Comprehensive Assessment and Obstacle Factor Recognition of Waterlogging Disaster Resilience in the Historic Urban Area

Abstract

As climate change intensifies, cities are experiencing more severe rainfall and frequent waterlogging. When rainfall exceeds the carrying capacity of urban drainage networks, it poses a significant risk to urban facilities and public safety, seriously affecting sustainable urban development. Compared with general urban built-up areas, they demonstrate greater vulnerability to rainfall-induced waterlogging due to their obsolete infrastructure and high heritage value, making it imperative to comprehensively enhance their waterlogging resilience. In this study, Qingdao’s historic urban area is selected as a sample case to analyze the interaction between rainfall intensity, the built environment, and population and business characteristics and the mechanism of waterlogging disaster in the historic urban area by combining with the concept of resilience; then construct a resilience assessment system for waterlogging in the historic urban area in terms of dangerousness, vulnerability, and adaptability; and carry out a measurement study. Specifically, the CA model is used as the basic model for simulating the possibility of waterlogging, and the waterlogging resilience index is quantified by combining the traditional research data and the emerging open-source geographic data. Furthermore, the waterlogging resilience and obstacle factors of the 293 evaluation units were quantitatively evaluated by varying the rainfall characteristics. The study shows that the low flooding resilience in the historic city is found in the densely built-up areas within the historic districts, which are difficult to penetrate, because of the high vulnerability of the buildings themselves, their adaptive capacity to meet the high intensity of tourism and commercial activities, and the relatively weak resilience of the built environment to disasters. Based on the measurement results, targeted spatial optimization strategies and planning adjustments are proposed.

1. Introduction

Urban waterlogging, defined as the excessive surface water accumulation in metropolitan areas during extreme precipitation events or short-duration torrential rains, manifests as delayed drainage discharge in topographical depressions. This hydrological phenomenon exhibits the characteristics of abrupt onset, extensive spatial distribution, and significant destructive potential. Empirical data from the past five decades reveal a global escalation in both frequency and intensity of extreme precipitation events, attributable to accelerated urbanization and climate change dynamics [1]. This trend has precipitated substantial infrastructure damage and economic losses across numerous cities worldwide, establishing urban waterlogging as a critical global disaster risk factor.

The exacerbation of urban inundation risks stems from dual systemic vulnerabilities: the increasing unpredictability of high-intensity rainfall patterns and the inherent limitations of conventional drainage infrastructure [2,3,4]. In response to these challenges, the United Nations Office for Disaster Risk Reduction (UNDRR) prioritizes preventive mitigation strategies through its International Strategy for Disaster Reduction (ISDR). The 2024 Global Assessment Report on Disaster Risk Reduction (GAR2024) Special Edition underscores the imperative of employing scientific methodologies to inform disaster prevention frameworks. Specifically, it advocates for evidence-based risk modeling and resilience-enhancing urban planning to optimize disaster impact reduction and strengthen systemic adaptive capacities.

Urban risk mitigation constitutes a critical component of contemporary urban governance, particularly given the catastrophic socioeconomic repercussions and escalating recovery costs associated with hydrological disasters [5,6,7]. Within this context, urban pluvial flood risk governance has evolved into a multidimensional discipline supported by an advanced research framework. This framework synergistically integrates statistical analysis with hydrological–hydrodynamic modeling and geospatial technologies (remote sensing (RS), geographic information systems (GIS), and global positioning systems (GPS)) [8,9,10]. The academic trajectory has progressively evolved through three distinct phases: (1) foundational hazard identification through hydrological pattern analysis [11,12,13]; (2) hydrodynamic simulation of inundation dynamics using coupled 1D–2D models [14,15,16]; and (3) quantitative risk assessment based on hazard-exposure-vulnerability matrices [17,18,19]. The current scholarly focus has shifted toward integrated risk evaluation systems that holistically address climate change projections and anthropogenic factors, including land-use modifications and infrastructure interdependencies [20,21,22].

Since Holling’s seminal conceptualization of ecological resilience in 1973, this paradigm has subsequently expanded into urban studies through interdisciplinary cross-fertilization [23]. Contemporary urban resilience is operationally defined as the capacity of metropolitan systems to absorb, adapt to, and transform under acute shocks (e.g., pluvial flooding) and chronic stresses, while maintaining essential socioeconomic functions [24,25]. This holistic perspective integrates socio-ecological-technical systems, with particular emphasis on critical infrastructure interdependencies and institutional adaptive governance [26]. Emerging as a pivotal subdomain, urban pluvial flood resilience has gained prominence due to its cascading impacts: flood events between 2010 and 2022 caused USD 327 billion in global damages, accounting for 32% of total flood-related economic losses (World Bank, 2023) [27]. The UNDRR Urban Resilience Index specifically identifies drainage system redundancy (≥70% capacity buffer) and emergency response latency (<30 min) as key performance indicators for waterlogging resilience [28].

Historic urban areas are generally located in the core areas of cities, containing a large number of historical buildings that have been exposed to the outdoors for prolonged periods [29]. As valuable non-renewable historical and cultural resources of cities, these heritage zones cannot be assessed using general urban waterlogging risk evaluation methods due to their advanced age and high heritage value [30]. Ferreira [31] proposed a comprehensive assessment method that combines historical disaster data with GIS technology to evaluate flood risks in Europe’s historic urban areas, providing a reference for risk management in similar regions. Wu [32] employed high-precision DEM grid data and vector data to quantitatively evaluate the relationship between rainfall intensity and ponding depth, developing a refined early warning model for rainstorm waterlogging risks in Beijing’s historical and cultural blocks. However, existing studies still lack in-depth analysis of the special characteristics of historical districts and insufficiently consider such comprehensive factors as the fragility of historical buildings, the impacts of tourism development, and the characteristics of disaster-bearing bodies [33].

