Establishing an ecological security pattern for urban agglomeration, taking ecosystem services and human interference factors into consideration

The assessment of ecological security patterns is a topic of conversation in landscape ecology in recent years. However, ecosystem services and human activities are seldom considered comprehensively in the assessment of ecological security patterns. The present study employs the Beijing–Tianjin–Hebei urban agglomeration as a study area, and uses ecological services to determine the ecological sources. The importance of ecological sources is classified based on logical coding and functional types of ecological services. The research combines regional characteristics to select and quantitatively calculate three human disturbance factors: soil erosion sensitivity, geological hazard sensitivity, and night lighting. Then the basic surface resistance of land use to limit migration is modified and ecological corridors are identified by combining these three disturbance factors. The results indicate that the sources of water production, soil and water conservation, and carbon fixation are mainly provided in mountainous areas, recreation sources are mostly distributed in the plains, and these ecological sources improve the maintenance of ecological corridors. The modification of resistance surfaces significantly changes the length of ecological corridors in Tianjin, Tangshan, Cangzhou, and Beijing, and the modified resistance surface improves the recognition of ecological corridors. This study provides a new research framework for identifying the ecological security patterns of urban agglomerations and provides scientific guidance related to ecological protection and urban planning for the Beijing−Tianjin−Hebei urban agglomeration.


INTRODUCTION
While urbanization provides beneficial results to modern civilization, it also creates a series of ecological and environmental problems, such as the loss of natural landscapes, a decline in ecosystem services, and the aggravation of environmental health risks, which can affect the sustainable development of cities (Han, Liu & Wang, 2015). Therefore, ensuring clear for urban development and construction. The problem is related to the rational use of land resources and the sustainable development of human society (Egoh & Reyers, 2008). In recent years, most studies have shown that clarifying the impact of ecological service categories on the importance of the ecological source areas is helpful in understanding the corresponding relationship between ecological source areas and ecological service functions. Moreover, understanding the relationships among multiple ecosystem services and the mechanisms behind these relationships will improve our ability to sustainably manage landscapes in order to provide multiple ecosystem services (Bennett, Peterson & Gordon, 2009) toward an integrated ecological sources management (Raudsepp-Hearne, Peterson & Bennett, 2010;Kareiva et al., 2007).
The next step in ecological security pattern development is to identify ecological corridors. In the extraction of ecological corridors, the most commonly used method is the GIS-based modeling approach proposed by Knaapen, Scheffer & Harms (1992) and improved by Yu (1999) based on the minimum cumulative resistance (MCR) model. GIS is a comprehensive technology for processing geospatial information. It can not only express visible spatial positional relationships, but also reflect spatial locations or related numerical values, words and charts that reflect the characteristics of things (Beni et al., 2011). Urban ecological planning is necessary to accurately judge and decide the spatial layout relationship between urban ecological environment and urban development and construction, and as a complex and comprehensive research field, it needs the abstract data of various spatial information to be processed (Aretano et al., 2015). GIS as a decision support tool for environmental protection and nature reserve protection can play a targeted role in spatial relationship processing and data analysis. The GIS-based modeling approach has been used for numbers of urban planning applications in China (Su et al., 2016;Dong et al., 2015), where resistance is the degree of impediment to species migration between landscape units. Currently, the development of resistance surfaces is generally based on developing a resistance coefficient related to ecological factors such as land cover type and slope (Fu et al., 2010). The method does not consider the spatial differences caused by specific ecological problems in different regions. The correction of the resistance surface is also based on nighttime lighting  or imperious index data (Peng et al., 2018c) that are used to enhance the spatial differences in resistance surfaces. The method considers the spatial differences caused by specific ecological problems in resistance surfaces, but is generally based on the qualitative evaluation of ecological security patterns for a single problem, the extent of the spatial coverage is generally aimed at the province (Peng et al., 2018a), city (Li et al., 2010), or county level (Yu et al., 2018). As a relatively complex ecosystem, an urban agglomeration involved various aspects of ecological security, each of which included many factors influencing urban conditions. A comprehensive evaluation cannot be done by studying ecological security based on a single factor .
The health and sustainable development of the Beijing-Tianjin-Hebei urban agglomeration is important ensurancing of china's national security. However, human activities become more frequent with intense urbanization, which leads to a series of ecological problems (Gong et al., 2009;Liu et al., 2017;Duan et al., 2015). Among them, soil and water conservation serve as an important ecological safeguards in support of economic and social development these should play an important role in the coordinated development of the Beijing-Tianjin-Hebei region (Chen et al., 2017;Miao et al., 2011). In addition, the fragile ecological base, changing climatic conditions, and intense activity frequently cause geological disasters in the Beijing-Tianjin-Hebei urban agglomeration (Meng et al., 2017). Soil erosion and geological hazard sensitivity indices reflect the degree of ecosystem response to human disturbance and natural environment changes (Ouyang, Wang & Miao, 2000). Night light data are widely used in the study of population density, economic development, and urban heat island effects, etc. These spatial measures-indicators are a comprehensive characterization of the intensity of human activity in an area (Mellander et al., 2015). In areas that are highly sensitive to soil erosion and geological hazards or that have nighttime light, the intensity of human activity is normally relatively high, which has a certain effect on regional the migration of urban species. Therefore, in setting the Beijing-Tianjin-Hebei urban agglomeration as the study area, this study employs an analysis of ecosystem services and human disturbance factors to construct ecological security patterns, with the goal of reducing the negative effects of human activities on the ecological security pattern of the Beijing-Tianjin-Hebei urban agglomeration. Hence, the research objectives were proposed as follows: (1) One objective was to determine the ecological source of various resources through a quantitative evaluation of water production services, soil and water conservation, carbon fixation services, as well as leisure and recreation services; (2) to determine the importance of dividing ecological sources by combining logical coding and ecological service function types; finally, (3) using the soil erosion sensitivity, geological hazard sensitivity, and nighttime light indices, we wanted to modify ecological resistance surfaces based on land use to identify ecological corridors.

