Cultural landscape clustering and zoning of traditional rural settlements Authors: Deyi Kong, Xinhui Fei, Ziyi Li Published: 2026 ABSTRACT Settlement cultural landscape serves as a prominent indicator reflecting regional disparities in cultural landscapes, among which the cultural landscape of traditional villages is the most direct and representative manifestation. From the cultural landscape perspective, this study deconstructs landscape characteristics and constructs a traditional village landscape classification system comprising 13 indicators across four dimensions: geographical environment, morphological layout, traditional building, and social culture. Using the K‑modes clustering method, 520 traditional village samples in the Min River Basin are classified into distinct cultural landscape types. By integrating these clustering results with city‑ and county‑level administrative divisions, dominant regional cultures, and river basin systems, the study further delineates eight traditional village cultural landscape zones: Jian River Northern Fujian Cultural Landscape Zone, Futun River Min–Gan Cultural Landscape Zone, Min River Middle Stream Eastern Min Cultural Landscape Zone, Min River Downstream Eastern Min Cultural Landscape Zone, Dazhang River Eastern Min Cultural Landscape Zone, You River Central Min Transitional Cultural Landscape Zone, Sha River Central Min Cultural Landscape Zone, and Jiulong River Min–Hakka Cultural Landscape Zone. This zoning provides a scientific reference for identifying village tourism characteristics, sustaining the inheritance of village cultural contexts, and advancing the concentrated and contiguous protection and utilization of traditional villages. Introduction Rural settlements emerged as a form of human habitation following the transition from prehistoric hunting-gathering societies to agrarian civilization, serving as the loci where agricultural producers congregated, labored, and thrived1. In China, rural settlements cover an area exceeding 2 million km², constituting an integral component of the national territory. Cultural landscapes represent the outcomes of human creative modification of the natural environment in response to societal needs, embodying the cultural characteristics and physiognomy of a given region2. Traditional settlements offer a crucial entry point for the study of cultural landscapes, among which traditional villages—condensing the wisdom and practical achievements of successive generations of working people—are the most direct and representative, possessing high historical, cultural, and artistic value3. In 2012, the release of the first batch of China’s Traditional Village List marked the official launch of state‑led conservation efforts. To date, 8155 villages have been included in the list and placed under legal protection. However, amid rapid urbanization, numerous villages that embody significant traditional historical and cultural values are gradually declining or even facing the risk of disappearance1. Once a traditional village is damaged, it may lead to the loss of core cultural assets and undermine the preservation and revitalization of cultural heritage. The vulnerability of these villages has become an internationally recognized concern4. How to rationally protect and develop traditional villages in the context of modernization has thus emerged as an urgent challenge, particularly for developing countries5,6. Faced with the reality of a large number and wide variety of traditional villages, conservation efforts targeting individual villages in isolation tend to disperse resources and, due to a lack of overall coordination, often lead to haphazard development. Consequently, research on the concentrated and contiguous protection and development of traditional villages has become a hot topic in recent years. Scholars have proposed conservation approaches such as cultural heritage clusters7 and regionalization8, articulating the concept of zone‑based (or cluster‑based) protection and focusing on protection methods and models9 to promote co‑construction, sharing, and diversified revitalization of contiguous zones, thereby forming a systematic and organic entity. This direction aligns well with the principles of integrated conservation and sustainable use emphasized in cultural landscape heritage10. As a distinctive heritage type and conservation concept, the cultural landscape offers inherent advantages for traditional village research. In terms of the composition of cultural landscapes, Kang, N. and Liu, C10. classified the cultural landscape of traditional villages in Anyi County into four aspects: natural, agricultural, living, and humanistic based on human perception. Wang, J. and Long, B11. suggested that the cultural landscape comprises two categories of landscape elements: material culture and social culture. Zheng, W. et al.12 further subdivided it into two major categories (material and intangible) and 15 subcategories, including village morphology, building materials, language, and festivals. Hu, Z. et al.13, from the perspective of settlement landscape genes, divided it into four parts: architecture, culture, environment, and layout. The above studies have laid an important foundation for the identification and classification of traditional village cultural landscapes. However, when these classification frameworks are directly applied to a study area like the Min River Basin, which is characterized by “numerous mountains, dense rivers and cultural diversity”, they often encounter many challenges. Based on this critical assessment, this study did not adopt any single model as a ready-made template. Instead, we regarded these previous classifications as a collection of candidate indicators, and then adjusted their categories, levels and identification criteria according to the unique geographical and cultural features of the basin. Within a certain spatial dimension, traditional villages exhibit significant intra-regional similarity and inter-regional differences. This spatial differentiation is the core research content for identifying the areal systematic characteristics of traditional villages and conducting regional systematic protection. Research on classification and regional division is not only a key approach to deeply grasp the regional traits of traditional villages but also an important foundation for promoting subsequent integrated protection and regional development14. Foreign scholars’ research on regional division originated from cultural geography, and the standards for regional division often take the comprehensive characteristics of cultural landscapes as indicators to conduct comprehensive regional division15. In recent years, research on the landscape regional division of traditional villages has mainly focused on three aspects. (1) Starting from different administrative units and cultural regions, it explores the regional concentration of traditional villages at different scales16,17. Specifically in China, relevant studies mostly focus on regions such as Jiangsu, Zhejiang, Anhui, Henan, and Sichuan provinces, with relatively little coverage of other regions18,19. (2) With cultural elements as the core, it conducts qualitative division of regions (clusters)26. Existing studies pay more attention to the clustering patterns of traditional villages in typical regions, ethnic groups, and ancient roads19,21, and explore qualitative analysis methods under different environments, ethnic groups, and cultures. (3) It uses specific indicators to quantify regional division28. Current methods for traditional village regional division mainly include the dominant factor analysis method2, multi-factor comprehensive analysis method23, hierarchical clustering method24 and geographic analysis and map overlay method25. Some studies have also integrated theories and methods related to landscape genes3,13,26. However, with the advancement of the trend of interdisciplinary integration, research methods from related fields such as geography, statistics, and computer science have been continuously integrated, and the research on traditional village regional division has gradually achieved a transition from qualitative analysis to quantitative research. The Min River is the largest river basin in Fujian Province and represents a significant cradle of culture in southeastern China. The basin features towering peaks, scattered and small plains, and a topographically enclosed and poorly accessible landscape. Despite these transportation constraints, villages along the river have historically relied on the Min River as a transportation hub, facilitating population mobility and cultural exchange. These villages, while exhibiting diverse landscapes, are interconnected, forming a distinctive cultural landscape system of high conservation and development value28. Nevertheless, existing research has largely focused on major tributaries29 and on individual or local village groups within county-level administrative units30, with emphasis on spatial distribution, influencing factors, settlement morphology, and dwelling types. Systematic studies targeting the entire traditional village system of the Min River Basin, particularly those addressing clustering and regionalization, remain relatively scarce. In the context of the Min River Basin, although the villages share a common riverine cultural lineage, they simultaneously exhibit pronounced differences in their natural environment and sociocultural characteristics. In order to move beyond fragmented, village-by-village conservation and to counteract the increasing homogenization of rural landscapes31, a systematic classification and zoning is urgently needed to provide a basis for regionally tailored ecological conservation and revitalization measures. This study takes 520 traditional villages in the Min River Basin as research samples to construct a landscape classification system for traditional settlements in the basin. The K-modes clustering method is employed to measure the overall dissimilarity among traditional villages. Based on the clustering results, the cultural landscape regions of traditional villages in the Min River Basin are delineated. By analyzing the characteristics and formation mechanisms of these landscape regions from the perspectives of the physical geographical environment and