The number of shrinking cities in the world is increasing, and the affected areas are spreading from developed countries to developing countries. From a global perspective, urban shrinkage has occurred or is occurring in highly urbanized and rapidly urbanizing areas. Since the mid-20th century, urban populations have begun to decline in cities that have developed into megacities driven by industrialization. For example, London was the world’s largest city for 100 years, but its population began to decline in the early 20th century. Many other industrial cities began to shrink at the same time, including Liverpool, Paris, Chicago, and Detroit. By the end of the 20th century, due to changes in the global political situation and changes in social systems, Eastern Europe had the most concentrated population declines. Studies have shown that between 1990 and 2000, the world’s urban population declined by a quarter, including a 40% decline in the urban population of Europe (Turok and Mykhnenko, 2007). In the 21st century, this trend has gradually spread to other developed and developing countries, such as Japan, South Korea, and South Africa. In State of the World’s Cities Report 2010, UN-Habitat identified approximately 50 cities in China that are shrinking.

The degree of urban shrinkage in China is gradually increasing, and the number and distribution of shrinking cities are increasing. In terms of the quantity of cities suffering from urban shrinkage, Long et al. (2015) used data from China’s fifth and sixth censuses to analyze the populations of 654 cities at or above the county level and identified 180 cities whose total population or population density is falling. Wu et al. (2015) used data from two censuses to identify population shrinkage of 18.8% in 181 areas of the Beijing-Tianjin-Hebei region and 44% in 282 areas of the Yangtze River Delta region based on four attributes. Using townships and sub-districts as well as city districts and counties as the spatial scales, Li et al. (2015) used population census and yearbook data to determine that 23.02% of 604 towns and sub-districts in the Pearl River Delta had negative average annual population growth rates. Zhang et al. (2016) conducted a study of 2,865 counties and county-level cities (or districts) in China, which showed that 26.71% of administrative units at and above the prefecture level and 37.16% of counties and county-level cities (or districts) have shrunk. China’s northeast region has the most concentrated and earliest instances of urban shrinkage, in a similar way to the Rust Belt in the Great Lakes region of the US. Population loss in resource-based industrial cities has led to the emergence of a series of urban shrinkage characteristics, such as economic recession, increased idle land, and unused infrastructure, such as in the cities of Chaoyang in Liaoning Province, Hegang in Heilongjiang Province, and Jilin in Jilin Province (Zhang and Xiao, 2020). Some urban shrinkage has also occurred in China’s northwest and eastern regions (Guo and Li, 2019). It is important we learn from international experience and carry out theoretical research and policy adjustments on urban shrinkage based on China’s national conditions and development path.

Urban shrinkage, like growth, depends on various historical, economic, political, and demographic factors, and it is essential to keep it in mind when pursuing rapid urbanization. The UN-Habitat pointed out in its State of the World’s Cities Report 2010 that it is wrong to assume that global urbanization means that all cities are developing. In fact, many cities in all regions of the world, especially in developed regions, are actually shrinking. While rapid urban growth is still the norm in most developing countries, shrinkage also occurs occasionally in those countries. An analysis of 1,408 cities in developing countries showed that 143 cities (10.2% of the total sample) experienced population decreases (i.e., negative growth) between 1990 and 2000. UN-Habitat stated in the World Cities Report 2020 that some cities and local governments have set up departments to welcome refugees and migrants in order to increase their populations, such as the German cities of Munich, Dusseldorf, Stuttgart, and Freiburg. For shrinking cities in Europe, North America, Japan and South Korea, which are experiencing ageing populations, low birth rates, and deindustrialization, immigration provides an opportunity to restore wealth and revitalize urban areas.

Urban shrinkage is no longer a new concept internationally, but the increase and spread of urban shrinkage around the world since the end of the 20th century has given research by domestic and foreign scholars the following regional characteristics and new features. (1) Urban shrinkage has increasingly obvious regional differences and unique features. Significant evidence shows that the global urban population is shrinking notably, especially in countries and regions with high levels of industrialization and urbanization. It occurred first in the UK and then spread to other countries in Europe, the US, Japan, and so on (Oswalt, 2012). There are also countries, like China, where urban shrinkage is occurring alongside increasing urbanization. (2) The content of urban shrinkage research is becoming more varied. There are different standards for defining cities around the world, and no consensus has been reached yet. Different development stages and political systems mean that urban shrinkage research can focus on economic, cultural, social, spatial, and environmental aspects of urban development. (3) Research on the influencing factors and process of urban shrinkage is developing. Research on urban shrinkage has gradually expanded from analysis of the phenomenon itself to discussions of its processes and mechanisms, from simple quantitative and regional research to dynamic observations of shrinkage, and from single-factor drivers, such as population or industry, to multi-factor feedback mechanisms involving economy, society, and the environment. (4) The number of areas of research and inter-disciplinary research have increased, and research methods have diversified.

The healthy and sustainable development of cities has long been a focus of international organizations, institutions, and countries, so international debate on urban shrinkage is spirited. As early as 2002, the research project “Shrinking Cities” launched by the German Federal Cultural Foundation, compared the similarities and differences of urban shrinkage in Germany with other countries as well as its social impacts. Shrinking Cities, edited by Philip Oswalt (2012), was the important outcome of the program, which expounded the urban shrinkage situation of Manchester and Liverpool in the UK, Detroit in the USA, Ivanovo in Russia, Halle and Leipzig in Germany, and Tokyo in Japan. The Shrinking Cities International Research Network (SCIRN), established in 2004, is a research consortium of 30 scholars and experts from 14 countries dedicated to conducting research on shrinking cities on a global scale, including definitions, causes, manifestations, spatial differences, and policy and planning interventions.

