The ultimate form of urbanization is the megacity, which is the typical form of urban structural characteristics and population agglomeration effects. However, the comparative studies on the urban form and expansion patterns of megacities worldwide are still insufficient. This study first analyzed the spatiotemporal changes of land use in megacities based on the Global Urban Land Use/Cover Composites with a 30 m spatial resolution (GULUC-30) by using the method of urban expansion intensity differentiation index (UEDI). Then, the fractal structure and expansion patterns of megacities were investigated using the area-radius model. Finally, urban land use efficiency was assessed through a linear relationship between urban land area and its interna population. The findings include: (1) Southeast Asia and China had experienced the most dramatic urban land expansion (3148.32 km² and 5996.26 km²) over the last 20 years, along with the average UEDI of both greater than 3. (2) In 2000-2020, morphological characteristics and sprawl patterns of megacities developed toward greater intensification and compactness, with the mean radial dimension increased from 1.54 to 1.56. Meanwhile, radial dimension difference showed a year-by-year decreasing trend, and the intra-city structure showed integration. (3) In 2020, urban land use efficiencies in both North America and Europe were low, with urban land area to its population ratio coefficients ranging from 0.89 to 4.11. On the contrary, high values happened in South Asia and Africa, from 0.23 to 0.87. Spatial patterns and morphological changes in megacities provide important reference values for the intensive utilization of urban land and its sustainable development. Controlling the scale of megacity expansion and promoting the balanced exploitation of urban land are the essential ways for building resilient cities and sustainable urban planning.

Scientifically regulating cultivated land fragmentation (CLF) to make sustainable intensification conducive to and adaptive to the local contexts is the goal of long-term efforts in the research and practice of CLF governance. Starting from the theoretical connotation of CLF and sustainable intensification of cultivated land use (SICLS), this study analyzed the theoretical mechanism of CLF affecting SICLS, and then attempted to construct a sustainable intensification-oriented CLF governance framework from the coupling perspective of "feature-influence-environment", and systematically discussed key issues, scenario decision-making mechanisms, spatial scale effects and operating mechanisms in CLF governance. The results show that: (1) CLF is a composite system covering multi-dimensional attributes such as resources, space, utilization and ownership, which has extensive and complex impacts on resource utilization, agricultural production, and landscape ecology, etc. CLF has become the core factor affecting SICLS through action pathways such as multi-scale and omni-direction, difference and uncertainty, and interaction. (2) The situation and extent to which CLF becomes a problem or a benefit depends on its impacts, the local contexts of physical geography, agroecology, social economy, etc., and the orientation of agricultural development. (3) The CLF comprehensive governance framework from the perspective of "feature-influence-environment" deconstructs five key issues, two relationships, two tendencies and three-level spatial scale interconnection mechanisms in complex CLF scenarios. This framework can provide new ideas for scientifically judging the governance orientation of CLF and optimizing the governance model, which is beneficial to improving the governance system of CLF and promoting the SICLS.

Both the food production value on arable land and the structure of food crop cultivation have impacted on national food security. Therefore, exploring changes in their spatial and temporal coupling patterns can help us to understand their relationships and to further improve food security. In this research, we took provincial-level regions in China as the object, and based on an in-depth interpretation of the connotation of the grain-to-arable value ratio and cropping structure, Simpson's index model, coupling coordination model and spatial auto-correlation are comprehensively used to study the spatial and temporal evolution patterns of grain-to-arable value ratio and cropping structure in the study area from 2004 to 2020, and the spatial and temporal dynamics between them in different regions are revealed in consequence. The results show that: (1) The spatial distribution of the grain-to-arable value ratio in all the provincial-level regions from 2004 to 2020 shows a double-decreasing characteristic, one from north to south and the other from west to east; (2) The spatial pattern of cropping structure shows a Hu-Huanyong-line-ridge in terms of spatial distribution, where Northeast, North and East China witness a significant decrease in crop diversity; (3) The degree of coupling coordination between the grain-to-arable value ratio and cropping structure is found between barely coupled and coordinated and highly coupled and coordinated, showing regular spatial clustering characteristics. In the central, western and northeastern China, there is a high-high concentration with matching advantage; in the southeast coastal region, there is a low-low concentration for matching improvement; and the overall distribution presents a pattern of high in the northeast and central regions, and low in the southeast. The study believes that, based on the heterogeneous characteristics of the spatial concentration and diffusion of factors, the synergistic evolution of food quality and diversity on arable land should be encouraged according to the local conditions, so as to promote the harmonious development of food structure and efficiency.

