Multifunctional rural transition along China’s Yangtze River Transect: Spatial differentiation and economic relevance

To understand rural sustainability, it is necessary to scrutinize the relationship between rural transition and economic growth. The article uses rural multifunctionality as an analytical lens through which to view the processes of the development of rural occupancy. There is a pressing need to ascertain how to quantify rural multifunctionality and reveal its spatial differentiation, as well as garner and investigate how multifunctional rural transition (MRT) responds to economic growth. This paper employed the concept of sample transect to compensate for data deficiencies in a long temporal series and established the indicator system from three different aspects-living function, production function, and ecological function-to measure MRT along China’s Yangtze River Transect. Our analysis showed that living function and production function display an increasing trend from underdeveloped western regions to eastern economically prosperous regions, and represent a high degree surrounding urban agglomerations, while economic growth only leads to a statistically insignificant decreasing trend in ecological function. The MRT resulting from multiple factors is much diverse, complex, and sophisticated; therefore, it should be understood within a framework incorporating both endogenous and exogenous factors. According to the results, it is thus important to formulate differentiated managerial countermeasures corresponding to the economic development level rather than the uniform regulations.


Introduction
Rural, a conception initially flat and uni-dimensional, has been made complex and ambiguous (Wilson 2001), especially in rural areas of the developed or rapidly urbanized economies (Murdoch 2000;Woods 2010). In the mixed process of urbanization and counter-urbanization, the new incomers including tourists, investors, and local residents enrich the demographic composition. Besides, the mixtures of human and non-human elements consisting of plants, animals, natural landscapes and socio-economic activities jointly constitute the hybridity of the countryside. The integration of agricultural production, processing, and manufacturing industry, as well as the business service industry, has altered the traditional production-centered activities. This process has restructured the production and living pattern, and has also driven the changes of rural environment. To clearly sketch the plurality of rural areas, "rural multifunctionality"-an integrated theoretical perspective-grows in popularity amid the strands of intellectual explanations to the rural changes. It derives from the research about the function variation of agricultural systems (CEC 1988), which is of great significance for sustainable rural development (Renting et al. 2009). Still, there remain reflections indicating that the broader dynamics of rural change were neglected as agriculture production is currently only one of the multiple functions rather than the central of rural (McCarthy 2005). Therefore, under the influence of hybridized theory (Cloke and Goodwin 1992), theoretical pluralism (Murdoch 2000) etc., the thinking of multifunctional agriculture critically evolves into another recognition which treats the multifunctionality as an increasingly important attribute in overall rural occupancy (Holmes 2006(Holmes , 2012.