Special issues on shrinking cities have been published in overseas journals including the International Journal of Urban and Regional Research in 2012, Build Environment in 2012, and European Urban and Regional Studies in 2015 (Zhang and Li, 2017). Special collection about “Shrinking Cities and Towns” have been published in the journal Urban Planning and Development in 2021，Urban Design International in 2013, about “Promoting Social Justice and Equity in Shrinking Cities” in Journal of Urban Affairs in 2020 (Li et al., 2021). To capitalize on international research progress regarding urban shrinkage and research on shrinking cities in China, in 2014, Chinese scholars organized and launched the China Shrinking Cities Research Network, which holds regular academic seminars on shrinking cities in China. As of January 2022, six sessions have been held. Chinese journals have also published special issues on shrinking cities, including Urban Planning International in 2020, Beijing Planning Review in 2019, Tropical Geography in 2019, Journal of Human Settlements in West China in 2018, Planner in 2017, and Modern Urban Research in 2015. The close attention paid to urban shrinkage in China and overseas has added support and impetus to the development of urban shrinkage research.

The Chinese government has issued new requirements and new policies regarding urban shrinkage research. The Key Tasks of New Urbanization Construction in 2019, released in March of that year by the National Development and Reform Commission of China, mentioned the development of shrinking cities for the first time, indicating that the government was starting to pay attention to the phenomenon in China. In April 2020, the National Development and Reform Commission mentioned shrinking cities again in the Key Tasks of New Urbanization Construction and Urban-Rural Integration Development in 2020: “We must coordinate the cultivation of new cities and shrinking cities so that they ‘lose weight’ but ‘gain strength’... safely reduce shrinking urban municipal districts, and prudently conduct research on and adjust shrinking counties (and county-level cities).” The two documents stated that urban shrinkage should not be ignored but should be faced up to with scientific research and policy control measures, changing the approach from development based on planned expansion to that based on maintenance or even cutting back. Urban shrinkage research is an important topic in the new era for the Chinese state to meet the dual conditions of internal adjustments and external pressure faced by urban development in China. It is also a new topic that can aid the optimization of the layout of China’s urban system, the modernization of urban governance capabilities, and the construction of sustainable cities.

The term “shrinking city” was first used in the US in 1947 (Gao, 2017). In 1988, German scholars Häußermann and Siebel formally applied the concept of “shrinking city” to their empirical research on the Ruhr region of Germany (Wu and Sun, 2017), which they used to describe the problem of urban economic decline in areas with significant population loss. As a result, urban shrinkage subsequently garnered widespread attention. SCIRN defines a shrinking city as a densely populated urban area of more than 10,000 people that has experienced population loss for two consecutive years and is experiencing a structural economic crisis (Wiechmann, 2008). This is the most accepted definition of a shrinking city at present. See Table 1 for different definitions of a shrinking city. In this study, we define urban shrinkage as the process of population loss, industrial and functional decline, and eco-environmental deterioration due to various natural and human factors that occur once a city develops to a certain point. It is the external manifestation of positive or negative changes in urban economy, society, culture, resource utilization, and sustainability. Unlike the negative connotation of “urban decline,” “urban shrinkage” is more neutral.

Methods for identifying urban shrinkage, such as the selection of spatial scales and evaluation indicators, are very diverse. The reason for this is differences in the zoning and administrative hierarchies of countries, resulting in dissimilar urban boundaries and scales, urban-rural division criteria, and demographic areas, which in turn has led to disagreement in academia concerning a unified standard for identifying urban shrinkage. For example, the US, the UK, and Germany have very different standards for defining cities, from small-scale communities to large-scale cities, making it difficult to compare urban populations between countries (Gao, 2017). There are at least seven spatial standards for determining urban shrinkage in China: administrative districts of cities at and above the prefecture level, municipal districts of cities at and above the prefecture level, counties and cities (districts), townships and sub-districts, urban towns, rural urbanization areas, and urban areas (Liu and Zhou, 2019). Scholars have used various population indicators and methods to determine whether a city is shrinking. Others have combined population data and economic data to identify urban shrinkage. Yet others have combined multiple indicators to comprehensively delineate shrinkage (Table 2). In response to these different approaches, some scholars have called for the use of a single indicator - population loss - to define urban shrinkage, as using too many indicators leads to conceptual differences. There are dissenters, however, who argue that the definition of urban shrinkage should consider multiple dimensions, as using just population is too one-sided and does not fully encapsulate the content of urban shrinkage (Li et al., 2017b).

Based on a summary of research by Chinese and overseas scholars, it is possible to make a more detailed delineation of types of urban shrinkage from the three aspects of spatial form, severity, and causes.

