Persistence Characteristics of Urban Vacant Land and its Driving Mechanism Based on High Spatiotemporal Resolution Remote Sensing Imagery: A Case Study of Chongqing, China

Persistence Characteristics of Urban Vacant Land and its Driving Mechanism Based on High Spatiotemporal Resolution Remote Sensing Imagery: A Case Study of Chongqing, China | Authorea try { document.documentElement.classList.add('js'); } catch (e) { } var _gaq = _gaq || []; _gaq.push(['_setAccount', 'G-8VDV14Y67G']); _gaq.push(['_trackPageview']); (function() { var ga = document.createElement('script'); ga.type = 'text/javascript'; ga.async = true; ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') + '.google-analytics.com/ga.js'; var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(ga, s); })(); Skip to main content Preprints Collections Wiley Open Research IET Open Research Ecological Society of Japan All Collections About About Authorea FAQs Contact Us Quick Search anywhere Search for preprint articles, keywords, etc. Search Search ADVANCED SEARCH SCROLL This is a preprint and has not been peer reviewed. Data may be preliminary. 17 April 2025 V1 Latest version Share on Persistence Characteristics of Urban Vacant Land and its Driving Mechanism Based on High Spatiotemporal Resolution Remote Sensing Imagery: A Case Study of Chongqing, China Authors : Wei Zhang 0000-0001-9029-7319 [email protected] , Zixuan Wang , Shuya Heng , Xiaolan Hu , and Yunxue Liu Authors Info & Affiliations https://doi.org/10.22541/au.174486448.84974465/v1 253 views 136 downloads Contents Abstract References Information & Authors Metrics & Citations View Options References Figures Tables Media Share Abstract Land vacancy during the processes of urban shrinkage and expansion is a major global urbanization challenge. However, research on the persistence of urban vacant land (UVL) is limited. This study analyzes the persistence characteristics of UVL at the parcel scale in Chongqing, China, and explores its spatiotemporal evolution patterns and driving mechanisms, using persistence cycle identification model, spatiotemporal cube, random forest, and high spatiotemporal resolution remote sensing imagery (2014–2024).The key findings are as follows: (1) There are 1,393 vacant parcels in Chongqing during 2014–2024, with an average duration of 6.67 years. Their total area is 5,127.73 hectares, accounting for 7.37% of Chongqing’s built-up area. (2) The spatiotemporal evolution of UVL aligns with urban spatial expansion, that supporting the wave - like circular succession hypothesis of UVL. (3) Urban construction, socio-economics, natural environment, and parcel characteristics differentially impact the persistence cycles of EVL and SVL. Interactions between natural environmental and parcel characteristic factors notably enhance their explanatory powers on vacancy duration. The “time-space-society” framework and persistence cycle identification model introduced here are hopefully to advance the formation and evolution process research of UVL, and offering robust theoretical and methodological support for its subsequent optimization and spatial governance. Persistence Characteristics of Urban Vacant Land and its Driving Mechanism Based on High Spatiotemporal Resolution Remote Sensing Imagery: A Case Study of Chongqing, China Wei Zhang 1,2,3 , Zixuan Wang 1 , Shuya Heng 1 , Xiaolan Hu 1 , Yunxue Liu 1 1 School of Geographical Sciences, Southwest University, Chongqing, 400715, China 1 New Liberal Arts Laboratory for Sustainable Development of Rural Western China, Chongqing 400715, China 1 Research Center for New Land-Sea Corridor and Regional Development, Southwest University, Chongqing, 400715, China Correspondence : Wei Zhang ( [email protected] ) Abstract: Land vacancy during the processes of urban shrinkage and expansion is a major global urbanization challenge. However, research on the persistence of urban vacant land (UVL) is limited. This study analyzes the persistence characteristics of UVL at the parcel scale in Chongqing, China, and explores its spatiotemporal evolution patterns and driving mechanisms, using persistence cycle identification model, spatiotemporal cube, random forest, and high spatiotemporal resolution remote sensing imagery (2014–2024).The key findings are as follows: (1) There are 1,393 vacant parcels in Chongqing during 2014–2024, with an average duration of 6.67 years. Their total area is 5,127.73 hectares, accounting for 7.37% of Chongqing’s built-up area. (2) The spatiotemporal evolution of UVL aligns with urban spatial expansion, that supporting the wave - like circular succession hypothesis of UVL. (3) Urban construction, socio-economics, natural environment, and parcel characteristics differentially impact the persistence cycles of EVL and SVL. Interactions between natural environmental and parcel characteristic factors notably enhance their explanatory powers on vacancy duration. The “time-space-society” framework and persistence cycle identification model introduced here are hopefully to advance the formation and evolution process research of UVL, and offering robust theoretical and methodological support for its subsequent optimization and spatial governance. Keywords: Urban Vacant Land ǀ Spatiotemporal Evolution ǀ Persistence Cycle ǀ Driving Mechanism ǀ Chongqing 1. Introduction Urban vacant land (UVL) is defined as land that remains in a state of disuse or underutilization within the urban boundary, encompassing various forms such as bare land, abandoned parcels, derelict building sites, and vegetated wasteland (Pagano & Bowman, 2000; Song & Li, 2019). As a critical component of urban spatial system, vacant land typically constitutes 10% to 15% of the total urban area (Newman et al., 2018). Globally, urban land vacancy representing a persistent and widespread structural phenomenon in the urbanization process, which is intrinsically linked to cyclical patterns of urban economic growth and decline, as well as the urban form expansion and contraction (Kremer & Hamstead, 2015). Substantial quantities of UVL not only signal inefficient urban land use but also precipitate issues such as neighborhood deterioration, elevated crime rates, and real estate depreciation (Prener et al., 2020). The adaptive reuse of UVL is closely aligned with the attainment of the United Nations Sustainable Development Goals (Goal 11: Sustainable Cities and Communities, and Goal 15: Life on Land) and constitutes a critical challenge in global urban governance (O’Callaghan, 2024). Given its significance, UVL has emerged as a focus of interdisciplinary inquiry and policy attention. Academic investigations have been systematically conducted across multiple research domains: empirical characterizations of its spatial distribution (Newman et al., 2016; Xu & Ehlers, 2022), multifaceted environmental, social, and economic impacts (Prener et al., 2020; South et al., 2023), underlying driving mechanisms (Lee et al., 2023), and sustainable utilization strategies (Qu et al., 2020; Rupp et al., 2022). These efforts have yielded substantial theoretical and empirical contributions, advancing both conceptual understanding and practical applications in UVL management. From the perspective of systems theory, urban land vacancy is not merely a static outcome but rather a dynamic process that evolves and accumulates over time. Given the strong linkage between the duration of UVL and its multidimensional impacts across social, economic, and ecological domains (Sedlar et al., 2018; Trigalet et al., 2016; Zhu et al., 2024), both the quantity and duration of UVL are identified as critical determinants for assessing urban decline trajectories (Han, 2019). Notably, some scholars have argued that the