Analysis of Land Cover and Landscape Pattern Change in Beijing-Tianjin-Hebei Region Based on Globeland30 Data

With the rapid economic development and urbanization process in Bei-jing-Tianjin-Hebei region, land use and landscape pattern have been changed. In order to reveal the temporal and spatial changes and provide scientific basis for the local land use planning and landscape pattern optimi-zation made by government, based on the classification results of Globeland 30 land cover data in 2000 and 2010, this paper quantitatively analyzed the land use transfer status and the dynamic change of landscape pattern in Bei-jing-Tianjin-Hebei by adopting the method of land use transfer matrix and landscape index. The results show that the land use and landscape patterns in Beijing-Tianjin-Hebei have been changed significantly from 2000 to 2010. Overall, land use in Beijing-Tianjin-Hebei region has been changed greatly, which shows that the cultivated land area has been decreased by 3721.67 km 2 , while the artificial surface area and grassland area have increased by 2721.29 km 2 and 657.09 km 2 , respectively. From the six partitions of Bei-jing-Tianjin-Hebei, it can be seen that the subareas with significant changes are Sand Control Farmland Protection Zone in Yellow River Flood Plain, Water Conservation Area for Sand Control in Mountain and Hilly Areas of Northwest Taihang Mountains, and Human Settlement Environment Maintenance and Farmland Protection Zone of Beijing-Tianjin-Hebei Urban Agglomeration. The arable land area of these three zones has decreased 572.77 irregular shape and discontinuous enhancement of landscape types, the landscape patch types increase and enrich, and the dominance shows a downward trend.


Introduction
Land use/land cover change (LUCC), as the main manifestation of human activities that significantly change the Earth's environmental system, is a hot and frontier area of global environmental change research [1] [2] [3] [4]. With the acceleration of economic development, population growth and urbanization, the role of human beings on the earth's surface has been strengthened, and the land use and landscape patterns on the earth's surface have been changing constantly [5]. Therefore, the analysis of land use and landscape pattern between regions plays a key role in the sustainable development of the region [6]. With the continuous progress of remote sensing technology, especially the emergence and renewal of land cover products [7]- [13], it provides an effective means of detection and data sources for the study of land use change at global and regional scales. In response to the above problems, scholars at home and abroad have carried out a lot of research and achieved fruitful results, such as Fan et al. [14] using land use dynamics model, land use transfer matrix and mobile window landscape index to analyze the spatial and temporal change process of land use landscape pattern in Shanghai from 1995 to 2015; Wang et al. [15] selected the upper reaches of Huaihe River to study the impact mechanism of land use and landscape pattern evolution on ESV. Zhang et al. [16] supported by remote sensing and GIS technology, used land use conversion matrix to analyze the land use pattern of Zhungeer Banner in two invariable years, and combined with soil erosion data to analyze the change of soil and water loss in the region. Lamine et al. [17] selected relevant indexes in FRAGSTATS to analyze the relationship between land use/land cover change and landscape fragmentation in the Mediterranean region. Most of these land use status and landscape pattern studies are based on cities or small areas as research areas, and few are evaluated from a whole region. Therefore, this paper selects the Beijing-Tianjin-Hebei region as the research area, and adopts the two methods of land use conversion matrix and landscape index to analyze the changes in land use status on a larger spatial scale.
In 2014, the National Basic Geographic Information Center launched Glo-beLand 30, a global 30 m resolution remote sensing cartographic data product for surface coverage includes products of 2000 and 2010. At present, the Chinese government has donated and opened up to the United Nations to share the data product [18] [19]. Due to the short time of occurrence of GlobeLand30 data, there are few studies on this high-resolution land cover data, most of which are aimed at the accuracy evaluation of the land cover products at global or national scales and land cover change [20] [21] [22] [23]. However, little is known about the landscape pattern analysis at regional scales using GlobeLand30 data. Based on these, this paper adopts the method of land use transfer matrix and landscape index to analyze the Beijing-Tianjin-Hebei region by using two-period data of Globeland 30 and reflects the transfer situation and structural feature of land use types in the region from 2000 to 2010 in a quantitative and intuitive way, so as to provide a reference for analyzing the law of land use transfer in the region and reveal the characteristics of landscape pattern change in Beijing-Tianjin-Hebei.

