Landscape Ecology Analysis of Traditional Villages: A Case Study of Ganjiang River Basin

Abstract

Traditional villages, rich in historical and cultural value, hold a high level of preservation value. In the process of urbanization, traditional villages face the crisis of decline, making it difficult to perpetuate the carried cultural heritage. The Ganjiang River Basin hosts numerous traditional villages with rich research value, making the study of their preservation and development in this region a significant topic. This paper, from the perspective of landscape ecology, employs geographic detectors to analyze the driving factors behind the emergence of traditional villages in the Ganjiang River Basin, summarizing the spatial distribution characteristics of traditional villages. A classification method based on village landscape features is adopted to categorize traditional villages in the Ganjiang River Basin, providing a reference for planning the preservation and development of traditional villages. The research results show that plain areas are more suitable for the continuation of traditional villages; a single suitable environmental element cannot provide an environment conducive to the development of traditional villages, which is the result of the combined effect of multiple suitable elements; the study has divided traditional village landscapes into nine types, with clear distribution differences among different types of villages; for different regions and types of traditional villages, it is necessary to balance development and protection tendencies and plan differently according to environmental characteristics.

1. Introduction

Since the 21st century, the urbanization rate in China has been increasing year by year. The rural population in China has been continuously migrating from villages to towns, the development vitality of Chinese villages has significantly declined, the rural ecological environment has been damaged, and the number of villages has decreased sharply [1,2]. Traditional villages, also known as ancient villages, are a cultural epitome of China’s small-scale agricultural society, possessing high research and conservation value. Accompanied by the continuous decline of rural society in China, traditional villages are also experiencing a crisis of survival. If this phenomenon is left unattended, it would be a significant blow to the national cultural revival, cultural confidence, and cultural identity [3,4]. Therefore, how to construct a scientifically effective management scheme to ensure the sustainable development of traditional villages is a focal point of concern for relevant research scholars [5].

The initial research on villages was reflected in the studies within the villages, including landscape genes [6,7,8,9], intangible culture [10], architectural heritage [11,12], tourism resources [13], protection and development [14,15], etc. This type of research mainly focuses on the cultural value and environmental characteristics of villages, with the research methods being mostly qualitative research and descriptive analysis. With the development of geographic information technology, research on villages began to focus on spatial structure [16,17], village environment [18], climatic characteristics [19], village site selection [20], distribution features [21,22,23,24], landscape patterns [25], and so on. The subjects and objectives of this type of research are relatively singular, failing to integrate macro and micro perspectives. In reality, protective planning for traditional villages mainly focuses on the preservation of architecture and cultural heritage, overlooking the importance of the surrounding environment and the morphology of villages [26]. This paper analyzes the driving factors and distribution laws of traditional villages from a macro perspective, constructs a method for classifying traditional village types from a micro perspective, and further reveals the impact of the ecological environment on traditional villages.

The landscape pattern of traditional villages is an important manifestation of village culture, history, and ecological characteristics [27,28]. In the development process of villages, the production and living styles of villagers, as well as policies and other related social factors, affect the landscape pattern around the villages [29,30], promoting the formation of village landscape patterns [31]. Analyzing the landscape pattern of traditional villages can reveal the spatial structure of the villages, reflecting the interconnection between the environment and people [32,33]. From a macro perspective, villages in different regions will present different landscape patterns due to spatial heterogeneity, a phenomenon particularly evident in areas with significant topographic variations [34]. Therefore, by integrating the local geographical environmental features, the impact of landscape on village development can be better demonstrated. Meanwhile, changes in landscape patterns affect biodiversity, soil and water conservation, and other ecological functions, which are crucial factors for preserving the integrity of ecosystems [35,36] and are the focal points in the protection of traditional rural settlements [37].

Village development relies on agricultural production, and due to the demand for irrigation, settlement migration has a strong association with the course of rivers; hence, villages in the same basin have a certain similarity in culture. The Ganjiang River Basin is rich in traditional village cultural resources, encompassing a total of 161 villages from the first to fifth batches of the List of Chinese Traditional Villages [20]. This paper takes the Ganjiang River Basin as the research area, using geographic detectors, correlation analysis, and kernel density analysis to analyze the spatial characteristics of traditional villages and the surrounding natural environment, exploring the driving factors of traditional village continuity. Simultaneously, based on land use data and topographic data, hierarchical clustering analysis is utilized for landscape type division of traditional villages, with development recommendations proposed. The research results can provide references for the protection and development of traditional villages.

