The Shennongjia Forestry District is one of the areas with the highest biodiversity in China. The complex topography exerts great influence on vegetation distribution in this region. This paper used the maximum entropy model (MaxEnt), digital elevation data, vegetation distribution map, and field-surveyed data, to study how the topographic characteristics affect local typical vegetation distributions at two scales, i.e., vegetation type and population levels. The relationship models between vegetation type and topographic factors, and between plant population and topographic factors were established respectively by quantifying the topographic ranges of vegetation types and plant species. Results show that: (1) the spatial distribution of different vegetation types was affected by different topographic factors. The distribution of coniferous forests was affected by elevation and coefficient of variation in elevation, the distribution of broad-leaved forests was controlled by elevation and aspect, and the distribution of shrubs was controlled by aspect and the slope of aspect. The factors affecting the distribution of grasses were various. (2) The elevation ranges of typical plant species were generally consistent with those of vegetation types. Specifically, 90% of coniferous forests were distributed at elevation between 1600 and 2600 m, and the typical populations of Abies fargesii and Pinus armandii were distributed at elevation of 1700~3200 m. 85% of the broad-leaved forests were distributed within the range of 1000 m to 2000 m in elevation, and the typical populations of Cyclobalanopsis glauca, Carpinus turczaninowii concentrated at elevation between1000 to 2000 m. 95% of shrubs occurred at slope of aspect within 0~40 degrees, and the typical populations of Rhododendron simsii and Rosaceae mainly occurred at slope of aspect less than 40 degrees. The relationship models used for vegetation types and plant species were different. The relationship between vegetation types and topographic factors was fitted using Gauss model. While the relationship between typical species and topographic factors was relatively complex, and the distribution patterns of different species were even different. (3) Vegetation distribution showed a rather weak relationship with typical slope characteristic. This study provides a basic reference for vegetation protection, vegetation restoration, and vegetation management in the Shennongjia region.

As one of the important characteristics of lakes, lake water level is an important indicator for evaluating the change of lake water storage capacity. As a new remote sensing monitoring technology, satellite altimetry technology has played an increasingly important role in monitoring the dynamic changes of lake waters, and has become an effective means for lake research, water resources investigation and wetland protection. This study took Hongze Lake, Gaoyou Lake and Dongting Lake as examples, using the Centralized Probability Density Function method (CPDF) to improve the accuracy of Jason-2 altimetry data, as well as analyze the correlation of precipitation and water level of each lake. Besides, based on the measured water level data we compared the accuracy of the original GDR data of the Jason-2 altimeter satellite and the satellite data processed by the CPDF method. Results show that: (1) The distribution of Jason-2 original GDR data points is sparse, most of the data are relatively concentrated, and there are certain periodic changes, but the evaluation results show poor accuracy, so the original GDR data cannot be directly used for lake water level monitoring. (2) CPDF method can greatly improve the accuracy of the water level data of the altimeter satellite. The evaluation results of Hongze Lake and Gaoyou Lake show that the RMSE decreased from 1.92m and 1.74m to 0.32 m and 0.36m, and the correlation coefficient increased from 0.28 and 0.04 to 0.85 and 0.72, indicating that Jason-2 altimetry data processed by the CPDF method can achieve higher accuracy in lake water level monitoring. However, it is worth noting that for the lake, which is with narrow north-south widths and large changes in daily water levels (such as Dongting Lake), the accuracy of the raw GDR improved by the CPDF method would be limited. (3) The precipitation of Dongting Lake had the strongest correlation with the water level, followed by Gaoyou Lake,the water level changes of Gaoyou Lake and Dongting Lake lag behind the precipitation by about 1 month and 1-2 months respectively. However, the precipitation of Hongze Lake was not significantly negatively correlated with the water level, which is partly due to the adjustment of water level by the Hongze Lake water conservancy project. This study is of great significance for obtaining lake water level values using altimetry satellites, and then for dynamic monitoring of lakes, especially in filling lake water level data in data-poor areas.