(1) Spatial form: Urban shrinkage in the West is usually classified based on the spatial form of the urban structure and population flows, which include punctured, donut-shaped, and cohesive. The punctured type is characterized by discontinuities in urban space, loss of population, vacant buildings, and mixed distribution of internal forms of the city. The doughnut type of urban shrinkage is when there is substantial outward migration of a population from the city center, while the population of the suburbs remains relatively stable. Conversely, the cohesive type of urban shrinkage is when the population of the city center remains stable, but there is population loss in the suburbs (Schetke and Haase, 2008; Blanco et al., 2009; Du and Li, 2017b; Wu and Sun, 2017). Chen et al. (2020) divided urban shrinkage in the Yellow River Basin into three types based on population census data: isolated point shrinkage, continuous shrinkage, and staggered shrinkage. Chen et al. (2021) used population data from 2005 and 2015 as well as a shrinkage index, geostatistical analysis, and geographic detectors to reveal that the pattern of population shrinkage in the Loess Plateau conforms with the “core-periphery” pattern, i.e., population agglomeration in the core area (city) and population loss in the peripheral area (counties).

(2) Severity: Based on an analysis of the state of urban population shrinkage in different time periods, Liu et al. (2020b) divided shrinking cities into four categories: continuous shrinkage, growth-to-shrinkage, shrinkage-to-growth, and continuous growth. Wu et al. (2020) looked at prefectural-level cities in Northeast China, and divided them into four types: severe shrinkage, moderate shrinkage, mild shrinkage, and non-shrinkage.

(3) Causes: Based on a case study of Hunan, Zhou et al. (2019) divided shrinking cities into three categories: hollow decay (in Taoyuan County), attacked center (in Shaodong County), and resource exhaustion (in Lengshuijiang City). Zhang et al. (2019a) investigated the total populations of prefectural-level cities from 2014 to 2017 and identified 26 shrinking cities, which they divided into four types based on their causes: poor geographical environment, resource depletion, irrational industrial structure, and siphon effect.

Scholars from different disciplinary backgrounds, including geography, planning, management, economics, demography, and environmental studies, have drawn on various theories, including life cycle, capital circulation, agglomeration economies, urbanization stages, supply and demand, population migration, and compact city theories, to analyze and solve the problem of urban shrinkage from multiple perspectives. For example, Robert and Beauregard (2009) conducted a historical quantitative analysis and theoretical evaluation of the reasons for three stages of population loss in large cities in the US, in which they stated that reasons for urban shrinkage vary with time. Scholars reached different conclusion on the driving factors of urban shrinkage based on their theoretical perspective. Table 3 shows some of the types of drivers identified, and research areas used, in Chinese studies that analyze urban shrinkage. Although scholars have identified various reasons for shrinkage, factors such as economy, population, resources, politics, globalization, and environment are commonly recognized factors among Chinese scholars (Yang and Sun, 2015).

Zhou and Qian (2015) identified the causes of urban shrinkage as population change (population evolution, age change, and migration), environmental change (climate change and environmental pollution), economic change (globalization, resource exhaustion, deindustrialization, housing market decline, uneven economic development, natural cycle of economic development, and technological progress in both production and transportation), political change (ideological change and policy decision mistakes), and spatial change (suburbanization, urban sprawl, satellite cities effect, and regional uneven development). Gao and Long (2017) used the city of Yichun in Heilongjiang Province, Northeast China as a case study and found that, compared with resource-based shrinking cities in the West, the city’s causes of urban shrinkage were more complex, as it was affected by a low regional birthrate, marginal location, and institutional changes, though it also suffered from problems that are common among resource-based cities. In the cases of both natural environmental driving factors and socioeconomic driving factors, internal and external variables gradually change from single-dimensional to multidimensional, and from a simple linear relationship to complex network and cyclical feedback modes. On the whole, research on the drivers of urban shrinkage has gradually revealed a complex process of multi-disciplinary, integrated, and multi-dimensional mutual feedback. In this study, we focus on analyzing urban shrinkage from the six aspects of population, industry, policy, major events, resource depletion, and transport openness.

(1) Population migration and loss are major factors affecting urban shrinkage. They are core criteria for measuring the degree of urban shrinkage. Due to the availability and continuity of population data, many scholars have used population indicators such as total population of a city or region, population structure, and changes in family size, to discuss patterns of urban population shrinkage, principles of spatial differentiation, and influencing factors of urban shrinkage. For example, Guo and Li (2019) used population changes of municipal and non-municipal districts of prefectural-level cities in China to analyze shrinking and potential shrinking cities and found that negative externalities of shrinking cities weaken in the course of development. Meng et al. (2020) used principal component analysis to analyze the factors driving urban shrinkage in the city of Siping, Jilin Province in China’s northeast based on the three major indicators of population, economy, and society, such as net population migration, natural population growth, and the proportion of employees in tertiary industries, and they found that there are correlations and interactions between the various factors. Hartt and Hackworth (2018) explored the impact of reduced household size on population loss in shrinking cities through an empirical analysis of the Rust Belt in the US and rejected the simple growth-decline binary of population loss in other literature. The study found no significant difference in the spatial distribution of elderly families and families with school-age children in shrinking cities, while single-person households are more likely to live in central areas of shrinking cities than families. They suggested that plans and development strategies for shrinking cities should reflect a range of demographic changes, such as outmigration and internal household adjustments. Wiechmann and Bontje (2015) argued that economic decline, deindustrialization, and high unemployment rates cause urban population loss and lead to urban shrinkage. Suburbanization and urban sprawl are also considered drivers of urban shrinkage (Wiechmann and Pallagst, 2012).

