Ecosystem protection and food security are often viewed as competing goals, creating a critical policy dilemma for agricultural regions. Here we show that grassland restoration under China's Grassland Ecological Compensation Policy significantly enhances maize yields through local climate regulation. Using a difference-in-differences design with county-level panel data, we find that restored grasslands reduce growing-season temperatures by 0.1 degrees C and increase precipitation by 11.48 mm, thereby mitigating heat and drought stress during critical reproductive stages. These changes extend the reproductive growing period by 0.93 days, increase yields by 7.76% (0.437 t ha-1) and reduce crop shortfall risk by 25.9%. Yield gains alone offset more than 80% of programme costs within 5 years and the additional production could alleviate 10% of the maize import deficit in the Northern Spring Maize Region. Our findings highlight ecosystem restoration as a scalable strategy for achieving climate-resilient food security in regions similar to our study area.

Vegetation productivity exhibits substantial spatiotemporal heterogeneity under the combined influence of climate change and anthropogenic disturbance. However, conventional long-term assessments focus on absolute changes, neglecting ecological heterogeneity and climatic confounding, thus limiting their utility for ecosystem management. From ecological and human activity perspectives, we established a reference-based framework using undisturbed ecosystems as ecological baselines, to detect subtle vegetation productivity dynamics across China between 2000 and 2020, capturing deviations between observed and reference productivity. The results showed that the RBRA significantly improved the detection of productivity declines, identifying a decline-affected area in China 39 percentage points larger than that of the CAA, particularly in temperate humid zones and northern regions with high human activity. By combining RBRA and CAA, four vegetation productivity change categories (sustained improvement, potential improvement, latent degradation, and sustained degradation) were identified. At the ecological zone level, the Tibetan Plateau emerged as the primary hotspot of relative decline (80.41%), while over 50% of the humid south-central region achieved sustained improvement. At the human activity zone level, high-intensity zones exhibited over 42% latent degradation, whereas low-intensity areas were dominated by sustained improvement (48-61%). Attribution analysis revealed that while climatic constraints remained the dominant drivers of vegetation productivity change, degradation risks driven by land-use change and economic expansion were more clearly captured by the RBRA framework, particularly in vulnerable and human-impacted regions. Overall, the reference-based framework enhances sensitivity to subtle vegetation change signals and provides a robust scientific basis for targeted conservation and adaptive ecosystem management.

Urbanization progressively intensifies the urban heat island (UHI) effect, rendering the quantitative evaluation of spatial ecological indicators critical for monitoring urban climate regulation services, particularly under severe land resource constraints. This study investigates four rapidly urbanizing municipalities in China—Beijing, Tianjin, Shanghai, and Chongqing—that have implemented extensive greening initiatives. We systematically analyzing the spatiotemporal evolution of the urban heat island (UHI) and urban green space (UGS) from 2000 to 2020, and quantitatively assess the contribution of UGS changes to UHI intensity. The results indicate that the total UHI area within built-up areas in these four municipalities expanded progressively with urbanization. The mean UHI intensity varying from 2.6 °C to 4.2 ℃, however, an inflection point occurred around 2015, capturing the sensitive ecological response to national greening interventions. Post-2015, the expansion of severe UHI areas slowed markedly in Beijing and Tianjin, while in Chongqing, these areas stabilized after a slight decrease. Spatial metrics analysis revealed that while UGS composition remains the paramount driver of urban cooling, the synergistic interaction between composition and configuration metrics significantly surpassed the contribution of any single metric alone. Overall, strategic urban governance and the dedication of land for greening in these municipalities have effectively promoted vegetation restoration and mitigated UHI effects. These findings provide replicable decision-support frameworks for land-constrained megacities globally to strengthen climate adaptation and promote sustainable development.

Identifying key targets for air pollution prevention and control while achieving synergy between pollution reduction and carbon mitigation is crucial for maximizing benefits to human health and the climate. To evaluate recent ambient air quality trends in Shenyang, this study analyzed atmospheric pollutant concentration data from 2020 to 2024, with a focus on two primary pollutants: fine particulate matter (PM2.5) and ozone (O3). The Positive Matrix Factorization (PMF) model was employed to identify the emission sources of PM2.5 and carbon dioxide (CO2), as well as their common contributing sources. The results show that in recent years, the overall ambient air quality in Shenyang has improved, while ozone pollution has intensified. Particulate matter (PM) remains the dominant factor contributing to air quality deterioration. Among northern Chinese cities, coal-related emission sources continue to be the main contributors to PM2.5. From the perspective of synergistic governance, coal combustion, vehicle exhaust emissions, industrial activities, and biomass burning all have significant impacts on the concentrations of both PM2.5 and CO2. Notably, vehicle exhaust emissions exert a substantial influence across all seasons throughout the year. These findings provide a scientific basis for Shenyang to implement targeted pollution control measures and address the complex challenge of air pollution.

