Town-village spatial restructuring in developed regions of China from the perspective of resident travel: a case of Lin’an district, Hangzhou

Abstract

The organization of town-village space in developed regions of China is undergoing profound restructuring, resulting in fundamental changes in the daily travel patterns of rural residents. However, current studies often focus on the static space itself, neglecting the dynamic interactions and functional shifts driven by the “users” of space. This study fills the gap by analyzing residents’ inter-village commuting and non-commuting travel links using mobile phone signaling data, combined with field research and questionnaire surveys. The findings reveal that the town-village spatial structure exhibits a combination of hierarchical and network patterns. First, the traditional “central place” model remains dominant, with central town centers forming radial connections with surrounding villages. However, localized networked structures have emerged, characterized by horizontal links among villages. Second, both commuting and non-commuting travel links have extended spatially, forming self-organized clusters with strong internal connections that do not strictly follow administrative boundaries. The linkages between rural and urban areas have also strengthened, following a logarithmic distance decay trend. Finally, the study highlights that technological advancements, industrial upgrading and diversification, and policy interventions are the main drivers of spatial restructuring. The findings provide valuable insights for town-village spatial planning and policy-making, emphasizing the need to consider actual functional relationships and resident behavior. These insights are important for rural areas in China and beyond in the context of rural transition.

Introduction

Towns and villages in China’s developed regions have undergone profound spatial structural changes over the past decades (Zhang et al., 2022). It has shown a faster pace and unique features under the national system, attracting the world’s attention (Chai et al., 2007; Long and Liu, 2016). From the 1980s to the 1990s, township enterprises¹ have successively sprung up in China’s eastern coastal areas, which has greatly promoted the industrialization of towns and villages from the bottom up. A typical example is the so-called “Zhejiang Model” (“浙江模式”)—originating from Zhejiang Province, China, which is a regional growth pattern driven by the private economy and grassroots entrepreneurship through market-oriented reforms (Zhao, 2009). With the growth of foreign trade and the implementation of the coastal opening policy, these township enterprises are increasingly integrated into the regional division of labor. Since the new century, the strengthening of the link to the global production system and intensive spillovers from metropolises to the surrounding areas has led to industrial specialization and spatial separation in rural areas, promoting the flows such as labor, capital, logistics, and information in a larger range (Woods, 2007, Yang et al., 2016). In addition, rural tourism has developed rapidly, benefiting from the natural resources of the countryside and the huge urban demand for leisure and culture, with pioneering regions starting as early as the 1990s. It further drives the shift in rural space use from production to consumption (Wu et al., 2024).

With these macro trends, the functions of towns and villages have been transformed from a single agricultural form to an agriculture-industry-service integration model, which has further driven the rural space to become diversified, interactive, and dynamic, distinct from the traditional, homogeneous, and isolated form (Tian et al., 2014, Yang et al., 2016, Wu et al., 2024). The urban-rural relationship has also gradually changed from unilateral connection to bilateral interaction with complementary functions (Zhang et al., 2022; Qian et al., 2025). Under such a transforming context, the town-village spatial structure and the mechanisms behind it need to be re-recognized.

From a micro perspective, these changes result in a fundamental transformation of residents’ working and living modes in town and village areas. This is because the economic, social, traffic, and other functional flows between regions foster people’s travel or movements, which in turn promote the reconstruction of the rural spatial structure (Woods, 2012; Chai et al., 2016). Moreover, due to the improvement of transportation and the increase in car ownership, rural residents now have greater accessibility and initiative to travel across villages, towns, and even cities, mainly driven by higher wages, public services, and consumer demand.

Regarding the policy context in China, 2017 witnessed a shift in emphasis towards the countryside when the Rural Revitalization Strategy² was put forward by the central leadership, with a series of strategies covering spatial configuration, industrial development, land management, human settlements, governance systems, etc., to support it (Liu et al., 2020a). Local rural planning has become more multifunctional, targeted, and implemented, from the focus on single land use and facility allocation to people-centered ideology (Liu et al., 2020a). Examples include the plan of the “beautiful rural cluster” in Guangzhou (Tang et al., 2024), the “rural functional unit” in Wuhan (Xiong, 2021), the “rural group” in Wuxi (Yuan and Wang, 2021), and the “urban-rural integrated development units” in Chengdu (Chen et al., 2020). They are similar in the idea that villages and towns can be merged into clusters and classified according to the 3 dimensions of “industry-residence-services” to achieve spatial control, resource sharing, industrial collaboration, and construction management. It helps reshape the rural spatial configuration and further promotes the free flow and rational allocation of urban and rural resources (Chen et al., 2020; Li et al., 2023).

Therefore, residents’ activities, especially cross-regional travel, are considered an important reflection of spatial structure and are widely used to reveal spatial patterns (Niu and Li, 2022; Gao et al., 2023). However, previous studies on rural space often focus solely on the physical space itself or rigid town-village hierarchical systems, ignoring their openness, connections, and dynamics under the functional shift. One of the main reasons is that the spatial relation or connection is difficult to measure, especially in rural areas where data is limited. On the other hand, the in-depth mechanisms of space evolution in research using big data are still vague due to the lack of qualitative analysis.