In the field of disaster prevention, mitigation, and relief, the proposal of resilience theory has opened up new avenues for enhancing urban flood resilience [34]. Adeyeye [35] developed a resilience collaboration conceptual framework integrating statistical data at macro and micro levels, providing resilience enhancement pathways for historic urban areas. Zhang [36] proposed a disaster-damage curve to quantify resilience thresholds, applicable to risk assessment in historic urban areas. However, most of the studies are based on qualitative analyses and lack quantitative data validation. When quantitative studies are conducted, the indicator system is too universal and lacks considerations for the ontological characteristics of historic buildings and the loss of architectural value. Referring to relevant research at the urban scale, a relatively mature research system for assessing waterlogging resilience or risk based on scenario simulation has been formed. For instance, real-time prediction frameworks such as CA-CNN [37], CA-LSTM early warning systems [38], and MGCNN (Multiple Geographical Units Convolutional Neural Network) [39] all contribute to exploring spatial characteristics and intrinsic mechanisms of waterlogging disasters, thereby offering technical references for historic urban areas as a unique regional context.

In order to address the insufficient exploration of the spatial differentiation driving mechanism of waterlogging resilience in historic urban areas in existing research, this paper takes Qingdao City, China, as an example, and adopts a comprehensive research method combining urban Internet open data and field survey to analyze the interaction between rainfall intensity, the built environment, and the characteristics of people and business and the mechanism of waterlogging disaster in historic urban areas by combining the concept of resilience. The development of a theoretical framework for UNDRR waterlogging resilience holds crucial importance in mitigating waterlogging risks and maintaining the stability of urban blocks. By integrating this framework while accounting for both the integrity and unique characteristics of historic urban areas, we propose a comprehensive indicator system specifically designed to evaluate waterlogging resilience in Qingdao’s historic urban area. In order to obtain accurate waterlogging simulation data under smaller-scale conditions, the CA model is used to construct the drainage system according to the actual arrangement of the pipe network and quantify the floodable depth of the historical urban area by simulating the inundation process under different rainfall scenarios. Finally, the obstacle degree model is introduced to explore the obstacle factors of flood resilience in Qingdao’s historic urban area, to accurately analyze the influencing factors limiting the flood resilience of Qingdao’s historic urban areas, and to propose optimization strategies.

2. Study Area and Data Description

2.1. Study Area

The study area is the historic urban area of Qingdao City, Shandong Province, covering a total area of about 28 km² (Figure 1). Qingdao’s historic urban area straddles the Shinan and Shibei districts and is characterized by a hilly terrain with higher elevations in the east and lower elevations in the west. The annual average rainfall in the study area is 680.5 mm, concentrated primarily in July and August, during which an average of 303.1 mm of rainfall occurs in each month, accounting for 45% of the annual total. The region has a temperate monsoon climate, and its old city exhibits significant maritime climate characteristics due to the direct influence of the marine environment, as well as the combined effects of southeast monsoon winds, ocean currents, and water masses.

The study area focuses on the old town centered around Zhongshan Road, which serves as the birthplace of Qingdao’s century-old culture. This area blends architectural and cultural characteristics from multiple countries and regions, including Germany, Japan, and Korea, forming a unique Qingdao cultural identity. During the 17 years of German colonial rule in Qingdao, a comprehensive water and sewerage system was established, with over 30,000 m of sewage pipes, approximately 40,000 m of rainwater pipes, and 10,000 m of mixed pipes laid during this period. These infrastructures account for about 45% of the current sewerage network in the old city, laying the foundation for Qingdao’s modern urban drainage system.

2.2. Data Sources and Acquisition

First, adhering to the FAIR data principles (Findability, Accessibility, Interoperability, and Reusability) and ISO 19115-1:2014 [40] metadata standards, the methodological workflow began with temporal data acquisition through web API protocols. During the observational period (December 2023, UTC+8), structured keyword queries, including Qingdao, pluvial flooding, extreme precipitation, historic urban fabric, built environment, and hydrological disaster, were systematically conducted via the Baidu Search Engine (https://baidu.com) using advanced Boolean operators (Figure 2). A hybrid analytical framework was implemented, combining machine learning-based feature selection (XGBoost algorithm with SHAP values > 0.75) and expert validation by three GIS specialists (Fleiss’ κ = 0.88). This process enabled rigorous spatiotemporal filtering of 23,780 raw geotagged records from 2011 to 2023, resulting in the generation of flood susceptibility maps.

Second, static spatial datasets were acquired, comprising Points of Interest (POI), drainage networks, building footprints, and road infrastructure. The POI dataset underwent preprocessing procedures including duplicate removal, geometric correction, and coordinate system transformation. Drainage network data, originally in CAD format, were preprocessed through format conversion and attribute enrichment. Building attributes (e.g., construction era, structural typology) were augmented via the Qingdao Historic District’s official WeChat mini-program database.

Ultimately, to address the limitations of web-crawled data accessibility, we conducted a comprehensive field campaign integrating on-site structural assessments and semi-structured interviews. The collected field data underwent systematic processing and were seamlessly integrated with existing geospatial datasets within an integrated geodatabase framework. The synthesized dataset encapsulates the key attributes detailed in

Table 1. Data sources.