(2) Industrial upgrading difficulties and economic recession caused by industrial transfer are prominent economic factors of urban shrinkage. Due to the dual effects of globalization and urbanization, urban development requires timely upgrades and industrial replacement to satisfy market demand. Du et al. (2019) found that the number of shrinking cities in the Pearl River Delta of China increased rapidly from 2007 to 2010, and argued that the transition from manufacturing to service industry effectively improved the economic resilience of cities after the global financial crisis, and suppressed the trend of continuous shrinking of cities. Technical barriers, lack of funds, talent restrictions, and environmental constraints can cause urban development to slow or stagnate, which leads to urban shrinkage phenomena such as a decline in urban economic vitality. Martinez-Fernandez et al. (2012) argued that globalization concentrates resources, critical infrastructure, and knowledge assets in global cities that continue to attract population and skills, forming a new global system of production, manufacturing, distribution, and consumption, while other towns, cities, and regions gradually shrink. Bernt and Rink (2010) found that deindustrialization is an important economic factor in the shrinking of many cities after studying urban shrinkage in Germany. It has also been found that economic recession can indirectly promote changes in population patterns within cities or outmigration, which may cause population shrinkage (Schatz, 2010; Reckien and Martinez-Fernandez, 2011). Economic globalization, various concepts of economic consumption, and economic systems can also cause urban shrinkage. In urban agglomerations, due to regional differences, cities at the bottom of the agglomeration system can become dependent shrinking cities or local shrinking areas within the urban agglomeration (Wu and Qi, 2021).

(3) Population outflows caused by policy easing, structural changes, and adjustments to administrative divisions are policy factors that can cause urban shrinkage. Changes to national systems and mechanisms or to plans and regulations, such as population relief policies and planning strategies, may cause some cities to experience a population outflow. Bontje (2004) described the development path of Leipzig in the 1990s and expounded the impact of urban institutional changes on shrinking cities in Germany, with specific attention paid to post-socialist cities. As an important measure to promote the coordinated development of the Beijing-Tianjin-Hebei region (Li and Xi, 2015), Beijing’s policy to relieve the capital of non-essential functions has brought population growth in Beijing under control and ensured its sustainable development, and areas surrounding Beijing have benefited from the relocation of industries and people from Beijing, laying sound foundations for their economic development. Although this policy adjustment has caused the population of Beijing to “shrink” temporarily, this differs from the economic structural crises that cause population outflows in shrinking cities. It is a part of the government’s measures to “streamline” and “enhance strength” in response to conflicts between humans and land and pressures on resources and the environment with a view to alleviating increasingly severe urban disease. He et al. (2017) presented two case studies of the oil-producing city of Daqing and the coal mining city of Pingxiang and found that although planning policies and practices have played a role in population loss in resource-based cities in China, they have benefitted the health of cities through long-term restructuring and development.

(4) A short-term decline in urban population caused by a major event is an uncontrollable factor of urban shrinkage. The occurrence of sudden and major events such as financial crises, climate change, and natural disasters may lead to a short-term fall in an urban population. Qin and Du (2017) analyzed the characteristics and driving mechanisms of urban shrinkage in areas of rapid rural urbanization in the Pearl River Delta urban agglomeration in China due to the financial crisis. They discovered spatial characteristics, such as vacant rental houses and sub-leasing of factories, and concluded that industrial transformation and upgrading as well as job losses were the main drivers of shrinking during the crisis. Based on historical documents and books and local chronicles on phenology and meteorology in China, Chu (1972) gave a general outline of climate change in China during the past 5000 years, pointing out major events that led to population changes, such as a sudden change in temperatures during the Eastern Han Dynasty and the wars of the Northern and Southern Dynasties. Many scholars have confirmed, using various research methods, the close correlation between changes in dynasties and anomalous weather (Zhang et al., 2008; Liu et al., 2009; Liu et al., 2016; Li et al., 2017a). Such research has indicated that sudden changes in climate affected the stability of ancient royal power, whether due to changes in crop growth and subsequent reductions in food, abnormal weather conditions, or war, causing urban populations to decrease.

(5) Depleted resources and energy and rigid constraints on the eco-environment are key factors that can cause urban shrinkage. Resource depletion and severe environmental pollution lead to population loss and rapid economic recession in resource-based cities, and they indirectly lead to urban shrinkage by affecting the restructuring and upgrading of industries and urban livability. Based on panel data of 116 resource-based cities in China, Li and Dewan (2017) found that most resource-based cities in China are total-factor energy inefficient, and to improve efficiency, it is necessary to adjust and upgrade their industrial structures, boost innovation, and extend leading industry networks. As such, a combination of production, education, and research should be carried out to enable resource-based cities to achieve sustainable development and avoid urban shrinkage. Wu et al. (2019) studied resource-based shrinking cities in Northeast China and found that higher urban economic efficiency led to less obvious vulnerability to urban livelihoods. This suggests that for resource-based cities, greater economic efficiency would contribute to the sustainability of the livelihoods of the urban population. Some studies have also shown that there is a paradox between population loss and economic and spatial growth in cities that rely on island resources to develop. Jiang et al. (2021) looked at Daishan Island in northeast Zhejiang Province, China and studied urban shrinkage and driving mechanisms from the perspectives of population and economy. Their results showed that between 2006 and 2016, the population of Daishan Island shrank imperceptibly amid economic and spatial growth. This paradox differs from Western concepts of urban shrinkage. This may be because the population outflow eased the pressure on island city resources, which were then directed toward industrial development.

