Development trends and driving mechanisms of coordination in the Digitalization–Urban development–low-carbon transition system

1 Introduction

Against the backdrop of intensifying global climate change and the accelerating transition toward sustainable development, cities are undergoing a critical transformation from traditional factor-driven growth to a green, low-carbon, and high-quality development paradigm (Xu et al., 2024; He et al., 2025). As the primary spatial units of energy consumption and carbon emissions, cities serve not only as the key arenas for achieving China’s carbon peaking and carbon neutrality goals but also as essential platforms for balancing economic growth, resource utilization, and environmental protection (Yu et al., 2024). Consequently, promoting coordinated development among Digitalization, Urban Development, and Low-carbon Transition has become a critical challenge for urban sustainable development.

Meanwhile, the rapid development of digital technologies, represented by big data, intelligent manufacturing, and artificial intelligence, is fundamentally reshaping resource allocation, industrial organization, and urban governance (Lai et al., 2025; Li L. et al., 2026; Zhao Y. et al., 2026; Yu et al., 2026). By improving information flows, fostering technological innovation, and facilitating industrial upgrading, Digitalization provides new momentum for urban sustainable transformation (Yuan et al., 2025). However, its effects are not unidirectional. On the one hand, digital technologies can improve resource allocation efficiency (Che et al., 2026), accelerate the diffusion of green technologies, and reduce energy consumption and carbon emission intensity. On the other hand, the construction of digital infrastructure and the operation of data centers may generate additional energy demand and increase energy consumption (Cao et al., 2026). Consequently, the relationships among Digitalization, Urban Development, and Low-carbon Transition extend beyond simple promotion effects. Instead, they constitute a complex system characterized by continuous evolution and reciprocal feedback through resource allocation, technological innovation, industrial upgrading, and green development. Nevertheless, the coordination mechanisms and evolutionary patterns underlying this system remain insufficiently understood.

Existing studies generally recognize that Digitalization, Urban Development, and Low-carbon Transition constitute three fundamental dimensions of urban sustainable development. However, current research has largely focused on relatively independent themes, such as digital empowerment, urban transformation, and low-carbon governance, and has gradually evolved from single-system analyses to investigations of dual-system interactions (Zhao X. et al., 2026). Regarding the relationship between Digitalization and Low-carbon Transition, most studies have shown that digital technologies can enhance energy efficiency by fostering technological innovation, optimizing resource allocation, and facilitating industrial upgrading, thereby promoting low-carbon development. Nevertheless, other studies have argued that the energy consumption associated with digital infrastructure construction and data center operation may offset these emission reduction benefits, highlighting the complexity and context dependence of Digitalization’s impacts on Low-carbon Transition (Huang W. et al., 2025). Concerning the relationship between Urban Development and Low-carbon Transition, existing studies generally suggest that economic growth, population agglomeration, and infrastructure improvement can enhance resource allocation efficiency and support green transformation. However, rapid urban expansion may simultaneously intensify energy consumption and environmental pressure, indicating that the relationship between Urban Development and Low-carbon Transition is neither linear nor uniform, but instead exhibits distinct stage-specific and regional characteristics (Shi et al., 2026; Ji et al., 2024). Building on these findings, recent studies have increasingly adopted coordination and coupling frameworks to examine interactions among multiple subsystems. Consequently, research has gradually expanded from dual-system relationships, such as economy–environment and Digitalization–Low-carbon Transition, to multi-system coordinated development, providing new perspectives for understanding the coordinated evolution of complex urban systems (Wang and Di, 2025).

Overall, although existing studies have substantially advanced the understanding of interactions among different urban systems, they remain largely confined to identifying system correlations and coordination levels, offering limited insights into how the long-term coordination among Digitalization, Urban Development, and Low-carbon Transition is formed, evolves over time, and gives rise to regional disparities (Feng et al., 2026). Specifically, three major research gaps remain. First, existing studies predominantly focus on unidirectional effects or dual-system relationships, with insufficient attention paid to the long-term interactions within the Digitalization–Urban Development–Low-carbon Transition (DUL) system. Second, current research mainly emphasizes development assessment or static coordination evaluation, while systematic investigations into the dynamic evolution and regional disparities of long-term coordination within the DUL system remain limited. Third, studies on driving mechanisms are largely restricted to the effects of individual factors, with inadequate consideration of multi-factor interactions and their dynamic changes, resulting in an incomplete understanding of the mechanisms underlying the long-term coordinated development of the DUL system. Therefore, it is essential to examine the long-term coordination among Digitalization, Urban Development, and Low-carbon Transition from a systems perspective and to further elucidate its dynamic evolution, regional disparities, and driving mechanisms.

To address the above research gaps, this study investigates 289 prefecture-level and above cities in China and develops an evaluation framework for the DUL system. Using panel data spanning 2010–2024, it systematically examines the dynamic evolution, regional disparities, and driving mechanisms of the long-term coordinated development of the DUL system. Specifically, the CRITIC-Entropy Weight Method is first employed to quantify the development levels of Digitalization, Urban Development, Low-carbon Transition, and the overall DUL system. Subsequently, the Local and Tele-coupling Coordination Degree (LTCCD) is adopted to characterize the long-term coordination and the evolution of coordination states within the DUL system. Building upon this, the Dagum Gini coefficient is applied to identify regional disparities in LTCCD and their sources. Finally, GeoDetector is employed to examine the driving factors of LTCCD spatial disparities and their interaction effects, thereby systematically revealing the evolutionary patterns and formation mechanisms underlying the long-term coordinated development of the DUL system.

Accordingly, this study aims to: (1) characterize the dynamic evolution of the development levels of Digitalization, Urban Development, Low-carbon Transition, and the DUL system; (2) identify the long-term coordination of the DUL system and the evolution of its coordination states; (3) reveal the regional disparities in LTCCD and their structural sources; and (4) investigate the key driving factors of LTCCD spatial disparities and their interaction effects, thereby providing empirical evidence for understanding the formation mechanisms underlying the long-term coordinated development of the DUL system.

Compared with existing studies, this study makes three main contributions. First, it develops a theoretical framework for the coordinated development of the DUL system by integrating Digitalization, Urban Development, and Low-carbon Transition into a unified analytical framework, thereby clarifying their long-term interactions and coordinated evolution and extending the systems perspective on urban low-carbon transition. Second, based on panel data for 289 prefecture-level and above cities in China from 2010 to 2024, this study introduces LTCCD to characterize the dynamic evolution of the long-term coordination of the DUL system, addressing the limitations of previous studies that primarily relied on static analyses based on single-year data. Third, by combining the Dagum Gini coefficient and GeoDetector, this study reveals the spatial disparities in LTCCD and their underlying mechanisms from the perspectives of regional disparities and driving mechanisms, providing new empirical evidence for promoting coordinated urban development in the context of Digitalization.

The remainder of this paper is organized as follows. Section 2 presents the theoretical framework and indicator system. Section 3 describes the research methods. Section 4 reports the empirical results. Section 5 discusses the main findings. Finally, Section 6 concludes the paper.

2 Theoretical framework and indicator system

2.1 Study area and data sources

This study focuses on prefecture-level and above cities in China (Figure 1), covering the eastern, central, western, and northeastern regions. Against the backdrop of rapid urbanization, green transition, and digital economy development, Chinese cities exhibit substantial disparities in economic development, infrastructure conditions, and low-carbon transition pathways (He et al., 2022). Therefore, the study area provides broad geographical coverage and effectively captures the overall characteristics and regional heterogeneity of Digitalization, Urban Development, and Low-carbon Transition across Chinese cities.

