Spatiotemporal Differentiation And Driving Mechanism Of Urban Shrinkage In China Based On Multi-source Data

Since the reform and opening-up was introduced,urbanization in China has advanced rapidly,with a significant increase in both the number and size of urban areas.However,urban development has now entered a phase of divergent growth,with an increasing number of cities experiencing shrinkage.China’s urban development is thus at the cusp of a new era characterized by concurrent urbanization and urban shrinkage.This study focuses on urban shrinkage at a national scale,employing county-level cities as the smallest unit of analysis.The study period spans from 2000 to 2020,with a detailed examination of two sub-periods: 2000–2010 and 2010–2020.Utilizing a diverse array of data sources,including census data,1 km ×1 km gridded GDP data,and nighttime light data,alongside a multifaceted structural framework encompassing scale,space,and urban-rural dynamics,this study systematically explores the spatiotemporal differentiation and driving mechanisms behind urban shrinkage in China across population,economic,and spatial dimensions.The key findings are as follows:(1)Drawing from the fifth,sixth,and seventh national population censuses,and employing methods such as the shrinkage model,spatial autocorrelation model,and standard deviational ellipse,this study delineates the spatiotemporal patterns of urban population-related shrinkage in China.The results indicate that:(A)The extent of urban population-related shrinkage in China is expanding and intensifying,with the process gradually advancing.The distribution of shrinkage is most widespread in the Northeast region,with provincial border areas and frontier regions being particularly affected.Significant positive spatial correlation exists between urban population growth and shrinkage;the distribution of the urban population-related shrinkage index is becoming more balanced,with its spatial center moving westward towards the Hu Line,and the area of the standard deviational ellipse gradually increasing.(B)In terms of the size,the advancement of the population-related shrinkage process does not alter the dispersed distribution pattern of urban populations,with smaller cities more likely to experience population-related shrinkage.Cities experiencing shrinkage are most numerous among smallsized cities,exhibiting a pattern of “dispersion in general and aggregation locally,” with Type II small-sized shrinking cities primarily located in provincial border areas,outnumbering Type I.There are fewer medium-sized shrinking cities,which are dispersedly distributed;among large-sized cities,only Type II cities exhibit population-related shrinkage,with the smallest number and least severity.(C)In terms of spatial patterns,urban population-related shrinkage shows a clear core-periphery structure,and there is a transmission mechanism of edge-type shrinkage → center-type shrinkage → whole-region shrinkage.Whole-region shrinking areas are mainly located in the Northeast region and are expanding;center-type shrinking areas’ range gradually narrowing;edge-type shrinking areas continue to increase in number as the process of urban shrinkage advances.(D)From an urban-rural structure perspective,the characteristics and trends of shrinking cities reflect the current situation in China where urbanization and urban shrinkage are occurring simultaneously,and the urban population-related shrinkage process is slower than in rural areas.On one hand,regions with increasing urban populations and decreasing rural populations are most numerous,with a reduction in the number of regions experiencing growth in both urban and rural populations;on the other hand,the number of regions experiencing shrinkage in both urban and rural populations(Type I)is decreasing while Type II is increasing,and the number of regions with decreasing urban populations and increasing rural populations is decreasing.(2)Based on the processing of diverse datasets,including China’s 1 km × 1 km gridded GDP data,Landscan global population data,global 30 m impervious-surface dynamic dataset,and nighttime light data,urban entities are identified to delineate the research scope of urban shrinkage in economic and spatial dimensions.The results indicate that:(A)While the scope of urban economy-related shrinkage is not as extensive as that of population-related shrinkage,its severity is more pronounced.Many of the shrinking prefecture-level cities are resource-based in the Northeast region,where the scope and degree of shrinkage significantly increased from2010 to 2020.The spatial distribution of economy-related shrinkage shows a pattern of“dispersion in general and aggregation locally,” with the Northeast region being the most extensive.Urban space-related shrinkage has the smallest scope and mildest degree,distributed in a dispersed pattern,with an increasing tendency to cluster in the western regions over time.(B)The spatial pattern of comprehensive urban shrinkage also exhibits a “dispersion in general and aggregation locally” distribution,concentrated mainly in the Northeast region and Inner Mongolia,showing significant positive spatial correlation.Based on the trajectory of the comprehensive urban shrinkage index,cities are classified into three types: potential,stage,and sustained shrinkage.Cities with potential shrinkage are the most numerous,concentrated mainly in the Northeast and Inner Mongolia;the number of stage shrinkage cities is second,and the proportion of prefecture level cities is relatively high;and sustained shrinkage cities are mainly resource-based and border cities.(C)Looking specifically at the differences in the process of urban shrinkage,population-related shrinkage occurs first,followed by economyrelated shrinkage and space-related shrinkage.The three dimensions interact and influence each other during the comprehensive urban shrinkage process.Based on the combination of population,economic,and spatial dimensions,cities undergoing comprehensive shrinkage are categorized into four types: transitional shrinkage,structural shrinkage,significant shrinkage,and full-dimensional shrinkage.Cities experiencing transitional shrinkage are the most numerous and their scope is expanding,concentrated mainly in the Northeast;significant shrinkage cities are next in number,also clustering in the Northeast;structural shrinkage cities are fewer,but their scope is rapidly expanding;full-dimensional shrinkage cities are the least numerous,mainly border and resource-based cities.(3)Utilizing Pearson’s correlation coefficient,ordinary least squares,spatial regression models,and geographically weighted regression,the present study analyzes the driving effects of different factors on urban shrinkage and their spatial differentiation characteristics to elucidate the driving mechanism behind.(A)Six indicators—population aging and fertility preferences,the proportion of the tertiary sector,level of social consumption,basic healthcare level,urban greening level,and internal investment strength—passed the significance tests in all four regression models from 2000 to 2020.The results also indicate a spatial differentiation in the effects of these factors on urban shrinkage.(B)The mechanisms driven by independent factors include: aging and low fertility leading to slow or negative urban population growth;regional economic disparities causing differences in residents’ employment income;resource depletion leading to shrinkage of city’s pillar industries;core cities’ siphonic effect concentrating resources in higher-tier cities;social environment differences affecting residents’ quality of life;industrial structure rigidity hindering urban development momentum;poor ecological environments prompting the loss of production factors;regional transport development accelerating the flow of urban factors;and the city’s administrative level and size establishing the basic pattern of urban shrinkage.(C)At the same time,there is a complex interplay among driving factors,which continuously propels the process of urban shrinkage.Resource depletion and industrial structure rigidity adversely affect economic development,which in turn leads to decreased resource demand and challenges in industrial upgrading.Poor ecological and social environments cause the loss of resources,further leading to a lack of economic development factors,while economic development supports the improvement of social and ecological environments.A superior social environment is a crucial factor attracting population inflow,but a continuous decline in population size can lower the urban rank and scale,leading to resource loss under the siphonic effect.Regional transport development further accelerates the loss of resources.(4)Drawing from the findings of this study,the driving mechanism of urban shrinkage as well as the experiences from typical developed countries such as Germany,the US,and Japan,specific countermeasures for addressing urban shrinkage in China are proposed.These include advocating for smart shrinkage,guiding cities towards “slimming and strengthening,”enhancing technological support to boost urban innovation capabilities,adjusting and optimizing industrial structures to develop alternative industries,controlling urban size rationally to improve land use efficiency,encouraging fertility scientifically to optimize population structure and quality,upgrading infrastructure to enhance urban resilience,strengthening urban division of labor and promoting regional collaborative development,and improving the social security system to optimize the business environment.