Data Type	Data Sources	Data Time
Waterlogging sites	Baidu search engine (http://baidu.com)	31 December 2023
POI	https://lbs.amap.com/	31 December 2023
Drainage network	Qingdao Municipal Bureau of Housing and Urban-Rural Development	31 December 2023
Building	https://lbs.amap.com/ Qingdao Municipal Bureau of Housing and Urban-Rural Development	31 December 2023
Road network	https://www.openhistoricalmap.org/	31 December 2023

.

In particular, it should be noted that for the purpose of analyzing the historic district with the analysis of key buildings in the area, the 38,357 buildings in the study area were divided into two types. One category is, in total, 2300 listed historic buildings, traditional landscape buildings, heritage conservation units, and key buildings with both of these attributes. The other category constitutes newly built commercial, general residential, and military sites that do not need to be assessed as priorities or are currently in a state of abandonment, totaling 36,057 properties, which have been grouped for attribution in this exercise.

To scientifically verify the accuracy and reliability of the waterlogging simulation results, this study constructs a multi-source data cross-validation system and takes the historical waterlogging point data as the core validation basis. By integrating user-generated content from social media and government-released waterlogging point information, we systematically organized the geographical coordinates and other key information of waterlogged points in areas surrounding architectural heritage sites in recent years, establishing a high-precision historical waterlogging point database. Simultaneously, using Geographic Information System (GIS) spatial analysis technology, we conducted spatial overlay analysis and attribute matching between the flooding simulation results (including inundation range and waterlogging depth data) and historical waterlogging point data. The simulation results were spatially evaluated by calculating the spatial overlap between simulated and actual waterlogged points.

3. Methodological Framework

This section may be organized using subheadings. It should provide a concise and precise description of the experimental results, their interpretation, and the experimental conclusions that can be drawn.

3.1. Subsection Overall Framework

This chapter delineates the methodological framework for assessing stormwater inundation resilience in historic urban areas, detailing the operational workflow of the developed assessment methodology (Section 3.1). The initial phase involved the systematic selection of resilience assessment indicators specific to pluvial flood hazards in historic urban areas. Subsequently, inundation depths within the historic district were simulated using Cellular Automata (CA) modeling, with the indicator system operationalized through integration of multi-source datasets (open-access geospatial platforms, statistical archives, and IoT monitoring networks). Normalization procedures were implemented, followed by entropy-weighted prioritization of resilience determinants (Section 3.2). Subsequently, a comprehensive resilience assessment was conducted for the historic district’s pluvial flood vulnerability using the established evaluation framework (Section 3.3). Finally, a Barrier Degree Model is established to systematically identify and analyze critical constraining factors that hinder waterlogging resilience enhancement in historic urban areas, thereby facilitating the formulation of targeted optimization strategies (Section 3.4).

Considering the unique hydrological risk profile and cultural heritage significance of Qingdao’s historic urban area, the methodological framework for assessing waterlogging resilience in historic districts was developed through a synthesis of the existing literature and site-specific contextual factors, integrating spatial configuration and conservation value parameters of the study area (Figure 3).

3.2. Selection of Metrics for Waterlogging Resilience in Historic Urban Areas

Screening metrics and constructing a metric system are the main methods of urban resilience-related metrics in the international arena, with both comprehensive evaluation-type and disaster process-type frameworks being assessed or measured by metrics. The selection of appropriate metrics is very important for urban waterlogging disaster resilience evaluation, which seriously affects the performance of the model, and the model indicator system should aim to completely and accurately reflect the resilience of the urban drainage system embodied in the disaster response processes. Additionally, metrics may be different due to the differences in watershed and terrain characteristics. Given the absence of standardized metrics for assessing resilience in historic urban areas and urban waterlogging disasters, this study addresses dimensional limitations in risk assessment by establishing architectural heritage (the primary protected entities within historic urban areas) as the core determinant of regional waterlogging disaster resilience. Accordingly, the indicator selection process prioritizes synthesizing existing research findings while rigorously adhering to principles of scientific validity, conceptual comprehensiveness, system representativeness, and operational feasibility. Furthermore, the framework explicitly incorporates Qingdao’s historic urban area’s unique socio-spatial characteristics, particularly its hazard mitigation capacity and the compounded socio-cultural-economic vulnerabilities arising from built environment degradation.

In particular, the risk index of waterlogging resilience reflects the likelihood of waterlogging under heavy rainfall scenarios. The risk of waterlogging resilience is assessed in conjunction with ($x111-x112$) indicators such as depth of standing water and density of sensitive points. The vulnerability index of waterlogging resilience reflects the degree of stress to which people and the built environment in the region are exposed to waterlogging hazards as vectors of response to disturbances in a disaster scenario. Combined with the adaptive index of waterlogging resilience, it reflects the regional flood mitigation capacity. The risk index is generally determined by the characteristics of the natural environment, which are difficult to change by perceived means, so measures to improve the level of resilience generally start from changing the vulnerability of the built environment and the population, as well as from improving the conditions of adaptability, which is primarily a response to disasters.

These factors interact with each other in urban waterlogging formation and resilience response, forming an interconnected relational network. Within this context, heightened hazard levels elevate the probability of damage to both affected populations and the built environment, necessitating improvements in systemic resilience alongside enhanced adaptive capacities through emergency response mechanisms and structural reinforcement measures. These elements maintain fundamental significance while operating as mutually independent components within urban waterlogging dynamics and resilience frameworks. Building upon this foundation, this study establishes a systematic hierarchical waterlogging resilience evaluation index system specifically designed for historic urban areas, incorporating their unique characteristics as detailed in

Table 2. Indicator system for evaluating resilience to waterlogging in historic urban areas.

