Introduction

As the global community progresses toward the United Nations 2030 Sustainable Development Goals (SDGs) and New Urban Agenda (NUA), cities are expected to exemplify resilience, safety, vitality, and inclusiveness (Caprotti et al. 2017; Cheng et al. 2021). Yet, many urban areas worldwide are grappling with the persistent challenge of urban shrinkage, which poses a significant threat to urban vitality—an important indicator for measuring the quality and potential of urban development (Jarzebski et al. 2021).

Urban shrinkage has been widely studied in Europe and North America (Turok and Mykhnenko, 2007; Wiechmann and Pallagst, 2012), but its increasing prevalence in developing nations, including China, has drawn considerable attention. As a major global economy in the late stages of urbanisation, China is transitioning from expansion to optimisation, and urban vitality is a cornerstone of this transformation (Jin, 2007). However, in many cities in China, population growth has stagnated or declined, and urban and regional shrinkage is an increasingly prominent issue. Long et al. (2015) identified and categorised 180 shrinking cities in China. Among these, cities in Northeast China—a historically robust industrial region—have suffered profound population loss and economic stagnation. Between 2010 and 2020, the three northeastern provinces collectively lost 10.99 million people and represent the most severely affected region in China (Huang and Wang, 2024). This trend threatens regional stability and poses substantial challenges to national development.

Amid mounting concerns over urban shrinkage, the Chinese government launched a series of revitalisation initiatives, most notably the ‘Northeast Revitalisation’ strategy, with the HCUA designated as a pilot area for recovery. Concurrently, researchers have sought to address urban shrinkage through the lens of urban vitality. Studies have focused on the internal dynamics of cities and examined factors such as the built environment, street density at the meso and micro scales, and their effects on urban vitality (Ye et al. 2018; Zhang et al. 2021). These studies include analyses of individual cities (Fan et al. 2021; Liu and Shi, 2022) and comparative studies across cities (Xia et al. 2020; Yue et al. 2021). However, these efforts still lack a systematic analytical framework and methodology and have failed to explore urban vitality in the context of urban shrinkage.

Against this backdrop, the study raises the following core research questions: Does urban shrinkage inevitably diminish urban vitality? Can different types of cities still exhibit multidimensional vitality under conditions of shrinkage? How is urban vitality spatially articulated across urban and regional scales, and what forms of heterogeneity emerge? What are the key factors that support or hinder the formation and enhancement of urban vitality?

To address these questions, this study foregrounds urban vitality in the context of shrinkage and proposes an integrated analytical framework spanning both urban and regional scales. It aims to investigate the spatial heterogeneity of urban vitality and its underlying drivers within urban agglomerations situated in representative shrinking regions of China. At the urban scale, a multidimensional evaluation system encompassing social, economic, environmental, and cultural dimensions is constructed, using statistical data to identify vitality patterns, spatiotemporal dynamics, and primary systemic obstacle factors. At the regional scale, the study incorporates grid-based spatial units and population heatmap data to examine the relationship between human activity intensity and urban vitality. Furthermore, multi-source geospatial big data are integrated to explore the influence of the built environment on urban vitality. This research seeks to provide policy insights and technical guidance for enhancing vitality in shrinking cities and promoting high-quality urban development, contributing to the broader global response to urban shrinkage.

Literature review

Jacobs (1961) introduced the concept of urban vitality as a hallmark of thriving urban environments by emphasising mixed land use, active streets, and the density of activities. Inspired by Jacobs’ urban theory, the relationship between urban vitality and the built environment has garnered sustained attention. Numerous studies have focused on theoretical and empirical examinations of urban vitality through the lens of physical, spatial, and street morphology (Li et al. 2022; Xia et al. 2022; Oliviera, 2016; Bosselmann et al. 1999). This attention arises from the direct and profound impact of the physical environment, and quantitative analysis of the factors that influence urban vitality has been an enduring topic of research. Early approaches relied on photographic documentation, manual scoring (Ewing and Cervero, 2001), and multiple linear regression models (Zarin et al. 2015). With advancements in research techniques and data availability, recent studies have employed big data sources such as point of interest (POI) data, GPS trajectories, mobile signalling, and nighttime light remote sensing to enhance analytical precision and comprehensiveness (Kim, 2020; Guo et al. 2021). Moreover, Jacobs’ ideas have been adapted into GIS-based urban vitality evaluation methods (Delclòs-Alió and Miralles-Guasch, 2018) that provide urban planners with actionable tools for enhancing urban liveability.

Current research on urban vitality has focused on developing countries (Paköz et al. 2022; Ruszczyk et al. 2023; Chen et al. 2023) and includes macro-scale studies at the national level (Liu et al. 2024), meso-level studies on the vitality of urban internal spaces (Shi et al. 2021), and micro-level research on waterfront areas, historic cultural districts, and transit hubs (Liu et al. 2021). These studies commonly focus on regions with high population density, complex functional structures, and rapid urbanisation. However, these focal areas highlight the lack of research in regions facing severe population decline and urban shrinkage. Additionally, urban agglomerations play a central role in regional economies, spatial planning, population movement, social development, and policy governance, and they drive sustainable urban development and global economic integration. Research on the vitality of urban agglomerations remains insufficient. As research progresses, scholars have recognised that urban vitality is not merely a reflection of spatial morphology and material layout; it is deeply connected to human activities, interactions, and emotional experiences within a city. For instance, Sung et al. (2013) evaluated the urban vitality of Seoul by collecting walking activity data and employing multiple linear regression models. However, traditional methods of assessing urban vitality have focused on the material environment of urban construction and have neglected the inclusion of non-built environment variables, which leads to incomplete conclusions.