2. Materials and Methods

2.1. Methodology

This study can be divided into 3 steps. Step 1 is ‘Build database’, which includes the construction of basic data, extraction of research areas, and data preprocessing. This step is the foundation of all the research in this paper, corresponding to Section 2.2 and Section 2.3 of the research method. Step 2 is ‘Geographic information analysis’, which contains the main content of this paper, including the extraction of drivers for the continuation of villages and the classification of traditional village types based on environmental elements. The research methods for this step correspond to Section 2.4, Section 2.5 and Section 2.6. Step 3 is ‘Analysis and Discussion’, based on the results of the previous step, which includes analyzing the influence of gatherings on the distribution of traditional villages, summarizing the distribution characteristics of different types of traditional villages, and proposing suggestions for protection and development, corresponding to Section 4 of this paper (Figure 1).

2.2. Study Area

The Gan River is located in the middle and lower reaches of the Yangtze River Basin, is the river with the widest coverage area in Jiangxi Province, and is an important source of irrigation for the agricultural development of the surrounding areas. The geographical conditions in the research area are complex with a high degree of spatial heterogeneity; the southern part is dominated by hilly and mountainous terrain with a higher elevation, while the central and northern parts are primarily comprised of plains and basins with lower elevations and terrain undulations, suitable for agricultural development. The scope of the Ganjiang River Basin in this study is determined based on a digital elevation model (DEM). By using the hydrologic analysis tool in ArcGIS software version 10.6, vector files in shapefile format are generated, which delineate sub-watersheds with ridgelines as boundaries. Finally, the scope of the basin is determined by comprehensively considering the spatial relationship between the main stream of the Ganjiang River and the sub-watersheds, in conjunction with the boundaries of the county-level administrative divisions of Jiangxi Province (Figure 2) [38,39].

2.3. Data Collection and Source

The data in this paper include digital elevation model data, land use data, and the list of traditional villages. The digital elevation model data are sourced from Earthexplorer—USGS (https://earthexplorer.usgs.gov/, accessed on 1 November 2023.); the land use data are sourced from the Chinese Academy of Sciences Data Center (https://www.casdc.cn/, accessed on 1 December 2022.); the list of traditional villages in China is sourced from the Digital Museum of Chinese Traditional Villages (https://www.dmctv.cn/directories.aspx, accessed on 1 December 2022.).

In this study, digital elevation data and land use data are standardized, unifying the geographic coordinate systems and spatial resolutions of each group of raster data. The geographic coordinate system is set to GCS_WGS_1984, and the spatial resolution is chosen as 30 m with resampling to ensure that the raster patches of each dataset overlap with each other. For the spatial data of traditional villages, each village is imported into ArcGIS as a point feature through spatial coordinate information to form a vector file. Ultimately, a basic database comprising the above contents is established.

2.4. Analysis of Driving Forces behind the Continuation of Traditional Villages

In studying the driving forces behind the continuation of traditional villages, utilizing a geographical detector can help us unveil the interrelationships between different natural factors and the spatial aggregation degree of traditional villages. The geographical detector includes differentiability, factor, and interaction detectors, among others, and is a statistical method to uncover the driving forces behind detecting spatial heterogeneity [40,41]. This step is implemented using the GD and dplyr packages in the R language.

The differentiation and factor detectors spatially overlay the spatial positions of traditional villages with natural element data, and natural elements are designated as multiple areas in the form of discrete datasets. A significance test is performed on the natural elements, where high significance implies that the natural elements have a high explanatory power regarding the distribution of traditional villages.

The differentiation and factor detector spatially overlay the spatial positions of traditional villages with natural element data, and natural elements are designated as multiple areas in the form of discrete datasets. A significance test is performed on the natural elements, where high significance indicates that the natural element has high explanatory power for the distribution of traditional villages. The specific calculation formula is as follows:

$$Q=1-\frac{{\sum }_{h=1}^L{N}_h{\sigma }_h^2}{{N\sigma }^2}=1-\frac{SSW}{SST}$$

(1)

In the formula, Q represents the explanatory power of natural environmental factors on traditional villages. The research area is divided into two layers, representing the independent variable X and the dependent variable Y, respectively. The entire domain of the independent variable X is divided into several sub-regions, represented by h, and L represents the entire domain of the dependent variable Y. N represents the total number of units in the entire domain, and $N_h$ represents the number of units in sub-region h. ${\sigma }^2$ represents the variance of Y in the entire domain, and ${\sigma }_h^2$ represents the variance of Y in sub-region h. SSW and SST represent the total variance of all sub-regions and the regional variance, respectively. The value range of Q is [−1, 1], representing the degree of explanation of X on Y, with a higher absolute value indicating a stronger explanatory degree. The sign (±) indicates the positive or negative correlation between them.

In this paper, the kernel density calculation results of villages are used as the dependent variable Y in the model, representing the degree of spatial aggregation of villages. High-value areas indicate that the region is more conducive to the continuation of traditional villages. Various environmental factors are used as the independent variable X in the model and are added to the computation in turn. This paper numbers the step of implementing the differentiation and factor detector as Experiment 1.