Primary Indicator Layer	Secondary Indicator Layer	Third Indicator Layer	Indicator Type	Weight
Dangerousness (x1)	Disaster exposure dangerousness (x11)	Depth of waterlogging (x111)	−	0.1612
		Density of sensitive sites (x112)	−	0.0315
Vulnerability (x2)	Vulnerability of population characteristics (x21)	Population density (x211)	−	0.0531
		Intensity of crowd concentration (x212)	−	0.0745
	Built-in characteristic vulnerability (x22)	Building density (x221)	−	0.1173
		Construction time (x222)	−	0.0847
		Building grade (x223)	−	0.0483
		Vulnerability of building structures (x224)	−	0.0568
		Business Complexity (x225)	−	0.0796
Adaptability (x3)	Shelter infrastructure (x31)	Accessibility of emergency shelters (x311)	+	0.0411
	Traffic level (x32)	Road density (x321)	+	0.0327
		Road accessibility (x322)	+	0.0344
	Medical rescue (x33)	Distance to hospital (x331)	−	0.0207
	Drainage network characteristics (x34)	Pipeline network node (x341)	+	0.0253
		Density of pipe network (x342)	+	0.0247
	Ecological environment (x35)	NDVI (x351)	+	0.0347
	Engineering measure (x36)	Building renovation and strengthening (x361)	+	0.0794

.

It is worth noting that in the resilience evaluation of waterlogging involved in this study, the scouring of the ground surface and the underlying buildings by heavy rainfall constitutes the primary disaster mechanism. Therefore, the original DEM data remain fundamental to the assessment framework, with slope parameters derived through computational processing of this dataset. In addition, given the substantial numerical disparities and varying measurement scales among indicators, standardization procedures were implemented to mitigate these discrepancies. Both positive and negative indicators underwent normalization using the extremum method’s standardization formulas as specified below.

The positive indicator processing formula is

$$A_i=\left(X_i-X_{min}\right)/\left(X_{max}-X_{min}\right)$$

(1)

The negative indicator processing formula is

$$A_i=\left(X_{max}-X_i\right)/\left(X_{max}-X_{min}\right)$$

(2)

where $A_i$ is the standardized value, $X_i$ is the original value of the indicator value, $X_{max}$ is the maximum value of the same original indicator, and $X_{min}$ is the minimum value of the same original indicator.

For the measurement of the historical urban waterlogging resilience index, an objective assignment method, the entropy weighting method (Entropy Weight Method, EWM) is adopted. To address the bias caused by the data themselves, the EWM method is introduced to calculate the absolute average rate of change, testing the sensitivity of the weights of each indicator to the evaluation results. This reveals the impact of changes in the weights of individual indicators on the resilience of historic urban areas to waterlogging and demonstrates the reasonableness of the entropy weighting method for indicator assignment. Finally, the evaluation index system and weights of waterlogging disaster resilience in historic urban areas are detailed in Table 2. According to the evaluation index system, statistical evaluation unit data are collected separately, and the data are normalized and standardized.

3.3. Simulation of Inundation Features Based on the CA Model

The Cellular Automata (CA) model is a lattice dynamics model with discrete time, space, and state and localized spatial interactions and temporal causality, with the ability to simulate the spatio-temporal evolutionary processes of complex systems. A typical CA model can be expressed as

$$CA= <W,B,S,R>$$

(3)

where $W$ is the geographic cellular space, which can be one-dimensional, two-dimensional, or multi-dimensional space. In this study, based on the elevation data and land use data raster cells, square tuples are used, and the length of the tuple side is determined by parametric rate determination. $B$ is the neighborhood type and the neighborhood of a tuple is the set of the tuple itself and its neighboring tuples; this study adopts the Moore-type neighborhood relationship, which consists of the tuple itself and the surrounding eight tuples to form the neighborhood. It should be noted that in order to satisfy the boundary requirements of the CA model, in this paper, the boundary conditions are set to the boundary of the study area, and the amount of water is not exchanged with the outside world. $S$ is the set of cellular states, which can be represented by two or more states. This study includes static state sets such as elevation and drainage capacity, as well as dynamic state sets such as runoff and water depth. $R$ is the evolution rule, which is the core of the CA model and explains how the state of the cell itself and the surrounding cells determine the state of the cell in the next moment. The evolution rule in this study includes the flow rate production rule and the confluence rule as follows.

In a rainfall event, the total amount of surface runoff from a metric cell can be obtained based on the SCS yield model, and the difference between the total rainfall and the total amount of surface runoff is the total retention of rainwater by that metric cell. Land use differences and water losses in different scenarios need to be considered in the surface flow depth calculation, and the runoff calculation in the area is divided into 14 scenarios, which are shown in

Table 3. Calculation formulae for runoff water depth in different built environment scenarios.

Built Environment Scenarios	Calculation Formula
Buildings (including key buildings)	$D=R-B-Q$
General hardened surfaces	$D=R$
Low-lying hardened surfaces	$D=R-D_{ds}$
Hollow ground	$D=R-I-D_{ds}$
Ground	$D=R-I$
Vegetation and ground	$D=R-P-I$
Vegetation and hollow ground	$D=R-I-D_{ds}-P$
Rainwater well	$D=R-Q$
General hardened surfaces and rainwater well	$D=R-Q$
Low-lying hardened pavement and rainwater well	$D=R-D_{ds}-Q$
Hollow ground and rainwater well	$D=R-I-D_{ds}-Q$
Vegetation and rainwater well	$D=R-P-Q$
Lakes	$D=R-D_{ds}$
Vegetation and lakes	$D=R-D_{ds}-P$

, taking into account the built environment of the historical urban area.

