Spatial Distribution Heterogeneity and Influencing Factors of Different Leisure Agriculture Types in the City

Abstract

Leisure agriculture is a crucial carrier for city agriculture and tourism growth. This study aims to explore the overall leisure agriculture and different types of leisure agriculture spatial sub-characteristics and their influencing factors. Taking the city of Xi’an, China, as an example, leisure agriculture was classified into four types: agritainments, agricultural parks, resorts, and rural homestays. According to this study, two ring zones and one core belt zone for leisure agriculture in Xi’an are dispersed unevenly and aggregated. Furthermore, geographic detectors and spatial principal components were employed as empirical techniques to investigate the primary factors influencing the spatial distribution of multiple leisure agriculture heterogeneity. The results about the influence mechanism indicate that the gross domestic product, population density, and accessibility to transportation dominated the overall spatial distribution, while the distance to scenic spots and water areas had a significant impact on the spatial distribution of some types. This research contributes to the reconstruction of leisure agriculture and promotes the sustainable development of agriculture and recreation by merging leisure agriculture spatial resources.

1. Introduction

1.1. Literature Review

Leisure agriculture links sustainability and the multifunctionality of natural resources and is an essential engine for agricultural development and city agricultural ecological restoration. The academic circles pay attention to agricultural tourism [1,2,3], mainly in rural areas, and advocate leisure agriculture and agritourism as the leading ways to develop rural areas. They contend that agricultural tourism still has much development potential and produces long-term economic effects in the face of rural poverty and regional economic imbalance [4,5]. However, for cities, not only are a large number of people gathered and land is in short supply [6,7], but agricultural stakeholders take the initiative to maintain agricultural land and agricultural activities on the premise of ensuring more economic income to sustain their livelihoods [8,9]. Therefore, balancing ecological and economic development, retaining the agricultural labor force and efficient use of agricultural land has become the major problem facing city agricultural land. Similar to rural agrotourism, the development of city leisure agriculture can be a better answer to this problem [10].

Leisure agriculture can be defined in various ways, but in essence, it involves agricultural activities that serve both agricultural and recreational purposes in the context of tourism [11,12]. Scholars have rigorously differentiated rural tourism from leisure agriculture tourism in some earlier studies [13,14], and they typically felt that rural tourism comprises leisure agriculture tourism. Due to geographical features and analytical viewpoint variations, leisure agricultural tourism is still tied to agrotourism, agri-leisure, eco-agriculture, and sightseeing agriculture [15,16,17,18].

The differences in the primary forms of leisure agriculture in different locations result from many environmental, social, and historical reasons. Italy has a unique regulatory and legislative mechanism for economic, social, cultural, and ecological issues related to leisure agriculture [19,20]. In the United States, leisure agriculture is more commonly known as citizens’ farms, which link farms and communities with mutual aid so that residents and farmers share the responsibility of operating and enhancing the profitability of farms [21,22]. In Taiwan, China, there are farms specializing in leisure agriculture that customize the corresponding experience projects according to the different interests of tourists [23]. In conclusion, combining components of agriculture and tourism in the area can encourage the multifunctional and multilayer usage of agricultural resources and maintain a balance between the ecological and the economic [24]. After a late start in China, leisure agriculture has continued to learn from the successful business models and advanced management experience of other nations, progressing from being heavily dependent on agricultural products to diversified management, high-end product positioning, and, finally, to recognizable towns and other new rural economic forms [25,26].

Several academics are concerned about the countermeasures and solutions surrounding leisure agriculture. Some have proposed similar promotional initiatives in response to problems with human resources, technology, and other factors [25,27]. Andéhn et al. pointed out that agri-tourism should embed multiple meanings for the destination to satisfy the various needs of stakeholders [28]. Kastenholz et al. proposed evaluating, selecting, and positioning tourism segments through a field survey of rural tourists to attract sustainable market segments [29]. While others concentrate on benefit evaluation studies of agritourism, contrasting leisure agriculture with other types of agricultural operations, arguing that leisure agricultural operations can achieve a deeper level of sustainability and create multiple cultural and socioeconomic benefits [30] or demonstrating the profitability of agritourism by outlining its operational capabilities during the COVID-19 pandemic [31]. Others, however, are more concerned with the operators and participants. For instance, Devesa et al. proposed that tourism motivation is the determinant of visitor satisfaction [32]; Kamau et al. demonstrated that poorer, younger, and less-connected farmers are less engaged in organic agriculture [33]; and Pachoud et al. demonstrated that effective interactions between stakeholders and visitors could promote the potential of farm tourism [34]. Some academics also contributed to studying leisure agriculture’s spatial distribution. For instance, Cui Jiaxing et al. spatially elaborated on agricultural picking orchards. They discovered five spatial clusters in China [35], and Xiang Yan et al. investigated the spatial distribution, evolution, characteristics, and influencing factors of 436 leisure agriculture brands in Hebei Province, China [36].