The precision and depth of research on urban vitality have improved significantly, but some issues remain underexplored. First, urban shrinkage offers a new perspective in the study of urban vitality. Since the mid-20th century, population decline in Western cities has attracted academic attention to the issue of urban shrinkage. Initially, research focused on defining shrinking cities, identifying their types, and exploring the causes, impacts, governance policies, and planning responses (Lotscher, 2005; Martinez-Fernandez et al. 2012; Schetke and Haase, 2008; Blanco et al. 2009; Cunningham-Sabot and Fol, 2007; Eraydin and Özatağan, 2021; Pallagst et al. 2021). As research expanded to include the built environment and liveability, a growing number of scholars linked urban shrinkage with urban vitality (Jiang and Sun, 2024). Urban shrinkage was historically viewed as a sign of declining urban vitality (Hollander et al. 2009), but recent studies suggest that the two are not necessarily in direct opposition (Feng et al. 2023), although discussions of this topic remain limited. Second, assessments of urban vitality should prioritise comprehensive frameworks and context-specific strategies, and studies must consider the interaction between human activity and urban space and how these factors respond together to urban shrinkage and influence urban vitality. In this study, we conducted a detailed analysis at both urban and regional scales and accounted for the influence of multiple factors on urban vitality. By integrating big data analytics, spatial statistical methods, and human behaviour data, the goal of this study was to provide new perspectives and methods for evaluating urban vitality in the context of urban shrinkage. For areas that typify shrinkage, we explored how to assess and enhance urban vitality while providing support for urban planning and policy-making.

Methodology, data, and research framework

Study area

The HCUA, which encompasses 11 major prefecture-level cities including Harbin, Qiqihar, Daqing, Suihua, Mudanjiang, Changchun, Jilin, Siping, Liaoyuan, Songyuan, and Yanbian (Yanbian Korean Autonomous Prefecture), was selected as the study area. The HCUA contains a core area of approximately 51,100 km2 (Fig. 1).

Fig. 1
figure 1

Study area.

Full size image

Methodology

Population change rate

This study uses 2016 and 2022 urban population data to calculate the change rate for each prefecture-level city. The formula is as follows:

$${C}_{\delta }=\frac{{P}_{2022}-{P}_{2016}}{{P}_{2016}}\times 100 \%$$
(1)

where Cδ denotes the population change rate, and P2022 and P2016 represent the urban population of a given city in 2022 and 2016, respectively.

Urban vitality indicator system

This study develops an urban vitality evaluation index system to systematically assess vitality across multiple dimensions, providing a scientific basis for evaluating the overall quality of urban development. The framework draws on Jacobs’ (1961) seminal notion that urban vitality emerges from the diversity generated through interactions between human activity and urban space, as well as Lynch’s (1984) perspective that urban vitality reflects a city’s capacity to sustain internal survival, growth, and development. It also incorporates the systemic definitions proposed by Lan et al. (2020) and Feng et al. (2023), both of which frame urban vitality as a multi-dimensional construct encompassing economic, social, environmental, and cultural domains, underpinned by strong inter-systemic synergies. Guided by the principles of multidimensionality, operability, and data accessibility, we construct an index system spanning four key dimensions: economy, society, environment, culture and technology. The framework comprises four vitality layers and 24 specific indicators (Table 1).

Table 1 Urban vitality indicator system at the urban scale.
Full size table

We also incorporated human activities in the construction of an evaluation system for urban vitality at the regional scale. By selecting Baidu population heatmap data and combining this data with the urban vitality indicator system, the HCUA was divided into a 3 km × 3 km grid to measure urban vitality at the regional scale, which was calculated as follows:

$${L}_{j}=\mathop{\sum }\limits_{i=1}^{n}{T}_{ij}\frac{{E}_{ij}}{{e}_{j}}\cdot\mathop{\sum }\limits_{j=1}^{m}{G}_{j}$$
(2)

where Lj represents the value of the dependent variable at grid location j, which indicates urban vitality at the regional scale. Tij represents the value of the independent variable for the grid at location j, which indicates the city-scale urban vitality of city type i at that grid. Eij is the value of the independent variable for the grid at location j, which indicates the area occupied by the cities of category i at that grid, where n = 1, 2, 3,…,11. ej is the area of the grid at position j. Gj is the value of the independent variable for the grid at location j, which represents the normalised population heatmap data values, where m = 1, 2, 3,…,36,776.

Entropy weighting

To avoid subjective bias, this study employs the entropy weight method for assigning weights, ensuring an objective and accurate evaluation. This helps to avoid randomness and overcome issues of information redundancy between multiple variables. Based on socioeconomic data from 2016 to 2022, we calculate the weights of urban vitality indicators and subsequently use radar charts to illustrate the evolution of urban vitality and the characteristics of each indicator dimension within the HCUA.

Obstacle degree model

This model was applied to assess the impact intensity of various hindering factors on urban vitality. This approach incorporates indicator deviation, factor contribution, and obstacle degree for analysis (Zhao et al. 2022). The contribution of factors is quantified through their weights, which measure the influence of individual elements on urban vitality. Indicator deviation denotes the disparity between an individual indicator and the ideal standard. Obstacle degree analysis is conducted at both the individual and criterion levels to illustrate the impact of evaluation indicators on overall urban vitality and the cumulative effect of individual obstacles within each criterion.