METHODOLOGY
Ecological security patterns combine landscape pattern with ecological processes to achieve regional ecological sustainability by comparing the importance of different landscape patterns (Peng et al., 2018d). The ecological security pattern of the Beijing-Tianjin-Hebei urban agglomeration constructed in this study consists of four steps: firstly, ecological services are used to determine the ecological sources. Second, the importance of the ecological sources is determined by combining the logical coding and identification of the type of ecological services. Thirdly, the basic resistance surface is modified with human interference factors. Finally, the ecological corridor is extracted by using the MCR model. The flow chart of this study is shown in Fig. 1.

Study area and data sources
The Beijing-Tianjin-Hebei urban agglomeration lies in the coastal area of Eastern China, and covers a total area of 218,000 km 2 . In this region, the Yanshan Mountains delimit the north, the North China Plain lies to the south, the Taihang Mountains stand to the west, and the Bohai Sea borders the eastern edge of the region (Fig. 2). The cold and snowy climate in winter gives way to an arid spring with high wind speeds and frequent sandstorms, summers are hot and rainy, which is conducive to rock and soil weathering. According to a survey, by 2015, a total of 7,255 geological hazards and hidden dangers were found in the study area, including 3,526 collapses, 770 landslides, 2,288 debris flows, and 671 ground collapses (Meng et al., 2017). The human population accounts for 7.8% of the total population of China, making this urban agglomeration a densely populated area. Data preprocessing and data processing are mainly done with GIS tools. The data sources are shown in Table 1.