The study period spans from 2010 to 2024. To construct a consistent panel dataset, only cities with relatively stable statistical standards and complete data for the major indicators throughout the study period were retained, including all four municipalities directly under the Central Government. Cities affected by administrative boundary adjustments, changes in statistical standards, or long-term missing data for key indicators that prevented the construction of a complete time series were excluded. Consequently, a final sample of 289 prefecture-level and above cities was obtained for the empirical analysis.

The data were primarily obtained from the China City Statistical Yearbook, China Statistical Yearbook, China Urban Construction Statistical Yearbook, China Environmental Statistical Yearbook, statistical yearbooks and statistical bulletins of individual provinces (autonomous regions and municipalities), as well as the database of the National Bureau of Statistics of China. To ensure data consistency and comparability, the statistical standards of all indicators were first harmonized, followed by consistency checks and indicator standardization to eliminate differences in measurement scales.

2.2 Theoretical framework

Digitalization, Urban Development, and Low-carbon Transition represent the three core dimensions of urban technological transformation, urban system evolution, and green development, respectively, and together constitute essential components of urban sustainable development (Chen et al., 2026; Tian, 2026). Rather than evolving as independent processes, these three subsystems form an integrated system characterized by mutual dependence and mutual support under the overarching goal of urban sustainable development. Specifically, Digitalization provides technological support and governance capacity for Urban Development and Low-carbon Transition; Urban Development offers the economic foundation and spatial platform for the application of digital technologies and the advancement of low-carbon transition; and Low-carbon Transition, through the demand for green development and resource–environment constraints, further promotes digital technological innovation and the optimization of urban development pathways. Therefore, the long-term coordinated development of the DUL system does not arise from the simple synchronous improvement of the three subsystems, but rather from their continuous functional complementarity, resource sharing, and feedback regulation. Driven by external forces such as the policy environment, technological progress, market mechanisms, and societal demand, the three subsystems gradually evolve into a DUL system characterized by long-term coordination through continuous interactions. The theoretical framework of the DUL system is illustrated in Figure 2.

The long-term coordinated development of the DUL system follows the fundamental logic of “digital empowerment–urban response–green feedback.” Digitalization, supported by digital infrastructure, the digital industry, and digital innovation, enhances information flow, knowledge spillovers, and resource allocation efficiency, thereby providing technological empowerment for Urban Development and Low-carbon Transition (Egba et al., 2026). Urban Development, through population agglomeration, economic growth, infrastructure improvement, and human capital accumulation, transforms the advantages of digital technologies into industrial upgrading, improved public services, and green investment capacity, thereby creating favorable conditions for Low-carbon Transition (Luo et al., 2026). Low-carbon Transition, by improving energy efficiency, strengthening environmental governance, and promoting the development of green industries, continuously alleviates resource–environment constraints and, in turn, generates feedback demand for the application of digital technologies and the optimization of urban development pathways, thereby constituting the green feedback mechanism (Fan and Wu, 2025; Wang et al., 2025). Throughout this process, Digitalization, Urban Development, and Low-carbon Transition perform distinct functions of technological empowerment, system response, and green feedback, respectively. Through continuous resource flows, technology diffusion, and system feedback, the interactions among the three subsystems are progressively strengthened, enabling the DUL system to exhibit sustained interactions and coordinated evolution.

As digital technologies continue to diffuse, urban capacities improve, and green governance advances, the relationships among the three subsystems gradually evolve from unidirectional promotion to bidirectional feedback, ultimately forming a dynamic coupling relationship. On the one hand, Digitalization continuously provides technological and informational support for Urban Development and Low-carbon Transition; Urban Development continuously strengthens the application of digital technologies and the capacity for low-carbon transition; and the demand for green development together with resource–environment constraints generated by Low-carbon Transition creates new conditions for the application of digital technologies and the optimization of urban development pathways (Tan et al., 2025). On the other hand, these interactions do not occur instantaneously but are progressively reinforced through the cumulative processes of resource allocation optimization, technology diffusion, green feedback, and system adaptation. Therefore, the coordinated development defined in this study does not refer to the simple synchronous improvement in the development levels of Digitalization, Urban Development, and Low-carbon Transition. Rather, it refers to a long-term coordination relationship that is gradually established through continuous interactions, dynamic coupling, and coordinated adaptation during the long-term evolution of the three subsystems. It should be noted that the above pathways represent the theoretical mechanism underlying the formation of long-term coordination in the DUL system. The LTCCD adopted in this study is intended to characterize the degree of system compatibility and the state of coordination evolution among the three subsystems, rather than to identify direct causal effects between them.

2.3 Indicator system

2.3.1 Digitalization

Based on the above theoretical framework, the indicator system for the DUL system was constructed, as presented in Table 1. In this study, Digitalization is defined as the comprehensive manifestation of urban digital infrastructure, digital industry development, and digital innovation capacity, reflecting the overall level at which digital technologies empower urban economic and social development (Zhao and Duan, 2026). From a theoretical perspective, Digitalization is not only manifested in the development of digital infrastructure, but also embodied in the industrialization of digital technologies and their capacity for innovation diffusion. Digital infrastructure provides the fundamental support for data transmission and information connectivity; the digital industry facilitates the deep integration of digital technologies with economic activities; and digital innovation determines the continuous diffusion and application of digital technologies. Together, these three dimensions constitute the core components of Digitalization. Accordingly, the Digitalization evaluation system was established based on three dimensions: digital infrastructure, digital industry, and digital innovation (Moghayedi, 2025).

TABLE 1. DUL system indicator system.

Specifically, the number of mobile phone subscriptions, the number of broadband Internet subscribers, and the density of long-distance optical cables respectively reflect the coverage of urban mobile communication, fixed broadband networks, and information transmission infrastructure, serving as important proxy indicators for measuring urban digital connectivity and the provision capacity of digital infrastructure. The proportion of telecommunications service revenue reflects the market scale of digital communication services and the level of digital economic activities, thereby measuring the development of the digital industry. The proportion of employees in the information transmission, software, and information technology service sectors represents the concentration of human capital and technological innovation support within the digital industry, providing an important basis for evaluating digital innovation capacity. These indicators correspond to the three dimensions of digital infrastructure, digital industry, and digital innovation, respectively, and together provide a comprehensive characterization of the level of urban Digitalization.

2.3.2 Urban development

In this study, Urban Development is defined as the comprehensive level of development reflected in population agglomeration, economic development, infrastructure, and public services, representing a city’s overall capacity to accommodate socioeconomic activities and support sustainable development (Ding et al., 2026). From a theoretical perspective, Urban Development is inherently a process of the coordinated evolution of population, economic, spatial, and public resources. Specifically, population agglomeration reflects a city’s capacity to concentrate production factors; economic development represents urban economic vitality; infrastructure determines the city’s spatial carrying capacity; and public services reflect its capacity to safeguard residents' wellbeing and sustain long-term development. Together, these four dimensions comprehensively capture the conceptual connotation of Urban Development. Accordingly, the Urban Development evaluation system was established based on four dimensions: population agglomeration, economic development, infrastructure, and public services (Chen et al., 2025).

Specifically, the urbanization rate reflects the level of population agglomeration and the city’s capacity to concentrate production factors. Per capita disposable income of urban residents reflects residents' income levels and the quality of urban economic development. Urban road area represents the provision of urban transportation infrastructure and the city’s spatial carrying capacity, providing the physical foundation for population mobility, economic activities, and public service delivery. Fixed asset investment in municipal public facilities reflects the city’s capacity to continuously invest in infrastructure and public service facilities, serving as an important proxy indicator for measuring infrastructure development and public service provision. The number of university students per 10,000 inhabitants reflects the level of educational resource provision and human capital accumulation, providing talent support for urban innovation and sustainable development. These indicators correspond to the four dimensions of population agglomeration, economic development, infrastructure, and public services, respectively, and together provide a comprehensive characterization of the level of Urban Development and its overall carrying capacity.

