Introduction

Chinese traditional villages, the world’s largest repository of historical and cultural heritage, hold significant historical, cultural, and economic value1. They play a crucial role in rural revitalization, promoting national culture, and bolstering cultural confidence2,3.

Recently, the Chinese government’s emphasis on rural areas and traditional culture has brought traditional villages to the forefront of rural studies4. At the micro level, researchers have enhanced the understanding of village types and architectural diversity by examining site-selection characteristics, public spaces, and architectural forms5,6. At the meso level, studies have explored the spatial form of villages, cultural landscape evolution, and land-use transformations, emphasizing the structural changes and spatial impacts driven by human activities7,8,9. At the macro level, scholars have analyzed the spatial distribution, protection effectiveness, and regional landscape traits of traditional villages, revealing regional differences and current development conditions10,11,12,13. Previous studies have specifically addressed the spatial distribution characteristics and formation factors of traditional villages at regional, watershed, and provincial scales11,14,15. These studies typically use spatial superposition analysis or qualitative research methods to investigate the factors influencing the spatial patterns of traditional villages. They identify slope, water systems, transportation, ethnic culture, and vegetation as key factors affecting the spatial distribution of these villages14,16,17. However, there is still a lack of quantitative analysis regarding the intensity or mode of influence of these factors. Numerous studies have concentrated on the underdeveloped ethnic minority areas in southwest China, suggesting that the challenging terrain and remote location have impeded economic progress while preserving many traditional villages. However, it remains uncertain whether a similar relationship exists in the more developed eastern regions of China. In conclusion, while significant progress has been made in studying the spatial patterns of traditional Chinese villages, several areas still require further exploration: (1) Studies predominantly focus on China’s less-developed provinces, identifying factors such as terrain, water systems, transportation, and ethnic minority populations as primary influences on village distribution18,19. However, it remains unclear if these findings apply to more developed provinces, necessitating further investigation in these regions. (2) Explanations of spatial distribution factors often rely on qualitative assessments or researchers’ subjective experiences, lacking quantitative methodologies. This results in an imprecise understanding of influential factors, hindering the identification of dominant ones and compromising scientific rigor. (3) Existing analyses approach the spatial pattern of traditional villages from a static perspective, neglecting its temporal evolution. Consequently, this oversight leads to an incomplete understanding of spatial distribution rules, diminishing the practical applicability of the findings. (4) These deficiencies have resulted in an inadequate grasp of the complexity and diversity involved in the growth and decline of traditional villages, thereby hindering theoretical integration and placing actual preservation efforts in a reactive position.

In summary, this study employed geographic detectors to systematically analyze 701 traditional villages in Zhejiang Province, aiming to: (1) identify the spatial distribution pattern of these villages and quantitatively determine the dominant factors influencing this distribution, thereby overcoming the limitations of previous qualitative analyses and enhancing understanding of their spatial arrangement; (2) explore the interactive mechanisms of natural and economic factors affecting the survival patterns of these villages; and (3) develop protection strategies for traditional villages in Zhejiang Province, providing insights applicable to similar initiatives in other regions.

Achieving these goals will enhance our understanding of traditional villages in developed regions, shed light on the complexity and diversity of their development and evolution in China, foster the integration of traditional village theories, and offer viable strategies for their preservation in developed areas.

Research area and methodology

Overview of the research area

Zhejiang is a key part of the Yangtze River Delta (Fig. 1). Spanning 105,600 km² and home to 66.7 million people (as of 2024), it encompasses 11 cities and 89 counties. Geographically, Zhejiang lies at the confluence of sea and land, featuring diverse terrains, numerous water system tributaries, and notable regional resource variations, fostering a rich environment for traditional village development. Economically, Zhejiang stands as one of China’s most affluent regions, consistently leading in per capita GDP and exemplifying developed areas in the country.

Previous research indicates a negative correlation between the spatial distribution of traditional Chinese villages and regional economic development, with location, economic structure, and terrain identified as key factors in the preservation of these villages. Unlike other provinces with numerous traditional villages, such as Yunnan and Guizhou(characterized by mountainous, isolated terrains and less developed economies), Zhejiang uniquely combines a high number of traditional villages with significant economic advancement. These phenomena suggest that the relationship between economic development and the preservation of traditional villages is not simply linear. By examining Zhejiang as a case study, we can explore the retention patterns of traditional villages in developed regions and identify the influencing factors. Comparing these findings with those from underdeveloped areas can promote the correlation and integration of knowledge about traditional villages and offer valuable insights for other similar regions to develop protection policies tailored to local conditions.