Simulating the surface flow dynamics is of great importance in disaster prevention and mitigation, which could benefit the land remediation, regional planning, environmental protection and water resource management. Therefore, in this paper the Compute-Unified-Device-Architecture (CUDA) is embedded into the TIN-based surface flow dynamic simulation algorithm to get a parallel method for simulating the surface flow discharge. The aim of this study is to rapidly and accurately perform the surface flow dynamic simulation at any position and any time to meet the actual application requirements. First, the proposed algorithm extracts the critical points and drainage network from a high-precision Digital Elevation Model (DEM) to generate a drainage-constrained Triangulated Irregular Network (TIN). Second, the flow direction of each triangular facet over the TIN is calculated by the coordinate data of the triangular facets. Third, the flow path network is traced by the flow direction and rainfall source points. Fourth, the flow velocity of rain dropsover the flow paths is obtained by the flow velocity calculation kernel function based on the manning formula. Finally, by combing the flow velocity with the pre-set time, the algorithm can rapidly simulate the flow discharge at any location by usingthe flow discharge count kernel function. The specific paralle lization process consists of the transmission mode of data, partition model of thread and implement ation of the flow velocity calculation kernel function and flow discharge count kernel function. Data transmission in the paralle lization process includes the input and output of data. It inputs the data of the DEM, rainfall source points and flow path network from the CPU to GPU and outputs the data of the flow discharge calculated by the above two kernel functions from the GPU to CPU. The two kernel functions are parallelized by the flow paths. Each thread handles a single flow path. As a result, flow paths are divided by the partitioning method to obtain numerous thread blocks and the number of the thread in each thread block is allocated by the computing power of the GPU. The Black Brook Watershed (BBW) located in the north-eastern of New Brun swick was selected as the study area. To validate its accuracy and efficiency, the proposed method was compared with TIN-based surface flow dynamic simulation method. The experimental results demonstrate that the proposed algorithm can greatly improve the simulation efficiency while maintaining the accuracy simultaneously and its acceleration ratio can reach up to 11.2. In addition, the parallel algorithm was compared with the Soil and Water Assessment Tool (SWAT) model to verify its precision. The experimental results indicate that the Nash coefficient of the parallel method is increased by 88% and the correlation coefficient is increased by 5%.

The rapid urbanization has constantly changed the transformation of land resources. With rapid growth of economic and population, urbanization has also led to many ecological and environmental problems. Simulating the mechanism of urban expansion and providing early warning of the risk of urban land use change in the future can help to regulate and control urban development. In addition, most urban expansion simulation studies select common drivers and uniform transformation rules for simulation and prediction. Insufficient consideration of spatial heterogeneity increases the simulation error.This paper proposed an urban expansion simulation model based on the geospatial partition and Future Land Use Simulation （FLUS） model for simulating and predicting complex land use change. The model used multiple indicator data for spatial clustering in township streets and grid units, and divided Pearl River Delta into 10 sub-regions. Urban expansion simulation was performed under unpartitioned and partitioned scenarios with geospatial partition results. The simulation results of the urban expansion in the Pearl River Delta from 2005 to 2015 show that the simulation accuracy under the partitioned （FoM=0.2329, increase 9%） scenario is significantly higher than that of the unpartitioned scenario. Land use conversion potentials in different districts display spatial differences. Combining geospatial partition with FLUS model can simulate urban land use change more effectively. The Markov chain model was used to predict the number of future land types. The model was further applied to simulate land use changes in the Pearl River Delta in 2025, 2035 and 2045. Based on the impact of urban expansion on urban form, ecology and intensity, this paper constructed an urban expansion early warning indicator evaluation system to assess the alert of urban expansion. Furthermore, this system can predict the level of urban expansion deterioration and provide a scientific reference for urban development planning and monitoring. Based on the simulated results of land use change between 2025 and 2045, an early warning analysis of urban expansion in Pearl River Delta was conducted at the partition and city level. The results show that the urban expansion of most cities in the region will reach above the middle and heavy level in 2045. Dongguan will always maintain heavy level in the future. Urban expansion in Pearl River Delta is not optimistic. The analysis suggests that strengthening macro-control on urban expansion in Pearl River Delta to alleviate alarm of future urban expansion.