In summary, in terms of research content, the geospatial analysis of leisure agriculture is less involved, and the research perspective and methodology should be expanded and innovated. Some kinds of literature analyze the characteristics of the overall leisure agriculture development based on research data [29,37], and the exploration of the influencing factors of leisure tourism focuses on a single perspective [38,39], which lacks the consideration of the regularity of the distribution of leisure agriculture and the spatial roles of the influencing factors. Some of them only elaborate on the relationship with leisure agriculture from statistical data [40], weakening the spatial heterogeneity of leisure agriculture resource endowment or ignoring the spatial correlation effect between different types of leisure agriculture and driving factors [41].

For these questions, this paper separately studies the overall and different types of leisure agriculture, using ArcMap 10.8-related operations, as well as geographic detectors and spatial principal components and other analytical pathways, to deal with multiple types of data, to explore the characteristics of the spatial distribution, and to quantitatively calculate the degree of influence of the factors on the spatial distribution of the overall and different types of leisure agriculture. Finally, the general rules of leisure agriculture distribution and influencing factors are summarized, and policy suggestions for the restructuring and layout optimization of leisure agriculture and business suggestions for leisure agriculture stakeholders are put forward to further promote the spatial resource integration of leisure agriculture.

1.2. Definition of Concepts

Various types of tourism can support the capitalization of agricultural resources and the stability of its economy [42]. For the different types of leisure agriculture, priority should be given to clarifying the differences and linkages between the types in research. Currently, leisure agriculture is classified differently based on the goals and objectives of academic study. Leisure agriculture was divided into three categories by Lu Y et al. [43]: city-dependent, agricultural-dependent, and nature-dependent. Pitchayadejanant et al. classified it into six categories: agricultural product display, tourists’ participation in agricultural activities, rustic boarding, agricultural knowledge training, agricultural product distribution, and agricultural business products [44]. Lu Lin et al. classified it into eight categories: farmhouses, rural homestays, rural boutique hotels, campgrounds, scenic paths, traditional villages or towns, idyllic complexes, and tourist towns [45]. This study separated leisure agriculture into four groups based on scale; product type; and management style: resort, rural homestay, agritainment, and agricultural park.

• Resort: The resort is a popular vacation spot that offers a variety of services, including lodging, recreational facilities, and entertainment options. Typically situated in picturesque locations, resorts boast a plethora of amenities such as swimming pools and golf courses to cater to the leisure and vacation requirements of tourists. Their main objective is to provide visitors with a complete relaxation and entertainment experience.
• Rural homestay: A rural homestay is a place where rural or city residents transform their homes into places to receive tourists, offering simple but unique accommodations. Rural homestays usually offer bed and breakfast personalized services and focus on providing family-style, friendly service.
• Agritainment: Agritainment is usually a place where farmers or farm owners transform their farmland into a place to receive tourists, which is small in scale and easy to manage and operate. It provides tourists with agricultural experience, farming participation, food and lodgings, and other services. Agritainment focuses on showcasing rural life, experiencing agricultural culture, and providing opportunities to spend time with nature.
• Agricultural park: An agricultural park is a place where agricultural production and operation and sightseeing tourism are combined. It can be a large farm or agricultural park, providing activities such as agricultural sightseeing, agricultural product picking, and farming experience. Agricultural parks focus on showcasing modern agricultural development, with large tracts of land managed in a unified and centralized manner.

Resorts, rural homestays, agritainments, and agricultural parks are all tourism and leisure venues to meet people’s needs for rural life, natural environment, and leisure vacation. They are all related to farmland, the natural environment, and agricultural culture, providing a rural experience and leisure services at different levels and in different ways. In contrast, it is more worthwhile to emphasize the differences between them. Agritainment focuses more on the agricultural experience and the display of rural culture, emphasizing interactions with farmers and participation in farming activities. Agricultural parks focus on agricultural production and the display of modern agricultural technology and business models. Rural homestays mainly provide lodging services and focus on a family-style living atmosphere. Resorts, on the other hand, are designed to provide a full range of leisure and recreational facilities to meet the high-end and extended vacation needs of tourists.