Growing agricultural water demand, driven by climate change and land-use intensification, is accelerating global water scarcity and threatening food and environmental security. This study quantifies spatiotemporal changes in crop water requirements (CWR) and irrigation water requirement (IWR) from 1980 to 2017 for wheat, maize, and soybean. A corrected FAO crop coefficient method was used to estimate global CWR, while the logarithmic mean Divisia index (LMDI) was applied to decompose its drivers into climate and crop area changes. IWR was calculated to evaluate the increasing water stress in four representative river basins: the Haihe (HRB), Yellow (YRB), Mississippi (MRB), and Ganges (GRB) river basins. Multiple linear regression models were used to identify dominant drivers of water stress. Results show that from 1980 to 2017, CWR increased significantly for maize (+210 x 108 m3) and soybean (+523 x 108 m3) primarily due to crop area expansion, while wheat CWR declined (-109 x 108 m3). Area growth contributed over +850 x 108 m3 to global CWR increases. At the basin scale, IWR rose notably in HRB, YRB, and GRB, but declined in MRB. Regression analysis confirms that crop area change was the dominant driver of variations in IWR, particularly for soybean in HRB and maize in YRB, while precipitation exerted strong negative effects in some regions. This study provides a scalable framework for diagnosing agricultural water stress and its key drivers, supporting climate adaptation and irrigation planning under global change.

Shanxi Province carries a profound historical and cultural accumulation and diversified regional and cultural characteristics. Comprehensive research on the spatial Characteristics of intangible cultural heritage in Shanxi Province and its influencing factors is of great academic significance for the protection and inheritance of China’s traditional cultural. This research focuses on 1,178 items of intangible cultural heritage in Shanxi Province, utilizing GIS spatial analysis techniques such as nearest neighbor index, kernel density estimation, and geographical detecto to investigate the spatial distribution patterns, and influencing factors of 1,178 intangible cultural heritage items documented across 11 municipal-level administrative units within Shanxi Province. Research shows: Shanxi’s intangible cultural heritage is generally distributed in a significant cluster state, and its spatial distribution characteristics are highly dense in central Shanxi, significantly clustered in southern Shanxi, and relatively dispersed in northern Shanxi, forming an overall structural feature of “four-core and two-belt”. In terms of influencing factors, cultural environmental factors have the most significant impact on the distribution of intangible cultural heritage, followed by socioeconomic factors. Natural geographic factors have a relatively minor direct impact on the distribution of intangible cultural heritage, but they indirectly influence it through interactions with other factors, which indicates that the distribution of intangible cultural heritage is the result of the joint action of natural support, social and economic evolution and cultural ecological construction. Finally, research proposes: the sustainable development of intangible cultural heritage in Shanxi Province should be realized by strengthening the construction of cultural institutions, improving the standard of education and promoting the integration of culture and tourism.

Accelerating urbanization intensifies urban heat exposure and extends its influence beyond summer to other seasons. However, fine-scale evidence on seasonal thermal dynamics and associated population exposure remains limited. By integrating MODIS daily land surface temperature (LST) with population density, we developed a four-quadrant framework to characterize the seasonal and spatial patterns of LST and Heat Exposure Index (HEI) in Shanghai, the commercial and financial center of mainland China. Results revealed a significant warming trend in Shanghai’s LST, with an average increase of 0.087 °C·yr− 1, most prominently in spring (0.175 °C·yr− 1). While urban new towns and rural areas experienced rapid warming, urban cores showed significant cooling trends (SlopeLST = -0.10 °C·yr− 1), likely associated with green infrastructure and urban renewal policies. Seasonally, summer exhibited the highest heat exposure risk (HEI = 6.84), with over 82