historical–cultural background, the spatial differentiation patterns are revealed. Our study is structured around the following research questions: (1) What are the key indicators for constructing a cultural landscape classification system that is suited to the geo-cultural context of traditional rural settlements in the Min River Basin? (2) Based on these indicators, what cultural landscape types can be identified? (3) How can these settlement types be further aggregated into a zoning scheme, and what are their typical landscape characteristics and formation mechanisms? This study deepens the theoretical understanding of traditional rural settlements in the Min River Basin, innovates the technical pathways for classifying and analyzing settlement morphology, and provides direct practical value for the conservation and development of rural areas in the Min River Basin. Methodology Study area The Min River Basin is located in northern Fujian Province, with a north-south length of 360 km, an east-west width of 280 km, and a drainage area of 61,000 km². Its upper reaches are composed of the Jian River, Futun River, Jin River, Jiulong River, and Sha River; the middle reaches consist of the middle stream, You River, and Gutian River; and the lower reaches are made up of the down stream and Dazhang River (Fig. 1). The basin administers 24 counties, supporting more than one-third of Fujian Province’s population29. The complex and diverse natural geographical and socio-cultural environment of the basin has nurtured highly distinctive landscape characteristics of traditional villages. Since 2012, a total of 520 villages in the basin have been included in the national-level and Fujian provincial-level traditional village lists. These villages have complete data records, are well-preserved, and cover a wide range of samples and diverse types—providing favorable conditions and abundant empirical data for selecting typical cases. Fig. 1Sit of study area. The map was generated by the authors using ArcGIS Desktop 10.8 (Esri, Redlands, CA, USA; https://www.esri.com/en-us/arcgis/products/arcgis-desktop/overview). Sit of study area. The map was generated by the authors using ArcGIS Desktop 10.8 (Esri, Redlands, CA, USA; https://www.esri.com/en-us/arcgis/products/arcgis-desktop/overview). Data source The data in our study is derived from four sources: (1) Sample data: Taking the main stream and tributaries of all levels of the Min River as the basin scope, a total of 520 traditional villages included in the national-level and Fujian provincial-level traditional village lists were sorted out as research samples. (2) Literature materials: Identification data for the index system of traditional village landscape gene identification was obtained through databases such as CNKI, Web of Science, and VIP, as well as books related to traditional villages. Basic data for traditional village regional division was acquired from atlases. For example, data on the river system distribution and geomorphic characteristics of traditional villages in the Min River Basin was obtained from Fujian Geographic Records32, and data on the distribution of language and ethnic subgroups in traditional villages was obtained from Fujian Provincial Records: Dialect Volume33. (3) Survey data: Some landscape gene identification data of traditional villages was directly obtained through field surveys of typical traditional villages in the Min River Basin, especially for samples lacking literature data. (4) Geographic materials: Digital Elevation Model (DEM) data related to the Min River Basin was obtained from the Geospatial Data Cloud Platform of the Chinese Academy of Sciences; data such as administrative division boundaries and river systems was downloaded from the National Geographic Information Resource Catalog Service System; geographic coordinates of traditional villages were picked via the Baidu Maps Coordinate Picker; and high-resolution satellite image data of traditional villages was acquired through platforms such as Ovi Interactive Maps and Google Earth. Study methodology Construction of the classification system for traditional villages in the Min River Basin This study selects landscape classification indicators for traditional villages with the primary aim of distinguishing differences in landscape features among traditional villages, following the principles of consensus, representativeness, and regional specificity8,11. Drawing on the landscape gene identification criteria and regionalization indicator categories documented in the existing literature, and integrating the unique landscape characteristics of traditional villages in the Min River Basin, a landscape gene identification indicator system for traditional villages in the Min River Basin is established. Specifically, the selection of gene categories for traditional villages mainly references the classification system of traditional settlement landscape genes proposed by Liu, P13., the feature deconstruction classification criteria proposed by Wei, K. et al.24, and the landscape classification indicators proposed by Sa, Q. et al.3. Accordingly, the cultural landscape of traditional villages is categorized into four constituent levels: geographical environment, morphological layout, traditional building, and social culture. Geographical environment is the most critical factor shaping the landscape of traditional villages. The complex and rugged terrain of the Min River Basin gives rise to a vertically varied three-dimensional climate. Village landscapes within the same region or basin tend to exhibit greater similarity due to ethnic migrations and cultural exchanges. Therefore, four indicators—river systems, topographic elevation, slope and aspect—are selected. These indicators comprehensively capture the macro-environmental characteristics of traditional villages in the Min River Basin and are amenable to data standardization and quantitative analysis. The identification data are primarily obtained through overlay analysis with thematic maps or Digital Elevation Model (DEM) data of the Min River Basin. Morphological layout represents the macroscopic spatial organization of traditional village landscapes. It is primarily reflected through three indicators: village layout, street layout, and courtyard layout. Among these, village layout captures the macroscopic spatial characteristics of traditional villages and, to a certain extent, influences the layout of streets and alleys. In contrast, courtyard layout is less affected by regional differences; multiple courtyard layout types may coexist within the same village, and similarities in courtyard layout can be observed across different regions. Therefore, village layout is adopted as the primary influencing factor and subjected to data standardization and clustering analysis, while street and alley layout and courtyard layout serve as auxiliary identification factors. The identification data are obtained through multiple approaches, including satellite imagery, morphological theory, literature review, image data, and field surveys. Traditional buildings are the micro-scale construction entities of the village landscape in terms of space. Due to the influence of factors such as climate, culture, and resources in different regions, the residences exhibit distinct differences in building form, building material, building wall, and roof form. Among them, building form and building wall are the core identification indicators because in the macroscopic regional classification, building form and building wall will to some extent determine the roof form and building materials. The identification data are mainly obtained through comprehensive acquisition based on literature review and on-site investigation. The influence of social culture on traditional village landscapes is both subtle and profound, primarily reflected through two indicators: language and ethnic subgroup and dominant culture. Among these, Language and ethnic subgroup are relatively easy to reflect some potential cultural characteristics and are also easy to identify; while dominant culture has a certain multi-level nature and ambiguity. Therefore, language and ethnic subgroup is regarded as data-based data, and dominant culture is regarded as descriptive data. The identification data is obtained not only by overlaying with specialized maps but also through comprehensive analysis and judgment based on literature review, field surveys, and other methods. The landscape classification indicator system for traditional villages in the Min River Basin is constructed from 4 levels and 13 indicator factors (Table 1). The selection of these 13 indicators follows a systematic, three‑tier screening logic designed to balance comprehensive landscape representation with the practical demands of data‑based clustering. First, an initial pool of candidate indicators was extracted from established landscape gene frameworks3,13,24 to ensure theoretical grounding. Second, indicators were filtered for their capacity to reflect the basin’s defining features—“numerous mountains, dense river systems, and cultural diversity”—eliminating those that showed minimal cross‑village variation in the Min River context (e.g., climate‑zone indicators, which are largely homogeneous basin‑wide). Third, and most critically for the socio‑cultural dimension, indicators were evaluated for their quantifiability and spatial identifiability. “Language and ethnic subgroup” was retained as a data‑based indicator because it can be unambiguously mapped to discrete, spatially continuous units using linguistic atlases and verified through field surveys, thus meeting the input requirements of K‑modes clustering. “Dominant culture,” while inherently more ambiguous, was preserved as a descriptive indicator precisely because it captures the multi‑layered, non‑discrete cultural identities—such as Neo‑Confucian versus Hakka influences—that are essential for interpreting the clusters, even if they cannot be reduced to a single categorical input without loss of nuance. This two‑track treatment allows socio‑cultural differentiation to enter both the algorithmic clustering and the qualitative interpretation, a design that was compared against alternatives (e.g., excluding social culture entirely, or forcing all cultural variables into crude binary codes) and found