Thus, this study proposes a novel approach that combines mobile phone signaling data to map spatial linkages and field research to uncover the mechanisms of the phenomenon, and tests it in an empirical case context. The remainder of this paper is organized as follows: Section “Literature review” is the literature review, elaborating on the spatial cognition evolution, the study of town-village space from the behavioral perspective, and the transformation of rural planning in China. “Study area and methods” introduces the study area, data sources, and methodology. In “Characteristics of the town-village spatial structure,” town-village spatial structure is quantitatively measured using commuting and non-commuting travel links identified by mobile phone signaling data, including the characteristics of the network structures, clusters, and rural-urban links. In “Drivers of town-village spatial restructuring,” we explore the drivers of town-village spatial restructuring and the mechanisms through questionnaire data analysis and qualitative field research. Based on the research results, “Discussion” discusses how the town-village spatial restructuring connects to Western theories and the implications for the town-village spatial planning.

Literature review

Spatial restructuring: cognitive shift towards hybrid relations

With the deepening of the regional division of labor from the mid-20th century (Krugman, 1991), the understanding of space has evolved from an initial static, undifferentiated entity to a dynamic, interconnected, and hybrid construct rooted in the continuous reshaping of relationships. Ullman (1957) suggests that no region can exist and develop in isolation, but through the material exchange and mutual collaboration between regions can obtain a win-win situation. The central place theory proposes the spatial relationships that central places with different orders provide goods and services for their corresponding rural hinterland, though the relationships between central places and regions beyond the hinterland were not established (Christaller and Baskin, 1966). Friedmann’s (1966) work on core-periphery dynamics further highlights the evolving spatial interactions within regional systems. Harvey (1973) introduces a social production idea that space is shaped by the relationships it contains, which vary across scales and are constantly in a process of reshaping. Under the “time-space compression” brought about by globalization and advanced communication technologies, Castells’s (1996) “space of flows” concept marked a more significant shift, emphasizing the dominance of relational networks (the “space of flows”) over geographical proximity (the “space of places”). This paradigm change has been supported by subsequent empirical studies on city networks (Alderson and Beckfield, 2004; Derudder and Witlox, 2009; Taylor et al., 2013), which highlight the importance of various flows - traffic, economic, information, and population - in shaping spatial networks.

The above changes have also influenced the realm of rural studies, as the countryside has gradually become open and has been reconstructed with the diverse material exchanges with the outside world (Woods, 2007; Fink et al., 2013; Yang et al., 2016). The relational perspective views rural space as a multi-authored, multi-faceted, and co-constituted space, shaped through local-global interconnections and their place-specific manifestations (Heley and Jones, 2012). It advocates for a relational rural epistemology that recognizes the co-constitutive production of rural space through material and discursive phenomena, processes, and practices, highlighting the theoretical pluralism in comprehending new rural space in its multidimensional complexity (Heley and Jones, 2012; Qian et al., 2025). Similarly, the “global countryside” illustrates that the full picture of rural space can only be seen through a hybrid and networked approach at the interregional level, rather than through a single site or urban-centric approach (Woods, 2007; Woods, 2009). With the rural transition began in the Global North before progressing to the Global South (Wu et al., 2024), the rural space in China is also increasingly understood as a dynamic and multifaceted phenomenon that transcends static, predefined socio-economic and cultural attributes, and is constantly being redefined and reimagined through its socio-spatial interactions with the urban (Qian et al., 2025).

However, Hoggar and Paniagua (2001) criticize the indiscriminate use of the term “rural restructuring,” which has become a catch-all phrase in academic discourse. To more accurately determine whether and to what extent rural restructuring has genuinely occurred, they propose viewing restructuring as an “approach” rather than a pre-set “fact”, and this approach necessitates a holistic and thorough examination of various dimensions, encompassing economic diversification, social stratification, cultural evolution, and spatial reorganization (Hoggar and Paniagua, 2001, Long and Liu, 2016).

Collectively, rural spatial restructuring is a complex process manifested through the interplay of multiple dimensions such as infrastructure networks, production and lifestyle changes, the mobility and heterogeneity of socio-spatial relations, and the dynamic nature of space itself (Long and Liu, 2016), which challenge traditional understandings of rurality (Qian et al., 2025). Yet the restructuring concept often lacks precision and theoretical depth, or the support of empirical evidence. Therefore, an up-to-date spatial interpretation framework for rural space should be established with the hybrid connections between spaces and the logic of functional organization behind them.

Rural spatial analysis from the behavior perspective

Building on the cognitive shift discussed in “literature review,” the understanding of rural space has long been enriched by examining the dynamic activity space composed of residents’ daily behavioral activities (Chai et al., 2016). These activities, with their diverse functions, continuously (re-)shape the rural space (Wang et al., 2021). This perspective traces back to the emergence of behavioral geography and spatiotemporal geography in the 1960s and 1970s under the Western humanist trend, which emphasized the spatial and temporal manifestations of human behavior, highlighting individual agency and the relationship between activities (especially travel or movement) and space (Hägerstraand, 1970, Massey, 1999). Studies in this area cover daily space-time patterns of residential, commuting, consumption, recreation, social visits, and public service access, in which great regularity is found (Ahas et al., 2010; Gao et al., 2023). Some research explores spatial organization through diverse travel behaviors (Asakura and Iryo, 2007, Verhetsel et al., 2018) or identifies different zones based on activity patterns (Fu and Wang, 2021). These studies highlight the functional connectivity patterns revealed by different types of travel links, reveal the rapid transformation of urban lifestyles, as well as people’s decision-making mechanisms in undergoing socio-economic transition.

