Spatial Distribution and Influencing Factors of Leisure Agriculture Resources in Southern Jiangsu Region Based on Multi-Source Data

Abstract

Leisure agriculture has become an essential driver of rural revitalization in China, yet most existing studies focus on provincial or municipal levels and rely on single-method approaches, leaving a gap in understanding spatial distribution patterns and driving mechanisms in highly urbanized metropolitan regions. This study addresses this gap by constructing a comprehensive leisure agriculture database for southern Jiangsu using multi-source data, including POI (Point of Interest), statistical yearbooks, and GIS datasets. Kernel density estimation, nearest neighbor index (NNI), geographic concentration index (GCI), and ordinary least squares (OLS) regression with VIF testing were applied to analyze spatial clustering and influencing factors. Results reveal that leisure agriculture resources exhibit significant clustering with a clear “core–periphery” pattern, concentrated in urban–rural transition zones. Agricultural output value and the number of A-level scenic spots significantly promote clustering, whereas GDP, population, and transportation density show weaker explanatory power. Theoretically, this study integrates multiple spatial statistical methods into a comprehensive analytical framework, enriching the understanding of leisure agriculture evolution under metropolitanization. Practically, it provides empirical evidence to support the optimization of leisure agriculture resource allocation, inform rural revitalization policies, and guide coordinated urban–rural planning in developed regions.

1. Introduction

Since 2000, China’s fast economic development has been matched by an increase in the demand for green ecological areas, cultural experiences, and healthy leisure activities among both urban and rural populations. As a result, leisure agriculture has grown significantly and is now a key factor in industry diversification, farmer income growth, and rural economic change.

As a novel agricultural production model, leisure agriculture combines agricultural production conditions with corresponding landscape resources to develop recreational, sightseeing, and tourism activities. It has become a crucial pathway for the integration of rural tertiary industry and farmers’ income enhancement under China’s current Rural Revitalization Strategy [1]. The suburban leisure agriculture model in the Yangtze River Delta region, the culture-integrated agritourism in the Pearl River Delta, and the ecological wellness agricultural projects in Sichuan’s hilly areas in practice all exemplify the typical “agri-tourism integration + regional characteristics” development model and facilitate the extension of rural industries from primary agriculture to comprehensive service sectors [2,3].

The conceptual definitions, development models, evolution of spatial patterns, and influencing factors of leisure agriculture have all been thoroughly studied by domestic scholars from a variety of angles. Agritourism upgrading, agricultural-cultural-tourism integration, and rural service industry extension are the main industrialization pathways in China, according to Ren and Dong’s thorough review of leisure agriculture [1]. In addition, some foreign scholars have examined the symbolic analysis of rurality in the British countryside [4], while others have investigated the development of tourist accommodation enterprises on farms in England and Wales to illustrate the rural tourism economy [5]. Studies on tourist motivation and satisfaction also provide important insights into destination loyalty and its implications for leisure agriculture and tourism [6]. For example, Ren et al. achieved spatial optimization of leisure agriculture resources by analyzing the spatial distribution differences and influencing factors of leisure agriculture resources in Shandong Province [7]. Luo et al. carried out a methodical examination of the spatial distribution of leisure agriculture resources in Jiangsu Province based on multi-source data [8]. Li et al. took 213 leisure agriculture and rural tourism demonstration sites in Guangxi as research objects to systematically analyze the spatial distribution of leisure agriculture and rural tourism destinations in Guangxi and the determinants of their spatial distribution by employing mathematical statistical analysis and Geographic Information System (GIS) spatial analysis [9]. Other researchers, like Chang et al. [10], studied Chengdu’s spatial patterns empirically and categorized leisure agriculture resources from the perspective of agricultural resource ontology. However, Hu analyzed various factors that impacted Shanghai’s leisure agriculture development and calculated the concentration and abundance levels in different regions, which provided zoning-based spatial management strategies for municipal leisure agriculture planning [11].