duration of UVL may hold greater significance than its quantity (Newman et al., 2022). However, despite its importance, the persistence characteristics of UVL have been surprisingly underexplored in existing research. This neglect is primarily evident in three aspects: (1) Conceptual definition gaps. Literature review on the definitions of UVL (Supplementary Tab.1) reveals that existing conceptualizations are broadly categorized into three groups: first, land-use-based definitions emphasizing its current undeveloped state; second, land-cover-oriented classifications identifying typologies such as bare land, vegetated wasteland, abandoned artificial surfaces, and derelict structures; and third, descriptive definitions relying on subjective criteria or contextual judgments to enumerate primary vacant land types. Notwithstanding this terminological diversity, the vast majority of definitions lack explicit temporal parameters, failing to incorporate the vacancy duration into their conceptual frameworks. (2) Epistemological limitations. UVL represents a complex socio-spatial phenomenon, characterized by intertwined physical attributes and social dynamics that evolving over time. However, under the dominance of a “growth-oriented” paradigm and binary analytical frameworks, urban land vacancy has been systematically positioned as the antithesis of developmental progress. Consequently, it is often framed through pejorative terminology, including “bad place” (Freire, 2020), “dead space” (Coleman, 1982), “zombie property” (Silverman et al., 2013), “urban cracks” (Loukaitou-Sideris, 1996), “lost space” (Brown, 1988), or “residue of the urban fabric” (Jeldes, 2020). Within this narrow cognitive framework, UVL is rarely conceptualized as a complex, dynamic evolutionary process (Dimitrakou, 2021); instead, it is reductively framed as a transitional void between “disuse” and “reuse”, disregarding its inherent potential as a site of adaptive transformation and socio-ecological resilience. In this context, the only quest is a convenient and practical “restart button” that is powerful for resolving the multifaceted issues and challenges posed by UVL, thereby propelling the urban economy back to growth trajectories. (3) Deficiency in empirical analysis. The systematic investigation and monitoring of UVL represent a critical foundational task in urban studies. Although a few developed countries, such as the United States, maintain relatively comprehensive official survey datasets (Newman et al., 2016), most countries or regions lack systematic data collection or public dissemination of such information. As a result, researchers predominantly rely on remote sensing imagery, aerial photography, and other geospatial data sources to obtain UVL data. Despite these efforts, empirical studies on UVL remain constrained by limitations such as restricted data accessibility, high production costs, and complex analytical processes. Consequently, the majority of existing empirical analyses focus on static cross-sectional assessments at single time points (Smith et al., 2017). Such reliance on single-phase remote sensing data inherently eliminates the temporal dimension which is critical for understanding the dynamic land-use processes, leading to incomplete and contextually impoverished datasets. These fragmented, ahistorical datasets not only provide misleading evidence to policymakers, but also impede progress in UVL research. For example, empirical studies on the drivers of UVL have identified a wide range of factors, including deindustrialization, market conditions, personal wealth, and plot physical characteristics (Newman et al., 2016); long-term population decline and demographic composition (Prener et al., 2020); spatial proximity (Deng & Ma, 2015); urban infrastructure development (Abe et al., 2014); market failures (Brueckner and Helsley, 2011); education levels and housing values (Lokhande & Xie, 2023); public regulations (Morandé et al., 2010); and disordered development patterns (Saizen et al., 2006). Such disparate and often contradictory findings not only create ambiguities for policy formulation but also diminishing the practical utility of research outcomes. In summary, the formation and evolution process of UVL is a critical research topic in urban geography and land use science. A deeper exploration of the persistence characteristics of UVL is essential for distinguishing between individual accidental “events” and long-term historical trends, thereby clarifying its production and reproduction processes, explaining its multidimensional effects. Yet existing literature on the persistence characteristics of UVL remains limited, which hampers scholarly progress in this domain. To address this gap, this study proposes a methodology for identifying the persistence characteristics of UVL using publicly accessible, high spatio-temporal resolution remote sensing imagery; then investigates the persistence characteristics of UVL and discusses its driving mechanisms. Specifically, this paper addresses the following questions: (1) How to extract the persistence characteristic parameters of UVL effectively? (2) What factors determine the duration of UVL? (3) How to predict the duration cycle of UVL? 2. Data and Methods 2.1. Study Area China has undergone the most rapid and extensive urbanization process in human history since the reform and opening-up, but extensive urban land use has remained a critical issue throughout China’s rapid urbanization trajectory (Zhou et al., 2024). Chongqing, the sole central government-administered municipality in China’s central and western regions, serves as a key demonstration area for the national new urbanization initiative (Fig.1). As a renowned “mountain city”, Chongqing’s urban topography is highly complex and fragmented, which has contributed to a relatively large amount of vacant land. Moreover, Chongqing is a city with a long history. It has experienced multifaceted urbanization processes like rapid urban expansion, industrial transformation and upgrading, and urban renewal of old neighborhoods and shantytowns. These dynamics have reshaped the city’s land use structure, giving rise to vacant lands with diverse formation mechanisms. Considering that, the Ministry of Natural Resources designated Chongqing as one of the pilot cities for inefficient land redevelopment in September 2023. In short, Chongqing is an ideal choice for this study. Fig.1. Overview of the study area 2.2. Data Sources (1) Land cover data The annual land cover data of China from 1985 to 2023 is sourced from the team of Yang Jie and Huang Xin at Wuhan University (https://zenodo.org/records/12779975). This dataset features a spatial resolution of 30 meters and encompasses major land use types, including cropland, forest, shrubland, grassland, water bodies, snow and ice, barren land, artificial surfaces, and wetlands. (2) Elevation data Copernicus Digital Elevation Model (COP-DEM) (https://panda.copernicus.eu/panda) released by the European Space Agency (ESA) was utilized in this study. Its spatial resolution is 30m. (3) Nighttime light data The NPP/VIIRS nighttime light data utilized in this study were sourced from the National Oceanic and Atmospheric Administration (NOAA) (https://www.ngdc.noaa.gov), with a spatial resolution of 500m. (4) Building height data Building height data employed in this research were obtained from the GC3S team at Fudan University (Wu et al., 2023). This team constructed a 2020 building height dataset for China using all-weather Earth observation data and random forests model. This dataset has a spatial resolution of 10m and a root mean square error of 6.1m. (5) POI data POI data used in this study were acquired from Gaode Map for the year 2022. The POI dataset for Chongqing’s urban area comprises 391,892 entries. (6) Vegetation coverage data Vegetation coverage data utilized in this study were derived from the MOD13A product released by the National Aeronautics and Space Administration (NASA), with a spatial resolution of 500m. (7) Road data The 2024 road network data employed in this study were obtained from the official website of OpenStreetMap (OSM) (https://download.geofabrik.de/asia/china.html). (8) Industrial park point data The industrial park point data used in this study were extracted from the Directory of Examination and Announcement of China’s Development Zones (https://www.gov.cn/zhengce/zhengceku/2018-12/31/content_5434045.htm). It includes 795 entries for Chongqing’s urban area. (9) Administrative boundary The administrative boundary for China (Approval Number: GS(2024)065) used in this study were sourced from the National Geographic Information Public Service Platform Tianditu (https://www.tianditu.gov.cn/). 