Urban primacy index

The Eleven Index Method of Urban Primacy was employed to examine the distribution of urban scales within a region and to analyse the structural dynamics of the urban system. A threshold of 1 was established for calculation, and values exceeding this limit indicate a primacy advantage. The formula for calculation is as follows:

$$S=\frac{2{P}_{1}}{({P}_{2}+{P}_{3}+{P}_{4}+\mathrm{..}.+{P}_{11})}$$
(3)

where P1, P2,…,P11 were the sizes of the city in a certain bit order after city sizes were sorted from largest to smallest.

Standard deviation ellipse

This method was utilised to explore the directional distribution and trends in geographic elements. It describes spatial characteristics through the mean centre, standard distance, and azimuth (Benita et al. 2021). In this study, ArcGIS Pro was employed to analyse the spatial distribution and central migration of urban vitality. We selected a 1-standard-deviation ellipse to cover 68% of the elements, which revealed the spatiotemporal evolution characteristics.

Multiple linear regression analysis

Multivariate linear regression analysis was used to explore the relationship between the dependent variable and multiple independent variables. This study focused on urban vitality and considered the influence of factors such as economics, society, environment, culture, and technology. We utilized POI data to calculate indicators such as a functional mix index and selected transportation and green space as candidate variables. Exploratory regression was conducted using ArcGIS Pro, with filtering criteria of Radj2 ≥ 0.8, a variance inflation factor (VIF) < 10, and a minimum AICc value. Finally, multifactor correlations were performed with SPSS 25 to explore the intrinsic relationships between the variables.

Data sources

The data used in this study include administrative division data, road network data, population heatmap data, statistical data, and points of interest (POI) data. Administrative division data were obtained from the Resource and Environmental Science and Data Center of the Chinese Academy of Sciences (https://www.resdc.cn). Road network and building data were sourced from the OpenStreetMap (OSM) platform (https://www.openstreetmap.org/). Population heatmap data were collected from the Baidu Heatmap Data Platform (https://lbsyun.baidu.com) for April, July, and November 2023, and January 2024, covering the same day each month and yielding 12,475,421 records. Statistical data were derived from the China Statistical Yearbook (https://www.stats.gov.cn/sj/ndsj/), Jilin Province Statistical Yearbook (http://tjj.jl.gov.cn/tjsj/tjnj/), and Heilongjiang Province Statistical Yearbook (https://tjj.hlj.gov.cn) for the years 2017, 2020, and 2023. POI data were retrieved via the Amap API (https://www.amap.com), covering the HCUA region in November 2022, with 985,498 valid records retained.

Research framework

This study develops an analytical framework for assessing urban vitality under conditions of urban shrinkage (Fig. 2). The framework is designed to be adaptable and context-sensitive, supporting empirical application across different regional settings. The overall workflow comprises three main components. The first component focuses on the collection and preprocessing of multi-source data, providing a foundation for subsequent analysis. The second component establishes a dual-scale analytical approach encompassing both urban and regional levels. Based on the identification of shrinking cities, a multidimensional urban vitality evaluation system is constructed at the urban scale, with objective weighting performed using the entropy method. Comprehensive vitality assessments are conducted within prefecture-level administrative units using statistical data, and vitality levels are classified via the natural breaks method to reveal spatial patterns and temporal evolution. System radar charts are then employed to illustrate the multidimensional vitality profiles across the 11 cities, while the obstacle degree model is used to identify constraining dimensions within the vitality system and highlight key bottlenecks. At the regional scale, a grid-based analytical unit is developed by integrating geospatial big data sources—including population heatmaps, road network density, and POI data—to capture diurnal rhythms and seasonal fluctuations in spatial vitality. This enables a deeper investigation into the coupling mechanisms among human activity, the built environment, and vitality distribution, and further identifies key spatial factors influencing vitality performance. The third component synthesises the multidimensional results to advance theoretical understanding of urban vitality and inform spatial governance in shrinking cities.

Fig. 2
figure 2

Flowchart.

Full size image

Results

Identification and key characteristics of urban shrinkage

There is currently no universally accepted definition of shrinking cities in domestic or international research; however, population decline is widely regarded as the primary criterion. As a direct outcome of industrial decline, deteriorating living environments, and other structural factors, population loss serves as the most immediate and reliable indicator of urban shrinkage. Accordingly, this study identifies shrinking cities based primarily on demographic indicators.

Drawing on the socioeconomic characteristics of China’s ‘new normal’ development stage, we categorise cities into three types according to population change rates (Cδ): rapid-growth cities (Cδ ≥ 6%), growth cities (0 < Cδ < 6%), and shrinking cities (Cδ ≤ 0%) (Ma et al. 2021). It is important to note that this study defines population shrinkage as a long-term and sustained process rather than a short-term fluctuation. Based on changes in urban population over a 6-year period, we identify 10 out of the 11 prefecture-level cities in the HCUA as shrinking between 2016 and 2022—namely, Harbin, Daqing, Mudanjiang, Qiqihar, Jilin, Suihua, Songyuan, Siping, Liaoyuan, and Yanbian—while only Changchun experienced population growth (Fig. 3).

Fig. 3
figure 3

Spatial distribution of shrinking and growing cities in the HCUA.