The aforementioned research primarily concentrates on leisure agriculture resources within discrete administrative boundaries at provincial or municipal levels [12]. However, municipal- or economic zone-level studies are usually constrained by administrative boundaries, focusing on resource endowments, land use, and policy planning within specific jurisdictions. By contrast, metropolitan studies emphasize cross-regional dynamics such as population agglomeration, consumer spillover, and urban–rural functional restructuring [13,14]. These differences arise from variations in research scale and driving mechanisms, which distinguish metropolitan contexts from municipal ones. Nevertheless, there is still little research on the leisure agriculture resources in economic zones or metropolitan areas [15]. Consequently, this study utilized multi-source data to further explore and analyze the spatial distribution characteristics and influencing factors of leisure agriculture resources within the region, concentrating on the southern Jiangsu region. The research focuses on the spatial distribution characteristics and optimized utilization of leisure agriculture resources, exploring innovative paths for achieving coordinated development of rural tourism and agricultural tourism in highly urbanized areas. It seeks to promote their leisure agriculture resources towards rational spatial allocation and multifunctional and high-quality development. The findings seek to optimize the allocation of leisure agriculture resources in southern Jiangsu, formulate rural development policies tailored to local conditions, and eventually accomplish rural revitalization and integrated urban-rural development. It also aims to provide a scientific reference for leisure agriculture in similar optimization economic zones [7].

2. Materials and Methods

2.1. Study Area

The southern Jiangsu region is located south of the Yangtze River in Jiangsu Province, consisting of five prefecture-level cities: Nanjing, Wuxi, Suzhou, Changzhou, and Zhenjiang. It is a typical region with leading economic development and urban-rural integration in China (Figure 1). The region enjoys superior geographical positioning since it is situated in the core of the Yangtze River Delta urban agglomeration. The study area presents a clear stepped terrain feature, gradually sloping from west to east. With the Taihu Lake Plain as the core in the east, the dense water networks have developed agriculture, while the western area is dominated by hilly topography with a mild and humid climate suitable for developing diversified agriculture and recreational tourism. A wealth of agricultural and tourism resource foundations has been fostered by this varied topography and biological environment, creating favorable conditions for the clustered development of leisure agriculture [16,17].

The Southern Jiangsu region is economically developed, with China’s most successful industrial parks and advanced manufacturing capabilities in integrated circuits, pharmaceuticals, and other industries. Meanwhile, the region serves as an important tourism hub and distribution center for Jiangsu Province and the country because it is well-known for its beautiful environment and flourishing tourism industry. It is a leading demonstration zone for Jiangsu Province to implement high-quality development, which is ahead of other areas nationwide. With its strong manufacturing foundation, comprehensive transportation networks, and comparatively high income, the southern Jiangsu region exhibits strong local consumption potential and tourism reception capacity [18]. Southern Jiangsu has steadily developed a variety of leisure agriculture development models, including “suburban metropolitan,” “agri-tourism integration,” and “ecological wellness” types, in tandem with the growth of rural tourism and the restoration of rural spatial functions in recent years, making it an important demonstration area for promoting agricultural industry integration and the Rural Revitalization Strategy in Jiangsu Province and nationwide [19].

Specifically, leisure agriculture in southern Jiangsu can be categorized into three major types. The first is the suburban metropolitan type, represented by farm stays, picking gardens, and agritourism parks in areas such as Jiangning (Nanjing) and Wuzhong (Suzhou), which rely on strong metropolitan consumer markets and meet the short-distance leisure needs of urban residents. The second is the agri-tourism integration type, found in places like Yixing (Wuxi) and Wujin (Changzhou), where traditional agricultural production is combined with sightseeing, education, and recreational experiences to form comprehensive “agriculture + tourism” models. The third is the ecological wellness type, mainly distributed along the Taihu Lake rim and hilly areas, focusing on ecological landscapes, vacation resorts, and health-related functions, reflecting a trend toward upscale and multifunctional development. These diversified types not only reveal the distinctive characteristics of leisure agriculture under different locational conditions but also highlight the varied development pathways in the metropolitan context of southern Jiangsu.

The study area is selected based on these characteristics as the research subject, which facilitates the identification of spatial distribution patterns and influencing factors of leisure agriculture resources within developed metropolitan areas. In addition to offering practical experience and reference for other similar regions, this approach deepens theoretical understanding of regional rural transformation pathways.

