Classification and Spatial Pattern of Rural Multifunction from the Diversify of Construction Land Use: Taking Tongshan District of Jiangsu Province as a Case Study

The development of rural area in suburban (RAS) is comprehensively influenced by the mixed processes of non-farming, local urbanization and rural industrialization and has distinctive multifunctional features on the socioeconomic and resource utilization. Multifunctional reconstruction in the RAS is mainly manifested by changes in the quantity and structure of the construction land use (CLU). We took Tongshan District of Jiangsu Province as the study case. Following the logic of identifying, clustering, and diagnosing, we assessed the functional features of CLU and analyzed the differentiation of the scale and structure of CLU in RAS, with the aim of identifying the types and characteristics of multifunctional development in these regions. This study found that: 1) typical RAS were experiencing multifunctional space reconstruction, and their multifunctional value was emerging and highlighted. The CLU was rapidly expanding, and the types of CLU were evolving from single to multiple uses; 2) based on the dominant function of diverse rural CLU, we classified RAS into five types, dominated by the following functions, respectively: commercial services, industrial production, residential living, basic support and public service; 3) according to the dominant function and spatial pattern, the development types of RAS were classified into agricultural service, comprehensive services, and integrated industries service, showing a circular spatial pattern from the suburbs to distant suburbs. The results can provide policy implication by creating new development impetus from rural multifunction perspective to promote the RAS revitalization.


Introduction
With socioeconomic development, the tight connection between agricultural production and rural development was historically broken first in the developed western world. The interests on rural areas of different social groups have increasingly diversified, and the rural multifunction grown in prominence . Rural regions' function is not merely agricultural production; it extends to multiple levels and dimensions, especially under the postmodernism and post-positivist context. The concept of rural multifunction was pro-posed in the 1980s and is now guiding the rural transformation and development around the globe (Stola, 1984;Wilson, 2001;Holmes, 2006). Over the past two decades, research on multifunctional agriculture (Wilson, 2007), multifunctional land use (De Groot, 2006;Brouwer and Van Der Heide, 2009), and multifunctional rural transformation (Holmes, 2006) was popular in the West. The deepening of multifunctional agricultural studies promotes the development of rural multifunction. Rural areas' original function of agricultural production is supplemented with social and cultural functions, ecological regulation, and recreation. Moreover, this added multiple functional value becomes increasingly prominent with the development of industrialization and urbanization as it is needed more strongly by the emerging middle class in this stage of development. Rural multifunction serves as an internal driving force for rural development and it constructs a new theoretical framework to guide the rural transformation and development practice (Kurkalova, 2005). Thus, nowadays multifunction becomes the vision and idea for rural reconstruction .
Numerous studies on rural multifunction were conducted based on different scales and perspectives. After more than two decades of extensive efforts, abundant achievements were achieved in clarifying the theory, causes, and evolutionary process of rural multifunction (Wilson, 2001;2010;Holmes, 2006). Moreover, multifunctional rural development gradually become an important theory for assessing the evolution of rural areas in various countries. In recent years, the research is mainly focusing on three aspects: clarifying the connotation of multifunction in rural areas (Holmes and Argent, 2016), exploring the evolutionary process and dynamic mechanism of the transformation of each function (Willemen et al., 2010), and analyzing the interactions and relationships among different functions, respectively. Although domestic scholars started their research relevant late, their study achievement are abundant as well. In terms of research content, domestic scholars focus on the spatial differentiation characteristics of rural multifunction (Gu et al., 2019), the classification of functions (Xu and Fang, 2019), and the impact mechanisms of different functions (Li et al., 2014;Tan et al., 2018). However, the regulatory paths of rural functions are less studied Qu et al., 2020). Regarding research methods, a popular approach is determining the magnitude of the function indices and analyzing the aggregation and coordination degrees of these functions based on a rural multifunction evaluation system (Yang et al., 2019), supplemented with the Gini coefficient and the coupling coordination degree model (Wang and Tang, 2018). Regarding the research scale, most studies focus on meso and macro scale. While research on the micro village level is relatively weak (Jiang et al., 2021).
Rural area in suburban (RAS) is more affected by the non-farming, local urbanization, and industrialization process from multiple aspects due to its spatial proximity to cities. Multifunctional development is an inevitable trend in the transition development of these rural areas. Consequently, the function of the agricultural production-oriented space gradually evolves into a multifunctional space, with the elements of production, living, ecological regulation, consumption, and cultural tourism (Ma Xiaolong et al., 2019). Land use is essentially a mirror of socioeconomic activities. Different development stages and types correspond to different construction land utilization statuses and components. Transformations in socioeconomic development precipitate specific land use changes, on the contrary, the land use pattern reflects the characteristics of transformation in socioeconomic development. Rural construction land use (CLU) is a spatial carrier for the living and production of farmer households (Ma et al., 2021), and changes in its scale, components and structures mainly reveals the multifunctional development features of RAS. Based on the important role of CLU in reflecting rural development multifunction, and according to current land use status and pattern and structure of CLU in RAS, we established a theoretical framework to illustrate the evolutionary relationship between the rural multifunction and CLU structure, then summarized the gradient differences and characteristics of CLU in RAS and explored the multifunctional features and patterns in the RAS. We took Tongshan District in Jiangsu Province as the empirical study to analyze the spatial differentiation characteristics of the multifunction in RAS based on the quantity and structure of CLU. We then identified the dominant functions of each village and classified the type of multifunctional development in the RAS, aiming to provide a policy basis for promoting suburban revitalization and integrated urban-rural development.