Full size image

The HCUA exhibits a spatial pattern of widespread shrinkage with marked clustering, reflecting the combined effects of regional imbalance and structural transition. Three main factors underpin this trend: First, peripheral geographic locations limit the integration of some cities into the region’s core development axis. For instance, Qiqihar and Yanbian lie on the western and eastern peripheries of Heilongjiang and Jilin provinces, respectively. Their distance from the HCUA’s central corridor and main industrial belt, coupled with limited transportation connectivity, industrial synergy, and policy spillovers, has led to economic marginalisation and continued population decline. In Yanbian’s case, cross-border labour outflows have further exacerbated shrinkage. Due to cultural and linguistic ties with neighbouring countries and sluggish local economic growth, many residents seek long-term or permanent employment opportunities in South Korea and Japan—creating a disembedded migration pattern that renders the region distinctive in the broader discourse on urban shrinkage in Northeast China. Second, resource-based cities face stalled industrial transitions and lack endogenous growth momentum. Cities such as Qiqihar, Liaoyuan, and Songyuan, long reliant on coal, petroleum, and other extractive industries, have experienced industrial hollowing, job losses, and youth outmigration as resources are depleted and reform policies fall short—resulting in typical patterns of shrinkage. Third, the siphoning effect of core cities exacerbates passive shrinkage in surrounding small- and medium-sized cities. The concentration of resources in Harbin and Changchun—particularly in education, healthcare, and transport—has exerted a strong gravitational pull on talent and investment, weakening the competitiveness of adjacent cities such as Suihua, Songyuan, Siping, and Liaoyuan in both industrial upgrading and population retention. These cities have suffered from a ‘shadow under the light’ effect during the regional integration process, where development concentrates in the core while peripheries are left behind. This pattern of passive shrinkage underscores the pressing issue of regional inequality in Northeast China.

Urban-scale evaluation of vitality and systemic obstacles in the HCUA

Overall pattern and multi-system analysis of urban vitality

In evaluating urban vitality at the urban scale, this study selects three critical time points—2016 (the year the HCUA was established), 2019 (prior to the COVID-19 pandemic), and 2022 (the post-pandemic recovery period marked by new policy interventions)—to assess the vitality levels of 11 cities. Based on the results, cities were categorised into three levels—high, medium, and low—using the natural breaks method. Overall, urban vitality showed an upward trend over time; however, significant disparities persisted across cities. As provincial capitals, Changchun and Harbin consistently outperformed other cities, with vitality scores of 0.823 and 0.808 in 2022—three to four times higher than most other cities—demonstrating a pronounced primacy effect, respectively. In contrast, although cities such as Daqing and Qiqihar experienced some improvement, the majority of non-core cities remained in the low vitality range, highlighting a pronounced imbalance in development (Fig. 4).

Fig. 4
figure 4

Comprehensive urban vitality levels across HCUA cities (2016–2022).

Full size image

Taken as a whole, the HCUA has developed a spatial structure characterised by a ‘dual-core, two-belts, and one-axis’ configuration: the ‘dual-core’ refers to Harbin and Changchun as vitality centres; the ‘two belts’ encompass the Harbin–Daqing–Qiqihar–Mudanjiang (Ha–Da–Qi–Mu) belt and the Changchun–Jilin–Tumen River (Chang–Ji–Tu) belt; and the ‘one axis’ connects Harbin and Daqing (Fig. 5). This configuration reflects a preliminary stabilisation of vitality hierarchies within the agglomeration, but the unbalanced development pattern has, to some extent, limited overall coordination and efficiency.

Fig. 5
figure 5

Spatiotemporal evolution of urban vitality across HCUA cities (2016–2022).

Full size image

To further investigate the internal structure of urban vitality, the study employs a radar chart analysis based on a four-dimensional index system, covering economic (a), social (b), environmental (c), and cultural (d) dimensions. Results show that Changchun performs strongly in the economic and cultural dimensions (scoring 24.24 and 23.02, respectively, in 2022), while Harbin stands out in the social and cultural dimensions (20.57 and 25.32, respectively). Both cities present relatively balanced, near-circular radar profiles, indicating a high degree of multidimensional coordination. In contrast, other cities exhibit unbalanced vitality structures. Daqing exhibited noticeable improvement in the economic dimension (rising from 8.08 in 2016 to 12.29 in 2022), but saw a decline in social vitality, revealing structural shortcomings. Yanbian Prefecture peaked in environmental vitality in 2019 (13.45), but declined sharply to 4.75 by 2022, reflecting significant volatility. Its cultural dimension, however, remained relatively stable, demonstrating a certain degree of resilience. Qiqihar and Suihua showed notable improvement in cultural and environmental vitality, contributing to their overall rank increases to fourth and eighth, respectively. However, both cities still face shortcomings in economic and social fundamentals. Cities like Songyuan and Mudanjiang exhibited modest fluctuations in environmental vitality, with limited overall progress. Meanwhile, Liaoyuan and Siping consistently ranked low across dimensions, with 2022 social and cultural scores both below 3, reflecting a sustained decline in structural vitality (Fig. 6).

Fig. 6
figure 6

Radar-chart of urban vitality evolution across HCUA cities (2016–2022).

Full size image

The ‘dual-core’ cities of the HCUA, Changchun and Harbin, consistently exhibited high levels of urban vitality. Primacy index analysis reveals their dominant positions across six key domains: population, economy, industry, services, transport, and culture and technology. Changchun ranks first in industrial primacy (1.100), reflecting a strong manufacturing base, particularly in the automotive sector, which serves as a key driver of its overall vitality. Harbin, by contrast, leads in services and science and technology, with primacy indices of 1.874 and 1.472, respectively, underscoring its strengths in tertiary sector development and innovation capacity. Their complementarity in industrial structure and functional positioning constitutes a core pillar of support within the urban agglomeration.