In rural contexts, research often examines the living and production spaces formed by residents’ daily activities in various natural and cultural settings (Guo et al., 2020). Studies in China often aim to inform planning, focusing on identifying patterns, evolution, and drivers of change under considerations of human-land relationships (Wang et al., 2021). The impact of globalization, urbanization, and informatization on rural living space has also garnered increasing attention (Zhang et al., 2022; Wang et al., 2023). However, traditional data collection methods, such as activity diaries, questionnaires, interviews, and censuses, often fail to capture the complexity of activity space in the face of rural population diversification and spatial heterogeneity (Gao et al., 2023).

The advent of mobile phone signaling big data has significantly enhanced spatial-temporal granularity, overcoming limitations of low spatial accuracy, small sample size, and poor timeliness, and offering advantages in investigating rural spatial structure (Liu et al., 2020b; Gao et al., 2023), especially given the widespread mobile phone ownership among rural residents. While mobile phone signaling data has been widely used in urban research since 2010, its application in rural areas remains limited. Notable studies include Eagle et al. (2009), who compared behavioral dynamics in rural and urban societies using mobile phone data, finding that rural and urban communities differ dramatically in terms of personal network topologies and behavioral characteristics such as travel. Using similar data, Liu et al. (2020b) analyzed rural residents’ daily activity patterns by frequency, diversity, and spatial range. Through movement links measured by mobile phone signaling data, Zhao and Xu (2019) identified complex network characteristics and cluster-form spatial connections in a town of Zhongshan City, China. Niu and Li (2022) also used this data to measure inter-town commuting and non-commuting travel linkages in the Shanghai suburb and identified clusters within, from which they proposed to optimize the town circles’ spatial configuration in a networked manner. The above explorations demonstrate the feasibility and potential of mobile phone signaling data in rural spatial analysis.

Study area and methods

Study area

Lin’an District is located in the outer suburbs of Hangzhou metropolis, Zhejiang Province, eastern China (Fig. 1). It includes 18 town- (street-) level administrative units, covering an area of 3126.8 km². It had 635,000 permanent residents in 2020, including 258 thousand rural residents. Since 2017, Lin’an’s administrative level has been upgraded from county level to city level, meaning that it officially joined the Hangzhou city district. Due to the location of the developed areas and the solid historical industrial foundation, Lin’an’s rural economy is relatively prosperous. The per capita disposable income of rural residents is higher than the average value of the eastern region and far higher than that of the central, western, and northeast regions in China. Under the guidance of the “Two Mountains Theory” (“两山理念”)³, Lin’an’s development mode is in the process of a shift from extensive expansion to more emphasis on green development and quality improvement. Thus, Lin’an’s new economy, especially the e-economy and eco-tourism, has developed vigorously in the past 10 years. Moreover, promoted by the rich local governments, the construction of settlements, public service facilities, and infrastructure in rural Lin’an is also at a high level, forming a new model of “beautiful towns and villages”.

Fig. 1

The study area (Based on 2017 administrative divisions).

In summary, rural Lin’an has been transforming from self-organized development relying on natural resources to spatial reconstruction under the influence of external forces such as urbanization, industrialization, informatization, and radiation of the metropolis. Whilst it also retains rurality in geography, society, and industry. Therefore, with a diverse urban-rural-dual nature, the Lin’an rural area makes a typical case for our study.

Towns and villages in China are regarded as closely interconnected and indivisible systems that are often studied as a whole. The town-village area studied in this paper covers 13 towns or 298 administrative villages in Lin’an District. Among them, Yuqian Town and Changhua Town are the two central towns (“中心镇”) with comprehensive functions, assuming the central and western sub-centers of Lin’an, respectively. The Lin’an city area (the 5 street-level administrative units) located in the east is the main center, which is depart from the town-village area. Under a town-level administrative division, there are two levels of administrative villages (“行政村”)—town center village(s) (“镇区”) and other general villages.

Methods and data

Data source and link identification

In this study, we use mobile phone signaling data to identify two kinds of resident travel links - commuting links and non-commuting links, while the former refers to travel from home to the workplace. The latter involves travel from home to a variety of public services, mainly including out-of-class education, medical care, shopping, and recreation.

The mobile phone signaling data were obtained from China Mobile⁴ for 30 consecutive days in April 2017, including 20 working days and 10 rest days. The data fields include user ID, time, and base station code. The total number of communication base stations within the study area is 2126, with an average spatial coverage of 1.47 km² each, covering 90.1% of the villages. Twenty-five villages with no base stations distributed were excluded from the follow-up calculation.

In the 30 days, when a user stayed for the longest time and at least 2 h from 19:00 to 8:00 the next day, and this routine had the largest number of repeated days and at least 60% of the total days, then the location of this base station was taken as the user’s home. The location identified in the same way from 9:00 to 18:00 during the daytime was regarded as the user’s workplace. When a user traveled from home and then stayed within the range of a destination base station for more than 30 min, and this base station was neither the home nor workplace base station, then this base station was defined as a non-commuting destination. We considered a link between the user’s home and workplace as a commuting link, and a link between home and the non-commuting destination as a non-commuting link. A total of 79,500 commuting links and 1,159,500 non-commuting links were identified within the study area.