2.3.3 Low-carbon transition

In this study, Low-carbon Transition is defined as the process through which cities promote the coordinated development of socioeconomic progress and resource–environment sustainability by improving energy efficiency, reducing carbon emission intensity, and enhancing environmental quality, thereby reflecting the overall level of urban green and low-carbon development (Tamasiga et al., 2026). From a theoretical perspective, Low-carbon Transition is not only manifested in the reduction of carbon emissions, but also embodied in the systematic improvement of energy utilization patterns, environmental governance capacity, and ecological development. Specifically, energy efficiency reflects the transformation of resource utilization patterns; carbon emissions represent emission reduction performance; environmental governance reflects the government’s capacity for green governance; and ecological development represents the foundation of green development. Together, these four dimensions comprehensively capture the overall level of Low-carbon Transition. Accordingly, the Low-carbon Transition evaluation system was established based on four dimensions: energy efficiency, carbon emissions, environmental governance, and ecological development (Zhang et al., 2025).

Specifically, energy consumption per unit of GDP and carbon emissions per unit of GDP respectively reflect urban energy use efficiency and carbon emission intensity, serving as the core proxy indicators for evaluating low-carbon development performance. PM₂.₅ concentration reflects the level of urban air pollution control and can be used to characterize improvements in environmental quality. The proportion of environmental governance expenditure to total fiscal expenditure reflects the intensity of governmental investment in environmental governance, indicating the policy support capacity for advancing urban low-carbon governance. The green coverage rate of built-up areas reflects the level of urban ecological development and the capacity for ecological environment improvement. These indicators correspond to the four dimensions of energy efficiency, carbon emissions, environmental governance, and ecological development, respectively, and together provide a systematic characterization of the level of Low-carbon Transition.

3 Methodology

To systematically investigate the evolutionary patterns, regional disparities, and driving mechanisms underlying the long-term coordinated development of the DUL system, this study develops an integrated analytical framework consisting of development assessment, coordination evaluation, disparity decomposition, and mechanism identification. Specifically, the CRITIC-Entropy Weight Method is employed to measure the development levels of Digitalization, Urban Development, and Low-carbon Transition, as well as the overall development level of the DUL system. Subsequently, LTCCD is adopted to characterize the long-term coordination among the three subsystems and the evolution of coordination states. Building upon this, the Dagum Gini coefficient is applied to decompose the regional disparities in LTCCD and identify their sources. Finally, GeoDetector is employed to identify the key driving factors of LTCCD spatial disparities and their interaction effects. By integrating these methods, this study provides a systematic understanding of the long-term coordinated development of the DUL system from the perspectives of dynamic evolution, regional disparities, and multi-factor synergy. The overall research framework is illustrated in Figure 3.

3.1 CRITIC-entropy weight method

The weighting scheme directly affects the reliability of the composite index, whereas a single objective weighting method may introduce methodological bias. The Entropy Weight Method emphasizes the degree of information dispersion among indicators and effectively reduces subjective interference, but it does not account for the correlations among indicators. In contrast, the CRITIC method captures both the contrast intensity and conflict among indicators, although it is more sensitive to sample distribution. To simultaneously account for information content and the structural characteristics of the indicator system, this study employs the CRITIC-Entropy Weight Method to determine the weights of the DUL system, thereby improving the robustness and interpretability of the composite index (Guo et al., 2025). The calculation procedures are as follows:

• Entropy weight method

First, standardize the data since differences in data dimensions can affect the accuracy of calculations. Therefore, the raw data need to be normalized to dimensionless values. The normalization method uses the extreme value method, as shown in Formulas 1, 2.

x i j = r i j − Min i r i j Max i r i j − Min i r i j (1)

x i j = Max i r i j − r i j Max i r i j − Min i r i j (2)

Among them, Max i r i j and Min i r i j represent the maximum and minimum values, respectively, under the same indicator i. x i j represents the value of the j-th city under the i-th indicator after normalization. For positive indicators, normalization is performed using Formula 1, where a higher value in the indicator indicates better performance; for negative indicators, normalization is performed using Formula 2, where a lower value in the indicator indicates better performance.

The entropy value e i of the indicator is calculated using Formula 3:

e i = − k ∑ j = 1 m p i j · ⁡ ln ⁡ p i j (3)

Among them, the coefficients k and p i j are calculated using Equations 4 and 5 respectively.

k = 1 ln ⁡ m (4)

p i j = x i j ∑ j = 1 m x i j (5)

The weight w i of a single indicator is calculated using Formula 6

w 1 i = 1 − e i / ∑ i = 1 n 1 − e i (6)

Where w i is the weight value of index i, and n is the total number of indices.

• 2. CRITIC weighting method

The CRITIC method considers both the contrast intensity and conflict among indicators, and the amount of information of the indicators is defined by Equation 7.

C i = σ i ∑ k = 1 m 1 − r i k (7)

Where σ i represents the standard deviation of index i, and r i k denotes the correlation coefficient between indices i and k. The CRITIC weight is calculated as shown in Equation 8.

w 2 i = C i ∑ i = 1 m C i (8)

• 3. CRITIC-entropy weight combination weighting

To integrate the complementary strengths of the two objective weighting methods, the final weights were determined using a linear combination approach, as expressed in Equation 9. Since there is currently no sufficient theoretical basis for assigning a higher combination weight to either objective weighting method, an equal-weight combination scheme (0.5:0.5) was adopted to avoid introducing additional subjective preferences while simultaneously accounting for both indicator dispersion and inter-indicator conflict. Furthermore, a sensitivity analysis was conducted by varying the combination proportions of the two weighting methods to examine whether the equal-weight setting would materially influence the evaluation results.

w i = γ w 1 i + 1 − γ w 2 i (9)

Where γ represents the weight proportion of the entropy weight method. To examine the influence of the equal-weight combination of the CRITIC method and the entropy weight method on the evaluation results, the combination coefficient of the entropy weight method was varied from 0 to 1 with an interval of 0.1. The Spearman rank correlation coefficients between the city rankings of the D, U, L, and DUL indices under each alternative weighting scheme and those under the baseline scheme were then calculated. The results are presented in Table 2. Within the reasonable perturbation range of γ = 0.3–0.7, the minimum Spearman rank correlation coefficients for the D, U, L, and DUL indices were 0.9982, 0.9994, 0.9864, and 0.9925, respectively, indicating a high degree of consistency across all four evaluation results. Even under the pure CRITIC or pure entropy weight schemes, the rankings of the D, U, and DUL indices remained highly correlated, with the corresponding correlation coefficients for the DUL index reaching 0.9741 and 0.9471, respectively. Overall, varying the combination proportions of the two objective weighting methods did not substantially alter the overall ranking structure of cities, indicating that the evaluation results are not dependent on the specific equal-weight setting of 0.5:0.5.

TABLE 2. Spearman correlation coefficients of evaluation results under different weighting combinations.