Ecological sources
Ecological service function assessment Assessment of the importance of ecosystem services frequently serves as a basic method used to identify regional sources of ecological services. This is done by analyzing variation in regional ecological services and identifying the areas that are most important for the maintenance of typical regional ecosystems. Therefore, the present study quantitatively identifies and evaluates services of water production, water and soil conservation, carbon fixation, and those related to leisure and recreation of the urban agglomeration in combination with an analysis of the characteristics of the natural environment of the Beijing-Tianjin-Hebei urban agglomeration.
In particular, the function of water production is based on the water cycle. Water production depends on the rainfall in a grid unit minus the actual evapotranspiration (Miao, Ni & Borthwick, 2010). In the present study, the Integrated Valuation of Ecosystem  where A c is the amount of soil and water conservation, t/(hm 2 ·a); A p is the potential amount of soil erosion; A r is the actual amount of soil erosion; R is the factor of rainfall erosivity, MJ·mm/(hm 2 ·h·a); K is the factor of soil erodibility, t·hm 2 ·h/(hm 2 ·MJ·mm); L and S are the topographic factors; L are the factor of slope length. S is the slope factor, and C is vegetation cover management factor. The service of carbon fixation was evaluated basing on the principle of light energy utilization. The estimation formula is shown in Eq. (2): Where PAR is photosynthetically active radiation, the unit is MJ/m 2 ; FPAR is the absorption component of the vegetation layer to the incident photosynthetically active radiation, no unit; (ε Ã is the maximum light energy utilization rate, the unit is gC/MJ; f(t) is the influence factor of temperature on the utilization of light energy, no unit; f(w) is the influence factor of water on the utilization of light energy, no unit; R a is the autotrophic respiration of vegetation, the unit is gC/m 2 . The leisure and recreation service mainly considers the wetlands of the Beijing-Tianjin-Hebei urban agglomeration. Buffer zones of pedestrian walking distance to these wetlands of 5, 10, 15, 20, and 25 min were analyzed, which are included in different levels of leisure and recreation, aiming at emphasizing the effects of human activities (Peng et al., 2018b).

Ecological source level
Based on quantitatively evaluating the ecosystem services of the studied region and identifying the ecological source areas, logical coding is introduced to identify the ecosystem services of different ecological sources. This coding is used to represent one or more types of ecosystem services contained in the ecological source area. The logical coding in this study uses "1" or "0" to indicate that a grid has or does not have a particular ecosystem service, respectively. In the present study, the raster computing function of ArcGIS ver. 10.2 was combined to obtain the spatial pattern of the logical coding of each ecological service category. Equation (3) provides the logical coding system: in which T is the logical coding value of column j and row i in the raster image of logical analysis results, which indicates one or more ecosystem service categories, and n is the number of ecosystem services. The present study looked at four ecosystem services. Therefore, n = 4 and G1 ij , G2 ij , G3 ij , and G4 ij are the logical coding of the corresponding raster types on the raster images of the services of water production, water and soil conservation, carbon fixation, and leisure with recreation, respectively. For example, if the type logical coding is "1101," it means that the ecological source area has the services of water production, water and soil conservation and leisure, and does not have the service of carbon fixation.
In the present study, the ecological source areas were classified as described above based on the logical coding to determine the importance of each ecological source area. If three or four of the ecological services are coded using the number "1," the area is determined to be a primary ecological source; if only two or one are coded using the number of "1," the area is determined to be a secondary or tertiary ecological source.