Interaction detectors can be used to identify the interactions among multiple natural elements. By comparing the Q values under the interaction of two natural elements (Q(X1∩X2)) with the Q values of individual natural elements (Q(X1) and Q(X2)), it can be determined whether the interaction between the two groups of natural elements has enhanced or weakened the benefit on the distribution aggregation of traditional villages. This paper numbers the step of implementing the interaction detector as Experiment 2.

2.5. Study on Spatial Distribution Characteristics of Traditional Villages

There is a mutual relationship between the spatial distribution of traditional villages and cities [42], and there might be a partial association between the number of regular villages in a region and the number of traditional villages. This paper has developed a method for studying the distribution characteristics of traditional villages and other spatial elements. The sampling of experimental data is performed by establishing a 1 km grid over the river basin area and selecting grids that contain traditional villages and adjacent grids as the basic data. The relationships among these three groups of data are calculated through correlation analysis. Correlation analysis is our primary statistical method used to explore the relationships between research variables. Correlation analysis is the primary statistical method used to explore the relationships between research variables, quantifying the strength and direction of linear relationships by calculating the Pearson correlation coefficient. The value range of the Pearson correlation coefficient is from −1 to 1, where −1 indicates a perfect negative correlation, 1 indicates a perfect positive correlation, and 0 indicates no linear relationship. The specific calculation formula is as follows:

$$r=\frac{n\left(\sum xy\right)-\left(\sum x\right)\left(\sum y\right)}{\sqrt{\left(n\sum x^2-{\left(\sum x\right)}^2\right)}\left(n\sum y^2-{\left(\sum y\right)}^2\right)}$$

(2)

Herein, x and y are the values of the two variables, respectively, and n is the number of observations.

This paper numbers the step of implementing the correlation analysis as Experiment 3.

Kernel density analysis is conducted on the three research objects by comparing the spatial distribution of traditional villages with the kernel density of the other two objects, and preliminary conclusions are drawn. The kernel density analysis method is a non-parametric way to estimate the probability density function of a random variable. It estimates the probability density function of a continuous random variable through smoothing processing. It does not depend on a pre-determined specific distribution form (such as normal distribution), making it suitable for various types of data distributions [43]. The basic formula is as follows:

$$f\left(x\right)=\frac{1}{n}{\sum }_{i=1}^n{K}_h\left(x-x_i\right)$$

(3)

In the formula, f(x) represents the calculated probability density function; n represents the total number of samples; $x_i$ represents an individual sample; and $K_h$ is the kernel function, defining the weight distribution around $x_i$.

This method utilizes the Kernel Density tool in the ArcGIS platform [44]. By inputting a point element layer, a spatial distribution map of the kernel density of traditional villages in the Gan River Basin is obtained, thereby displaying the spatial features of village distribution within the basin. This paper numbers the step of implementing the kernel density analysis and comparison as Experiment 4.

2.6. Classification of Traditional Village Landscape Types

This paper has constructed a method for classifying traditional village types based on landscape characteristics. It involves data collection on the proportion of land use types (cultivated land, forest land, water bodies, and construction land) within a 1 km buffer zone of all traditional villages in the basin, as well as topographic elements such as elevation, slope, and terrain undulation within a 1 km buffer zone of all traditional villages in the basin, and applied hierarchical cluster analysis to classify the landscape types of traditional villages. Cluster analysis can be used to identify groups of samples with similar predefined variable patterns. Cluster analysis does not distinguish between dependent and independent variables, thus allowing for the grouping of all variables to reveal different patterns within the groups [45,46,47]. The research is based on topographic data and land use type ratio data, using the ggplot2 and ggtree packages in RStudio, and conducting hierarchical cluster analysis on each traditional village in the Gan River Basin, analyzing the similarity in different unit composition structures while reflecting the differences in feature spatial distribution and cluster aggregation characteristics [48]. This paper numbers the step of implementing cluster analysis as Experiment 5.

3. Results

3.1. Extraction of Environmental Elements in Traditional Villages

3.1.1. Distribution Characteristics of River Elements

Rivers are an important water source for agricultural production in villages, and some significant waterways also serve the purpose of water and land transportation. The Euclidean Distance tool in ArcGIS is used to calculate the river distribution data, obtaining river buffer data (Figure 3b). Based on the calculation results, the majority of traditional villages have a relatively close relationship with rivers. A total of 60 villages have a close relationship with rivers, with a straight-line distance of less than 1 km, accounting for 37.28% of the total; 72 villages have a distance of between 1 km and 5 km from rivers, accounting for 44.72% of the total; 29 villages have a distance of over 5 km from rivers, accounting for 18.01% of the total.