The long-term evolution and underlying drivers of urban-rural ozone (O3) differences across China remain unclear. Here, we identify a national-scale reversal in the urban-rural Maximum Daily 8 h Average (MDA8) O3 difference (urban minus rural), shifting from higher O3 in rural areas (-2.3 mu g m-3 in 2013) to higher O3 in urban areas (1.2 mu g m-3 in 2020) based on the reanalysis data, indicating increasingly intensified O3 exposure in densely populated cities. Using chemical transport model simulations, we demonstrate that changing meteorological conditions account for a notable positive shift in the urban-rural O3 difference across nearly half of the Chinese provinces during 2013-2019. Meanwhile, larger reductions in urban nitrogen oxides (NO x ) emissions relative to rural areas emerge as the dominant anthropogenic driver, contributing 4.9 mu g m-3 in summer and 3.7 mu g m-3 in winter to increasing urban-rural O3 difference. This transition has reshaped ozone-NO x -volatile organic compounds (VOC) sensitivity, shifting optimal O3 production zones from rural to urban areas across most regions in summer. Our findings highlight the need for strengthened urban NO x mitigation strategies while accounting for the growing ozone risk associated with intensifying UHI effects.

China faces the dual challenge of improving ambient air quality while mitigating climate change, yet synergistic control strategies are often formulated at national or provincial scales due to the lack of detailed county-level emission information. This study integrates multi-source data within a unified framework to develop a comprehensive county-level emission inventory of major air pollutants and CO2 for China. The results reveal pronounced spatial heterogeneity in emissions. A total of 929 counties, accounting for approximately 33% of all counties and 24% of China's land area, contribute about 80% of national CO2 emissions and 41 - 67% of air pollutant emissions, indicating the emergence of county-level joint emission hubs across multiple pollutants. Emission intensities are substantially higher in small and medium-sized cities than in large-sized cities, underscoring their underestimated environmental pressures during economic growth. Comparison with conventional GDP-based allocation approaches shows that emissions in small and medium-sized cities are systematically underestimated, whereas those in large-sized cities are overestimated. These findings demonstrate the limitations of economic proxy-based allocation methods and provide a refined evidence base for targeted, synergistic air quality management and carbon mitigation at city and intra-city scales in China.

Urban green spaces (UGSs) provide essential ecological functions in highly urbanized landscapes, yet city-scale associations between soil conditions and vegetation greenness remain insufficiently understood due to limited field-based evidence. This study investigated Shanghai, China, by integrating long-term Enhanced Vegetation Index (EVI) time-series data from 2000 to 2020 with field soil sampling from 183 UGS sites, covering 18 edaphic properties. Together with vegetation baseline, human activity, and climate covariates, we assessed the relative and spatially heterogeneous associations of soil conditions with vegetation greenness dynamics (VGD) and current vegetation greenness (VG). Shanghai's UGSs showed substantially stronger greening than the broader urban background, with a mean VGD of 0.074 decade−1 compared with 0.017 decade−1 citywide, and 87.98% of sampled sites exhibited positive greening trends. Park green spaces maintained higher VG and stronger VGD than roadside green spaces, while VGD and VG were significantly but only moderately associated, indicating that they represent distinct dimensions of UGS performance. Attribution analysis showed that vegetation baseline and soil properties were the dominant contributors to VGD, explaining 17% and 11% of the variance, respectively. For VG, soil emerged as the leading determinant, independently explaining 15% of spatial variation, with associations distributed across multiple edaphic dimensions rather than controlled by a single soil property. Spatial modeling further revealed scale-dependent soil associations: soil–VGD associations were strongest in intermediate and outer urban zones, whereas soil–VG associations were more outward-oriented and more frequently negative. These findings provide city-scale empirical evidence that soil conditions are associated with urban greenness patterns in spatially heterogeneous ways, highlighting the need for soil-sensitive UGS assessment and management across urban gradients.

Long-term monitoring of vegetation dynamics is essential for assessing ecosystems resilience and responses to climate change.The China Ecosystem Research Network(CERN)is a national long-term observation network that provides an ecological baseline across Chi-na's major ecosystems.This study analyzed vegetation trends and their driving factors from 2000 to 2024 across 34 CERN field stations and their surrounding areas.An intercomparison of multiple NDVI products revealed substantial inconsistencies.MODIS NDVI exhibited su-perior temporal stability and spatial coherence and was therefore selected for long-term analysis.The mean NDVI at CERN stations(0.419)was 37.8％higher than the national av-erage(0.304),and showed significantly faster greening trends(0.022/10 a)compared to suburban(0.013/10 a)and rural(0.017/10 a)areas,reflecting effective vegetation restoration and stable ecosystem management.Vegetation changes at CERN field stations were pre-dominantly governed by climatic rather than anthropogenic factors.Among different ecosys-tem types,air temperature(AT),sunshine duration(SD),and relative humility(RH)were the dominant drivers in farmland,forest,and wetland.For grassland,AT and SD were the main drivers,whereas AT and RH exerted the strongest influence in the desert ecosystem.These findings confirm the representativeness of CERN stations and underscore the predominant role of climatic factors in shaping long-term vegetation trajectories across China's diverse ecosystems.