Overview of the Study Area
The Beijing-Tianjin-Hebei region is located at 113˚04'E -119˚53'E longitude and 36˚01'N -42˚37'N latitude, east of Bohai Sea, west of Taihang Mountains, north of Yanshan Mountain, high in the northwest and low in the southeast, with an area of about 217 thousand km 2 , mainly including Beijing, Tianjin and Hebei Province. [24] Based on the comprehensive analysis of the existing literature [25] [26], this paper divides the Beijing-Tianjin-Hebei area into six parts, they are

Data Sources and Processing
The surface coverage data used in this paper is Globeland30 data with 30 m resolution of 2000 and 2010. Supported by 863 national key projects, the product was developed by the National Basic Geographic Information Center in conjunction with 18 units, such as Beijing Normal University, Tsinghua University and the Institute of Remote Sensing application (IRSA) of Chinese Academy of Sciences [27] [28], using WGS84 coordinate system and UTM projection. Its classified images are mainly multi-spectral images with 30m resolution, including multispectral images of China Environmental Disaster Reduction Satellite (HJ-1), Beijing 1 (BJ-1) and Landsat TM5, ETM+ [18]. Ground cover can be divided into 10 categories: cultivated land, forest, grassland, shrub land, wetland, water body, artificial surface, bare land, glacier and permanent snow cover, tundra. The overall accuracy in 2000 is 85.8%, the Kappa coefficient is 0.79, the overall accuracy in 2010 is 86.9%, and the Kappa coefficient is 0.81 [28]. The Globeland30 data used in this paper is downloaded from the National Basic Geographic Information Center, which website is http://ngcc.sbsm.gov.cn/.
In GIS software, the data in the study area is obtained by projection transformation, image mosaic, research area clipping and so on. Because shrub land and bare land occupy less area in Beijing-Tianjin-Hebei area, and there are no glaciers, permanent snow and tundra areas in the study area, it is necessary to reclassify these data and eventually merge them into seven types of land cover, namely cultivated land, forest, grassland, wetland, water body, artificial surface and others.

Land Use Transfer Matrix
The transfer matrix of land use reflects the dynamic process information of the transformation between different land types of area at the beginning and the end of a certain period in a certain region. It includes not only the static data of different types of area at a certain time point in a certain region, but also the more abundant information of different types of area transfer at the beginning and at the end of the period. The general form of land use transfer matrix is [29] [30]: 11 12 1 where S represents area; n is the number of land use types before and after transfer; i, j (i, j = 1, 2, ••• n) represents the land use types before and after the transfer, respectively; S ij denotes the area of the class i before the transfer to the class j after the transfer.

Landscape Index
Landscape index refers to quantitative indicators which can highly condense the information of landscape pattern and reflect some characteristics of landscape structural composition and spatial allocation [31]. Due to the high correlation between landscape pattern indexes, the representative, easy to quantify and relatively sensitive indicators should be selected according to the characteristics of the study area, geographical environmental factors and research needs, combined with the ecological meaning of each indicator. [32] [33] The landscape pattern can be analyzed from three different levels: patch level, patch class level and landscape level. Based on the existing literature [34], this paper intends to use landscape level index to analyze the fragmentation and spatial heterogeneity of the Beijing-Tianjin-Hebei region. Patch Density (PD), Edge Density (ED), Landscape Shape Index (LSI), Contagion Index (CONTAG), Shannon Diversity Index (SHDI) and Shannon Evenness Index (SHEI) are selected to study Landscape pattern change of Beijing-Tianjin-Hebei. The formula for calculating the specific index is shown in Table 1.