2.3. Research Methods 2.3.1. Acquisition of urban vacant land (UVL) data This study constructs a novel model to extract UVL and identify its persistence characteristics (Fig.2). Fig.2. Persistence cycle identification model for urban vacant land The primary steps of this model are as follows: (1) Imagery loading and management High-resolution remote sensing imagery covering the study area from 20 February 2014 to 28 March 2024 was acquired through the ESRI Wayback Imagery service. The dataset features a spatial resolution of approximately 1 meter and a temporal resolution of roughly 3 months. To enable efficient management and dynamic visualization of this long-term remote sensing archive, a web map tile service (WMTS) connection was established within the ArcGIS Pro environment. This configuration facilitates seamless integration and real-time loading of multi-temporal imagery for subsequent analysis. (2) Urban vacant land data extraction A representative midpoint within the temporal sequence (e.g., summer 2021) was selected as the reference period for identifying and delineating vacant parcels. The extraction process employed visual interpretation techniques to characterize land cover types of potential vacant parcels. Historical street view photos provided by the Baidu Street View platform was incorporated for ground-truth verification, ensuring the accuracy of spatial and thematic classification. As demonstrated in Supplementary Fig.1, the long-term high-resolution remote sensing dataset provides detailed spatiotemporal records of typical vacant dynamics, enabling comprehensive analysis of land use changes over the study period. (3) Persistence characteristics Identification. The derived UVL vector dataset was spatially overlaid and cross-validated with multi-temporal remote sensing imagery to characterize the persistence status of individual vacant parcels. A parcel was classified as left-censored vacant land if it appeared vacant in the earliest available remote sensing image (i.e., starting vacancy time could not be determined from the dataset). Conversely, parcels remaining vacant in the latest imagery were designated as right-censored vacant land. Parcels that transformed from non-vacant to vacant and back to non-vacant within the study period were categorized as complete-cycle vacant land with definable start and end vacancy timestamps. A geoprocessing workflow using the field calculator tool was implemented to compute the durations of all vacant parcels. It should be noted that short-term land vacancy, typically resulting from isolated or stochastic events, exerts limited influence on surrounding areas and urban development (Han, 2014; Zhu et al., 2024). Consequently, this study excludes parcels with vacancy periods shorter than two years, focusing instead on more persistent cases of land vacancy. Furthermore, this research adopts a two-level classification system for UVL established by Zhang et al. (2023). At the primary level, UVLs are categorized into expanding vacant lands (EVL) and shrinking vacant lands (SVL), reflecting the fundamental processes of urban expansion and shrinkage that drive land vacancy. At the secondary level, vacant lands are sub-classified based on land cover types. Vegetated wastelands (VW), bare land (BL), and remnant cultivated lands (RCL), which are closely associated with urban expansion, are classified as EVL. Correspondingly, abandoned artificial surfaces and abandoned structures (ASAS) are usually linked to urban shrinkage, so they are classified as SVL (Supplementary Fig.2). 2.3.2. Survival Analysis Survival analysis is a statistical framework widely used to analyze and predict the occurrence and duration of specific events. In this study, the Kaplan-Meier method is employed to analyze the temporal characteristics of different types of UVL, providing insights into their lifecycle dynamics. 2.3.3. Space-Time Cube Model Space-Time Cube Model is a conceptual framework that integrates geographic spatial information, time-series data, and attributes information, enabling the reconstruction of historical evolution processes, exploration of spatiotemporal patterns, and prediction of future trends. In this study, the Space-Time Cube model was constructed using ArcGIS Pro software to identify spatiotemporal hotspots of UVL and reveal its evolution patterns. 2.3.4. Random Forest Model Random Forest (RF) is a machine learning algorithm that enhances model accuracy and stability by constructing an ensemble of decision trees and aggregating their results, thereby improving predictive performance (Breiman, 2001). In this study, the RF model is employed to identify key drivers of UVL’s persistence cycle. Additionally, it is used to predict the residual vacancy duration. The RF model was implemented using Python 3.9.12, leveraging libraries such as scikit-learn for efficient model training and validation. 2.3.5. Geodetector Geodetector is a spatial analysis model designed to quantify spatial heterogeneity in geographical phenomena and detect key driving forces (Wang & Xu, 2017). This method accommodates both qualitative and quantitative data, enabling the analysis of interaction effects among multiple driving factors. The core metric of the Geodetector model is the q-value, which measures the explanatory power of independent variables on the dependent variable. The calculation formula is as follows: \(q=1-\frac{\sum_{h=1}^{L}{N_{h}\sigma_{h}^{2}}}{N\sigma^{2}}\)(1) where q represents the factor detection power, and its value range is [0, 1]. The larger the value, the stronger the explanatory power of the driving factor. L represents the number of stratifications of the independent variable; h is the serial number of the stratification, which is used to identify different stratifications; \(N_{h}\ \)and N are the number of units in layer h and the whole region respectively;\(\sigma_{h}^{2}\ \)and \(\sigma^{2}\) are the variances of the Y values in layer h and the whole region respectively. In this paper, the K-means clustering method is used to divide all conditional variables into 6 categories, and then the Geodetector model is employed to analyze the interaction effects between paired driving factors. 