2.2. Data Source

The following variables and data used in this investigation are listed in

Table 1. Data categories, types, and sources used in this study.

Data Categories	The Required Data	Source
Socio-economic Factors	Population Data	Jiangsu Statistical Yearbook (http://tj.jiangsu.gov.cn/) (accessed on 22 March 2025)
	Economic Data
	Density of Expressway Network	National Geomatics Center of China (http://www.ngcc.cn/) (accessed on 18 April 2025)
	Tourist Spot Data	JiangSu Culture and Tourism Department (http://wlt.jiangsu.gov.cn/) (accessed on 5 April 2025)
Natural Factors	DEM (30m)	Geospatial Data Cloud (https://www.gscloud.cn/) (accessed on 20 April 2024)
Leisure Agriculture Points	25 April 2024	Gaode Map (https://www.amap.com/) (accessed on 25 April 2024)
Base map data	8 March 2024	National Geomatics Center of China (http://www.ngcc.cn/) (accessed on 8 March 2024)

. Two main factors are taken into account when choosing the variables:

The selection of explanatory variables is grounded in a “resources–location–market–environment” framework, which has been widely applied in leisure agriculture and rural tourism studies [20]. Specifically, agricultural, forestry, animal husbandry, and fishery output value reflects the resource endowment, providing the material foundation for leisure agriculture development. GDP and population represent market demand, capturing consumer capacity and potential. The number of A-level scenic spots characterizes tourism attractiveness, highlighting the synergy between tourism and leisure agriculture. Transportation density indicates locational accessibility, which strongly influences tourist mobility and clustering patterns. Average elevation is included to account for natural environmental conditions, as topography and water resources may affect the suitability and distribution of leisure agriculture. By incorporating these variables, this study builds a comprehensive framework to analyze the multidimensional driving mechanisms of leisure agriculture spatial distribution [21].

2.3. Methods

In this study, we employed kernel density estimation (KDE), nearest neighbor index (NNI), and geographic concentration index (GCI) in combination. KDE is suitable for identifying local hotspots and spatial clustering patterns of leisure agriculture POIs; NNI provides a quantitative assessment of whether the overall spatial pattern tends toward clustering or dispersion; and GCI reflects the degree of regional concentration across administrative units, thereby revealing disparities among different subregions. The empirical results indicate that leisure agriculture exhibits a typical clustered distribution, with high-density hotspots formed in Nanjing, Wuxi, and Suzhou (KDE results). The overall NNI value is less than 1, further confirming the clustering tendency, while the GCI analysis shows that resources are mainly concentrated in a few core cities, highlighting a pronounced regional imbalance. These three methods are complementary, enabling the capture of both micro- and macro-scale spatial characteristics, thus avoiding the bias that may arise from relying on a single indicator.

2.3.1. Kernel Density Estimation (KDE)

To reveal the spatial clustering patterns and density variations in the leisure agriculture resource in the study area, we used the kernel density estimation (KDE) tool of the Spatial Analyst Tools extension of ArcGIS Pro (v3.2, Esri Inc., Redlands, CA, USA). KDE has long been used for calculating the density value of unknown regions based on the distribution of specific spatial elements based on the kernel function [22,23]. The general formula for KDE is as follows:

$$f\left(x\right)={\displaystyle \frac{1}{n{h}^2}}\sum _{i=1}^{n}K\left({\displaystyle \frac{d_i}{h}}\right)$$

(1)

In the formula, $d_i$ represents the distance between the $i$ leisure agriculture POI and the estimation point; $h$ is the bandwidth parameter, controls the search radius, and affects the smoothness of the kernel function; and $K\left(\cdot \right)$ denotes the kernel function, which is usually Gaussian or Epanechnikov. In this study, the bandwidth $h$ was set to 10 km, following previous studies on the spatial distribution of rural tourism and POI. This value provides a balance between oversmoothing and excessive local variation and effectively captures the clustering patterns of leisure agriculture resources in the study area.