Rural multifunction
The understanding and utilization of rural multifunction has gone through a process from simple to complex, and the rural multifunction classification system has gradually evolved from three basic functions of economics, environment and society to different scales of subfunctions (Stola, 1984;Knickel and Renting, 2000;Sánchez-Mateos, 2018). For example, it is generally believed there are five basic functions of rural areas include five types: agricultural production function, economic development function, social security function, ecological conservation function and cultural inheritance function (Long et al., 2022). Exactly, according to different research purposes, the refinement of classification has become the mainstream, and agricultural production (Prändl-Zika, 2008) and other subfunctions have been derived from the three basic functions. They formed the diverse and multi-scale classification systems. Among these, the most influential one (Helming et al., 2008) is to divide economics, environment and society into nine subfunctions based on the South African National Seed Organization plan. However, most of the above classification methods are based on the fact that the classification subject is agricultural productive villages. Most of the classifications are limited to concrete and intuitive land use methods, without considering the new multifunction such as commercial services (Argent et al., 2007), tourism (Vukovic et al., 2012), and lacking in unified classification standards. At the same time, due to the multi-scale of rural multifunction, it is necessary to further investigate the details of production and service functions when needing subdivision. RAS is a typical area for multifunction research particularly, so it is worth mentioning that in the process of urbanization and non-agricultural development, the multifunction of CLU can better reflect the multifunction of rural areas, while industrial, public services and other functions cannot be identified by rural agricultural land, also the relevant research results are still rare.
As mentioned above, the rural multifunction basically includes five types. However, specifically, for every category, there are lots of detailed functions and their interactive functions, needing further and substantive exploration rather than staying on the rough classifica-tion. For the RAS, there are many studies concerning its multifunction of leisure, tourism, eco-agriculture, even culture and education, while neglecting its urban-oriented functions. As a transition development area, RAS provides an ideal sample to investigate the spatial evolution of urban-oriented functions. Especially, the CLU data from the Third National Land Survey, which is so far the detailed land uses about rural socio-economic activities, makes this investigation possible. Therefore, we mainly classified the RAS with five major functions as following: commercial service, industrial production, residential living, public service, and basic support based on the diversity of CLU.