3.2 Local and tele-coupling coordination degree model (LTCCD)

Building upon the conventional Coupling Coordination Degree (CCD) model, this study incorporates the inverse distance spatial weighting method with reference to previous studies to further calculate the LTCCD of the DUL system (Liu Y. et al., 2025). Compared with the conventional CCD model, the LTCCD framework objectively incorporates the intensity of spatial interactions through a distance-decay function, thereby distinguishing within-system coupling from telecoupling effects among cities. The LTCCD is employed to characterize the long-term coordination relationship and its evolutionary characteristics among Digitalization, Urban Development, and Low-carbon Transition. A higher LTCCD indicates a more coordinated development relationship among the three subsystems; however, it does not imply that any one subsystem exerts a direct causal effect on the others. The calculation formula is as follows Equation 10:where U E 1, U E 2, and U E 3 denote the development levels of Digitalization, Urban Development, and Low-carbon Transition, respectively, and C U E 1 , U E 2 , U E 3 represents the coupling degree of the three subsystems. Before calculating the coupling degree, U E 1, U E 2, and U E 3 are normalized to ensure that all values range from 0 to 1. Subsequently, the comprehensive development index and the coordination degree are calculated. Assuming that the three subsystems are of equal importance, the calculation formula is as follows Equations 11–13:

C U E 1 , U E 2 , U E 3 = U E 1 × U E 2 × U E 3 3 U E 1 + U E 2 + U E 3 3 (10)

T = ∑ i = 1 n α i × U E i , T = ∑ i = 1 n α i = 1 (11)

α 1 = α 2 = α 3 = 1 / 3 (12)

D = C × T (13)

Building upon the CCD model, this study employs ArcGIS and incorporates the inverse distance weighting interpolation method to calculate the local coupling degree and tele-coupling degree of the DUL system, which are subsequently integrated to obtain LTCCD. The local and tele-coupling degrees are defined follows Equations 14–18: where C s i denotes the LTCCD of city i. In this study, the local and tele-coupling components are assumed to contribute equally to the dynamic coordinated development of the DUL system; therefore, the corresponding weights are set as μ = ε = 1 / 2. Here, i denotes a given city, n is the total number of cities, and k is an integer ranging from one to n (k ≠ i). C tele i , k represents the tele-coupling effect between city i and city k, while W i k denotes the inverse distance weight between the two cities, which is defined as follows:where d i k denotes the distance between the administrative centers of cities i and k, and P is the distance decay exponent. Following previous studies, P is set to 2.where T s i denotes the local and tele-coupling comprehensive development index of the DUL system. T 1 and T 2 represent the local and tele-coupling comprehensive development indices, respectively. T tele i , k denotes the tele-coupling interaction between cities i and k.where C s i, T s i, and D s i denote the local and tele-coupling degree, the local and tele-coupling comprehensive development index, and the LTCCD of the DUL system, respectively. Based on the value of D s i, the coordination level is classified into ten categories, as presented in Table 3.

C s i = μ C U E 1 i , U E 2 i , U E 3 i + ε ∑ k = 1 , k ≠ i n W i k × C t e l e i , k + ∑ k = 1 , k ≠ i n W k i × C t e l e k , i 2 (14)

C t e l e i , k = 3 U E 1 , i × U E 2 k × U E 3 k 3 U E 1 i + U E 2 k + U E 3 k (15)

W i k = d i k − p ∑ k = 1 n d i k − p (16)

T s i = μ T U E 1 i , U E 2 i , U E 3 i + ε ∑ k = 1 , k ≠ i n W i k × T t e l e i , k + ∑ k = 1 , k ≠ i n W k i × T t e l e k , i 2 (17)

D s i = C s i × T s i (18)

TABLE 3. Classification of coupling coordination levels.

3.3 Inequality analysis

To characterize the uneven development of LTCCD across cities, this study employs the Dagum Gini coefficient to quantify and decompose regional disparities in LTCCD. The Gini coefficient effectively reflects the degree of dispersion in the distribution of samples, with values ranging from 0 to 1. A higher value indicates greater disparities in development among cities, whereas a lower value indicates smaller disparities and a more balanced development level (Yang et al., 2025). The specific calculation formulas such as Equations 19–28:

G = ∑ j = 1 k ∑ h = 1 k ∑ i = 1 n j ∑ r = 1 n h y j i − y h r / 2 n 2 y ¯ (19)

Y ¯ h ≤ … Y ¯ j ≤ … ≤ Y ¯ k (20)

The overall Gini coefficient G can be decomposed into three components: the contribution from within-region differences G w, the contribution from between-region differences G n b, and the contribution from excess density G t, with G = G w + G n b + G t. G j j represents the Gini coefficient of region j, G j h denotes the Gini coefficient between region j and region h, and D j h signifies the mutual influence between regions j and h.

G j j = 1 2 Y ¯ j ∑ i = 1 n j ∑ r = 1 n j y j i − y j r n j 2 (21)

G j h = ∑ i = 1 n j ∑ r = 1 n h y j i − y h r n j n h Y j + Y h (22)

G w = ∑ j = 1 k G j j P j S j (23)

G n b = ∑ j = 2 k ∑ h = 2 j = 1 G j h P j S h + P h S j D j h (24)

G t = ∑ j = 2 k ∑ h = 1 j − 1 G j h P j S h + P h S j 1 − D j h (25)

D j h = d j h − P j h d j h + P j h (26)

d j h = ∫ 0 ∞ d F j y ∫ 0 y y − x d F h x (27)

P j h = ∫ 0 ∞ d F j y ∫ 0 y y − x d F j x (28)

In the formula, P j = n j / n, S j = n j Y ¯ j n μ, function F represents the cumulative probability density function of LTCCD. d j h denotes the difference in LTCCD values between regions, which is the mathematical expectation of the sum of all sample values where y i j − y h r > 0 in regions j and h. P j h represents the excess first-order moment, which is the mathematical expectation of the sum of all sample values where y i j − y h r < 0 in regions j and h.

3.4 Driving mechanism

To further identify the key driving factors underlying the spatial differentiation of the DUL system, this study employs the GeoDetector model. Based on the theory of spatial stratified heterogeneity, this method measures the explanatory power of independent variables on the spatial distribution of the dependent variable by comparing intra-group variance with overall variance. It has the advantages of not requiring linear assumptions and being capable of identifying multi-factor interaction effects (Guo et al., 2026; Li X. et al., 2026).

• Variable setting

Considering multiple dimensions such as economic foundation, factor endowments, and policy environment, six categories of driving factors are selected: GDP per capita (X1), the share of the tertiary industry (X2), resident population size (X3), the number of newly authorized patents (X4), general public budget expenditure of local governments (X5), and the level of information infrastructure (X6). Specifically, X1 and X2 are used to characterize regional economic development and industrial structure; X3 reflects city size and factor agglomeration effects; X4 represents regional innovation capacity and technological supply; X5 measures government intervention capacity and the level of public resource allocation; and X6 captures the foundation of digital development and its supporting role in improving resource allocation efficiency. These variables collectively characterize urban development conditions from multiple dimensions, including economic level, industrial structure, agglomeration effects, innovation capacity, infrastructure, and institutional regulation, thus providing comprehensive support for uncovering the driving mechanisms of the coordinated evolution of the DUL system.

• b. Factor detection

Factor detection was employed to quantify the explanatory power of each driving factor for the spatial heterogeneity of the coupling coordination degree, with the q-statistic serving as the corresponding measure. The q-statistic in GeoDetector evaluates the degree of spatial consistency and explanatory power between the explanatory factors and the spatial distribution of LTCCD, rather than representing causal effects. Therefore, the results are interpreted as the relative explanatory effects of different factors on the spatial heterogeneity of LTCCD, rather than as evidence of strict causal relationships. The calculation formula is as follows Equation 29:

q = 1 − ∑ h = 1 L N h σ h 2 N σ 2 (29)

Where q ∈ 0 , 1, h denotes the stratified unit, L represents the number of stratifications, N h and N represent the number of samples in the h th class and the total number of samples, respectively, and σ h 2 and σ 2 represent the within-stratification and total variance, respectively. A larger q value indicates a stronger explanatory power of the factor on the spatial distribution of the dependent variable. The statistical significance of the q-values for all explanatory variables was evaluated using permutation tests. Only factors that passed the significance test were retained for subsequent analysis.

• c. Interactive probing

Building on the factor detection results, interaction detection is further conducted to examine the synergistic mechanisms among different driving factors. By comparing the q-statistics of individual factors with those obtained from the combination of two factors, the types of interaction effects can be identified. This analysis helps to reveal nonlinear driving pathways arising from the joint influence of multiple factors. The classification of interaction types is presented in Table 4.