2.3.2. Nearest Neighbor Index (NNI)

The Nearest Neighbor Index (NNI) is a widely used method for evaluating the spatial proximity and dispersion of point features in geographical space [24]. It reflects the spatial distribution pattern of the features by calculating the ratio between the observed mean nearest neighbor distance and the expected distance under a random distribution. The spatial pattern of leisure agriculture POIs in the study area is measured in this study through the NNI. The formula is expressed as

$$D_1={\displaystyle \frac{1}{n}}\sum _{i=1}^n D_i$$

(2)

$$D_2={\displaystyle \frac{1}{2\sqrt{\frac{n}{S}}}}$$

(3)

$$R={\displaystyle \frac{D_1}{D_2}}$$

(4)

where $D_1$ denotes the observed mean nearest neighbor distance, $D_2$ refers to the expected distance under a random distribution, $n$ is the number of POI, and S is the total area of the study region. The NNI R interprets the distribution pattern as follows: R = 1 denotes a random distribution, R < 1 suggests a clustered distribution, and R > 1 implies a uniform or scattered pattern. NNI is particularly suitable for identifying spatial proximity and dispersion of leisure agriculture POIs, which provides important insight into their spatial organization.

2.3.3. Geographic Concentration Index (GCI)

A quantitative indicator of a phenomenon’s level of spatial concentration is the Geographic Concentration Index (GCI). It effectively captures the inter-city disparities in the distribution of leisure agriculture resources. Compared with KDE and NNI, which emphasize local clustering, GCI provides a macro-level measure of spatial concentration, making it suitable for identifying cities with highly aggregated resources versus those with more dispersed patterns [25]. This study applies the GCI to assess the inter-city distribution concentration of leisure agriculture POIs in the study area. The index is calculated as follows:

$$G={\sum }_{=1}^n{\left({\displaystyle \frac{X_i}{T}}\right)}^2$$

(5)

In this equation, $X_i$ represents the number of leisure agriculture resources in the $i$ prefecture-level city; $T$ is the total number of such resources in the study area, and $n$ denotes the number of cities considered. The higher the GCI value, the higher the concentration. It reflects a uniform distribution among cities when the index equals ${\displaystyle \frac{1}{n}}$; values above ${\displaystyle \frac{1}{n}}$ suggest spatial concentration, whereas values below it indicate a more dispersed layout.

2.3.4. Construction of the Regression Model

A multivariate regression model based on the Ordinary Least Squares (OLS) method is constructed to examine the influence of various factors on the spatial distribution of leisure agriculture in the study area [26]. This model introduces a set of socio-economic and environmental indicators as explanatory variables, with the number of leisure agriculture POIs at the county level serving as the dependent variable. The general form of the model is given as:

$$Y={\beta }_0+{\sum }_{i=1}^k{\beta }_i{X}_i+\epsilon$$

(6)

where $Y$ denotes the dependent variable (namely, the number of leisure agriculture sites), ${\beta }_0$ is the intercept term, ${\beta }_i$ represents the coefficient of the $i$ explanatory variable $X_i$ (like GDP, population, highway density), and $\epsilon$ is the error term, assumed to be normally distributed with constant variance and zero mean. OLS regression, which is especially well-suited for cross-sectional data analysis, is renowned for its interpretability and consistency in parameter estimation. It allows for a straightforward identification of the direction and significance of each factor’s impact on spatial distribution patterns and has been widely utilized in empirical studies on rural tourism and agricultural spatial organization.

3. Results

3.1. Composition and Spatial Differentiation of Leisure Agriculture Resource Types

In domestic research, leisure agriculture resources are commonly classified into several functional categories. For instance, Ren et al. proposed a widely adopted framework that distinguishes agricultural sightseeing, agricultural experience, agricultural technology, rural culture, and characteristic villages and towns [1]. This classification reflects both functional orientations and industrial characteristics and has been applied in multiple empirical studies across China. Building on this framework, the present study further refines the classification into five primary and 17 subcategories to systematically capture the diverse characteristics of leisure agriculture resources in southern Jiangsu in

Table 2. Classification of leisure agriculture resource types.