to maximize both analytical rigor and cultural fidelity. Eventually, a landscape gene identification database was formed. Table 1Classification Indicators and Attributes of Traditional Villages in the Min River Basin.Village Cultural Landscape CompositionLandscape classification indicatorsData SourceData TypeQuantificationGeographical environmentalRiver systemsOverlay and identification of specialized basin mapsData-basedCategorical: 1 = Jian River, 2 = Futun River, 3 = Jin River, 4 = Jiulong River, 5 = Sha River, 6 = Min River Middle Stream, 7 = You River, 8 = Gutian River, 9 = Min River Down Stream, 10 = Dazhang RiverTopographic elevationIdentified via DEM data in ArcGISData-basedCategorical: 1 = Plains (H < 200 m), 2 = Hills (200 m ≤ H < 500 m), 3 = Low Mountains (500 m ≤ H < 1000 m), 4 = Medium Mountains (H ≥ 100 m)SlopeIdentified via DEM data in ArcGISData-basedCategorical: 1 = Plains (0° < S ≤ 0.5°), 2 = Gentle Slopes (0.5° < S ≤ 2°), 3 = Moderate Slopes (2° < S ≤ 5°), 4 = Slopes (5° < S ≤ 15°), 5 = Steep Slopes (15° < S ≤ 35°), 6 = Cliffy Slopes (35° < S ≤ 55°)AspectIdentified via DEM data in ArcGISData-basedCategorical: 1 = Shady Slopes (0° < A ≤ 90°; 270° < A ≤ 360°),2 = Sunny Slopes (90° < A ≤ 270°)Morphological layoutVillage layoutIdentified based on satellite images and morphological theoryData-basedCategorical: 1 = Clustered-style, 2 = Fortified-style, 3 = Beaded- style, 4 = Dispersed-styleStreet layoutIdentified comprehensively based on satellite images and morphological theory, combined with the characteristics of village layoutDescriptiveCategorical: 1 = Grid-style, 2 = Contour-following Style, 3 = Main-road-connecting Style, 4 = Flexible-style, 5 = Mixed-styleCourtyard layoutIdentified comprehensively based on satellite images and morphological theory, combined with the characteristics of village layoutDescriptiveCategorical: 1 = Inline-style, 2 = Enclosed-square Style, 3 = Concave-style, 4 = Compound-styleTraditional buildingsBuilding formLiterature review, field surveys, image archives, observation recordsData-basedCategorical: 1 = Town houses, 2 = Courtyard houses, 3 = Courtyard well, 4 = Tang Heng houses, 5 = FortressesBuilding materialComprehensive identification based on Literature Review and Field Surveys, combined with Building Form, supplemented by Image Archives and Observation RecordsDescriptiveCategorical: 1 = Earth, 2 = Brick, 3 = Stone, 4 = WoodBuilding wallLiterature review, field surveys, image archives, observation recordsData-basedCategorical: 1 = Wooden Wall, 2 = Blue-tiled wall,3 = Entirely rammed earth wall, 4 = Boot-and-hat wall, 5 = Tile-coated wall, 6 = Rammed earth wall, 7 = Composite material wall, 8 = Stuccoed wallRoof formLiterature review, field surveys, image archives, observation recordsDescriptiveCategorical: 1 = Overhanging gable apex, 2 = Horse-head-shaped firewall top, 3 = Five elements-shaped firewall top, 4 = Saddle-shaped firewall topSocial culturalLanguage and ethnic subgroupIdentified by overlaying with specialized language and ethnic subgroup maps and combining with field surveysData-basedCategorical: 1 = Wu dialect, 2 = Nanping dialect, 3 = Min-Hakka dialect, 4 = Min-Gan dialect, 5 = Eastern Min dialect, 6 = Central Min dialect, 7 = Northern Min dialect, 8 = Transitional Min dialect, 9 = Min-Hakka-Gan dialectDominant cultureLiterature review, field surveys, comprehensive analysisDescriptiveCategorical: 1 = Neo-Confucian Culture, 2 = Gan Culture, 3 = Central plains culture, 4 = Hakka culture, 5 = Immigrant culture, 6 = Mountain forest culture, 7 = Mixed culture, 8 = River-city culture, 9 = Maritime culture Table 1Classification Indicators and Attributes of Traditional Villages in the Min River Basin.Village Cultural Landscape CompositionLandscape classification indicatorsData SourceData TypeQuantificationGeographical environmentalRiver systemsOverlay and identification of specialized basin mapsData-basedCategorical: 1 = Jian River, 2 = Futun River, 3 = Jin River, 4 = Jiulong River, 5 = Sha River, 6 = Min River Middle Stream, 7 = You River, 8 = Gutian River, 9 = Min River Down Stream, 10 = Dazhang RiverTopographic elevationIdentified via DEM data in ArcGISData-basedCategorical: 1 = Plains (H < 200 m), 2 = Hills (200 m ≤ H < 500 m), 3 = Low Mountains (500 m ≤ H < 1000 m), 4 = Medium Mountains (H ≥ 100 m)SlopeIdentified via DEM data in ArcGISData-basedCategorical: 1 = Plains (0° < S ≤ 0.5°), 2 = Gentle Slopes (0.5° < S ≤ 2°), 3 = Moderate Slopes (2° < S ≤ 5°), 4 = Slopes (5° < S ≤ 15°), 5 = Steep Slopes (15° < S ≤ 35°), 6 = Cliffy Slopes (35° < S ≤ 55°)AspectIdentified via DEM data in ArcGISData-basedCategorical: 1 = Shady Slopes (0° < A ≤ 90°; 270° < A ≤ 360°),2 = Sunny Slopes (90° < A ≤ 270°)Morphological layoutVillage layoutIdentified based on satellite images and morphological theoryData-basedCategorical: 1 = Clustered-style, 2 = Fortified-style, 3 = Beaded- style, 4 = Dispersed-styleStreet layoutIdentified comprehensively based on satellite images and morphological theory, combined with the characteristics of village layoutDescriptiveCategorical: 1 = Grid-style, 2 = Contour-following Style, 3 = Main-road-connecting Style, 4 = Flexible-style, 5 = Mixed-styleCourtyard layoutIdentified comprehensively based on satellite images and morphological theory, combined with the characteristics of village layoutDescriptiveCategorical: 1 = Inline-style, 2 = Enclosed-square Style, 3 = Concave-style, 4 = Compound-styleTraditional buildingsBuilding formLiterature review, field surveys, image archives, observation recordsData-basedCategorical: 1 = Town houses, 2 = Courtyard houses, 3 = Courtyard well, 4 = Tang Heng houses, 5 = FortressesBuilding materialComprehensive identification based on Literature Review and Field Surveys, combined with Building Form, supplemented by Image Archives and Observation RecordsDescriptiveCategorical: 1 = Earth, 2 = Brick, 3 = Stone, 4 = WoodBuilding wallLiterature review, field surveys, image archives, observation recordsData-basedCategorical: 1 = Wooden Wall, 2 = Blue-tiled wall,3 = Entirely rammed earth wall, 4 = Boot-and-hat wall, 5 = Tile-coated wall, 6 = Rammed earth wall, 7 = Composite material wall, 8 = Stuccoed wallRoof formLiterature review, field surveys, image archives, observation recordsDescriptiveCategorical: 1 = Overhanging gable apex, 2 = Horse-head-shaped firewall top, 3 = Five elements-shaped firewall top, 4 = Saddle-shaped firewall topSocial culturalLanguage and ethnic subgroupIdentified by overlaying with specialized language and ethnic subgroup maps and combining with field surveysData-basedCategorical: 1 = Wu dialect, 2 = Nanping dialect, 3 = Min-Hakka dialect, 4 = Min-Gan dialect, 5 = Eastern Min dialect, 6 = Central Min dialect, 7 = Northern Min dialect, 8 = Transitional Min dialect, 9 = Min-Hakka-Gan dialectDominant cultureLiterature review, field surveys, comprehensive analysisDescriptiveCategorical: 1 = Neo-Confucian Culture, 2 = Gan Culture, 3 = Central plains culture, 4 = Hakka culture, 5 = Immigrant culture, 6 = Mountain forest culture, 7 = Mixed culture, 8 = River-city culture, 9 = Maritime culture Classification Indicators and Attributes of Traditional Villages in the Min River Basin. Methods for the zone division of traditional villages in the Min River Basin Traditional cultural regionalization methods mostly take the qualitative analysis of dominant geographical and cultural factors as the core and then conduct data statistics and classification based on this. However, such methods have the problem of insufficient objectivity and accuracy. On the other hand, relying solely on quantitative clustering methods tends to lead to drawbacks such as confusion between primary and secondary factors and overemphasis on the integration of multiple indicators. Given the large geographical span and diverse cultural types of the Min River Basin, our study draws on the principles regarding cultural regionalization proposed by cultural geographers—including relatively consistent or similar cultural landscapes, similar regional cultural development processes, basically contiguous cultural regions, and having a core element that reflects regional cultural characteristics34—to conduct research on the regionalization of traditional villages in the Min River Basin. In this study, the identification results of 8 data-based indicators of 520 traditional villages were subjected to standardization and quantitative calculation via K-modes clustering, so as to obtain the macro-level regionalization results of traditional villages in the Min River Basin. However, due to the natural drift and regional variation of the identification factors, there exist characteristics such as long-distance scattered distribution of villages with the same attributes and short-distance mixed distribution of villages with different attributes to a certain extent35. Therefore, on the basis of macro-level regionalization, it is necessary to conduct qualitative analysis by combining the identification results of descriptive data-including street layout, courtyard layout, building material, roof form, and dominant culture—to further clarify the regionalization attributes and areal boundaries of traditional villages. Finally, a complete regionalization process is accomplished by combining quantitative and qualitative analyses. (1) Data Standardization Processing. The 8 categories of data-based indicators from the landscape features identification results of traditional villages in the Min River Basin serve as the fundamental basis for subsequent clustering analysis. However, due to differences in information types and dimensions among these data, they cannot be directly used for statistical analysis. This study employs Python for data preprocessing: Z-score normalization is applied to numerical variables to eliminate dimensional effects, and one-hot encoding is used for categorical variables. Given that the objective is to cluster the original samples (N = 520) while preserving their distributional characteristics, and considering the relatively balanced sample size and the limited impact of class imbalance on the analytical results, no additional frequency weighting is performed. Through the above preprocessing, the inconsistency in variable types and dimensional scales within the data matrix is resolved, thereby meeting the requirements for subsequent data