enhancing urban resilience and strengthening cities’ ability to adapt and recover from disasters have become hot topics globally. Although urban resilience evaluation has been studied from different dimensions, the study of urban resilience from a practitioner’s perspective has received less attention. In this study, based on 1464 valid samples of practitioners in Huangshi City, a structural equation model (SEM) was applied to evaluate urban resilience. The evaluation indicators framework was selected from the economy, ecology, society, and infrastructure dimensions. The findings show that (1) the SEM model provides a scientific basis for establishing an index system for the comprehensive evaluation of urban resilience, and the corresponding correlation coefficients help determine the relative contribution of each indicator. (2) Social resilience accounts for the largest proportion of the entire evaluation system, followed by infrastructure resilience, ecological resilience, and economic resilience. (3) Taking Huangshi City as an empirical research case, the results show that the resilience assessment method based on SEM is feasible, with the resilience of Huangshi City showing an upward trend from 2013 to 2022. Finally, some plausible measures to improve urban resilience based on the evaluation results are discussed.

1. Introduction

Cities, as hubs of dense population, are characterized by a significant concentration of human resources, economic activities, and resource consumption [1]. According to a UN-Habitat report, approximately 56% of the global population lived in urban areas in 2021, with this rate expected to increase to 68% by 2050 [2]. Urbanization is particularly rapid in developing countries, driven by rural-to-urban migration for better job opportunities and living conditions [3]. In China, the urbanization rate increased from 17.9% in 1978 to approximately 65% in 2022 [4]. While this unprecedented urbanization has spurred industrial growth and improved quality of life, it has also made cities increasingly vulnerable to risks, particularly in the context of climate change and natural disasters [5,6,7]. In response to these challenges, the concept of urban resilience has become increasingly prominent in urban planning and urban building. The agenda of urban resilience has gained significant attention from international organizations and countries worldwide [8,9,10]. Notable examples include the “Making Cities Resilient” campaign initiated by the United Nations Office for Disaster Risk Reduction (UNDRR) and the “City Resilience Program” launched by the World Bank, etc. [11]. The Chinese government also places great emphasis on resilient cities and has undertaken a series of explorations [12]. For instance, in 2021, the 14th Five-Year Plan and the 2030 Long-Term goals were released, explicitly proposing to “align with new concepts and trends in urban development, carry out pilot demonstrations of urban modernization, and build livable, innovative, smart, green, humanistic, and resilient cities” [13,14].

The term “resilience” comes from the Latin word “resilio”, which initially meant “to return to the original state” [15]. In 1973, Holling, a Canadian ecologist, introduced the concept of resilience in his research [16]. In 1990, the concept of resilience was applied to urban research and expanded into studies on cities and disaster prevention [17]. This research defines urban resilience as a city’s and its sub-systems’ ability to absorb initial damage, reduce the effects of disturbances, adapt to changes, and manage systems that enhance future adaptive capacity, ultimately leading to sustainable urban development [18,19]. Thus, assessing and evaluating urban resilience is an essential foundation for comprehensively understanding and further enhancing the resilience of cities [20].

Evaluating urban resilience could deepen the understanding of its composition while also helping city planners and policymakers establish a clear perception of the current level of urban resilience and determine the next steps for resilience planning [21]. An accurate evaluation of urban resilience bears substantial theoretical and practical implications for urban risk management [22]. Existing research on urban resilience evaluation has concentrated on two main themes. The more widely applied method is quantitative evaluation, which involves establishing an urban resilience indicator system and quantitatively evaluating urban resilience based on subjective judgment from experts’ knowledge and experience or by calculating indicator weights through sample data analysis [23]. A smaller portion of research employs qualitative evaluation, which, in situations where resources are limited, and data are difficult to obtain, evaluates urban resilience by identifying alternative indicators or using methods such as interviews and surveys.

In terms of the urban resilience evaluation method, the suitability of the research approach directly affects the reliability and validity of the evaluation outcomes [24]. The determination of indicator weights is a crucial step in urban resilience assessment. Currently, commonly used methods are primarily categorized into two groups. The first involves experts assessing the relative importance of each evaluation indicator based on their expertise and then deriving indicator weights after comprehensive processing, with examples including the Delphi method [25], the expert investigation method, and the analytic hierarchy process. The second category comprises objective methods that directly determine the weight of each evaluation index based on its characteristics, such as the grey correlation method, the entropy weight method [26], and the multivariate statistical method. While these existing studies have made foundational contributions to assessing and improving urban resilience, it has also been found that these evaluations have primarily relied on the knowledge and experience of scientists, scholars, and experts [27] or relied on the analysis of sample data, yet ignoring the voices of practitioners directly involved in implementation resilient cities [28]. This may result in the allocation of weights to indicators that are not aligned with the actual situation. Therefore, involving various practitioners in decision-making processes fosters comprehensiveness, democracy, and better alignment with the needs of urban residents while also facilitating the integration of resources, knowledge, and experiences [29]. However, despite practitioners being directly involved in building resilient cities and possessing a wealth of practical experience and expertise in urban systems, their perspectives have not yet been taken into account in current research on the assessment of urban resilience.

At present, studies on urban resilience from the practitioners’ experience have mainly focused on identifying the concept of resilience. For instance, Adriana conducted in-depth interviews with international non-governmental organization staff working on disaster resilience in developing countries to understand the practical applications of the concept [27]. George Babington utilized a qualitative research methodology, conducting semi-structured interviews with 20 practitioners in Ghana to investigate their perspectives on urban resilience and its implications for theory and practice [29]. Lorenzo explored the perspectives of European academics and city practitioners on the meanings and principles of urban resilience. His aim was to deepen the understanding of the concept’s evolution and to bridge the gap between theory and implementation [30]. Amer used Qazvin as a study area to investigate urban resilience in urban management, considering the views of urban managers and experts actively involved in crisis management [5]. Therefore, examining urban resilience from the practitioners’ perspective can offer a more accurate depiction of real-world scenarios and enhance the practical relevance of the research.

