Research on the form and formation mechanisms of Tujia traditional settlement in the Wuling Mountains, China

Abstract

The conservation and sustainable development of traditional settlements have garnered global attention in the context of rapid urbanization and modernization. The traditional settlement of the Tujia ethnic minority in China carries profound historical culture and humanistic memories, serving as living cultural treasures that embody distinctive regional characteristics. Taking the 110 Tujia traditional villages in the Wuling Mountains as an example, this study aligns closely with the rural construction objectives of the national rural revitalization strategy, and explores the quantification method of settlement form oriented to the continuation of folk culture, which can effectively guide the villages in carrying out rational positioning and protective planning. This study first constructed a traditional settlement form quantitative research framework based on field survey and previous research foundations. Then, the shape index, spatial dispersion, and fractal dimensions methods were used to explore the form characteristics of Tujia traditional villages in terms of boundary shape and spatial structure. While the Grey Relational Analysis method was employed to the influences of natural and humanistic environments on settlement forms, revealing the formation mechanisms of it. The results indicate that there are three boundary forms of Tujia traditional villages: band, clump and finger-like, with the finger-like being the preferred form of the Tujia people in the process of building settlements in different landscapes. The spatial structure has three types: moderate agglomeration, high agglomeration and loose type, and the first is the main type, that is, the level of structuring is mostly medium, medium building density. The Tujia traditional settlement form is mainly affected by the river system, topography and transportation, and the distance from the river has the greatest influence on it. In general, the village form is a pattern that is rationally chosen and adapted to the joint influence of natural, social and human factors. The implications of this study are significant for understanding traditional village dynamics, promoting sustainable development, and refining quantitative methods for rural studies.

Introduction

Chinese traditional settlements, also known as ancient villages, refer to “villages that were formed early, possess rich cultural and natural resources, have certain historical, cultural, scientific, artistic, economic, and social values, and should be protected”. Traditional settlements are rich in historical information and cultural landscapes and are the greatest heritage left by China’s agricultural civilization. Its forms reflect their physical-spatial manifestations, and are the embodiment of the historical and cultural elements of settlements at the material level. Traditional village authentically reflect the wisdom of people’s production and living practices in traditional society, with the unique spatial patterns encapsulating the essence of agrarian civilization and vibrant spatial-cultural memories. In recent years, particularly in developing countries, the deepening globalization process coupled with accelerated urbanization and ‘beautiful countryside construction’ have led to the relentless erosion of traditional settlements by modernization. Many are being renovated or replaced, facing irreversible decline or outright disappearance^1,2. As far as China is concerned, following its 2010 survey, the Research Center of Chinese Village Culture once again carried out a field survey of traditional villages in the Yangtze and Yellow River Basins, and has made a return visit to the 1033 villages in the basin that have been included Chinese Traditional Villages List. The data show that the number of villages decreased from 1033 to 572, a total of 461 disappeared in just four years (approximately 11.2% annual decline or 9.6 villages disappeared monthly)³. This rapid disappearance underscores the severe challenges and dilemmas confronting the preservation and development of traditional settlements, and how to protect and pass has become an important issue facing the whole world today⁴.

As a branch of urban morphology research and development, the space cognition of traditional settlement form is not a new topic. As early as Kohl⁵ analyzed the relationship between settlement form and topography. After the 1960s, as computer technology developed, a combination of quantitative and qualitative methods was gradually adopted⁶. In recent years, to more accurately investigate the growth, evolution, and forms of traditional settlement and to preserve the historical context of villages form an objective and scientific perspective, these methods have been further developed. Research achievements have flourished annually, spanning interdisciplinary fields such as ecology, geography, environmental science, anthropology, sociology, and archaeology. Yun⁷ transformed the spatial composition of traditional settlement into mathematical models for quantitative analysis and conducted structural quantification of settlement spatial configurations. Xincheng⁸ (abbreviated as the Pu-style quantification method) employing the morphological index method, delved into the two-dimensional general morphology of traditional settlements, yielding a refined and quantifiable approach for morphological comparisons, classifications, and analyses. Yifan⁹ extended Xincheng Pu’s quantitative models for settlement boundary shape by utilizing computer programming and Delaunay triangulation to extract settlement boundary contours, thereby enriching the quantitative study of traditional settlement. In an illustrative case of 15 traditional villages in southern Ye et al.¹⁰ applied a synthesis of morphological index and cluster analysis to scientifically demarcate the spatial morphology of these settlements. Most of these studies have conducted quantitative analyses of spatial morphology indexes at a smaller and more precise level. Besides, quantitative research methods frequently leverage technological platforms such as ArcGIS¹¹ and theoretical frameworks like spatial syntax¹² and fractal geometry¹³. Additionally, scholars have explored factors influencing the formation and development of traditional settlement, including natural environments, social contexts, and cultural landscapes^14–16. For example, Ke¹⁷ explored the influence of natural and human factors on the planar form of settlements.