Travel networks building and analysis

To depict the spatial distribution of resident travel, we built commuting and non-commuting networks by drawing on the complex network analysis approach (Niu and Li, 2022; Gao et al., 2023). Specifically, the geometric center points of villages served as network nodes, the linking lines among them as edges, and the sum of two-way travel links as the weight of the edges. The one-way movement trajectories from Village A to Village B and from Village B to Village A were summed into a single two-way link between Village A and Village B. Only inter-village links were retained, and intra-village travel links were excluded. Moreover, the links between villages and urban areas (Lin’an City and Hangzhou Central City) were summed up in the same way.

To quantify the hierarchical characteristics of inter-village travel links, we defined an indicator—“proportion of top town center links (PTCL)”—for each village, i.e., the proportion of links between a village and the maximum-link town center to the total amount of external links to the village. For a village, the larger the PTCL, the more dominant the village’s linkage with the top town center, and vice versa, for a more evenly split linkage with neighboring villages. For all the villages studied, the more villages with high PTCL, the more centralized and hierarchical the whole linkage structure is, and vice versa, the more decentralized and flatter it is. All linkage values here refer to two-way linkages. The formula is as follows:

$${{PTCL}}_{a}={L}_{a\mathrm{max}}/\mathop{\sum }\limits_{i=1}^{n}{L}_{{ai}}$$

where PTCL_a is the PTCL of Village A, L_{a max} is the maximum value in the links between Village A and town centers, L_ai is the value of links between Village A and Village i, and $\mathop{\sum }\nolimits_{i=1}^{n}{L}_{{ai}}$ represents the sum of Village A’s links with the other villages.

To further measure the strength of linkages between villages and urban areas, we define the “proportion of links to the city (PLC)”. The formula is as follows:

$${{PLC}}_{a}={L}_{{ac}}/\mathop{\sum }\limits_{i=1}^{n}{L}_{{ai}}$$

where PLC_a is the proportion of links between Village A and the urban area (Lin’an City or Hangzhou Central City), L_ac is the amount of Village A links to Lin’an City or Hangzhou Central City.

Network cluster identification

We employed the Louvain method, a community-detecting algorithm proposed by Blondel et al. (2008), to detect clusters in the networks. The algorithm goes through several iterations, trying to add individual nodes to the communities (clusters) that can maximize modularity, until no node change in the network can improve the modularity⁵, at which point the algorithm automatically stops. As the iteration proceeds, smaller clusters are gradually merged. We substituted the commuting network and the non-commuting network into the algorithm. Both networks eventually went through three iterations and demonstrated 12 and 10 clusters, respectively. The final modularity reached 0.77 and 0.67, respectively. The modularity is high, meaning that the intra-cluster links are very tight and the inter-cluster links are relatively weak.

Field research and questionnaire distribution

After big data mapping, we further dig into the drivers of town-village spatial restructuring and the mechanisms behind them. Firstly, we selected 4 representative towns in the Lin’an district: Yuqian, Changhua, Tianmushan, and Taihuyuan. Next, 6–8 villages in each town, including town-center villages and general villages at varying distances from them (32 villages in total), were selected for in-depth field research. We visited the town governments, village committees, typical rural households, and some enterprises to investigate details of local industries, town and village construction, and residents’ daily lives.

Meanwhile, a questionnaire was distributed to a random sample in these villages from 12 July to 20 July 2021. Questionnaires were distributed to rural residents face-to-face by members of the investigation team, with about 20 questionnaires in each town-center village, and 6–10 questionnaires in each general village. The questionnaire mainly included questions on the respondents’ basic personal and family information, employment, and frequency of travel for various purposes. A total of 307 questionnaires were collected, with 299 valid ones. The valid questionnaire rate was 97%.

In addition, another questionnaire was distributed to city residents of Hangzhou via Wenjuanxing—a Chinese professional questionnaire online platform—between 27 July and 31 July 2021. This questionnaire mainly contained aspects such as respondents’ basic information, and frequency and purpose of traveling to the countryside. A total of 210 questionnaires were collected. After excluding the samples of respondents living outside Hangzhou urban areas, 164 questionnaires were considered valid. The questionnaire validity rate was 78%.

Characteristics of the town-village spatial structure

The hierarchical form dominates, while the network pattern appears

From the commuting and non-commuting networks of the Lin’an town-village area (Fig. 2), their patterns both primarily show a hierarchical “central place” form, where the centers of each town form radial connections with the surrounding villages. Among them, the linkage patterns of the two central towns—Yuqian and Changhua—are more centralized. They are also linked to a larger range of areas beyond their township boundaries. However, a networking trend has simultaneously emerged. Some villages have formed closer horizontal links with each other, especially in the areas adjacent to Lin’an City, where relatively dense and flat networks have formed. In other words, residents travel to and from villages of the same level to a great extent, rather than just between general villages and town centers. This finding suggests a mismatch with the traditional “central place” model of town-village spatial structure.

Fig. 2

Commuting and non-commuting networks based on resident travel links.