Where $D$ is the depth of surface runoff water after various water loss processes, $R$ is the amount of rainfall, $B$ is the depth of water on the roof of a building that is full of stagnant water, $D_{ds}$ is the depth of puddle storage water, $I$ is the depth of infiltration water at that time of the day, $Q$ is the amount of water drained per unit of time, and $P$ is the depth of water retained by vegetation.

$$P=P_{max}\left[1-\mathrm{e}\mathrm{x}\mathrm{p}\left(-0.046L_{LAI}R/P_{max}\right)\right]$$

(4)

$$P_{max}=0.935+0.498L_{LAI}+0.00575{L_{LAI}}^2$$

(5)

$$B=ciA$$

(6)

$$I=I_c+\left(I_0{-I}_c\right)\mathrm{e}\mathrm{x}\mathrm{p}\left(-kt\right)$$

(7)

where $P_{max}$ is the maximum retention volume water depth, $L_{LAI}$ is the leaf area index, $c$ is the runoff coefficient, $i$ is the rainfall intensity, $A$ is the roof area, $I_c$ is the steady infiltration rate, $I_0$ is the initial infiltration rate, and k is the infiltration parameter of the infiltration curve. c is taken from reference [41], and $I_c$, $I_0$, and $k$ are taken from reference [42].

Three basic rules are identified based on the hydrodynamic mechanisms involved in the surface convergence process, as follows:

(1)

Determine whether to perform water allocation by selecting a central cell within the study area whose elevation difference $\Delta h_i$ is greater than 0 with other cells i in its neighborhood and if the water depth of this cell is greater than 0 and there a water level elevation in the domain exists that is less than that of this cell (the sum of the water depth and the ground surface elevation), then water allocation will be performed;

(2)

Determine the direction of water flow. Water distribution exists only between the central cell and the cell with the lowest water level elevation in the domain, and if more than one cell with the lowest water level elevation is included, water flow in more than one direction is calculated;

(3)

The final determination of the amount of water to be allocated. The amount of water to be allocated is subject to three conditions: the amount of water in the central cell, the difference in water level elevation of the cell and the velocity of the water flow, the amount of water allocated to the downstream cell by the central cell is not higher than its own amount of water, and water allocation stops when the water level elevation of the central cell and the downstream cell are on a par with the water level elevation of the downstream cell.

As the historic urban areas are generally old urban areas with a high degree of completion and environmental consistency, this paper tries to shorten the time step as much as possible in order to guarantee the accuracy of the model calculations, which reduces the efficiency of the model calculations, assuming that the rainfall runoff in the unit time step is a steady flow, and Manning’s formula is used to calculate the flow rate between cells [43], with simulated between cells of size 30 m and a time step of 1 min. In particular, as the historical urban areas are generally small in size, the rainfall runoff flow rate is generally a fixed value without spatial differentiation, which is calculated by the following formula:

$$v=h^{2/3}{s}^{1/2}/r$$

(8)

where $v$ is the velocity of water flow (m/s), $h$ is the depth of water (m), $s$ is the slope, and $r$ is Manning’s roughness coefficient.

3.4. Formatting of Mathematical Components

Drawing on the principles of the natural hazard risk rating methodology, the waterlogging resilience assessment methodology for historic urban areas was improved with the following formulae:

$$Res=V/P$$

(9)

$$V=\sum _{j=1}^n{w}_j{x}_{ij}$$

(10)

where $Res$ is the comprehensive resilience index of the historic urban area, taking into account the possibility of waterlogging disaster and the resilience of the carrier and rescue capacity. $P$ is the likelihood of waterlogging disaster; this paper uses the meta-cellular automata model to simulate the condition of the water flow, reflecting the likelihood that the study area suffers from the risk of waterlogging. $V$ is the resilience to waterlogging in the historic urban area, $w_j$ is the weights of the indicators, $x_{ij}$ is the indicator normalized value, and $j$ is the number of indicators.

According to Formula (10), the normalized urban waterlogging resilience index can be divided into five levels as shown in

Table 4. Measurement level of waterlogging resilience in historic urban areas.

Waterlogging Resilience Classification	Waterlogging Resilience Index
Higher	0.405381 < Res ≤ 0.472806
High	0.472807 < Res ≤ 0.525328
Moderate	0.525329 < Res ≤ 0.564724
Low	0.564725 < Res ≤ 0.600678
Lower	0.600679 < Res ≤ 0.663721

.

3.5. Obstacle Model

On the basis of the assessment of waterlogging resilience in historic urban areas, the obstacle model [2] is introduced to analyze and summarize the key factors hindering the enhancement of waterlogging resilience in historic urban areas in order to formulate optimization measures in a targeted manner, so as to improve the current waterlogging resilience of historic urban areas in a more effective way, to enable the government to better cope with urban problems caused by waterlogging hazards, and to realize the sustainable development of historic urban areas. The formula for calculating the obstacle degree is as follows:

$$O_i={\displaystyle \frac{W_i\left(1-x_i\right)}{{\sum }_1^n{W}_i\left(1-x_i\right)}}$$

(11)

where $O_i$ denotes the degree of obstruction of indicator i within that catchment and $W_i$ denotes the weight of the indicator on the resilience to waterlogging in the historic urban area.