Responsible Editor: Roula Inglesi-Lotz
Rural multifunctionality performs a non-absolute and neutralized response oscillating between agriculture-centered "productionism" and the "post-productionism" (Murdoch and Pratt 1993) with anti-agriculture philosophy (Wilson 2001), as it treats agricultural production as a requisite but deems it insufficient to portray the whole picture of rural. This means that there exists rare direct link between rural multifunctionality and traditional agricultural production, and multifunctional rural development might be at the expense of or concert with agriculture production (Gelats et al. 2015). It is noteworthy that "rural multifunctionality" is distinguished from "multifunctional rural" which is seemingly an ideal post-production pattern and contains some imaginative elements. Rural multifunctionality is prone to be a developing state of rural regions, fully embodied in a more specific and tangible transition concluding with various positions of incomplete changes. This transition is also the process of liberating the concept of rural from its previous close connection to a g r i c u l t u r e ( H o l m e s 2 0 1 2 ) . G e n e r a l l y , t h e "multifunctionality" integrally represents the rural "diversity, non-linearity and spatial heterogeneity" (Almstedt et al. 2014), constructing a networked and innovative fabric of rural, yet differs from another definition of "rurality" which assembles overcomplicated and metaphysical elements (Woods 2010).
The course of history in western rural has witnessed that the multifunctional rural transition (MRT) was inextricably knit with the economic background. At the beginning of post-war economy recovery in the 1950s, the western rural centered on agricultural production for a long time (Wilson 2001). When finishing the economic capital accumulation, the occurrence of non-farm large-scale enterprises was reported in some developed countries, changing the Fordist rural production pattern (the so-called Fordist transition) (Cloke and Goodwin 1992). Since the 1970s, cleaner agricultural production through contamination control was enshrined in laws (e.g., U.S.EPA 2013) and campaign-style rural beautification (Thomas 2020) was launched in some leading western countries, which endows rural development with ecological protection contents. In the following decades, rural was shaped into a pastoral idyll, triggering the rural gentrification action of "counter-urbanization" (Phillips 2010). That is, the urban middle class spontaneously moved to the hinterlands, which to s o m e e x t e n t i n d i c a t e d t h a t t h e f u n c t i o n o f consumption and recreation gradually became prevalent in rural regions.
Following the path of theoretical development, empirical research emerges from agriculture sector. Relevant studies mainly contributed to the assessment and development of agriculture multifunctionality (Kizos et al. 2011;Wilson 2008), landscape multifunctionality (Gulickx et al. 2013), land use multifunctionality (Siciliano 2012), etc. As the theory evolves and enriches with rural variation, a substantial body of research and commentary provided a more complete articulation of the concept of MRT (Holmes 2010). The transition of agrarian issues including multifunctional livelihood, planting structure (Martin and Lorenzen 2016), farmland use, and rural settlement (Qu et al. 2017) was further investigated to portray the plurality of rural change. Nevertheless, there are three insufficient points of these researches. One is that the developed economies are overfocused (Cui et al. 2020), while the similar but delayed functional changes in developing countries were drawn less research attentions. Second, limited available works (Yang et al. 2019;Yang et al. 2020) have carried out comprehensive analysis towards rural multifunctionality, but mostly remained pragmatical evaluations based on the given administrative units, revealing a distinct paucity of theoretical exploration. Moreover, rare existing literature shed light on MRT, and, moreover, how MRT responds to regional economic growth remains unclear.
To address this gap, this paper adopts multivariate analysis which is one of the common techniques to deal with indicators, and draws China's Transect of Yangtze River as study area. The river transect not only demonstrates gradual economic characteristics but also presents obvious differentiation of natural conditions, which determines the production and lifestyle pattern. Therefore, taking the Transect of Yangtze River as the research object is helpful to further understand the spatial pattern of rural multifunctionality and its reciprocity with economic growth, through which the general mechanism of MRT can be further disclosed. The remainder of this paper is organized as follows: "Multifunctional transition in rural China: A policy perspective" reviews the historical trajectory of MRT in China from the perspective of policy transformation. In the "Hypothesis" section, the hypothesis is put forward based on theoretical analysis. Then, this paper proceeds with a description of the study area and methodology. "Results" elicits the results. Discussions are provided in the "Discussion" section. The last part presents and concludes our findings.
Multifunctional transition in rural China: a policy perspective MRT in contemporary China is determined by socioeconomic level and national policy. It has experienced a spiral growth course that can be divided into four phases ( Fig. 1):  The new founded People's Republic of China once made efforts to promote the transformation from traditional rural society towards the process of industrialization; therefore, the prices of agricultural products were artificially lowered to secure the capital accumulation of heavy industry Wang et al. 2014). In light of such implementation of agricultural "price scissors" institution, the unsubstantial agricultural foundation and even the "rural malaise" were shaped at this stage.
Phase II (1978II ( -1993 Since the reform and opening-up policy in 1978, the de-collectivization, which aimed to endow private sectors with well-delineated property rights, caused the disillusionment of the Maoist model that nationalized all production resources by establishing cooperatives and unified marketing (Bing et al. 2018). With the help of land adjustment and foreign capital, the liberated individual economy fueled the townships and village-owned enterprises (TVEs) covering various fields (Wang et al. 2008), such as agriculture, industry, commerce, construction, transportation, and service sector. In the golden age (mid-1980s to mid-1990s), the total output value of these rural enterprises exceeded that of agriculture, and the swift annual growth rate of TVE employees amounted to 9% (Qu et al. 2009). The MRT germinated during this period.
Phase III (1994III ( -2004 In 1994, tax sharing reform revamped central-local revenue-sharing arrangements, achieving a recentralization of the fiscal system, but causing the local governments to fail in making ends meet (Rogers 2014). Making matters worse, the industrialization and urbanization profited from the rural land transaction through legitimated approaches conferred by Land Management Law. Also, increasing competition from both overseas and domestic markets promoted rural enterprises to be restructured or collapsed. The rural area was no longer able to act as an ideal place to hold labor forces (Qu et al. 2009). Hence, millions of farmers migrated to economically prosperous areas. Rural industrialization peaked in 2000, while this prosperity was gradually overshadowed by the tremendous rural outmigration , thus delaying the MRT.
Phase IV (2005-present) Since 2004, the central government refocused on rural recession, which was highlighted in the No. 1 Central Document . Then, the Eleventh Five-Year Plan claimed a new rural revival strategy-Construction of New Socialist Countryside (CNSC). It aimed to renovate rural areas with the guidelines "production is developed, life is high-standard, the environment is clean, and management is democratic" (Rogers 2014). The contribution of CNSC to rural multifunctionality lies in encouraging the diversification of the rural economy, improving the environment and life quality. In the following years, coordinated urban-rural development has become a major political issue. These central government's documents successively addressed the following: terminating agricultural tax, increasing farmers' incomes, developing modern agriculture, advancing the "Building Beautiful Countryside" scheme (Rogers 2014), etc. Issues related to prospering rural areas enhanced production and living functions, and at the same time, underlined the protection of agro-environment (Cai and Yu 2018). In 2012, another new strategy named New-type Urbanization was proposed to ameliorate rural degradation and promote rural sustainability . The guidelines were updated to narrowing the gaps between urban and rural, and realizing urban-rural integrated development. The central government's goals transformed from GDP-centered development to ecofriendly growth, highlighting the integrated development of different industries relying on rural environment and resources . In 2017, Rural Vitalization, with the guidelines of "thriving business, pleasant living environment, social etiquette and civility, effective governance and property," was implemented to prevent rural declines and to promote rural transformation development (Long et al. 2019). Such external supports delineated the trajectory of rural development, which gradually accelerated the MRT.