Resistance surface
The concept of an ecological resistance surface refers to the degree of impediment to species migration between landscape units. The spatial differences of specific ecological problems in small-scale areas can often be solved by constructing desirable ecological resistance surface based on the single-factor correction of land cover types. However, as a relatively large urban agglomeration, land managers of the Beijing-Tianjin-Hebei region should not only consider the ecological problems confined to specific areas, but should also analyze the region as a whole. Geological conditions in the Beijing-Tianjin-Hebei urban agglomeration are complex and fragile, landslides, debris flows, and serious soil erosion occur frequently in the region. Increasingly human activities interfere with the natural landscape conditions, this is especially true as it relates to the expansion of urbanization. These factors will have a great impact on species migration and regional ecological security and stability. Therefore, in this study, sensitivity indices tied to geological hazards, soil erosion, and night light data are introduced to modify the basic ecological resistance surface based on land use. In addition, the ecological red line corridors were manually vectorized in the Beijing, Tianjin, and Hebei Red Lines of Ecological Protection in 2018. Initially, ecological resistance is assigned to 0 and finally superimposed on the basic resistance surface, then the revised resistance surface is converted into raster format. Among, the above three sensitivity indices, that of soil erosion adopts the evaluation indexes of precipitation erosivity, soil erodibility, slope length, and surface vegetation cover, which have been proposed based on the red line technical guidelines for ecological protection of Yang et al. (2016). The sensitivity index of geological hazards used nine indicators to calculate the total distribution weight of the layers ( Table 2). The weight of each index was determined by analytic hierarchy process (AHP) (Saaty, 1987). The AHP is completed by Yaahp software, which is a comprehensive evaluation assistant software based on the AHP and fuzzy comprehensive evaluation method. The consistency test of the results of the AHP is carried out where the consistency index value is 0.1056, because the calculation of geological hazard sensitivity includes nine indicators, namely n = 9, and the table is found, random index = 1.45, consistency ratio = 0.0728 < 0.1, indicating the judgment matrix has satisfactory consistency. In this paper, with the sensitivity analysis in Yaahp software, the number of sampling points is set to 100, and the corresponding total ranking weights when the middle layer elements change from 0 to 1 are calculated. The basic sensitivity indicators of disaster, terrain and other are 0.0423, 0.0317, and 0.0255, respectively. This indicator highlights the maximum range of weight change of each option caused by the change of the current element weight. Finally, the three factors were normalized and replaced by the specific ecological resistance surface correction as follows in Eq. (4): where R i is the ecological resistance coefficient of grid i based on the water loss sensitivity, night lighting, and geological hazard sensitivity indices, NL i is the night light of grid i, NL a is the average night light of the land use type a corresponding to grid i; SE i is the water and soil erosion sensitivity index of grid i; SE a is the average soil erosion sensitivity index of land use type a corresponding to grid i. GS i is the sensitivity index of geological hazards of grid i and GS a is the average geohazard sensitivity index of land use type a corresponding to grid i, and R is the basic resistance coefficient of land use type corresponding to grid i.
Since it is impossible to determine which factors have a greater impact on the resistance surface of an urban agglomeration, the weight of night lighting, soil and water loss sensitivity index, and geological hazard sensitivity index is set to 1:1:1, that is w b :w c :w d = 1:1:1.

Ecological corridors
Constructing ecological corridors can solve the problem of fragmentation related to ecological sources and enhance the connectivity among ecological sources. With the rapid development of the Beijing-Tianjin-Hebei urban agglomeration, the islanding and fragmentation of the local landscape has increased dramatically. The MCR model was first proposed by Knaapen, Scheffer & Harms (1992). This model considers three factors, source, distance, and landscape interface, to calculate the cost to a species caused by moving from a source to a destination, and it extracts the resistance trough between two adjacent "sources" and the most easily connected low-resistance channel as a corridor between ecological safe habitat sources as shown in Eq. (5): where MCR is the MCR value; D ij is the spatial distance of species from source J to landscape unit i, R i is the resistance coefficient of landscape unit i; and f is the positive correlation between MCR and ecological process.