2. Materials and Methods

2.1. Research Area Overview

The city of Xi’an is located in the west-central region of China, between longitude 107°40′–109°49′ E and latitude 33°42′−34°45′ N. The area under its jurisdiction consists of 11 municipal districts and two counties, covering a total area of 10,752 square kilometers. In recent years, the populations of many new first-tier cities have continued to expand massively against the backdrop of diminishing population increments in China’s four traditional first-tier cities. Xi’an, as a representative of the new first-tier cities, is a famous city for tourism in China, which has welcomed more than 350 million domestic and international visitors, bringing in 500 billion RMB in total in 2022. A wide range of source markets, a well-developed tourism industry, and rich, natural landscapes provide favorable conditions for the development of leisure agriculture. According to the current distribution of different leisure agriculture in Xi’an (Figure 1), a large number of leisure and tourism agriculture has been developed in some areas, such as the city’s outskirts and the northern foot of the Qinling Mountains. Based on the above, this paper selected Xi’an City, Shaanxi Province, China, as a research case.

Various forms of leisure agriculture have different geographical and numerical distribution patterns (Figure 2). It can be seen that most are agritainments, followed by resorts, agricultural parks, and rural homestays, in that order. Regarding the distribution of leisure agriculture among the various counties and districts, Chang’an District has the most significant number, while Beilin District has the lowest.

2.2. Data Resources

The terms resort, picking garden, farm, agritainment, rural homestay, and leisure homestay as keywords have been used to identify sites of interest in Baidu API (application programming interface). After eliminating duplicate data, such as parking lots, apartment-style bed and breakfasts, and express hotels without agricultural qualities, the study focused on 1348 vector points collected as the main subject. The GSCloud extracts numerous photos of Xi’an’s elevation, which are then mosaiced together to create the digital elevation model (DEM). To support the pertinent analysis of the study, the normalized difference vegetation index (NDVI) is linked to the findings of the current data processing [46]. The information from the Xi’an Statistical Yearbook on the regional GDP, population, and government spending on the arts and tourism is transformed into a raster format in ArcMap 10.8 before being discretized and categorized. To gauge the level of governmental support for the growth of leisure agriculture, sort the number of notification papers connected to the industry that the district and county governments in Xi’an issued from 2014 to 2023. Water and road network data are extracted through Open Street Map and converted into vectors. Refer to the Technical Standard of Highway Engineering (JTG B01-2014), assigning the value of 70 km/h for national highways, 60 km/h for provincial highways, 30 km/h for county highways, and uniformly assigning the value of 15 km/h for other areas, to calculate the time from each raster to the corresponding district and county government as the result of transportation accessibility. The point of interest (POI) of the district and county government here is also in Baidu API. A-grade scenic spot data were vectorized to the panoramic spot data published by the Shaanxi Provincial Department of Culture and Tourism (

Table 1. Sources and descriptions of the data.

Data Type	Data Description	Data Source
POI data	point of interest	https://lbsyun.baidu.com/, accessed on September 2022
DEM data	2022 elevation raster data with a 30 m spatial resolution	https://www.gscloud.cn/ accessed on 2023
NDVI data	2020 observations with a time resolution of one year and a spatial resolution of 30 m	https://doi.org/10.1016/j.rse.2019.111395.
Socioeconomic data	gross regional product (GDP), regional population data, government investment in cultural and tourism industry,	http://tjj.xa.gov.cn/tjnj/2022.htm accessed on 2022
Degree of policy support	Statistics on relevant documents	https://www.xa.gov.cn/index.html accessed on June 2023
Road network, water area vector data	road network vector polyline data, water vector polygon data	https://www.openstreetmap.org accessed on 2022
Provincial Map of China Review	No. GS(2020)4619, Vector Data	http://bzdt.ch.mnr.gov.cn/index.html accessed on 2022
A-grade scenic spot	including all scenic spot vector point data for Xi’an 1A, 2A, 3A, 4A, and 5A	http://whhlyt.shaanxi.gov.cn/ accessed on 2022

).

2.3. Methods

2.3.1. Kernel Density Analysis

$$\widehat{f}\left(x\right)=\frac{1}{{nh}^d}\sum _{i=1}^{n}K\left[\frac{1}{h}\left(x-x_i\right)\right]$$

(1)

Kernel density is a method for measuring the degree of agglomeration and distribution of elements and spatial correlation, which can respond to the distribution law of the research object and is a commonly used method for spatial analysis [47]. In addition, it can handle high-dimensional data and has the advantage of smoothness, but it should pay attention to the boundary sensitivity and parameter setting. For this study, the visualization results of the kernel density analysis can show the effect of leisure agriculture agglomeration and the focused location of agglomeration and intuitively describe its spatial distribution pattern. The number of leisure agricultures is given by n, the kernel density function is provided by K, x − x_i gives the geographic straight line distance between each leisure agriculture, and the smoothing index of the kernel density measurement is given by h.