Climate change and intensified human activities are fundamentally altering terrestrial ecosystems, inducing complex and nonlinear shifts in the coupling between vegetation structure and function. Despite observed global greening, the extent to which these structural gains translate into enhanced ecosystem function remains unclear. This study proposes an integrated framework that combines temporal trajectory analysis (trends and shapes) of vegetation structure and function. Based on the joint temporal dynamics of Fractional Vegetation Cover (FVC) and Gross Primary Productivity (GPP), we identified the coupling and spatial heterogeneity of vegetation structural-functional trajectories across China from 2000 to 2020. Our analysis revealed a nationwide increase in both trend of vegetation structure (FVC, 57.8% of pixels) and function (GPP, 91.5% of pixels). Notably, over half of China's territory (54.91%) exhibited a coupled enhancement, predominantly in humid regions of the northeast and south. In contrast, functional-leading trajectories with increasing GPP were prevalent in arid and high-altitude regions (33.69%), suggesting that productivity gains may occur under relatively constrained structural trajectories. Conversely, trajectories of coupled decline and structure-leading change were rare (3%), largely confined to decertifying or rapidly urbanizing areas. These spatial patterns were found to be moderated by hydroclimatic conditions and ecological baselines, while climate-human interactions were linked to greater regional disparities. Specifically, coupled enhancement reflects strong climate-vegetation synergy, whereas coupled decline points toward localized disturbances rather than macro-scale drivers. Meanwhile, structure-leading trajectories are tied to baseline vegetation status, and functional-leading patterns signify hydro-thermal compensatory productivity. Our results demonstrated that vegetation greening is not synonymous with functional improvement, highlighting the critical need to integrate functional metrics into ecosystem assessments. The proposed framework offered a tractable approach for long-term ecosystem monitoring, quality evaluation, and targeted restoration management.

Providing widespread and equitable greenspaces for urban residents is of great significance for promoting sustainable development and public well-being. Recent studies indicated that some cities had experienced greening, but the temporal dynamics of greenspace and human exposure remain unclear due to the lack of long-term trajectory analysis. This study takes Shanghai, a rapidly developing megacity in China, as the study area to investigate the temporal trajectories of greenspace dynamics and human exposure equality with long-term, high-resolution enhanced vegetation index (EVI) data and population data. The Venn conceptual model was employed to assess how greenspace coverage and population distribution influence human greenspace exposure and equality along the urban-rural gradient. Our findings revealed that despite a decline in EVI in Shanghai from 1984 to 2023, recovery in greenness and greening trend was found in 55.7% areas, which accounted for 87.3% in the urban core and 53% in the suburb, driving significant green recovery in highly urbanized regions. Over the past 20 years, population-weighted greenspace exposure in Shanghai increased, following a spatial pattern of "increase-decrease-increase" from the urban core outward. However, human greenspace exposure equality was not increased with the green recovery around the 2000s. The positive influence of greenspace coverage in the suburb and rural areas was insufficient to offset the negative influence of population distribution, failing to increase equality. Our study clarified the temporal changes in greenspace dynamics and human exposure during urbanization, and highlighted the need to strengthen the human-nature relationship to increase greenspace exposure equality.

Accurate measurement of greenhouse gas (GHG) emissions from livestock is essential for developing effective emission reduction strategies. Herein, we used unmanned aerial vehicles (UAVs) and deep learning technology to estimate methane and nitrous oxide emissions from enteric fermentation and manure management in pastoral regions. Refined GHG emission factors were derived by considering animal weight, feed quality, breeding methods, and grassland types, thereby improving the precision of livestock GHG emission measurements. Based on the refined emission factors, average household-level livestock emissions were estimated at 110.6 tons CO2e. Potential emission reductions from different strategies, including reducing livestock numbers, improving feed quality, and changing breeding methods, together with their associated costs and benefits, were evaluated to identify the optimal mitigation strategy. Among the tested options, improving feed quality showed the best balance between emission reduction potential and cost-effectiveness. In particular, converting low-quality forage to medium-quality forage could reduce emissions by 45.8 tons CO2e per household, with an estimated regional mitigation potential of 17.57 million tons CO2e in our study area. We also discussed appropriate carbon pricing mechanisms for mitigating livestock-related GHG emissions. This study offers valuable guidance for developing effective emission reduction strategies for pastoral livestock.