Fig. 1
Fig. 1
Full size image

The location of the study area in China and the distribution of 701 samples.

Figure source: Created by the author based on ArcGIS9.3.1 software. Figures 2, 3, 4, 5, 6, 7 and 8 are from the same source as Fig. 1.

Data source

The traditional village names align with the six versions of the “Catalog of Chinese Traditional Villages” published by the Chinese government. Geographical coordinates, altitudes, and related data are obtained from the 91 Map Assistant(https://www.91wemap.com/index.htm).

Data on county-level administrative boundaries, water systems, DEM, and roads are sourced from the Resource and Environment Science and Data Center of the Chinese Academy of Sciences. Economic and social data for this research are derived from Zhejiang Provincial Government’s statistical yearbooks.

Main research methods

This study primarily utilized the nearest neighbor distance method to ascertain the distribution patterns of traditional villages20. The geographical concentration index was employed to assess the level of concentration of these villages14. Additionally, the kernel density method was used to map the clustering areas of traditional villages21,22,23. Furthermore, the GeoDetector method was implemented to identify the influencing factors.

Spatial differentiation is a key aspect of geographical phenomena, with GeoDetector serving as a valuable tool for detecting and analyzing spatial variations in geographic elements and their underlying drivers24. The fundamental premise of GeoDetector is based on the notion that when an independent variable significantly impacts a dependent variable, they are likely to demonstrate spatial similarity in their distribution patterns25.

GeoDetector encompasses four detector types, among which this study focuses on the factor detector and interaction detector. The factor detector is utilized to evaluate the influence of a specific factor on the spatial differentiation of a particular indicator. On the other hand, the interaction detector is employed to explore the synergistic effects between different risk factors. It investigates whether the combined independent variables, denoted as X1 and X2, either strengthen or weaken the explanatory capacity of the dependent variable, Y. This analysis is conducted through the following formula:

$$\:q=\raisebox{1ex}{$\left(N{\sigma\:}^{2}-\sum\:_{h=1}^{L}{N}_{h}{\sigma\:}_{h}^{2}\right)$}\!\left/\:\!\raisebox{-1ex}{$N{\sigma\:}^{2}$}\right.$$

where h = 1, …, L represents the stratification layers, Nh denotes the number of units in layer h, N is the total number of units in the entire area, \(\:{\sigma\:}_{h}^{2}\) and σ2 are the variances of Y within layer h and across the whole area, respectively. The range of Y is bounded between 0 and 1. The measure of the strength of an influencing factor’s effect, denoted as q, takes a value between 0 and 1, with values closer to 1 indicating greater strength of the influencing factor.

Analysis and results

The agglomeration characteristics of two cores and multiple points and the Obvious Spatial differences in county-level differentiation

(1)Clustered spatial distribution type.

The Average Nearest Neighbor tool in ArcMap 10.2 was utilized to determine the average nearest neighbor distance of traditional villages in Zhejiang Province. The analysis revealed an actual average nearest neighbor distance of 0.048 degrees, contrasting with the theoretical average nearest neighbor distance of 0.072 degrees. The nearest neighbor index R was calculated to be 0.660, indicating a spatial clustering pattern as it is less than 1. The Z-score was found to be −17.17, with a corresponding P-value of 0, further supporting the significant spatial clustering of traditional villages in the region. This spatial clustering signifies the concentrated distribution and enduring presence of traditional villages within specific areas of Zhejiang Province.

(2)Remarkable agglomeration characteristics.

The geographical concentration index serves to demonstrate the level of clustering of traditional villages in Zhejiang. The calculated geographical concentration index for traditional villages in Zhejiang Province is 16.91, surpassing the value of 10.59 for a uniform distribution. This result suggests a pronounced regional agglomeration pattern among traditional villages in Zhejiang Province, indicating the presence of a notable agglomeration zone.

(3)Cluster in the mountainous areas of southwestern Zhejiang.