2.3.2. Spatial Gini Coefficient

$$G=\frac{1}{2n^2\overline{s}}\sum _{i=1}^n\sum _{j=1}^n\left|s_i-s_j\right|$$

(2)

The spatial Gini coefficient is often used as a standard measure of the unevenness of spatial distribution in geography, reflecting the degree of inequality of geographic phenomena among different geographic units [48]. It is simple to calculate and easy to understand, but its accuracy is affected by sample size. Therefore, the study is only used when exploring the homogeneity of all recreational agriculture. n refers to the number of road networks, and s_i and s_j refer to the density of elemental points in place i and place j. In the research, the statistical quantities of leisure agriculture in each district and county of Xi’an were calculated using the districts and counties as the basic units.

2.3.3. Nearest Neighbor Index Method

The method can quantitatively assess the degree of aggregation of point samples, but the lack of standardization, the tendency to ignore proximity relationships, and the limitation of the influence of sample size may lead to less accurate results [49]. The study addresses this issue by incorporating the coefficient of variation for additional clarification similar to the Gini coefficient, which is the only one used for all leisure agriculture studies. Its essence is

$$R=\frac{{\overline{r}}_i}{{\overline{r}}_e}$$

$${\overline{r}}_e=\frac{1}{2\sqrt{\frac{n}{A}}}$$

(3)

R is the nearest neighbor index, r_i refers to the nearest neighbor distance, r_e refers to the theoretically most relative neighbor distance, n refers to the number of leisure agriculture, and A is the area of Xi’an City. R > 1 tends to be uniformly distributed, R = 1 tends to be randomly distributed, and R < 1 tends to be cluster-distributed.

2.3.4. Coefficient of Variation

$$R=\sqrt{\sum {(S_i-S)}^2/n}$$

$$CV=\frac{R}{s}$$

(4)

The study uses the coefficient of variation of polygons to validate the results of the nearest neighbor index. S stands for the polygon area, n for the number of polygons, and V for the mean polygon area. CV stands for the coefficient of variation. According to the relevant division [50], for CV < 33%, the point elements are uniformly distributed; for 33% ≤ CV ≤ 64%, the point elements are randomly distributed; for CV > 64%, the point elements are clustered.

2.3.5. Riley’s K Function

Riley’s K function is a kind of multi-distance spatial cluster analysis, inscribes the degree of spatial correlation among all the observations based on the distribution of the sample points [51]. Its visual representation is excellent and suitable for the point-of-interest data in this study but not for the polyline and polygon data. ArcMap 10.8 was used to calculate the distance between each type of leisure agriculture point and the points for analysis, to get the distance statements and graphs, and to analyze the relationship of the observed value curves with the expected value and the high and low curves of the confidence intervals. The principle is as follows:

$$L\left(d\right)=\sqrt{\frac{A\sum _{i=1}^N\sum _{j=1,j\ne i}^{N}k\left(i,j\right)}{\pi N\left(N-1\right)}}$$

(5)

When Observed > ExpectedK and ObservedK > HiOonfEnv, point factor significance clustering is observed. For Observed < ExpectedK and ObservedK < HiOonfEnv, the point elements are significantly discrete, and the other results are randomly distributed.

2.3.6. Analysis of Accessibility

The study simplifies the accessibility model by weakening the effect of the gradient and other factors on the speed and replacing the actual speed with the general travel speed of the road. The core principle of the model is to superimpose the ratio of distance traveled and general travel speed of different classes of roads between the origin and the destination to characterize the accessibility class, which is realized by ArcMap 10.8 to calculate and express.

$$A_i=\sum _{m=1}^n\frac{S_{ij}}{v}$$

(6)

S is the distance between the distribution point of leisure agriculture and the point data of the district and county government, m is the number of types after the classification of the road network, and v is the speed attribute value of different roads.

2.3.7. Analysis of Geographic Detector

The geographic detector is suitable for studying multiple sample values and is implemented based on the software to explore the main factors that influence the spatial distribution of geographic phenomena [52]. It can identify nonlinear relationships but more demands on the data at the same time. In this study, the geographic detector was chosen to explore the extent to which each factor affects the overall recreational agriculture:

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

(7)

P is the degree to which the measure variable X explains the spatial differentiation of attribute Y, h is the number of layers, N is the number of units, and σ² is the regional variance. In this study, the p-value measures whether the test passed or not, while the q-value portrays the degree of influence of the driver.