Primary Category	Subcategory	Quantity	Proportion
Recreational Sightseeing	Agricultural sightseeing parks, leisure farms, resort areas, natural scenic spots, forest parks, wetland parks	649	9.01%
Agricultural Experience	Ecological experience gardens, picking gardens, rural (fishing) homestays	2405	33.39%
Agricultural Technology	Agricultural science and technology parks, agricultural demonstration zones, crop/livestock production bases	3521	48.88%
Rural Cultural Heritage	Agricultural culture parks, folk culture parks, agricultural culture exhibition parks	198	2.75%
Characteristic Villages and Towns	Historic and cultural villages/towns, modern agricultural demonstration villages/towns	430	5.97%

.

The five resource types show different clustering patterns when viewed from the standpoint of spatial distribution (see Figure 2). Agricultural technology resources are primarily concentrated in the Nanjing-Zhenjiang border area, while agricultural experience resources are distributed, with Changzhou’s northern area serving as the core and extending in a belt-like pattern into adjacent areas. Leisure sightseeing resources are heavily concentrated in the border regions of Zhenjiang, Changzhou, and Wuxi. Rural culture resources are scattered in a point-like distribution, with a concentration in southern Wuxi. Featured villages and towns show relative sparsity in northern Nanjing but concentrate in the Zhenjiang, Changzhou, and Suzhou regions, demonstrating strong location dependency and historical-cultural inheritance characteristics.

The spatial distribution of different leisure agriculture resources in the southern Jiangsu region shows considerable variation (Figure 2), which reveals the differentiated spatial distribution patterns of each main category of leisure agriculture resources in this region. Figure 2 shows that leisure sightseeing resources are primarily concentrated in the border regions of Zhenjiang, Wuxi, and Changzhou, creating a core aggregation belt with noteworthy periphery concentration, while Nanjing and Suzhou have a fairly balanced distribution. Moreover, it indicates that agricultural experience resources are distributed with northern Changzhou as the center and gradually diffuse outward in an east–west-oriented belt-like structure. This may be because such resources (like agritourism farms and fishing lodges) are predominantly located in suburban areas away from urban centers. Figure 2 also demonstrates that agricultural technology resources are concentrated in the intersection of Zhenjiang and Nanjing and radiate to surrounding areas; Zhenjiang has a higher density across the entire municipality compared to other cities, while the periphery exhibits a lower density. Influenced by specific historical and cultural backgrounds, rural culture resources show scattered point-like distribution, with the densest area in southern Wuxi. Northern Nanjing has almost no resources for characteristic villages and towns; instead, these resources concentrate in Zhenjiang, Changzhou, Wuxi, and Suzhou, where they form multiple density clusters. A multi-core, outward-radiating pattern is generally seen in leisure agriculture resources, with high-density locations encompassing approximately 80% of the study area. Although the distribution in Suzhou is comparatively homogeneous, the most significant clustering area is located at the junction of Zhenjiang, Changzhou, and Wuxi.

3.2. Overall Distribution Pattern: Clustering and Concentration

The overall R value for leisure agriculture in the study area, as determined by the Nearest Neighbor Index (NNI), is approximately 0.50 (R < 1), suggesting a significant clustering distribution pattern. Figure 3 illustrates the county-level spatial distribution of the Nearest Neighbor Index (NNI) for leisure agriculture in the study area. First, most counties fall within the low to moderately low classes, indicating a predominantly clustered rather than random distribution pattern across the study area. In terms of spatial variation, very-low and low NNI values are concentrated in the urban–rural transition zones of metropolitan areas, such as Jiangning District of Nanjing, Jurong City of Zhenjiang, Wuzhong District of Suzhou, and Yixing City of Wuxi. These areas are strongly influenced by proximity to large urban markets and transportation corridors, leading to closely spaced POIs and highly clustered local groupings. Second, Moderate NNI values are mainly observed in the Yangtze River corridor and several inland counties, including the Wujin District of Changzhou and Danyang City of Zhenjiang. These regions still show clustering tendencies, but the intensity is relatively weaker, reflecting a balance between localized concentration and partial dispersion. In contrast, high and very-high NNI values occur sporadically in the southern and peripheral margins of the study area, such as parts of southern Changzhou and other fringe counties of southern Jiangsu. Leisure agriculture in these areas is less developed, with sparse and scattered POIs, making their spatial distribution closer to random or even tending toward uniformity.