4. Results and Discussion

4.1. CA Model Accuracy Validation

To validate the scientific validity and rational basis of the model, ensuring its capacity to accurately represent real-world conditions, rigorous testing of inundation characteristics derived from the developed CA simulation model is essential. This study employs a historical validation methodology to verify the model’s alignment with observed hydrological phenomena, thereby enhancing its fidelity to physical realities (Figure 4). Comparing the relative errors between the simulation results and the historical data, in general, the deviation between the calculation results and the actual values of this method is less than 15%, which can meet the needs of small-scale calculations.

Considering the feasibility of the Head/Tail breaks method in the division of evaluation units [44,45,46], we adopt this method for the division of evaluation units in Qingdao’s historic urban area, and the criterion for the Head/Tail breaks method is that two neighboring units are not in the same functional unit. Qingdao’s historic urban area was divided into 293 evaluation units in order to maximize the preservation of the regional spatial morphological integrity and to accurately classify the minimum evaluation units. A waterlogging risk model based on the CA model was constructed and used to simulate the inundation characteristics of the study area using the designed 2 h precipitation at 1 in 5a, 10a, 20a and 100a. The combination of its maximum inundation characteristics with the possibility of the simultaneous occurrence of astronomically high tides is used for the assessment and visualization of the factor depth of waterlogging in the hazard dimension. The spatial distribution of waterlogging risk levels and recorded historical waterlogging points is shown (Figure 5). The spatial distribution of stormwater accumulation risk levels is generally consistent with the distribution of recorded waterlogging points. The areas with high and relatively high risk of waterlogging are concentrated in the areas around Taiping mountain and Baguan mountain, the eastern area of Zhushui mountain, and the southern and northern coastal areas (during the astronomical tide period).

Overall, apart from coastal areas, the areas with high risk of waterlogging are mainly distributed along the hills, which are more affected by the topography of the area. When heavy precipitation comes, rainwater will converge from the hill downward to the foot of the hill, and waterlogging is prone to occur when water accumulates at the bottom of the hill in excess of the drainage capacity of the area. It should be noted that the areas with low and lower risk of waterlogging are concentrated in the north-western part of the Zhushui mountain, the western and north-western part of the Guanxiang mountain, and the north-western and central part of the historic district. These areas are relatively elevated and have a dense drainage network that can convey stormwater away relatively quickly.

In order to be able to quantitatively assess the accuracy of the waterlogging risk levels, the waterlogging risk levels of the corresponding locations of the waterlogging points were counted.

Table 5. Statistics on the spatial distribution of waterlogging sites.

Risk Level	Number of Flooded Sites
Low	2
Relatively low	14
General	15
Relatively high	11
High	0

lists the waterlogging risk levels for the corresponding locations of the 42 waterlogging points. Among them, 11 points are located in the higher waterlogging risk level, 15 points in the medium waterlogging risk level, 14 points in the lower waterlogging risk level area, and 2 points in low the waterlogging risk area and about 61.90 per cent of the waterlogging points are located in the medium and above waterlogging risk area.

4.2. Level of Resilience to Waterlogging

In order to analyze the spatial distribution characteristics of different dimensions of waterlogging resilience in Qingdao’s historic urban area, this study determined indicator weights through the application of the entropy weighting method. Subsequently, spatial superposition of quantitative assessments for each evaluation dimension was performed to calculate composite resilience values across three critical components: dangerousness, vulnerability, and adaptability. Following normalization procedures, the analytical outcomes are visualized through spatial mapping (Figure 6).

At the dangerousness level (Figure 6a), the areas with the highest risk of waterlogging are mainly located in the southern coastal area, which has low topography and is vulnerable to the dual risk of catchment water from the northern side of the area as well as seawater back-up (astronomical tides). In addition, Yan’an Road on both sides and the peripheral area of Zhushui mountain are also at high risk due to the relatively high topography of the surrounding area, rainwater catchment, sparse drainage network, and insufficient rainwater evacuation capacity. The low-risk areas for waterlogging can be divided into three main groups. These areas, such as Dalian Road, Jiaxiang Road, and Weihai Road, have relatively high terrain and dense population on both sides. The other risk levels are distributed from high to low along the cluster to the outer layers, showing a clear circled distribution. Combined with the site characteristics, it can be seen that, excluding the factor of seawater inundation under the condition of astronomical tides, the risk mainly depends on the low topography and high density of historical buildings, with poor rainwater infiltration, as well as the reduction in open space caused by the high-intensity development in the neighborhood, which increases the risk of waterlogging.

At the vulnerability level (Figure 6b), the most vulnerable regions are concentrated in the Sifang Road historical and cultural block, Zhongshan Road historical and cultural block, Guantao Road historical and cultural block, both sides of Leling Road, both sides of the eastern section of Weisi Road, both sides of Xikang Road, and both sides of Weihai Road, with an overall distribution along the south–west to north–east line. And the areas of lower vulnerability are mainly located in the north–south coastal region. The highly fragile regions are densely populated with historical buildings and have a high intensity of commercial and tourism development, especially the Sifang Road, Zhongshan Road, and Guantao Road historical and cultural block, which have many historical buildings with a high level of protection, and the construction years of the buildings are mainly concentrated in the end of the 19th century to the beginning of the 20th century, with many wooden buildings, which are more likely to be destroyed. Moreover, the above area carries a large number of commercial, financial, cultural, and tourism functions, while the presence of several texture fracture zones makes the crowd gathering flow line elongated, exacerbates the deterioration of historical buildings that are more vulnerable to disturbance, and highlights the brittleness characteristics.