The results in Fig. 3 show that in commuting and non-commuting networks, villages with PTCL greater than 0.3 accounted for 40.0% and 8.3%, respectively, while villages with PTCL less than 0.1 accounted for 31.6% and 60.4%, respectively. It indicates that although there are still some villages that form major links with a single town center, some have built evenly matched links with other town centers or villages. Therefore, the dominance of the town center, with which they have major links, is not clearly shown. Specifically, commuting links are more confined within the township, similar to the “central place” system within each township. In contrast, non-commuting links are more extensive, more dispersed, and less hierarchical. It is worth noting that for a village, the top-linked town center is not necessarily in the town to which it belongs, i.e., a village may exceed its township to form its major link with other town centers (mainly the two central towns). This is the case for 16% and 74% of the commuting and non-commuting networks, respectively. This result is somewhat contrary to the conventional notion of the hierarchical town-village system in China, also deviating from the assumptions of the central place theory, which we will elaborate on in the discussion section.

Fig. 3

Numerical distribution of PTCL of the commuting and non-commuting network (in descending order according to the PTCL of each village).

Links extend on a large scale while forming clusters at the local level

Analysis of travel distances

In the Lin’an town-village area, the average non-commuting travel distance⁶ is longer than that of commuting (see Table 1): when all links are calculated, the average commuting distance is 1322 m, while the average non-commuting distance is 1397 m—slightly higher than the former. When only inter-village links are considered, the average non-commuting distance (2045 m) is significantly higher than the average commuting distance (1625 m). This result indicates a higher proportion of long distances in non-commuting travel, although this may be influenced by the different identification algorithms of the two types of trajectories.

Table 1 Average commuting and non-commuting travel distances in different regions.

Next, we do the horizontal urban-rural comparison. Firstly, the second section in Table 1 shows that the average distance for both types of travel in Lin’an City is lower than that in the Lin’an town-village area. One possible reason for this is that the Lin’an City itself is small in size, so long-distance travel is mainly linked to the Hangzhou central city. Secondly, the average commuting distance in the Hangzhou central city is considerably higher than that in the Lin’an town-village area, regardless of whether the calculation includes intra-village links. Only for the inter-village non-commuting average distance is the former lower than the latter. This is possible because the distribution density of various public service facilities in the rural area is lower than that in the city, making it unavoidable to travel longer distances to access the services.

As the work-residence relationship is considered an important aspect in judging the spatial structure configuration, we take the commuting distance as a key indicator. From Table 2, the daily commuting distances of rural residents are mostly short, i.e., links of less than 3 km account for the majority. It is worth noting that longer-distance commuters are also beginning to appear and have become a non-negligible number. When all links are calculated, the proportion of commuting links longer than 3 km is as high as 33.2%, those longer than 5 km are 17.9%, and those larger than 10 km still make up 4.8%. A total of 11.2% are inter-town links. Whilst among inter-village links, the proportions of all the above long-distance links are even higher.

Table 2 Percentage of commuting links of different distances in the Lin’an town-village area.

Analysis of network clusters

Clusters in our study can be interpreted as network substructures formed by self-organized people flow, representing rural functional zones promoted by the demand of residents. The commuting clusters are indicators of the work-residence balance situation, while the non-commuting clusters illustrate the patterns of linkages between homes and rural public services and consumption places. Corresponding to the structure of the linkage networks in Fig. 1, each cluster often contains a core (usually the town centers) and a hinterland that are closely interconnected.

Figure 4 shows that the clusters identified from the commuting and non-commuting networks differ in spatial patterns. There are 12 smaller and more equitable-in-size clusters identified from the commuting network. Each cluster has a balanced work-residence relationship within it, with shorter commuting distances. By comparison, the non-commuting network forms 10 clusters with a greater variation in size. Among them, the boundaries of the two clusters, with Yuchen and Changhua towns as the cores, respectively, noticeably extend beyond the town boundaries and include the surrounding villages of other towns. While the rest of the clusters are smaller.

Fig. 4

Distribution of clusters after the third iteration of the commuting and non-commuting network.

Interestingly, the clusters of both commuting and non-commuting networks do not completely match town administrative boundaries. A few clusters, especially the two central clusters, break through their town boundaries and extend outward, while some split into two clusters within a township. This finding indicates that the administrative factor plays an inconclusive role in separating the cluster organization of resident travel in rural areas. Instead, it may be influenced by a variety of other factors. Additionally, the modularity of the commuting network (0.77) is higher than that of the non-commuting network (0.67), indicating that the former has a more distinct feature of clustering.

Linkages to the city strengthen, following the distance decay law

From the analysis above, the daily travel ranges of rural residents have extended far beyond a walkable range, making it necessary to analyze the links between rural and urban areas. Figure 5 illustrates the results of the regression analysis on two indicators: 1) the distance between each village and the city, and 2) PLC. It is found that the farther the distance between a village and Lin’an City, the lower the percentage of its commuting and non-commuting links with Lin’an City, indicating that the geographical proximity of residents’ travel has not been surpassed to a large extent. In other words, rural residents living close to the city rely more on the city for access to jobs and higher-level public services, thus exhibiting greater cross-border mobility. Whereas villages farther away from the city mainly form a stable work-residence and services-residence balance with neighboring towns and villages.

Fig. 5

The regression analysis on villages’ urban-rural distance and PLC.

Furthermore, this distance decay is not linear but approximates a logarithmic function curve. This means that when the distance increases to a certain degree, the regression curve gradually flattens out, so the travel proportion to Lin’an City does not decrease much with the distance increase. The goodness of fit of commuting links is better than that of non-commuting links, with R² values of 0.60 and 0.56, respectively. It indicates that the commuting decay trend is more pronounced, while the latter is less sensitive to geographic distance, probably due to the greater occasionality in service and leisure travel.