Kernel density analysis was utilized to delineate the geographical areas of traditional villages clustering in Zhejiang Province. This analysis resulted in the production of a kernel density map illustrating the distribution of traditional villages in the province (Fig. 2). The agglomeration structure of traditional villages in Zhejiang Province is characterized by a pattern of “two cores and multiple points.” The two primary cores are situated in the southwestern mountainous regions, with one core centered in the Han-dominated area of Songyang County and the other core located in the She ethnic enclave at the intersection of Longquan County and Jingning County. Additionally, multiple points of traditional village agglomeration are dispersed throughout the low mountainous and hilly terrains, encompassing Xianju County, Lanxi County-Jiande County, as well as Yiwu and Dongyang.

From a watershed standpoint, traditional villages are mainly clustered in the upper regions of the Ou River, distinguished by steep mountains and geographically confined terrain, and in the low mountainous and hilly zones of the middle and upper sections of the Qiantang River. Conversely, the eastern watersheds of the Tiaoxi, Lingjiang, and Cao’e rivers show minimal presence of traditional villages.

From an ethnic perspective, as the only minority-inhabited county in Zhejiang Province, Jingning She Autonomous County has a concentrated number of traditional villages, exhibiting distinct characteristics of ethnic minority aggregation.

Fig. 2
Fig. 2
Full size image

Kernel density of traditional villages in Zhejiang Province.

Fig. 3
Fig. 3
Full size image

The county-level differences in the spatial distribution of traditional villages in Zhejiang Province.

(4)Significant differences among counties.

Zhejiang Province exhibits significant variations in the distribution of traditional villages among its counties, with a clear decrease from the mountainous southwest to the northeastern plains (Fig. 3). For instance, Songyang County, located in the southwest region, stands out with the highest concentration of traditional villages, totaling 78. Following closely behind, Jingning and Longquan Counties have 56 and 50 traditional villages, respectively. Moving towards the central hilly area, counties like Lanxi, Wuyi, and Jiande each host approximately 20 traditional villages. In contrast, the northeastern coastal plains demonstrate a scarcity of traditional villages, with certain areas completely lacking such settlements.

Single-factor identification and quantitative analysis based on geodetector

Single factor identification

Villages, as the basic units of human habitation, reflect a combination of natural and socioeconomic influences that shape their distribution and survival. However, the relative strengths of these individual factors and their interactions remain unclear. In this study, GeoDetector was used to quantify the primary drivers and interaction effects governing the distributional characteristics of traditional villages in Zhejiang Province.

The geographical detector method identifies the driving factors of spatial heterogeneity by examining whether the spatial distributions of an independent variable (X) and a dependent variable (Y) are similar, indicating a significant impact of X on Y24. Geodetector comprises four analytical modules; in this study, we use the factor detector to identify the principal determinants of village distribution. We also apply the interaction detector to evaluate how the explanatory power of factors changes when they interact. The GeoDetector quantifies spatial association with the q-statistic to assess how much an independent variable explains the spatial variance of the dependent variable. Statistical credibility is evaluated with p-values; p < 0.05 denotes high significance, and p < 0.1 indicates notable significance, thereby supporting the reliability of the results.

Based on relevant research and the specific context of Zhejiang Province19, the distribution pattern of traditional villages has changed across historical periods. In the earliest phases of formation, natural factors such as terrain, slope, and hydrology governed site selection. Over time, location, transportation, economic development, and industrial structure exerted growing influence and contributed to the decline of these villages. This study selected 11 indicators, covering natural and socio-economic dimensions, as determinants. Quantitative indicators were discretized into categorical variables using the natural breaks method in ArcGIS 10.2 software25,26. Using geographic detectors and county-level administrative units, the study quantitatively evaluated each factor’s explanatory power for the spatial distribution of traditional villages (Table 1).

Table 1 Explanatory power of each factor (q - value).
Full size table

Based on the q-values and p-values presented in Table 1, it is evident that among the 11 factors, average elevation (q-value of 0.468), per capita disposable income (q-value of 0.363), investment in social fixed assets (q-value of 0.327), proportion of primary industry (q-value of 0.279), and economic aggregate (q-value of 0.265) exerted the most significant influence on the spatial distribution pattern of traditional villages in Zhejiang.