The Future Land Use Simulation (FLUS) model is a new model for simulating multiple land-use changes, and has a broad application prospect. This paper improved the FLUS model by incorporating a spatial autocorrelation factor into the Artificial Neural Network (ANN) module of FLUS, selected thePearl River Delta region as the case study area, and validated the improved FLUS model based on the land use data of 2009 and 2015, as well as a series of driving factors. Three future land-use scenarios in 2035: the baseline scenario, cultivated protection scenario, and ecological protection scenario, were simulated using the improved model. The results showed that: (1) After incorporating the spatial autocorrelation factor, the model had better predictive powerfor the occurrence probability distribution ofeach land use. The ROC values of cultivated land, forestland, water area,construction land, and unused land increased from 0.819, 0.928, 0.885, 0.855, and 0.861 to 0.857, 0.934, 0.890, 0.863 and 0.978, respectively. (2) The simulation accuracy of the improved FLUS model was improved. The Kappa value increased from 0.732 to 0.744, and the FOM value increased from 0.077 to 0.106. (3) The scenario simulation results indicated that under all three scenarios, forestland and construction land would increase, whereas cultivated land would decrease. Apparent differences also existed in the simulated change sizes and locations of each land use type under different scenarios. Under the baseline scenario, construction land would expand rapidly at the expense of a large amount of cultivated land. Under the cultivated land protection scenario, cultivated land area would remain at a reasonable level, the expansion of construction land would alleviate, and the land use layout would tend to be reasonable. Under the ecological protection scenario, cultivated land, forestland, and water area would be well protected, the layout of construction land would be more rational, and the land use sustainability in the study area would be improved significantly.

As China has been vigorously promoting the new-type urbanization and implementing spatial planning policies, city clusters have become the urbanization frontier. But the contradiction between economic development and ecological protection restricts the sustainable urbanization. As the basis of urban development, the changes of land quantity and distribution have a significant impact on the structure and functioning of ecosystem. Therefore, land ecological security assessment of city clusters and predicting future land use patterns of city clusters based on such an assessment are crucial to the sustainability of city clusters. The assessment of land ecological security is the basis of the optimal allocation of land resources, and land use simulation is an important method to predict the trend of land use changes. Combination of the two methods can provide reference for optimizing land use patterns and protecting the eco-environment. By studying the city clusters around Poyang Lake, this paper analyzed land ecological security patterns and changes of the city clusters. Based on the results of land ecological security assessment, the paper set up the business-as-usual scenario and the ecological protection scenario. By combining the multinomial logistic regression and Multi-Criteria Evaluation method (MCE), the paper constructed the CA-Markov model to predict land use patterns in 2030 under the two scenarios, and conducted comparative analyses. Results show: (1) In 2005, 2010, and 2015, the average ecological security of the city clusters around Poyang Lake were 0.574, 0.573, and 0.571, respectively. The security level were low in the middle and high in the east and west in the spatial layout. (2) In 2030, the newly developed urban lands under the business-as-usual scenario will mainly occur in Jiujiang, Shangrao, and Nanchang. Under the ecological protection scenario, the lands for constructing towns and other construction purposes will be restricted properly for compact urban growth. (3)Under the ecological protection scenario, the area of high ecological security zones will be 39.39% larger than that under the business-as-usual scenario,and it tends to be distributed more evenly. The ecological security of the city clusters, including areas around Poyang Lake, the central part of Jiujiang, Xinyu and Ji'an, will be effectively protected. It is hoped that the present study can serve as a reference for the land use planning and ecological protection of thecity clusters around Poyang Lake.

：Delineation of urban growth boundary is an important means to prevent the disorderly spread of urban land, protect the ecological environment of the open space outside the cities, and realize the smart growth of cities, which is of great significance to the healthy and sustainable development of cities. Currently, the studies on urban growth boundary delineation mostly use model method, such as Cellular Automata (CA) model, a relatively mature model, to simulate future urban patterns. However, these studies mostly focuses on the delineation of urban growth boundary. There is little effort on quantitative delineation of the inertia boundary of urban growth. The delineation of inertial boundaries can not only reserve a certain space for urban development, but also improve implementation efficiency in urban planning. Based on this, we proposed a method based on the kinetic energy theorem of the mechanics. The FLUS model and the Morphological Erosion and Dilation method (MED) were used to delineate the urban growth boundary. Slope and land use type were used as the frictional force to delimit the inertial boundary. The FLUS model inherits the Artificial Neural Network (ANN) and the Cellular Automaton(CA) model for simulating and predicting the future urban landscape. The MED was used for the clustering of urban neighborhoods to merge into a single area, while eliminating small but isolated urban plaque. In our paper, we selected Fuzhou, a coastal developed city with obvious changes in recent years, as the study area. We simulated the land use patterns from 2000 to 2015 to verify the accuracy of the model. The overall accuracy of land use simulation in 2015 was 0.9389 and the Kappa coefficient was 0.9165. We further predicted the land use pattern in 2027, and delineated the urban growth boundary and inertia boundary in 2027. Results show that the FLUS model and MED can effectively simulate land use and better fit the growth boundary of urban land use. Using the method of kinetic energy theorem for reference, the inertia boundary of a city can be well delineated according to the expansion resistance in different directions and the expansion intensity in different directions, which provides practical operability and reference values for future study.