At present, the research on the Tujia traditional settlements mainly focuses on the study of a single administrative region or individual settlements. Dianyu¹⁸ classified Tujia settlements in southwestern Hubei of the Wuling Mountains area into three types: business-based settlements, kinship-based settlements, and topography-based settlements, exploring the morphological characteristics of each type. Ting¹⁹ conducted a comprehensive investigation into the historical evolution of Tujia architecture in western Hunan, examining its adaptive development patterns under the combined influences of natural environments, economic conditions, and sociocultural factors. Zhaoming²⁰ studied the formation and developmental dynamics of Tujia villages in northeastern Guizhou, uncovering their typical morphological traits. Weibo²¹ analyzed the spatial characteristics and formative mechanisms of Tujia traditional settlements in southeastern Chongqing, emphasizing the construction philosophies. However, comprehensive research covering the entire Wuling Mountains area remains absent, particularly in quantitative analyses of settlement morphology. Thus, this study’s quantitative analysis of Tujia traditional settlement morphology across the entire Wuling Mountains area addresses a critical gap in current research, demonstrating significant innovation.

In the field of quantitative research on settlement form, there are still the following four research gaps, which are also the significance of this study: (1) at present, few people study the spatial form of mountainous settlements of ethnic minorities. Mountainous settlements have extremely high theoretical research and practical application value; (2) the quantitative research method of settlement morphology is in its infancy, with relatively few research achievements, and there is a lack of application research results of the Pu-style quantification method. This paper optimizes the virtual boundary scale of the Pu-style quantification method, which is also a new breakthrough in the study of mountain settlements; (3) most studies have been limited to measuring spatial morphological elements of village, have yet to delve deeply into the interaction mechanisms between settlement forms and environmental factors, with insufficient breadth and quantity of research subjects. Based on a large sample size, this paper supplements the research on the correlation between settlement patterns and environmental factors; (4) a systematic quantitative framework for traditional settlement morphology research has not yet been established. This paper constructs a quantitative research framework system for mountain settlement morphology, which also provides a reference for mountain settlement research, and the research framework is universally applicable. Building on field investigations and existing research foundations, this study constructs a quantitative research framework for traditional settlement form, which can reveal the logic behind the traditional village material form formation. By this framework, we can conduct a more comprehensive quantitative analysis of the traditional village form. Taking 110 mountainous Tujia traditional settlements as examples, this research conducts quantitative analysis of settlement morphology and investigates the impacts of natural and cultural environments on settlement forms, which is a critical topic in urban research. This work aims to provide scientific and quantitative methodologies for studying the form features of traditional settlement and to offer theoretical and practical foundations for rural planning, construction, and sustainable development. It also offers reference for similar mountain areas, which is conducive to the promotion of rural revitalization.

Materials

Study area

The Tujia ethnic minority has resided for generations in the central section of China’s natural east–west geographical divide, nestled between the Dongting Lake Plain, Jianghan Plain, and Sichuan Basin—specifically within the Wuling Mountains spanning the contiguous areas of Hunan, Hubei, Guizhou provinces, and Chongqing municipality, the hinterland of central China (the Wuling Mountains are both a geographic category, and a historical administrative area, a complete historical geographic unit ). It is China’s largest inter-provincial ethnic minority settlement area, located at approximately 107.78° E to 111.36° E and 27.59° N to 31.39° N, covering approximately 90,000 km², and includes 30 county-level units: Xiangxi Tujia and Miao Autonomous Prefecture and Zhangjiajie in northwestern Hunan; Enshi Tujia and Miao Autonomous Prefecture, Wufeng and Changyang Tujia Autonomous Counties in southwestern Hubei; Qianjiang, Xiushan, Youyang, Pengshui, and Shizhu counties in southeastern Chongqing, and Tongren area in northeastern Guizhou (Fig. 1). As a core zone of China’s ethnic diversity, the Wuling Mountains are home to the Tujia people, whose settlements reflect centuries of cultural heritage and environmental adaptation. Characterized by rugged terrain, underdeveloped infrastructure, and diverse microclimates, these traditional villages exhibit unique spatial morphologies shaped by topography, migration patterns, and socio-cultural practices. This study selects 110 representative Tujia villages (Fig. 1) from 313 Tujia traditional villages on the Chinese Traditional Villages List (the traditional villages inherently meet the following criteria: the integrity of traditional architectural features, the preservation of traditional features in location and layout, and the active inheritance of intangible cultural heritage.). They were selected based on criteria including diversity of traditional villages, balanced regional distribution, dominance of topography and geomorphology, spatial originality and intactness of settlements, and boundary integrity. And from July 2022 to May 2024, in-depth research was conducted on 110 settlements. The research content was not limited to drone shooting, settlement plan drawing, architectural surveying, photo recording, etc. All selected villages retain self-organized, minimally-altered layouts, offering distinct morphological contrasts for quantitative analysis of environmental interactions and cultural preservation strategies (Fig. 2).