analysis. (2) K-modes Clustering Analysis. Clustering analysis serves as a commonly used statistical method in sample classification research, and different data types correspond to different suitable clustering analysis methods. In this study, the identification data of traditional villages belong to categorical attribute data, which not only have the characteristics of discrete variables but also involve a large number of samples. Therefore, the K-modes clustering method—developed on the basis of the K-means clustering method-is selected, and the relevant clustering calculations are completed with the help of the Python programming language. The core of the K-modes clustering method lies in the measurement of dissimilarity of categorical data; it replaces the concept of “distance” in the K-means algorithm with dissimilarity. Within the K-modes algorithm system, the smaller the dissimilarity value, the smaller the corresponding distance36. According to the dissimilarity calculation function in the K-modes algorithm, the dissimilarity between traditional village samples x and y can be defined as37:\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{array}{l}d(x,y)={\sum\:}_{j=1}^{m}\frac{{n}_{xg}+{n}_{yg}}{{n}_{xg}{n}_{yg}}\delta\:({x}_{g},{y}_{g})\\\:\delta\:({x}_{g},{y}_{g})=\{\begin{array}{ll}0&\:({x}_{g}={y}_{g})\\\:1&\:({x}_{g}\ne\:{y}_{g})\end{array}\end{array}$$\end{document} d(x,y) represents the overall dissimilarity between samples x and y; m\:represents the number of indicators contained in a sample; \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\delta\:({x}_{g},{y}_{g})$$\end{document} is defined as follows: when the attribute value of the g-th indicator of sample x is the same as that of sample y, the distance between\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{x}_{g}$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${y}_{g}$$\end{document} is 0; when the attribute values are different, the distance between \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${x}_{g}$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${y}_{g}$$\end{document} is 1. \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${n}_{xg}$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${n}_{yg}$$\end{document} respectively represent the frequencies of the attribute values of the g-th indicator of samples x and y among all samples to be classified. When the g-th attribute classification feature values (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${x}_{g}$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${y}_{g}$$\end{document}) of traditional villages x and y are rarer among the attributes of all traditional villages to be classified, the greater the weight of this landscape classification feature indicator g in the function for calculating the dissimilarity between traditional villages x and y. Correspondingly, the greater the dissimilarity between x and y in terms of this attribute classification indicator g, which makes it easier to regionalize them. The K‑modes clustering was implemented using the klaR package in R (version 1.7‑3). For each candidate K value, n_init = 50 random initializations were performed, and the solution with the lowest within‑cluster dissimilarity was retained. The maximum number of iterations was set to 100; convergence was typically reached within 6–12 iterations. The random seed was fixed (set.seed = 2024) for reproducibility. Regarding class imbalance, the most frequent attribute level in any single indicator accounts for no more than 34% of the sample, and the built‑in inverse frequency weighting in the K‑modes dissimilarity function further mitigates the influence of dominant categories. (3) Elbow Method. The Elbow Method is mainly used to determine the number of clusters for traditional villages in the Min River Basin. The determination of the number of clusters in K-modes clustering, commonly referred to as the selection of K value, can usually be done through random initialization or manual setting. Different initial cluster centers and K values may lead to different clustering results. In this study, the Elbow Method is adopted to determine the most appropriate K value. The Elbow Method establishes different K-modes models by examining the changes in the Sum of Squared Errors (SSE) corresponding to different K values (number of clusters). As the number of clusters increases, the sample division becomes more refined, the aggregation degree of each category becomes higher, and the Sum of Squared Errors (SSE) will gradually decrease. The Sum of Squared Errors (SSE) is the core indicator of the Elbow Method:\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$SSE = \sum\limits_{{t = 1}}^{k} {p \in \:l_{t} } |p - m_{t} |^{2}$$\end{document} \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${l}_{t}$$\end{document} denotes the t-th cluster, p represents a sample point in \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${l}_{t}$$\end{document}, and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${m}_{t}$$\end{document} is the centroid of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${l}_{t}$$\end{document} (i.e., the average value of all samples in this cluster), which reflects the quality of the clustering effect. When the K value is lower than the actual number of clusters, as the K value increases gradually, the degree of aggregation of each cluster improves, thus the SSE decreases significantly. However, when the K value is close to or reaches the actual number of clusters, the enhanced error reduction effect brought by increasing the K value will weaken significantly, leading to a sharp drop in the rate of SSE decrease. Subsequently, as the K value continues to increase, the SSE tends to stabilize. Therefore, the relationship graph between SSE and K value takes on an elbow shape, and the K value corresponding to this “elbow” is regarded as the optimal number of clusters for the data. In our study, the Elbow Method was used to analyze the 8 main features after data processing to determine the optimal K value. The sum of squared errors (SSE) corresponding to different K values was calculated using the klaR package in R. As shown in Fig. 2, a pronounced elbow point appeared at K = 8. To further validate the rationality of this K value, K‑modes clustering was performed for K = 7, 8, and 9, and the clustering results were interpreted based on the practical geographical significance of traditional village landscapes. The results showed that at K = 8, each cluster exhibited a clearly distinguishable cultural–geographical differentiation pattern; at K = 7, some landscape types were overly merged; and at K = 9, redundant groups lacking interpretability emerged. By combining the elbow method with practical interpretation, the 520 villages were ultimately classified into eight cultural landscape types. Fig. 2Determining the number of clusters using the Elbow Method. Determining the number of clusters using the Elbow Method. Results Identification results a of classification indicators for traditional village landscapes in the Min River Basin We converted the traditional villages in the Min River Basin into abstract “point” elements, obtained their precise longitude and latitude data via Baidu Maps, and then completed geoinformation processing on the ArcGIS platform. Meanwhile, we integrated the data-based identification results, and finally realized the visual presentation of multiple categories of landscape genes, including river systems, topographic elevation, slope, aspect, village layout, building form, building wall, as well as language and ethnic subgroup. In terms of river systems (Fig. 3a), the traditional villages in the Min River Basin are mostly concentrated in the ranges of river systems such as the Jian River, Futun River, the middle stream, and Dazhang River, forming multiple densely distributed zones within the basin. In terms of topographic elevation (Fig. 3b), the traditional villages in the Min River Basin have an average topographic elevation of 542 m, a maximum topographic elevation of 1316 m, and a minimum topographic elevation of 0 m. Among these traditional villages, 18.27% are located in plains (95 villages), 37.12% in hills (193 villages), and 44.16% in low mountains and medium mountains (232 villages). This indicates that the traditional villages in the Min River Basin have the characteristic of being built against mountains. In terms of slope (Fig. 3c), the traditional villages in the Min River Basin have an average slope of 18.92°, a minimum slope of 0.48°, and a maximum slope of 82.67°. As the slope increases, the number of villages in the basin shows a trend of a sharp increase followed by a sharp decrease, reaching its peak in the range of 5°-15° with a total of 294 villages. In addition, 349 traditional villages are distributed in the range of 5°-35° in terms of slope, accounting for 67.12% of the total number. This further confirms the characteristic of the basin’s villages being built against mountains. In terms of aspect (Fig. 3d), 308 (59%) of the traditional villages in the Min River Basin are located on sunny slopes, and 202 (41%) are located on shady slopes; the number distribution of the two is relatively close. To a certain extent, this indicates that due to the constraints of the Min River Basin’s complex geographical conditions—where mountains account for roughly 80%, water bodies 10%, and farmland 10%, a rough description of the region’s geographical composition—the flat space available for village construction is very limited. It is difficult for village site selection to fully take into account the sunlight demand, so “sunlight” is not a decisive factor. This precisely reflects the ancestors’ certain adaptive compromise to natural constraints. In terms of village layout (Fig. 3e), dispersed-style traditional villages are the most numerous but distributed relatively scattered, which is mainly related to the mountainous topography and geomorphology of the Min River Basin; Clustered-style traditional villages are concentrated in the lower reaches of the Min River, which is somewhat associated with the alluvial plain topography there; Fortified-style traditional villages are mainly distributed along the Sha River, which is mainly influenced by the immigrant defense culture; Beaded-style traditional villages are