3.1.2. Terrain Elements and Distribution Characteristics

Utilizing the buffer tool in ArcGIS to delineate a 1 km radius buffer for the 161 traditional villages in the study area and calculating the average elevation within each village buffer area using statistical tools to represent the overall terrain features surrounding the villages (Figure 3d–f). Within the river basin, there are 87 traditional villages located at a distance of 0–2000 m from rivers, accounting for 54.04% of the total. These villages have a close relationship with the rivers (Figure 4a). The highest average elevation of all villages is 707.26 m, and the lowest is 13.86 m. There are 80 villages with an average elevation below 100 m, accounting for 32.92%; 41 villages with an average elevation between 100 and 200 m, accounting for 26.71%; 36 villages with an average elevation between 200 and 500 m, accounting for 34.78%; and 4 villages with an average elevation above 500 m, accounting for 5.59% (Figure 4b). The highest average slope across all villages is 24.65°, and the lowest is 1.99°. There are 64 villages with an average slope below 5°, accounting for 39.75%; 65 villages with an average slope between 5 and 10°, accounting for 40.37%; 25 villages with an average slope between 10 and 15°, accounting for 15.53%; and 7 villages with a slope above 15°, accounting for 4.35% (Figure 4c). The average undulation for all villages is 8.91°, with the highest value being 13.64° and the lowest value being 2.47°. Among these, 49 villages have an average undulation below 5°, accounting for 30.43%; 89 villages have an average undulation between 5 and 10°, accounting for 55.28%; and 23 villages have an average undulation above 10°, accounting for 14.29% (Figure 4d). The terrain features of traditional villages in the Ganjiang River Basin are closely related to the spatial heterogeneity of the area. Villages located in the central plain region of the basin tend to have lower elevations and gentler terrains, while traditional villages in the southern hilly area have higher average elevations and steeper slopes.

3.1.3. Distribution Characteristics of Land Use Types

Using a 1 km buffer zone around traditional villages as the range, the area proportions of each land use type are calculated to represent the composition structure of different land use types around the villages (Figure 5). The average proportion of cultivated land in all villages is 53.51%; there are 90 villages with a proportion exceeding 50% and 3 villages with a proportion less than 10%. Among the 161 villages, the highest proportion of forest land area is 98.53%, and the lowest is 0%. There are 45 villages with a forest land area proportion greater than 50%; 32 villages have a proportion less than 10%. Hence, the average proportion of construction land in villages is 7.66%; there are 13 villages with a proportion higher than 20% and 122 villages with a proportion less than 10%. Cultivated land and forest land are the land use types with the highest area proportions in traditional villages of the Ganjiang River Basin, while the proportions of construction land and water bodies are lower. Due to the large area of flat cultivated land in the central part of the basin, in most traditional villages in this area, cultivated land has the highest land use type proportion. In the traditional villages mainly characterized by hilly and mountainous terrain in the southern part of the basin, forest land has the highest proportion among land use types, and most villages in this area are concentrated in mountain streams or foothill areas.

3.2. Analysis of Driving Forces behind Traditional Village Distribution

3.2.1. Divergence and Factor Detection

The spatial distribution of traditional villages varies under different natural geographical conditions, but the varying impact strength of different influencing elements needs further research and judgment. This part of the research content corresponds to Experiment 1. This paper utilizes geographical detectors for divergence and factor detection to analyze the driving forces behind traditional village distribution. In the experiment, the average elevation, average slope, average terrain undulation, proportion of cultivated land area surrounding the village, proportion of forest land area, proportion of construction land area, proportion of water area, and river distance were selected as assessment factors X, and choose the core density value of traditional villages as the dependent variable Y, to evaluate the explanatory power of natural environmental factors on the continuation of traditional villages. The extracted influencing factors of traditional villages are imported into R for calculation, and after discretizing the continuous data, the impact of each factor on the distribution of traditional villages is computed.

In differentiation and factor detection, the Q value represents the degree to which factor X explains the dependent variable Y. Four groups of data—average elevation, average terrain undulation, average slope, and cultivated land area—show credibility with Q values of 0.4154, 0.3092, 0.2842, and 0.0718, respectively. However, the four groups of data consisting of forest land area, construction land area, water area, and river distance did not pass the credibility test, with Q values of 0.0041, 0.0034, 0.0012, and <0.0001, respectively, indicating a lack of credibility. Average elevation is the most significant factor affecting the distribution of traditional villages, followed by average terrain undulation, average slope, and cultivated land area (Figure 6).

3.2.2. Interaction Detection

Interaction detection is used to identify the conditions where the explanatory power of factor Y may increase or decrease when different evaluation factors X act on dependent variable Y together; the results obtained through interaction detection are shown in

Table 1. Interaction detection of factors in traditional villages.