In summary, this research attempted to bridge the gap between science, policy, and practice by exploring urban resilience evaluation from the practitioners’ perspective. This study took Huangshi City as a research area to investigate the viewpoints of practitioners charged with executing disaster resilience programs and compared them with the perspectives of scholars and policymakers. It aimed to provide empirical support and policy recommendations to enhance urban resilience. In 2013, the Rockefeller Foundation initiated the “100 Resilient Cities (100RC)” project (“100 Resilient Cities Project” was pioneered by the Rockefeller Foundation. The project aims to help cities around the world to withstand external shocks and disasters by developing and implementing resilience plans and providing related technical support, information sharing, and innovative financial and infrastructure development opportunities to cities), which widely promoted the concept of resilient cities globally and attracted broad participation from numerous cities. Huangshi City was selected for the Rockefeller Foundation’s 100RC project and thus is an ideal area for studying urban resilience evaluation from the practitioners’ perspective.

The rest of this paper is organized as follows: Section 2 provides a brief description of the research area and research methodology. Section 3 presents data analysis results of SEM. Section 4 takes Huangshi City as an example to calculate urban resilience. Section 5 concludes the paper by discussing the contributions, recommendations, limitations, and future research directions.

2. Research Methodology

2.1. Research Area

Huangshi City is located in the southeastern part of Hubei Province, on the southern bank of the middle reaches of the Yangtze River, and borders Wuhan City to the west. It functions as a sub-central city within the Wuhan metropolitan area, the largest urban agglomeration in Central China [31]. The city’s location is depicted on national and provincial maps in Figure 1. Following the 2012 flood in Beijing, China intensified its focus on resilient city construction. In response to longstanding risks of flooding, rainstorms and potential threats such as resource depletion and infrastructure vulnerability, Huangshi City has undertaken resilience enhancement measures, focusing on its water, economic, and residential systems. After years of sustained efforts, improvements have been observed in urban water quality, alleviation of urban development pressures, and overall enhancement of city livability. In addition, in 2013, the Chinese government issued the “National Sustainable Development Plan for Resource-Based Cities”. This policy covers 262 RBCs, accounting for 40% of the country’s cities, of which 69 are resource-exhausted cities [32]. Huangshi City was designated by the State Council of China as part of the second batch of resource-exhausted cities in the country. Due to prolonged resource depletion, these cities adopt aggressive resource exploitation methods, leading to imbalanced industrial structures and a deficiency in urban functions. This results in significant economic disparities, low overall development levels, a dominance of extractive industries, minimal scientific and technological investment, weak innovation, and notably inadequate resilience to address internal and external risks and disruptions [33]. Fostering resilience in resource-exhausted cities necessitates a distinct approach from other urban areas [34]. Leveraging these advancements and the special characteristics of Huangshi City, practitioners engaged in the resilient Huangshi City construction project have developed a shared understanding of urban resilience, making Huangshi City an ideal case for studying urban resilience.

2.2. Urban Resilience Index Construction

The establishment of an indicator system is the primary step in assessing urban resilience [35]. The most common urban resilience evaluation systems are based on three categories: characteristics of urban resilience [25,36], stages of urban resilience [37,38,39], and sub-system composition of cities [40,41,42,43]. In China, urban resilience assessment studies have generally adopted a research framework based on urban sub-system composition. These assessments focus on specific types of disaster resilience, typically centered around natural disaster risks such as climate change, geological hazards [44], and floods [45]. Therefore, this paper referred to the research of Zhou and Wang to construct an urban resilience evaluation system based on four sub-systems: economic, ecological, social, and infrastructure. It selected 32 indicators to evaluate urban resilience based on the criteria of breadth, practicability, and scientific and practical applicability [33,34,46], as outlined in Table 1. All indicators were positive indicators, and the data were primarily sourced from the Huangshi Statistical Yearbook, time spanning from 2013 to 2022; missing data for individual years were filled using interpolation.

Economic resilience denotes the capacity of a city’s economic systems and economic agents to effectively and flexibly respond to hazardous shocks and minimize losses [47]. Economic strength and economic stability reflect the ability to provide security for urban systems, and the stronger they are, the faster the city recovers after a crisis. Therefore, economic resilience is primarily demonstrated by the strength and stability of a city’s economy [48].

Ecological resilience is manifested in a city’s ability to regulate itself when facing natural shocks, as well as its ecological carrying capacity and recovery capability in response to human-induced impacts. It emphasizes achieving urban development with minimal ecological costs, thereby promoting the harmonious development of human and ecological systems. A robust urban ecosystem can significantly mitigate the impact of natural disasters and enhance urban resilience [49].

Social resilience emphasizes a city’s ability to cope with external pressures resulting from long-term social changes. It involves the urban system learning from crisis experiences to provide better protection and disaster mitigation methods in the future [50]. Compared with economic resilience, social resilience places greater emphasis on a city’s long-term performance [51].

In urban infrastructure, a comprehensive water supply, communication systems, and transportation systems play a crucial role in mitigating the impact of extreme events. This is the most direct indication of urban infrastructure’s ability to rapidly respond to and recover from disaster impacts during a city crisis, measurable through evacuation, security, and communication capabilities [52,53].

Table 1. An index system for comprehensive evaluation of urban resilience.