Hypothesis
The above discussion regarding China's rural transformation development has revealed that relevant policies, which are parallel with sustained economic changes, are highly coupled with MRT. Western experiences suggest that the effect of economic growth is not only influenced by the chronicled background but also determined by the regional discrepancies (Râmniceanu and Ackrill 2007). Therefore, rural multifunctionality needs to be considered from both historical and geographical perspectives.
Regional economic development stimulates the population aggregation and attracts more migrants to get employed in urban regions through considerable revenue, reliable public services, high transportation accessibility (Kilicoglu et al. 2020a(Kilicoglu et al. , 2020b, etc. In turn, the improvement of life quality corresponding to the regional development stage would give rise to various demands in terms of diet changes, tourism, leisure, and exquisite environment, as well as the yearning for a peaceful life. Generally speaking, in the underdeveloped areas, subsistence production is emphasized, while the demand for enjoyment and self-development in the developed areas is higher, which follows Maslow's hierarchy of needs (Sirgy 1986). Building on this rule, an inference can be drawn that economic growth is associated with urban residents' demands for rural areas, further driving the MRT.
Regional development is highly correlated with the regional location advantage, which affects the industry distribution and development direction. High-efficiency enterprises tend to be arranged in coastal regions of China, which attracts more investments and has high mobility. Following the rule of industry transfer, industries with high pollution or low valueadded would be moved to inland regions. Therefore, there exists a stratified development stage from coastal to inland regions (Martin and Ottaviano 1999). As discussed above, the driving effect of regional economic growth on rural transformation is expected to be highly heterogeneous.
Owing to the increasing land demand, economic development inevitably compresses the environmental space. Urban sprawl has resulted in that tremendous ecological land was encroached. Inverted U-shaped relationships could be observed among economic growth and arable land occupation (Song 2014), deforestation (Zhang 2000), wetland destruction, nature reserves overexploitation (Ke et al. 2011), etc. Specifically, the first halves of these curves commonly exhibit strong antagonism between economic development and environmental protection. After reaching the turning point, enhanced environmental awareness and advanced technology might be helpful for environmental conservation (Ricci 2007). This implies that the general development mode of "treatment after pollution" in China is indeed subject to regional disparities. Nonetheless, predatory development is still a major threat to rural sustainability. Based on the above arguments, hypothesis 1 is proposed as follows: H 1 : Rural multifunctionality would be strengthened with regional economic growth.
Rural multifunctionality is characterized by the endemic, intense conflicts among multiple values, most commonly occurring in peri-metropolitan zones (Holmes 2012). It implies that the urban area should be taken into account when understanding the rural multifunctionality. One reason is that urban area serves as the role of node and hinge and influences the surrounding rural area (Marshall 1989;Zhang et al. 2019). This kind of impact can be viewed as a gravitational interaction, which changes with the size and distance of the urban and rural area. Another explanation is that rural areas provide goods and services for urban residents (Gu and Pang 2008). To avoid traffic congestion and environmental pollution, a growing number of large energy-consuming industries move to rural areas, which endows the rural area with the industry sectors. Meanwhile, a vast amount of labor force migrates as the industries transfer and relocate around the industrial park in the suburb, stimulating the emergence of buildings and other infrastructures. Additionally, the urbanization and industrialization result in diet change and the emphasis on food quality. Among them, the demand for more nutritious food promotes the development of modern agriculture, while the need for daily leisure and short-distance travel accelerates the development of the tertiary industry, such as rural homestay. In contrast to urban, rural is a representation of healthy, relaxing, and recreational paradise (Woods 2010). Accordingly, we hypothesize that, H 2 : The near-urban rural areas display a higher multifunctionality level than those of remote wildernesses.