Evolutionary relationship between rural multifunction and CLU
The reasons why rural CLU was employed to reflect the rural multifunctionality were that: the rural CLU is the focus of socioeconomic activities, and the characteristics of rural socioeconomic activities including their multifunctionality were finally embodied on the rural CLU. There is some corresponding relationship between rural multifunction and CLU in Fig. 1. At the stage of traditional agriculture, rural residence was constrained by the farming radius and distributed around the agricultural production space. The quantity of CLU was less, and its structure was simple, with residential land in absolute dominance. At the stage of agricultural industrialization, land for facilities and services increased considerably. The agro-processing industry developed rapidly, and industrial land is gradually separated from residential space (Wang and Li, 2011), forming an independent production space. The quantity of rural CLU increases sharply, and the types and structures of land use also diversifies. The production space is clustered, and the service spaces become centralized. Additionally, the functions of recreation, wellness service, and ecological regulation emerge (Yao and Ma, 2019). The transfer of industries out of urban areas has accelerated the growth of rural CLU, making RAS also the function space for regional urbanization and industrialization. At the stage of developing urban and rural integration, non-agricultural functions are gradually enriched in rural areas (Long et al., 2022). With the rapid emergence of facilities for tourism and leisure, multifunction have been well organized (Zhang and Li, 2020). Under the continuous influence of urbanization and non-farming, the multifunc-tionality of CLU in RAS mainly manifests in the residential space which tends to be community-oriented. Also, due to the escalating urban demand for rural goods and services, massive CLU have been reconstituted by improving infrastructure and providing new service facilities, the functions and structures of CLU tends to be optimized . Multifunctional rural development provides a key entry point for sustainable rural transformation, while the demand for ecological products and tourism services accelerates the scale growth and structural optimization of CLU in RAS and promotes its process of urban-rural integration. Nowadays, the land use, spatial pattern, and functional structure of RAS in China are undergoing post-production transformation and diversification, thus, there is an urgent need to apply the theory of rural multifunctionality to guide the reconstruction of RAS (Fang and Liu, 2015).

Study area and data sources
Tongshan District is located in the northwest of Jiangsu Province, at the junction of the provinces of Jiangsu, Shandong, and Anhui. As a core part of the Xuzhou Metropolitan Area, Tongshan District is a typical suburban area surrounding about three quarters of Xuzhou's central urban region. Under the dual effect of urban and rural activities, the CLU in this region changed drastically, undergoing a multifunctional evolutionary process in RAS. The land use, industry type, and spatial structure all exhibit representative regional characteristics (Fig. 2). In 2010, the previous Tongshan County promoted to be a district. Along with administrative organizational changes on such a large geographical scale, many functions of the main urban areas of Xuzhou gradually shifted to Tongshan, resulting in a significant adjustment in the quantity and structure of CLU. The southern Tongshan District bordering the main urban areas of Xuzhou City undertook the previous city center' s many urban functions. After years of development, this region merged with the main urban areas of Xuzhou. However, as a whole, limited by the large administrative area of Tongshan District, the orientation of township development, and its resources endowment, the urban functions transferred from the central urban areas can not radiate throughout the suburban. The multifunctional development of the suburbs and the distant rural areas exhibits gradient characteristics. The CLU in Tongshan District covers 425.78 km 2 , accounting for 21.23% of the total study area. Rural residential land and industrial land account for a high proportion of CLU. The spatial gradient difference within CLU is significant. In recent years, influenced by the progress of non-agriculturalization, urbanization, and industrialization, rural CLU exhibited noticeable trends of spatial diversity and structural evolution. So, there exists an urgent need to analyze the spatial differentiation characteristics of multifunction in RAS and classify rural regions' development in order to optimize the functional layout of CLU in RAS and guide the development orientation, functional enhancement, and specialized development of rural areas. In this study, 317 administrative villages in Tongshan District were taken as units to demonstrate the spatial pattern of rural multifunction, and research data included the vector data of CLU and administrative borders retrieved from the 1 : 5000 land use map of 2019 (http://zrzy.jiangsu.gov.cn/xz/gtzx/bsyw/202108/t20210830_1119744.htm).