(6) Accessibility and cultural factors are catalysts of urban shrinkage. The accessibility and connectivity of a transport network affect the spatial pattern, migration, and flow of a population. For China’s central cities and most growing provincial capitals, transportation is an important way to attract people, strengthen trade links, promote economic growth, and enhance urban vitality. For shrinking urban areas, transportation connectivity may accelerate urban shrinkage as well as provide potential development opportunities. Empirical results have shown that the construction of high-speed rail has exacerbated urban shrinkage in China, but its impact on urban populations may have a time lag (Deng et al., 2019). Thus, transportation can pose challenges to the development of shrinking cities, widen the development gap between cities, and further polarize developed and shrinking cities. In terms of culture, it has been posited that racial discrimination can induce urban shrinkage, such as in the case of “white flight” in the US (Zhang et al., 2021).

Other factors, such as location and the eco-environment and its protection, may also lead to urban shrinkage. Cheng et al. (2019) studied the shrinkage of 141 Chinese border cities and concluded that 26% of them were experiencing shrinkage, mainly in the northeast. Antoni and Djuki (2020) found that urban shrinkage may be connected to tightening local environmental protection and ecological regulations. Conversely, they argued that cities may also shrink when laws are missing or poorly enforced, citing forestry management in Mexico as an example, and severe environmental pollution may directly cause urban shrinkage, just as severe drought caused Australia’s inland cities to shrink. Liu (2021) studied the relationship between city shrinkage and haze pollution from both theoretical and empirical aspects. The results showed that the intensification of haze will aggravate urban shrinkage, and labor force played a mediating role in the process of urban shrinkage affected by haze. Khavarian-Garmsir et al. (2019) discussed that climate and environmental change can reduce cities’ livable in the Middle East, indirectly making residents move out and exacerbating urban shrinkage. Although the impact of environmental factors on urban shrinkage is not as obvious and direct as economic factors, it can be long-term or short-term and lead to various types of shrinking cities.

Increasing numbers of urban shrinkage scholars believe that the phenomenon should be considered as a holistic complex system because it is a cumulative and self-reinforcing cyclical process with indistinguishable causal relationships (Wu and Sun, 2017). Urban shrinkage can cause negative effects, such as an economic downturn (Lin et al., 2017), declining employment (Reckien and Martinez-Fernandez, 2011), brain drain (Anja, 2016), increased housing vacancies (Couch and Cocks, 2013), and other social issues. Some scholars, however, believe that urban shrinkage can bring positive effects and provide cities with opportunities, such as by improving the eco-environment, spaces, housing, and transportation (Haase et al., 2014).

Governments and planning circles around the world have traditionally regarded growth of population and the economy as well as spatial expansion as paradigms of urban development and planning. It is widely believed that intervening in urban decline can help a city return to growth after a short period of recession or urban shrinkage, regardless of a city’s own resilience and environmental carrying capacity. The “Youngstown 2010” plan proposed in 2005 represents a breakthrough from the old growth paradigm, providing a new perspective and strategy for solving urban problems, i.e., accepting urban shrinkage, creating a high-quality living environment, and stimulating economic redevelopment through human and organizational structures (Safford, 2009; Wiechmann and Pallagst, 2012; Rhodes and Russo, 2013). As such, there are two main attitudes toward urban shrinkage: those who regard it as a short-term stage of urban growth, and those who acknowledge and accept it as a long-term reality.

The views of governments and planning communities is gradually changing in response to urban shrinkage, with a shift from the dominant growth paradigm for dealing with shrinkage to focusing on the qualitative improvements to urban areas, or from passive decline to considered shrinkage (Zhu et al., 2019). Song et al. (2020) believed that urban shrinkage was one of the reasons which caused urban vacant land. The lack of management of urban vacant land would reduce the ecological benefits and functions of the city, suggesting establishing a monitoring, regulation and differentiated restoration mechanism. In order to solve a series of economic and social problems such as urban vacant land caused by international and domestic shrinking cities, and to actively face the negative effects of future urban shrinkage, scholars at home and abroad put forward strategies to deal with urban shrinking. Zhang and Li (2017) concluded that European and American countries were the main areas to discuss this problem, including right sizing/downsizing and making shrinking city planning based on social equity theory, generalized that Hollander et al. (2009) proposed shrinking city planning should pay attention to land use, eco-environment, equity, infrastructure and density issues, and prospected that China’s future shrinking city research should focus on the formation mechanism of various types of shrinking cities, change thinking and planning paradigm, and seek regional cooperation, development comparison and theoretical research. Zhang et al. (2017), according to China’s national conditions, believes that China’s countermeasures lie in the adjustment of urban development ideas and strategies, the implementation of coordinated man-land welfare policy for urbanization, the establishment of institutions and mechanisms in line with the inventory planning era, the shaping of flexible urban spatial structure, and the promotion of “double urban repairs” (ecological restoration and urban betterment).

(1) The existing literature has paid little attention to the eco-environmental effects of urban shrinkage, which can be divided into the two general categories based on their timing: immediate and latent.