：The ecological environment of the agro-pastoral ecotone in Northwest China is extremely fragile. Due to global warming, the frequent droughts in this region have seriously affected agriculture and animal husbandry, which is not conducive to the sustainable socio-economic development of the agro-pastoral ecotone there. It is the special location and ecological value that determine the region has an important strategic significance in socio-economic development and ecological environment protection in China. The usage of land resources plays an important role in not only the ecological environment, but the socio-economy. Therefore, the purpose of this paper is to achieve ecological environmental protection and economic development by optimizing the allocation of limited land resources in the region. The application of multi-objective genetic algorithm and FLUS model can improve the land use optimization configuration from many aspects (i.e., quantitative structure, spatial distribution, comprehensive benefits). In addition, the combination of the genetic algorithm and multi-objective programming model can provide more alternative options for it. In this paper, the multi-objective genetic algorithm and FLUS model are used to simulate the land use change of the region in 2025. By setting up four scenarios (i.e., natural development, ecological protection priority, economic development priority and eco-economic equilibrium), we explore how to optimize land use allocation on the condition of considering both the protection of ecological environment and the development of social economy. The results suggest that the eco-economic equilibrium optimization scheme shows great advantages in both the quantitative structure and spatial distribution of land use types, and its comprehensive benefits are superior to the other three scenarios. On the premise of reasonably limiting the speed of economic development, this scenario has enabled the ecological construction to develop at stable speed. Its economic benefits have increased by 8.96% compared with those under the ecological protection priority scenario, and its ecological benefits have increased by 0.77% compared with those under the economic development priority scenario. The scenario achieves the coordination between ecological protection and economic development.The eco-economic equilibrium optimization scheme has shown great potential in many aspects, such as promoting the development of unused land, assisting in the adjustment of land use structure, and driving the optimization of industrial structure. The optimal allocation of land use provides decision-making assistance for the future ecological environment construction and economic development planning of the agro-pastoral ecotone in this region. In addition, to solve the contradiction between economic development and ecological construction in the agro-pastoral ecotone, the government should strictly co-ordinate arrangements for various types of land, while enhancing the comprehensive utilization of land resources.

The urban-agricultural-ecological space zoning (referred to as the "three zones") is the core content of land space planning. It is important for scientific and rational planning and using of limited land resources. The former researches set up various indicator systems for "three zones", mainly based on the current regional land use and socio-economic development status. But they rarely incorporate future land use changes into the "three-zone" delineation process, making the results less forward-looking in guiding practice. To make up the gap of current research, we propose a new "three zones" delineation method in this paper. It is based on the simulation of land use scenarios and combines the advantages of indicator system and decision tree data mining, which is different from the traditional method of "indicator system - weighted comprehensive value". This paper selected Wuhan as the research area, and used this method to explore the spatial differences of the "three zones" under different land use scenarios in the future. We simulated four land use scenarios of Wuhan in 2035 (Natural development scenario, Farmland protection scenario, Ecological protection scenario and Balance development scenario) based on the land use of Wuhan in 2015. After that, we combined the constructed indicator system and multiple textual references to select typical samples for decision tree training. Then using the classification rule set generated by the decision tree (86.4% average accuracy) to identify the "three zones" spatial categories of the research units. Finally, we obtained the "three zones" distribution under different land use scenarios. Compared with the similar former researches, the method proposed in this article is more reasonable and feasible and can be used in specific research and practice. Besides, we found that: (1) there are obvious differences in the area, spatial distribution and main change types of the "three zones" caused by different land use scenarios. So, it is indeed necessary to incorporate future land-use changes into the "three zones" delineation process. (2) The spatial distribution of "three zones" in different land use scenarios shows similar characteristics. The differences in spatial distribution of "three zones" in different land use scenarios are mainly the border area where the main land use function changes. Therefore, these areas are the key areas that land space planning should focus on.