Fig. 1.

Geographical distribution of 110 Tujia ethnic minority traditional villages in China (Note: the map was generated by QGIS 3.16 (https://qgis.org/).

Fig. 2.

Selected photographs of typical Tujia traditional villages (some samples, this image was taken by the author D.R.)

Data sources

The research data include traditional settlements spatial coordinates, settlements satellite image, Digital Elevation Model (DEM) data, rivers, roads, geomorphic types, and field survey photographs. The first six datasets are vector data. The spatial coordinates data of traditional villages were obtained by crawling POI from the authoritative website Guihuayun (http://guihuayun.com/). DEM data were sourced from the Geospatial Data Cloud (https://www.gscloud.cn/search). River, road, and geomorphic type data were obtained from the Resource and Environment Science and Data Center of Geographic Sciences and Resources Research, CAS (https://www.resdc.cn). Settlements satellite image data came from Google Earth with an accuracy of 18, facilitating the later addition of vector data such as building outlines and enclosing walls in the research area. Field surveys and drone photography were conducted in July to August 2022 and March to July 2023 to supplement and refine the village planar patches. ArcGIS 10.8 was employed to calibrate the accuracy of all vector data, ensuring validity and consistency.

Methods

Based on literature analysis and field investigations of Tujia traditional settlements, a quantitative research framework for settlement form was constructed (Fig. 3). This study categorizes traditional villages form into boundary shape and spatial structure. Shape indices were applied to analyze boundary shape, while spatial fractal dimensions, spatial dispersion, and building density were used to evaluate spatial structure. Grey relational analysis was adopted to determine the correlation between environmental factors and settlement form. These four methods are combined to jointly solve the problems of the morphological characteristics and formation mechanisms of Tujia traditional settlements. All data processing and analysis were conducted in ArcGIS, Rhinoceros, and IBM SPSS Statistics.

Fig. 3.

The framework of the traditional village form quantitative research.

Shape index

Village boundary reflects the adaptive growth and development of settlement. In existing studies and administrative divisions, it is mainly defined by natural boundaries (e.g., roads, rivers) and land use boundaries (e.g., redlines)²². In this study, settlement boundaries are defined as closed contours formed by the physical boundaries of buildings and the virtual boundaries of interstitial spaces between structures. For virtual boundary quantification, this research builds on Xincheng Pu’s methodology^8,23 which employs three-tier virtual boundary scales (100 m, 30 m, 7 m). However, the Tujia traditional settlements are typical mountain settlements, where site selection and morphological formation are profoundly shaped by the rugged terrain. Xincheng Pu’s proposed virtual boundary scales (100 m, 30 m, 7 m) were derived from traditional villages in Zhejiang Province in China, which are characteristic of plain settlements. These Zhejiang villages developed without topographic constraints, featuring compact building layouts. Consequently, the virtual boundary scales are unsuitable for mountainous settlements and necessitate optimization. Using drone-captured imagery and high-resolution satellite images, planar patches of settlements were delineated in AutoCAD, focusing on buildings and enclosing walls. A 3D point cloud model was created in Rhinoceros 7.0 to test virtual boundary scales (7 m, 12 m, 14 m, 30 m, 50 m, 80 m, 100 m, 150 m, 200 m)^7,24. After validation, select three scale levels of 14 m, 50 m, and 150 m (Fig. 4) as the maximum distance that the virtual border can span, and define them as the small border, the middle border, and the large border.