mainly distributed along the Jian River, and the hilly and gully landform is the main reason for their beaded-style layout. In terms of building form (Fig. 3f), town houses, as the earliest and most fundamental architectural style in human history, are distributed throughout the entire basin; traditional villages dominated by courtyard wells have the widest distribution across the basin, covering river areas such as the Jian River, Futun River, the middle stream, and the down stream. These courtyard wells typically feature a square outline, enclosed appearance, and a structural characteristic of short width and long depth; Traditional villages centered on courtyard houses are mainly distributed in the middle and down reaches of the Min River, with structural features of short depth and wide width. The exteriors of courtyard houses are usually enclosed by waist-high courtyard walls, creating working spaces such as grain-drying yards; Traditional villages dominated by Tang Heng houses are primarily distributed in the Jiulong River area. This building form has a symmetrical layout, with the main Tang house in the center, supplemented by two, four, or more rows of side Heng houses symmetrically arranged on the left and right, all facing the central axis of the main Tang house, and it has a rectangular structural feature; Traditional villages centered on fortresses are mainly distributed in the Sha River area. This building form is tall and enclosed, with defensive properties, forming a unique building form. In terms of building walls (Fig. 3g), wooden walls, as the most fundamental wall type, are distributed throughout the entire basin; Rammed earth walls are the most widely used wall material in the Min River Basin, concentrated in areas such as the Jian River, Gutian River, the middle stream, and Dazhang River; Traditional villages with blue-tiled walls are mainly distributed in the Futun River and Jiulong River regions, and the use of this material is somewhat related to the profound influence of Gan culture in this area; Composite material walls are concentrated in the You River area, which is somewhat related to the prevalence of mountain forest culture in this region; Traditional villages dominated by stuccoed walls are mainly distributed in the down stream, and river city culture is the main reason for this phenomenon. In terms of language and ethnic group (Fig. 3h), the Jian River area is dominated by northern Min dialect; the Futun River area is dominated by Min-Gan dialect; the Sha River area is dominated by central Min dialect; the middle stream area is dominated by Nanping dialect; the down stream area is dominated by eastern Min dialect. Among these areas, different dialect sub-groups are also interspersed, creating a diverse linguistic environment. Fig. 3Identification results of classification indicators for traditional village landscapes in the Min River Basin. The map was generated by the authors using ArcGIS Desktop 10.8 (Esri, Redlands, CA, USA; https://www.esri.com/en-us/arcgis/products/arcgis-desktop/overview). Identification results of classification indicators for traditional village landscapes in the Min River Basin. The map was generated by the authors using ArcGIS Desktop 10.8 (Esri, Redlands, CA, USA; https://www.esri.com/en-us/arcgis/products/arcgis-desktop/overview). Clustering results of traditional villages in the Min River Basin Clustering calculations were performed on 520 traditional villages in the Min River Basin by dividing them into 8 categories using K-modes. The calculation results were combined with the longitude and latitude of the traditional villages, and geoinformation data visualization was conducted via ArcGIS, resulting in the clustering results shown in Fig. 4. As can be seen, the 8 categories of traditional villages each exhibit a distinct clustered distribution, with obvious basin characteristics. Traditional villages in river areas such as the Jiulong River, Sha River, and You River are densely distributed, forming small-scale clustered areas; traditional villages in the Jian River area and Futun River area have formed two large-scale clusters. Based on the K-modes clustering results of data-based data, the traditional villages in the Min River Basin were divided into 8 major zones. On this basis, individual traditional villages that are scattered within large-scale clusters of this type but do not belong to this type are directly classified into this type; for traditional villages located at the junction zones or transition zones between two types with ambiguous classification relationships, their classification type is comprehensively determined by combining the identification results of descriptive data; for two types of traditional villages with a significant distance between them, regionalization and boundary revision are carried out with reference to basin boundaries or administrative boundaries, thus forming the final regionalization type results. Fig. 4K-modes clustering results of traditional villages in the Min River Basin. The map was generated by the authors using ArcGIS Desktop 10.8 (Esri, Redlands, CA, USA; https://www.esri.com/en-us/arcgis/products/arcgis-desktop/overview). K-modes clustering results of traditional villages in the Min River Basin. The map was generated by the authors using ArcGIS Desktop 10.8 (Esri, Redlands, CA, USA; https://www.esri.com/en-us/arcgis/products/arcgis-desktop/overview). Regional division results of traditional villages in the Min River Basin The division of cultural landscape zones of traditional villages should follow the principles of landscape feature similarity, cultural background consistency, and geographic spatial continuity. Based on the results of the qualitative and quantitative identification of the cultural landscape of traditional villages in the Min River Basin, this study divides the basin into eight cultural landscape zones. To fully capture the specific human–environment relationships and regional characteristics of each zone, the naming emphasizes environmental orientation, the association with human activities, and the dominant ethnic and folk culture27,38. In line with the pronounced river‑basin characteristics observed in this study, the eight zones are named using the format “river name + cultural zone identity” (Fig. 5). Fig. 5Result of regional division of traditional villages in the Min River Basin. The map was generated by the authors using ArcGIS Desktop 10.8 (Esri, Redlands, CA, USA; https://www.esri.com/en-us/arcgis/products/arcgis-desktop/overview). Result of regional division of traditional villages in the Min River Basin. The map was generated by the authors using ArcGIS Desktop 10.8 (Esri, Redlands, CA, USA; https://www.esri.com/en-us/arcgis/products/arcgis-desktop/overview). Jian River Northern Fujian Cultural Landscape Zone Zone I is the Jian River Northern Fujian Cultural Landscape Zone. This region is predominantly characterized by mountainous and hilly terrain. Traditional villages are primarily situated at topographic elevations ranging from 200 to 1000 m above sea level, with a minority located above 1000 m. Slopes vary between 5° and 35°, and the number of villages on sunny slopes is roughly equivalent to those on shady slopes. Village layouts exhibit a beaded style influenced by the terrain, while street layouts are mostly of a flexible-style. Courtyard layouts are predominantly of the hui-shaped and ao-shaped types, featuring enclosed-square style and concave-style. Roofs commonly employ the horse-head-shaped firewall top. Construction materials consist primarily of earth and wood, characterized by rammed earth walls with boot-and-hat, stone wall foundations, and brick door-hooded gates. These architectural features represent adaptations to the mountainous, limited-flatland geography of northern Fujian and the region’s humid, rainy climatic conditions. The dominant language in this zone is Northern Min dialect. Neo-Confucianism and tea culture constitute the principal cultural foundations, giving rise to vibrant folk activities such as Mountain Calling, Panting Bridge Walking, Candle Bridge Lanterns, and so on39. A notable characteristic of villages in this region is the presence of academies and schools established to promote Zhu Xi’s Neo-Confucianism. Owing to its location in the hinterland of the Wuyi Mountains and its topographically enclosed setting, villages in this zone have been preserved, largely avoiding the large-scale demolitions and reconstructions that occurred elsewhere in earlier periods. Concurrently, tea culture has provided sustained economic support for the region, while clan structures and folk activities have maintained community cohesion, mitigating disintegration following rural depopulation. Tea culture originated in the Wuyi Mountains. Historically, the influx of merchants engaged in tea trade was substantial; attracted by the scenic landscapes and driven by operational requirements, merchants with moderate wealth established settlements along the tea trade routes. The constructed enclosed-square style courtyards were relatively large in scale, characterized by horse-head-shaped gables with graceful profiles and intricate patterned details, reflecting a synthesis of indigenous architectural traditions with exogenous influences. In the context of contemporary development, this region can leverage the tea culture industry chain to transform traditional villages into vibrant spaces integrating tea culture and tourism, thereby establishing a virtuous cycle linking heritage conservation with local livelihoods. Futun River Min–Gan Cultural Landscape Zone Zone II is the Futun River Min–Gan Cultural Landscape Zone. This region is predominantly characterized by mountainous terrain, with topographic elevations mostly above 500 m and slopes generally exceeding 5°. The number of traditional villages situated on sunny slopes and shady slopes is approximately equal. Influenced by the mountainous topography, village layouts are predominantly of dispersed-style. Street layouts are primarily contour-following style and flexible-style, while courtyard forms are mainly of the concave-style and enclosed-square style, with roof forms mostly being horse-head-shaped firewall top. Building materials are mainly brick and wood, consisting of blue-tiled walls, blue-tiled wall foundations, and