However, it is important to note that this ‘core–periphery’ structure has yet to yield a ‘trickle-down’ effect. Cities such as Suihua, Siping, Liaoyuan, and Songyuan have long remained at low vitality levels, showing little evidence of benefiting from the growth of core cities. Moreover, although Yanbian and Songyuan have maintained relatively high environmental vitality scores, their economic and social dimensions have seen little improvement. This suggests that in contexts of low human activity intensity, relatively intact ecosystems may persist, but are not easily translated into comprehensive and sustainable urban vitality. In sum, the evolution of urban vitality within the HCUA reflects a spatial structure of ‘strong cores, weak peripheries, and marked differentiation’. Cities exhibit high heterogeneity in both vitality levels and dimensional structures, and the asynchronous evolution of multidimensional vitality—alongside imbalances in resource allocation—presents a major challenge to regional coordination.

Identification of obstacle factors in multidimensional vitality system

Using the obstacle degree model, we further identified the primary constraints within the multidimensional urban vitality system. Setting the obstacle degree threshold at 5, ten significant hindering factors were screened (Table 2). The results indicate that cultural and technological indicators consistently ranked at the top, with library collections (D1), the number of university students (D2), and government expenditure on science and technology (D4) showing persistently high obstacle degrees. For instance, D1 ranked second in 2016 with an obstacle degree of 10.74%, and reappeared in the top three in 2022 (8.87%). D4 topped the list in both 2019 (10.84%) and 2022 (9.24%), suggesting persistent shortcomings in cultural infrastructure, talent retention, and innovation capacity, which hinder the endogenous growth of urban vitality. Moreover, the obstacle degree of local fiscal revenue (A4) has steadily increased—from 5.03% in 2016 to 6.78% in 2022—reflecting weak fiscal autonomy and rising pressure on local budgets. This indicates inadequate investment in public services and spatial governance, forming a key bottleneck in the enhancement of urban vitality.

Table 2 Main urban-scale obstacles to vitality in the HCUA (2016–2022).
Full size table

By contrast, the degree of constraint imposed by several economic and infrastructure-related factors—such as gross regional product (A3), road area (B7), and hospital beds (B1)—has declined in the overall obstacle ranking. This suggests that the HCUA has reached a basic threshold in economic development and infrastructure provision, with diminishing marginal returns in their contribution to vitality. Meanwhile, new drivers such as tourism development and improved public services are gradually emerging. Overall, the cultural subsystem consistently exhibited the highest obstacle degrees across all years, significantly exceeding those of environmental and social dimensions, indicating a structural, persistent, and dominant constraint (Fig. 7). These findings highlight the urgent need to address bottlenecks in cultural and technological capacity, fiscal autonomy, and public service investment, and to shift from investment-led to innovation-driven and service-oriented urban development.

Fig. 7
figure 7

Obstacle degrees within the urban vitality system of the HCUA (2016–2022).

Full size image

Regional-scale evaluation of vitality and influencing factors in the HCUA

Evaluation of urban vitality through population heatmap data

At the regional scale, this study conducts a fine-grained assessment of urban vitality using grid-based spatial units, revealing the spatial relationship between human activity intensity and vitality distribution. The results indicate that urban vitality across the HCUA exhibits pronounced temporal and spatial variations. Shaped by a combination of climatic conditions, industrial rhythms, and population dynamics, vitality levels fluctuate in response to both diurnal and seasonal cycles.

Urban vitality follows a distinct diurnal rhythm, peaking during the day and declining at night. Between 3 a.m. and 5 a.m., vitality is concentrated in central urban areas but drops to its lowest overall level due to low temperatures. In contrast, midday (11 a.m.–1 p.m.) and early evening (5 p.m.–7 p.m.) witness increased resident mobility and heightened commercial activity, which boost vitality and facilitate its diffusion from urban cores to peripheral zones (Fig. 8). Although the night-time economy primarily targets younger populations, sustained youth outmigration and accelerated demographic ageing have disrupted this alignment, resulting in a structural mismatch between consumption demand and service supply. This, in turn, weakens both the sustainability and spatial reach of night-time urban vitality.

Fig. 8
figure 8

Hourly distribution of urban vitality at the regional scale.

Full size image

Beyond daily cycles, population heatmap data reveal seasonal patterns of urban vitality, with higher levels in spring and summer and notable declines in autumn and winter. Variations in temperature and daylight duration not only affect mobility patterns but also influence seasonal urban lifestyles (Fig. 9). These fluctuations are not solely attributable to climate but rather reflect the interplay of resident behaviour, industrial activity, and demographic dynamics. First, mild temperatures and abundant sunlight in spring and summer facilitate outdoor activities and the use of public spaces, substantially enhancing residents’ motivation for leisure, consumption, and social interaction—particularly invigorating vitality in peripheral areas. In contrast, harsh winter suppresses non-essential travel, concentrating vitality within well-equipped central zones and resulting in a visibly contracted spatial pattern. Second, demographic structures further intensify both diurnal and seasonal disparities. The region’s significant youth outmigration and population ageing undermine the consumer base necessary to sustain night-time and winter economies. Especially during high-consumption periods (e.g. evenings or holidays), imbalances in population structure shift vitality further toward daytime activity and spatial concentration in core urban areas.