Simulation of urban land use change is the scientific basis for optimizing land resource allocation, and improving its refinement and reliability is helpful to accurately grasp the development trend of urban land use. This is immensely crucial for accurate regulation of urban land resources. The simulation of land use change based on remote sensing classification is macroscopic and simple. However, it is difficult to apply this approach to reveal the change in urban land use social functions, as well as the source and mechanism of spatial scale effect in the refined simulation at block scale. This study identified the refined urban land use characteristics by combining remote sensing images and POI data. Moreover, the optimal spatial scale combination was calibrated for refined land use simulation with the response surface method. Based on the optimal spatial scale combination, the refined simulation of future land use change was performed by using the CA-Markov model. Considering the Wuhan core urban area as an example, the results demonstrate that: (1) POI-based refined urban land use identification method can deeply analyze the social functions of urban construction land, which greatly improves the traditional remote sensing-based macro interpretation of land cover. (2) Optimal spatial scale combination of CA-Markov model for refined land use change simulation in the study area is at the cell size of 30 m and neighborhood size of 7 using the Von Neumann neighborhood type, at which the reliability of refined land use change simulation can be improved. The results of the response surface design can effectively distinguish not only the main sources of the spatial scale effect, but also the magnitude of their influence and the positive or negative effects on the simulation accuracy in the refined simulation process. (3) It is predicted that by 2025 the construction land scope of the study area will continue to expand to the periphery with various types of land interlaced, and the spatial pattern of land use will become more fragmented.

As an important factor affecting vegetation coverage, the terrain is of great importance to vegetation restoration evaluation and ecological restoration planning. The terrain niche index (TNI), a comprehensive index proposed to measure the combined topographic effect, has incomparable advantages over any single terrain factor. However, the whole study area was taken as a unit reference for the TNI calculation in the previous algorithms, which did not consider the spatial heterogeneity of terrain factors. As a result, the old TNI cannot well reflect the local environment of vegetation growth. Inspired by the geographical environment similarity emphasized in the Third Law of Geography, this study integrated the "proximity of spatial location" and the "similarity of environmental characteristics", and modified and improved the traditional TNI model in the following two aspects: (1) local window algorithm was introduced to highlight the local terrain features; and (2) the new TNI was calculated in each habitat unit formed by the combination of different environmental factors to eliminate the interference of environmental heterogeneity on the topographic effect. The case study shows that the improved model can increase the explanatory power of TNI on vegetation coverage, and the correlation coefficient between TNI and vegetation coverage can be maximized when considering the habitat unit and adopting the optimized local window. In addition, among different habitat factor combinations, it can be found that locations with higher temperature values and less precipitation values show a much closer correlation between the new TNI and vegetation coverage, which indicates that the more the water-deficient areas are, the more the sensitive vegetation coverage is to the local terrain conditions. As a result, this study is expected to provide new indicators for vegetation restoration evaluation and planning.

The Qinling-Daba Mountains constitute the bulk of the north-south transition of China. The transitions and patterns of plant species distribution must be determined to understand the spatial variation and complexity of this region and demarcate the boundary between the warm temperate and subtropical zones of China. We analyzed variations in plant species richness, relative importance value (RIV) and typical species abundance in a north-to-south direction based on plant community filed survey data. There were 163 sample sites along three north-south transect lines in the eastern, middle, and western Qinling-Daba Mountains. Furthermore, the differences in RIV between subtropical and temperate species (SND-RIV) revealed the dominance of northern and southern plant species and explained spatial vegetation patterns. The results show that: (1) along the eastern (Sanmenxia-Yichang), middle (Xi'an-Dazhou) and western (Tianshui-Guangyuan) transects, the richness and RIV of subtropical plant species increased while those of temperate plant species decreased from north to south. In the eastern transect, temperate plant species richness and RIV were the highest at Shennongjia and Funiu Mountains, respectively, because of the high elevation there. In the middle transect, subtropical plant species richness and RIV were the highest in Daba Mountain. In the western transect, richness and RIV were higher for subtropical compared with temperate plant species in the south of Longnan. (2) In the eastern, middle, and western transects, the crisscross areas of the northern and southern plant species were ~180 km, ~100 km, and 50-60 km, respectively. (3) The eastern, middle, and western transects differed in terms of their north-south plant species transition patterns and the factors driving them. Decreases in subtropical plant species from south to north were related to the decreases in annual precipitation in the eastern and western transects. In the middle transect, mean annual temperature had a slightly greater influence on plant species distribution than the moisture index.