Variable	Average Elevation	Average Slope	Average Relief Amplitude	Cultivated Land Area	Forest Land Area	Water Area	Construction Land Area
Average slope	0.4790-Eb *	NA	NA	NA	NA	NA	NA
Average relief Amplitude	0.5031-Eb	0.3621-Eb	NA	NA	NA	NA	NA
Cultivated land area	0.4402-Eb	0.3614-En	0.389-En	NA	NA	NA	NA
Forest land area	0.4943-Eb	0.4045-En	0.4085-En	0.224-En	NA	NA	NA
Water area	0.4815-En *	0.342-En	0.3718-En	0.1197-Eb	0.1647-En	NA	NA
Construction land area	0.5328-Eb	0.4382-En	0.4354-Eb	0.2729-En	0.2488-En	0.179-Eb	NA
River distance	0.5406-En	0.5219-En	0.4952-En	0.184-En	0.3289-En	0.1113-En	0.2767-En

* “-En” indicates nonlinear enhancement when two groups of elements interact; “-Eb” indicates bivariate enhancement when two groups of elements interact; “NA” stands for “No Data”. Bivariate enhancement refers to the interaction of two factors being higher than the effect value of a single factor, while nonlinear enhancement refers to the combined effect of two factors being higher than the sum of the effect values of single factors;

. This part of the research content corresponds to Experiment 2.

The top five interactions among evaluation factors are river distance and average elevation, construction land area and average elevation, river distance and average slope, average terrain undulation and average elevation, and forest land area and average elevation, with Q values being 0.5406, 0.5328, 0.5219, 0.5031, and 0.4943, respectively. The five pieces of data on cultivated land area, forest land area, construction land area, water area, and river distance exhibited weaker explanatory power or did not pass the test in divergence and factor detection while showing stronger explanatory power and all passing the test in interaction detection. The study found that the eight evaluation factors used can generate stronger driving forces under interaction, enhancing the explanatory power for the continuation of traditional villages.

3.3. Spatial Distribution Characteristics of Traditional Villages

3.3.1. Study on Spatial Distribution Correlation of Traditional Villages

This research establishes a 1 km grid covering the study area and selected 501 grids, including those containing traditional villages and their surrounding adjacent grids, as experimental subjects. The study counted the number of traditional villages, the total number of villages, and the proportion of construction land area in each grid. This part of the research content corresponds to Experiment 3. The experiment utilized the ggplot2 and GGally packages in R language software version 4.3.2 as tools for correlation analysis. Based on the aforementioned data, distribution curves for three types of elements were calculated, as well as scatter plots and Pearson indices for each pair of elements. The results indicated that the number of traditional villages within the grid is weakly positively correlated with the number of ordinary villages and weakly negatively correlated with the proportion of construction land area (Figure 7). However, these correlations did not pass the credibility test, implying that the correlation between these two contrasting sets of elements is not significant and lacks a clear mutual relationship. The correlation between ordinary villages and construction land is weakly positive with higher significance, suggesting that a larger area of construction land in a region correlates with a lower number of ordinary villages.

3.3.2. Impact of Ordinary Villages and Construction Land Distribution on Traditional Villages

This section conducts a kernel density analysis of traditional villages in the study area (Figure 8a) and extracts values from the kernel density analysis of traditional villages, ordinary villages, and construction land for comparison. This part of the research corresponds to Experiment 4. The results show that the highest kernel density value is 0.0172, with the Ji’an area in the central part of the basin being the most densely distributed region for traditional villages. To further study the impact of the spatial characteristics of ordinary villages on the distribution of traditional villages, we conducted a kernel density analysis of 134,969 villages in the study area (Figure 8b). The results indicate that the highest kernel density value is 2.10, with the most concentrated village distributions in the northern part of Ganzhou and the western part of Yichun. The terrain in this area is mainly hills and mountains, with villages characterized by numerous, densely distributed, and small-scale settlements. Central Ji’an and southern Nanchang are also areas with dense village distribution. The terrain in these areas is mainly plains and basins, with villages characterized by large buildings and arable land areas. Overall, the distribution characteristics of ordinary villages show significant differences from those of traditional villages.

To analyze the impact of construction land on the distribution of traditional villages, the study conducted kernel density analysis on construction land using 1 km grid squares to calculate the area ratio of construction land in the region, converting grid data with statistical result fields into point features, and using the area ratio as the population value (Figure 8c). Areas with high kernel density indicate a concentrated distribution of construction land, representing large urban areas. The kernel density values of major municipal districts in the basin, such as Nanchang, Ganzhou, Ji’an, and Yichun, range between 0.2 and 0.56, with Nanchang reaching the highest at 0.5642. Using the Extract Values To Points tool to extract kernel density values to traditional villages, the results show that the highest kernel density value in traditional villages is 0.1454, with 98.14% of them having a value lower than 0.1. This indicates that traditional villages are mostly located away from urban areas, and the area ratio of large cities in the region has a certain negative impact on the spatial distribution of traditional villages.