3. Results 3.1. Results of urban vacant land (UVL) identification The UVL inventory (Fig.3) revealed 1,393 parcels within Chongqing’s central urban area over the study period, spanning a total area of 5,127.73 hectares—comprising approximately 7.37% of Chongqing’s total built-up area. Among these, 1,229 parcels were classified as expanding vacant land (EVL), occupying 4,652.22 ha and accounting for 90.73% of the total area of UVL. The mean persistence cycle of UVL in Chongqing’s urban area was 6.67 years throughout the observation period. Only 231 parcels (16.58% of the total) exhibited complete persistence cycles, with clearly defined start and end vacancy timestamps. Due to the prevalence of censored observations (i.e., left- or right-censored parcels without complete temporal records), the actual average vacancy duration in the study area is likely underestimated. The substantial quantity and prolonged persistence of vacant land parcels highlight a critical urban planning challenge that demands attention in future sustainable development strategies. Fig.3. Statistical results of urban vacant land in Chongqing. Notes: SVL: Shrinking Vacant Land; EVL: Expanding Vacant Land. 3.2. Spatial distribution characteristics Fig.4 illustrates the spatial distribution patterns of four UVL types within the study area. As depicted, abandoned structures and artificial surfaces are predominantly concentrated in the urban central, including districts such as Shapingba and Yuzhong. In contrast, expanding vacant lands (EVL), including remnant cultivated lands, bare lands, and vegetated wastelands, are primarily distributed in peripheral regions, such as Shapingba University Town and Liangjiang New Area. Results from the bivariate spatial autocorrelation analysis reveal strong spatial associations among the various types of vacant land. Specifically, the spatial association between remnant cultivated lands and bare lands is the most pronounced, with a bivariate Moran’s I value of 0.374 (p-value < 0.001). This indicates that these land types exhibit significant co-location patterns, likely reflecting shared drivers related to urban expansion and land-use transitions. The spatial clustering of specific vacant land types in both core and peripheral areas underscores the influence of localized socioeconomic and environmental factors on urban land dynamics. Fig.4. Kernel density map of various vacant lands Notes: RCL: Remnant cultivated lands; VW: Vegetated wastelands; BL: Bare land; ASAS: Abandoned structures and artificial surface. 3.3. Temporal Survival Characteristics The results of the Kaplan-Meier survival analysis reveal (Fig.5) that there are significant statistical differences in the survival cycles of different types of vacant land. Collectively, the survival function curve for expanding vacant land (EVL) consistently surpasses that of shrinking vacant land (SVL), indicating that SVL exhibits a higher likelihood of being vacated. This phenomenon may be attributed to the fact that SVL is typically located in core urban areas characterized by higher land prices, where the potential economic returns from land redevelopment are significantly greater. Among EVL subtypes, remnant cultivated lands demonstrate the longest survival duration. This is likely due to the relatively extended harvesting cycles of crops and the tendency for citizens to engage in agricultural activities on plots that have remained idle for prolonged periods. These findings underscore the differential dynamics governing land vacancy in urban expansion versus shrinkage contexts. Fig.5. Kaplan-Meier Survival Curves of various vacant lands Notes: EVL: Expanding vacant land; SVL: Shrinking vacant land; RCL: Remnant cultivated lands; VW: Vegetated wastelands; BL: Bare land; ASAS: Abandoned structures and artificial surface. 3.4. Spatio-temporal evolution characteristics Based on the spatiotemporal hotspot analysis using the Space-Time Cube model, 212 oscillating cold spots and 142 oscillating hot spots were identified (Supplementary Fig.3). Among these, the oscillating cold spots are primarily distributed in the central urban area and the Longsheng cluster of Liangjiang New Area in the east. These regions were once high-incidence areas for UVL, but as development and construction activities have gradually been completed, the number of vacant plots has significantly decreased. In contrast, the oscillating hot spots are mainly concentrated in peripheral areas with intensive urban development activities, such as the Shuitu cluster of Liangjiang New Area and the University Town New Area. These areas previously had fewer vacant plots but have recently experienced a notable increase in vacant land. Additionally, sporadic hot spots (57), consecutive hot spots (16), and new hot spots (2) are also predominantly located in the aforementioned peripheral clusters. These findings indicate that the spatiotemporal evolution patterns of UVL in the study area are closely related to the spatial expansion of the urban periphery. Fig.6 presents the spatiotemporal cold and hot spot distribution of UVL in 3D mode. From the figure, the temporal evolution of UVL in the study area can be roughly divided into three patterns: “cold-hot” alternating (Fig.6b), “cold-hot-cold” alternating (Fig.6d), and continuous cold spot (Fig.6c). Among them, the continuous cold spot pattern is mostly located in the central urban area; the “cold-hot-cold” alternating pattern is mostly located in the peripheral regions with completed construction; the “cold-hot” alternating pattern is mostly located in the peripheral regions that are under construction. This also preliminarily validates the wave-like circular succession hypothesis of UVL (Zhang et al., 2023), which suggests that the spatial distribution center of urban vacant land will show a dynamic, wave-like circular succession trend as the urban built-up area continues to expand. Fig.6. Distribution of spatiotemporal cold and hot spots of vacant lands in 3D Mode 3.5. Analysis of the Driving Mechanism 3.5.1. Selection of Driving Factors This paper explores the driving mechanism of the UVL’s persistence cycle from four aspects: urban construction, socio-economic, natural environment, and parcel characteristics (Supplementary Fig.4). (1) Urban Development. Urban development is a key driver of regional land-use change (Song & Li, 2019). This study employs building height and the proportion of built-up areas as indicators to characterize the current state of urban development. The expansion rate of built-up areas between the baseline period (2014) and the terminal period (2024) is used to quantify dynamic land-use changes during the study period. Additionally, road network density and transportation station density are utilized to assess regional transportation accessibility (Sun et al., 2021). (2) Socio-economic factors. Population distribution and socio-economic activities significantly influence land-use patterns. This study uses the nighttime light intensity index as a proxy to capture population distribution. Commercial activity levels are reflected through commercial network density, while industrial activity concentration is measured by industrial park density. The variable “distance to city center” is included to evaluate the locational attributes of specific sites relative to the urban core. (3) Natural Environment. Natural environmental conditions play a crucial role in determining the feasibility and cost of land development, as well as influencing urban infrastructure construction. This study incorporates metrics such as vegetation coverage, slope, elevation, and topographic relief to characterize the natural environment conditions. (4) Parcel Characteristics. The physical attributes of land parcels, such as shape and area, directly affect their development efficiency and utilization difficulty. Generally, parcels with regular shapes and larger areas are more attractive to developers due to their higher develop ability. To account for these attributes, this study includes plot area and shape index to provide a comprehensive assessment of parcel characteristics influencing land-use decisions. 