The Hengduan Mountains is one of the most biologically diverse areas in the world, geographically located in the eastern part of the Qinghai-Tibet Plateau and the upstream of several major rivers. These mountains are ecologically conservated, with an extremely important ecological barrier function. Since 1990s, China has implemented several significant ecological restoration and construction projects in the research area. Due to limited data and lack of field investigations, few studies explored the integrity and long-term vegetation changes in this area. Hardly any of them discussed the systematical relationship between vegetation changes and elevation factors of the Hengduan Mountains. In view of filling this research gap, we used the Climate Change Initiative Land Cover (CCI-LC) dataset for the period 1992-2020 and the Normal Difference Vegetation Index (NDVI) dataset from MODIS for the period 2000-2020 to reflect changes in vegetation distribution and activity, respectively. Here, we used transfer matrices to investigate the transition relationships among different vegetation types, and used Theil-Sen Median trend analysis to explore the spatial and temporal trends of vegetation activities. We also used the partial correlation analysis to quantify the influencing factors of vegetation activities. Herein, the relationship between vegetation coverage area and mean elevation change was also investigated. The results showed that: (1) The most widely distributed vegetation in the study area contained evergreen coniferous forests and shrub-grass types. Vegetation changes were distributed in river valleys, and southern low-altitude areas. The grasslands were mainly converted to evergreen coniferous forests to gradually improve the vegetation coverage. This showed that ecological protection policies played an essential role. However, vegetation coverage changes are generally in the initial stage of policy implementation. (2) The overall vegetation activity showed an increasing trend. In unchanged vegetation regions, vegetation activities increased in more than 75% of the area, and more than 20% of the area increased significantly (P < 0.05). The changing trend of forest vegetation was more significant than that of grassland. (3) The environmental factors that significantly impact vegetation activities are climate change and topographic slope. Although vegetation coverage was enhanced by increasing temperatures, the precipitation influenced the vegetation in hot and dry valleys. Nearly 25% of the total research area shows a decreasing. This decreasing trend is due to natural disasters such as landslides and mudslides caused by much precipitation and steep mountain slopes. These findings can provide a reference for the evaluation of ecological protection policies, the comprehensive risk assessment of natural disasters, and the prediction of vegetation changes under the influence of future climate change in the Hengduan Mountains.

The investigation of the dynamic changes of vegetation growth in arid areas is important to the study of the formation, development, and evolution of desertification. By using the Sen+Mann-Kendall method to detect the change of the time series trend, this paper analyzed the spatio-temporal variation characteristics of the vegetation cover in the potential extent of desertification in China (PEDC) from 2000 to 2020, especially in the forest and grass-covered area, based on the data set of annual vegetation coverage obtained by the improved directional dimidiate pixel model. The research results show that: (1) The average vegetation coverage of the PEDC is 0.284, the improved vegetation coverage estimation result well reflects the vegetation coverage, and the estimation accuracy is 86.98%. The vegetation growth situation is getting better, among which the arid area has the most prominent performance, with a significant increase of 48%, while the average increase in the sub-humid arid area is 0.1. (2) The ecological restoration projects of the forest and grass sector are effective, but vegetation restoration is a long and slow process. In terms of forest and grass area, the increase is small (0.002%) from 2000 to 2010 and relatively large (0.371%) from 2010 to 2020. (3) The forest quality of PEDC has improved most obviously among all vegetation types, while the grassland is relatively stable, and the areas of forestland and grassland which have seen significant increase account for 76.4% and 71.8% respectively from 2000 to 2020. Among them, the vegetation coverage of forestland increased by 0.15 in the sub-humid arid region, while that of grassland increased by only 0.06. This study can provide an in-depth understanding of the long-term vegetation growth status in the PEDC, especially in the forest-grass coverage area, and provide important information for the further formulation and implementation of various ecological projects.