• d. Variable discretization

TABLE 4. Factor interaction results.

Given that the GeoDetector requires explanatory variables to be categorical, continuous variables are discretized using the natural breaks (Jenks) method and uniformly classified into five categories. This approach maximizes between-group differences while minimizing within-group variance, thereby improving the robustness and interpretability of the detection results.

4 Results

4.1 Spatiotemporal evolution of the DUL system

4.1.1 Evolution of subsystem

Figure 4 illustrates the development levels and spatiotemporal evolution of the three subsystems from 2010 to 2024. Overall, all three subsystems exhibited pronounced temporal evolutionary characteristics; however, significant differences were observed in their growth rates, spatial diffusion patterns, and degrees of regional disparity. These findings indicate that the three subsystems followed distinct evolutionary trajectories during China’s urban sustainable transition, rather than exhibiting completely synchronized development.

From the perspective of temporal evolution, the violin distributions of all three subsystems generally shifted toward higher values over the study period, indicating a continuous improvement in the overall development level of Chinese cities. Among them, the Digitalization subsystem exhibited the most pronounced rightward shift in its distribution, with the pace of improvement accelerating after 2018, reflecting the continued advancement of digital infrastructure, the widespread adoption of information and communication technologies, and the expansion of the digital economy. In comparison, the Urban Development subsystem maintained a steady upward trend, with its distribution center gradually shifting upward and the number of high-value cities continuously increasing, indicating a progressive enhancement in comprehensive urban development capacity. The Low-carbon Transition subsystem also exhibited stable growth; however, changes in its distribution pattern were relatively modest, suggesting that low-carbon transition was characterized by gradual long-term improvement rather than rapid short-term advancement.

The spatial evolution patterns further corroborate the temporal trends described above. High-value areas of the Digitalization subsystem remained predominantly concentrated in the eastern coastal regions, including the Yangtze River Delta, the Pearl River Delta, and the Beijing–Tianjin–Hebei region, and gradually expanded toward regional growth poles such as the Middle Yangtze River and the Chengdu–Chongqing region. The spatial pattern evolved from isolated clusters to broader urban agglomerations while maintaining a pronounced east–west gradient, with higher levels in the east than in the west. The spatial pattern of the Urban Development subsystem remained relatively stable. 1n addition to the eastern coastal areas, high-value cities were widely distributed across provincial capitals and regional central cities in central and western China. Although medium- and high-value areas continued to expand, the overall pattern of spatial disparity changed only marginally, indicating that comprehensive urban development exhibited strong path-dependent characteristics. In contrast, the Low-carbon Transition subsystem exhibited a spatial evolution pattern distinct from those of the other two subsystems. Besides several eastern coastal cities, high-value areas were also distributed across Northeast China and some resource-based cities, resulting in a more dispersed spatial pattern and more pronounced regional heterogeneity. These findings suggest that the spatial distribution of Low-carbon Transition may be associated with multiple factors, including resource endowment, industrial structure, and environmental governance conditions.

4.1.2 Evolution of the comprehensive DUL index

Figure 5 presents the spatial distribution and temporal evolution of the DUL composite development index from 2010 to 2024. Overall, the DUL composite development index exhibited a steady upward trend throughout the study period, indicating a continuous improvement in the overall development level of the DUL system. Nevertheless, from a spatial perspective, considerable disparities in the composite development level persisted across cities.

From the perspective of temporal evolution, the violin distribution of the DUL composite development index generally shifted gradually toward higher values, indicating a continuous improvement in the overall development level of the DUL system across Chinese cities from 2010 to 2024. Meanwhile, the distribution showed no evident convergence, with a pronounced upper tail persisting throughout the study period. This suggests that the composite development level increased more rapidly in some cities, whereas most cities remained concentrated within the low- and medium-level ranges. Overall, although the DUL system followed an upward developmental trajectory, the pace of improvement varied considerably across cities, indicating that urban sustainable transition continued to exhibit a certain degree of developmental divergence.

From the perspective of spatial evolution, the DUL composite development level exhibited a clear spatial pattern characterized by higher levels in eastern China and lower levels in central and western China. In 2010, high-value cities were primarily concentrated in the eastern coastal regions and a few regional central cities, whereas most cities in western China and parts of central China remained at relatively low development levels. By 2018, medium- and high-value areas had expanded, with the composite development level increasing markedly in several provincial capitals and regional central cities in central China. This pattern indicates that the spatial distribution of the DUL composite development level gradually extended from the eastern coastal core toward inland regional growth poles. By 2024, high-value areas became further concentrated within major urban agglomerations, including the Yangtze River Delta, the Pearl River Delta, and the Middle Yangtze River region. Meanwhile, several central cities in western China also exhibited noticeable improvement, although the overall east–west spatial gradient remained evident.

Taken together with the evolutionary characteristics of the three subsystems, the changes in the DUL composite development index comprehensively reflect the expansion of Digitalization, the accumulation of Urban Development, and the improvement of Low-carbon Transition, rather than the development status of any single subsystem. However, owing to differences in the evolutionary rates and spatial diffusion patterns of the three subsystems, the composite development level continued to exhibit pronounced regional heterogeneity despite its overall improvement. These findings indicate that urban sustainable transition is reflected not only in the enhancement of individual subsystem development levels but also in the changing relationships among Digitalization, Urban Development, and Low-carbon Transition during their long-term evolution. Therefore, it is necessary to further examine the evolutionary characteristics of the LTCCD of the DUL system.

4.2 Evolution trend of LTCCD

4.2.1 Spatiotemporal evolution of LTCCD

Figure 6 illustrates the spatial evolution of the LTCCD of the DUL system during the study period. Overall, the LTCCD exhibited a continuous upward trend, with the number of cities at low coordination levels gradually decreasing and those at medium and high coordination levels steadily increasing. This indicates that the long-term coordination relationship among Digitalization, Urban Development, and Low-carbon Transition became progressively stronger over time. Nevertheless, pronounced spatial disparities in the LTCCD persisted across cities, suggesting that the coordination evolution of the DUL system remained characterized by significant regional imbalance.

From the perspective of temporal evolution, most cities exhibited relatively low LTCCD values in 2010, indicating that the long-term coordination status among Digitalization, Urban Development, and Low-carbon Transition remained relatively weak during the early stage of the study period. Between 2014 and 2018, the overall LTCCD increased, accompanied by a growing number of cities at medium coordination levels. Improvements in coordination status gradually extended from the eastern coastal regions to several cities in central China. After 2020, the number of cities with relatively high coordination levels increased further. By 2024, the overall coordination status had improved substantially compared with that in 2010, indicating that the long-term coordination relationship of the DUL system continued to improve throughout the study period.

From the perspective of spatial patterns, the LTCCD generally exhibited a distribution characterized by higher levels in eastern China, followed by central China, and lower levels in western China. The eastern coastal regions, particularly the Yangtze River Delta, the Pearl River Delta, and the Beijing–Tianjin–Hebei urban agglomeration, consistently maintained relatively high coordination levels, indicating a comparatively high degree of coordination among Digitalization, Urban Development, and Low-carbon Transition in these regions. Several provincial capitals and regional central cities in central China experienced gradual improvements in LTCCD, suggesting a certain degree of spatial expansion in medium- and high-coordination areas. In contrast, most cities in western China remained at relatively low coordination levels throughout the study period. This spatial pattern may be associated with regional differences in development foundations, infrastructure conditions, and transition capacity.