Demarcating permanent prime farmland is an effective way to realize farmland protection, intensive land use and food security. With the challenging of rapid urban expansion, the completion between urban expansion and farmland protection is unavoidable. It is urgent to reasonably demarcating permanent prime farmland by synergizing farmland protection and urban expansion. In view of this conflict, taking Wuhan as an example, this paper carries out the demarcation of permanent prime farmland by using LESA (Land Evaluation and Site Assessment) method and LANDSCAPE ( LAND System Cellular Automata model for Potential Effects) model, attempting to achieve balanced development of urban expansion and farmland protection, which provides reference for scientific demarcation of permanent prime farmland under the background of multiplanning integration, and provides support for prime farmland protection and optimizing of land allocation. Firstly, comprehensive evaluation and classification of farmland in Wuhan is carried out through LESA, then based on this result, simulation of demarcation of permanent prime farmland and urban expansion is carried out through LANDSCAPE model, at last, the results of the two methods are compared in quantity, quality and spatial patterns. The results show that permanent prime farmlands demarcated in two methods have little difference in quantity and quality. The area of permanent prime farmland demarcated by LESA method is 243 259 hm ², 45.63% of which land fertility grade of cultivated land is between grade 1 and grade 3. The area of permanent prime farmland demarcated by LANDSCAPE model is 243 200 hm ², 45.77% of which land fertility grade of cultivated land is between grade 1 and grade 3. There are more than 80% of the permanent prime farmland that land fertility grade is between grade 1 and grade 5 in both methods, which shows that the total quality of permanent prime farmland is good in either method. However, the result of LANDSCAPE model is better in spatial patterns, which permanent prime farmland demarcated by LANDSCAPE model distributes concentrated and is regular in shape. It's remarkable that 15.8% of permanent prime farmlands demarcated by LESA are overlapped by urban constructive land in the processing of urban expansion, which has been effectively avoided by LANDSCAPE model. The possible cause of this conflict are as follows: (1) The permanent prime farmland demarcated by LESA method is based on ranking of integrated score, in which the requirements for concentrated fragmentation and shape index of permanent prime farmland are not considered; (2) The conflict of farmland protection and urban expansion is not considered in LESA method.

Urban rail transit possesses significant impacts on land use change and urban development. This study applies Future Land Use Simulation Model (FLUS) to reproduce land use changefrom 2000 to 2010 in the Mid-Zhejiang urban agglomeration based on GlobeLand30 datasets. The simulation results in 2010 show that the FLUS model can reproduced a realistic land use pattern with an overall accuracy of 89.74% and FOM 29.86%. A Markov chain is then used to predict the total land demand in 2030 for predicting future land use scenarios. We design two scenarios: the scenario of business-as-usual (BAU-scenario) and the scenario based on planned urban rail transit sites (RTS-scenario). Within 5 km from the urban rail transit, the RTS-scenario yields a significant effect on built-up areas with an increasing expansion intensity, where the newly built-up areas are allocated in the suburb sand are greater than that produced by BAU-scenario by 45.25 km ².The newly built-up cells mainly occupy high-quality farmland. The farmland transformed to built-up area is higher in RTS-scenario than in BAU-scenario by 33.34 km ².We categorize the built-up expansion intensity (BUI) into five levels: lowest, low, medium, high and highest. The BUI for RTS-scenario is higher than that for BAU-scenario because the former’s proportion of expansion intensity above the lowest level is 3.70% greater than of latter. Spatial patterns for forest, grassland and water are similar between both scenarios. This study not only indicates that FLUS can be used to capture land use change and predict future scenarios, but also helps to examine the effects of urban rail transit site plansin the Mid-Zhejiang urban agglomeration.