Immediate eco-environmental effects are those that appear straight away in the process of urban shrinkage. The existing literature has tended to focus on the impact of urban shrinkage on carbon emissions, and the main conclusions are that urban shrinkage reduces total carbon emissions as well as carbon emission efficiency. Xiao et al. (2019) divided 55 cities in Northeast China and the Yangtze River Delta into three groups of shrinking cities and three groups of growing cities to explore changes in carbon emission patterns. They found that rapidly shrinking cities had the lowest total CO₂ emissions, while the CO₂ emissions of slightly and moderately shrinking cities gradually declined during the study period (2003-2015), and shrinking cities brought about a decline in urban carbon emissions. It was found that CO₂ emissions in rapidly shrinking cities showed an inverted U-shaped trend (Tong et al., 2022). However, shrinking cities tend to be less energy efficient than growing cities (Liu et al., 2020a) because depopulation and dispersion tend to increase per capita costs of daily travel and infrastructure network maintenance (Moss, 2008). Zeng et al. (2022) found that city shrinkage had a negative impact on CO₂ emission efficiency, every 1% increase, efficiency decreases about 2.1%. Meanwhile, it indirectly curbed the CO₂ emission efficiency by reducing urban R&D expenditure and urban compactness. In addition to carbon emissions, urban shrinkage can theoretically have an immediate effect on urban air pollution and water pollution, but there only a few empirical studies can be found. Such as urban shrinkage and the increase of air pollution degree had negative relationships, directly and indirectly. Local economic growth had a positive spatial spillover effect on air quality improvement in neighboring cities, while population factors had a negative effect (Qiang et al., 2021).

Latent eco-environmental effects are those that can be harnessed through planning and governance. The existing literature tended to focus on the potential ecological value of land made idle by urban shrinkage. Vacant houses and abandoned facilities due to urban shrinkage will eventually be demolished for safety and maintenance reasons to create vacant land (Frazier and Bagchi-Sen, 2015), which can be transformed into green spaces to provide ecosystem services to the city and improve the urban environment and the quality of life of residents (Wu and Li, 2019). Lauf et al. (2014) simulated future land use development in metropolitan Berlin, Germany based on the two scenarios of urban shrinkage and growth and found that ecosystem services were significantly improved under the shrinkage scenario. A simulation by Frazier and Bagchi-Sen (2015) of reclaiming vacant spaces caused by urban shrinkage as natural areas substantially increased urban green space connectivity. Ecosystem services are considered an important element of the shift toward sustainable cities, and vacant spaces caused by urban shrinkage provides the opportunity to achieve this through ecosystem services (Haase et al., 2014; Nassauer and Raskin, 2014). The ecosystem services that these potential green spaces can provide include regulation of the local climate and air quality, carbon sequestration and storage, and preservation and enhancement of urban biodiversity (Haase et al., 2014). Research has shown that urban brownfield sites formed by urban shrinkage have huge potential ecological value. Urban shrinkage provides an opportunity for the sustainable transformation of cities from an ecological perspective, and ecology is particularly suited to providing solutions (Herrmann et al., 2016). Paying adequate attention to the ecological effects of urban shrinkage is vital for the long-term development of cities.

(2) A reduction in human activities in urban areas has a positive impact on the eco-environment. There are very few studies on the eco-environmental effects of urban population loss, industrial shrinkage, and land use reduction, with most studies focusing on the positive environmental effects of greening open spaces, such as improved species diversity, as well as shrewd use of idle land resources. Nevertheless, Peng et al. (2022) found that loss of population in shrinking cities alleviated urban heat island. Marinello et al. (2021) demonstrated during the Covid-19 pandemic, measures such as border closures and lockdowns adopted by countries greatly limited the distances people travelled, which improved local air quality. The overall trend was a decrease in concentrations of particulate matter, nitrogen dioxide, sulfur dioxide, and carbon monoxide, and a significant increase in tropospheric ozone. This indirectly indicates that as the intensity of human activities weakens, the reduction in space for human activities is likely to lead to an improvement in the air quality of shrinking cities.

(3) Vacant urban land provides opportunities for improving the eco-environment of cities. While substantial population loss poses challenges to urban economic development, it also presents opportunities for advancing sustainable development goals, with ecologists having long recognized the advantages of vacant space in cities (Herrmann et al., 2016). Scholars have conducted a lot of research on idle land and vacant buildings generated by urban shrinkage, with the focus on the relationship between green infrastructure construction and improvement of ecosystem service value as well as the relationship between changes in species diversity, vacant buildings, and vegetation coverage. Turo et al. (2020), for example, found a correlation between bee abundance and proximity to larger greenspaces within an urban landscape, showing that shrinking cities can create favorable conditions for the provision of valuable habitats. Christopher et al. (2018) evaluated the ecological and sociological benefits of urban spontaneous vegetation in shrinking cities, aimed at reducing negative impacts and improving sustainability of shrinking cities development. Previous studies have argued that urban green space coverage is highly correlated with vulnerable groups and is an indicator of environmental justice. However, green space coverage in shrinking cities can have different meanings, and research has shown that vegetation richness is positively correlated with the residential vacancy rate. This suggests that quantitative vegetation indexes alone cannot accurately explain the link between environmental justice and ecology (Deng and Ma, 2015; Kirsten et al., 2018; Berland et al., 2020). Dustin et al. (2018) argued that agroecology - an alternative economic and governance model that focuses on natural resource management - create the potential for social-ecological innovations that can achieve sustainable transformations in shrinking cities. Riley et al. (2018) see ecology as a solution to the transformation of shrinking cities, as natural vegetation in open spaces contributes to the regulation of ecosystem functions and services. Ma and Long (2018) used a right size model and proposed methods and targets for shrinking cities in China to develop green infrastructure, land banks, and cooperative planning.