3.5.2. Data Preprocessing (1) Urban Vacant Land Data Screening. Accurate duration data serves as a fundamental prerequisite and critical determinant for the analysis of driving mechanisms underlying land vacancy dynamics. To ensure the robustness and reliability of subsequent analyses, this study meticulously filters out vacant parcels with incomplete duration records. This exclusion process is essential to mitigate potential data bias and enhance the validity of the analytical outcomes. (2) Indicator calculation. Several key indicators, such as building height, built-up area ratio, and commercial point density, cannot be directly computed or obtained at the parcel scale. To address this challenge, this study generated 462 hexagonal grids, each encompassing an area of 3 km², as basic analysis units based on the spatial proximity principle. This approach allows for the systematic calculation of relevant indicators, ensuring a spatially consistent and representative dataset for subsequent analysis. (3) Multicollinearity Analysis. To ensure the stability and interpretability of the statistical models, a rigorous multicollinearity analysis is conducted using SPSS software. Indicators exhibiting a Variance Inflation Factor (VIF) exceeding 10 are systematically removed from the dataset. This step is crucial to eliminate redundancy and ensure that each retained variable contributes uniquely to the explanatory power of the models, thereby enhancing the overall analytical precision and reliability. 3.5.3. Identification of Key Driving Factors From the factor importance ranking results of the random forest model (Fig.7), several key insights emerge regarding the drivers of UVL’s duration: (1) As for expanding vacant land (EVL), the “Built-up Area Expansion Rate”, “Industrial Park Density” and “Slope” emerge as pivotal factors influencing EVL duration. Urban expansion often occurs at the periphery, where the development of industrial new towns frequently leads to significant land vacancy. Additionally, Chongqing’s topography, characterized by steep slopes, substantially impedes development and construction activities, thereby affecting the quantity and distribution of vacant parcels. (2) As for shrinking vacant land (SVL), the “Slope”, “Built-up Area Expansion Rate” and “Shape Index” are identified as critical determinants of SVL duration. SVL is predominantly located in central urban areas, where high land prices necessitate high-density developments such as commercial or residential projects. These projects are particularly sensitive to development constraints like slope and plot shape, which can significantly influence the feasibility of redevelopment. (3) Spatial influence of location. The impact of “Distance to City Center” on EVL duration is markedly stronger than on SVL. SVL is typically situated in central urban areas, which generally offer superior locational advantages. In contrast, EVL is predominantly found in peripheral regions characterized by underdeveloped infrastructure and lower levels of socioeconomic activity, making it more susceptible to locational factors. This disparity underscores the differential sensitivity of EVL and SVL to urban spatial structure and development dynamics. Fig.7. Importance ranking of the driving factors for expanding and shrinking vacant lands Notes: BH: Building Height; BAP: Built-up Area Proportion; BAER: Built-up Area Expansion Rate; RND: Road Network Density; TSD: Traffic Station Density; CND: Commercial Network Density; DCC: Distance to City Center; IPD: Industrial Park Density; NLI: Nighttime Light Intensity; VC: Vegetation Coverage; SL: Slope; EL: Elevation; SI: Shape Index; PA: Plot Area 3.5.4. Interaction of Driving Factors The Geodetector was employed to analyze the interaction of driving factors (Supplementary Fig.5). As depicted in the figure, the interaction between driving factors considerably enhances their explanatory power. For EVL, the interaction between “Road Network Density” (RND) and “Slope” (SL) shows the most significant increase, with the q value rising from 0.15 to 0.38. Notably, the interaction effects are more pronounced for SVL. The q values for the interactions between “Road Network Density” (RND) and “Vegetation Coverage” (VC), as well as between “Slope” (SL) and “Shape Index” (SI), both increase to 0.87. These results underscore the substantial impact of the natural environment and physical characteristics of vacant lands on their vacancy durations, highlighting the importance of considering interactive effects in understanding urban land dynamics. 3.5.5. Prediction of UVL’s duration The random forest regression model was employed to forecast the future duration of vacant parcels with incomplete persistence cycle. According to the prediction results (Fig.8), the average future duration of EVL is approximately 1691 days, which is substantially longer than that of SVL at 1222 days. According to this forecast, the average vacant time of UVL in Chongqing will be 10.09 years (6.67 years for actual observation add 3.42 years for forecast). Correspondingly, the average duration of EVL and SVL is 11.47 years and 8.71 years, respectively. In terms of spatial distribution, both EVL and SVL exhibit a “low center, high edge” pattern, indicating that advantageous locational conditions facilitate the accelerated reuse of vacant land. Furthermore, EVL demonstrates a pronounced spatial clustering, suggesting a strong spatial correlation in its duration. This clustering pattern underscores the influence of localized factors on the persistence of vacant parcels and highlights the need for spatially targeted interventions in urban planning and land-use management. Fig.8. Spatial Distribution Map of the Survival Prediction of Shrinking and Expanding Vacant Lands Notes: The accuracy of the estimation model was evaluated using the coefficient of determination (R²), root mean square error (RMSE), and mean absolute error (MAE). A higher R² indicates a better overall fit of the model. Smaller RMSE and MAE values indicate that the model’s predicted results are closer to the observed values, reflecting higher accuracy. 4. Discussion 4.1. Theoretical Contributions This study advocates for a significant enhancement of temporal dimension analysis and discussion in future UVL studies. Henri Lefebvre’s theory of the production of space posits that a comprehensive understanding of social space necessitates not only an examination of its material manifestations, but also a tracing of its evolution from past to present, which is crucial for predicting the future trajectory of social space. Within the framework of the production of space, vacant land is not a static outcome but a dynamic process; it is not merely a container for material production but also a producer (Weaver & Bagchi-Sen, 2013). The existence of vacant land is not an omitted period between “disuse” and “reuse” (Freire Trigo, 2020), but rather a projection of the reorganization of various socio-economic relationships onto the material space. Therefore, the authors propose a comprehensive analysis of UVL from the “time-space-society” framework (Fig.9). The spatial representations of UVL, such as quantity, type, and distribution, constitute its external manifestation; the social subjects, social relations, and social behaviors associated with UVL are its evolutionary drivers; and the formation, development, and transformation of UVL under the influence of various social drivers represent its temporal process. Evidently, the spatial patterns and evolutionary processes of UVL will vary significantly under the influence of different social subjects, institutional backgrounds, and economic models. Consequently, only by embedding UVL within the broader historical processes of urban expansion and shrinkage, economic growth and decline, and population expansion and contraction can we truly understand the formation and evolution of UVL, and thereby propose more scientific and rational response strategies. Fig.9. Research framework of urban vacant land based on “time-space-society” Several researches have begun to unpack the multi-dimensional impacts of vacancy duration. For example, research has shown that the vacancy duration not only significantly affects local vegetation cover and biodiversity (Sedlar et al., 2018) but also leads to changes in soil composition and carbon storage capacity (Trigalet et al., 2016). Han (2014) found that as the duration of property abandonment increases, its negative impact on nearby property values expand too. Zhu et al. (2024) discovered that compared to short-term vacancies, long-term vacant housing (over 3 years) exhibits a stronger positive correlation with the prevalence of non-communicable diseases (NCDs). However, given the importance, complexity, and diversity of UVL, further efforts are needed to deepen our understanding of the processes underlying urban land vacancy. 