Study area
The east-central-west gap is one of the most typical regional differences in China: the east leads in economic development, the central comes second, and the west is in the last place. Yet the highly unbalanced east-west economic development differences are intertwined with the unmatched south-north differences, which undoubtedly disturbs the correlation analysis. Considering this, the Transect of Yangtze River (TYR) is chosen to investigate the links between economic growth and rural multifunctionality, which can also compensate for data deficiencies in a long temporal series.
As depicted in Fig. 2, the TYR continuum covers 343 counties, which are inhabited by more than 1/3 of the Chinese population. Its total length is about 3600 km starting from Tibet, and ends in Shanghai, passing through Qinghai, Yunnan, Sichuan, Chongqing, Hunan, Hubei, Jiangxi, Anhui, and Jiangsu, and covering the eastern, central, and western regions. The TYR has been proven to be a useful tool to reveal regional differences in physical, environmental, social, and economic aspects (Long et al. 2007). As the core region of the Yangtze River Economic Zone defined in the 13th Five-Year Plan (The State Council 2016), the TYR is more meaningful for rural renaissance and new-type urbanization.

Index system and data source
Previous studies identified inter-related aspects and indicators about MRT, such as agricultural production, food regimes, and environmental impacts, which are adopted into the conceptual model of "production-consumption-protection" (Holmes 2012). Yet rural modes of occupancy in the west have limited relevance to China's rural areas, as the dissonance in the social system, economic development stage, land management mode, etc. Building on the review of relevant literature in western countries, we constructed a framework from the dimensions of "livingproduction-ecology" by considering the characteristics of China's rural areas (Yin et al. 2017), which can be encapsulated as three dimensions (Table 1).
Living function mainly refers to the capability of providing living and leisure space for human beings, as well as satisfying material life and spiritual demand. The former is embodied in material goods, such as residents' living space security, income, and consumption level, while the latter is reflected in the supply of corresponding supporting facilities and services in the aspects of culture, education, etc. Accordingly, the relevant indicators (LFI) are selected. Then, the production function index (PFI) takes the industrial structure as an essential dimension, which includes direct and indirect production activities. The direct one mainly relies on farmland to obtain material products such as grain crops, cash crops, and livestock products, while the indirect one depends on land construction for achieving goods and service products such as the second and third industries. Finally, the ecological function is the natural surfaces of land use, which plays a role in maintaining ecosystem stability and the capability of self-maintenance. Normalized difference vegetation index (NDVI) (Pekkan et al. 2021), topological slope, and the value of ecosystem services are chosen for measuring ecological function. The NDVI and topological slope are sometimes replaceable for direct measurement of environmental conditions as the latter requires substantial data with multiple sources.
The data is derived from the "China County Statistical Yearbook (National Bureau of Statistics 2015)," and the yearbooks of those 11 provinces, as well as the annual statistics from the governmental reports. Other data on roads, terrain, and vegetation coverage, etc. were collected from Data Center for Resources and Environmental Sciences (CAS 2015). The spatial distribution of each index is shown in Fig. 3.