Functions of the rural CLU
To analyze the multifunctional evolution features of RAS, we referred to the National Current Land Use  Fig. 1 The evolutionary relationship between rural multifunction and construction land use (CLU) diversity Classification Standard (Ministry of Natural Resources, 2017), where the CLU is classified into Class I land uses, including commercial land, industrial, mining & storage land, residential land, public management and public service land, and transportation land, and their subordinate Class Ⅱ land uses illustration are also given, according to different functions. We then classified the RAS in sub-regions with five major functions: commercial services, industrial production, residential, public services, and basic support ( Table 1).
The structure of rural CLU not only reflects the differences in rural economic and industrial structures, but also characterizes the rural area's functional layout, revealing the level of rural development to a certain extent. In this study, the information entropy of various types of construction land in the rural area (Tan et al., 2017) was applied to measure the function indices, and for layout analysis, the spatial aggregation of different functions was evaluated with reference to the spatial Gini coefficient. We determined the dominant functions   Note: the numbers in parentheses in the first column of are the weights obtained by the information entropy method of each evaluation unit to generally identify the dominant functions of the rural area.

Function evaluation methods for the construction land in RAS
Information entropy is often used as a quantitative indicator of system information (Li and Guo, 2002). The quantitative structure of construction land in RAS characterizes different rural functions, which can be identified based on the information entropy of various types of construction land. The calculation formula is as follows: where H i is the information entropy of the i-th rural function, and a ij is the proportion of j-th type of land in the i-th rural function, and the relationship between land use types and rural function is in Table 1. The greater the value of H i , the more complex the of the i-th rural function; the smaller the value, the more monotonous the structure of rural construction land with regards to the i-th rural function. After eliminating the differences in index through standardization, we applied information entropy to determine the weight of each function index of the rural area (He et al., 2016) and then calculated the values for the functions of commercial services, processing and production, residential, public services, and basic support functions using the weighted summing method.
where x' ij is the standardized value of the j-th evaluation index for the i-th function; W j denotes the weight of the j-th rural function; S i represents the individual function, with a value ranging from 0 to 1, and the larger the value, the stronger the specific function. The indices of the five functions are summed to obtain the comprehensive function index of the rural area.

Spatial analysis method for multifunctional land use in RAS
This method evolved from the Gini coefficient for measuring spatial distribution. It is used to assess the degree of aggregation for the geographical distribution of various functions on village scale (Ma et al., 2007). Each function is treated as an object to be individually evaluated. The higher the value, the more concentrated the spatial distribution of the function. The calculation formula is: where S im denotes the value of function m in the i-th village, and S M is the total value of function m; n denotes the number of villages, and n = 317; G m represents the Gini coefficient of function m, with a value between 0 and 1, and the larger the value, the more aggregated the function, while the smaller the value, the more dispersed the function.

Identifying the functions of RAS
Differences exist in the manifestation and intensity of different functions in RAS. The dominant function reflects the regional characteristics, thus playing a critical role in rural development. Whether a function is the dominant one can be determined based on the relationship between each function value of the evaluation unit and the average value of that type of function (Xiong et al., 2021).
where C ij is the difference between the value of the i-th function of village j and the average value of the function. S ij denotes the value of function i of village j. S I is the total value of function i of the 317 villages. If C ij > 0, the evaluated function is the dominant one; if C ij ≤ 0, then that function is not in dominance. If there is only one dominant function in the evaluation unit, its development belongs to the single function-led type; if a village unit has at least two dominant functions, it is of an integrated development type; if there is no dominant function, it is of a lagged development type.