wooden door-hooded gates. The enclosed-square style courtyard well structure represents an adaptation to the mountainous, rainy climate and the functional requirements of family-centered living, while the dispersed layout pattern responds to the scattered distribution of constructible land in mountainous environments. This region is mainly distributed within the Min-Gan dialect and the Min-Hakka-Gan dialect area. Deeply influenced by Gan (Jiangxi province) culture, it has fostered highly distinctive folk activities such as Fan God Dance, Victory Drum Dance, Nuo Dance, and so on40. Owing to its location in the hinterland of the Wuyi Mountains at the juncture of Fujian and Jiangxi provinces, characterized by complex terrain and relative transportation isolation, local cultural customs and practices have been preserved to the present day, largely shielded from external disturbances. Represented by the “Great Sage Culture” with distinctive Fujian-Taiwan characteristics, the cultural landscape of traditional villages also adheres to Confucian ritual order. Residential architecture is dominated by enclosed-square style courtyards, characterized by square and regular forms with understated and elegant color schemes. In conservation efforts, emphasis should be placed on maintaining the continuity of cultural inertia to prevent excessive tourism development from disrupting the existing ritual spaces and ceremonial traditions. Min River Middle Stream Eastern Min Cultural Landscape Zone Zone III is the Min River Middle Stream Eastern Min Cultural Landscape Zone. This region is predominantly characterized by low-mountain and hilly landforms. Traditional villages are mostly situated at topographic elevations ranging from 200 to 1000 m, with slopes predominantly between 5° and 35°, and the majority are located on sunny slopes. Village layouts are predominantly of the dispersed-style, while clustered traditional villages have formed in relatively flat river valley areas, reflecting the scattered distribution of constructible land characteristic of low-mountain and hilly terrain. Street layouts are similarly influenced by terrain, mostly presenting a flexible-style; however, traditional villages in river valley plains typically exhibit a grid pattern. To adapt to the mountainous, rainy climate and the requirements of family living, architectural forms are dominated by enclosed-square style, with roof forms mostly adopting the horse-head-shaped firewall style. Building materials are mainly earth and wood, consisting of entirely rammed earth wall, stone wall foundations, and wooden door-hooded gates41. Linguistically, the region is mainly home to Eastern Min dialect and Nanping dialect. Owing to its location along the middle reaches of the Min River Basin, this zone has been influenced by diverse cultures—including Central Plains culture, mountain forest culture, and immigrant culture—resulting in the formation of a hybrid cultural tradition. The region is endowed with abundant forest resources and a rich diversity of snake species. Through prolonged coexistence, local inhabitants have venerated snakes as deities, engaging in their worship and ritual propitiation, thereby giving rise to a distinctive snake culture. This has spawned folk activities imbued with rustic mountain characteristics, such as the Snake King Festival, Taiping Drum Dance, and Bench Dragon Dance. The preservation of villages in this region does not rely on isolation or a single economic base, but rather on the cultural resilience engendered by the integration of diverse cultural traditions. The superimposition of Central Plains culture, mountain forest culture, and immigrant culture has endowed these villages with the capacity to absorb external changes and transform them into distinctive local characteristics, thereby sustaining their vitality through historical transformations. Min River Downstream Eastern Min Cultural Landscape Zone Zone IV is the Min River Downstream Eastern Min Cultural Landscape Zone. This region is predominantly characterized by plain landforms, with topographic elevations mostly below 200 m and slopes predominantly ranging from 0° to 15°; most villages are located on sunny slopes. In terms of village layout, the clustered type is predominant. Street layouts mostly present a grid-style pattern, while courtyard layouts are mainly enclosed-square style. Building forms are dominated by courtyard wells, with roof forms mostly being saddle-shaped firewall top. Building materials are mainly earth and wood, consisting of stuccoed walls, stone wall foundations, and wooden door-hooded gates, reflecting the spatial order of the settlement of the clan. Linguistically, the primary language in this region is Eastern Min dialect. Adjacent to the East China Sea and benefiting from convenient and developed water transportation, this region has fostered river-city culture and marine culture, and retains the distinctive customary activity of Deity Procession42. Historically, large-scale southward migrations of northern populations occurred on multiple occasions, and owing to its advantageous geographical location, this region became the preferred entry point for Han Chinese migrating into Fujian via coastal routes. Immigrants integrated northern architectural traditions with local vernacular dwellings, introducing exaggeratedly sculpted fire-sealing gables on the basis of enclosed-square style courtyard. Subsequently, with the southward shift of the economic center and the development of the Maritime Silk Road, residential architecture grew in scale, while carving, painting, and molded plasterwork became increasingly refined, giving rise to architectural characteristics that are simultaneously grand in overall conception and meticulously detailed in execution. The preservation of villages in this region is primarily sustained by the combined forces of economic prosperity and the clan system. Historically serving as the frontline for immigration into Fujian and a key node along the Maritime Silk Road, this region accumulated substantial economic resources, enabling continuous maintenance and renewal of vernacular architecture. Concurrently, the spatial order of the clan settlements strengthened the cohesion of the village, enabling it to maintain its basic layout during the process of modernization. Dazhang River Eastern Min Cultural Landscape Zone Zone V is the Dazhang River Eastern Min Cultural Landscape Zone. This region is predominantly characterized by mountainous landforms, with traditional villages mostly situated at topographic elevations above 500 m (some exceeding 1000 m) and slopes predominantly ranging from 5° to 35°. Influenced by the mountainous terrain, village layouts are predominantly of the beaded-style, while street layouts demonstrate flexible-style configurations. Courtyard layouts are mainly of the inline-style and concave- style. Building forms are dominated by courtyard houses, with roof forms mostly being saddle-shaped firewall top. Building materials are mainly earth and wood, consisting of rammed earth walls with tile-coated wall, stone wall foundations, and wooden door-hooded gates, reflecting the constraints of limited constructible land in mountainous environments. Linguistically, the primary language in this region is the Eastern Min dialect, influenced by river-city culture43. Owing to its location in the upper reaches of the Dazhang River basin, characterized by extensive mountainous terrain, some villages have been preserved to the present day by virtue of their topographic advantages, having largely escaped large-scale demolitions and reconstructions that occurred elsewhere in earlier periods. Among these are distinctive defensive structures—fortresses (villagers built the walls with rammed earth and gravel, and set up watchtowers and fortresses around the courtyards) - constructed during the Ming and Qing dynasties to resist foreign invasions. Concurrently, folk activities such as Deity Procession and Dragon Procession have also been sustained to the present day. In contemporary development, villages of this type must exercise caution against excessive development facilitated by improved transportation access. The defensive landscape should be regarded as a core conservation priority to prevent the loss of their inherent security logic resulting from spatial opening. You River Central Min Transitional Cultural Landscape Zone Zone Ⅵ is the You River Central Min Transitional Cultural Landscape Zone. Located in the mountainous interior of Fujian Province, the region is dominated by hilly and mountainous landforms. Traditional villages are mostly situated at topographic elevations above 500 m, with some exceeding 1000 m, and slopes are generally greater than 5°. The numbers of villages on sunny slopes and shady slopes are roughly equal. Village layouts are predominantly dispersed-style, and street layouts mostly present a inline-style. Courtyard layouts are mostly enclosed-square style and concave-style, adapting to the dispersed distribution of constructible land in mountainous areas. Building forms are dominated by courtyard houses, with roof forms mostly being overhanging gable apex. Building materials are primarily wood, giving rise to the architectural characteristic of “thousands of pillars touching the ground.” The combination of composite material walls, stone wall foundations, and gate-tower style gates responds to the self-sufficient production and living needs of the mountain forest region. Due to relatively late development, the population consists mainly of migrants from other areas. Folk culture and architectural forms exhibit multi‑cultural fusion characteristics. Linguistically, the region is mainly distributed within the Transitional Min Dialect area. Mountain forest culture prevails, fostering folk activities with both traditional heritage and regional characteristics, such as the Dragon Lantern Parade and Zhuzi Memorial Ceremony44. This self-sufficient production and lifestyle, relying on mountain forests, results in low external dependence of villages, demonstrating strong autonomy in the process of modernization. Sha River Central Min Cultural Landscape Zone Zone Ⅶ is the Sha River Central Min Cultural Landscape Zone. The region is dominated by hilly and mountainous landforms, with traditional village topographic