Fig. 4.

14 m, 50 m, and 150 m 3D point data model connection diagram (Shemihu Village).

The shape index (S) is a mathematical index extensively applied in landscape ecology, it’s based on the shape index of a compact shape (circle, square, rectangle, or other regular polygons, etc., as needed) as a reference standard. For this study, an ellipse with equal area and aspect ratio was selected as the reference shape. The "shape deviation degree" is calculated by comparing the perimeter of a closed contour to that of an equivalent-area circle^24–26. Building on village quantification models and parametric programming in Rhinoceros 7.0, the perimeter and area of closed contour patches, along with the aspect ratio of their minimum bounding rectangles, were determined using 14 m, 50 m, and 150 m virtual boundary scales, the specific way of drawing boundaries is shown in Fig. 5. The formula for the shape index is as follows:

1

where P is the perimeter of the traditional village border (m), A is the area (m²), and λ denotes the aspect ratio (length-to-width ratio) of the minimum bounding rectangle of the closed contour. The final shape index S is determined as the weighted average of S_s, S_m, and S_l (represent shape indices for the small, middle and large border, respectively). The classification criteria for settlement boundary shapes are detailed in Table 1.

Fig. 5.

Plotting method of settlement boundary.

Table 1.

Traditional village boundary form determination.

S	Form	Boundary aspect ratio λ	Type of boundary form
S ≥ 2	Finger-like form	λ < 1.5	Clump-prone, finger-like village
		1.5 ≤ λ < 2	Finger-like village with no clear inclination
		λ ≥ 2	Band-prone, finger-like village
S < 2	Banded form	λ < 1.5	Clump village
		1.5 ≤ λ < 2	Band-prone clump village
		λ ≥ 2	Banded village

Spatial dispersion

Spatial dispersion (W) indicates the degree of building aggregation within settlements and serves as a key indicator for evaluating settlement morphology²⁷. The calculation formula is:

2

where S is the area of the closed contour (m), and P is the perimeter of the boundary (m). A smaller W value indicates a higher degree of clustering within the settlement, reflecting a more compact and aggregated layout. Conversely, a larger W value signifies greater spatial dispersion, suggesting a fragmented or elongated distribution of buildings. This metric enables systematic comparisons of morphological patterns in settlements.

Fractal dimension

The fractal theory, a significant branch of modern nonlinear science and theoretical geography, has been widely applied in geographical studies since 1975^28–30 and has achieved breakthroughs in interdisciplinary research with architecture. Scholars propose that fractal geometry serves as a classical measurement approach in fractal-related architectural studies, enabling the analysis of structural compositions and spatial morphology at the building scale³¹. There are three calculation methods of fractal dimension: area-perimeter relation, box-counting method, and area–radius relation. For village public spaces, which exhibit structural characteristics as distinct spatial patches, the area-perimeter relation is selected to compute the spatial fractal dimension. The formula is defined as:

3

Among them, D represents the fractal dimension of public spaces, P is the perimeter of the spatial patch (m), and A is the patch area (m²). The D reflects the complexity of the settlement spatial structure, with a theoretical range of 1 to 2. The closer D is to 2 indicates higher spatial compactness and complexity, suggesting diverse and intricate public spaces, and the sense of spatial experience is more enriched. That is to say, the larger the value of D, the stronger the structure of the village space, and the weaker the opposite.

Grey relational analysis (GRA)

The grey relational analysis is a method that uses grey relational grades to describe the strength, magnitude, and order of relationships between factors. It evaluates the degree of influence among system factors or the contribution of factors to the system’s primary behavior by calculating grey relational grades. The core principle involves comparing the geometric similarity between a reference sequence (representing the system’s characteristic behavior) and comparative sequences (representing influencing factors). If the geometric shape of a comparative sequence closely aligns with the reference sequence, their grey relational grade is higher, and vice versa^32,33. The computational steps are as follows:

1. Define sequences. The reference sequence x₀ = {x₀(j)|j = 1,2,…,n} = {x₀(1), x₀(2),…,x₀(n), the comparative sequences x_i = {x_i(j)|j = 1,2,…,n} = {x_i(1),(2),…x_i(n)},i = 1,2,…n, where j denotes discrete time points.