Weight calculation and quantifying rural multifunctionality
Weight determination is a pivotal process when assessing each sub-functions. Although there exist   several different methods that have been widely used to set the weights, error or deviation sometimes occur. To obtain more accurate results, three methods consisting of the Analytic Hierarchy Process (AHP), Delphi method, and Entropy method are used to compute the weights (Fox 2017). A multi-factor comprehensive evaluation was employed to calculate each sub-function and the comprehensive index of rural multifunctionality. The formula is as follows: where RMI is the rural multifunctionality index, I i represents indicator i, and W i indicates the weight of indicator i. The standardized indicator values range from 0 to 1, and the calculated indexes were finally standardized by using the max-min method.

Exploratory spatial data analysis (ESDA)
ESDA is a descriptive and visual-spatial distribution technique to identify and distinguish geographic data features. On this basis, outliers are then observable for the follow-up spatial associations' discovery and clustering, or spatial hot spot patterns. The indicators of Global Moran's I and Local Moran's I are used in this paper for analyzing spatial agglomeration characteristics of rural multifunctionality. The former is a commonly used metric, essentially a standardized spatial covariance, to evaluate whether there is a significant spatial autocorrelation from the overall level. The latter is used to identify and analyze the local spatial differentiation pattern. They were accomplished by using the following expressions: Fig. 3 The spatial distribution of the selected variables. The Quantile method was used to classify the variables where I and I i correspond to Moran's I and Local Moran's I respectively; N depicts the unit umber of researched counties; W ij is the spatial weight matrix; x i and x j denote the RMI of county i and j; x is the average value of RMI; s x 2 means the variance.

Spatial characteristics of rural multifunctionality
As diagrammatically represented in Fig. 4a, the high-value areas of LFI are chiefly concentrated in the middle and lower reaches of the Yangtze River, especially the Yangtze River Delta. And the Chengdu-Chongqing Economic Circle also presents a relatively high living function. This phenomenon roots in the location advantage, which creates more opportunities and provides more resource endowments for industry development, naturally attracting more people to get employed. These highly urbanized regions have a strong capacity to offer local residents desirable medical care, transportation accessibility, and social security, presenting high consistency with that in Fig. 3. The spatial distribution of the high-value area of PFI is generally similar to that in those areas with the prominent living function, but Anhui and Hubei segments are out of the rule. It is mainly attributed to the fact that these regions are located in the middle and lower reaches of the Yangtze River Plain, with a dense population and high ratio of agricultural sector, but underdevelopment of secondary and tertiary industries. Besides, the areas with high EFI are concentrated in the hinterlands of the urban agglomerations. The less human activities and the higher vegetation coverage there provide better ecological services. Nevertheless, these hinterland areas located in mountainous areas with steep terrain are unconducive to infrastructure construction, industrial agglomeration, and urban development, leading to the growing pains of living and production functions. It should be noted that the headwater area of the Yangtze River is located in the Tibetan Plateau with an average elevation exceeding 4000 m, with sparse population and poor vegetation coverage. Thus, the rural functions there in all aspects are at the lowest.
The zone with high RMFI is mostly distributed in the Yangtze River Delta, Yichang-Jingzhou City Circle, and regions surrounding Kunming City. The highest economic and social function in the Yangtze River Delta raises its overall score, while the areas around Yichang-Jingzhou and Kunming weigh the higher eco-function and medium-high living and production function indexes to achieve the high gross value.
Moran's I is 0.50, indicating a significant agglomeration effect of RMFI. Specifically, HH (high-high units are located in the first quadrant, the counties with high RMFI surrounded by that with high RMFI) is mainly distributed in the Yangtze River Delta and Yichang; the majority of HL counties (high-low counties and located in the fourth quadrant, the counties with high RMFI surrounded by that with low RMFI) is scattered in the southwest of China; LH (low-high counties are located in the second quadrant, the units with low RMFI surrounded by that with high RMFI) units are in the minimum amount, appearing as two clusters in the middle-lower Yangtze Valley; the LL (low-low counties are located in the third quadrant, the cities with low RMFI surrounded by that with low RMFI) units are widely spread in the western Sichuan Province and Qinghai Province (Fig. 5), which almost coincides with the areas with high EFI.