Among natural factors, altitude has the strongest explanatory power for the distribution of traditional villages. In the economic and social factors, per capita disposable income、Social fixed asset investment、proportion of primary industry were the most influential, whereas the proportion of secondary industry and the urbanization rate had a lesser impact.

Analysis of the influence process and the direction of action

(1)Topographic slope.

In Zhejiang Province, the physical geography, particularly topography and water systems, significantly influence the formation and conservation of traditional villages. An examination mapping the spatial distribution of these villages with the Digital Elevation Model (DEM) reveals that regions with high village density are predominantly situated in mountainous areas characterized by substantial elevation changes and steep inclines. In contrast, flatlands and low-gradient hills exhibit fewer traditional villages (see Fig. 4). Counties harboring over 50 traditional villages have an average elevation of 664.5 m, those hosting 20 to 35 villages average 420 m, and counties with less than 10 villages are positioned below 200 m in elevation. During the formative period, the rugged terrain provided a safe haven for individuals engaged in resource competition or facing military defeats. This challenging landscape, characterized by its rough and uneven topography, served as a natural barrier, offering protection and security to those seeking refuge or engaged in conflicts. As societies evolved, however, this very rugged terrain posed obstacles to economic progress and the development of transportation infrastructure. Mountainous regions became isolated, hindering connectivity with external influences and thereby fostering the preservation of traditional material environments and historical cultures. Consequently, the isolated terrain played a crucial role in safeguarding the authenticity of traditional villages27.

Regarding water systems, an analysis of buffer zones (refer to Fig. 5) reveals that 43.8% of villages are situated within 100 m of rivers, while 93.4% are located within 300 m. This close proximity indicates a strategic placement of traditional villages near water sources. These water bodies were not only vital for sustaining daily life and supporting production activities but also served as primary transportation routes, underscoring their significance in shaping the spatial organization of these settlements.

Fig. 4
Fig. 4
Full size image

Traditional villages are mainly concentrated in mountainous areas with high altitudes and steep slope.

Fig. 5
Fig. 5
Full size image

The concentrated distribution of traditional villages has a significant hydrotropism.

(2)The level and structure of economic development.

Economic development negatively impacts the preservation of traditional villages. In the counties of Zhejiang, a negative correlation is evident between the number of traditional villages and the levels of economic development. Statistical analysis indicates that economic indicators are lower in areas with a higher concentration of traditional villages. Specifically, in regions with more than 10 traditional villages, these economic indicators fall below the provincial average (Table 2), while the proportion of the primary industry exceeds the provincial average. In agriculturally dominated areas characterized by low economic development and resident income, limited development activities and minimal external influences contribute to the preservation of traditional villages. In contrast, in economically advanced and accessible regions, frequent urban-rural construction accelerates the erosion of ethnic culture, village landscapes, and traditional architecture, leading to a decline in the number of preserved traditional villages.

Table 2 The relationship between the number of traditional villages and the economic development of counties.
Full size table

(3)Social fixed asset investment.

Social fixed asset investment encompasses economic activities associated with the construction and acquisition of fixed assets, including buildings, infrastructure, and transportation equipment.

Fig. 6
Fig. 6
Full size image

Coupling of social fixed asset investment and the number of traditional villages in the county.

fixed asset investment serves as a crucial indicator for evaluating regional economic development. GeoDetector analysis revealed a significant influence of fixed asset investment on the distribution pattern of traditional villages (q-value = 0.327). Further statistical analysis demonstrated a negative correlation between total fixed asset investment at the county level and the spatial distribution of traditional villages. Specifically, counties with higher total fixed asset investment exhibited fewer traditional villages, as depicted in Fig. 6. In counties with more than 20 traditional villages, the average fixed asset investment was markedly lower at 18.243 billion yuan, in contrast to the provincial average of 53.799 billion yuan. Conversely, counties with fewer than 10 traditional villages reported an average fixed asset investment of 61.970 billion yuan, significantly exceeding the provincial average. While fixed asset investment is instrumental in promoting infrastructure development, housing renewal, and enhancing village landscapes, it simultaneously increases the risk of traditional village disappearance. This trend results in the loss of traditional architectural styles and structures, thereby continuously reshaping the distribution pattern of traditional villages28.

(4)Highway density and the radiation of central cities.