3.4. Determination and Analysis of Landscape Types of Traditional Villages

3.4.1. Characteristics of Environmental Elements in Traditional Villages

Landscape features reflect the surrounding environment of a village, embodying the conditions for production and living in the village. A correlation study on the seven environmental elements used in the research revealed that all 12 sets of correlation indices have high significance. Three sets of topographical data show a strong correlation between cultivated land and forest land. They are negatively correlated with cultivated land with indices of −0.644, 0.648, and 0.509, respectively, and positively correlated with forest land with indices of 0.682, 0.686, and 0.530, respectively (Figure 9). There is also a negative correlation between topography, water areas, and construction land, but the degree of explanation is lower. This is somewhat related to the lower proportion of water areas and construction land in the basin. Based on the above results, it is known that in high-altitude areas, there is less cultivated land, water areas, and construction land but more forest land, whereas in low-altitude areas, the situation is reversed.

3.4.2. Delimitation of Traditional Village Landscape Types and Numerical Ranges of Environmental Elements for Each Type

The evaluation data for classifying traditional village landscape types consist of two items: topographic data and land use data. The topographic data include elevation, slope, and terrain undulation. A 1 km buffer zone centered on the village is set as the research scale, extracting the regional average values of three topographic data items for 161 traditional villages and utilizing these three average topographic data to represent the landscape features surrounding the villages. The extraction of land use data selected the 2020 land use type map of the Ganjiang River Basin as the basic data. The area proportions of cultivated land, forest land, construction land, and water bodies within the 1 km buffer zone of 161 traditional villages are extracted separately, serving as the landscape features representing the land use around the villages. Through cluster analysis, 161 traditional village samples were classified into several distinct clusters, with the data characteristics of samples within each cluster being similar. According to the clustering results of each cluster, the landscape types of traditional villages in the Ganjiang River Basin are divided into nine types: A1, A2, A3, B1, B2, C, D, E, and F (Figure 10).

This part of the research content corresponds to Experiment 5. The land use type proportion numerical features of village types A1, A2, and A3 are similar, with cultivated land occupying more than 50%, forest land proportion between 20 and 50%, and low water body proportion. The proportion of construction land in A2 is between 10 and 20%, while in A1 and A3, it is below 10%. For B1 and B2, the forest land proportion is greater than 50%, the cultivated land proportion is between 20 and 50%, and the water body proportion is low. The proportion of construction land in B2 is higher than in B1. Category C has a higher proportion of construction land and a lesser proportion of water bodies. Category D has the highest forest land proportion, Category E has the highest cultivated land proportion, both have extremely low water body proportion, and Category E has a higher construction land proportion than Category D. The main features of Category F are high water body proportion, high cultivated land proportion, and lower forest land proportion. Category D has the highest average elevation and slope. Categories A2, A3, and B2 belong to a higher class in terms of average elevation, but Category A2 has relatively lower slope values compared to the other two categories. Categories A1, B1, C, E, and F have the lowest average elevation, and the terrain is more gentle in Categories E and F (Figure 11 and Figure 12).

Based on the analysis results of the data characteristics within the aforementioned clusters, nine traditional village landscape types have been delineated: A1 Plain Cultivated Land Mixed Type, A2 Hilly Cultivated Land Large Settlement Mixed Type, A3 Hilly Cultivated Land Small Settlement Mixed Type, B1 Plain Forestland Mixed Type, B2 Hilly Forestland Mixed Type, C Plain Construction Land Type, D Mountain Forestland Type, E Plain Cultivated Land Type, and F Waterfront Type. Details are provided in

Table 2. Overview of the quantity and characteristics of traditional village landscape types.

Type	Code	Quantity	Feature Description
Plain Cultivated Land Mixed Type	A1	11	Gentle terrain, abundant cultivated land landscape, other landscapes occupy a certain proportion.
A2 Hilly Cultivated Land Large Settlement Mixed Type	A2	9	Medium elevation, terrain with certain undulations, large proportion of cultivated land and construction land.
Hilly Cultivated Land Small Settlement Mixed Type	A3	11	Medium elevation, certain undulations in terrain, smaller village size, abundant cultivated land landscape, other landscapes occupy a certain proportion.
Plain Forestland Mixed Type	B1	9	Medium to high elevation, steep terrain, large proportion of forestland, other landscapes occupy a certain proportion.
Hilly Forestland Mixed Type	B2	16	Low elevation, gentle terrain, large proportion of forestland, other landscapes occupy a certain proportion.
Plain Construction Land Type	C	4	Low elevation, gentle terrain, large proportion of construction land, near to cities.
Mountain Forestland Type	D	22	High elevation, steep terrain, large proportion of forestland, low proportion of other landscapes.
Plain Cultivated Land Type	E	69	Low elevation, gentle terrain, large proportion of cultivated land, low proportion of other landscapes.
Waterfront Type	F	10	Low elevation, gentle terrain, large proportion of water bodies, near to large water channels.