2. Calculate relational coefficients. For each time point j, compute the relational coefficient &_i(j) between x₀ and x_i:

   4

   where |x₀(j) − x_i(j)| is denoted as the absolute difference between the x₀ series and the x_i series at point j. min_imin_j|x₀(j) − x_i(j)| and max_imax_j|x₀(j) − x_i(j)| denote the minimum and maximum differences across all sequences, ρ ∈ [0, 1] is the distinguishing coefficient, generally ρ = 0.5.
3. Compute relational grade. The grey relational grade r_i for each comparative sequence is the average of its coefficients:

   5

Usually, r_i ≥ 0.8 indicates strong association, indicating significant influence. 0.5 ≤ r_i < 0.8 indicates moderate association, suggesting partial relevance. r_i < 0.5 indicates weak or negligible association^34,35.

Data-processing methodology

Since the data of each indicator have different units and dimensions, to eliminate the impacts of varying magnitudes and dimensional differences among indicators on computational results, this study adopted the range standardization method for data normalization³⁶. The calculation formula is as follows:

6

where X_i represents the standardized value of the indicator for settlement i, and X, X_min, X_max, denote the original, minimum, and maximum value of the indicator, respectively.

Results

Villages boundary shape

Based on the shape index method (1), three-tier boundaries (14 m, 50 m, and 150 m) were delineated for 110 sample settlements to calculate shape index values (S_s, S_m, S_l) (Fig. 6). Since the shape indices derived from the three-tier boundaries varied, the settlement boundary shape index (S) was obtained through weighted averaging: S_l(1.4352) < S_m(2.1992) < S_s(4.0452), with S_m/S_l = 1.5323, and S_m/S_s = 0.5437. Thus, S = S_s × 0.5437 × 0.25 + S_m × 0.5 + S_l × 1.5323 × 0.25. Ultimately, λ and S were determined (Fig. 7). The planar morphology of Tujia traditional villages encompasses six types: banded, band-prone clump, clump, band-prone finger-like, clump-prone finger-like, and finger-like with no clear inclination. Finger-like villages dominate (54.5%), including band-prone finger-like villages (26.4%), clump-prone finger-like (18.2%), and finger-like with no clear inclination (10%). Clump villages account for 24.6%, banded account for 20.9%. Village boundaries exhibit significant fragmentation, reflecting the adaptation of Tujia traditional settlements to the complex mountainous terrain, resulting in intricate and fragmented morphological patterns.

Fig. 6.

The small, middle and large boundaries of typical villages.

Fig. 7.

Shape index relationship and boundary morphology statistics of 110 villages.

Villages spatial structure

Through the determination of the three-layer boundary-closed geometric shapes mentioned above, it was found that the middle border most closely approximates the planar morphology of villages, and its calculation results are more accurate (Fig. 6). Therefore, the middle border of 50 m was selected as the spatial scale for quantifying village spatial structure analysis. Based on previous research^23,37,38 and field investigations of traditional Tujia settlements, the boundary of village public space patches was defined as the active space within approximately 2.5 m outside the middle border (Fig. 8), which makes the public space form a pattern. Using the spatial fractal dimension Eq. (3), the D values of sampled settlements were calculated and analyzed (Fig. 9). The results show that the D fall within a range of 1.2367 to 1.4802, with a mean of 1.3991. Applying the "68–95–99.7" rule of normal distribution, the standard D was determined as [1.3567, 1.4414], this interval is classified as medium fractal dimension villages. Those below 1.3567 as low, and those above 1.4414 as high. Medium and low fractal dimension settlements account for 80.9%, with low comprising 13.6% and medium 67.3%. High fractal dimension villages are relatively rare, representing only 19.1%. This indicates that Tujia traditional settlement predominantly exhibit medium fractal dimensions. Low settlements are characterized by sparse buildings and simple spatial structures, while high settlements feature complex spatial configurations and strong structural organization, offering richer spatial experiences and higher activity levels. Medium settlements lie between these extremes. Low, medium, and high fractal dimensions can respectively represent weak, moderate, and strong structural characteristics of settlements, confirming that Tujia traditional settlements are primarily moderate structural features.

Fig. 8.

Boundaries of village public space patches schematic.

Fig. 9.

Classification statistics of fractal dimension and spatial structure of 110 villages.