The relationship among the sub-functions
It can be concluded from Fig. 6 that there exist a strong synergy and positive correlation between living function and production function, whereas both living function and economic function showed weak negative correlations with ecological function. It can be easily understood that a prosperous rural economy is always accompanied with a strong capacity of housing dense population and providing sound public service. In some cases, plain regions, where more rural population inhabits, present high living function, while the production function is not quite high as the nonagricultural production activity is stagnant. This can be manifested in the major grain-producing regions in Anhui and Hubei. The hinterlands of urban agglomeration, including Yangtze River Delta city clusters and Chengdu-Chongqing metropolitan area, undoubtedly demonstrate intense socio-economic activities, but always are encroached by accelerating urban sprawl. Thus, the ecological space will shrink, and this can signify that the tradeoff exists between socio-economic performance and ecological conservation. Conflicts of construction land and ecological land emerged as prominent in eastern regions, which are preferentially developed in terms of both national and regional strategies. The abovementioned empirical evidence shows that there exist strong tradeoffs between socioeconomic function and ecological function, which is consistent with the study conducted by Duan et al. (2018). The negative impact of regional development on ecological environment resulted from the irrational resource utilization, especially land resources. The encroachment of farmland and the decline of forestland are caused by the increase of rural construction land, which gives rise to the shrinkage of ecological space. This implies that most of the counties are still in the primary development stage and the coordinated development of socio-economic activities and ecological protection has not been realized.

Coupling relationship between economic growth and rural multifunctionality
To establish the interrelations between economic growth and rural multifunctionality, and substantiate our abovementioned hypothesis, correlation analysis was used (see Fig. 7). Results show that LFI, PFL, and RMFI demonstrate significant positive relations with PCGDP, and the R 2 exceeds 0.5, which signifies LFI, PFL, and RMFI gradually increase as the regional economy advances. While, economic growth only leads to a statistically insignificant decreasing trend in the ecological function. The possible explanation of this phenomenon is that intense contradiction has yet to appear in underdeveloped regions, and the socio-economic development level is relatively low. In economically developed areas, land use is highly contested with limited space that can be exploited. Thus, the relations between socio-economic function and ecological function, to some extent, are determined by regional ecological carrying capacity. Figure 8 depicts the spatial tendency of different indexes along the YRT, and the evolution trend is in line with the results in Fig. 7, proving the assumption from the spatial perspective. Generally, the LFI, PFI, and RMFI experience wavelike increases from the upper reach to lower reach, peaking at the surrounding areas of urban agglomeration and bottoming at remote wildness. As for LFI, it first increases, reaching a peak around Kunming, and then experiences a slight decrease until Chongqing, and finally proliferates. Living function is linked with population distribution. The central cities naturally have large hinterland with more rural residents. For instance, Wuhan has a rural population of 2.19 million people. In view of PFI, there is a slight increase in Qinghai and Chongqing, after which it goes up dramatically. In western China, the harsh natural environment inhibits the industry development and the population carrying capacity is low. Thus, the western rural regions are still in the lagging stage of multifunctional rural development. Differently, EFI presents drastic fluctuations over the whole transect but generally shows a decreasing trend. It is obvious that the ecological condition varies along this transect, which causes the spatial differences of the EFI. Besides, the development intensity differs, and it makes the resource utilization intensity different, thus making the ecological disturbance different. These factors jointly contribute to the irregular variation of EFI. RMFI remains stable at approximately 0.35 between Qinghai and Guiyang, and then roughly displayed a considerable growth.