Fig. 9
figure 9

Seasonal variation of population thermal values over a 24-h period.

Full size image

Building on this, we further examine the spatial dynamics of seasonal vitality using the standard deviation ellipse model (Fig. 10). The results reveal that urban vitality is relatively constrained and spatially stable during spring and autumn, whereas in winter and summer it expands significantly, with distribution becoming more dispersed and directionally pronounced. This spatial expansion is not only a result of broader resident activity ranges but is also closely linked to the stimulation of seasonal tourism economies.

Fig. 10
figure 10

Standard deviation ellipse distribution of seasonal vitality in the HCUA.

Full size image

In summer, the cool climate attracts vacationers seeking respite from the heat, while in winter, ice and snow-based resources and tourism products trigger short-cycle sightseeing consumption. These seasonal economic activities—such as ski resorts and ice lantern festivals—promote the temporary prosperity of related service sectors, including accommodation, catering, transportation, and entertainment. As a result, urban vitality extends into peripheral areas and enhances their level of engagement. According to the ellipsoidal parameters, the major and minor axes in summer reach approximately 128.34 and 75.53 km, respectively, with the lowest flatness value of the year (~0.41). This indicates that urban vitality during this season exhibits the widest spatial distribution, the strongest degree of expansion, and the most pronounced directional tendency. Winter follows, with slightly smaller ellipsoidal dimensions but still a notable degree of elongation. In contrast, the major axes in spring and autumn measure approximately 125.55 and 122.77 km, with flatness increasing to around 0.44 and 0.47, suggesting a more compact spatial distribution of vitality. Meanwhile, the centre of urban vitality shifts along a northeast–southwest axis during both summer and winter, further reflecting its distinct seasonal dynamics.

Overall, the HCUA demonstrates a broader spatial reach, higher degree of dispersion, and noticeable centroidal shifts in winter and summer, reflecting the profound role of seasonal change in shaping the spatial configuration of urban vitality—by influencing resident behaviour, adjusting industrial rhythms, and amplifying demographic effects.

Importantly, the regional-scale vitality structure echoes and reinforces the patterns identified at the urban scale. High-density vitality clusters are anchored in the ‘dual-core’ centres of Harbin and Changchun, with linear extensions radiating outward to form a ‘dual-core–satellite’ configuration. The central zone serves as the vitality engine of the HCUA, linking surrounding hotspots into a spatial network that supports agglomeration development. In contrast, peripheral areas in the east and west remain fragmented. The eastern region is characterised by dispersed vitality points, impeding the formation of larger clusters, while the western region exhibits small-scale, isolated concentrations with limited integration. These findings underscore the enduring core–periphery structure of the HCUA and highlight the persistent spatial inequality in vitality distribution, even amidst seasonal and temporal fluctuations.

Driving factor analysis using multi-source geospatial big data

Building on existing research on urban vitality and in response to analytical needs at the regional scale, this study selects eight candidate variables to examine the driving forces behind urban spatial vitality. Drawing on recent empirical studies and analytical frameworks, these variables represent critical aspects shown to shape urban vitality (Li et al. 2022; Sun and Jiang, 2024; Zhang et al. 2024). Specifically, the selected variables—road network density, building density, functional mix, residential service facilities, transportation facilities, economic facilities, green space facilities, and administrative facilities—capture key dimensions of the built environment, including spatial morphology, service accessibility, and urban governance (Fig. 11).

Fig. 11
figure 11

Influencing factors identified by multiple linear regression.

Full size image

To identify the most explanatory combination of variables, this study employed correlation analysis and exploratory regression. Table 3 presents three representative combinations of four variables each, all of which passed multicollinearity diagnostics and demonstrated high explanatory power (R2 > 0.8) with acceptable variance inflation factors (VIFs < 10), indicating strong model fit without significant multicollinearity.

Table 3 Exploratory regression results on driving factors.
Full size table

Building on the model evaluation and analysis of variable interrelationships, four key driving factors were ultimately retained: road network density, building density, residential service facilities, and transport service facilities. Using urban vitality as the dependent variable, these four predictors were entered into a multiple linear regression model in SPSS 25. The model achieved an R2 of 0.844, suggesting that 84.4% of the variance in urban vitality is explained, and the ANOVA results confirmed the model’s statistical significance (P = 0.000 < 0.05; Table 4).

Table 4 Model summary.
Full size table

Further regression results (Table 5) indicated that all four variables have positive coefficients, confirming their significant positive association with urban vitality. Specifically, a well-structured urban form, efficient transportation systems, and comprehensive residential and transport services collectively provide the foundational support for enhancing both internal connectivity and external interactions, thereby reinforcing urban vitality at the regional scale.

Table 5 Regression coefficients.
Full size table

Discussion

This study aims to provide a new analytical approach for understanding urban vitality in the context of urban shrinkage and to offer empirical evidence for the transformation-oriented governance of shrinking cities. It seeks to clarify the intrinsic relationship between shrinkage and vitality, examining whether the two are linked through a singular negative causal chain. The findings indicate that urban shrinkage and vitality are not in a simple linear opposition (Feng et al. 2023; Jiang and Sun, 2024); rather, their interactions vary significantly across dimensions and city types.