.

3.4.3. Spatial Distribution Characteristics of Traditional Village Landscape Types

Traditional villages in different spatial locations exhibit distinct characteristics in terms of topography and land use type area proportions. From the perspectives of land use proportion and topography, the study classified traditional villages into nine categories, each showing clear differences in landscape features. For instance, A1, A2, and A3 type traditional villages all have large areas of cultivated land. A1 and A2 have relatively larger areas of construction land, while A2 and A3 are located at higher altitudes with steeper slopes. In terms of quantity, Ji’an City and Ganzhou City have the most type A villages, with 11 and 10, respectively. Among them, A1 is concentrated in the central Ji’an region of the basin. A2 and A3 are more dispersed without significant clustering. B1 and B2 type traditional villages have more forest land, with B1 being at a higher average altitude and mainly distributed in the southern and northern parts of Ganzhou City, while B2 is at a lower average altitude and mainly in central Ji’an City. Type C traditional villages are characterized by large areas of construction land, are few in number, close to urban spaces, and are mainly concentrated in the central and northern parts of the basin. Type D traditional villages have flat terrain and strong transportation advantages. However, due to their distance from major urban areas, these villages have not been further influenced by modern social production and lifestyle, retaining traditional culture and a large number of historical buildings. This type of village is distributed in central Ji’an City and northern Yichun City. Famous traditional villages in the Ganjiang River Basin, such as Meibei Village, Piaxia Village, and Baihe Village, belong to this type. Type E traditional villages have a large proportion of forest land and are located at higher average altitudes and slopes. These villages are smaller and more vulnerable to external impacts, with some, like Chouxi Village, having been relocated due to natural disasters, making it difficult to continue their traditional culture. Type F traditional villages have a larger proportion of water areas, are flat in terrain, and are distributed around major rivers (Figure 13). These categories each have their unique forms, contributing to the diverse landscape and appearance of traditional Chinese villages.

4. Discussion

4.1. A Favorable Development Environment Makes Traditional Villages More Sustainable, and Areas with a Larger Base of Villages Do Not Give Rise to More Traditional Villages

Villages often tend to conform to mainstream society in their modes of production and living under the impact of social development, losing their original cultural traditions. Traditional villages are the products of the continuation of village cultural material over thousands of years. Identifying the main reasons for the emergence of traditional villages from a landscape ecology perspective is a key research focus of this paper. The average elevation, average slope, and average terrain undulation around the villages reflect the flatness of the land surrounding the villages; flat land is more suitable for agricultural production and village expansion, providing a better transportation environment. The proportion of cultivated land, forest land, construction land, and water area around traditional villages represents the overall landscape environment of the villages; the area of cultivated land, forest land, and water reflects the agricultural production conditions of the villages and the area of construction land reflects the scale and volume of the villages. The distance to river channels reflects the ability of villages for water transportation and irrigation [49,50]. The research results of environmental elements of traditional villages using the geographical detector indicate that the combined effect of land use type area, distance to rivers, and topography has a stronger impact on the continuation of traditional villages. A suitable natural environment and convenient transportation are the foundations of regional development [51]. This result also indirectly indicates that a single element cannot be the driving force for the continuation of traditional villages. The continuation and development of traditional villages must be the result of the combined effects of multiple suitable environmental factors.

The research finds that traditional villages and ordinary villages have distinctly different spatial distribution characteristics. The southern and northwestern parts of the basin are the most concentrated areas of ordinary village distribution, while the distribution of traditional villages is mainly concentrated in the central part of the basin. The central basin’s Jitai Plain has flat terrain, convenient transportation, and sufficient irrigation water sources, making it suitable for agricultural production and village development and more likely to produce larger-sized villages. For a traditional village to continue for a thousand years and still exist today, a larger size and stable social structure are quite important. Our research results indicate that the central part of the basin, with its suitable development environment, has a higher number of traditional villages with a more concentrated distribution. From the perspective of village landscape types, the villages concentrated in the central region are mainly of the Plain Cultivated Land Mixed Type (A1), Plain Forest Land Mixed Type (B1), Plain Construction Land Type (C), and Plain Cultivated Land Type (D). These types of villages have larger cultivated land areas compared to other types, and the surrounding terrain is relatively flat. These favorable development environments are important factors for the continuation of traditional villages.

In the southern and northwestern parts of the basin, due to large terrain undulations, land that is difficult to develop, and inconvenient transportation, there is a lack of suitable development conditions. A small area of land suitable for development at high altitudes often has many villages, which, due to limited development space and inconvenient transportation between them, present a dense spatial distribution. The distribution of traditional villages in these areas is more dispersed, resulting in the starkly different spatial distribution characteristics of the two types of villages in the basin. Therefore, it is evident that suitable development space is more conducive to the formation of traditional villages, while a larger base of ordinary villages in the region does not significantly promote the emergence of traditional villages.