Dimension	Indicator	Properties	Justification
Economy	EC1. Per Capita GDP	+	[54,55]
	EC2. Growth Rate of Fixed Assets Investment	+	[54]
	EC3. Total Value of Imports and Exports	+	[55]
	EC4. Disposable Income per Capita	+	[52]
	EC5. Proportion of Tertiary Industry in GDP	+	[33,54]
	EC6. Amount of Actual Utilized Foreign Capital	+	[52]
	EC7. Added Value of High-tech Industries	+	[56]
	EC8. Average Wage of Employed Staff and Workers	+	[34]
	EC9. Total Retail Sales of Consumer Goods	+	[34,57]
	EC10. Number of Industrial Enterprises Above Designated size	+	[52,58,59]
Ecology	EN1. Greening Rate of Built-up Area	+	[54,55]
	EN2. Park Green Area	+	[55]
	EN3. Days Meeting Air Quality Standards	+	[60]
	EN4. Centralized Wastewater Processing Rate	+	[54]
	EN5. Comprehensive Utilization Rate of Industrial Solid Waste	+	[26,54]
	EN6. Harmless Treatment Rate of Waste	+	[1,54,61]
Society	SO1. Urbanization Rate	+	[54,60]
	SO2. Number of Urban Employee Basic Endowment Insurances	+	[60]
	SO3. Number of Urban Employee Basic Medical Insurances	+	[60]
	SO4. Number of People Covered by Unemployment Insurance	+	[55]
	SO5. Proportion of Science and Technology Expenditure to Public Budget	+	[26,57]
	SO6. Number of Health Institutions	+	[34]
	SO7. Number of Health Staff	+	[52,55]
	SO8. Proportion of Social Security and Employment Expenditure to Public Budget	+	[55]
	SO9. Public Library Book Collections	+	[55,59]
	SO10. Proportion of Education Expenditure to Public Budget	+	[34,55]
Infrastructure	IN1. Urban Drainage Pipeline Length	+	[55,62]
	IN2. Number of Private Cars	+	[26,54]
	IN3. Number of Internet Accesses	+	[54,63]
	IN4. Number of Mobile Phone Subscribers	+	[30,34,54]
	IN5. Length of Highways	+	[34,54]

Table 1. An index system for comprehensive evaluation of urban resilience.

Dimension	Indicator	Properties	Justification
Economy	EC1. Per Capita GDP	+	[54,55]
	EC2. Growth Rate of Fixed Assets Investment	+	[54]
	EC3. Total Value of Imports and Exports	+	[55]
	EC4. Disposable Income per Capita	+	[52]
	EC5. Proportion of Tertiary Industry in GDP	+	[33,54]
	EC6. Amount of Actual Utilized Foreign Capital	+	[52]
	EC7. Added Value of High-tech Industries	+	[56]
	EC8. Average Wage of Employed Staff and Workers	+	[34]
	EC9. Total Retail Sales of Consumer Goods	+	[34,57]
	EC10. Number of Industrial Enterprises Above Designated size	+	[52,58,59]
Ecology	EN1. Greening Rate of Built-up Area	+	[54,55]
	EN2. Park Green Area	+	[55]
	EN3. Days Meeting Air Quality Standards	+	[60]
	EN4. Centralized Wastewater Processing Rate	+	[54]
	EN5. Comprehensive Utilization Rate of Industrial Solid Waste	+	[26,54]
	EN6. Harmless Treatment Rate of Waste	+	[1,54,61]
Society	SO1. Urbanization Rate	+	[54,60]
	SO2. Number of Urban Employee Basic Endowment Insurances	+	[60]
	SO3. Number of Urban Employee Basic Medical Insurances	+	[60]
	SO4. Number of People Covered by Unemployment Insurance	+	[55]
	SO5. Proportion of Science and Technology Expenditure to Public Budget	+	[26,57]
	SO6. Number of Health Institutions	+	[34]
	SO7. Number of Health Staff	+	[52,55]
	SO8. Proportion of Social Security and Employment Expenditure to Public Budget	+	[55]
	SO9. Public Library Book Collections	+	[55,59]
	SO10. Proportion of Education Expenditure to Public Budget	+	[34,55]
Infrastructure	IN1. Urban Drainage Pipeline Length	+	[55,62]
	IN2. Number of Private Cars	+	[26,54]
	IN3. Number of Internet Accesses	+	[54,63]
	IN4. Number of Mobile Phone Subscribers	+	[30,34,54]
	IN5. Length of Highways	+	[34,54]

2.3. Method

2.3.1. Mearament Scales

The measurement scales used in this study were adapted from the previously mentioned evaluation index system. The survey questionnaire comprises two sections: one focusing on respondents’ personal information and the other assessing the importance of various indicators. A pilot study involving 15 questionnaires was conducted with practitioners in Huangshi City. They were encouraged to provide feedback if they encountered any difficulties in comprehending or completing the questionnaire. Subsequently, the questionnaire items were reviewed for length, clarity, and language simplicity. Respondents were instructed to rate the indicators of urban resilience based on their practical knowledge and experience. The questionnaire scores are divided into five levels (1~5): “very unimportant”, “unimportant”, “generally important”, “important”, and “very important”. The corresponding measured variables for each latent variable are shown in

Table 2. The latent variables and measured variables of the structural equation model (SEM).