However, for the regression analysis of the villages to the Hangzhou Central City, no such decay pattern is observed. It is presumably because when the travel distance exceeds a certain threshold, the travel frequency is very low, depending more on the random personal demand without a daily pattern.

We further calculated the proportion of the scale of the rural-urban link. Of residents living in the Lin’an town-village area and having a stable workplace outside of the same village, 8.6% have their workplace in Lin’an city. For the non-commuting links, 16.7% of rural residents have travel destinations in Lin’an City, and 2.8% have travel destinations in Hangzhou Central City. Moreover, the scale of the urban-to-rural links is also considerable: the number of commuters living in Lin’an City and employed in the town-village area is about twice that in the opposite direction. The number of non-commuting travelers from Lin’an City to the town-village area is about 40% higher than that in the opposite direction, while the number of non-commuting travelers from Hangzhou Central City to the Lin’an town-village area is nearly four times that in the opposite direction. It is shown that with the overall improvement of urban and rural economy, the urban-rural dualism is gradually being broken, and regular daily travels across urban and rural areas have been formed.

Drivers of town-village spatial restructuring

Multiple exogenous and endogenous factors are driving the changes in the town-village area revealed above, including natural resources, economic factors, social factors, policy measures, and transportation improvements. This section summarizes major drivers from the spatial relational perspective into three aspects: transportation and information technology improvement, industrial reconstruction, and changes in physical space resulting from policies.

Time-space compression due to technological advancements

Limited by the backward productivity level and poor traffic conditions, previous villages were often centered on settlements, with farming distance determining the village range. Therefore, the traditional rural landscape in China was homogeneous, independent, and greatly affected by natural factors such as mountains and rivers. Nowadays, however, the significant improvement of transportation infrastructure and communication technology has greatly weakened the geographic constraints on village and town development (Yang et al., 2016). This phenomenon, described as “time-space compression”, triggering the reconstruction of town-village landscapes and the formation of town-village networks.

Firstly, the improvement of rural transportation has provided residents with long-distance accessibility, which has significantly expanded the scale of live-work-service cells. By 2016, the motorization rate of rural roads in Lin’an reached 100%, and public transport accessibility in natural villages (the lowest division level under the administrative village) in the region reached 99%. Since 2020, an east-west expressway has been constructed across Lin’an District, and a subway line has been built connecting to Hangzhou Central City. The private car ownership rate of rural residents has also greatly improved. According to our questionnaire data, over 70% of the rural families in Lin’an own their cars and can easily travel between towns and villages, even to Lin’an City and Hangzhou Central City. However, it is worth noting that rural public transport is still characterized by shortages and long intervals, and some natural villages still lack public transport stations. This results in inequitable accessibility for the elderly, juveniles, low-income earners, and other vulnerable groups without cars or driving capability.

Secondly, informatization has indirectly influenced spatial restructuring by changing the ways of production, consumption, and social interaction among rural residents. The first point to mention is rural production in the network era. Towns and villages have greater chances to explore their talent resources and comparative advantages, thus easily participating in the regional division of labor and creating more rural employment (Wu et al., 2021, Wang et al., 2023). Typical cases include “Taobao villages” and “internet-famous villages”⁷These have emerged in recent years. Some towns and villages play leadership roles and reorganize their surrounding areas. For example, Bainiu Town in Lin’an has become a base for e-entrepreneurship incubation, e-commerce training, and sightseeing. To support Bainiu, the nearby Daoshi Town serves as the main production base, Longgang Town and Qingliangfeng Town are the main processing and logistics bases, and Changhua Town is the exhibition and tourism center. Thanks to online shopping platforms and logistics networks, the products of these town groups are sold nationwide.

Another point concerns rural life. Online shopping has partly replaced the commercial functions of town centers and nearby urban areas. Our questionnaire shows that the rural residents in Lin’an, especially young and middle-aged groups, have become accustomed to online shopping. In addition, the prevalence of online chatting has not only expanded the social network but also transferred kinship and geographical ties in rural areas. In addition, Fig. 6 illustrates that travel for “visiting relatives and friends” is still frequent among rural residents in Lin’an. Hangzhou urban residents with family members living or working in towns and villages have a higher travel frequency between urban and rural areas (see Fig. 7). It is suggested that though the urban-rural gap is sewn by modern technologies, the traditional rural social relations have not disintegrated, but to some extent, been strengthened.

Fig. 6

The proportion of different types of travel intensity of Lin’an town and village residents at three spatial levelsFootnote

To quantify the travel frequency in the questionnaire data, we set weights according to different frequency intervals - almost daily at 1.00, 3-5 times per week at 0.60, 1-2 times per week at 0.20, 1-2 times per month at 0.06, 1 time in two or three months at 0.02, 1-3 times per year at 0.01, and no travel at 0.00. Then the sample size was weighted and summed to obtain the travel intensity.

.

Fig. 7

Questionnaire statistics of Hangzhou city residents traveling to towns and villages a. Travel purposes b. Comparison of the travel frequencies of respondents with or without family members in towns or villages.

