Objective Metropolitan areas have become the primary mode of urbanization in the new era. The development of urbanization in metropolitan areas should consider the activation of suburban land and the needs of urban residents for green space. Urban agriculture represented by citizen farms can provide an essential experiential green living interaction space, and thus promote the sustainable development of urban agriculture land blocks and contribute to the construction of experiential urban green open spaces. Existing researches on urban agriculture and citizen farms, both at home and abroad, mainly focus on the historical theory, classic case studies and development of citizen farms, as well as the use and spatial distribution of urban agricultural lands. However, these researches mostly rely on literature analysis, current situation investigation, and questionnaire survey, and seldom adopt big data and quantitative analysis methods to identify the characteristics and responses of citizen farms. At present, relevant researches are mainly conducted from the perspective of single citizen farm or single urban scale. It is essential to grasp citizen farm planning and construction methods from the perspective of metropolitan area and urban cluster to promote the sound and orderly development of citizen farms.

Methods The Tokyo metropolitan area in Japan is home to 30.67% of the total citizen farms in the country, which is an ideal object for the research on citizen farm especially due to its vast market demand and development potential. Firstly, this research gathers original citizen farm information from official website of the Ministry of Agriculture, Forestry and Fisheries of Japan and transforms such information into longitude and latitude information. As to data statistics, repeated data distribution points are calculated only once, while distribution points that cannot be located are either manually corrected or removed, with all data finally obtained being visualized using ArcGIS software. Secondly, the research regards all citizen farms in the Tokyo metropolitan area as point features, and determines the spatial layout types thereof using the nearest neighbor index. Additionally, the research adopts the kernel density estimation method to reveal the agglomeration trends of citizen farms, and ellipse analysis of the standard deviation to determine their diffusion directions and degree of dispersion in space. This approach can help identify the spatial distribution pattern characteristics of citizen farms in a mature metropolitan area. On this basis, the research visualizes the differences between citizen farms in the distribution of five attributes, including the number of citizen farms, average land occupation area of citizen farms, average number and area for sub-divisions of citizen farms, and presence of management or guidance personnel. Finally, the research adopts the number of citizen farms per capita in various cities within the Tokyo metropolitan area as a distribution indicator for citizen farms, and adopts the correlation analysis and spatial overlap methods to investigate the internal mechanisms affecting the distribution of citizen farms across the four levels of transportation, economic development, population composition, and policies. The findings of this research may provide valuable insights for the development of citizen farms in metropolitan areas in China.

Results The research investigates the spatial distribution and attribute distribution characteristics of urban citizen farms in the Tokyo metropolitan area, and identifies the internal distribution mechanism thereof. The main findings are as follows. 1) Spatial distribution: Citizen farms in Tokyo and Kanagawa Prefecture exhibit a patchy pattern, while those in Saitama Prefecture have a combination of patchy and point distributions, and those in Chiba Prefecture are mainly point-distributed at the city's edge. In terms of land use, citizen farms are patchy in high-density built-up areas, densely point-distributed in suburban areas, and scattered in natural environments. The citizen farms in the Tokyo metropolitan area are centered in Chofu City in Tokyo and show a north-south diffusion trend. The citizen farms obviously present a ring-shaped multi-layered structure, and the number thereof overall displays a "first-increase-then-decrease" trend. The first layer is 0-30 km away from the central point, in which the number of citizen farms increases in fluctuation. The second layer is more than 30 km away from the central point, and the number of citizen farms distributed therein decreases in fluctuation. 2) Attribute distribution: The number of citizen farms is generally higher in the western and southwestern sides of the Tokyo metropolitan area, and the average area is generally smaller in Tokyo and Kanagawa Prefectures. The cities with the largest average area are mostly distributed in Saitama Prefecture, and the average area of citizen farms in Chiba Prefecture is generally larger. The spatial distribution characteristics of the average number and average area of sub-divisions for urban citizen farms are similar. As to the management of citizen farms, those in the central part of Tokyo and the southern part of Kanagawa Prefecture are relatively well-managed, while the management of land in Chiba Prefecture and Saitama Prefecture is relatively loose. 3) Analysis of influencing factors: The spatial distribution of citizen farms is influenced by four factors at the urban agglomeration level: transportation, economic development, population composition, and policies. The density of rail transportation is significantly positively correlated with the number of citizen farms, while the density of expressways has no significant correlation. Per capita income is significantly positively correlated with the number of citizen farms, and the per capita income in the aggregation area of citizen farms is relatively low. The number of citizen farms is significantly positively correlated with the population size, and significantly negatively correlated with the degree of population aging. The per capita citizen farm area is significantly negatively correlated with the population size and has no significant correlation with the degree of population aging. Policy systems play a guiding and restrictive role in the development of citizen farms.

Conclusion After analyzing the spatial distribution and influencing factors of citizen farms in the Tokyo metropolitan area in Japan and comparing them with the current development status and governance issues of urbanized cities in China, this research has identified the prospects and trends of urban citizen farms. In terms of site selection, various factors such as transportation construction, economic level, and population development should be considered comprehensively. The synergies of urban agglomerations in such aspects as transportation and policies, should be fully utilized to select areas with good external environment for construction of urban citizen farms, so as to create a demonstration effect. To ensure the stable and sustainable development of urban citizen farms in China, policy guarantees should be put in place. However, relevant policies should be based on the national conditions of China with a focus on land use and the combination of urban citizen farms and urban green space classification. Specific normative policies for urban agriculture and urban citizen farms should be developed to safeguard the interests of operators and users. In conclusion, the development of urban citizen farms has great potential in promoting sustainable urban development and improving the quality of life of urban residents. Through careful site selection and policy support, China can create a favorable environment for the development of urban citizen farms and realize the full potential of this emerging form of urban agriculture.