The overlay analysis of county road density and the number of traditional villages reveals a negative correlation, indicating that areas with higher road density generally contain fewer traditional villages (Fig. 7). Road traffic directly influences the accessibility and developmental prospects of these villages. Increased road density enhances connectivity between traditional villages and external regions, thereby elevating the likelihood of investment, development, and construction activities. Consequently, this process contributes to the decline of traditional villages.

A buffer analysis of central cities demonstrates that traditional villages are predominantly located on the periphery of urban areas (Fig. 8). The results indicate a decrease in the number of traditional villages as the distance from the central city diminishes. Only 11.7% of villages are found within a 10 km radius of the central city, whereas nearly 30% are situated more than 20 km away. On average, there is approximately a 25% reduction in the number of traditional villages for every 5 km closer to the central city.

The county seat, functioning as the economic nucleus of each county, acts as a catalyst for regional economic advancement. Proximity to the central city results in heightened economic spillover effects, fostering increased economic activities and village development. This phenomenon poses challenges in preserving traditional village landscapes and cultural heritage. Conversely, villages located further away from the central city experience greater transportation isolation, limiting external influences and thereby facilitating the conservation of traditional settlements.

Fig. 7
Fig. 7
Full size image

The relationship between the density of county-level road networks and the number of traditional villages, The greater the road network density, the fewer the number of villages.

Fig. 8
Fig. 8
Full size image

Spatial relationship between central cities and traditional villages, 70% of traditional villages are distributed more than 20 km away from central cities.

Multifactor cross-comprehensive analysis based on geodetector

Multi-factor cross-identification

Single-factor detection identifies the factors with the greatest explanatory power. In contrast, factor interaction detection assesses whether this explanatory power is enhanced or diminished when specific factors interact29.

Cross-detection reveals that explanatory power significantly improves when factors interact (Table 3). For instance, the interaction between “Average elevation” and “Proportion of primary industry” yields a q-value of 0.579, indicating that the spatial distribution of traditional villages in Zhejiang Province results from the combined influence of multiple factors. Further analysis shows that when each factor is cross-referenced with average elevation and the proportion of primary industry, the explanatory power notably increases. This finding suggests that altitude and an agricultural economy are crucial to the formation and preservation of traditional villages. The agriculture-based economic structure supports the development of these villages, while altitude has helped mitigate their decline.

The proportion of the secondary industry, by itself, offers limited explanatory power. Nevertheless, when it interacts with other factors, its explanatory power significantly increases, highlighting the necessity for a comprehensive analysis of all contributing elements.

Comprehensive analysis of multiple factors

By integrating single-factor identification and multi-factor cross-analysis, we determine that the spatial retention pattern of traditional villages in Zhejiang is shaped by a combination of various factors, evolving through two main phases: site selection and distribution, and dynamic evolution. Initially, natural elements like terrain, slope, and water systems primarily influenced site selection, while an agriculture-based economy facilitated the establishment and widespread distribution of these villages. Over time, regional economic changes, increasing incomes, and urbanization drove significant renewal efforts, especially in areas with convenient transportation and proximity to cities. This era witnessed frequent investments in construction by both social and private entities, resulting in the destruction of numerous cultural landscapes, historical structures, and artifacts. Consequently, the number of traditional villages underwent continuous adjustments, leading to their scarcity in economically advanced regions characterized by flat terrain, significant social investments, high per-capita disposable income, and proximity to urban areas.

Discussion and conclusions

Discussion

Understanding spatial distribution patterns and identifying influencing factors are essential for conserving and utilizing traditional villages. This study makes three main contributions to the research on spatial distribution and influencing factors of traditional villages: (1) Methodologically, unlike previous qualitative assessments, this study uses GeoDetectors to scientifically identify and.