Industrial upgrading and diversification

Since the 1980s and 1990s, substantial bottom-up industrialization in rural Lin’an has promoted the transformation of the industry structure and formed local industrial clusters. This is not a rare case in the developed regions of China (Zhang et al., 2022). The original single agricultural production has evolved into a comprehensive industrial structure that integrates agricultural production, agricultural product processing, cultural tourism, e-commerce, scientific research, etc. For example, originating from bamboo planting, the bamboo shoot industry in Taihuyuan Town has developed into an industrial cluster including bamboo seeding research and breeding, bamboo shoot gardening technology research and promotion, bamboo science popularization, ecological teaching, and rural technician training. The pecan industry in the west of Lin’an has also extended from original planting and processing to e-commerce platform sales, leisure agriculture, and cultural tourism.

The reconstruction of the industry has changed local labor markets and the connectivity between work and residence, which in turn has also reshaped the rural spatial structure. The advancement of the new economy, especially e-commerce, has been proven to intensify the expansion and integration of rural land use and the spatial network connections between towns and villages (Wang et al., 2023). In our study, on the one hand, different from urban areas, the locality and informality of town and village industries determine that the distance between Lin’an rural residents’ occupations and residences is mainly short. From our questionnaire data, about 95% of rural employees live nearby. Among them, about 26.8% work part-time in enterprises near their homes, and also operate small shops, farmhouse restaurants, homestays, online shops, or cultivate their own farmland. On the other hand, some long-distance commuters, even across urban and rural areas, have emerged. Specifically, people living in rural areas find jobs in city enterprises, institutions, or governments, commuting in the “pendulum” mode, which used to be characterized as the urban commuting pattern.

Concentration of built-up space driven by policies

First of all, this spatial concentration is manifested in industrial land use. Like most rural areas in the developed regions along the southeast coast of China, benefiting from a relaxed policy environment, cheap land rent, and low labor prices, Lin’an has been industrialized for decades. The early rise of the private economy in towns and villages has led to a point-like sprawl of industrial land. However, in recent years, guided by top-down policies that encourage industry agglomeration, township industries have begun to relocate to industrial parks, mainly neighboring town centers, especially in central towns.

A similar concentration trend is seen in the spatial layout of public service facilities. The allocation of public service facilities in rural Lin’an traditionally follows the hierarchical model, so as to match the settlement layout. However, a myriad of policies in recent years have contributed to the rapid and large-scale implementation of merging townships and the relocation and combination of rural settlements. From 2001 to 2011, the number of townships in Lin’an was reduced from 39 to 13, and the number of administrative villages was reduced from more than 600 to around 200. A typical response to it is that most of the village-level kindergartens and primary schools have been or will be revoked. The two central towns, Yuqian and Changhua, have public service facilities of larger scale and better quality in rural Lin’an. These facilities are equipped with senior middle schools, junior middle schools, district-level hospitals, hub bus stations, large shopping malls, and entertainment facilities. They attract more people from surrounding towns and villages who have medical, shopping, entertainment, and travel needs. In contrast, other general towns mostly own only town governments and farmers’ markets. Thus, their importance in service provision or even their overall centrality has been greatly weakened.

The above changes in physical space have led to a concentration of daily flows toward the two central towns. The results are presented in the mobile phone signaling data mapping in “The hierarchical form dominates, while the network pattern appears”. The central place form is dominant when viewed from the regional level, within which hierarchical transformations and structural distortions occur. As the industrial parks, residential communities, and commercial centers of the two central towns are more clearly zoned, the intra-core network densification is also significant.

Discussion

The spatial restructuring of the town-village area is not merely a transformation of land functions but also involves changes in spatial functional relationships, which are closely related to the work and life of rural residents. From agricultural civilization to industrial civilization and then to the informatization era, towns and villages in developed regions in China have experienced earlier and more drastic spatial changes, challenging and extending traditional central place theory. There are two main points to discuss in this section.

The first point addresses the trend of networked restructuring in modern rural spaces. In Christaller’s central place model, central places of different orders provide commercial, administrative, and transportation functions of different levels, while the hinterlands or “complementary regions” are in the position of receiving services (Christaller and Baskin, 1966). Aligning with this hypothetical framework, the long-standing town and village planning in China has built a hierarchical system of “central town—general town—central village—general village”. Most previous studies on China’s town-village system and village layout planning have also followed this logic, but limited by its strict assumptions, static nature, and lack of consideration for natural-human interactions, it struggles to fully explain the observed spatial patterns in modern rural areas (Wang and He, 2014, Cao et al., 2020).

Thus, the mixture of hierarchical and networked spatial structures in rural Lin’an calls for a more dynamic and linkage-oriented approach to understanding spatial organization, which should not be limited to hierarchical patterns, not only in scale but in function. This is especially true for developed regions, where the flow of factors is more frequent and hybrid (Zhang et al., 2022). The spatial restructuring of Lin’an originated from reforms in rural residents’ production and living approaches. Therefore, there is no unitary spatial model, such as the central place or the spatial network, that can precisely and comprehensively depict its town-village spatial organizations. Moreover, the strongly interconnected clusters suggest that spatial interactions are not solely dictated by hierarchical relationships but are also influenced by functional interdependencies and social networks. This is consistent with the shift towards a more macro and networked view of spatial organization, as noted by Castells (1996) and Woods (2007). The expansion of residents’ daily reach found in our study is also in line with the observations by Harvey (1973) that technological progress leads to “time-space compression,” reducing the constraints of geographical distance. It highlights the need to incorporate technological factors into spatial analysis and policymaking. Spatial planning should focus on the cluster characteristics in networks - integrate towns and villages by the three major functions of living, production, and services into functional clusters.