General characteristics of the construction land in RAS
According to the distance from the urban center, scatter diagrams reflecting the proportional structure of construction land are plotted with distance intervals of 2 km. The five types of construction land reflecting rural multifunction in the study area present the following characteristics: (1) Overall scale of construction land in the RAS: the construction land in Tongshan District has an overall scale of 425.78 km 2 , accounting for 21.23% of the total study area. In 6.62% of the villages, the proportion of construction land is less than 10.00%, and these villages are mainly located in the distant suburban; 59.62% of the villages have a construction land proportion in the range of 10.00%-21.23% and are mostly located in the middle suburbs; 29.97% of the villages have a construction land proportion in the range of 21.23%-50.00% and are mainly located in the suburbs near the central urban; 3.79% of the villages, most of which are the seats of the subdistrict offices, have a construction land proportion above 50.00%. Up to 66.24% of the villages in Tongshan District have a construction land proportion that is lower than the average. The further from the central urban, the lower the levels of urbanization and industrialization, the greater the proportion of agricultural land, and the more obvious of the rurality.
(2) Structure of construction land in the RAS: the structure of different types of construction land is highly heterogeneous, with significant differences among villages. The types of construction land are unevenly distributed, and the land use pattern is polarized. Residential land is still the dominant land use, with an average proportion of 52.32%. In some villages, residential land accounts for up to 86.65% of the total construction land. The proportion of commercial land is generally low, with an average of only 2.89%, and 21.13% of the villages have no land for commercial facilities. The average proportion of industrial production land is 21.19%, but its distribution is also polarized. In some administrative villages, the proportion of this land use is as high as 78.54%, but 11.04% of the villages have no industrial land. The proportion of public service land is relatively low, with an average of 3.85%, and 36.91% of the administrative units have less than 1.00% of land for public services. The distribution of infrastructure land is relatively homogeneous, with a consistent proportion of about 20.00%. Among all the evaluation units, there are 47 villages with less than three types of construction land, accounting for 14.83%; 156 villages have three to five types of construction land, accounting for 49.21%; 114 villages have six types of construction land, accounting for 35.96%. The more diverse the construction land, the higher the multifunction of RAS.
(3) Gradient differentiation of rural CLU: the proportional structure of construction land is affected by the distance from the central urban (Fig. 3). The peak distribution of different land use proportions appears in sequence spatially, with prominent gradient characteristics. Commercial land is closely related to land use multifunction. It is the first land type to show the peak pro-   Fig. 3 Gradient differentiation of construction land structure of villages in Tongshan District, Jiangsu Province. And the points represent villages, and different colors represent different clusters of villages classified by their distances to city center portion, accounting for more than 20.00% at a distance 10-12 km, and its proportion gradually stabilizes at distances farther than 22 km. The proportion of infrastructure land peaks at a maximum of over 30.00%, and the value decreases significantly between 10 km and 16 km and gradually stabilizes over 16 km, as the effect of distance is attenuated. Industrial land accounts for a high proportion within 14 to 30 km, and its proportion gradually declines after peaking at 18 km. The proportion of public service land does not show an obvious peak, and overall change is small. The closer to the central urban, the higher the proportion of public service land. The proportions of four types of construction land decrease in gradient with increased distance from the central urban. Only the proportion of residential land shows an increasing trend as the distance increases. The scale and structure of construction land can well characterize multifunctional development. The farther from the central urban area, the higher the proportion of residential land, the more monotonous the type of construction land, and the lower the multifunction of RAS.