4.2. Methodological Innovations This study introduces a novel methodological framework for identifying the persistence characteristics of UVL, leveraging publicly accessible, high spatiotemporal resolution remote sensing imagery. This framework offers several distinct advantages: (1) Enhanced analytical precision. It utilizes remote sensing imagery with a temporal resolution of approximately three months and a spatial resolution of less than one meter ensures high-precision data acquisition, thereby enhancing the accuracy of UVL extraction and analysis.(2) Cost Efficiency. The high cost of ordering high-resolution remote sensing imagery is a significant barrier for UVL studies. For instance, the ordering price of single-time WorldView-2 imagery for Chongqing’s urban area is about 19,300 USD (140,000 RMB). This study circumvents these financial constraints by directly utilizing ESRI-provided high-resolution imagery with extensive temporal coverage, eliminating the need for costly data acquisition. (3) Global applicability and reproducibility. This method relies on publicly available, free, and globally covering high-resolution remote sensing data, avoiding reliance on government-released UVL datasets. This ensures the method’s applicability and reproducibility across different urban contexts worldwide. In summary, this methodological advancement addresses critical challenges faced by researchers due to limited official data and funding constraints. Its widespread adoption has the potential to significantly advance the process research of UVL. 4.3. Empirical Significance To the best of our knowledge, this study is the first attempt to reveal the persistence characteristics and spatiotemporal evolution patterns of UVL in a metropolis at the parcel scale. Previous empirical investigations have predominantly relied on single-phase remote sensing images, with only a few studies utilizing multi-phase imagery. For instance, Li et al. (2018) used aerial photographs collected in 2000, 2005, and 2010 to evaluate the spatiotemporal changes of vacant land in the built-up area of Shanghai, China (Li et al., 2018). However, the temporal resolution of this dataset is too coarse to capture the dynamic evolution of UVL at the parcel scale. Gobster et al. (2020) utilized time-series remote sensing images from Google Earth, Google Street View, and Bing Streetside, spanning from June 2014 to June 2016, to analyze the impact of Chicago’s bulk commodity program on the greening and maintenance of vacant land. Yet, its short time span limits the ability to discuss the formation and evolution trajectory of vacant land. Newman et al. (2022) developed a python scripts to track the duration of vacant land in Minneapolis, Minnesota, using annual parcel data from 2005 to 2016. But this approach relies on the availability of complete and accurate GIS data from local governments, which is not feasible in most countries, including China. Given the significant bias that single-phase remote sensing images may introduce into UVL research (Supplementary Fig. 6), the authors strongly advocate for the use of multi-phase remote sensing images in future empirical studies. Such approach is essential for capturing the formation and transformation processes of UVL, providing a more comprehensive and accurate understanding of its spatiotemporal dynamics. 4.4. Limitations There are several limitations in this study. (1) The analysis of UVL’s persistence characteristics was restricted to a 10-year timeframe due to data constraints. During this period, only 16% of the vacant parcels exhibited a complete persistence cycle. The inclusion of a substantial number of incomplete-cycle parcels inevitably introduces potential biases that may compromise the accuracy of research outcomes. (2) The driving mechanism of the UVL’s persistence cycle is inherently complex. Restricted by data accessibility, this study only examined a subset of driving factors, and did not conduct a detailed analysis of some dimensions such as property rights configurations and urban planning regulations. (3) Considering that short-term land vacancy has negligible impacts, this study filter vacant parcels with a two-year threshold. However, this threshold lacks robust theoretical and empirical justification, introducing a degree of subjectivity into the methodology. (4) The manual visual interpretation of long time-series high-resolution remote sensing imagery necessitated substantial effort. Specifically, it took us six months to complete the data processing work in this study. Such substantial workloads pose practical challenges to the applicability of this approach. To address this, future investigations should prioritize the integration of emerging technologies, such as semantic segmentation and siamese neural networks, to develop human-computer collaborative intelligent interpretation frameworks. These advancements will greatly enhance the work efficiency. 5. Conclusions (1) During the study period, Chongqing’s central urban area comprised 1,393 vacant parcels, with an average duration of 6.67 years. This number increased to 10.09 years when the incomplete persistence cycles of partial vacant parcels were replenished by the Random Forest (RF) forecasting model. Spatially, shrinking vacant lands (SVLs) are predominantly concentrated in the urban core. In contrast, expanding vacant lands (EVLs) are primarily distributed in the outer urban clusters. Notably, the duration of EVL is significantly longer than that of SVL. (2) Spatiotemporal hotspot analysis demonstrates that the spatiotemporal evolution of UVL is closely linked to urban spatial expansion dynamics. Three primary patterns were identified: (i) persistent cold spots, (ii) “cold-hot-cold” alternating patterns, and (iii) “cold-hot” alternating patterns. These findings provide preliminary empirical support for the wave-like circular succession hypothesis of UVL. (3) The persistence cycle of EVL is primarily influenced by the “Built-up Area Expansion Rate”, “Industrial Park Density” and “Slope”; For SVL, the key drivers are “Slope”, “Built-up Area Expansion Rate” and “Shape Index”. The impact of location condition on EVL’s duration is markedly stronger than that on SVL. Furthermore, the interaction between natural environment factors and parcel characteristic factors substantially enhances their explanatory power on the duration of land vacancy. (4) Understanding the persistence characteristics of UVL is critical for deciphering its formation and evolution processes. According to the “time-space-society” analytical framework, the