While shrinkage is widespread across the HCUA, core cities such as Harbin and Changchun have maintained high levels of vitality. Moreover, some shrinking cities—such as Qiqihar and Suihua—have exhibited signs of vitality resurgence, particularly in cultural and environmental dimensions. This indicates that vitality enhancement does not necessarily depend on overall growth, but may instead occur through forms of ‘asymmetric restoration’ or ‘dimension-specific reinforcement’. From a spatial perspective, most shrinking cities are concentrated in ‘peripheral’ and ‘urban fringe zones’ (Chen et al. 2024), where they are significantly influenced by the ‘siphoning effect’ of core cities (Feng et al. 2023). Nevertheless, vitality distribution within the region remains spatially fragmented, with limited evidence of a ‘trickle-down’ effect.

Notably, population decline has a direct impact on social and consumption-related vitality (Ma et al. 2024), while cultural and environmental vitality appear relatively resilient. In some cities suffering from substantial outmigration, these dimensions have remained comparatively stable. For example, Daqing and Qiqihar have established vitality-supporting systems through environmental improvements and investment in cultural infrastructure during the process of industrial restructuring. Similarly, Yanbian has experienced long-term population loss due to cross-border labour migration, resulting in weakened economic and social vitality. Yet, its cultural vitality has remained resilient, driven by strong ethnic identity and the development of border tourism. These cases not only highlight the distinct trajectories of border cities under shrinkage pressure, but also complement the existing literature by drawing attention to underexplored regional contexts (Ma et al. 2021; Feng et al. 2023; Jiang and Sun, 2024). They also resonate with emerging growth pathways—such as urban renewal, tourism development, and ecological initiatives—suggesting that vitality reconfiguration may occur through non-traditional growth logics and value redistribution.

Thus, shrinkage should not be equated with inevitable decline. Its impact on urban vitality is shaped by a complex interplay of development trajectories, industrial structures, demographic transitions, and institutional responses. It is therefore essential to identify both structural vulnerabilities and revitalisation potential based on city-specific characteristics, and to promote differentiated governance and spatial transformation under the paradigm of ‘smart shrinkage’.

These findings cannot be fully identified through a single-scale perspective, but rather emerge through the dual-scale urban–regional analytical framework introduced in this study. This framework not only addresses the limitations of previous related studies that have relied heavily on administratively defined units (Sun et al. 2023; Zhang et al. 2024) but also extends the understanding of vitality to the regional scale, offering a more comprehensive insight into the dynamics of urban agglomerations. To facilitate urban management and planning, it is necessary to examine urban vitality from the perspective of administratively defined units. Simultaneously, we move beyond traditional administrative boundaries by focusing on urban agglomerations from a regional perspective. In the context of deepening globalisation, cities are increasingly transcending their formal territorial boundaries. Regional-scale analysis facilitates the integration of high-precision, multi-source geospatial data (Liu et al. 2022, 2023), enabling a more nuanced understanding of the relationship between human activity and urban vitality while addressing prior limitations in capturing mobility dynamics. Adopting human activity as the analytical unit offers a novel lens for urban spatial research (Liu et al. 2015).

Through this dual-scale analysis, we identify a ‘dual-core–satellite’ structure of urban vitality within the HCUA. At the urban scale, the two core cities exhibit clear advantages across vitality dimensions, while low-vitality cities form a north–south strip encircling the cores, creating a centripetal spatial pattern. This observation is further corroborated by regional-scale analysis and aligns with previous research findings that have highlighted pronounced spatial differentiation and central agglomeration tendencies in the region (Chen, 2020; Du et al. 2022). These patterns suggest an initial stabilisation of vitality disparities within the urban agglomeration, while also revealing problems of regional disconnection and structural imbalance. The multidimensional evaluation system also enables the identification of structural shortfalls and coordination features in vitality, thereby supporting more targeted local governance and transformation planning. In particular, vitality levels are shaped by the dual influence of regional homogeneity and internal heterogeneity. During the early stages of agglomeration development, shared histories, resources, and policy directions tend to result in convergent vitality levels. Over time, however, differentiated strategies contribute to intra-city divergence and selective improvement. For instance, Daqing has developed an economy-driven vitality axis based on petroleum resources, while Qiqihar has improved its vitality ranking—from seventh in 2016 to fourth in 2022—by strengthening environmental and social dimensions. These cases demonstrate that urban vitality evolves through dynamic reconstruction shaped by both regional structures and local adaptation.

Building on this framework, we further analysed population heatmap data, which revealed a pronounced ‘strong daytime, weak nighttime’ pattern and a clear tendency toward ‘warm-seeking and cold-avoiding’. Low temperatures, population ageing, and youth outmigration have suppressed night-time and winter vitality, whereas summer and daytime vitality exhibited broader spatial diffusion, with notable improvements in peripheral cities. These findings illustrate the complex interaction between climate conditions, human activity, and spatial vitality rhythms, while also highlighting the latent potential for revitalising vitality through night-time and winter economies.

Finally, through obstacle degree identification and regression analysis, this study reveals the core bottleneck factors within the urban vitality system, as well as the key built environment variables influencing urban vitality. At the urban scale, cultural vitality consistently emerges as a primary constraint, while economic weakness also proves to be a major limiting factor—highlighting the deep-rooted structural restrictions imposed by factors such as cultural and technological capacity, public services, and fiscal capacity. At the regional scale, elements of the built environment—including road network density, building density, residential service facilities, and transportation facilities—are identified as key contributors to vitality. These factors help enhance spatial accessibility and foster the clustering of human activities, thereby providing foundational spatial support to address the resource hollowing effects associated with urban shrinkage.