elevations mostly ranging from 200 to 1000 m and slopes mostly between 5° and 35°. The numbers of villages located on sunny slopes and shady slopes are roughly equal. Village layouts are predominantly fortified-style, street layouts are predominantly grid-style, and courtyard layouts are mostly enclosed-square style or the compound-style. Building forms are dominated by fortresses, with roof forms mainly including overhanging gable apexes and five elements-style firewall top. Building materials are mainly earth and wood, consisting of entirely rammed earth walls, stone wall foundations, and stone-built fortress gates. The fortified‑style layout and fortress architecture directly respond to the survival needs of clan‑based communal living and self‑defense in an immigrant society, while the enclosed‑square courtyard embodies the spatial order of clan aggregation. Linguistically, the region is primarily Central Min dialect. Due to the predominantly immigrant composition of the population and the influence of migrant culture, the unstable social environment inclined the ancestors toward conservative and stability‑seeking cultural tendencies. They mostly lived in clan clusters, forming a unique defensive landscape—the earth fortresses—while also fostering dynamic folk activities with dramatic intensity, such as Anzhen Jing Drum, Bench Dragon Dance, and Fire Walking43. The driving force for the preservation of villages in this region stems from the dual reinforcement of defensive architecture and the clan system. The earth fortress is not merely a physical defensive structure but also an externalization of the immigrant society’s psychology of “clustering for self‑protection.” This defense mechanism, integrating internal and external elements, endowed the villages with the capacity to resist external risks. Consequently, the villages were preserved during turbulent periods, and in peacetime, the cohesiveness of the clan system has maintained their spatial integrity. Jiulong River Min–Hakka Cultural Landscape Zone Zone Ⅷ is the Jiulong River Min–Hakka Cultural Landscape Zone. The region is dominated by hilly and mountainous landforms, with topographic elevations mostly ranging from 200 to 1000 m (some traditional villages have elevations exceeding 1000 m) and slopes mostly greater than 5°. The numbers of villages located on sunny slopes and shady slopes are roughly equal. Influenced by the terrain, village layouts are predominantly dispersed‑style, while some villages adopt fortified‑style layouts. Street layouts are mainly grid‑style or flexible‑style, and courtyard layouts are mostly compound‑style. Building forms are dominated by Tang Heng houses, with roof forms mainly including horse-head-shaped firewall top and five elements-shaped firewall top. Building materials are mainly blue tiles and wood, consisting of blue-tiled walls, sanhe soil wall foundations, and gate-tower style gates. The dispersed layout adapts to the scattered distribution of constructible land in mountainous areas, while the Tang Heng houses and compound courtyard respond to the Hakka people’s deep‑seated needs for clan‑based communal living and ancestral worship. The “nine halls and eighteen wells” type residence, evolved from the courtyard house, is a concentrated expression of these needs. Linguistically, the region is primarily the Min‑Hakka dialect. Deeply influenced by Hakka culture, it has fostered folk customs such as Lantern Carrying, and Giant Dragon Procession44. The preservation of villages in this region is rooted in the deep structure of Hakka culture. After long‑distance migration, the Hakka people’s concepts of clan and family were strengthened and consolidated through struggles with nature and society, forming an extremely strong cultural cohesiveness. As most villages are located deep in the mountains with limited interaction with the outside world, they have retained northern folk customs, attire, and culture over the passage of time, earning the reputation as “living fossils” for the study of ancient Central Plains culture. This preservation mechanism, centered on cultural identity, enables the villages to maintain the continuity of traditional space and rituals through clan power, even under relatively economically disadvantaged conditions. Discussion and conclusion Discussion Indicator selection and its localized adaptation. Indicator selection and its localized adaptation. Indicator selection and its localized adaptation. In the traditional village landscape classification system, the identification of classification indicators is the core foundation for cluster analysis and zoning. Differences in indicator selection and criteria for determining identification results may lead to divergent clustering and zoning outcomes. For example, Xia, H. et al. argued that topography and building decoration are priority landscape factors in traditional settlement zoning45; Liu, Q. and Shang, B. emphasized natural environmental factors such as topography and climate as the basis for regional division46; and Jiang, B. et al.47 highlighted morphological cognition, structural order, transportation layout, and boundary shaping as possessing significant explanatory power in multi‑ethnic plateau settlements. To date, no unified set of identification indicators has been established, and most studies define indicator categories and hierarchical levels according to their specific research needs. Amid the global research trend of cultural landscape conservation and feature identification, the tension between a “universal framework and local adaptability” remains a core concern. Whether the zoning operational tools developed by UNESCO48 or the local landscape character classification systems proposed by European scholars such as Vallés-Planells, M49., the direct application of these paradigms to regions with complex terrain and diverse cultures often overlooks local geographical context and cultural specificity. This tension is particularly evident when our approach is contrasted with existing basin‑scale studies. In the iuzhou River - Yunhe River Basin in Zhejiang Province, Zhang, L. et al.50 relied primarily on GIS spatial analysis and morphological typologies, foregrounding settlement patch characteristics while giving less weight to sociocultural indicators such as language, ritual practices, or ethnic identity. In the broader Xiangxi Region, where water networks and flat terrain predominate, scholars have tended to prioritize hydrological connectivity and economic‑functional indicators over topographically sensitive variables, reflecting a “plain‑logic” adapted to low‑relief landscapes12. Internationally, studies along the Mississippi River corridor have emphasized land‑use and ecological landscape metrics, while research on Rhine River settlements has been framed through historic landscape characterisation (HLC) systems that assume relatively stable administrative boundaries and long‑term land‑parcel continuity5. These frameworks, whether domestic or international, have been developed in basins where either topographic simplicity, administrative homogeneity, or a single dominant cultural tradition allows more streamlined indicator sets. The Min River Basin is defined by simultaneous extremes: “numerous mountains” that fragment settlement space, “dense river systems” that simultaneously connect and divide, and “cultural diversity” that layers Min, Hakka, Gan, and She traditions within a single watershed. Our four‑level framework (geographical environment, morphological layout, traditional buildings, social culture) is therefore not merely a localised checklist but a deliberate structural response to this “compressed diversity.” Unlike single‑axis systems, our design treats geographical, morphological, architectural, and sociocultural variables as equally constitutive, ensuring that river‑system connectivity, dialect boundaries, and architectural typologies all enter the clustering process on a comparable footing. This design, while rooted in the Min River Basin, offers a transferable principle for other mountainous, multi‑ethnic watersheds—such as the Upper Mekong or the Salween River Basin in Southeast Asia—where similar geographical and cultural compression renders single‑dimension frameworks inadequate. Our study adopts “localized adaptation of international experience” as its guiding principle. We did not directly apply existing general standards; instead, we reviewed regionalization experiences from multiple disciplinary backgrounds and then adjusted identification factors and indicators according to the distinct characteristics of the Min River Basin—namely “numerous mountains, dense river systems, and cultural diversity.” This approach aligns with Jiang, B. et al.51. When processing identification results, we further selected the most typical features in cases where multiple landscape characteristics coexist within a single village, thereby improving accuracy and achieving greater landscape gene recognition than traditional generalized identification methods52. (2)Cluster analysis and its methodological considerations. (2)Cluster analysis and its methodological considerations. Cluster analysis and its methodological considerations. Cluster analysis has effectively advanced traditional village zoning from qualitative description to quantitative research. For instance, Huang, Y., et al.53 applied space syntax and GIS to analyze the morphological evolution of the ancient city of Hangzhou. In recent years, the K‑modes method advocated by Fan, Y54. has been widely employed for large‑volume classification of traditional rural settlements. However, K‑modes clustering results are sensitive to data characteristics and the number of clusters, directly affecting zoning outcomes. To mitigate this, we tested different n_init values using the elbow method to determine the optimal cluster number. It should be noted that although K‑modes can generate objective clustering results, it demands high data quality and may yield mechanically rigid groupings. Therefore, we further coupled quantitative K‑modes analysis with qualitative judgment to adjust and correct the clustering outcomes, thereby delimiting the final zone boundaries. (3)Natural environment as a primary shaping mechanism. (3)Natural environment as a primary shaping mechanism. Natural environment as a