We conducted a lab-in-the-field experiment to examine the performance of different management strategies on the sustainable management of public grasslands in China. We designed three experimental interventions, including reward, punishment, and communication, to identify the effectiveness of different treatments on the grassland conservation, livestock production, and household welfare. Results reveal that communication is the most effective in achieving grassland conservation and enhancing household welfare. Compared with reward, communication can achieve better grassland conservation outcomes without external monetary transfers and reduce income inequality among herders. Compared with punishment, communication can achieve similar grassland conservation outcomes without lowering herders' financial payoffs. Mechanism analyses show that communication encourages herders to conduct low-intensity grazing by increasing their beliefs in others with low-intensity grazing. Our results highlight the importance of communication on common-pool resource management.

The urban environment is the "natural laboratory" of the global ecosystem, and it has complicated effects on vegetation growth, including direct effects (land use transformations) and indirect effects (climatic environment changes). However, the long-term responses of vegetation to urbanization and its associated controlling factors across different spatial scales, from pixels to regions, remains unknown. Here, we unraveled the dual influence of urbanization on vegetation growth and its potential drivers along the urban development gradients in China with satellite observations of leaf area index (LAI) during 2000-2020. The results showed that 65.68 % of pixels in whole China exhibited an increasing trend in vegetation growth, with prominent greening (as indicated by LAI increases) in rural background areas (0.198/10a), significant greening in urban core areas (0.0343/10a), and significant browning in suburban areas (-0.0391/10a). As the process of urbanization intensified, the relationship between urbanization and vegetation growth became increasingly complex, transitioning from linear to non-linear interaction. The overall direct effects of Chinese cities were negative and increased annually. Meanwhile, the positive indirect effects of urban environments on vegetation growth initially declined and then recovered. Cities with high urbanization level (urbanization rate, ULp >70 %) had higher indirect effects (0.24 %) and growth offsets (0.98 %) than that with moderate (ULp = 60 %-70 %) and low urbanization levels (ULp <60 %) (0.18 %, 0.10 %). In economically developed cities, land use changes from construction to vegetation, influenced by urban policies and management strategies, positively impacted urban greening. Overall, more urbanized cities (ULp >70 %) experienced vegetation growth enhancement due to more intense land use changes, whereas less urbanized cities (ULp <70 %) showed the opposite trends. Understanding the direct and indirect effects of urbanization on vegetation growth is crucial for devising effective urban planning and environmental conservation policies. It can help guide future urbanization processes and minimize adverse effects on the natural environment.

Arid inland basins represent critical hotspots of intensified conflict among water resources, ecological integrity, and economic development on a global scale. The coevolution of groundwater systems and land use patterns plays a pivotal role in shaping regional sustainability trajectories. This study synthesizes multi-source data spanning 2000 to 2020 from the Wuwei Basin, located within the Shiyang River watershed in China, to elucidate the synergistic dynamics between hydrological and land use transformations. Key findings reveal: (1) Around 2010, a significant structural shift in land use occurred, transitioning from production-oriented expansion to ecologically driven priorities. This shift was characterized by a reduction in cultivated land, increased utilization of artificial surfaces, and accelerated ecological restoration efforts. These changes were jointly influenced by enhanced water governance frameworks and spatial planning policies. (2) Groundwater levels exhibit marked spatial variability. While stability is maintained in piedmont and discharge zones, persistent overdraft has led to pronounced declines in transitional and distal recharge areas. This heterogeneity is primarily governed by the interplay of hydrogeological factors—such as recharge capacity and aquifer permeability—and anthropogenic pressures, including the extent of cultivated land and intensity of groundwater extraction. Notably, these patterns cannot be explained solely by the proportion of cultivated land or total extraction volumes. (3) A positive feedback mechanism—termed the “gain-loss regime shift”—has been identified in the discharge zone, where simultaneous increases in groundwater extraction and water-level recovery are observed. However, human activities have disrupted the natural coupling between precipitation and groundwater recharge, resulting in a significant attenuation of recharge rates (exceeding 80%). These findings offer a robust scientific basis for implementing spatially differentiated water resource management strategies and optimizing land use in arid basin environments. The implications extend beyond regional contexts, contributing to broader efforts in harmonizing human–environment interactions globally.