At the adaptability level (Figure 6c), the high adaptability regions are mainly distributed on both sides of Yunnan Road, Zhongshan Road historical and cultural block, Sifang Road historical and cultural block, Guantao Road historical and cultural block, Shanghai Road-Wuding Road historical and cultural block, the northern part of Wudi Road historical and cultural block, the both sides of Leling Road, and the both sides of Weihai Road. The lowest low-adaptation regions are mainly distributed in the outer coastal region and both sides of Longtan Road, both sides of the northern section of Suzhou Road, the eastern side of Longshan Road, the eastern side of Jingshan Road, and the southern side of the western section of Huangshan Road, presenting a trend of an increase along the southwest-northeast and a decrease in the outward extension of the adaptability. Low-adaptation areas are mainly due to the high level of building deterioration and the presence of a certain number of unrepaired buildings, as well as poor road density and accessibility, which affects the evacuation of people. These regions are far away from hospitals, and the level of medical assistance is weak. In addition, the drainage network is relatively sparse and the terrain is low, so if affected by astronomical tides, the water cannot be drained, and these areas are not well adapted to waterlogging.

To analyze the spatial distribution characteristics of the comprehensive resilience of waterlogging disasters in Qingdao’s historic urban area, this paper combines the evaluation results of hazard, vulnerability, and adaptability dimensions and calculates the comprehensive resilience value of waterlogging disasters. And based on ArcMap, the natural breaks method is used to divide it into five grades, from low to high as the low resilience level (0–0.472806), relatively low resilience level (0.472807–0.525328), general resilience level (0.525329–0.564724), relatively high resilience level (0.564725–0.600678), and high resilience level (0.600679–0.663721) and the results are shown in Figure 7.

There is an obvious unevenness in the development of waterlogging resilience in Qingdao’s historic urban area. In terms of quantity, the distribution of waterlogging resilience levels in Qingdao’s historic urban area is uneven, with the highest number of general resilience units, 96, accounting for about 32.76%. Relatively high resilience levels have 75 units, accounting for the second highest percentage, about 25.60%. In order, there are 54 relatively low-resilience units, accounting for about 18.43%. High resilience units have 46, with about 15.70%. Finally, there are 22 low-resilience units, accounting for the least percentage, about 7.51%. It can be seen that Qingdao’s historic urban area waterlogging resilience level of general and above grade accounts for 74.06% and the overall view of Qingdao’s historic urban area resilience is high, but there are more prominent problems in the local area to be further improved.

In terms of spatial distribution, the high-resilience regions in Qingdao’s historic urban area are mainly divided into three major clusters: the central and western clusters, the central and northeastern clusters, and the central and southwestern clusters of the study area. These three groups of drainage pipe networks and drainage wells provide dense convenient transportation, closer to hospitals, emergency shelters, and most of the buildings have been retrofitted and reinforced. In general, these three groups are significantly more adaptable than other areas. Low and relatively low resilience coastal distribution is mainly due to the low topography of the region, the lower density of the drainage network, and proximity to the sea. The result is that these regions are susceptible to astronomical tides, causing seawater to back up, and are highly susceptible to waterlogging disasters, with the risk of waterlogging being significantly higher than in other regions. Overall, other waterlogging resilience levels range from high to low along the three major high-resilience group distributions in a circular type distribution.

4.3. Factor Analysis of Resilience Obstacles in Waterlogging

Based on the evaluation of the resilience level of waterlogging in Qingdao’s historic urban area, the obstacle model was introduced in order to determine the main factors affecting the waterlogging resilience in Qingdao’s historic urban area. According to the Formula (11), the top three obstacle factors in the indicator system were calculated, and the results are shown in Figure 8. In terms of dangerousness, both depth of waterlogging and density of sensitive sites appear in the first three barrier factors, and in particular, depth of waterlogging is the most dominant obstacle factor, which is present in all 281 evaluation units, and this factor is mainly distributed in the first three barrier factors and appears most frequently as the first obstacle factor. In terms of vulnerability, only population density and intensity of crowd concentration are distributed among the first three barrier factors, with population density appearing with the highest frequency, totaling 119 times, mainly as the third barrier factor. In terms of adaptability, pipeline network node, density of pipe network, NDVI, and building renovation and strengthening appear in the top three obstacle factors, with pipeline network nodes appearing with the highest frequency, totaling 102 times, mainly as the third obstacle factor. Building modification and reinforcement appear second most frequently, totaling 97 times, mainly as the third obstacle factor. Overall, Qingdao’s historic urban area suffers from a dense population, insufficient drainage capacity, high risk of waterlogging, insufficient vegetation cover, and lack of reinforcement measures for buildings, which affect the resilience of the historical urban area to waterlogging, so the resilience of Qingdao’s historic urban area can be strengthened in the future by measures in terms of channelling the flow of people, enhancing the drainage capacity, and increasing the indicator cover and the renovation and reinforcement of the buildings.

4.4. Strategies for Spatial Optimization and Solutions for Enhancing Resilience

Rapid urbanization has also changed the natural hydrological cycle of urban areas. Many historical buildings are facing the risk of homogenization, functional decline, environmental degradation, and even complete destruction. Based on the results of the above analysis, the following waterlogging resilience enhancement strategies are proposed using Qingdao’s historic city as a sample.