By arranging the fractal dimension values of public spaces in ascending order and constructing a relational diagram (Fig. 9) between building density, spatial dispersion, and fractal dimension values for sampled villages. It was observed that as fractal dimension values increase, building density shows fluctuations but exhibits an overall gradual upward trend. Higher building density implies greater settlement compactness, which correlates with elevated fractal dimension values and stronger internal organizational structure. Spatial dispersion displays significant fluctuations, where higher dispersion typically indicates sparser building layouts. Theoretically, this should correspond to lower fractal dimension values. However, empirical analysis revealed that settlements with high dispersion may still exhibit medium or high fractal dimensions. This suggests a positive correlation between fractal dimension and building density, while the relationship with dispersion remains inconspicuous, likely influenced by topographic factors.

Formation mechanisms of settlement form

Based on the quantitative analysis of settlement form in the previous section, a strong correlation exists between natural and socio-environmental factors and settlement morphology. To further elucidate the holistic characteristics of traditional Tujia settlement morphology and clarify the interrelationships and influencing factors among morphological elements. The ten environmental factors were selected: average elevation, average slope, aspect, RDLS, landform, the relationship with river (Table 2), river width, the distance from river, per capita GDP, and the distance from road based on the existing research review and the actual situation of Tujia traditional settlements. The GRA (4)(5) are utilized to conduct a causal analysis of the morphological characteristics of the sample settlements. Settlement boundary maps delineated in AutoCAD and georeferenced satellite imagery from Google Earth were imported into ArcGIS10.8, the use of partitioning statistics to obtain the village’s environmental factors data, and imported into IBM SPSS Statistics 25 with the settlement morphology quantitative data, and all data were standardized using Eq. (6). Bivariate analysis tools were utilized to investigate relationships between settlement morphology and environmental factors (Table 3).

Table 2.

Location of settlements in relation to rivers.

Settlement-river relationships	Distance between settlement and river (m)	Standardized grade
Distant the river	≥ 1000	1
Near the river	500–1000	2
Cross the river	–	3
Perpendicular to the river	≤ 500	4
Single-side parallel riverside	≤ 500	5

Table 3.

Correlation analysis of village form and environmental factors.

		Average elevation	Average slope	Aspect	RDLS	Landform	Relationship with the river	River width	Distance from river	Per capita GDP	Distance from road
Boundary shape	p	0.037	− 0.049*	− 0.062	0.060*	− 0.082	0.204*	− 0.037	0.098*	0.081	0.243*
	Sig	0.704	0.047	0.519	0.042	0.392	0.033	0.701	0.043	0.402	0.011
Fractal dimension	p	0.013	− 0.222*	− 0.005	− 0.056*	0.034	0.188*	0.037	− 0.417**	0.018	0.004
	Sig	0.892	0.020	0.962	0.042	0.724	0.050	0.704	0.000	0.855	0.971

The probability of significance is less than 0.01, and the correlation between the two variables is very significant, expressed by "**".

The significant probability is less than 0.05, and the correlation between the two variables is significant, which is represented by "*".

The probability of significance was greater than 0.05, and there was no significant correlation between the two variables.

Correlation of boundary morphology with environmental factors

As shown in Table 3, the settlement boundary shape exhibits no significant correlation with average elevation, aspect, landform, river width and per capita GDP. However, it demonstrates significant relationships with average slope, RDLS, relationship with river, distance from river, and distance from road. Specifically, negative correlation with average slope: higher average slopes within settlement boundaries correlate with larger boundary shape indices, reflecting planar morphologies approximating finger-like (Fig. 10). Slope area, topographic constraints settlement expansion, making clustered construction challenging. Consequently, the Tujia people preferentially utilize flatter areas for housing, resulting in dispersed, finger-like configurations. Clump villages typically occupy gentle slopes, facilitating radial expansion. Banded and finger-like villages can be located in different slope environments, including ridge-aligned development, riverside linear layouts, or scattered distributions on flat terrain. Positive correlations with RDLS, relationship with the river, distance from river and distance from road. It shows that the greater the RDLS, the more the planform tends to be banded (Fig. 10). In rugged terrain, buildings align with contours or roads to minimize construction challenges, producing elongated patterns. Clump villages dominate low RDLS areas, while finger-like adapt to diverse relief conditions, demonstrating finger-like suitability for mountainous environments. Distant-river settlements tend toward finger-like patterns, single-side parallel riverside settlements favor banded development. Settlements farther from rivers exhibit more clustered morphologies due to reduced fluvial curvature influence (Fig. 10). Notably, all settlements lie within 2000 m of rivers, except outliers. Greater distances to roads correspond to smaller boundary shape indices, favoring banded and clump villages. Proximity to roads in mountainous terrain often results in finger-like patterns due to road curvature and constraints from adjacent farmland or topography. Remote villages, less influenced by roads, expand along contours, favoring banded.