Latent Variables	Observed Variables
Economy	EC1. Per Capita GDP
	EC2. Growth Rate of Fixed Assets Investment
	EC3. Total Value of Imports and Exports
	EC4. Disposable Income per Capita
	EC5. Proportion of Tertiary Industry in GDP
	EC6. Amount of Actual Utilized Foreign Capital
	EC7. Added Value of High-tech Industries
	EC8. Average Wage of Employed Staff and Workers
	EC9. Total Retail Sales of Consumer Goods
	EC10. Number of Industrial Enterprises Above Designated size
Ecology	EN1. Greening Rate of Built-up Area
	EN2. Park Green Area
	EN3. Days Meeting Air Quality Standards
	EN4. Centralized Wastewater Processing Rate
	EN5. Comprehensive Utilization Rate of Industrial Solid Waste
	EN6. Harmless Treatment Rate of Waste
Society	SO1. Urbanization Rate
	SO2. Number of Urban Employee Basic Endowment Insurances
	SO3. Number of Urban Employee Basic Medical Insurances
	SO4. Number of People Covered by Unemployment Insurance
	SO5. Proportion of Science and Technology Expenditure to Public Budget
	SO6. Number of Health Institutions
	SO7. Number of Health Staff
	SO8. Proportion of Social Security and Employment Expenditure to Public Budget
	SO9. Public Library Book Collections
	SO10. Proportion of Education Expenditure to Public Budget
Infrastructure	IN1. Urban Drainage Pipeline Length
	IN2. Number of Private Cars
	IN3. Number of Internet Accesses
	IN4. Number of Mobile Phone Subscribers
	IN5. Length of Highways

.

2.3.2. Data Collection

The survey was administered to practitioners actively engaged in the implementation of the “100 Resilient Cities (100RC)” project in Huangshi City. In this study, a practitioner is defined as an individual possessing technical knowledge and expertise in urban resilience and who is actively involved in implementing the 100RC project [29]. According to the Huangshi Municipal People’s Government Office on the Establishment of Huangshi City Resilient City Construction Leading Group Notice, seven organizations have been identified as key practitioners involved in the construction of resilient cities in Huangshi City [64]. Additionally, purposive and snowball sampling techniques were employed to select research participants [27]. Thus, both the community and practitioners from other organizations were also included.

The survey questionnaires were distributed via online web links using “wen juan xing” from April to July 2024. After a three-month period, a total of 1541 questionnaires were collected from the survey. Two types of questionnaires were excluded: those with excessively short response times and those with highly similar answers. Subsequently, 77 invalid questionnaires were discarded, leaving a total of 1464 valid ones. The detailed characteristics of the sample are shown in

Table 3. Detailed profiles of the sample (N = 1464).

Measure	Item	Number (Person)	Percentage (%)
Gender	Male	755	51.57%
	Female	709	48.43%
Age	Below 30	56	3.83%
	30~44	1389	94.46%
	45~59	10	0.68%
	60 and over	9	0.61%
Education	Junior middle school and lower	8	0.55%
	High school	27	1.84%
	College degree	95	6.5%
	Bachelor’s degree	957	65.51%
	Master’s degree and over	375	25.61%

.

2.3.3. Data Analysis

Structural Equation Modeling (SEM) is a method used to establish, estimate, and test causal relationship models. It offers a detailed analysis of the impact of individual indicators on the overall model and the interrelationships among these indicators [65]. Additionally, SEM has the advantage of handling multiple dependent variables, simultaneously estimating the structure and relationships between factors. By considering the influence of error factors, SEM addresses the limitations of factor analysis and allows for the precise estimation of relationships between observed and latent variables. It can accurately quantify the impact of different indicators on urban resilience [51]. Therefore, this study utilizes SEM to establish a resilience assessment framework for cities and to determine weights, aiming to improve upon shortcomings in previous methods of constructing indicator systems and determining weights. The following are the specific steps for degerming weights:

Step 1. Calculate the Weight of Evaluation Indicators

Structural Equation Modeling (SEM) is a commonly used method to test the relationships between variables in a hypothetical model. Path coefficients represent the strength and direction of the relationships between variables in the model, reflecting the influence of each factor on the outcome variable [45]. Therefore, calculating weights using path coefficients is theoretically reasonable, as this method directly reflects the importance of each factor. In this study, the calculation results from SEM are used to determine the weights of each indicator. The formulas for these calculations are as follows:

$${\mathsf{\lambda }}_i={\displaystyle \frac{\left|{\beta }_i\right|}{\sum _{i=1}\left|{\beta }_i\right|}}$$

(1)

In the equation, λ_i represents the weights of different indicators relative to the overall objective, while β_i denotes the coefficients of direct correlation between latent variables and latent variables or between measured variables and latent variables in the structural equation model (SEM).

Step 2. Data standardization

Due to each indicator having its unique characteristics, affecting its measurement scale and quantitative level, it is necessary to standardize each indicator before conducting the empirical study. This paper uses the Min-max normalization to standardize the raw data. A linear transformation to the original data is applied, mapping the results to the interval [0,1], allowing for comparison based on a unified standard. The processing method is as follows:

$$x_i^{\prime }={\displaystyle \frac{x_i-min\left(x_i\right)}{max\left(x_i\right)-min\left(x_i\right)}}$$

(2)

In the formula, x_i represents the raw data; x_i’ represents the standardized data; max(x_i) and min(x_i) represents the maximum and minimum values of the raw data, respectively.

Step 3. Calculate urban resilience

The following formula is used to determine the urban resilience of Huangshi City and the resilience of its sub-systems.

$$\left|E\right|=\sqrt{{\sum }_{i=1}^m{\lambda }_i{\chi }_i^{\prime }}$$

(3)

In the formula, E represents the comprehensive urban resilience index; λ_i is the weight of the i-element relative to the overall goal; and χ_i′ is the standardized value of the i indicator.