In terms of temporal differences, overseas studies on urban shrinkage began approximately 60 years before Chinese ones. Overseas research on urban shrinkage began in the middle of the 20th century, and it has gradually grown and received greater attention internationally since the end of the 20th century. Chinese research on urban shrinkage, on the other hand, began in the early 21st century, and the topic’s popularity has grown steadily since 2010.

In terms of spatial differences, the phenomenon of urban shrinkage has gradually spread from developed countries in Europe and the US to developing countries such as China. As China’s urbanization process has gradually slowed in recent years as a result of diminishing demographic dividends, changes to the industrial structure, and environmental and resource constraints, leading to the emergence of urban shrinkage, which may spread in the future and coexist with urban growth for a long time. As a result, academics and politicians should continue to pay attention to urban shrinkage, accurately define related concepts of urban shrinkage in China, carry out multidisciplinary and multi-perspective research using a range of methods, prudently adjust development policies in response to shrinking in specific cities or regions, and formulate corresponding planning strategies.

International research on the eco-environmental effects of urban shrinkage mainly focuses on changes in the total amount and efficiency of carbon emissions, ecosystem service value of urban green space, reuse of urban brownfields or idle land (mainly converted into urban green space), and the potential ecological value of green spaces for climate regulation, carbon sequestration and storage, and enhancement of biodiversity. Chinese research has focused on land use in shrinking cities and its ecosystem service value and impact on energy efficiency. Very few relevant studies have shown that, theoretically, urban shrinkage can have both positive and negative effects on the eco-environment. Identifying these positive and negative effects and their interaction mechanisms is crucial to enhancing urban sustainable development and improving the urban eco-environment. There is also an urgent need for more in-depth and comprehensive systematic research. Some progress has been made in research on urban shrinkage internationally, but due to differences in economic systems, social systems, urbanization processes, and national conditions, urban shrinkage in other countries and urban shrinkage in China are significantly different in such areas as their connotations, driving mechanisms, research focal points, relevant indicators, and policy requirements. In both Chinese and overseas research though, there has been very little attention paid to the relationship between urban shrinkage and the eco-environment, while urbanization in China is having an increasingly significant impact on the eco-environment. Within this context, there is an urgent need to study the eco-environmental effects of urban growth and shrinkage as well as the positive mitigating impacts eco-environmental improvements can have on urban shrinkage.

A review of literature on the influencing factors and driving mechanisms of urban shrinkage in China and overseas revealed that the vast majority of scholars focus on analysis of human, social, and economic elements, but there has been insufficient consideration of and too few indicators that involve the eco-environment. Research on influencing factors has focused on population, economic, social, policy, location, and transportation, and research on indicators has focused on population, industry, natural conditions, and infrastructure, with few studies choosing eco-environmental influencing factors and indicators. More studies have focused on the correlation between eco-environmental factors and urban shrinkage recently, such as the green coverage rate of built-up areas, wastewater discharge, SO₂ emissions, and harmless treatment of domestic waste, but the focus wavers and they lack sufficient analysis of underlying mechanisms. Creating livable environments is a basic feature of China’s new type of urbanization, and it is an important reflection of an improving standard of living. Thus, it is important to incorporate eco-environment factors and indicators into analysis of urban shrinkage.

Urban shrinkage is the result of the combined effect of multiple factors in a specific period. In addition to population loss and economic recession, it also usually involves spatial, social, and environmental dimensions. There is still no consensus on a definition of urban shrinkage among Chinese and overseas scholars. It is still a matter of debate whether it best defined by a single dimension, such as population, or whether it is more accurate to define it using multiple dimensions, such as population, economy, and environment. We recommend that in future research, the comprehensive, systematic, and scientific nature of urban shrinkage should be considered when defining the phenomenon. Moreover, when quantitatively measuring shrinking cities, in addition to population, economic, and social indicators, eco-environmental indicators should be included in index systems. A combination of big data, remote sensing image data, and statistical data should be investigated and applied to research on urban shrinkage. In the course of research, concepts and national conditions should be clarified to prevent the misuse of concepts.

The phenomenon of urban shrinkage is a complex, dynamic process in which economic, social, and environmental factors interact with and influence each other. Changes in multiple regional internal and external factors directly affect a city and produce shrinkage, but there are also many potential factors that indirectly cause urban shrinkage. At the global level, urban shrinkage is driven by multiple natural and human environmental factors in different regions, including resource depletion, technological progress, policies, cultural differences, geographical location, accessibility, globalization, and suburbanization. At the local level, it is affected by specific natural resource endowment conditions and the effects of population, economy, ecology, and space under the development background of different policy systems (Figure 1). It is necessary to combine multi-disciplinary knowledge to place objects of research within the different scales of global, regional, national, and local to analyze the modes and intensity of interactions between influencing factors according to local conditions, as well as to use complex models and computer methods to study in-depth the dynamic driving mechanisms of urban shrinkage.