primary shaping mechanism. The natural geographical environment functions as the primary shaping mechanism for the cultural landscape regionalization of traditional villages in the Min River Basin, echoing the findings of Huang, Y., et al.53 in the Nanxi River Basin. First, terrain and village layout are closely coupled. In the mountainous upper reaches, villages predominantly assume compact, contour-adapted layouts in valley and foothill locations, whereas in the broader downstream plains, settlements tend toward more dispersed, river‑parallel arrangements. This terrain–layout relationship constitutes a fundamental axis of differentiation across the eight landscape zones. Second, cultural context and architectural form are intimately linked. The river network not only provided the geographical basis for the linear distribution of settlements but also enabled cross‑regional cultural transmission through navigability, giving rise to a diffusion gradient of “mainstream propagation and tributary variation” in architectural techniques, dialects, and ritual practices27,30. For instance, the distribution of earth forts and fortified dwellings is strongly correlated with Hakka cultural influence zones, while building ornamentation styles and spatial organization patterns shift along dialect boundaries, reflecting the intertwined transmission of language and architectural know‑how. Third, mountain barriers filter cultural traits and reinforce spatial clustering of specific landscape features. The Wuyi Mountains, as a geographical barrier, fostered an enclosed environment for the independent development of ancient Min culture and shaped an “edge preservation” mechanism that retains more ancient Chinese features in the Northern Min dialect area8,55. This linguistic conservatism co‑varies with the preservation of archaic architectural typologies in the same zone. Meanwhile, the Jufeng and Daiyun Mountains acted as historical barriers between the upper and lower Min River, and their steep terrain hindered large‑scale modern construction, objectively facilitating “passive protection” and the persistence of “cultural islands”56. Defensive structures are highly coupled with mountainous terrain, reflecting an immigrant society’s historical trajectory of “migration–settlement–defense” and displaying a spatial logic akin to the synergistic defense model proposed by Jiang, B., et al.51. In sum, the enduring landscapes of the basin—river, mountain, and topography—are not static backdrops but have continuously shaped settlement layout, dialect differentiation, architectural typology, and cultural transmission. These indicator‑level correlations, between terrain and layout, and between cultural context and architectural form, serve both as boundary conditions for cultural evolution and as crucial anchors for historical interpretation. (4)Recommendations for practical application. (4)Recommendations for practical application. Recommendations for practical application. The identified cultural landscape zones can directly inform practice. In spatial planning, they provide a scientific basis for defining the developmental orientation and land‑use control rules of different village clusters, helping to avoid the “one‑size‑fits‑all” approach that leads to landscape homogenization. For heritage protection, the zoning scheme enables the design of “cultural landscape corridors” along major river systems, facilitating the contiguous conservation of shared architectural heritage, dialect culture, and ritual traditions. In tourism development, each zone’s distinctive cultural landscape identity—such as the Min–Gan transitional character of the Futun River area or the Min–Hakka blended culture of the Jiulong River area—can be translated into thematic tourism routes and interpretive narratives, fostering differentiated and sustainable rural tourism while strengthening local cultural continuity. Limitations It must be acknowledged that the challenges encountered in landscape gene identification at such a basin scale are not specific to this study, but reflect a shared difficulty in the global academic community52. From traditional villages in the Nile River Basin to vernacular settlements along the Mississippi River, large-scale regional identification consistently faces biases arising from wide scope, high sample heterogeneity, and complex cultural stratification4,36. In striving to integrate international experience with the local realities of the Min River Basin, this study has pursued objectivity and comprehensiveness. Nevertheless, several limitations remain. First, regarding data coverage, the analysis is based on 520 officially recognized traditional villages. Although these represent the core protected units of the Min River Basin, certain undocumented but culturally significant vernacular settlements may have been excluded, which could affect the completeness of the zoning picture. Second, concerning indicator selection, even with the localized adaptation of landscape gene indicators, the determination of indicator categories, hierarchical assignments, and the selection of the “most typical” feature among coexisting characteristics inevitably involves a degree of subjectivity. This may lead to the inaccuracy of certain identification factors or produce debatable results, particularly given the basin’s pronounced internal cultural diversity. Third, with respect to historical source reliability, cultural background information and language distribution were derived from provincial gazetteers and foundational cultural geography works33,34. These reflect administrative records and broad-brush regional syntheses rather than fine-grained field surveys; for frontier and transitional zones, the cultural labels may thus represent generalized classifications that require refinement through future ethnographic fieldwork. These limitations highlight the continuing tension between universal frameworks and local specificity in cultural landscape regionalization. At the same time, the experience gained in this study offers a localized, practical reference for improving heritage identification methods in large international river basins, and the relevant methods merit further in-depth refinement in future work. Conclusion Taking the Min River Basin in China as a case study, this research identified the cultural landscape characteristics of traditional rural settlements, conducted cluster analysis on the identification results, and delineated distinct cultural landscape zones with an exploration of their landscape characteristics. The theoretical contributions of this study are twofold. First, it demonstrates that the systematic identification of cultural landscape features can bridge the analytical gap from individual village morphology to macroscale regional patterns. By constructing a localized landscape gene indicator system and following the logical chain of “individual analysis—group induction—regional exploration,” the study reveals the multi-dimensional landscape structure of traditional villages in a complex river‑basin setting, thereby enriching the theoretical framework of cultural landscape regionalization. Second, by integrating K‑modes quantitative clustering with qualitative correction, the study refines the methodological toolkit for zone-based settlement classification, offering a replicable pathway for other regions where categorical landscape data dominate and where mechanical clustering alone tends to blur culturally transitional boundaries. The main findings are as follows: (1) Cultural landscape feature identification serves as a robust means for analyzing the characteristics of traditional rural settlements. In this study, the customized indicator system successfully supported the analysis of individual landscape traits of 520 traditional villages and, when combined with clustering and qualitative reasoning, enabled the extraction of group‑level patterns. This “individual‑to‑group” path provides a concrete analytical template for traditional village regionalization in other regions. (2) The K‑modes clustering method, applied to categorical attribute data, functions as the core quantitative engine in traditional settlement zoning. In the Min River Basin, following the workflow of “landscape feature identification—data processing—K‑modes clustering—qualitative adjustment—zoning,” the method established the principal typology and spatial template of traditional village zones. This methodology can be transferred to other basins and countries undertaking similar settlement‑scale classifications. (3) Traditional villages in the Min River Basin are delineated into eight cultural landscape zones: Jian River Northern Fujian Cultural Landscape Zone, Futun River Min–Gan Cultural Landscape Zone, Min River Middle Stream Eastern Min Cultural Landscape Zone, Min River Downstream Eastern Min Cultural Landscape Zone, Dazhang River Eastern Min Cultural Landscape Zone, You River Central Min Transitional Cultural Landscape Zone, Sha River Central Min Cultural Landscape Zone, and Jiulong River Min–Hakka Cultural Landscape Zone. The zonation exhibits clear “within‑zone convergence and between‑zone divergence” across dimensions of environment, layout, architecture, and culture. The spatial differentiation patterns uncovered provide a scientific basis for tailoring protection and development strategies to each zone. Future research could proceed in several directions. First, the landscape gene indicator system, currently comprising 13 indicators, could be expanded to include finer‑grained intangible cultural traits—such as festival calendars, oral traditions, and craft techniques—so as to capture a more complete spectrum of cultural landscape values. Second, cross‑basin comparative studies, for instance between the Min River Basin and the Mekong River Basin or the Danube River Basin, could test the transferability and robustness of the identification–clustering–zoning pathway and reveal universal versus basin‑specific mechanisms of cultural landscape formation. Third, enhancing the quantitative rigor of boundary delineation, possibly by incorporating machine‑learning‑based sensitivity analyses or participatory community mapping, could further reduce the subjectivity inherent in qualitative correction steps and increase the credibility of zone boundaries for policy adoption.