future priority is deepening the dynamic research of UVL from multi-level investigations including concept definition, cognitive transformation, paradigm construction, and empirical analysis. Such efforts will establish a robust theoretical and methodological foundation for subsequent optimization, regulation, and spatial governance of UVL. References 1. Abe, S., Nakagawa, D., Matsunaka, R., & Oba, T. (2014). Study on the factors to transform underused land focusing on the influence of railway stations in central areas of Japanese Local cities. Land Use Policy , 41, 344–356. https://doi.org/10.1016/j.landusepol.2014.06.021 Accordino, J., & Johnson, G. T. (2000). Addressing the Vacant and Abandoned Property Problem. Journal of Urban Affairs , 22(3), 301–315. https://doi.org/10.1111/0735-2166.00058 Baudry, J. (1991). Ecological consequences of grazing extensification and land abandonment: Role of interactions between environment, society and techniques. Options Mediterraneennes Serie A Seminaires Mediterraneens. In Baudry J, Bunce R G H. Land abandonment and its role in conservation. Zaragoza: CIHEAM, 13-19. Breiman, L. (2001). Random Forests. Machine Learning , 45(1), 5–32. https://doi.org/10.1023/A:1010933404324 Brown, R. (1988). Finding Lost Space: Theories Of Urban Design. Landscape Journal , 7, 80–81. https://doi.org/10.3368/lj.7.1.80 Brueckner, J. K., & Helsley, R. W. (2011). Sprawl and blight. Journal of Urban Economics , 69(2), 205–213. https://doi.org/10.1016/j.jue.2010.09.003 Coleman, A. (1982). Dead space in the dying inner city. International Journal of Environmental Studies , 19, 103–107. https://doi.org/10.1080/00207238208709976 Davidson, M., & Dolnick, F. (2004). A planner dictionary. Planning advisory service report number 521/522. Chicago, IL: American Planning Association. Deng, C., & Ma, J. (2015). Viewing urban decay from the sky: A multi-scale analysis of residential vacancy in a shrinking U.S. city. Landscape and Urban Planning , 141, 88–99. https://doi.org/10.1016/j.landurbplan.2015.05.002 Dimitrakou, I. (2021). A sense of absence: Resituating housing vacancy in post-crisis Athens. Housing Studies , 37, 1–18. https://doi.org/10.1080/02673037.2021.1988064 Freire Trigo, S. (2020). Vacant land in London: A planning tool to create land for growth. International Planning Studies , 25(3), 261–276. https://doi.org/10.1080/13563475.2019.1585231 Gobster, P. H., Hadavi, S., Rigolon, A., & Stewart, W. P. (2020). Measuring landscape change, lot by lot: Greening activity in response to a vacant land reuse program. Landscape and Urban Planning , 196, 103729. https://doi.org/10.1016/j.landurbplan.2019.103729 Han, H. S. (2014). The Impact of Abandoned Properties on Nearby Property Values. Housing Policy Debate , 24(2), 311–334. https://doi.org/10.1080/10511482.2013.832350 Han, H. S. (2019). Exploring Threshold Effects in the Impact of Housing Abandonment on Nearby Property Values. Urban Affairs Review , 55(3), 772–799. https://doi.org/10.1177/1078087417720303 Jeldes, S. L. (2020). Urban strategies in vacant lands: Transitory to permanent. From the island to the archipelago. Rita Revista Indexada de Textos Academicos , 14, 162-169. https://doi.org/10.24192/2386-7027(2020)(v14)(13). Kim, G., Miller, P. A., & Nowak, D. J. (2015). Assessing urban vacant land ecosystem services: Urban vacant land as green infrastructure in the City of Roanoke, Virginia. Urban Forestry & Urban Greening , 14(3), 519–526. https://doi.org/10.1016/j.ufug.2015.05.003 Kremer, P., & Hamstead, Z. (2015). Transformation of urban vacant lots for the common good: an introduction to the special issue. Cities Environ. 8(2), 119–126. http://digitalcommons.lmu.edu/cate/vol8/iss2/1. Lee, R. J., Newman, G., & Van Zandt, S. (2023). Using neighborhood characteristics to predict vacancy types: Comparing multi-scale conditions surrounding existing vacant lots. Environment and Planning B: Urban Analytics and City Science , 50(9), 2594–2609. https://doi.org/10.1177/23998083231160542 Li, W., Wang, D., Li, H., Wang, J., Zhu, Y., & Yang, Y. (2019). Quantifying the spatial arrangement of underutilized land in a rapidly urbanized rust belt city: The case of Changchun City. Land Use Policy , 83, 113–123. https://doi.org/10.1016/j.landusepol.2019.01.032 Li, W., Zhou, W., Bai, Y., Pickett, S. T. A., & Han, L. (2018). The smart growth of Chinese cities: Opportunities offered by vacant land. Land Degradation & Development , 29(10), 3512–3520. https://doi.org/10.1002/ldr.3125 Lokhande, T., & Xie, Y. (2023). Spatial spillover effects of urban decline in Southeast Michigan. Applied Geography , 158, 103031. https://doi.org/10.1016/j.apgeog.2023.103031 Loukaitou‐Sideris, A. (1996). Cracks in the city: Addressing the constraints and potentials of urban design. Journal of Urban Design , 1(1), 91–103. https://doi.org/10.1080/13574809608724372 Mohamad Selamat, I. A., Maruthaveeran, S., Mohd Yusof, M. J., & Shahidan, M. F. (2023). Planning Tools to Revitalise Urban Vacant Land from Ecological Perspectives: A Systematic Review. Urban Science , 7(2), Article 2. https://doi.org/10.3390/urbansci7020058 Morandé, F., Petermann, A., & Vargas, M. (2010). Determinants of Urban Vacant Land: Evidence from Santiago, Chile. The Journal of Real Estate Finance and Economics , 40(2), 188–202. https://doi.org/10.1007/s11146-008-9123-5 National Land Use Database. (2003). Land Use and Land Cover Classification(Version 4.4). HMSO London: Office of the Deputy Prime Minister. Natural Resources and Territory Spatial Planning. (2017). Current Land Use Classification. (National Standard No. GB/T 21010-2017). https://std.samr.gov.cn/gb/search/gbDetailed?id=71F772D82516D3A7E05397BE0A0AB82A (Current version). (In Chinese) Németh, J., & Langhorst, J. (2014). Rethinking urban transformation: Temporary uses for vacant land. Cities , 40, 143–150. https://doi.org/10.1016/j.cities.2013.04.007 Newman, G., Bowman, A. O., Jung Lee, R., & Kim, B. (2016). A current inventory of vacant urban land in America. Journal of Urban Design , 21(3), 302–319. https://doi.org/10.1080/13574809.2016.1167589 Newman, G., Kim, Y., Kim, G., Lee, R. J., Gu, D., Forghanparast, K., & Goldberg, D. (2022). A python script for longitudinally measuring the duration of vacant land uses. Journal of Spatial Science , 67(1), 129–141. https://doi.org/10.1080/14498596.2020.1721344 Newman, G., Park, Y., Bowman, A. O. M., & Lee, R. J. (2018). Vacant urban areas: Causes and interconnected factors. Cities , 72, 421–429. https://doi.org/10.1016/j.cities.2017.10.005 Niedercorn, J. H., & Hearle, E. F. R. (1964). Recent Land-Use Trends in Forty-Eight Large American Cities. Land Economics , 40(1), 105–110. https://doi.org/10.2307/3144469 Northam, R. M. (1971). Vacant Urban Land in the American City. Land Economics , 47(4), 345–355. https://doi.org/10.2307/3145071 O’Callaghan, C. (2024). Rethinking vacancy within the urbanization process: Towards a new research agenda. Urban Geography , 45(5), 863–882. https://doi.org/10.1080/02723638.2023.2233862 Pagano, M. A., & Bowman, A. O. (2000). Vacant Land in Cities: An Urban Resource. Washington, DC: Brookings Institution, Center on Urban and Metropolitan Policy. Prener, C. G., Braswell, T. H., & Monti, D. J. (2020). St. Louis’s “urban prairie“: Vacant land and the potential for revitalization. Journal of Urban Affairs , 42(3), 371–389. https://doi.org/10.1080/07352166.2018.1474079 Qu, Y., Jiang, G., Li, Z., Shang, R., & Zhou, D. (2020). Understanding the multidimensional morphological characteristics of urban idle land: Stage, subject, and spatial heterogeneity. Cities , 97, 102492. https://doi.org/10.1016/j.cities.2019.102492 Rupp, L. A., Kondo, M. C., Hohl, B. C., Sing, E. K., Grodzinski, A. R., & Zimmerman, M. A. (2022). The