With the visual landscape quality (VLQ) as the evaluation index, scenic roads are among the efforts for the integrative development of tourism and transportation. The current progress in computer vision techniques provides an opportunity for exploring the VLQ. Technique revolution brings about a new methodology to cut across the long-standing dichotomy of subjectivity and objectivity in landscape quality research. Taking the Simao-Xiaomengyang highway in Yunnan as a case study, this paper develops a multi-camera configuration approach to capture and synthesize landscape images fitting with human field of view, and adopts the image recognition technique to acquire the pixel data. Based on these data, we conceptualized the non-scale semantic differential analysis to calculate the landscape qualities from the dimensions of "naturalness—artificiality", "diversity—coherence", and "openness—deepness". The results show that the VLQ in the road in the case study is distinctive, which is highly natural, relatively coherent and open. The VLQ is different in various driving directions, but the differences diminish with the expansion of observation distance. Based on the empirical results, we propose to set up branch byways to consolidate the image of the scenic road, as well as increase differentiated construction measures between directions of the road. The case study results are intuitive and accurate in the aspect of the scenic road VLQ evaluation, which indicates that our methodology is advantageous in integrating subjectivity and objectivity. This method would help the road administrators to formulate policies in planning, construction and management, hereby promoting the high-quality development of scenic roads in China.

The continuous remote sensing monitoring of land use/cover change at the national scale is of great scientific significance for land spatial planning and blueprint drawing of "Beautiful China" in the new era. Landsat 8 OLI, GF-2 and other satellite remote sensing data were used to develop the data of land use change across China during 2015-2020 and China Land Use/Cover Dataset in 2020 (CLUD 2020) by integrating remote sensing big data and expert knowledge-assisted human-computer interaction interpretation methods. Long time series land use dynamic database at a 30-m resolution in China was established at 5-year interval in the end of 1980s-2020. On this basis, the general trend, regional differences and main characteristics of land use change in 2015-2020 were revealed from national and regional scales. The research indicated that integrating vegetation cover change at a 30-m resolution and land change information generated by remotely sensed big-data cloud calculation into the expert human-computer interaction interpretation can effectively improve the efficiency of mapping and the accuracy of land use change detection. The overall accuracy of CLUD 2020 first-level type mapping reaches 95%. In general, the intensity of territorial development entered a stable state compared with 2010-2015. During the period, the cropland continued to decrease. Nationwide farmland was encroached by urban development and construction, paddy fields in Northeast China continued to decrease, and cultivated land in Xinjiang was reclaimed in the south and abandoned in the north. The built-up land continued to increase, showing a spatial pattern that "the expansion of built-up land changed from the agglomeration of coastal areas and mega and large cities in 2010-2015 to the surrounding sprawl of large, medium and small towns in the central and western regions in 2015-2020". Although the area of natural ecological land for forest and grass continued to decrease nationwide, the intensity decreased compared with 2010-2015. Under the continuous impact of climate change, the area of water in the Qinghai-Tibet Plateau increased significantly. The pattern of land use change is closely related to the national macro strategy for high-quality development during the "13^th Five-Year Plan" period (2016-2020) and ecological civilization construction, as well as the impact of climate change.