3. The Results of Structural Equation Modeling (SEM)

3.1. Measurement Model

3.1.1. Reliability Analysis

The reliability of scale data are generally assessed using two indicators: the internal consistency coefficient (Cronbach’s α) and composite reliability (CR) [66]. Cronbach’s Alpha coefficients of all variables were analyzed using SPSS 26.0 software, and the result showed that Cronbach’s Alpha coefficients exceeded 0.7, indicating that the scales used in this study have internal consistency and reliability. Additionally, according to the standards, the CR value must reach at least 0.7 to indicate good composite reliability. In this study, the CR values for the dimensions were 0.963, 0.954, 0.969, and 0.956, all of which exceeded 0.7. This indicates that all dimensions have good composite reliability. The results are shown in

Table 4. Test results of Cronbach’s Alpha, AVE, and CR.

Latent Variable	Observation Variable	Standardization Factor Loading	Cronbach’s Alpha	AVE	CR
Economy	EC1	0.668	0.958	0.724	0.963
	EC2	0.843
	EC3	0.860
	EC4	0.838
	EC5	0.889
	EC6	0.884
	EC7	0.893
	EC8	0.874
	EC9	0.877
	EC10	0.859
Ecology	EN1	0.839	0.958	0.775	0.954
	EN2	0.858
	EN3	0.897
	EN4	0.892
	EN5	0.899
	EN6	0.895
Society	SO1	0.874	0.972	0.778	0.969
	SO2	0.853
	SO3	0.854
	SO4	0.894
	SO5	0.908
	SO6	0.898
	SO7	0.856
	SO8	0.879
	SO9	0.867
	SO10	0.91
Infrastructure	IN1	0.896	0.954	0.783	0.956
	IN2	0.895
	IN3	0.884
	IN4	0.851
	IN5	0.884
	IN6	0.897

.

3.1.2. Validity Analysis

After verifying the reliability of the questionnaire, its validity was analyzed. The validity analysis includes four aspects: content validity, structural validity, convergent validity, and discriminant validity. In terms of content validity, the scales used in this study are based on mature scales published in international journals, which ensures the content validity of this study to a certain extent. In terms of structural validity, convergent validity, and discriminant validity, this study was tested using confirmatory factor analysis with Amos 27.0 software. Before this, we used SPSS 26.0 to conduct the KMO test and Bartlett’s spherical test. The KMO value was 0.984, and Bartlett’s Spherical Test Coefficient was 67,268.611, with a significance level of 0.000. Combining the above indicators, it can be concluded that the scale data of this study are well-suited for factor analysis.

The results of the confirmatory factor analysis showed that the CMIN/CF was 4.16, which is within the acceptable range of less than 5 [65], and the RMSEA was 0.044, which is within the excellent range [67]. Additionally, the test results for GFI, AGFI, NFI, IFI, TLI, and CFI all reached an excellent level above 0.9. The values of PNFI and PCFI were 0.794 and 0.799, respectively, both greater than 0.5, within the standard range. These results indicate that the overall model fit is good and that the scale has good structural validity [68]. The test results are shown in

Table 5. Results of the model fitness test.

Fitting Index	Criterion		Actual Value	Result
	Acceptable	Good
CMIN/DF	<5	[2,3)	4.160	Acceptable
GFI	[0.7,0.9)	>0.9	0.939	Good
AGFI	>0.5		0.921	Good
RMSEA	<0.1	<0.07	0.044	Good
NFI	[0.7,0.9)	>0.9	0.975	Good
IFI	[0.7,0.9)	>0.9	0.981	Good
TLI	[0.7,0.9)	>0.9	0.977	Good
CFI	[0.7,0.9)	>0.9	0.981	Good
PNFI	>0.5		0.794	Good
PCFI	>0.5		0.799	Good

.

Given that the urban resilience scale CFA model is suitable, the convergent validity and composite reliability of each dimension of the scale were further tested. The testing process involved calculating the standardized factor loadings of each measurement item on their corresponding dimensions using the established CFA model. Then, using the formulas for AVE and CR, the convergent validity and composite reliability values for each dimension were calculated [69]. According to the standards, the AVE value must reach at least 0.5, and the CR value must reach at least 0.7 to indicate good convergent validity and composite reliability [70]. As shown in Table 3, the AVE values for the four dimensions in this urban resilience scale validity test were 0.958, 0.958, 0.972, and 0.954, all exceeding 0.5. The CR values for the dimensions were 0.963, 0.954, 0.969, and 0.956, all exceeding 0.7. This indicates that all dimensions have good convergent validity.

In the discriminant validity test, the standardized correlation coefficients between each dimension were all less than the square root of the AVE values corresponding to those dimensions, indicating good discriminant validity among the dimensions [71]. The results are shown in

Table 6. Correlation coefficient and discriminant validity of latent variables.

Dimension	Average Value	Standard Deviation	Economy	Ecology	Society	Infrastructure
Economy	3.561	0.654	0.851
Ecology	3.827	0.668	0.658 **	0.880
Society	3.772	0.664	0.682 **	0.736 **	0.882
Infrastructure	3.759	0.688	0.648 **	0.669 **	0.800 **	0.884

Note: The diagonal bold number in the matrix represents the square root of AVE, and the standardized correlation coefficients between latent variables are indicated below. ** p < 0.01.

.

3.1.3. Common Method Bias (CMB)

The responses to this questionnaire are all from professionals involved in urban resilience planning and construction in Huangshi City. The single source of data may cause the research conclusions to be influenced by common method bias [72]. Therefore, this paper used the single-factor confirmatory factor analysis method, incorporating all scale items into the single factor of economic resilience to test for common method bias. As shown in Table 5, compared with the original fitting model, the model fitted using the single-factor confirmatory factor analysis had a poorer fit and did not meet the reference standards. Therefore, there was no serious common method bias in this study. The test results are shown in

Table 7. Test results of the common method bias.