Land Use Transfer in Beijing-Tianjin-Hebei Region
Based on the land cover remote sensing image classification results of the two phases of GlobeLand30, and adopting the land use transfer matrix method, the GIS tools were used to generate the transfer matrix between land use types in Beijing-Tianjin-Hebei area, as shown in Table 2. It can be seen from Table 2 that from 2000 to 2010, land transfer in the entire Beijing-Tianjin-Hebei region mainly occurred between cultivated land, artificial surface and grassland, while the changes between forests, wetlands and water bodies were smaller, and their changes were 349.72 km 2 , 175.23 km 2 , −189.94 km 2 . During the ten-year period, the total area of arable land decreased by 3721.67 km 2 , and its transfer-out area was 7813.11 km 2 , of which the area converted into grassland and artificial surface

Land Use Transfer in the Subareas of Beijing-Tianjin-Hebei
Similarly, using the land-use transfer matrix method and the GlobeLand30 land cover classification images of the six districts of Beijing-Tianjin-Hebei, transfer matrices between the land-use types of the six districts were generated. In order to more intuitively display the changes in the land use types and the magnitude of the changes in each subarea from 2000 to 2010, a percentage accumulation chart was used, as shown in Figure 2.

Landscape Index Analysis
Through the landscape index analysis method, the landscape index bar charts of six districts of Beijing-Tianjin-Hebei were generated, as shown in Figure 3 Figure 3, we can see that: The PD values of Sand Control Farmland Protection Zone in Yellow River Flood Plain, Human Settlement Environment Maintenance and Farmland Protection Zone of Beijing-Tianjin-Hebei Urban Agglomeration and Ecological Maintenance Area of Water Conservation in Hilly Area of Yanshan Mountain increase, indicating that the degree of fragmentation of the three areas is gradually increasing. At the same time, the area of artificial surface in these three areas has increased considerably. Therefore, it may be due to the strengthening of urbanization activities, which weakens the ability of landscape to resist natural disasters, and reduces the carrying capacity of ecological environment.

Conclusions
The Beijing-Tianjin-Hebei region is taken as the research area, and the study area is divided into six zones. Using the two-period data of Globeland30 of 2000 and 2010, the land use transfer status and the dynamic change of landscape pattern in Beijing-Tianjin-Hebei region from 2000 to 2010 are quantitatively analyzed by adopting the methods of land use transfer matrix and landscape index analysis and using GIS technology and FRAGSTAT software.
1) From the perspective of land use change, the land use change in Beijing-Tianjin-Hebei from 2000 to 2010 is obvious. The general trend of land use change in the whole Beijing-Tianjin-Hebei region is the decrease of the area of arable land and waterbody, and the increase of the area of forest, grassland, wetland and artificial surface. Among them, cultivated land, artificial surface and grassland are the three land use types that have changed most. From the six zones of Beijing-Tianjin-Hebei, it can be seen that the partitions with significant changes are Sand Control Farmland Protection Zone in Yellow River Flood Plain, Water Conservation Area for Sand Control in Mountain and Hilly Areas of Northwest Taihang Mountains, and Human Settlement Environment Maintenance and Farmland Protection Zone of Beijing-Tianjin-Hebei Urban Agglomeration. Generally speaking, the decrease of cultivated land area and the increase of man-made surface area in Beijing-Tianjin-Hebei region coincide with the urbanization process in Beijing-Tianjin-Hebei region.
2) From the perspective of landscape pattern change, the patch density in the southeast and northwest of Beijing-Tianjin-Hebei region increases from 2000 to 2010, indicating that the degree of fragmentation in these regions increased gradually; the increase of edge density indicates the increase of patches and discontinuity of landscape types in Beijing-Tianjin-Hebei area, and the increase of landscape shape index indicates that landscape patches in the study area are complex and irregular in shape; the increase of the contagion index in the southeastern part of Beijing-Tianjin-Hebei region indicates that the dominant patches in the landscape form a good connection, while the increase of Shannon diversity index and Shannon evenness index in Beijing-Tianjin-Hebei region indicates the increase of patch types and the uniform distribution of patch types in the landscape.
3) This paper focuses on the landscape pattern evaluation and analysis of the fragmentation and spatial heterogeneity of Beijing-Tianjin-Hebei region by using the landscape level index. Therefore, the next step can be evaluated by selecting the landscape components and the biodiversity of the landscape, as well as the landscape index reflecting the direction and strength of human activities. At the same time, this paper evaluates the ecological zoning of Beijing-Tianjin-Hebei region, but not through the selection of the scale of typical cities, which is the focus of the next work.