With the continuous enhancement of regional connectivity, the indirect land resources use through trade in goods and services plays an increasingly important role in the utilization of land resources. However, relevant research in this field is still in its infancy, and few papers have addressed this issue. This paper, for the first time, uses the multi-region input-output model to calculate the embodied land in the 30 provinces (autonomous regions and municipalities) and 8 regions of China from the perspective of regional trade, and further analyzes the spatial pattern characteristics of embodied land flows. The results show that the amount of embodied land occupied by China's inter-regional trade accounts for 21.39% of the country's total land, and an average of 38.54% of China's provincial land demand is met by land export from other provinces. More than 80% of land consumed by Beijing, Tianjin and Shanghai are from other provinces. Heilongjiang, Inner Mongolia, Xinjiang, and Hebei are the provinces with the largest net exporter of cultivated land, forest and grazing land and aquatic land, fishery land, and construction land, respectively (the outflows account for 42.26%, 27.53%, 38.66% and 35.28% of the corresponding land types in the province); and Guangdong, Shandong and Zhejiang are the main net importer. The flow of agricultural land (cultivated land, forestland, grazing land and aquatic land) mainly generally shows a shift from west to east and from north to south. The regions in northeast and northwest China have the largest scale of agricultural land outflow, mainly into East and South China. In addition, 8.43% of cultivated land, 7.47% of forestland, 6.41% of grazing land, 6.88% of aquatic land and 18.35% of built-up land in China are provided for foreign use through international trade. This paper provides feasible ideas and a theoretical basis for solving the contradiction between land use and ecological protection, and realizing sustainable land use in China.

Exploring the types of spatial patterns of urban land use in the Guangdong, Hong Kong and Macao since the reform and opening up in the late 1970s, and the multi-dimensional and interactive driving mechanism for the evolution of urban land spatial pattern, is important for promoting the construction of the Guangdong-Hong Kong-Macao Greater Bay Area, and building a world-class bay area and urban agglomeration. This study divided and defined the evolution of urban land patterns types in the Greater Bay Area from the perspectives of area, shape and distribution characteristics based on fuzzy clustering. Using geographic detectors, this study measured the multi-dimensional driving forces and interactions of urban land use changes in the study area. The results showed that: (1) In terms of spatial pattern, there are five types of urban land use: fractal/complex large area, fractal small area, complex small area, simple small area and scattered area. The area scale and shape complexity of urban land had a certain positive correlation. (2) Socio-economic factors such as population, GDP, local fiscal revenue, fixed asset investment, and road network density had significant impacts on urban land area and morphological changes. The impacts of core cities on the development of urban agglomerations became apparent after 2000. In the later stage of the research, the differences among the effects of various influencing factors on the changes in the area, distribution and spatial pattern types of urban land use tended to be weakened, but the influences of social and economic factors on the shape characteristics increased. (3) The interactions of influence on the change of urban land area characteristics were the most complex, with four types of nonlinear enhancement, synergy enhancement, single factor weakening and nonlinear weakening; the interactions of influence on shape, distribution and spatial pattern changes were mainly manifested as synergistic enhancement and non-linear enhancement. Interactions between core city traction, road network density, river network density and most factors produced a "1+1 > 2" effect, reflecting the strong promotion effects by the core cities of urban agglomerations and transportation advantages.

In recent years, earth observation and remote sensing technology has enabled a better assessment of the terrestrial vegetation dynamics and improved our understanding of critical global environmental change issues. Various long-term satellite derived vegetation datasets have been produced at the global scale. However, a systematical review of these new resources from the perspective of various end-users is still lacking. We review the history of remote sensing derived global vegetation datasets by focusing on satellite missions and corresponding data products, and paying special attention to the sources, characteristics and connections of them. These products have undergone remarkable transitions, i.e. (1) from vegetation spectral index to biophysical characteristics, (2) from traditional optical remote sensing to Solar-induced Chlorophyll Fluorescence (SIF) and satellite radar system, and (3) from single-sensor-derived products to datasets fusing multiple remote-sensing sources. In the future, in order to improve the quality of existing products considerably, including its temporal and spatial resolutions, accuracy, consistency and continuity, new datasets can be developed through fusing multiple data sources based on the raw data of long-term satellite observations. Regions and ecosystems characterized by unique natural conditions should be given more attention in the way of developing algorithms. Meanwhile, multiple vegetation traits need to be monitored. In addition, the development of sharing platforms for remote-sensing datasets, enabling targeted vegetation monitoring, is urgently required so as to facilitate standard data processing and distribution services, and deliver data quality information to end-users in an efficient and transparent way.