Table 3 The q-value after factor cross-detection.
Full size table

quantify the factors affecting the spatial distribution of traditional villages in Zhejiang. While earlier studies emphasized terrain slope and water systems as primary influences30,31,32, our findings reveal that in developed areas, factors like per capita disposable income, per capita GDP, and social fixed-asset investment are more significant. This offers a valuable reference for similar regional studies. (2) Theoretically, our research suggests that the spatial distribution of traditional villages results from the complex and dynamic interplay of multiple factors, rather than a single determinant. In Zhejiang, the spatial distribution of traditional villages has experienced two key phases: initial site selection and distribution, followed by dynamic evolution. Initially, natural factors such as topography, slope, and water systems were central in determining village locations, while an agriculture-based economy facilitated their formation and distribution. In the subsequent evolution phase, economic transformation, rising incomes, and increased urbanization led to significant renewal of traditional villages, especially in areas with good transportation links and proximity to cities. This development often involved substantial social and private investment, leading to the destruction of many cultural landscapes, historical buildings, and relics. (3) In terms of case knowledge, in developed regions, the spatial patterns of traditional villages evolve alongside regional economic growth, contrasting with those in underdeveloped areas. Beyond topographic factors, economic indicators like social fixed-asset investment and per capita disposable income significantly impact the preservation of these traditional villages.

To protect and utilize traditional villages in Zhejiang, we propose the following strategies: (1) Due to their spatial clustering, a cluster protection approach is essential. Traditional villages exist as interconnected settlement clusters rather than in isolation. Protection measures should be tailored to the ethnic attributes, architectural styles, and historical context of each region through zoning and classification. For example, district-specific guidelines for restoring traditional structures and incentive policies can be developed to ensure effective preservation. (2) The negative impacts of rural modernization on these villages must be mitigated. While introducing new technologies, funding, and talent to support rural revitalization is crucial, large-scale development investments, especially those from government public funds that harm traditional villages, should be limited. In areas with dense concentrations of traditional villages, the government should establish investment restrictions or negative construction lists to minimize construction-related damage. (4) Increasing owners’ motivation to safeguard traditional buildings while securing their financial interests is crucial. A significant challenge in conserving traditional villages is the misalignment between the conservation entity and the property owner, particularly for large public or residential structures with complex property rights. To address this issue for buildings with multiple owners, it is recommended to facilitate the transfer of property rights and designate a clear responsible party. In this framework, the village collective would allocate a portion of its revenue to property owners as compensation for maintaining and restoring traditional structures. (5) Additionally, the roster of protected traditional villages should be expanded. Special attention should be given to surveying and documenting villages in remote mountainous regions to ensure that valuable yet overlooked heritage sites receive appropriate protection.

Conclusion

  1. (1)

    The spatial distribution of traditional villages in Zhejiang Province is mainly clustered, following a “two cores and multiple points” pattern. These villages are largely concentrated in the southwestern mountainous areas, such as Songyang and Jingning Counties, while they are sparsely located in the northeastern plains.

  2. (2)

    Among the 11 influencing factors, average elevation, per capita disposable income, social fixed asset investment, proportion of primary industry, GDP, and per capita GDP significantly shape the spatial distribution of traditional villages. In contrast, water system density, urbanization rate, and proportion of secondary industry have a lesser impact. Of the more influential factors, terrain elevation positively correlates with the number of traditional villages, while the others negatively affect it.

  3. (3)

    Cross-detection and identification reveal that the spatial distribution of traditional villages in Zhejiang Province is influenced by multiple factors. The explanatory power notably increases when each factor intersects with the proportion of the primary industry, highlighting the agricultural economy as essential for preserving these villages. Traditional villages are particularly numerous and densely clustered in remote rural areas marked by high altitudes, low per capita disposable income, limited investment in social fixed assets, and significant distances from central cities.

  4. (4)

    The spatial arrangement of traditional villages in Zhejiang Province originates from historical evolution. Initially, the terrain slope and river systems determined their layout. Over time, economic and social factors have continuously influenced and modified this distribution. The dominant factors and their impacts vary across different periods, requiring a precise understanding during preservation efforts.

This study uses counties as the unit of analysis to identify key factors influencing the retention patterns of traditional villages from 11 identified factors. While this approach has limitations, such as uneven spatial analysis scales and insufficient exploration of multicollinearity, it remains significant. By employing a quantitative model, the study identifies factors affecting the retention of traditional villages and proposes the multi-stage characteristics of their formation and evolution in Zhejiang Province. This contributes to the theory of spatial distribution of traditional villages and strengthens the scientific basis for protection policies. Future research could adopt a finer spatial analysis scale to improve identification accuracy and use additional statistical methods for cross-validation. This would enhance the scientific validity and applicability of factor identification and help establish a more comprehensive regional evolution model for traditional villages.