Fitting Index	Single-Factor Confirmatory Factor Analysis Value	Actual Value	Criterion
			Acceptable	Good
CMIN/DF	25.263	4.160	<5	[2,3)
GFI	0.571	0.981	[0.7,0.9)	>0.9
AGFI	0.511	0.921	>0.5
RMSEA	0.122	0.044	<0.1	<0.07
NFI	0.827	0.975	[0.7,0.9)	>0.9
IFI	0.833	0.981	[0.7,0.9)	>0.9
TLI	0.821	0.977	[0.7,0.9)	>0.9
CFI	0.833	0.981	[0.7,0.9)	>0.9
PNFI	0.774	0.794	>0.5
PCFI	0.779	0.799	>0.5

.

3.2. Structural Model

Structural models can identify potential correlations between exogenous latent variables and the causal relationships between exogenous and endogenous variables. The magnitude of the path coefficients reflects the correlation between different latent variables and urban resilience [73]. Following execution, the calculation results of the structural equation model are depicted in Figure 2.

The findings indicate that economic resilience (β = 0.875) exerted a significantly positive effect on urban resilience. Ecological resilience (β = 0.910, t = 28.211, p < 0.001) can positively influence urban resilience. Social resilience (β = 0.995, t = 28.352, p < 0.001) had a significant positive relationship with urban resilience. Infrastructure resilience (β = 0.954, t = 28.679, p < 0.001) is positively correlated with urban resilience. Therefore, it can be considered that these four dimensions all have a positive impact on urban resilience.

Economic resilience is positively correlated with per capita GDP, growth rate of fixed assets investment, total value of imports and exports, disposable income per capita, proportion of tertiary industry in GDP, amount of actual utilized foreign capital, added value of high-tech industries, average wage of employed staff and workers, total retail sales of consumer goods, and number of industrial enterprises above designated size; the path coefficients were 0.668, 0.843 (t = 33.288, p < 0.001), 0.860 (t = 31.549, p < 0.001), 0.838 (t = 30.724, p < 0.001), 0.889 (t = 32.352, p < 0.001), 0.884 (t = 32.096, p < 0.001), 0.893 (t = 32.466, p < 0.001), 0.874 (t = 29.443, p < 0.001), 0.877 (t = 29.731, p < 0.001), and 0.859 (t = 28.211, p < 0.001), respectively; ecological resilience is positively correlated with the greening rate of built-up areas, green park areas, days meeting air quality standards, centralized wastewater processing rate, comprehensive utilization rate of industrial solid waste, and harmless treatment rate of waste; the path coefficients were 0.839, 0.858 (t = 58.569, p < 0.001), 0.897 (t = 47.922, p < 0.001), 0.892 (t = 46.813, p < 0.001), 0.899 (t = 47.535, p < 0.001) and 0.895 (t = 47.316, p < 0.001), respectively; society resilience is positively correlated with the urbanization rate, the number of urban employee basic endowment insurances, the number of urban employee basic medical insurances, the number of people covered by unemployment insurance, the proportion of science and technology expenditure to the public budget, the number of health institutions, the number of health staff, the proportion of social security and employment expenditure to public budget, public library book collections, and the proportion of education expenditure to public budget; the path coefficients were 0.874, 0.853 (t = 70.765, p < 0.001), 0.854 (t = 55.114, p < 0.001), 0.894 (t = 52.181, p < 0.001), 0.908 (t = 50.428, p < 0.001), 0.898 (t = 46.169, p < 0.001), 0.856 (t = 51.693, p < 0.001), 0.879 (t = 50.024, p < 0.001), 0.867 (t = 46.932, p < 0.001), and 0.910 (t = 48.350, p < 0.001); infrastructure resilience is positively correlated with urban drainage pipeline length, the number of private cars, the level of internet access, the number of mobile phone subscribers, the length of highways, and the number of health care beds; the path coefficients were 0.896, 0.895 (t = 56.030, p < 0.001), 0.884 (t = 54.522, p < 0.001), 0.851 (t = 49.995, p < 0.001), 0.884 (t = 54.391, p < 0.001), and 0.897 (t = 51.058, p < 0.001).

4. Empirical Results in Huangshi City

4.1. The Weight of Indicators

Based on the results of the structural equation model, the model’s fit indices are within the standard range. The overall model fits well, indicating that the path coefficients can effectively reflect the relationships between variables. Therefore, the path coefficients can be used to calculate weights, with the weight results for various indicators calculated using Equation (1), as shown in

Table 8. The weights of each indicator.

Dimensions	Dimension–Weight	Indicators	Weight1	Weight2
Economy	0.2300	EC1	0.0787	0.0181
		EC2	0.0994	0.0229
		EC3	0.1014	0.0233
		EC4	0.0988	0.0227
		EC5	0.1048	0.0241
		EC6	0.1042	0.0240
		EC7	0.1052	0.0242
		EC8	0.1030	0.0237
		EC9	0.1034	0.0238
		EC10	0.1012	0.0233
Ecology	0.2400	EN1	0.1589	0.0381
		EN2	0.1625	0.0390
		EN3	0.1699	0.0408
		EN4	0.1689	0.0405
		EN5	0.1703	0.0409
		EN6	0.1695	0.0407
Society	0.2700	SO1	0.0994	0.0268
		SO2	0.0970	0.0262
		SO3	0.0971	0.0262
		SO4	0.1017	0.0275
		SO5	0.1033	0.0279
		SO6	0.1021	0.0276
		SO7	0.0974	0.0263
		SO8	0.1000	0.0270
		SO9	0.0986	0.0266
		SO10	0.1035	0.0279
Infrastructure	0.2600	IN1	0.1688	0.0439
		IN2	0.1686	0.0438
		IN3	0.1666	0.0433
		IN4	0.1604	0.0417
		IN5	0.1666	0.0433
		IN6	0.1690	0.0439

. Dimension–Weight represents the weights of economy, ecology, society, and infrastructure on the overall resilience of the city; Weight1 represents the weights of the different indicators on their corresponding dimensions, with Weight2 representing the weights of the different indicators on the overall resilience of the city.