Overall, the NNI results reveal a clear core–periphery gradient: low-value clusters around core-city outskirts, moderate values in transitional inland areas, and higher values toward the outer periphery. This pattern highlights the strong relationship between leisure agriculture distribution and the metropolitan spatial structure, where core and peri-urban zones emerge as clustering hotspots, while remote peripheries remain more scattered due to weaker market attraction and diffusion effects.

Figure 4 illustrates the spatial distribution of the Geographic Concentration Index (GCI) for leisure agriculture in southern Jiangsu. Overall, most counties fall within the moderate to high classes, indicating a significant tendency toward concentration at the regional scale rather than an even distribution.

Spatially, high and very-high GCI values are concentrated in Jiangning District of Nanjing, Wuzhong District of Suzhou, Yixing City of Wuxi, and Jurong City of Zhenjiang. These locations, situated in close proximity to major metropolitan centers, benefit from favorable accessibility and strong market demand, leading to pronounced clustering of leisure-agriculture resources and forming the core development zones. Moderate GCI values are observed in the Yangtze River economic belt and certain inland cities, such as the Wujin District of Changzhou and Danyang City of Zhenjiang. These areas demonstrate a medium level of concentration, with some clustering effects but not as pronounced as in the metropolitan core. In contrast, low and very-low GCI values are distributed in the outer peripheries of the study area, including northern Nanjing, northern Suzhou, and several remote suburban counties in southern Jiangsu. In these regions, leisure-agriculture POIs are scattered, reflecting low levels of spatial concentration and a marginal position in the overall pattern. Overall, the GCI results highlight a distinct core–periphery structure: resources are highly concentrated in the metropolitan core, moderately concentrated in the secondary surrounding zones, and weakly concentrated in the remote periphery. This pattern corresponds closely with the metropolitan hierarchy, transportation accessibility, and consumer market size in the study area.

3.3. Spatial Hotspot Identification: Kernel Density Distribution Analysis

The spatial hotspot patterns of various resource types are further shown by the findings of kernel density analysis (Figure 5). Leisure agriculture in southern Jiangsu generally exhibits a multi-core spatial structure radiating along transportation corridors and urban peripheries, with hotspots primarily distributed in the Changzhou-Wuxi-Zhenjiang junction belt. Each main resource type’s hotspot distribution largely conforms to Jiangsu Province’s “Three Zones and Five Belts” leisure agriculture layout planning, which highlights the spatial clustering advantages of areas with convenient suburban transportation and abundant tourism resources.

Spatial hotspots of various leisure agriculture resources are revealed by kernel density analysis (see Figure 3). It presents kernel density distribution maps for five types of leisure agriculture and the overall distribution in southern Jiangsu, with a positive correlation between color depth and resource distribution density. Several high-density hotspot zones have been created by various forms of leisure agriculture in the study area, which are mostly concentrated around central cities and along transportation corridors. These results are consistent with previous studies on the “Three Zones and Five Belts” spatial pattern for leisure agriculture that Jiangsu has proposed. Three major zones (like the metropolitan leisure agriculture zone, peri-hilly leisure agriculture zone, and peri-lake wetland zone) and five creative leisure agriculture belts along the Yangtze River, coast, Grand Canal, ancient Yellow River course, and Longhai Railway line are included. Additionally, kernel density analysis further confirms the distribution patterns shown in Figure 2—the clustering tendencies of various leisure agriculture in the study area’s major cities and corridors.

3.4. Multicollinearity Test: Variance Inflation Factor (VIF)

This study conducted VIF tests on all explanatory variables to ensure the robustness of regression analysis. The findings demonstrate that the direct entry into OLS modeling is allowed since all variable VIF values are much below 10 (

Table 3. Multicollinearity diagnostics of independent variables for leisure agriculture (VIF).

Variable	VIF Value	Multicollinearity Present
GDP	2.378	No
Population	3.070	No
Highway Destination	1.147	No
Gross Output of Agriculture, Forestry, Animal Husbandry, and Fishery	1.387	No
Average Elevation	1.570	No
Number of A-level Scenic Spots	1.658	No

) and show no serious multicollinearity problems among variables.