Fig. 10.

The relationship between boundary morphology and environmental factors.

The analysis reveals significant correlations between settlement boundary shape and five environmental factors: average slope, RDLS, relationship with river, distance from river, and distance from road. Employing GRA (3)(4), the relational coefficients between these factors and boundary shape were calculated (Table 4). All factors exhibit relational coefficients exceeding 0.5, confirming their substantial influence on the formation of settlement boundary shape. The ranked order of relational coefficients is distance from river (0.81611) > distance from road (0.71026) > relationship with river (0.69010) > average slope (0.68949) > RDLS (0.65862). Distance from the river has the greatest degree of influence, as water is indispensable for production and daily life, settlements closer to rivers exhibit morphology more directly shaped by fluvial dynamics (e.g., banded patterns along riverbanks). Secondary influence of road proximity, the average slope and RDLS is relatively weak. The traditional Tujia stilted dwellings architectural adaptations to mountainous terrain significantly mitigate topographic limitations, enabling flexible spatial organization even on steep slopes. This explains why slope and RDLS have weaker correlations with settlement boundary shape.

Table 4.

The degree of correlation between environmental factors and boundary morphology.

Environmental factors	Correlation coefficient
Average slope	0.68949
RDLS	0.65862
The relationship with the river	0.69010
The distance from river	0.81611
The distance from road	0.71026

Correlation of spatial structure with environmental factors

As indicated in Table 3, the fractal dimension exhibits no significant correlation with average elevation, aspect, landform, river width, per capita GDP, or distance from road. However, it shows significant relationships with average slope, RDLS, relationship with river, and distance from river. Specifically, negative correlations exist between fractal dimension and average slope, RDLS, and distance from river. This indicates that as slope steepness, terrain ruggedness, or distance from river increases, the fractal dimension decreases, reflecting weaker structural organization in settlements, characterized by sparse building layouts. Steep slopes predominantly host weakly structured settlements, moderate structural strength occurs in gentler slopes, and strongly structured settlements are typically located closer to rivers (Fig. 11). Mechanistically, steep terrain limits construction feasibility, necessitating selection of flatter, buildable sites, which reduces spatial compactness. Additionally, reduced reliance on water resources in river-distant areas allows for more arbitrary spatial arrangements. Positive correlation is noted with relationship with river. Distant river villages exhibit smaller fractal dimension and weaker spatial structural, whereas single-side parallel riverside settlements display larger fractal dimension and stronger structural features, indicating tighter building clustering near waterways (Fig. 11). This underscores the critical role of rivers in shaping compact, organized layouts due to their indispensable role in daily life and production.

Fig. 11.

The relationship between spatial structure and environmental factors.

The analysis demonstrates that the spatial structure of settlements is influenced by average slope, RDLS, relationship with river, and distance from river. Utilizing GRA (3)(4), the relational coefficients between these factors and the fractal dimension were calculated (Table 5). All factors exhibit relational coefficients exceeding 0.5, confirming their great impact on settlement spatial structure. The ranked order of relational coefficients is distance from river (0.81450) > average slope (0.71024) > RDLS (0.65586) > relationship with river (0.55588), distance from river had the greatest, followed by average slope and RDLS, and least by relationship with river.

Table 5.

The degree of correlation between environmental factors and spatial structure.

Environmental factors	Correlation coefficient
Average slope	0.71024
RDLS	0.65586
The relationship with the river	0.55588
The distance from river	0.81450