Notably, the improvement in LTCCD did not exhibit a uniform nationwide expansion but was primarily concentrated within the eastern coastal urban agglomerations and several inland regional central cities. This finding indicates that the coordinated development of the DUL system is not determined solely by improvements in the development levels of the individual subsystems, but is more closely associated with the establishment of a stable long-term coordination relationship among Digitalization, Urban Development, and Low-carbon Transition. Overall, the LTCCD of the DUL system continued to increase from 2010 to 2024, while the regional gradient remained evident, indicating that urban sustainable transition in China has entered a stage of improving coordination, although the process of coordination evolution continues to exhibit pronounced spatial heterogeneity.

4.2.2 Evolution of coordination stages

Figure 7 further illustrates the evolution of the classification structure of the long-term coordination status of the DUL system from 2010 to 2024. Compared with the continuous LTCCD values, the coordination classification provides a more intuitive representation of the stage-wise transition of cities from maladjustment to coordination. Overall, the coordination status of the DUL system gradually shifted from high incoordination toward moderate incoordination, over the study period. This pattern indicates a continuous improvement in the coordination status among Digitalization, Urban Development, and Low-carbon Transition.

From the perspective of the classification structure, cities were mainly concentrated in the High incoordination and Moderate incoordination states in 2010, with High incoordination accounting for a relatively large proportion, indicating that most cities had not yet formed a relatively stable system coordination relationship during the early stage of the study period. As time progressed, the proportion of cities in High incoordination continuously declined, while Moderate incoordination gradually became the predominant type and remained dominant after 2016. This indicates that the coordination status of the DUL system did not undergo an abrupt transition, but instead experienced a gradual shift from High incoordination to Moderate incoordination. Meanwhile, the proportions of cities in the higher coordination states, including Mild incoordination, Basic coordination, and Favorable coordination, gradually increased, with this trend becoming more pronounced after 2020. Although the proportion of cities that had fully entered higher coordination stages remained limited, the classification structure of the coordination status gradually shifted from being dominated by incoordination in the early period to a pattern characterized by the coexistence of Moderate incoordination, Mild incoordination, and Basic coordination. This indicates that the coordination evolution of the DUL system exhibited a stage-wise improvement, although high-level coordination has not yet been achieved on a large scale.

It is worth noting that the proportions of cities in the Extreme incoordination and Low coordination states remained relatively low throughout the study period, indicating that most cities were not in a state of extreme incoordination but were primarily concentrated in intermediate transitional stages. This finding is consistent with the spatial evolution results of the LTCCD presented above, indicating that the long-term coordination relationship of the DUL system improved overall during the study period, although the improvement process was gradual.

4.3 Analysis of regional differences in LTCCD

To further reveal the evolutionary characteristics of regional disparities in the LTCCD of the DUL system across Chinese cities, the Dagum Gini coefficient was employed to dynamically decompose the overall disparities and their sources. The results are presented in Figure 8. Overall, the regional disparities in LTCCD exhibited a gradual convergence trend from 2010 to 2024; however, disparities in coordination levels among regions remained the primary source of overall spatial inequality.

From the perspective of the evolution of overall disparities (Figure 8a), the overall Dagum Gini coefficient of the LTCCD decreased from 0.119 to 0.105 during the study period, indicating that the long-term coordination level among Chinese cities became gradually more balanced as Digitalization, Urban Development, and Low-carbon Transition continued to advance. Further decomposition revealed that inter-regional disparities exhibited the largest decline, decreasing from 0.058 to 0.048, and constituted the primary source of the reduction in overall disparities. By contrast, intra-regional disparities remained at a relatively low level throughout the study period and showed only a slight downward trend, while transvariation density changed only marginally, indicating that both the coordination disparities within regions and the degree of inter-regional overlap remained relatively stable over the study period.

From the perspective of the contributions of different disparity sources (Figure 8b), inter-regional disparities consistently constituted the largest component of the overall regional disparities, with their contribution remaining above 45% throughout the study period, substantially exceeding those of intra-regional disparities and transvariation density. Meanwhile, the contribution of inter-regional disparities declined slightly over time, whereas the contribution of transvariation density exhibited a modest upward trend, indicating that, as regional coordination levels continued to improve, development gaps associated with traditional regional boundaries gradually weakened, while the overlap in coordination levels among different regions became increasingly pronounced. Nevertheless, the overall contribution structure changed only marginally, and the development gradient among regions remained the primary manifestation and structural source of the spatial disparities in LTCCD.

Further comparison of the evolution of the Dagum Gini coefficients within and between regions (Figure 8c) reveals that intra-regional coordination disparities generally exhibited a declining trend across all regions during the study period. Among them, the eastern region consistently recorded higher intra-regional disparities than the central, northeastern, and western regions, indicating that although cities in the eastern region achieved relatively higher coordination levels, they also exhibited more pronounced internal heterogeneity. In terms of inter-regional disparities, the Dagum Gini coefficient between the eastern and western regions remained the highest throughout the study period, indicating that the coordination gap between eastern and western China continued to represent the most prominent feature of regional disparities in LTCCD. This was followed by the disparities between the eastern and central regions and between the eastern and northeastern regions, whereas the disparities among the central, western, and northeastern regions were relatively small. Notably, the Dagum Gini coefficients for all regional pairs declined to varying degrees, indicating a certain degree of regional convergence in the long-term coordination level. However, the gradient pattern of regional coordinated development has not fundamentally changed.

4.4 Dynamic driving mechanisms of LTCCD

To further identify the driving factors and interaction effects underlying the long-term LTCCD of the DUL system, GeoDetector was applied to conduct factor detection and interaction detection using four representative years (2010, 2016, 2020, and 2024). The results are presented in Figure 9. In the interaction matrix, the diagonal elements represent the explanatory power (q-values) of individual factors, whereas the lower triangular matrix shows the explanatory power of pairwise interactions. It should be noted that the q-values of all explanatory factors were evaluated using permutation tests and were statistically significant at the 1% level (P < 0.01; see Table 5), indicating that each factor significantly explained the spatial heterogeneity of LTCCD.

TABLE 5. Statistical significance of factor detection results.

The factor detection results indicate that all explanatory variables contributed to the spatial heterogeneity of LTCCD, although their explanatory power changed considerably over time. In 2010, most factors exhibited relatively high explanatory power. In particular, the number of authorized patents (X4), local general public budget expenditure (X5), digital infrastructure (X6), and resident population size (X3) showed the highest q-values, suggesting that innovation capacity, government resource allocation, digital infrastructure, and population agglomeration jointly provided strong explanations for the spatial heterogeneity of LTCCD during the early stage. As the study period progressed, the explanatory power of GDP per capita (X1) and the proportion of the tertiary industry (X2) gradually declined, with their q-values decreasing to 0.2018 and 0.1078, respectively, by 2024. This result indicates that the independent contribution of economic scale and industrial structure to the spatial heterogeneity of LTCCD weakened over time. By contrast, authorized patents (X4), local general public budget expenditure (X5), and resident population size (X3) consistently maintained relatively high explanatory power throughout the study period, highlighting the persistent importance of innovation capacity, government regulation, and population agglomeration in promoting the long-term coordinated development of the DUL system. Digital infrastructure (X6) exhibited a stage-dependent pattern. Its q-values reached 0.6167 and 0.5743 in 2010 and 2016, respectively, indicating a strong explanatory effect during the early stage of the study period. By 2024, however, its explanatory power had declined to 0.3184. This result suggests that, as digital infrastructure became more widespread, its direct contribution to spatial heterogeneity gradually diminished, while its role increasingly depended on interactions with other factors.