The Variance Inflation Factor (VIF) for each independent variable is calculated in the study’s regression analysis to test for multicollinearity among variables. Serious multicollinearity is generally indicated when VIF values are above 10, which is detrimental to model robustness. The VIF values for variables in the southern Jiangsu leisure agriculture regression model are presented in

Table 4. OLS regression results for leisure agriculture distribution.

Variable	Coefficient	Standard Deviation	p Value	t Value	VIF
Constant	−93.451	40.952	0.029	−2.165	Y
GDP	0.021	0.014	0.147	1.246	N
Population	0.203	0.384	0.601	0.473	N
Highway Density	−19.905	193.879	0.919	−0.081	N
Gross Output of Agriculture, Forestry, Animal Husbandry, and Fishery	0.000	0.000	0	7.329	Y
Average Elevation	1.378	1.902	0.474	0.935	N
Number of A-level Scenic Spots	15.350	7.154	0.040	2.231	Y

. Results show all variable VIF values are well below 10, and highway density (1.15), the output value of agriculture, forestry, animal husbandry, and fishery (1.39), average elevation (1.57), and number of A-level scenic areas (1.66) exhibit relatively low VIF values, which indicates weak linear dependence among these variables. Even GDP (2.38) and the population (3.07), however, keep VIF values within acceptable ranges. Therefore, there is no significant multicollinearity issue between explanatory variables, ensuring the robustness of regression coefficient estimation and the reliability of statistical results. Although GDP and other macroeconomic indicators do not significantly affect site distribution, their underlying consumption foundations, locational advantages, and institutional guarantees remain worthy of attention. This implies that deeper pathways of structural mechanisms and micro-agent behavior effects on spatial structure evolution should be the main areas of future study.

3.5. OLS Regression Analysis

This study seeks to identify overall influencing mechanisms of leisure agriculture resource distribution in southern Jiangsu and emphasize overall trend identification at the regional scale rather than location-specific spatial heterogeneity. Ordinary least squares (OLS) regression modeling was therefore selected for factor estimation and analysis. The OLS model, a classical global regression method, can be used to identify overall linear relationships among variables at the regional scale. Geographically weighted regression (GWR) and other spatial local models were not introduced in this study since it focuses more on connections between overall regional development characteristics and macro-driving variables.

The impact of various factors on the quantity of leisure agriculture at the district/county level (Table 4) was analyzed through OLS multiple linear regression modeling. The results indicate that the output value of agricultural, forestry, animal husbandry, and fishery (p < 0.001) and the number of A-level scenic areas (p = 0.040) positively affect leisure agriculture. This reflects that agricultural infrastructure and tourism resource clustering can attract and drive leisure agriculture. It is possible that variable selection, spatial heterogeneity, or scale effects prevented other variables—like GDP, population, and highway density—from passing significance tests. The regression model performs well overall, with R² = 0.7701, a significant F-value, and a non-significant Koenker (BP) test, indicating excellent model fit, absence of heteroscedasticity, and normally distributed residuals.

4. Discussion

This study adopts multi-source data to build a database of leisure agriculture resources in southern Jiangsu. The factors identified here have important implications for policymakers, local residents, and regional development. For example, transportation accessibility directly shapes how leisure agriculture sites are allocated and how efficiently they operate, while the level of economic development and population density reflects local demand potential and investment capacity. Together, these findings provide a solid evidence base for improving land-use planning, infrastructure development, and rural revitalization strategies not only in Jiangsu Province but also in other regions.

The results indicate that the spatial clustering of leisure agriculture in the study area is driven primarily by the structural interaction between agricultural production and tourism resources, rather than by traditional macroeconomic factors. In the broader context of China’s urbanization and rural revitalization, this underscores the importance of fostering synergies between agriculture and tourism instead of pursuing expansion based solely on scale or investment. The strong clustering observed in peri-urban transition zones illustrates the “urban spillover–rural reception” dynamic, while A-level scenic spots function as anchors that generate spillover effects benefiting surrounding villages. Together, these mechanisms provide meaningful guidance for local governments, farming communities, and visitors, highlighting the contribution of leisure agriculture to urban–rural integration.