Implications

The primary theoretical contribution of this study lies in its evaluation of urban vitality under conditions of shrinkage through the development of a dual-scale analytical framework that bridges individual cities and the broader regional context. We argue that the transformative potential of urban agglomerations stems from spatial tensions and intercity coupling effects triggered by uneven development. While such internal asymmetries may lead to regional fragmentation, they also create opportunities for functional complementarity and structural realignment. In this regard, the proposed dual-scale framework not only elucidates the multidimensional vitality of individual cities but also uncovers their relative positions and functional roles within the broader intercity network—effectively overcoming the limitations of conventional spatial analyses based on administrative boundaries. Building on this, the study further moves beyond static interpretations of the built environment by integrating multi-source spatial data, such as population heatmaps and POI distributions, to address insufficient consideration of dynamic spatial processes in traditional vitality assessments. These contributions offer a new lens to transcend the dominant linear narrative of ‘shrinkage equals decline’ and advance a more nuanced understanding of the complex interplay between urban shrinkage and vitality.

Beyond its academic significance, this study also provides practical insights for enhancing vitality across urban agglomerations. To promote overall development, it is essential to advance cultural and technological innovation, upgrade industries, and improve infrastructure and public services to enhance urban competitiveness, liveability, and resilience. In view of significant seasonal fluctuations, locally tailored and diversified industrial pathways should be explored. Given the marked spatial heterogeneity of vitality, governance strategies must facilitate the functional spillover of core cities and strengthen regional coordination through industrial guidance, resource integration, and fiscal rebalancing. For low-vitality cities, leveraging local resource endowments to develop distinctive industries—particularly in cultural and tourism sectors with greater resilience—may offer critical entry points. Targeted optimisation of these subsystems may, in turn, lift overall vitality and facilitate sustainable transitions within the urban agglomeration. This study not only deepens the understanding of urban agglomeration development in Northeast China but also offers a valuable analytical approach and governance framework for other similarly shrinking regions. The transferability of this framework requires full consideration of local spatial conditions, institutional contexts, and objective environmental factors—such as climate, geographical features, and transportation accessibility—so as to formulate context-specific strategies that enhance its practical feasibility and policy relevance.

Limitations and future perspectives

Several limitations remain, primarily stemming from current data availability and the practical considerations involved in application. First, although the construction of the urban vitality evaluation system aimed to be comprehensive and minimise subjectivity, the validity and relative contribution of certain indicators require further validation. While the existing four-dimensional structure offers general applicability, it risks becoming overly generic and lacking explanatory power if not adequately contextualised within specific urban development settings. Advancing the theoretical foundations and developing models with greater explanatory depth will be essential for future studies. Second, although population heatmap data effectively capture patterns of human activity, the inherent mobility and volatility of such data may undermine the temporal stability and analytical precision of the results. At the regional scale, future studies should consider integrating higher-resolution spatial units, such as 1 × 1 km grids, along with emerging data sources like mobile signalling, GPS trajectories, and street-level imagery to refine interpretations of urban vitality and strengthen its policy significance. Moreover, the complex and evolving relationship between urban shrinkage and vitality merits further investigation. Classifying shrinkage trajectories and analysing their associated vitality pathways could enhance the capacity of academic research to inform targeted governance strategies and revitalisation efforts in shrinking cities.

Conclusions

Set in the HCUA of Northeast China, this study develops a dual-scale framework to explore the underlying relationships linking urban shrinkage and vitality, and to identify the dynamics and driving factors shaping urban vitality in the context of shrinkage. The main findings are as follows:

  1. (1)

    Urban shrinkage and urban vitality are not inherently in linear opposition. The empirical results reveal pronounced urban shrinkage within the HCUA, with 90.91% of major cities experiencing sustained population decline between 2016 and 2022. This pattern underscores the pervasiveness and persistence of shrinkage across the region in recent years. Nevertheless, core cities have maintained high levels of vitality, and certain shrinking cities have exhibited signs of vitality recovery in specific dimensions. This suggests that vitality can be sustained and enhanced under appropriate conditions, particularly through institutional responsiveness and adaptive adjustment.

  2. (2)

    Urban vitality displays marked spatial disparities and temporal rhythms. Spatially, the HCUA follows a ‘dual-core-satellite’ vitality structure. Temporally, it exhibits clear patterns of ‘strong daytime, weak nighttime’ and ‘warm-seeking, cold-avoiding’, with seasonal fluctuations characterised by winter suppression and summer expansion, driven by the interplay of social behaviour, climatic conditions, and economic structures.

  3. (3)

    The influencing factors of urban vitality differ across scales. At the urban scale, constraints stem primarily from underdeveloped cultural–technological systems and limited fiscal autonomy. At the regional level, the built environment plays a more prominent role, with road network density, building density, residential service facilities, and transport facilities emerging as key drivers underpinning vitality.

  4. (4)

    The urban–regional dual-scale framework proposed in this study serves not only as a technical evaluation tool but also as an analytical lens for interpreting the complex dynamics of urban vitality under shrinking conditions. By combining statistical indicators with multi-source spatial data, the framework integrates both administrative and regional perspectives, moving beyond rigid boundary constraints to provide a transferable methodology with contextual adaptability for identifying the spatiotemporal patterns and driving factors of urban vitality. We contend that, as urban spaces continue to evolve, recognising the vitality embedded within shrinkage offers a renewed perspective for interpreting the dynamics of cities.