Functional differentiation of the construction land in RAS
The individual function indices and the comprehensive function index of villages were calculated using Eqs.
(1) and (2), and the function indices were classified into five levels, from small to large, using the natural breaks method in ArcGIS. Using Eq. (3), it was obtained that the spatial Gini coefficients of the commercial service, industrial production, residential living, public service, and basic support functions are 0.231, 0.199, 0.189, 0.202, and 0.178, respectively. The spatial aggregation degree of the construction land with the five functions is generally low, indicating a dispersed layout, and the spatial distribution of these functions varies significantly (Fig. 4). The following results can be obtained: 1) driven by the radiation of the central urban, the highvalue regions with a commercial service function are distributed in clusters at the southwest of the central urban. Medium-and high-value regions are also scattered in the villages in the northwest and southeast of the district, where the township governments are seated. Commercial services are highly correlated with the levels of urbanization, industrialization, and non-agriculturalization. The closer to the central urban, the stronger the commercial service function.
2) The industrial function exhibits obvious gradient differentiation characteristics. The high-value areas are mainly distributed in a semiannular region surrounding the central urban. The northern region has a good industrial and mining base, and the industrial function there is generally stronger than that in the southern region.
3) The level of residential function is high in the northwest and southeast and low in the central part, and the high-value areas are mainly distributed in the distant rural regions. The closer to the central urban area, the more diverse the type of construction land, and the lower the dominance of residential land. 4) The index of the basic support the public service function is generally low. Medium-and highvalue areas are scattered around the central urban and in the villages where the township governments are seated, while the low-value areas are predominantly distributed in the northwest and southeast. Overall, the public service capacity is relatively low in RAS. 5) The highvalue public service function is distributed in a strip centered in the town of Tongshan on the periphery of the central urban. The southeastern part of Tongshan District is led by the economic and high-education development zones, with a high flux of transit traffic, and the basic support function there is significantly higher than that in the northwest. 6) The spatial distribution of the comprehensive function is generally characterized by a decreasing gradient from the urban center to the periphery, with a local pattern of 'one core and multiple points'. The core functions of rural construction land have three aspects: residence, production, and service. In this study, we categorized the commercial and industrial functions as production, the residential function as residence, and the infrastructure and public service functions as service. Thus, the five functions were categorized into three core functions (Fig. 5a), and the service function was further divided into three sub-functions: transportation services, cultural services, and administrative services (Fig. 5b). The ternary diagrams plotted intuitively reveal the overall structure and interrelationship of the three core functions and three service functions of rural construction land, where coordinate axis number represents the corresponding function shares of the total three service function.
Regarding the overall pattern of function combinations, the residential function still dominates, while the production function varies significantly among differ-ent villages, and the service function is obviously low in the study area. According to the agglomeration and distribution features of the residence, production, and service functions in the ternary diagram, structural differences and spatial interactions among the three functions in RAS of Tongshan District are apparent, exhibiting the characteristics of 'high residence function, low service function, and great variations in production function'. Specifically, 1) a large number of data points are clustered in the lower right region of the ternary diagram, with some data points scattered in the lower left region, indicating that the service function is highly correlated with the residence function but weakly correlated with the production function. 2) The residence function is still the dominant function, and the service function in RAS is basically to support the residence function.
Based on the distribution characteristics of the transportation, cultural, and administrative services in the ternary diagram, it can be concluded that the aggregation degree of these three types of services is low, showing the features of 'high levels of transportation services, balanced cultural services, and low levels of administrative services'. Specifically, 1) a large number of points are scattered in the upper left region of the ternary diagram, with a few points distributed in the lower left region, indicating that the transportation service function is generally high in RAS due to transit traffic  The ternary diagrams of production, service and residence service and transportation, cultural, and administrative service in villages of Tongshan District in Jiangsu Province, where coordinate axis number represents the corresponding function shares and that the cultural service level varies significantly among the villages because of the geographical location and uneven economic development.
2) The level of the administrative service function is relatively stable, showing a scattered distribution pattern. The differences among the villages are minor, but the proportion of this service is limited.