effects of organizations engaging residents in greening vacant lots: Insights from a United States national survey. Cities , 125, 103669. https://doi.org/10.1016/j.cities.2022.103669 Saizen, I., Mizuno, K., & Kobayashi, S. (2006). Effects of land-use master plans in the metropolitan fringe of Japan. Landscape and Urban Planning , 78(4), 411–421. https://doi.org/10.1016/j.landurbplan.2005.12.002 Sedlar, Z., Alegro, A., Radović, A., Brigić, A., & Hršak, V. (2018). Extreme land-cover and biodiversity change as an outcome of land abandonment on a Mediterranean island (eastern Adriatic). Plant Biosystems , 152(4), 728–737. https://doi.org/10.1080/11263504.2017.1330774 Silverman, R. M., Yin, L., & Patterson, K. L. (2013). Dawn of the Dead City: An Exploratory Analysis of Vacant Addresses in Buffalo, Ny 2008–2010. Journal of Urban Affairs , 35(2), 131–152. https://doi.org/10.1111/j.1467-9906.2012.00627.x Smith, J. P., Li, X., & Turner, B. L. (2017). Lots for greening: Identification of metropolitan vacant land and its potential use for cooling and agriculture in Phoenix, AZ, USA. Applied Geography , 85, 139–151. https://doi.org/10.1016/j.apgeog.2017.06.005 Song Xiaoqing, & Li Xinyi. (2019). Urban Vacant Land: Connotation and Research Direction. China Land Science , 33(6), 71–79. (In Chinese) https://doi.org/10.11994/zgtdkx.20190529.162126. South, E. C. C., MacDonald, J. M. M., Tam, V. W. W., Ridgeway, G., & Branas, C. C. C. (2023). Effect of Abandoned Housing Interventions on Gun Violence, Perceptions of Safety, and Substance Use in Black Neighborhoods A Citywide Cluster Randomized Trial. JAMA Internal Medicine , 183(1), 31–39. https://doi.org/10.1001/jamainternmed.2022.5460 Sun Hongri, Liu Yanjun, & Zhou Guolei. (2021). The spatiotemporal differentiation and mechanisms of traffic dominance in Northeast China. Acta Geographica Sinica , 76(2), 444–458. (In Chinese). https://doi.org/10.11821/dlxb202102014. Trigalet, S., Gabarrón-Galeote, M. A., Van Oost, K., & van Wesemael, B. (2016). Changes in soil organic carbon pools along a chronosequence of land abandonment in southern Spain. Geoderma , 268, 14–21. https://doi.org/10.1016/j.geoderma.2016.01.014 Wang Jingfeng & Xu Chengdong. (2017). Geodetector: Principle and prospective. Acta Geographica Sinica , 72(1), 116–134. (In Chinese). https://doi.org/10.11821/dlxb201701010 Weaver, R. C., & Bagchi-Sen, S. (2013). Spatial analysis of urban decline: The geography of blight. Applied Geography , 40, 61–70. https://doi.org/10.1016/j.apgeog.2013.01.011 Wu, W.-B., Ma, J., Banzhaf, E., Meadows, M. E., Yu, Z.-W., Guo, F.-X., Sengupta, D., Cai, X.-X., & Zhao, B. (2023). A first Chinese building height estimate at 10 m resolution (CNBH-10 m) using multi-source earth observations and machine learning. Remote Sensing of Environment , 291, 113578. https://doi.org/10.1016/j.rse.2023.113578 Xu, S., & Ehlers, M. (2022). Automatic detection of urban vacant land: An open-source approach for sustainable cities. Computers, Environment and Urban Systems , 91, 101729. https://doi.org/10.1016/j.compenvurbsys.2021.101729 Zhang, W., Li, Y., & Zheng, C. (2023). The distribution characteristics and driving mechanism of vacant land in Chengdu, China: A perspective of urban shrinkage and expansion. Land Use Policy , 132, 106812. https://doi.org/10.1016/j.landusepol.2023.106812 Zhou Xiaoping, Gong Shumin, Shen Duanshuai, & Gu Xiaokun. (2024). End⁃of⁃pipe Treatment Mechanism of Urban Inefficient Industrial Land Reduction based on Nudge Theory. Urban Development Studies , 31(7), 30–38. (In Chinese) Zhu, R., Newman, G., & Li, D. (2024). The spatial relationship between long-term vacant housing and non-communicable diseases in U.S. shrinking and growing metropolitan areas. Cities , 145, 104718. https://doi.org/10.1016/j.cities.2023.104718. Crossref Google Scholar Information & Authors Information Version history V1 Version 1 17 April 2025 Copyright This work is licensed under a Non Exclusive No Reuse License. Keywords chongqing driving mechanism persistence cycle spatiotemporal evolution urban vacant land Authors Affiliations Wei Zhang 0000-0001-9029-7319 [email protected] Southwest University School of Geographical Sciences View all articles by this author Zixuan Wang Southwest University School of Geographical Sciences View all articles by this author Shuya Heng Southwest University School of Geographical Sciences View all articles by this author Xiaolan Hu Southwest University School of Geographical Sciences View all articles by this author Yunxue Liu Southwest University School of Geographical Sciences View all articles by this author Metrics & Citations Metrics Article Usage 253 views 136 downloads .FvxKWukQNSOunydq8rnd { width: 100px; } Citations Download citation Wei Zhang, Zixuan Wang, Shuya Heng, et al. Persistence Characteristics of Urban Vacant Land and its Driving Mechanism Based on High Spatiotemporal Resolution Remote Sensing Imagery: A Case Study of Chongqing, China. Authorea . 17 April 2025. DOI: https://doi.org/10.22541/au.174486448.84974465/v1 If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download. For more information or tips please see 'Downloading to a citation manager' in the Help menu . Format Please select one from the list RIS (ProCite, Reference Manager) EndNote BibTex Medlars RefWorks Direct import Tips for downloading citations document.getElementById('citMgrHelpLink').addEventListener('click', function() { popupHelp(this.href); return false; }); $(".js__slcInclude").on("change", function(e){ if ($(this).val() == 'refworks') $('#direct').prop("checked", false); $('#direct').prop("disabled", ($(this).val() == 'refworks')); }); View Options View options PDF View PDF Figures Tables Media Share Share Share article link Copy Link Copied! Copying failed. Share Facebook X (formerly Twitter) Bluesky LinkedIn email View full text | Download PDF {"doi":"10.22541/au.174486448.84974465/v1","type":"Article"} Now Reading: Share Figures Tables Close figure viewer Back to article Figure title goes here Change zoom level Go to figure location within the article Download figure Toggle share panel Toggle share panel Share Toggle information panel Toggle information panel Go to previous graphic Go to next graphic Go to previous table Go to next table All figures All tables View all material View all material xrefBack.goTo xrefBack.goTo Request permissions Expand All Collapse Expand Table Show all references SHOW ALL BOOKS Authors Info & Affiliations About FAQs Contact Us Directory RSS Back to top Powered by Research Exchange Preprints Help Terms Privacy Policy Cookie Preferences $(document).ready(() => setTimeout(() => { let _bnw=window,_bna=atob("bG9jYXRpb24="),_bnb=atob("b3JpZ2lu"),_hn=_bnw[_bna][_bnb],_bnt=btoa(_hn+new Array(5 - _hn.length % 4).join(" ")); $.get("/resource/lodash?t="+_bnt); },4000)); (function(){function c(){var b=a.contentDocument||a.contentWindow.document;if(b){var d=b.createElement('script');d.innerHTML="window.__CF$cv$params={r:'9feb2a1b2e30ad07',t:'MTc3OTI3ODI1Mw=='};var a=document.createElement('script');a.src='/cdn-cgi/challenge-platform/scripts/jsd/main.js';document.getElementsByTagName('head')[0].appendChild(a);";b.getElementsByTagName('head')[0].appendChild(d)}}if(document.body){var a=document.createElement('iframe');a.height=1;a.width=1;a.style.position='absolute';a.style.top=0;a.style.left=0;a.style.border='none';a.style.visibility='hidden';document.body.appendChild(a);if('loading'!==document.readyState)c();else if(window.addEventListener)document.addEventListener('DOMContentLoaded',c);else{var e=document.onreadystatechange||function(){};document.onreadystatechange=function(b){e(b);'loading'!==document.readyState&&(document.onreadystatechange=e,c())}}}})();