2.3.8. Analysis of Spatial Principal Component

Principal component analysis (PCA) is often used in chemistry, medicine, and crop-related evaluation [53,54]. In addition, we implement the spatial principal component analysis using ArcMap and statistical methodologies. The degree of explanation of the dependent variable by the independent variables was assigned to several uncorrelated principal component factors and then spatially expressed. According to the weighted coefficients, the value of each raster-independent variable after standardization is weighted by multiple rasters to express the strength of the spatial role of the factors. The spatial distribution of the principal component factor role is examined to determine the score result value for each principal component.

3. Results

3.1. Spatial Distribution of Leisure Agriculture Characteristics

3.1.1. Spatial Pattern of Leisure Agriculture

POI is a type of spatial point data representing an actual geographic entity in the form of data containing information such as name, category, longitude, latitude, and address [55]. The findings of resorts, rural homestays, agritainments, and agricultural parks are obtained using kernel density processing and a standard deviation ellipse analysis (Figure 3).

Overall, leisure agriculture in Xi‘an is mainly distributed in two ring areas and one core belt area. The two ring-shaped zones refer to the leisure agriculture distributed in Beilin District, Yanta District, New City District, and Lianhu District within the third ring of the core city area, mainly in the form of rural homestays; the other ring-shaped zone refers to the leisure agriculture distributed in Lintong District in the northeastern outskirts of the city around the Terracotta Warriors of the First Qin Emperor’s Mausoleum and Huaching Pond, mainly in the form of agritainment. The core belt-shaped area refers to the distribution of leisure agriculture clusters formed around the northern foothills of the Qinling Mountains, with scenic spots such as the Xi’an Qinling Safari Park and the Taiping National Forest Park. Leisure agriculture there is dominated by small-scale agritainment and some agricultural parks. Zhouzhi County and Yanliang District have fewer leisure agriculture options than other areas. To summarize, the belt and core zones are intertwined to form the present distribution of leisure agriculture in Xi’an City.

3.1.2. Concentration and Equilibrium of Spatial Distribution of Leisure Agriculture

The study counted the number of recreational farms in each district and county and used the Lorenz curve to explore the uniformity of their distribution (Figure 4). The whole curve presents a downward convex trend, and the Gini coefficient of the distribution of leisure agriculture formats is 0.47498. According to the relevant expression [56], a result of 0.4–0.5 indicates a large gap, which comprehensively demonstrates that the distribution of leisure agriculture in various districts and counties of Xi‘an is unbalanced.

The Lorentz curve only shows the nonuniformity of the distribution of leisure agriculture, and the nearest neighbor index can be used to indicate whether it is concentrated. The average nearest neighbor tool was used in ArcMap 10.8 to show that leisure agriculture is clustered (Figure 4). Then, by the coefficient of variation test, we get 286.76% > 64%. Thus, the conclusion is correct, and the clustering effect is evident.

The uniformity is described using an optimized hot spot analysis (Figure 5), with red spots denoting hot regions of the distribution of leisure agriculture, blue spots indicating places with a lower distribution of leisure agriculture, and others not statistically significant. The distribution of rural homestays is clustered, and the majority of them are located in the central area of the city, where they have a convenient transportation environment and good infrastructure, making them a more popular choice of tourist accommodations. The results show that resorts are primarily distributed in the middle of the city. The agricultural park is large in scale and dispersed throughout all districts and counties, with the core of its distribution primarily in the Lianhu, Yanta, Chang’an, and Baqiao Districts. Agritainments are widely distributed. Most are spread around the northern foot of the Qinling Mountains, with less distribution in Zhouzhi County. Uneven distribution can be seen across several forms of leisure agriculture.

Additionally, Riley’s K function was used to examine the clustering and chose a confidence level of 90% (Figure 6). In general, the resorts are more significant than the predicted values, and the high values of the study’s confidence intervals are at near ranges, indicating an agglomerated distribution. Additionally, in the longer length, the predicted value is higher than the observed value, showing a random distribution. The observed values for the remaining rural homestays, agritainments, and agricultural parks are significantly clustered and substantially higher than the high values of the anticipated values and confidence intervals. In conclusion, the various forms of leisure agriculture exhibit uneven and clustered features comparable to the general distribution.

3.2. Elements Influencing the Spatial Distribution of Leisure Agriculture’s Heterogeneity

3.2.1. Basis for Establishing the Impact Factor

The variables influencing the spatial distribution of leisure agriculture are viewed from three perspectives: social, regional, and natural. Transit accessibility, the distance to watersheds, and the presence of A-class scenic sites serve to describe the regional factors. Natural factors are primarily examined in terms of elevation and vegetation cover factors. Social factors are divided into policy, demographic, and economic aspects.