The interaction detection results further demonstrate that strong synergistic effects generally existed among the explanatory factors. The explanatory power of pairwise interactions consistently exceeded that of individual factors, indicating that the spatial heterogeneity of LTCCD was jointly shaped by multiple interacting factors rather than any single determinant. In 2010, combinations such as X3∩X6, X4∩X5, and X4∩X6 exhibited relatively high explanatory power, suggesting that the interactions among population agglomeration, innovation capacity, government investment, and digital infrastructure played an important role in the coordinated development of the DUL system during the early stage. In 2016, the explanatory power of X3∩X6 and X5∩X6 increased further, highlighting the growing interaction effects between digital infrastructure, population agglomeration, and government investment. In 2020, X3∩X6, X4∩X6, and X4∩X5 continued to exhibit relatively high explanatory power, reflecting the persistent joint effects of innovation, digitalization, and government regulation. By 2024, although the explanatory power of several interaction terms involving X6 declined, combinations such as X4∩X5, X3∩X5, and X3∩X4 remained among the strongest explanatory factors. These findings suggest that the synergistic interactions among innovation capacity, population agglomeration, and government regulation gradually became the dominant pathway underlying the spatial heterogeneity of LTCCD.

5 Discussion

5.1 The dynamic characteristics and formation mechanism of DUL system coordinated evolution

The results indicate that the three subsystems all exhibited continuous improvement during the study period, although their evolutionary rates and spatial diffusion patterns differed substantially, whereas the LTCCD increased continuously. These findings suggest that the long-term coordinated development of the DUL system does not require the three subsystems to exhibit synchronized growth; rather, it reflects the gradual formation of a more stable coordination relationship through their continuous interactions. Therefore, the LTCCD characterizes the state of long-term coordination evolution among Digitalization, Urban Development, and Low-carbon Transition, rather than the direct causal effects among the three subsystems.

This evolutionary pattern can be further interpreted in conjunction with the theoretical framework developed in this study. Digitalization continuously improves information flow efficiency and resource allocation capacity through digital infrastructure, digital industry, and digital innovation, thereby providing technological support for Urban Development and Low-carbon Transition. Urban Development, supported by population agglomeration, infrastructure improvement, and enhanced public services, continuously strengthens the capacity to accommodate the diffusion of digital technologies and the advancement of green transition. Low-carbon Transition, in turn, reinforces the demand for green development by improving energy efficiency, strengthening environmental governance, and promoting the development of green industries, while providing continuous feedback for the application of digital technologies and the optimization of urban development patterns (Tian et al., 2025; Huang M. et al., 2025; Liu X. et al., 2025). As the cumulative effects of resource flows, technology diffusion, and green feedback continue to unfold, the three subsystems gradually evolve from one-way interactions to continuous interaction and dynamic adaptation, thereby contributing to the progressive strengthening of the coordination relationship within the DUL system.

Further comparison of the DUL composite development index and the LTCCD shows that both exhibited a continuous upward trend during the study period; however, they reflect fundamentally different conceptual meanings. The DUL composite development index represents the development status of the three subsystems, whereas the LTCCD reflects the degree of coordination compatibility and the state of long-term coordination evolution among Digitalization, Urban Development, and Low-carbon Transition. Therefore, an improvement in the composite development level does not necessarily imply the automatic formation of a coordination relationship. Only when Digitalization, Urban Development, and Low-carbon Transition continuously strengthen resource sharing, information flows, and system feedback can the coordination status of the DUL system continue to improve. This also explains why the LTCCD continued to increase despite the different developmental trajectories of the three subsystems.

5.2 Regional differential evolution and spatial differentiation formation mechanism

The gradual convergence of the overall LTCCD together with the persistent regional gradient indicates that the coordination evolution of the DUL system in China is characterized by overall improvement accompanied by persistent regional disparities. The eastern region has consistently maintained a relatively high coordination level, not merely because of its higher level of economic development, but more importantly because it possesses relatively well-developed digital infrastructure, urban carrying capacity, and low-carbon governance foundations, resulting in more balanced development conditions across the three subsystems. Stronger digital connectivity and a more robust innovation foundation are conducive to information flows and technology diffusion, while more advanced urban infrastructure and public services provide the supporting conditions for the application of digital technologies and the advancement of the green transition. Meanwhile, greater investment in environmental governance and a higher level of green technology adoption enhance the compatibility between Low-carbon Transition and Urban Development (Lv et al., 2026). Consequently, the eastern region is more likely to maintain a higher level of coordination status within the DUL system.

In contrast, some cities in central and western China still exhibit deficiencies in digital infrastructure, innovation resources, public services, and low-carbon governance capacity. As a result, the development foundations and evolutionary trajectories of the three subsystems remain inconsistent, and their LTCCD generally remains lower than that of the eastern region. Nevertheless, the coordination levels in central and western China continued to improve throughout the study period, suggesting that the development of regional central cities, increased infrastructure investment, and the expansion of digital service coverage have, to some extent, narrowed the development gaps among the three subsystems. Since digital infrastructure, industrial restructuring, and low-carbon governance capacity are all characterized by long-term accumulation (Nie et al., 2025), such improvements are more likely to be reflected in a gradual catching-up process rather than the rapid elimination of the long-standing regional gradient.

The relatively large intra-regional disparities observed in the eastern region further indicate that a higher average regional coordination level does not necessarily imply greater internal consistency in regional development. The eastern region includes not only core cities with relatively advanced digital economies, service industries, and green governance capacity, but also peripheral cities characterized by a relatively high proportion of traditional industries and comparatively slower progress in Low-carbon Transition. Consequently, substantial differences exist among cities in the pace of digitalization diffusion, industrial upgrading, and green governance, giving rise to pronounced internal heterogeneity (Li C. et al., 2026). By contrast, most cities in central, western, and northeastern China are at relatively similar stages of development and exhibit comparable coordination levels, resulting in relatively small intra-regional disparities. These findings suggest that the regional disparities in LTCCD are associated not only with the long-standing differences in development foundations among eastern, central, and western China, but also with the divergence in urban transition pathways and the pace of coordination evolution within each region.

5.3 Multi-factor interaction mechanisms underlying the spatial heterogeneity of LTCCD

The GeoDetector results indicate that the spatial heterogeneity of LTCCD cannot be adequately explained by any single development condition, but instead exhibits a pronounced pattern of enhanced multi-factor interactions. The single-factor detection results show that the explanatory power of different factors changed dynamically over the study period, indicating that the key conditions underlying the spatial heterogeneity of the long-term coordination status of the DUL system were not fixed, but continuously adjusted as the urban sustainable transition progressed. Meanwhile, the independent explanatory power of traditional factors, such as economic scale and industrial structure, generally declined, suggesting that economic growth or industrial restructuring alone can no longer adequately explain the spatial disparities in the long-term coordination status of cities. Instead, the formation of LTCCD increasingly reflects the systemic characteristics arising from the joint effects of multiple development conditions during the coordinated evolution of Digitalization, Urban Development, and Low-carbon Transition.

Innovation capacity, population agglomeration, and government regulation consistently exhibited relatively high explanatory power across multiple periods, indicating that technological supply, factor agglomeration, and institutional support constitute important conditions for explaining the spatial heterogeneity of LTCCD (He and Liu, 2022). In conjunction with the theoretical framework developed in this study, innovation capacity facilitates the diffusion of digital technologies and the application of green technologies, thereby providing technological support for the empowerment of Urban Development and Low-carbon Transition through Digitalization. Population agglomeration enhances knowledge spillovers, market size, and resource flows, thereby creating a more favorable environment for the application of digital technologies and the advancement of the green transition. Government regulation, through fiscal investment, infrastructure development, and the provision of public services, improves the operating conditions of the system and provides institutional support for the sustained coordination of the three subsystems (Zhang et al., 2024). These three conditions correspond to different functional dimensions of the theoretical framework, namely, digital empowerment, system support, and governance support, and jointly underpin the long-term coordination evolution of the DUL system, rather than relying on any single factor in isolation.