Discussion

The Tujia traditional settlement patterns are primarily influenced by natural and social conditions such as rivers, topography, and transportation, yet these represent only external driving mechanisms. Intrinsic mechanisms like Feng Shui concepts, cultural customs, and clan culture have also profoundly shaped the formation of settlement morphology. Traditional settlement patterns emerge as rational and adaptive forms shaped by the interplay of natural, social, and cultural factors (Fig. 12). The natural environment is the background foundation of traditional village forms: settlements and their surroundings engage in a dynamic equilibrium of active and passive interactions³⁹, or villages take the initiative to transform nature by leveling the base site and planting trees, or passively rely on nature by building villages based on the mountainous terrain and building houses according to local conditions. Among natural factors, slope gradient, topographic relief, and river systems exert significant influence. Steeper slopes and greater relief reduce suitability for construction, leading to dispersed settlement layouts as the Tujia people selectively settled in favorable locations. Settlements were established along rivers, serving both as crucial defensive barriers and as sources of daily water for production and living. Social economy is the driving force for the development of traditional village form: transportation infrastructure facilitates external connections and the exchange of advanced construction technologies, thereby influencing settlement morphology. Traditional culture is the spiritual connotation of traditional village form: site selection follows the Feng Shui concepts of “backing mountains, facing watercourses while harmonizing yin-yang energies”. The Tujia’s complex spiritual system, centered on polytheistic worship⁴⁰, fosters clan cohesion, promoting kinship-dominated settlements, and form a social relationship network of close relations and mutual supervision. This internal cohesion creates a centripetal force within villages, resulting in a “large dispersion with small clustering” pattern despite mountainous constraints. Additionally, influenced by migration culture, frequent interactions with other ethnic groups gradually diluted the Tujia’s kinship-based traditions. Consequently, while clan buildings like ancestral halls exist, settlements are seldom centripetally arranged around them. Therefore, it can be said that the morphological characteristics of Tujia traditional settlement are the result of “selection and adaptation” to the topographical and cultural environments for thousands of years.

Fig. 12.

Formation mechanism of Tujia traditional village form characteristics.

Conclusions

Based on literature review and field investigations of 110 Tujia traditional settlements, this study established a quantitative research framework for settlement form. Building upon this framework, the boundary morphology and spatial structural characteristics of these villages are explored, while uncovering their formative mechanisms. The main conclusions are as follows:

1. Tujia traditional settlement boundary morphology manifest as banded, clump, or finger-like forms, with finger-like predominating, suggesting that it’s the preferred of the Tujia in the process of building settlements in different landforms. Boundary morphology shows significant correlations with average slope, RDLS, relationship with the river, distance from river, and distance from road. The influencing factors rank as: distance from river > distance from road > relationship with the river > average slope > RDLS.

2. Tujia traditional settlement spatial structures exhibit weak, moderate, and strong organizational characteristics, predominantly moderate. These are typically characterized by medium building density and low-to-medium spatial dispersion, reflecting an overall appropriate level of structural organization in Tujia settlements—neither excessively scattered nor overly compact. This demonstrates the Tujia people’s environmental adaptability and wisdom in mountainous habitation. Spatial structure significantly correlates with average slope, RDLS, relationship with the river, and distance from river. The influencing factors rank as: distance from river > average slope > RDLS > relationship with the river.

3. The Tujia traditional settlement forms are mainly restricted by the local topography (including rivers and terrain) and roads, representing external driving mechanisms. However, intrinsic mechanisms like Feng Shui concepts, cultural customs, and clan culture also critically influence settlement formation. The natural environment is the background foundation, social economy is the motive power for developing and traditional culture is the spiritual connotation of traditional village form. Settlement form should be a rationally chosen and adapted form under the combined influence of natural, social and cultural factors.

This research not only provides a novel cognitive framework for traditional settlements, and but also establishes a quantitative study system for settlement morphology, which deepens the public’s understanding of the diverse values of villages. The results show that revealed the logic behind the formation of the Tujia traditional village material form and realized the systematic quantification of the traditional settlement form. The research conclusion has the universality and applicability characteristics of ethnic minority settlements in the Wuling Mountains Area. The findings and methodologies presented can serve as a reference for conservation and sustainable development of traditional villages, applicable to similar contexts other mountain settlements. On this basis, urban planners and cultural heritage protectors can better control the protection of traditional settlement cultural heritage, so as to realize the concentrated and contiguous protection and sustainable development of traditional settlements. Rural planning and design can also respond promptly to the development of modern society to achieve true cultural heritage and protect world cultural diversity. However, due to the vast and complex research scope of this study and the large sample size of the empirical research, the information covered is extremely complex. Constrained by the limitations of individual research, this study needs to be further deepened and improved in future research work. In particular, the quantification methods of cultural factors need to be further improved and enhanced.

Author contributions

D.R. and Y.L. have contributed to the study conception and design. Material preparation, data collection, and data analysis were performed by D.R., L.X, Z.B and Y.C. The first draft of the manuscript was written by D.R., and all authors commented on previous versions of the manuscript. All authors have read and approved the final manuscript.

Data availability

The datasets used and/or analysed during the current study available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Informed consent

This article does not contain any studies with human participants performed by any of the authors.