Through ecological function transformation and management system innovation, local governments may support economic restructuring and the green change of the social growth mode to support the integrated development of rural tourism [57]. GDP can typically measure a region’s income and expenditure levels, which affect people’s willingness to consume leisure agriculture and the degree of consumption. The population plays a role in the main body of operations and consumption, influencing the development of leisure agriculture to some extent. The distance from A-class scenic areas and bodies of water, along with favorable accessibility circumstances, can improve tourist appeal and operate as an internal mechanism to encourage the growth of leisure agriculture. In natural settings, leisure agriculture primarily depends on agricultural development, and the spread of this industry follows natural laws. In summary, the determination of the impact factor can follow these rationales (Figure 7).

3.2.2. Empirical Analysis of Impact Factors

Extract the data of the variables after being gridded in ArcMap and discretize them using the natural breakpoint method, then import them into Geo_Detector for analysis. The p-values were all zero and significant at the 0.001 level, which shows the method may be used for exploration, and it is a high level of accuracy in the factor’s interpretation of the geographical distribution.

From the factor detector results (

Table 2. Results of the factor detector.

Project Level	Factor Level	q Statistic	p-Value
Social factor	Population (X1)	0.256802	0.000
	GDP (X2)	0.179414	0.000
	Government investment in culture and tourism (X3)	0.151682	0.000
	Degree of policy support (X4)	0.13882	0.000
Nature factor	NDVI (X5)	0.028301	0.000
	DEM (X6)	0.081927	0.000
Regional factor	Distance to water area (X7)	0.054637	0.000
	Distance to A-grade scenic spot (X8)	0.165136	0.000
	Transportation accessibility (X9)	0.234289	0.000

), it is clear that the order of the strength of the influence factors is population factor > transportation accessibility > GDP > cultural and tourism investment > degree of policy support > distance to scenic spots > DEM > distance to water area > NDVI, with population factor and transportation accessibility having the most extensive influence effects. The interaction detector demonstrated (Figure 8) that the distribution of leisure agriculture was better explained by a combination of factors than by a single item.

Using the geographical detector to produce a logical impact is challenging because of the limited number of samples and a large number of specified elements for each leisure agricultural research, so the spatial principal component is chosen as the analysis method. The impact factors passing a KMO (Kaiser–Meyer–Olkin) value of 0.5 and the Bartlett sign values of 0.000 met the analytical requirements. Nine criteria were reduced to four components corresponding to the data’s interpretation. The square value of the factor loadings reflects the contribution of each variable to the principal component, and the square value of the principal component loadings can introduce the degree to which the correlates contained in the principal component factor explain the spatial distribution of different types of leisure agriculture (

Table 3. Results for the spatial principal component analysis.

Types	PC Factor	Total	Variance Percentage	Cumulative Percentage	Dominant Factor
Resorts	PC1	3.426	38.064	38.064	Transportation accessibility, NDVI
	PC2	1.699	18.878	56.941	GDP
	PC3	1.389	15.435	72.376	Distance to A-grade scenic spot
	PC4	0.896	9.952	82.328	Government investment in culture and tourism
Rural Homestays	PC1	3.028	33.644	33.644	Transportation accessibility, NDVI
	PC2	1.813	20.144	53.787	GDP
	PC3	1.132	12.573	66.361	Distance to A-grade scenic spot
	PC4	0.78	8.663	75.023	Distance to the water area
Agritainments	PC1	1.717	19.076	19.076	Degree of policy support
	PC2	1.392	15.468	34.545	Distance to the water area
	PC3	1.29	14.33	48.874	Transportation accessibility
	PC4	1.039	11.547	60.421	DEM
Agricultural Parks	PC1	2.522	28.027	28.027	DEM, Transportation accessibility
	PC2	1.583	17.584	45.611	GDP
	PC3	1.189	13.207	58.818	Distance to the water area
	PC4	1.029	11.432	70.25	Degree of policy support

).

ArcMap 10.8 was utilized to depict the influencing factor fishing nets, which were uniformly 1000 m × 1000 m, to visualize the strength of the principal component factor impacts spatially. After calculating the average value for each grid, the principal component factor scores were batch-calculated to produce the spatial distribution map of the principal component factor impacts (Figure 9).

The spatially weighted effect of each influence factor corresponds to the geographical characteristics of the spatial distribution of various types of leisure agriculture. Resorts are mainly concentrated in the central part of Chang’an District, with a better ecological environment and adjacent to the northern foot of the Qinling Mountains; most of the rural homestays are distributed in the city center area, and some of them are concentrated around the scenic spots, with good accessibility and policy support guiding them to make them more dependent on the tourist attractions; there are a large number of agritainments, and they are widely distributed, and rich, natural resources combined with the dense population and policy guiding them help to form the phenomenon of aggregation and distribution; the agricultural parks cannot be developed without agricultural resource conditions and policy support; the Shuangyi District government’s guidance and construction documents for leisure agriculture in the past decade are much higher than those of other regions, making its leisure agriculture development stand out in the context of similar natural resources and regional conditions.