Insights into the spatial heterogeneity of rural multifunctionality
The unprecedented process of urbanization with a remarkable urban population boom results in an increasing demand for rural goods and services, driving the multifunctional transition in the suburbs. Our study provides evidence of multifunctional rural development in urban suburbs based on the quantitative evaluation, which is consistent with the findings of Gu et al. (2019). This work reveals the "urban-driving" effect of MRT at large scale covering different economic development stages, rather than focusing on a specific urban back-land with limited areas. The increasing trend of living function and production function identified along the TYR also validates the hypothetic associations. It has been recognized that MRT mainly results from external powers, including industry development, regional development policy, and foreign investment. There is no doubt that the impact caused by local characteristics cannot be ignored. Thus, the determinants can be encapsulated as endogenous and exogenous factors. The former incorporates the forces that add fuel to rural development momentum and stimulate the external demand of rural goods and services. Urbanization and industrialization are the general schematics of external powers from urban areas. This echoes Lewis' dual development theory, which underlines that the main role of rural economy is to produce products for urban consumption (Lewis 1954;Qin et al. 2020). The latter contains resource endowments, landforms, climatic conditions, and infrastructure construction, as well as the location conditions. Following the neoclassical growth theory, resource endowment and infrastructure have been shown as significant factors, and have the most explanatory power for rural development (Aggarwal 2018). It can be easily proved in western regions with fragile ecological conditions that socio-economic activities are not only greatly constrained by highland climate, steep terrain, low traffic accessibility, etc., but also limited by low urbanization level, sparse population, and deficient investment. Therefore, rural multifunctionality is a combined result of multiple factors and should be understood within an intricate network comprising both endogenous and exogenous powers. The spatial heterogeneity of rural multifunctionality Fig. 6 Graph matrix of the indexes of three dimensions revealed by this paper is in sync with the basic organization of a land use system that the traditional agricultural location theory depicts (Fig. 9). Production cost (including labor input), transportation cost, and land rent play a crucial role in the land use pattern Thünen outlined (O'Kelly and Bryan 1996). The technic progress in transportation and the increasing multiple demands of urban and rural residents have changed the determinants of industry distribution, especially in the suburb regions. In line with the traditional location theory, the labor-and capital-intensive planting industry occupies the outskirts of the city, which targets at providing food-oriented products, such as vegetables and fruits. Current improved transportation conditions enable the location selection of cash crops more flexible. The escalating demand of leisure and recreation creates more opportunities for the suburb regions to transform the traditional planting or breeding industry into a combined industry morphology integrating the first, secondary, and tertiary industry. The land use pattern is accordingly reorganized and intensive agriculture may be arranged further than ever before.