Spatially, leisure agriculture in the study area follows a polycentric clustering pattern with peripheral gaps, as high-density areas are concentrated in peri-urban zones such as Jiangning, Wuzhong, Yixing, and Jurong. This distribution underscores the locational advantages of urban fringes in meeting leisure demand and is consistent with the urban spillover–rural reception dynamic.

The geographic concentration index indicates that leisure agriculture resources in southern Jiangsu are moderately concentrated, with clustering more evident in secondary or sub-center cities such as Jiangning, Wuzhong, and Wujin rather than in traditional urban cores. This pattern is consistent with recent studies on the spatial distribution of agritainment in China, which show that rural tourism destinations tend to cluster in economically developed regions while remaining more dispersed in peripheral areas [27]. Comparable findings across other Chinese provinces also suggest that terrain constraints and urban proximity play a decisive role in shaping distinct tourism clusters [28]. From a theoretical perspective, this tendency diverges from Christaller’s classical “center–periphery” model and aligns more closely with the asymmetric metropolitan development paradigm proposed by Fujita and Krugman.

In regions with advanced infrastructure and high levels of socioeconomic development, the marginal effects of conventional macroeconomic indicators appear to be weakening. Instead, structural drivers—particularly the integration of agricultural resources with tourism industries—are becoming decisive. This observation is consistent with international evidence: in peri-urban areas of Europe and North America, multifunctional agriculture has been found to prosper not through scale expansion but through synergies between farming and recreational services (Zasada, 2011; Bryant & Johnston, 1992) [15,29]. Studies from Canada highlight that peri-urban farms often succeed through diversification into tourism, direct marketing, and leisure-oriented services rather than traditional production metrics (Bryant & Johnston, 1987) [30]. In the study area, similar patterns can be observed at the Zhenjiang–Changzhou–Wuxi junction and in southern Suzhou and Liyang, where dense agricultural resources and scenic landscapes converge, underscoring the importance of agri-tourism integration in shaping spatial clusters.

Although leisure agriculture clusters have formed in local areas, problems of convergent development paths, repetitive business formats, and increasing ecological pressure still exist. This indicates that the current development paradigm remains constrained by “scale-first” logic. On the one hand, fierce homogeneous competitiveness and ecological resource overload are problems in high-density clustering areas. On the other hand, potential areas in periphery zones fail to coordinate development with core areas due to a lack of efficient policy support and resource linkage mechanisms, resulting in significant spatial “fracture zones.” This phenomenon aligns well with theoretical perspectives on “spatial development imbalance” proposed by Crescenzi and Rodríguez-Pose [31].

5. Conclusions

This study integrates multi-source POI data, kernel density analysis, nearest neighbor index, geographic concentration index, and OLS regression modeling to examine the spatial distribution and influencing factors of leisure agriculture resources in the study area. The main conclusions are as follows:

(1)

Leisure agriculture in the study area exhibits a highly uneven spatial distribution, characterized by multi-core clustering and a clear core–periphery structure. Urban sub-centers and urban–rural transition zones are the primary areas of concentration, while central urban districts present evident spatial blank zones.

(2)

Spatial statistical results confirm significant clustering, with first-tier core areas forming in districts such as Wujin, Xishan, and Dantu, and several second-tier hotspots emerging around sub-centers and ecological zones.

(3)

Agricultural resource endowments and tourism attractiveness are the key drivers shaping the spatial distribution of leisure agriculture, while the effects of macroeconomic indicators such as GDP and population density are less significant in this developed region.

The main contribution of this research lies in constructing a systematic spatial database and establishing an analytical framework that integrates spatial pattern identification, clustering measurement, and driving factor analysis. The findings provide empirical support for promoting coordinated regional development and guiding the spatial governance of leisure agriculture. These contributions were achieved by assembling multi-source spatial data (POIs, statistical yearbooks, and GIS layers) into a comprehensive database, and by employing established spatial analytical techniques (KDE, NNI, GCI, and OLS regression) to reveal both fine-scale clustering patterns and broader regional concentration dynamics.