Spatial pattern of ecological sources
Spatial pattern of ecological services The spatial pattern of a single ecosystem services can reflect the effects of different ecological processes on regional ecological security. The assessment results of ecosystem services were divided into five levels from low to high by the natural breakpoint method as follows: generally, unimportant and slightly, moderately, highly, and extremely important (Fig. 3), at the same time, the area and area ratio occupied by important areas of various ecosystem service was calculated (Table 3). In the Beijing-Tianjin-Hebei urban agglomeration the water production service was generally higher in mountainous and hilly areas than that of plain and plateau areas, it is also higher in the western and eastern areas than in the central areas ( Fig. 3A; Table 3). Most of the extremely important areas are located west of Handan and Xingtai, northwest of Baoding, and in a few scattered areas of the eastern coastal areas including Qinhuangdao, Tangshan, and Cangzhou. Most of these areas are located in the windward slopes of the Yanshan and Taihang Mountains that have a large amount of topographical relief and receive relatively higher amounts of precipitation, so the overall water production capacity is strong there. The total amount of the moderately and highly important areas for the water production service is not very different, the proportions of these areas were 8.37% and 8.26%, respectively, and they were mainly distributed in the central plain area. Moreover, the spatial extent of generally important areas was the largest, accounted for 46.97%, indicating that this urban agglomeration belongs to an area of serious water shortage.
In the Beijing-Tianjin-Hebei urban agglomeration the distribution of the soil and water conservation of area gradually decreased from the Yanshan and Taihang mountains to the surrounding areas ( Fig. 3B; Table 3). The aforementioned area was the largest covering a total amount of 167,863.23 km 2 , accounting for 78.31% of the total study area. Except for extremely important areas, the proportion of other grades was not very different. The combined areas of the moderately, highly, and extremely important regions was 35,058.85 km 2 , accounting for only 16.36% of the total area, indicating that the overall soil and water conservation capacity of the region is relatively weak. As an important ecological area of soil and water conservation in the Beijing-Tianjin-Hebei urban agglomeration, these areas are rich in precipitation, while the factors of soil erodibility are small, or the vegetation cover has a good ability to regulate and store precipitation; that is, these areas have a strong ability conserve water and soil.
In the Beijing-Tianjin-Hebei urban agglomeration the carbon fixation service is higher in the north and lower in the South and higher in the west and lower in the east ( Fig. 3C; Table 3). The spatial differences among the important regions are relatively large. The areas of the highly and extremely important regions were 9,691.46 km 2 , and 20,117.15 km 2 , respectively, with the area ratios of 4.48% and 9.31%, respectively. They were mainly distributed in a small part of Handan, Xingtai, Shijiazhuang, Baoding, and Beijing and most of Chengde. Most of the land use types in these areas were forest, with abundant vegetation coverage, a humid climate, and good hydrothermal conditions, which can better meet plant growth requirements than other habitats. The area of moderately important areas covered 77,111.86 km 2 which accounted for 35.67% of the study area. Most of moderately important areas were surrounded by highly and extremely important areas. The areas of generally and slightly important areas were 30,006.56 km 2 and 79,240.88 km 2 which accounted for 13.88% and 36.66% of the study area, respectively. These areas are basically developed but unused areas that are mainly distributed in the central, southern, and eastern parts of the region. The leisure and recreational service of the Beijing-Tianjin-Hebei urban agglomeration presented a multi-point distribution pattern as a whole ( Fig. 3D; Table 3). This service was mainly distributed in the plateau area of Zhangjiakou, the eastern area of Beijing, the southeastern area of Tianjin, and along the coastal area. The spatial extent of the most important area was 8,701.79 km 2 , and it accounted for 4.03% of the entire area.