Implications for multifunctional rural planning
The results revealed the differentiation of the three dimensions of rural occupancy and identified the hotspots of each dimension, which supported the statement of Iker et al. (2018) that it is not an easy task to satisfy numerous demands within one specific region. This denotes that the appropriate spatial arrangement of functional topology is necessary according to the location and economic development level. Maslow's hierarchy of needs provides an appropriate theoretical perspective of explaining the functional heterogeneity in different regions with varying development stages. Demand change is a dynamic process, which is in line with socioeconomic development stage. The rural development goals in developed areas are under transformation towards common and comprehensive rural development objectives. Substantial parts of the previously direct supporting for agricultural production are transformed to multiple and diversified rural industries. Understanding the coupled relations between economic growth and demand evolution is the basis for making out the dominated function arrangement of rural regions (Cetin 2015). The results show clear differences between areas with different trajectories or rural traits and help identify the principal function of a different region. Therefore, counties located in the outskirts of megacities are supposed to strengthen the living and production function, while emphasizing the protection of ecological space. Strict restriction of farmland occupation should be enforced. Relying on a favorable environment, unique countryside landscape, and the huge market from urban residents, rural tourism might be a better choice for achieving rural sustainability. It has been found that there exists tremendous low-efficiency use of industrial land in the suburbs of Shanghai (Gu et al. 2018). Countermeasures should be taken to reduce inefficient industrial land use to improve the rural ecological environment. Concerning remote underdeveloped counties, promoting agricultural modernization and reinforcing ecological conservation should be underlined, as population outflow is an inevitable trend against the context of accelerated urbanization. Furthermore, regional sustainability requires the collaboration of different areas, which means some counties take living and production as their main responsibility, and other regions need to take full use of their natural conditions to highlight the ecological value, thus realizing regional multifunctional development.
In China's current rural planning, de-agriculturalization has been a pervasive ideological trend, which attempts to separate rural areas into construction land and agricultural production land, and seldom considers the impacts of agricultural land on rural development. In most cases, agricultural land is regarded as ecological matrix; however, its social and cultural functions are ignored. This implies that an understanding of rural development land use as a multifunctional practice is necessary, which is connected with a call for "integration" and "diversification" (Mölders 2014).

Limitations and outlooks for future research
A multi-factor comprehensive appraisal method was employed in this paper to evaluate rural functions, while several major uncertainties should be highlighted when using this method. First, there are no uniformly prescribed guidelines for selecting the most representative criteria. Secondly, thresholds for the overall indicators are unavailable in existing studies to support setting those for the indicators. Besides, this becomes difficult as they vary with regional conditions and socioeconomic development levels.
Although the transect was proved to be an effective tool for compensating data deficiencies in a long temporal series, the unavailability of data denoting detailed sub-functions restricts indicator setting and fails to represent multidimensionality, such as the leisure function and landscape function. Besides, the cultural function is difficult to be quantified because it should be analyzed through the methods from behavioral geography or sociology. Therefore, research about the appraisal b. Fig. 9 Land use system observed by von Thünen (a) and the spatial evolution of rural multifunctionality from urban areas to rural remoteness (b) method, variable setting, and how to delineate cultural functions needs to be highlighted, thus providing a more comprehensive and holistic understanding of rural multifunctionality. Importantly, developing multifunctionality must be considered in relation to the farm aspects, as farm level is the smallest spatial unit for the implementation of multifunctional decision-making processes. Therefore, a microscale analysis is urgently needed and it should not be merely focused on mesoscale by using the statistical data, but underlines the household behavior related with diversification, intensification, or environmental conservation.

Conclusion
Attempting to understand how rural multifunctionality responds to economic growth, we established indicator systems with three dimensions for measuring MRT: living function, production function, and ecological function, and quantitatively examined their links to economic growth, as well as delineated the spatial pattern by employing exploratory spatial data analysis (ESDA). The concept of transect was applied to compensate for data deficiencies in a long temporal series. Various conclusions were drawn: Living function and production function display an increasing trend from underdeveloped western regions to eastern developed regions and represent a high degree surrounding urban agglomerations, while economic growth only leads to a statistically insignificant decreasing trend in ecological function. These findings confirm our hypothesis that rural multifunctionality will undergo specific stages corresponding to a particular economic developmental level, and the "urban-driving" effect plays a significant role in the spatial agglomeration of counties with high rural multifunctionality. Both endogenous and exogenous factors contribute to rural multifunctional transition, which should be understood within a comprehensive framework.
Despite the limitations and uncertainties, our findings not only highlighted the difficulty of quantifying rural multifunctionality but also provide empirical evidence of the relations between economic growth and MRT. Moreover, the trade-off or mutual relations among the three dimensions of rural occupancy can be used for formulating regional spatial planning. Importantly, managerial countermeasures and policy recommendations concerning rural vitalization and sustainable rural development should be designed by considering the economic development level rather than the uniform regulations.
Availability of data and materials The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.