The explanatory role of digital infrastructure exhibited distinct stage-specific characteristics. As digital infrastructure continued to improve, its spatial distribution gradually became more balanced, resulting in a decline in its single-factor explanatory power. This suggests that the independent explanatory role of digital infrastructure in the spatial heterogeneity of LTCCD gradually weakened. However, the interaction detection results show that the combinations of digital infrastructure with innovation capacity, population agglomeration, and government regulation consistently exhibited relatively high explanatory power. This indicates that digital infrastructure primarily serves as a fundamental platform supporting digital connectivity, information flows, and resource allocation. Its role is more fully realized when combined with innovation resources, urban development foundations, and the institutional environment, thereby providing stronger support for the long-term coordination evolution of the DUL system.

Further examination shows that the explanatory power of two-factor interactions consistently exceeded that of individual factors throughout the study period, indicating that the spatial heterogeneity of LTCCD was consistently characterized by stable multi-factor synergistic explanatory patterns rather than being dominated by any single factor. Meanwhile, the combinations of factors with relatively high explanatory power varied across different stages, suggesting that the key development conditions underlying long-term coordination evolution continuously adjusted as the urban sustainable transition progressed. Therefore, the GeoDetector results are more appropriately interpreted as revealing the relative explanatory roles of different development conditions in the spatial heterogeneity of LTCCD, rather than as evidence of direct causal effects of any individual factor on long-term coordinated development. Overall, the long-term coordination status reflected by LTCCD represents the systemic outcome of the continuous interactions among multiple development conditions during the long-term evolution of Digitalization, Urban Development, and Low-carbon Transition, rather than the independent effect of any single factor.

5.4 Research limitations and prospects

Although this study systematically investigated the spatiotemporal evolution, regional disparities, and driving mechanisms of the long-term LTCCD of the DUL system, several limitations should be acknowledged. First, this study focused on prefecture-level and above cities, emphasizing the coordinated evolution of the DUL system at the city scale. Consequently, spatial heterogeneity within individual cities was not examined. Future research could incorporate finer-resolution spatial data to investigate intra-urban coordination patterns. Second, this study identified the driving factors and interaction effects underlying the spatial heterogeneity of LTCCD using GeoDetector. However, while GeoDetector effectively evaluates the explanatory power of influencing factors, it cannot establish strict causal relationships. Future studies may combine spatial econometric models or causal inference approaches to further validate the underlying driving mechanisms. Finally, the analytical framework developed in this study was constructed based on Chinese cities. Given the substantial differences in development stages, institutional contexts, and digitalization processes across countries, the generalizability of the proposed DUL analytical framework requires further examination. Future comparative studies based on cross-country datasets could further examine its applicability under different institutional and developmental contexts.

6 Conclusions and policy implications

6.1 Main conclusions

Using panel data for 289 prefecture-level and above cities in China from 2010 to 2024, this study systematically investigated the spatiotemporal evolution, regional disparities, and driving mechanisms of the DUL system. The main findings are summarized as follows:

• The overall development level of the DUL system increased steadily during 2010–2024, while the three subsystems followed distinct asynchronous evolutionary trajectories. Digitalization exhibited the most pronounced spatial expansion, Urban Development maintained steady growth, and Low-carbon Transition progressed gradually. These findings suggest that coordinated development of the DUL system is achieved through continuously strengthened interactions among the three subsystems rather than synchronized subsystem growth.
• LTCCD of the DUL system improved continuously throughout the study period, evolving progressively from imbalance toward coordination. Nevertheless, pronounced spatial heterogeneity persisted, with eastern China consistently exhibiting higher coordination levels than central and western China. These findings suggest that improvements in coordination capacity do not necessarily eliminate long-standing regional development gradients, but instead promote a gradual transition toward more balanced coordinated development.
• Regional disparities in LTCCD exhibited an overall convergence trend; however, inter-regional disparities remained the dominant source of spatial inequality. Although coordination gaps among regions gradually narrowed during the study period, the disparity between eastern and western China consistently remained the largest, while eastern China exhibited greater intra-regional heterogeneity than the other regions. These findings suggest that the spatial evolution of LTCCD reflects a multi-scale pattern in which overall convergence coexists with persistent inter- and intra-regional heterogeneity.
• The explanatory structure of the spatial heterogeneity of LTCCD exhibited pronounced dynamic changes over time. The explanatory power of individual factors varied across different periods, with traditional factors such as GDP per capita and industrial structure showing an overall decline, whereas innovation capacity, government regulation, and population agglomeration consistently exhibited relatively high explanatory power. Moreover, pairwise interactions consistently showed stronger explanatory power than individual factors, indicating that the spatial heterogeneity of LTCCD is characterized by a stable multi-factor interaction pattern rather than being dominated by any single factor.

6.2 Policy implications

This study reveals the dynamic evolution, regional disparity patterns, and driving mechanisms underlying the coordinated development of the DUL system, highlighting that enhancing long-term system coordination requires simultaneous attention to overall system integration, regional coordination, and multi-factor synergy. Based on these findings, the following policy implications are proposed.

• Adopt a system-oriented strategy to promote the coordinated development of the DUL system. Given the distinct developmental trajectories and evolutionary dynamics of the three subsystems, policy interventions should avoid prioritizing the rapid advancement of any single subsystem. Instead, greater emphasis should be placed on improving the coordination among Digitalization, Urban Development, and Low-carbon Transition. This can be achieved by deepening the integration of digital technologies with urban development and low-carbon governance, thereby fostering synergistic development and continuously enhancing the overall coordination level of the DUL system.
• Implement differentiated regional coordination strategies to reduce regional disparities in coordinated development. Given the persistent spatial pattern of higher LTCCD in eastern China and lower LTCCD in central and western China, region-specific policies should be formulated according to local development conditions. For central and western China, priority should be given to improving digital infrastructure, strengthening comprehensive urban development capacity, and enhancing low-carbon governance to gradually increase regional coordination capacity. For eastern China, greater attention should be paid to reducing intra-regional disparities by promoting more balanced coordinated development among cities. Such differentiated strategies would help narrow long-standing regional coordination gaps while fostering more balanced coordinated development across Chinese cities.
• Strengthen inter-regional collaboration mechanisms to promote more balanced coordinated development. Given that inter-regional disparities remain the primary source of the spatial heterogeneity of LTCCD, greater efforts should be made to enhance resource sharing, technological exchange, and collaborative governance across regions. In particular, advanced regions should be encouraged to disseminate successful development experiences, technological innovations, and governance practices to less-developed regions. These measures would facilitate the formation of inter-regional collaborative networks, thereby accelerating the diffusion of coordinated development and enhancing the overall coordination capacity of the DUL system nationwide.
• Establish a multi-factor collaborative governance framework to enhance the effectiveness of coordinated development. Given that LTCCD is fundamentally shaped by the synergistic interactions among multiple determinants, policy design should place greater emphasis on integrating fiscal support, technological innovation, population agglomeration, and digital infrastructure within a coordinated governance framework. Strengthening the complementarity among different policy instruments would facilitate the transition from fragmented policy implementation to integrated system governance, thereby enhancing the long-term coordination capacity and resilience of the DUL system.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fenvs.2026.1883112/full#supplementary-material

Supplementary material Supplementary material at the source

Supplementary material Supplementary material at the source

Supplementary material Supplementary material at the source

Supplementary material Supplementary material at the source

Supplementary material Supplementary material at the source

Supplementary material Supplementary material at the source

Supplementary material Supplementary material at the source

Supplementary material Supplementary material at the source

Supplementary material Supplementary material at the source

Footnotes

Edited by: Muniyandi Balasubramanian, Institute for Social and Economic Change, India

Reviewed by: Siliang Guo, Qilu Normal University, China

Weifeng Gong, Qufu Normal University, China

Yibo Yan, Hohai University, China