The spatial distribution of leisure agriculture is primarily influenced by population and GDP, with policy support and investments in cultural tourism showing differences for various forms of leisure agriculture. The population and GDP are the factors that affect the spatial distribution of recreational agriculture. Many results show apparent differences between districts and counties, and the values do not change continuously because the population, GDP, the degree of policy support, and government investment in culture and tourism are counted by districts and counties, as in the case of the rural lodging in the figure of PC2 (Figure 9). It is worthwhile to emphasize that government investment has a significant effect on the spatial distribution of the resorts, especially in districts and counties that are located on the outskirts of the city; and the degree of policy support produces prominent city periphery promotion and center inhibition effects on the spatial distribution of agritainments and agricultural parks.

The DEM and NDVI have a small impact on leisure agriculture; the former impacts the distribution of resorts and rural homestays, while the latter influences the distribution of agritainments and agricultural parks. According to the results of the spatial distribution of principal component factors (Figure 9), the distribution patterns of PC4 for rural resorts and PC1 for agricultural parks under the influence of elevation roughly conform to the characteristics of Xi’an City, which is high in the south and low in the north, with a distinct demarcation between plains and mountains. In areas with different vegetation indices, PC1 spatial distribution patterns are similar in resorts and rural lodges, with a high concentration of city development zones in the center and north and a low concentration of agricultural development zones and ecological environment safety control zones in the south. The values of DEM and NDVI were separated into different intervals to count the number of different forms of leisure agriculture in order to further expand the spatial action mechanism of the components (Figure 10). With the rising elevation, the quantitative benefits of agricultural parks and agrotourism become increasingly substantial. An average level of vegetation cover makes agriculture more conducive to leisure activities, but an excessive amount tilts land use features more in favor of ecological than economic factors.

The distribution of leisure agriculture is significantly influenced by transportation accessibility, whereas resorts and rural homestays are influenced mainly by scenic distance, and agricultural parks and agritainments are primarily influenced by water distance. Transportation accessibility can be used to explain the spatial distribution of all types of leisure agriculture. It also helps to create modern, convenient transportation conditions, proximity to markets, and other advantageous locations, all of which support the growth of the tourism industry and its physical space [58]. The number of leisure agriculture dispersed across the various classes was tallied (Figure 11) to further investigate the impact of transportation convenience on the distribution of leisure agriculture. Most are widely dispersed below rank 5, and less than 10% are found in isochronous circles higher than that. In conclusion, the distribution of leisure agriculture preferentially selects places with better accessibility within a comparable range of distances from metropolitan regions, improving market mobility for operators and participants and saving time on trips.

A-grade scenic areas positively influence local leisure agriculture and can exchange resources and support one another. The availability of water resources ensures agricultural irrigation and the diversification of leisure agriculture. The distributions of PC3 for resorts and rural homestays are characterized by cluster distribution because of the high-intensity influence of the distance of A-class scenic spots (Figure 9). The results in PC2 for agritainment and PC3 for agricultural parks as the dominant factor show that being influenced by water is more distributed in Lintong and Chang’an Districts. Additionally, data on the distribution of leisure agriculture about distances from scenic spots and water areas (Figure 12) reveal that this activity is mainly concentrated within a range of 2–10 km. At 16 km from water and 12 km from A-level scenic areas, the distribution of all types achieves a stable condition and does not noticeably alter as the distance increases.