Spatial pattern of the level of ecological sources
The extremely important areas of each ecosystem service were selected as ecological source areas. The spatial pattern of the logical coding of the ecological service categories is shown in Fig. 4. The area and proportion of logical coding of each ecosystem source type was also calculated ( Table 4).
The Beijing-Tianjin-Hebei urban agglomeration had eight types of ecological source areas ( Fig. 4A; Table 4). The area of ecological sources covered 82,605.78 km 2 , accounting for 38.21% of the total area. Among them, the plain area provided the source of leisure and recreation services, and the source logical coding containing this function is 0001, accounting for 0.92% of the area of this urban agglomeration. The mountainous area is the source of water production, soil and water conservation and carbon fixation services. Among the four types of ecological sources, carbon fixation sources accounted for the largest proportion.
Most of the mountain areas are first or second-level ecological sources (Fig. 4B). Plains are typically a third-level ecological source. The ecological source areas were primarily distributed in the regions of dense mountains and sparse plains. The numbers of ecological sources at different levels were counted. A total of 15, 38, and 97 primary, secondary, and tertiary ecological sources areas were identified, respectively. The spatial extent of primary, secondary, and tertiary ecological source areas was 27,999.21 km 2 , 25,378.32 km 2 , and 15,505.07 km 2 , respectively. Mountainous areas provided first and second-level ecological source areas, and the plains provided third level ecological source areas. The results indicate that the habitat quality of the Beijing-Tianjin-Hebei urban agglomeration is better than that of the plains area, and the source land of the plains area is severely fragmented.

Spatial pattern of ecological corridors and resistance surfaces
Resistance surface modification is the core part of the development of an ecological security pattern. The Beijing-Tianjin-Hebei urban agglomeration frequently experiences geological disasters, serious soil erosion, and frequent human activities. Therefore, this study uses soil erosion sensitivity, geological hazard sensitivity, and night lighting to modify the resistance surface directly based on land use patterns. The averages of the soil erosion sensitivity index, nighttime light, geological disaster sensitivity index, and resistance values of each city in the Beijing-Tianjin-Hebei urban agglomeration were calculated individually (Table 5). At the same time, the average indices of soil erosion sensitivity, nighttime light index, and geological hazard sensitivity of various land use types were calculated (Table 6).
Higher values for soil erosion sensitivity index (Fig. 5A) were mainly distributed in the Taihang Mountains, the intermountain basins in the northwest portion of the Hebei Province, and the Yanshan Mountains in north China. The average sensitivity index of soil erosion of each land use type was measured independently. The average soil and water loss sensitivity index values of Chengde, Beijing, Qinhuangdao, and Zhangjiakou were    (Table 6).
The highest values of the night light index were mainly distributed in the economically developed cities (Fig. 5B), with urbanized land having the largest average night light indices ( Table 6). The average night light index values of Tianjin, Langfang, Beijing, and Tangshan were relatively large (Table 5). The average night light of Chengde and Zhangjiakou were relatively small. At a regional scale, the distribution of the sensitivity index of geological hazards was highly correlated with the macro-geomorphological gradient and the micro-topographic characteristics of the landscape (Fig. 5C). The geological structure has a pattern of NE-SW and NW-SE directions, being highest in the NE and NW. The average geological hazard sensitivity index of Beijing and Chengde was relatively large, while that of Hengshui and Cangzhouwas relatively small ( Table 5). The high values of geological hazard sensitivity were mainly distributed inforestland in the mountainous regions (Table 6).
Based on the resistance value of land use in the Beijing-Tianjin-Hebei area determined by land use type and using methods described by Xie et al. (2015), the resistance value of land use in the Beijing-Tianjin-Hebei urban agglomeration and its corresponding landscape resistance value were determined based on the land use type (Fig. 5D). The resistance value of the forest land, wetland, grassland, farmland, unused land, and construction land were 10, 40, 80, 100, 500, 800, respectively. The revised resistance surface was calculated based on the basic resistance surface, night lighting, soil erosion sensitivity index, and geological hazard sensitivity index (Fig. 5E). The revised resistance surface exhibited significant changes within the same category, and had a relatively distinct spatial heterogeneity, which can more accurately characterize the resistance differentiation of biological migration in the Beijing-Tianjin-Hebei urban agglomeration. The average resistance values of Langfang, Tianjin, and Tangshan were relatively large, while those of Chengde and Zhangjiakou were relatively small (Table 5).
Using the Linkage Mapper plug-in, the ecological source and the modified resistance surface data were input into Arcgis 10.2 to obtain a distribution map showing the locations of ecological corridors (Fig. 5F). At the same time, the length and variation of ecological corridors in each ecological city before and after the resistance surface modification are calculated (Table 7).
The length of ecological corridors in some cities of the Beijing-Tianjin-Hebei region have changed significantly over time (Table 7). The length of ecological corridors in Tianjin, Tangshan, Cangzhou, and Beijing changed relatively greatly. Among them, Tianjin has experience major changes in night lighting and soil erosion, Tangshan and Beijing are greatly affected by geological hazards and night lighting, and Cangzhou is sensitive to geological hazards and soil erosion. There are 313 ecological corridors in the Beijing-Tianjin-Hebei urban agglomeration, with a total length of 9,399.59 km (Fig. 5F). The Yanshan and Taihang mountains have a relatively high density of ecological corridors, mainly because of the good quality of the habitats, the strongest correlation between various ecological corridors in the region, and the possibility is greater that various species are able to overcome resistance and migrate. The plains area of central and southern Hebei, has a very sparse distribution of ecological corridors. This primarily occurs because the land use types in this area are mainly farmland and urbanized land with relatively poor quality habitat, and the possibility of various species being able to overcome resistance to migration is relatively small.