Classification of the dominant function in RAS
The dominant function plays a leading role in the future development mode of the region. We used Eq. (4) for function type identification to classify the dominant function of each village in Tongshan District, and a total of six types are classified (Fig. 6).
(1) There exist 71 dominant comprehensive multifunction villages, accounting for 22.40% of all units in the study area, with a spatial pattern of 'surrounding the main urban as a whole and scattered locally'. These villages' industrial, commercial services, and basic support functions are evenly developed, and the overall development level of these functions is all high. These villages are mainly distributed in the towns of Liuxin, Tongshan, and Tangzhang, surrounding the central urban. Some are scattered in and around the villages where the township governments are seated. This type of village has an excellent location, mostly sitting in suburban areas and subjected to intense economic radiation from the central urban area. Such units receive lots of investments in infrastructure and public service facilities, resulting in rapid urbanization and industrialization and desirable residential conditions. Hence, they are model areas for promoting integrated urban-rural development.
(2) There exist four types of dominant single function villages, accounting for 62.14% of all units in the study area, namely dominant commercial service function, dominant basic support function, dominant residential living function, dominant industrial living function, respectively. Significant differences exist in the spatial distribution of these four types of single function dominant villages. The proportion of the residential living function dominant village is the largest, with a total number of 114, accounting for 35.96% of all units. They are concentrated in the remote areas in the northwest and southeast, for example, the villages in the towns of Huangji and Yizhuang, with relatively low levels of urbanization and industrialization and obvious agricultural service characteristics. 39 villages are industrial living function dominant, accounting for 12.30% of all units. These are mainly distributed on the north of Tongshan District, for example, in the towns of Liuquan and Liguo. Owing to these regions' rich mineral resources, they have a good industrial and mining base and convenient transit traffic conditions, leading to the clustering of industrial space and the formation of an industrial area. 10 villages are commercial service function dominant, accounting for 3.15% of all units. These are mainly distributed in suburban areas, for example, the towns of Hanwang and Sanpu, where rural tourism resources are rich, the ecological environment is good,  (3) There are 49 balanced multifunction villages, accounting for 15.46% of all units. They are concentrated in the fringe regions in the northwest and southeast of the study area, for example, the towns of Shanji, Huangji, and Heqiao. It is weakly driven by the radiation of the urban center, the levels of urbanization and industrialization there are relatively low. Moreover, the functions are generally low, and there is no prominent leading function. Agricultural services dominate, showing a significant feature of rurality. The comprehensive development level of these areas is relatively low.

Village development model based on multifunction identification
The classification of rural types on multifunction is the prerequisite for planning to restructure rural elements, which aims to change the spontaneous and relatively disorder pattern of rural multifunction. Besides their multifunction in status, it also involves their locations within the urban-rural fringe, spatial interrelation like cluster, contiguous and sporadic, and their future development directions from the upper levels of planning (Li et al., 2018;Yang et al., 2021). Therefore, we comprehensively considered these factors and classified three types of rural development (Table 2): the agricultural service type, the comprehensive services type, and the integrated industries service type (Fig. 7).
(1) Agricultural service type: this type of village is mainly located in the traditional agricultural regions, and promoting agricultural industrialization is the first priority for its future development. It is necessary to change the mode of agricultural development, actively conduct characteristic agricultural operation, and compensate for the shortcomings of agricultural infrastructure. The variety, quality, and market structure of agricultural commodities should be optimized. Moreover, flower and seedling cultivation and picking and experience gardens can be deeply integrated with rural tourism to construct an efficient, stable, and high-yielding agricultural system. Additionally, eco-environmental protection should be a basic focus to steadily realize rural revitalization.
(2) Comprehensive services type: industrial villages have excellent resource endowment and a certain industrial base. Future development should focus on accelerating the transformation of traditional industries, actively developing the recycling economy, and enhancing the industrial function with the various industrial zones and logistics parks as carriers to improve the rural area's competitiveness. Commuter villages have significant advantages in terms of traffic location. Many workers live in these villages and commute to work in the central urban area, so these areas should further improve their transportation infrastructure and health and education services to make the residential space and ecological environment more desirable. Villages with a good  (3) integrated industries service type: the future development of this type of village should exploit their location advantage of closing to the central urban area and capture the opportunity of the spillover of industries from central urban area to outskirts. In these areas, the development of high-tech industries and a modern logistics industry should be emphasized and should integrate into the value chain of non-agricultural industries, enhance the foundation of regional industrial development, strengthen industrial division and cooperation, form an industrial system complementary to that of the central urban area, equalize public services in the urban and rural areas, and promote rural revitalization and integrated urban-rural development.