4.2. The Empirical Evaluation Results of Urban Resilience in Huangshi City

After calculating the weights for each indicator, Equation (2) is used to standardize the raw data of each indicator. The overall resilience of Huangshi City and the resilience of each sub-system from 2013 to 2022 were calculated using Equation (3). The relevant results are presented in

Table 9. The overall resilience of Huangshi City and the resilience of each sub-system.

Year	Urban Resilience	Economy Resilience	Ecology Resilience	Society Resilience	Infrastructure Resilience
2013	0.5103	0.1174	0.1225	0.1378	0.1327
2014	0.5628	0.1295	0.1351	0.1520	0.1463
2015	0.5888	0.1354	0.1413	0.1590	0.1531
2016	0.5763	0.1326	0.1383	0.1556	0.1498
2017	0.6794	0.1563	0.1631	0.1834	0.1767
2018	0.7287	0.1676	0.1749	0.1968	0.1895
2019	0.7899	0.1817	0.1896	0.2133	0.2054
2020	0.8168	0.1879	0.1960	0.2205	0.2124
2021	0.8663	0.1992	0.2079	0.2339	0.2252
2022	0.9015	0.2074	0.2164	0.2434	0.2344

, while the trends of overall resilience and sub-system resilience are depicted in Figure 3. Figure 3 illustrates that urban resilience in Huangshi City showed a consistent upward trend from 2013 to 2022, increasing from 0.5103 in 2013 to 0.9015 in 2022, with an average annual growth rate of 7.66%. Aside from a slight decline between 2015 and 2016, this trend remained consistent. The minor decrease in 2015–2016 may be attributed to the severe rainstorm and flood disasters that affected 20 provinces in China in 2015. Due to the necessity for cities to allocate portions of their social, economic, and ecological resources to mitigate the adverse effects of nationwide natural disasters, the urban resilience level consequently declined during this period.

5. Discussion and Conclusions

5.1. Discussion

From Table 8, it can be seen that, for urban resilience, the weights of the economy, ecology, society, and infrastructure sub-system were 0.230, 0.240, 0.270, and 0.260, respectively, indicating that the weight of the social dimension and infrastructure dimension on urban resilience from the practitioners’ perspective is greater than that of the other two sub-systems, suggesting that the city’s capacity to handle external pressures from long-term social and ecological changes is the primary factor influencing urban resilience. Its influence on urban resilience is greater than that of the economic, ecological, and infrastructure dimensions.

Regarding economic resilience, the added value of high-tech industries and the proportion of the tertiary industry in GDP hold the largest weights, at 0.105 and 0.105, respectively. This highlights the importance of industrial structure on urban resilience and, from another perspective, reflects the significance of transforming and upgrading the industrial structure from primary and secondary industries to tertiary industries. Moreover, by shifting towards technology-intensive and high-value-added industries that possess higher innovation and competitiveness, cities can enhance their economic resilience. Following these are the amount of actual utilized foreign capital, total retail sales of consumer goods, and the average wage of employed staff and workers, with weights of 0.104, 0.103, and 0.103, respectively. Next are the total value of import and export, the number of industrial enterprises above a designated size, the growth rate of fixed asset investment and disposable income per capita, with weights of 0.101, 0.101, 0.099, and 0.099, respectively. The weight of per capita GDP is the smallest at 0.079.

In terms of ecological resilience, the comprehensive utilization rate of industrial solid waste holds the highest weight at 0.1703. This is followed by the days meeting air quality standards, the harmless treatment rate of waste, and the centralized wastewater processing rate with weights of 0.1699, 0.1695, and 0.1689, respectively. These indicators reflect the efficiency of urban environmental cleanliness and demonstrate the adaptability of the ecosystem in the face of disturbances. The two pollution treatment variables reflect the city government’s approach to pollution management. Higher treatment rates decrease the risk of secondary hazards from industrial waste, which can be persistent, insidious and challenging to reverse or even irreversible. Compared with these indicators, the weights for the park green area and the greening rate of built-up areas were the smallest at 0.1625 and 0.1589, respectively. The greening rate of urban built-up areas and the size of park green spaces indicate that the health of a city’s hydrological, geological, and regional climate systems is pivotal in mitigating the adverse effects of meteorological and geological hazards. Additionally, green spaces such as parks serve as emergency contingency sites during disasters.

In the aspect of social resilience, ranked according to their weights, are the proportion of education expenditure to public budget (0.1035), the proportion of science and technology expenditure to public budget (0.1033), the number of health institutions (0.1021), the number of people covered by unemployment insurance (0.1017), the proportion of social security and employment to public budget (0.1000), the urbanization rate (0.0994), public library book collections (0.0986), the number of health staff (0.0974), the number of urban employee basic medical insurances (0.0971), and the number of urban employee basic endowment insurances (0.0970). The indicator of the proportion of education expenditure and the proportion of science and technology expenditure to the public budget shows the value that cities place on these fields.

From the perspective of infrastructure resilience, the weights of the six selected indicators, namely the number of healthcare beds (0.1690), the length of drainage pipelines (0.1688), the number of private cars (0.1686), the length of highways (0.1666), the amount of internet access (0.1666), and the number of mobile phone subscribers (0.1604), do not differ significantly. The length of drainage pipes reflects the ability of drainage facilities to resist and absorb floods, and the importance of drainage pipes is highly related to the geographic location of Huangshi City, which is situated in the middle reaches of the Yangtze River and is prone to flooding. The number of hospital beds reflects the resilience to floods in terms of medical care.