Planning of ecological security pattern
The spatial extent of ecological source areas determined based on the ecological services they provide was 68,882.81 km 2 ; 63.16% of the ecological source areas was within the red line of ecological protection. The areas of overlap between the primary, secondary, and tertiary ecological sources and red line ecological protection were 18,580.57 km 2 , 14,066.25 km 2 , and 6,723 km 2 , respectively. The areas where the identified ecological source area overlapped with the existing ecological protection red line areas gradually decreased over time with the reduction of the level of ecological source area, indicating the rationality of the identification of ecological source areas. Figure 5F shows that the corridors between the plains are relatively long, the ecological source base is basically the third-level ecological source, and the area of ecological source is less. Therefore, it is suggested to add radiation sources on the basis of the original ecological source. The primary ecological source area is arcuate along Yanshan and Taihang mountains. Attention should be paid to the protection of habitats in the aforementioned areas due to their large size and the short ecological corridor. The secondary ecological sources are mainly distributed in the northern mountains of Heibei Province, here, regional habitat restoration should be given more attention, with the goal of forming an ecological security pattern for the Beijing-Tianjin-Hebei urban agglomeration with reasonable spatial distribution and enhanced ecological service functions.
Contrast with the existing ecological security pattern As noted above, the red line of ecological protection refers to China's strict land use control boundary demarcated by law for important ecological functional and sensitive areas and vulnerable areas, providing the last line of defense for national and regional ecological security. By geometric correction of the spatial pattern of this ecological protection red line, the delineation of this red line in the analyzed urban area was obtained by vectorization (Fig. 6A). In the Beijing-Tianjin-Hebei urban agglomeration the ecological protection red line area covers 62,329.78 km 2 , with a total length of the ecological red line corridor of 2,670.57 km. Because the ecological red line corridors are relatively small, this study defines expressways, rivers, and the ecological red line corridor as the current corridor. The habitat in this vector includes forestland, grassland, and wetland in the ecological red line area; it includes nature reserves, water conservation areas, large-scale ecological land use areas and is based on land use data in the Beijing-Tianjin-Hebei planning outline that were regionally vectorized and superimposed. The ecological source area within the study area was extracted and found to cover 4,689.72 km 2 . Combined with the revised resistance surface, an ecological corridor with human interference factors was generated (Figs. 7A, 7D, 7G, and 7J), and the ecological corridor considering the  Table 8. The distribution of ecological corridors was generally consistent when the three methods of delineating corridors were compared ( Fig. 6B; Table 8). Because of differences in human disturbance and ecological services available in some areas, in this study, the overlapping lengths of the ecological corridors identified by ecosystem services and human interference factors and the current corridor became significantly longer, indicating that comprehensive consideration of ecological services and human interference factors can improve the identification of ecological corridors in urban agglomerations.