Conclusions
The multifunction of RAS is significant, while RAS in China are experiencing rapid function space reconstruction and as the carrier of diverse socioeconomic activities, the CLU reflects the multiple functions of rural area to some extent. We took the CLU of Tongshan District in Jiangsu Province as the study object and the administrative villages as evaluation units to explore the pattern of rural multifunction. Then the development mode of RAS was classified and the spatial differentiation characteristics of each type of development was analyzed.
The main conclusions obtained are as follows: (1) Variations in the quantity and structure of CLU can well characterize the evolution of rural multifunction. Construction land in RAS typically undergoes a period of traditional agricultural production when residential land dominates, followed by a period of agricultural industrialization when residential land and industrial land are separated, and finally, a period of urbanrural integration when the type and function of construction land become diversified. As the carrier and intermediary of multiple social and economic activities, the construction land has spatial heterogeneity, thus changes in the CLU indicate that the land use, spatial pattern, and functional structure of RAS in suburban are undergoing a multifunctional development transition.
(2) The RAS in Tongshan District have various function combinations, showing a significant differentiation and multifunctional character. Due to differences in the rural areas' resource endowment, development policy, internal driving forces, and external environment, the villages are classified into six types: dominant commercial service function, balanced multifunction, dominant basic support function, dominant residential living function, dominant comprehensive multifunction, dominant industrial living function, respectively. Most of the units are dominated by a single function, and the type of units with multifunctional integrated development is less. We need to optimize the function space pattern by compens- ating for the weak functions, strengthening the leading functions, and finally improve the comprehensive function in order to realize urban-rural integration.
(3) The development mode based on multifunction identification of RAS in Tongshan District was classified into three types: the agricultural service, the comprehensive services, and the integrated industries service type, based on the dominant function types and according to the functional advantages and their future development direction. Overall, the types of development in RAS in Tongshan District exhibit a spatial differentiation pattern, featuring small circles engulfed in large circles. Villages classified as integrated industries service, comprehensive services, and agricultural service appear sequentially from the suburb to the distant suburb. Discontinuous and leapfrogging development was also observed in and around the villages where the township governments are seated, with small circles of integrated industries service, comprehensive services, and agricultural service types observable from inside to outside.

Discussion
Previous studies have classified rural multifunctional into production, living, culture and ecology categories, while without examining the spatial evolution of subcategories of rural multifunction, and ignoring the growth process of potential multifunction (Long et al., 2022). With the perspective effect of CLU on social and economic activities, this study investigates the commercial service, industrial production function, residential living function, basic support function, public service function in the detailed multifunctional category. This expands and deepens the study of rural multifunction, and makes the study focus on more sub-categories of multifunction, such as tourism agriculture and technology agriculture among agricultural functions, which undoubtedly enriches the connotation of rural multifunction theory. This study further proposes a multifunctional-based sustainable development model zoning based on the analysis of the multifunctional status quo to promote the practical application of rural multifunctional theory and its integration with spatial development strategies. It has been found that there is a gradient distribution pattern of production, living multifunction in the countryside around big cities, and this study further finds that the same gradient pattern exists for the more detailed industrial production functions and living functions. However, this is different from the general gradient pattern, it shows a gradient fluctuation pattern, which may be the result of the combination of the urban distance factor and the geographical location of the various township.
The main novelty of this paper is that we focus on the RAS which has the typical multifunction characteristic, and employ the refinement of CLU to describe rural multifunction. Since data of the detailed CLU classification was unavailable before, such research was neglected. Here, we used the high-resolution data of refined land use classification to investigate the pattern of multifunction of RAS. Also, we proposed the regulatory policies for rural multifunction from the perspective of CLU. There are also some deficiencies needing further probe. How to evaluate and measure rural multifunction is a key issue, here we employed the construction land as an intermediary to characterize the diverse rural socioeconomic activities, however, due to the lack of multi-temporal land use data and the statistical data of village-scale, the rural multifunction may not be completely excavated. Actually, data deficient is always a significant challenge for rural studies, because village is not the official statistical unit in China, and the observation of rural elements is unappreciated. The big data like mobile signal data, POI data and social media data, which can be available from unofficial way, provides new data for rural research. So, following further studies can be conducted: firstly, multi-temporal land use data can be joined to excavate the spatiotemporal variation; secondly multisource data should be combined to measure rural multifunction precisely; thirdly, the formation mechanism of influencing factors of the rural multifunction pattern should be explored.

Conflict of Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.