Urban Expansion (UE) is a critical parameter to examine urbanization level. In fact, UE is often accompanied with people, strategies, and urban elements extending over the old edge of city. Hence, UE is a highly complex socioeconomic phenomenon. Dynamic models are very effective to reveal the features and mechanisms of UE. In particular, Cellular Automata (CA) is one of the kinetic models to simulate the natural evolution process and famous for its broad range of applications. A large amount of research findings argue that CA is one of the most potential tools to understand UE. In this study, a multi-factor restricted expansion simulation model of CA (MCES-CA) was proposed according to CA's main features and the main influencing elements of UE. This paper took a middle-zone city named Hengyang City as the case study area. Hengyang City is located in southern Hunan Province of China, and has been positioning to develop into the regional central city of southern Hunan. Hence, it is high significant to reveal the main features and mechanisms of Hengyang City's expansion based on MCES-CA. Firstly, we downloaded LANDSAT images and ASTER DEMs through the internet, and collected Hengyang City's Planning datasets from the Planning Burau of Hengyang Government. Next, we developed the MCES-CA model by the Model Builder Tools of ArcGIS. Then, we ran it to observe the entire expansion process of Hengyang City from 2001 to 2017. We employed the Expanding Intensity Index (EII) and Expanding Velocity Index (EVI) to assess the MCES-CA model. Results show: (1) Hengyang City presented different expansion directions in different ages. (2) The spatial expansion process of Hengyang City was composed of two main stages. (3) The main factors of shaping the expansion process of Hengyang City were geo-environments and development planning. (4) The simulation results were in accordance with the real states of Hengyang's UE. Our findings suggest that MCES-CA is a potentially popular tool to capture the EU processes of cities because it is easy to build and complete. For future research, we can improve the simulation accuracy of MCES-CA by refining transformation rules.

Due to human activities and rapid urban expansion, land use/land cover has changed dramatically. The change has a great impact on the ecological environment and surface landscape. The change process of land use and cover change is not only affected by various factors such as nature and economy, but also an external representation restricted by the laws of human activities and natural factors. Therefore, it is of vital significance to study the change process of land use and land cover. For the monitoring and analysis of land use and cover change, the traditional method focuses on the study of the overall differences in land use structure in each spatiotemporal snapshot. This method cuts off the organic connection of land use units in the evolution process between different snapshots. Traditional research has the phenomenon of paying attention to pattern but neglecting process and emphasizing simulation but despising measurement. This paper takes the land change process composed of serial land use data as the core research object. The advantage of this is that the relevant land evolution units at different time snapshots can be considered as a unified whole. On this basis, this paper chooses the nearest spatiotemporal distance to measure the spatiotemporal agglomeration of the land use change process. First, through multiple experiments, the appropriate space-time grid was selected to segment the land use/land cover data in the study area. Secondly, based on the analysis, the typical land use change process was extracted. And then, for the land use change process object on the space-time cube, calculate the average nearest-neighbor spatiotemporal distance. The distance is compared with the distance value in the random mode based on Monte Carlo simulation, so as to judge the spatiotemporal aggregation of the land use change process in the study area. Finally, the results were tested for statistical significance. The land use data of Huainan mining area from 2008 to 2017 was used for empirical research. The land use evolution process from cultivated land to grassland in any two years was selected as a typical spatiotemporal evolution type. The results show that this type of land change in the study area has exhibited a spatiotemporal aggregation pattern in the past 10 years, but the pattern is not statistically significant. The research in this paper is helpful for grasping the evolution process of land use units in space and time, and to explore the potential spatiotemporal evolution patterns in the process of land use change.

For assessing urbanization level and urban environment, the mapping of impervious surface has become a research hotspot. Compared with single-phase imagery, time series mapping can depict temporal trends, which is of great significance for monitoring urban expansion. Based on the Google Earth Engine platform, this paper calculated BCI and NDVI using Landsat TOA data from 2000 to 2017, and determined their thresholds by an adaptive iteration method to extract the initial impervious surface. Then, Temporal Consistency Check (TCC) was performed to make the time series of impervious surface more reasonable. Results show that: (1) Adding NDVI to both BCI and TCC improved the quality of impervious surface mapping. (2) The average accuracy of impervious surface mapping in this paper was 90.4%, and the average Kappa coefficient was 0.812. (3) The impervious surface area of Guangzhou downtown nearly doubled from 2000 to 2017 with a decreasing growth rate. (4)The newly developed impervious surface mainly concentrated on the relatively backward outskirts of Guangzhou downtown. (5) Elevation, road density, and shopping mart density were the main factors influencing the expansion of impervious surface.