4. Discussion

In terms of the analysis of the spatial layout, POI data were selected as the object of the study compared to previous studies exploring the influencing factors of leisure agriculture [26,59]. The specific geographic position of the data can build a link between the spatial location and the function of influencing variables by correlating it to the values of various variables at the spot. The majority of research [35,60] has concentrated on leisure agriculture as a whole or a subset. To refine and expand the research, leisure agriculture should be characterized from the perspectives of size, product type, and business management. Additionally, we summarized the geographical distribution features of leisure agriculture and its various forms from distribution balance and agglomeration perspectives. Regarding the research methodology, various methods suitable for leisure agriculture or individual types were screened. In addition, compared with popular spatial analysis methods in recent years such as geographically weighted regression [61,62], geographic detectors [63,64], spatial autocorrelation [65], etc., the study of the contribution of spatial principal component dimensionality reduction expression factors to the spatial distribution of each type of leisure agriculture is valuable and enhances the space, embodied in the combination of numerical calculations and geographical visualization. As for the influence mechanism, comparing the previous scholars analyzed the spatial distribution of leisure agriculture from the perspective of spatial measurements [66,67] and quantified the degree of government support for leisure agriculture [33,68]. The results of the detection are consistent with some earlier studies, which found that a range of regional, natural, and social factors all affect the spatial distribution of leisure agriculture [69,70,71], with social factors, in this case, having a more substantial influence than regional and natural ones. Meanwhile, spatial principal component dimensionality reduction was used to thoroughly examine the primary influencing variables of the different types of spatial distribution, and the components’ contributions to each type’s spatial distribution were represented. It has been discovered, the degree to which various factors affect the spatial distribution of leisure agriculture and each type. For instance, regional characteristics like water areas, scenic spot distances, and DEM have minimal impacts on the spatial distribution of leisure agriculture but have a considerable impact on distinct kinds (Table 2 and Table 3). These situations are easily ignored, so the classification study of leisure agriculture is vital. Overall, the study of the heterogeneity of leisure agriculture is not only reflected in the unevenness of spatial distribution and the differences in the distribution of different types of leisure agriculture but also needs to pay attention to the spatial differences of the role of the influencing factors.

Based on the above discussion, more profound research can be carried out subsequently. At the macro level, the contribution value of impact factors can be used as weights to further establish the evaluation system of leisure agriculture [72]; at the micro level, the surplus phenomenon of leisure agriculture can be pinpointed, and a subject of leisure agriculture can be analyzed in terms of its comparative advantages, which means some suggestions may be given to further guide the transformation.

It is emphasized that there are some limitations to this study. The possibility exists that POI data may be partially captured, and the need for real-time access makes it difficult to summarize and explore patterns over longer timescales. In addition, leisure agriculture is considered a multidimensional concept involving first-hand agricultural activities, natural experiences, cultural characteristics, and traditional lifestyles [73]. Therefore, there may be a need for more detailed big data to support the consideration and study of several market factors, such as consumer and operator behavior and decision-making at the micro level in terms of influencing factors.

5. Practical Insights

This paper’s research provides some insights and suggestions for leisure agriculture development and operation:

• Concerning the government, the government should create appropriate forms of leisure agriculture and execute effective development strategies for various kinds of leisure agriculture based on local conditions. For instance, while agricultural attractions and parks require government guidance, resorts require more government investment. In addition, accessibility distance somewhat influences the distribution of leisure agriculture, and transportation facilities should be continually improved to offer city residents convenient services to experience rural life and encourage market flows.
• As for agricultural resources and the environment, water resources are crucial to agricultural tourism and should be subject to ongoing protection measures. In addition, agricultural resources are integrated and exploited to form a clustered agrotourism landscape. While operating and managing, ensure the sustainable development of the ecological environment and do not overdevelop.
• As for the parties involved in the operation, they should carry out appropriate forms of leisure agriculture by using geographical advantages, such as creating rural lodgings near beautiful areas or city centers and carrying out farmhouse activities close to the water. Additionally, the area can be chosen according to the business model decided for leisure agriculture.

6. Conclusions

Four categories were created from the gathered leisure agricultural POI data, each explaining the geographical distribution law and examining the degree and method of influencing elements’ spatial effects. This investigation led to the following conclusions:

Firstly, the distribution of leisure agriculture is generally unbalanced and exhibits the spatial features of clustering; at the moment, it exhibits the distribution pattern of two rings plus a belt. It is mainly distributed in the third ring of Xi’an, Shuangyi District, near the Terracotta Warriors of Qinshihuang Mausoleum and Huaqing Pond, as well as a large number of clusters distributed in the northern foothills of the Qinling Mountains.

Secondly, leisure agriculture in Xi’an was classified into four categories: agritainments, rural homestays, agricultural parks, and resorts, and it was found that the leisure agriculture industry represented by rural homestays is predominant in the center of the city and that agritainment is predominant in the northern foothills of the Qinling Mountains; the leisure agriculture parks and resorts are distributed more often in Chang’an District, which is in the southern outskirts of Xi’an.

Thirdly, for the spatial distribution of all types of leisure agriculture, the influence of the social factor is more vital than that of the regional factor and the natural factor, in that order.

Fourthly, the population, GDP, and transportation accessibility have a more balanced intensity of influence on all types of leisure agriculture. One or two forms of leisure agriculture are significantly influenced by the natural factor, the distance to attractions and waters in the regional factor, the degree of financial investment, and policy support for culture and tourism.

Fifthly, the influence on the spatial distribution of leisure agriculture in the core city area is mainly oriented toward the water areas, scenic spots, and social factors. While farther away from the central city area, the natural factors and accessibility become the main influencing factors.