4.2. Developing Protection Plans Based on Village Landscape Characteristics Is a Means for Sustainable Development

Traditional villages, as heritage with cultural attributes, have inherent vulnerability. Vulnerability manifests in the damage to structures caused by natural disasters and social activities. The vulnerability is evident in the damage inflicted on structures by natural disasters and social activities. The phenomenon of rural hollowing out caused by urbanization also significantly contributes to the overall decline and demise of villages [52]. The dual vulnerability and the irreversibility of traditional villages highlight the importance of protecting traditional villages. In the past few decades, China has witnessed a large number of abandoned villages. This phenomenon has triggered many socio-economic and environmental issues, such as abandoned cultivated land, stagnation or decline of rural development, disappearance of rural settlements, secondary vegetation succession, and landscape changes [53]. With the increase in urbanization and industrialization, the continuation of traditional villages and the inheritance of historical and cultural heritage are at risk. Landscape patterns are regarded as the texture and structure of traditional villages. Accurate detection and analysis of landscape patterns play a key role in understanding the socio-cultural environment and the relationship between humans and nature. This paper categorizes traditional villages through a generic classification method. The formulation of village protection measures should be combined with its own characteristics, devising protection and development strategies that align with its objective environment. Villages with suitable development environments, such as types A1, A2, B1, C, E, and F, should primarily focus on development and utilization, taking into full consideration various conditions like objective transportation conditions, village volume, and the status of village remnants [54]. Whereas for villages of types A3, B2, and D, due to limited developmental space, protection should be emphasized in planning to avoid excessive interference from human activities.

4.3. Uncertainty and Limitations

This paper has achieved some research results in the study of the driving forces behind the continuation of traditional villages, the classification of traditional village landscape types, and their distribution characteristics. However, there are still many uncertainties in the research process and results.

Firstly, in the study of driving forces using the geographical detector, the experiment chose the kernel density value of traditional villages as the Y value. This led to two experimental phenomena: one is that traditional villages in the central Jitai Basin area of the basin are densely distributed around Ji’an City, resulting in some areas with dense construction land and almost no traditional villages having higher than expected kernel density values; the second is that some traditional villages scattered along rivers have too low a kernel density value, failing to adequately reflect the importance of this river distance element for the continuation of these villages. These phenomena are due to the limitations of the experimental model and the scale effect of the research. Fortunately, in the interaction detection, these environmental elements obtained higher Q values, with results aligning with our expectations. This research suggests that interaction detection is more suitable for studying the driving forces of traditional villages. In future research, it is hoped that more research models will be compared and used to summarize more appropriate methods and approaches for the study of traditional villages.

Secondly, in the process of classifying traditional village landscapes, the environmental elements used and the data thresholds for classification types are also content worthy of further study. Among the elements, the area of cultivated land and forest land are important bases for the classification of landscape types in this study. However, judging solely from land use types ignores the complexity of vegetation-covered land. Including the NDVI index in the study or incorporating elements such as canopy density, forest types, and types of crops in cultivated land can make our classification more detailed and accurate. At the same time, traditional villages from different cultural backgrounds often have significant differences in village site selection and production and living styles. This paper’s research is based on geographical features and does not consider the cultural factors of traditional villages. The Ganjiang River Basin has a rich variety of traditional village cultures, with the most representative being the Lu-ling culture in the central part and the Hakka culture in the south. Future research will start from the perspective of landscape genes, classifying traditional villages by cultural characteristics and analyzing traditional village architectural styles and internal spatial patterns at a microscopic level to promote the inheritance and protection of culture.

5. Conclusions

In order to better protect the traditional villages in the Ganjiang River Basin, this study, based on ArcGIS software and the geographical detector model, analyzed the driving forces behind the continuation of traditional villages, the spatial relationship between traditional villages, ordinary villages, and construction land, and classified traditional villages into landscape types based on various environmental elements, summarizing the spatial distribution characteristics of each type.

The research results show that the spatial distribution of traditional villages is clustered, with clear distribution differences in different areas. Ordinary villages in high-altitude areas are more densely distributed and smaller in scale, while villages in plain areas are larger in scale and more likely to give rise to traditional villages. The interaction of factors such as altitude, rivers, slope, and topographical undulation around the village is the main driving force behind the formation of traditional villages. A single factor is unlikely to be the determining factor for the emergence of traditional villages; rather, the combined effect of multiple suitable elements is key to the continuation of traditional villages. Based on land use distribution and topographical characteristics, this paper divides traditional villages in the Ganjiang River Basin into nine types. Cultivated land and forest land are the most important landscape elements of traditional villages, and the distribution characteristics of different types of villages are distinctly different. For different areas and types of traditional villages in the basin, it is necessary to balance the tendencies of development and protection and propose different development measures according to local conditions.