Analysis of the existing literature showed that indicators for measuring the driving mechanisms of urban shrinkage mainly focus on social dimensions, such as population and the economy, with a lack of research on the impact urban shrinkage on eco-environmental factors and a lack of relevant indicators, such as the atmospheric environment, climate change, and the eco-environment. As a stage in the process of urbanization, does urban shrinkage have the same nonlinear coupling relationship as urbanization and eco-environment, meanwhile have different coupling states at different coupling stages (Fang et al., 2020)? There is still a lack of in-depth mechanism research. As people’s expectations on living standards change from quantitative to qualitative, the built environment of cities may become an important part of population migration in the future. The new findings suggested that city shrinkage had a dual impact on the changes of living environment. On the one hand, it improved residents’ life satisfaction by easing the living environment and pressure in the city; on the other hand, it reduced residents’ satisfaction by weakening employment, landscape and livability (Fu et al., 2022). Strengthening research on the interactions between urban shrinkage and the eco-environment and the principles behind them will be essential to understand accurately and fully the process of urban shrinkage. There is a complex feedback mechanism of interactions between urban shrinkage and the eco-environment. Urban shrinkage not only positively effects the eco-environment of a city by causing improvements to the internal and external environment; it also exacerbates threats to the urban environment and causes eco-environmental decline (Figure 2). Population loss and a decline in human activities reduces disturbance to the eco-environment. When the intensity of disturbance is below the environmental threshold, the self-renewal and self-regulation capabilities of the ecosystem will gradually emerge, which have a positive effect on the eco-environment. Thus, controlling urban shrinkage can, to a certain extent, create positive conditions for the next stage of urban growth. When the degree of disturbance exceeds the environmental threshold, the ecosystem’s capacity for self-repair and purification are undermined, leading to negative eco-environmental effects. Alternatively, when there is no human management and protection, ecosystem service functions are not fully realized, and severe ecological problems, such as soil erosion, occur (Song et al., 2018) and eventually create a vicious cycle between urban shrinkage and the eco-environment. This will then exacerbate the decline and shrinking of the city. As such, it is necessary to reveal the main controlling factors, response mechanisms, and principles of interactions between urban shrinkage and the eco-environment. It is also necessary to analyze systematically the positive and negative effects urban shrinkage has on the eco-environment and the inhibitory or catalytic effect of eco-environmental changes on urban shrinkage, so as to provide empirical support for sustainable urban development and regulation.

From 2019 to 2020, the National Development and Reform Commission mentioned urban shrinkage and the target of containing urban shrinkage on two successive occasions in their key tasks for new urbanization initiatives and urban-rural integration. This indicates that research and empirical understanding of the phenomenon of urban shrinkage needs to be strengthened, and a localized urban shrinkage theoretical framework and system urgently need to be established in China to assist with implementing national development strategies. Our review of relevant Chinese literature, however, revealed that there are still too few case studies of urban shrinkage in China, too few long-term follow-up studies, various index system for measuring urban shrinkage, no comprehensive system for measuring urban shrinkage, and no monitoring, early warning, or control mechanism for urban shrinkage. To fully understand the many connotations of urban shrinkage and be able to describe quantitatively how multiple dimensions act upon urban shrinkage, a priority should be to construct an urban shrinkage evaluation index covering the six dimensions of population, economy, society, resources, eco-environment, and space (Figure 3).

In this indicator system, population is the core factor for measuring whether a city is shrinking. For example, a reduction in population scale indicators, such as total population at year end and average annual population, can describe temporal and spatial differences and the population loss situation. The economy is an important gauge of whether a city’s development strength and potential are in decline. For example, a slowdown in economic growth indicates that a city’s future development may be in jeopardy and its economy may shrink. Society, space, resources, and eco-environment are the key factors that describe whether the inner spaces, urban functions, urban vitality, and quality of life of a city are in decline. It is necessary to combine subjective and objective interpretations to directly or indirectly determine the degree of and potential for urban shrinkage. It is worth mentioning that, as China’s new type of urbanization advances, social and environmental factors such as urban infrastructure construction, house prices, air quality, and educational resources will increasingly become important indicators of where residents wish to live (Li et al., 2018). As a result, creating a scientific and comprehensive evaluation index of urban shrinkage to understand the types of urban shrinkage and using advanced technical methods to monitor and warn of urban shrinkage as well as propose plans to guide sustainable development of shrinking cities are important measures with great practical significance.

The causes of urban shrinkage have differed in different historical periods in China and overseas, and its driving mechanisms are subject to complex causal cycles and feedback mechanisms. Clarifying types of urban shrinkage in China and strengthening case analysis of typical shrinking cities and research on policy and planning adjustments would supplement international urban shrinkage research with Chinese case studies. For cities with various types of shrinkage, a positive attitude of understanding and acceptance as well as scientific management and precise policy implementation would facilitate their healthy development. Following a review of relevant literature, we can divide shrinking cities into five categories: population shrinkage-dominated, industrial shrinkage-dominated, resource shrinkage- dominated, policy shrinkage-dominated, and other shrinkage types (Figure 4). We recommend carrying out research on urban shrinkage coping strategies according to the specific conditions of each category. Since shrinking cities have significant geographical differences, when formulating corresponding coping strategies, the urban development path - “sustained growth” or “considered shrinkage” - should be designed based on the future development goals of the city. Presently, most of China’s politicians and academics adhere to the growth-oriented urban development concept and are unable to view the phenomenon of shrinkage as neutral. Most cities in China and around the world are still urbanizing at a moderate or high speed, with shrinking cities only occurring in a few countries and regions. Studies have predicted that the global population will peak in 2064, and China’s population is forecast to decline rapidly starting in 2050, down to 732 million by 2100 (Vollset et al., 2020). Consequently, urban shrinkage should not be disregarded, and it warrants greater attention and more in-depth research.

To cope with more and more cities shrinking in the future, experiments and demonstrations are required of driving mechanisms of typical cases or different types of urban shrinkage. Summaries of the experiences and responses of shrinking cities could be of use to other shrinking cities in formulating practical containment plans and policies.