With the acceleration of industrialization and urbanization, air quality has become a global concern. Ozone (O₃), a significant secondary pollutant in the atmosphere, is increasingly recognized as a major environmental and public health threat. This study focused on Shenyang, a major city in Northeast China, analyzing the temporal variation characteristics of O₃ concentrations at five different functional sites from 2018 to 2021. The study also examined trends and periodicities in O₃ concentrations using the Mann-Kendall test, Theil-Sen slope estimation, and Morlet wavelet analysis. Additionally, stepwise multiple linear regression was employed to assess the correlation between O₃ levels and meteorological factors. The results indicated that, despite variations in annual O₃ levels across different sites, a general downward trend was observed from 2020 to 2021, suggesting a reduction in O₃ pollution. Furthermore, O₃ concentrations were found to peak during the summer, gradually increasing from January and reaching their highest levels in May, June, or July. The Mann-Kendall test results revealed that only O₃ levels at sites D and E exhibited a statistically significant downward trend. All five selected sites showed clear periodic fluctuations in O₃ concentrations. Furthermore, O₃ levels were significantly positively correlated with temperature and wind speed, and significantly negatively correlated with relative humidity.

Decoupling carbon emissions from economic development is crucial for sustainable development and achieving carbon peaking and neutrality goals. However, previous studies primarily assume linear relationships between influencing factors and carbon emissions, with limited attention to nonlinear effects at the urban scale. This study investigates the spatiotemporal evolution of carbon emissions across 232 Chinese cities from 2000 to 2022. Using the Tapio decoupling model, we analyze the relationship between carbon emissions and economic growth, as well as their spatiotemporal variations. Additionally, an XGBoost-SHAP model is applied to uncover nonlinear drivers of carbon emissions. The results show that, over the past 23 years, total and per capita carbon emissions have increased, while carbon emission intensity has declined. Northern cities, particularly in the Hohhot-BaotouOrdos-Yulin (HBOY) urban agglomeration and Northeast China, exhibit higher emissions and intensity compared to southern cities. From 2000 to 2022, over 94 % of cities achieved decoupling between economic growth and carbon emissions, though most were classified as weak decoupling. The proportion of cities achieving strong decoupling increased from 20 % to 40 % over three phases. Expansive negative decoupling and expansive coupling cities are concentrated in the HBOY region and Northeast China. Key determinants of carbon emissions include GDP, population size, industrial structure, and energy intensity. GDP and energy intensity have nearlinear positive effects, while industrial structure, green patents, and environmental regulations show inverted U-shaped relationships. Conversely, population size, environmental protection expenditures, environmental rights trading exhibits a U-shaped trend This study highlights the nonlinear effects of various factors on emissions and offers insights for emission reduction strategies in diverse regions.

Urban vegetation shows significant spatial differences due to the combined effects of natural and human factors, yet fine-scale evolutionary patterns and their cross-scale feedback mechanisms remain limited. This study focuses on the Yangtze River Delta (YRD), the top economic area in China. By integrating data from multiple Landsat sensors, we built a high—resolution framework to track vegetation dynamics from 1990 to 2020. It generates annual 30-m Enhanced Vegetation Index (EVI) data and uses a new Vegetation Green—Brown Balance Index (VBI) to measure changes between greening and browning. We combined Mann-Kendall trend analysis with machine—learning based attribution analysis to look into vegetation changes across different city types and urban—rural gradients. Over 30 years, the YRD’s annual EVI increased by 0.015/10 a, with greening areas 3.07 times larger than browning. Spatially, urban centers show strong greening, while peri—urban areas experience remarkable browning. Vegetation changes showed a city-size effect: larger cities had higher browning proportions but stronger urban cores’ greening trends. Cluster analysis finds four main evolution types, showing imbalances in grey—green infrastructure allocation. Vegetation baseline in 1990 is the main factor driving the long-term trend of vegetation greenness, while socioeconomic and climate drivers have different impacts depending on city size and position on the urban—rural continuum. In areas with low urbanization levels, climate factors matter more than human factors. These multi-scale patterns challenge traditional urban greening ideas, highlighting the need for vegetation governance that adapts to specific spatial conditions and city—unique evolution paths.