(1)

Reducing disaster ontology vulnerability based on micro-renewal and crowd diversion

At present, urban renewal in Qingdao’s historic urban areas has only begun to take on a connotative approach, with 420,000 m² of old buildings having been preserved and repaired. The next phase of work could introduce a strategy of incremental small-scale renovation, also known as micro-renewal, which aims to improve the disaster resilience of historic urban areas through local optimization and precise interventions, while preserving their historical features and cultural values to the maximum extent possible. At the same time, flood resilience in historic urban areas can be significantly improved by increasing crowd risk mitigation campaigns. Specifically, for the densely populated, narrow streets and alleys of historic urban areas and multi-level and multi-form disaster prevention and mitigation publicity and education can effectively enhance residents’ awareness of the risk of waterlogging and emergency risk avoidance capabilities. In particular, according to the climatic characteristics of its peak tourist season, it opens up the historical and cultural heritage for visits by limiting the flow of people at different times of the day.

(2)

Regional linkages and infrastructure optimization enhance overall resilience

Regional resilience enhancement in historic urban areas through spatial connectivity constitutes an inter-regional synergy strategy designed to holistically strengthen disaster mitigation capacity and sustainable development potential. This approach operationalizes through three synergistic mechanisms: resource integration from adjacent zones, functional configuration optimization, and cross-jurisdictional infrastructure co-management. The ability of urban areas to cope with flooding disasters will be significantly enhanced through systematic renewal and optimization of drainage, transportation, electricity, and other infrastructures. For example, the existing water system will be fully sorted out, permeable paving will be introduced without damaging the original urban color and spatial texture, and landscape cisterns will be added.

5. Conclusions

Historic urban areas are the roots and soul of urban culture, serving as both the origin of the city’s cultural lineage and the centralized manifestation of its distinctive character. However, with the passage of time, these areas have gradually developed increasingly complex and fragile characteristics, making their safety hazards and potential risks more prominent compared to general urban zones. In this study, we establish a waterlogging resilience evaluation system for historic urban areas by considering their architectural ontological characteristics and preservation value. This system is constructed through three key dimensions: dangerousness, vulnerability, and adaptability. By employing the CA model to simulate regional storm inundation patterns and incorporating emerging open-source data to complement traditional statistical and monitoring datasets, a quantifiable indicator system is proposed. Building upon this foundation, we implement the Head/Tail Breaks method for evaluation unit delineation, followed by a comprehensive assessment of waterlogging resilience in historic urban areas. Finally, we apply an obstacle model to identify critical constraints affecting waterlogging resilience in Qingdao’s historic area, proposing context-specific optimization strategies based on localized conditions.

The research findings demonstrate that (1) the two-dimensional surface inundation model based on the CA model framework can effectively simulate waterlogging characteristics in Qingdao’s historic urban area. The water depth distribution results derived from the CA-based surface runoff module in this study show fundamental consistency with established commercial software models, demonstrating less than 15% deviation between simulated results and measured flow values. However, the CA model retains certain limitations. While effectively simulating short-duration heavy rainfall events in plain urban areas, the model’s focus on localized neighborhood interactions may overlook broader systemic influences. Additionally, its capacity to simulate complex coupled hydrological processes remains constrained. (2) Building upon urban resilience theories and evaluation index systems established in domestic and international studies, this research constructs a resilience evaluation index system under waterlogging disaster scenarios for historic urban areas. The evaluation of Qingdao’s historic urban area reveals an overall resilience pattern: low-resilience zones predominantly cluster in southern and northern coastal areas, while high-resilience areas concentrate in central-western, northeastern-central, and southwestern-central urban regions. (3) Dominant factors influencing waterlogging resilience in Qingdao’s historic urban area exhibit notable consistency, primarily including population density, crowd aggregation intensity, density of sensitive points, water accumulation depth, drainage network density, pipeline node distribution, vegetation index, and building retrofitting measures. Consequently, targeted improvements in these dimensions are imperative to enhance comprehensive resilience. (4) The research methods and conclusions of this article contribute to a deeper understanding of urban waterlogging disaster risks, identify resilience barriers faced by specific entities in historic urban areas, and assist government agencies and relevant stakeholders in disaster risk mapping and the selection of appropriate reinforcement measures for architectural heritage.

Through empirical analysis, this study identifies critical priorities for waterlogging disaster prevention in Qingdao’s historic urban area and proposes context-specific optimization strategies, offering valuable insights for resilience-oriented urban planning in historic areas. However, limitations persist: the current index system requires refinement, as the exclusion of certain indicators due to data unavailability (e.g., detailed rainfall impact analysis on individual buildings) constrains methodological completeness. Although CA models have limitations compared to hydrodynamic models, such as assuming constant and non-dynamic outflows over time, it is still an easy and effective way to simulate flooding risk at a fast and large scale. Future research must address these gaps through three critical pathways: (1) developing high-resolution spatiotemporal assessments of precipitation impacts on structural integrity by coupling hydrological–hydrodynamic modeling with building-specific vulnerability curves; (2) establishing dynamic flood risk time-series prediction models that integrate real-time drainage system performance monitoring and machine learning-based early warning algorithms; and (3) constructing multidimensional analysis frameworks that systematically integrate IPCC SSP-RCP climate projection scenarios, urban morphological evolution patterns, and infrastructure aging parameters. This tripartite approach enables simultaneous quantification of acute waterlogging threats and chronic climate change impacts on urban systems.