The second point concerns the human-centered perspective and the value of human flow in re-depicting the rural spatial structure. The functional connections are reflected in human mobility. The logic of this lies in that various internal and external factors drive people’s movements. In turn, people and various flow elements such as capital, labor, information, and social relations that they “carry” reshape spatial structures. Our study shows that residents’ travel links, both commuting and non-commuting, significantly extend beyond traditional hierarchical boundaries, forming self-balanced clusters with strong internal connections. This highlights the importance of human behavior and mobility in shaping spatial structures.

The relationship between people and land in rural areas has long been a core issue studied. However, top-down town and village planning in China is often oriented towards arranging land functions and town-village hierarchies, with insufficient consideration for the actual land “users” and the fundamental changes in spatial relationships. Entering the people-centered New Urbanization process, meeting people’s needs is the premise of spatial configuration. How people “use” space rather than the space (land) itself is what most calls for attention. In the case of Lin’an, most rural residents have transferred their working and living ways while continuing to live in their rural houses. Accordingly, the towns and villages did reshape to accommodate their new demands, which were mainly observed in the central towns, county-level cities, and even beyond. For some residents who still maintain traditional rural farming lifestyles, policies and rural planning guaranteed basic facilities and infrastructure to ensure their daily work and living. By prioritizing people’s daily needs, planners can better accommodate the lived experiences and requirements of the population, leading to more effective and sustainable spatial planning outcomes.

In summary, Lin’an provides a glimpse of the towns and villages that are economically developed, highly accessible, and modernized in terms of work and living in China. The rural landscape becomes complex and cannot be simply summarized as restructured or not restructured. This study has made some advancements, perhaps on at least one side, in understanding the interplay between human activities and changes in the rural spatial organization. The technical approach promoted here can better accommodate this hybridity and dynamism and help promote sustainable spatial development.

Conclusion

This study investigates the spatial restructuring of town-village areas in developed regions of China, using Lin’an District of Hangzhou as a case. By analyzing residents’ daily travel links derived from mobile phone signaling data and combining them with questionnaire surveys and field research, we draw the following conclusions.

1) The spatial structure of the town-village area exhibits a combination of hierarchical and network patterns. While the “central place” model still dominates, with central town centers forming radial connections to surrounding villages, some localized networked structures have emerged, characterized by horizontal links between villages. This indicates a partial deconstruction of the traditional hierarchical model, as villages are increasingly connected to multiple centers and other villages rather than being solely dependent on a single town center.

2) Both commuting (home-work) and non-commuting (home-service) travel links have extended spatially, forming self-balanced clusters with strong internal links. These clusters do not strictly follow administrative boundaries, suggesting that functional connections play a more significant role in shaping rural spatial structure than administrative divisions. The commuting network forms 12 smaller and more equitable-sized clusters, while the non-commuting network forms 10 clusters with greater variation in size, indicating different patterns of work-residence balance and service access.

3) The linkages between rural and urban areas have strengthened, following the law of distance decay. Rural residents living closer to urban areas rely more on cities for jobs and higher-level services, forming regular daily travel patterns across urban-rural boundaries. However, this relationship is not linear, with a logarithmic decay trend observed for commuting links, indicating that the influence of distance diminishes as it increases. The proportion of rural-urban commuting and non-commuting links is significant, highlighting the increasing urban-rural integration.

4) The key drivers of town-village spatial restructuring are multifaceted. “Time-space compression,” led by technological advancements, has enabled residents to travel longer distances and establish remote communication, weakening the constraints of natural conditions and making networked connections possible. Industrial restructuring and diversification have transformed local labor markets, creating new employment opportunities and changing work-residence relationships. Policy interventions aimed at concentrating built-up space, such as relocating industries to parks and consolidating public service facilities, have directly influenced residents’ travel patterns.

This study makes several contributions to the understanding of town-village spatial restructuring in the context of socio-economic transformation. Firstly, it provides a novel methodological approach by utilizing mobile phone signaling data to quantify and map resident travel links, revealing the underlying spatial networks and clusters. This method offers a more detailed and dynamic view of spatial structure compared to traditional approaches. Secondly, the combination of quantitative data analysis with qualitative field research and questionnaire surveys allows for a comprehensive exploration of the drivers and mechanisms behind spatial changes. This integrated approach provides a deeper understanding of the complex interactions between human activities and spatial organization. Finally, the findings offer valuable insights for town-village spatial planning and policy-making, highlighting the need to consider actual functional relationships and resident behavior in addition to physical land use arrangements.

However, this study has some limitations. One crucial concern is that the mobile phone signaling data cannot fully capture the complex travel patterns of rural residents in China. These patterns, influenced by multiple, part-time, and informal employment, and tied to household roles, lack fixed or regular space-time commuting characteristics. For example, men typically engage in agricultural or industrial work, while women manage household tasks and service-oriented activities. The current data cannot fully capture the underlying purposes of these travel patterns. Additionally, the questionnaire data has a relatively small number of responses and limited geographic scope, which may affect the representativeness and generalizability of causal mining. Future research could focus on validating and refining the methods with data of larger coverage and more cases to assess their applicability and robustness. The development of models that incorporate the diverse factors influencing spatial restructuring would also be beneficial. The dynamic nature of space suggests that continuous monitoring and adaptive planning approaches will be essential for achieving balanced and sustainable urban-rural development in China and beyond under the transformation context.