5.2. Conclusions

Evaluating the resilience of cities could objectively assess their capacities to cope with disasters and post-disaster recovery, as well as provide valuable references for their planning, construction, operation, and management. This research directly attempted to evaluate urban resilience by conducting a questionnaire survey with practitioners involved in the “100 Resilient Cities (100RC)” project in Huangshi City, who had not been adequately considered in previous urban resilience assessments. This is a crucial step in understanding how the evaluation of urban resilience is carried out in practice. The key findings are summarized as follows.

First, the urban resilience evaluation index system was established from the four sub-systems of the city, and a questionnaire containing 32 observed variables was designed using structural equation modeling to propose a hypothetical model of the relationships among the indicators. The reliability and validity of the questionnaire were tested using SPSS 26.0 and AMOS 27.0 software, and the results showed that the questionnaire data could be used for model fitting. The model was analyzed, and the results showed that it fit well. The structural equation modeling (SEM) can quantitatively assess the relationships between economic, social, ecological and infrastructure resilience, as well as the overall resilience level of cities. The economic, ecological, social, and infrastructure resilience are positively correlated with urban resilience.

Second, calculating the weights of each evaluation index based on the correlation coefficients obtained from the SEM model provides an innovative method for improving urban disaster resilience. The results show that practitioners have different understandings and perspectives when conducting urban resilience assessments, prioritizing physical urban infrastructure and social protection measures, influenced by their varying experiences and implementation priorities. The research suggests that the impact weights of social resilience and infrastructure resilience on urban resilience are greater than those of economic resilience and ecological resilience.

Finally, based on the results of the model analysis, the weights of the evaluation indicators were calculated. This study then took Huangshi City as an empirical case and concluded that the level of urban resilience in Huangshi City showed an upward trend from 2013 to 2022. In addition, the resilience of individual sub-systems also increased during this decade, verifying the reasonableness of the constructed evaluation indicators and assumptions.

5.3. Theoretical Implications

This study makes several important theoretical contributions. First, it advances the understanding of urban resilience by integrating the perspectives of practitioners, which have been underrepresented in previous research. By highlighting the practical insights and experiences of those directly involved in urban resilience implementation, this study provides a more comprehensive view of urban resilience.

Second, this study enhances the methodological framework for evaluating urban resilience. By employing structural equation modeling (SEM), it offers a robust and systematic approach to assess the complex interrelationships between various factors influencing urban resilience. This methodological contribution can serve as a reference for future studies aiming to quantitatively evaluate urban resilience.

Third, the study extends the application of resilience theory to the context of rapidly urbanizing cities in developing countries. By focusing on Huangshi City, a participant in the “100 Resilient Cities” project, this research provides empirical evidence and insights that are particularly relevant to similar urban contexts experiencing rapid growth and associated challenges, especially for those cities with exhausted resources.

Overall, the study’s theoretical contributions lie in broadening the conceptual understanding of urban resilience, refining evaluation methodologies, and providing context-specific insights that enhance the applicability of resilience theory in diverse urban environments.

5.4. Practical Implications

Based on the evaluation of urban resilience and the voices of practitioners, this study proposes the following practical recommendations to further enhance urban resilience.

Firstly, enhancing and deepening collaboration with practitioners who specialize in developing and implementing strategies for urban resilience is proposed. This involves engaging with a diverse range of stakeholders, including urban planners, architects, engineers, environmental scientists, and policymakers, to create robust frameworks that address the challenges of urbanization, climate change, and disaster risk management. By fostering strong partnerships, sharing knowledge, and leveraging each other’s expertise, cities will become more adaptive, sustainable, and capable of withstanding various shocks and disasters.

Secondly, balancing urban economic development with infrastructure construction is proposed. Effective management of infrastructure investments, combined with strategic economic planning, ensures that cities can support growing populations and businesses while maintaining environmental sustainability. Achieving a balance between infrastructure construction and urban economic development contributes to creating resilient urban environments capable of addressing future challenges and promoting sustainable prosperity.

Thirdly, enhancing the soft power of cities, specifically social resilience, is proposed. This involves strengthening community networks, improving public health systems, and ensuring access to education and social services. By fostering a supportive and inclusive environment, cities can better withstand social and economic disruptions. Additionally, enhancing urban resilience can be achieved through industrial restructuring and digital transformation.

5.5. Limitations and Future Directions

Despite the progress made in evaluating urban resilience, this study has several limitations. First, the accuracy of the assessment results may be influenced by data availability and quality. This study relies on data collected from questionnaires, which may be subject to subjective bias. Additionally, the study sample is limited to Huangshi City, which may restrict the generalizability and applicability of the findings. Future research should consider expanding the sample to include more cities to validate the conclusions of this study.

Second, this study primarily uses quantitative methods for evaluation, while qualitative in-depth interviews and case studies could provide richer contextual information and practical insights. Future research could combine qualitative and quantitative methods to further explore the complexity and diversity of urban resilience.

Finally, this study focuses mainly on the perspectives of practitioners without fully considering the opinions of residents and other stakeholders. Future research should emphasize the participation of multiple stakeholders, integrating the views of different groups to form a more comprehensive framework for urban resilience assessment.

In conclusion, although this study provides new insights into urban resilience assessment, there is still room for improvement. Future research should expand and deepen in areas such as data collection, methodological application, time span, and stakeholder participation to more comprehensively and accurately assess and enhance urban resilience.