Determination of ecological sources and level of ecological sources
In terms of ecological source identification special protection areas of the Natura 2000 network that contain forests and agroforest mosaics were selected as core areas to be connected through ecological corridors at a regional scale in the Basque Country (Gurrutxaga, Lozano & Barrio, 2010). This method of directly selecting nature reserves as the core ecological source focuses on considering the functional properties of ecological land patches, but does not consider the spatial structure importance of the ecological source in the whole landscape and the relationship with the surrounding environment. This research effort main goal was to quantitatively evaluate the functions of ecological services, water production services, soil and water conservation services, and leisure services on the basis of absorbing the methods of considering the importance of ecological services in the past, taking into account the importance of the ecological source itself and its structure in the landscape pattern. In the selection of ecological service factors, Zhang et al. (2017) selected biodiversity service, soil conservation service, and water resource security combining the natural environment characteristics of Beijing-Tianjin-Hebei urban agglomeration. On this basis, this paper combines the research results of Peng et al. (2018d) and the ecological protection red line technical guidelines delineated by the Ministry of Environmental Protection of China. Thereby enriching the ecological service supply of the Beijing-Tianjin-Hebei urban agglomeration. In terms of ecological source identification, Zhang et al. (2017) proposed a new evaluation framework integrating ecosystem services importance assessment and landscape connectivity analysis with human ecological demand importance assessment to identify ecological sources. This paper introduces logical coding to establish the spatial relationship between the ecological source and its corresponding ecological services, and describes the spatial pattern of the two by spatial rasterization. Based on logic coding, the ecological service logic map and the ecological source map of importance are generated. From the ecological service logic map, what ecological services are available in the ecological source area can be judged by the location of "1," "0" and its "1" and "0." From the ecological source map of importance, the ecological source level can be derived. This is of great significance for the management of refined ecological sources.

Selection of human interference factors
In recent years, the modification of a resistance surface that considers specific ecological problems for specific regions has become an important area of research in the optimization of ecological security patterns. . Many scholars studied the influence of spatial differences caused by specific ecological problems in different regions on species migration; these factors are collectively referred to as human interference factors, which are used to express the degree of impact that human activities cause to natural environments. For small regions, the effects of a single factor on species migration can be considered for specific ecological environment problems. However, an analysis of the relatively large ecosystem of the Beijing-Tianjin-Hebei urban agglomeration needs to take into account the protection of biodiversity, the restoration of degraded ecosystems, and the sustainable development of the social economy. The purpose of the present study is to systematically help land managers solve regional environmental problems. The singular problems of environmental pollution or biological resources protection have been extended to the systematic analysis and comprehensive study of regional ecological environment problems (Franklin, 1993). Borders of mountains, plains and plateaus can be seen from Fig. 4B. The plateau is located in a small part of the northwest direction of the study area. The plain is located in most of the southwest direction, and the middle part of the plain and plateau is the mountain. The human interference factors selected in this paper include soil erosion sensitivity, geological hazard sensitivity, and nighttime light index. It can be seen from the results in Figs. 5A-5C that the areas of this high geological hazard while soil erosion sensitivity are concentrated in the mountainous areas of urban agglomeration, and the areas of high night light are concentrated in the plains area. The established resistance surface considers the human activity situation of the entire area of Beijing-Tianjin-Hebei. So, the results highlight that the selected human interference factor is relatively reasonable.

CONCLUSION
Based on ecosystem services and human disturbance factors, this study defined the ecological security pattern of the Beijing-Tianjin-Hebei urban agglomeration, which provides a new holistic framework. The main content of this study include the use of ecosystem services to determine ecological sources, and the establishment of the logical relationship between ecological sources and ecosystem services. Additionally, ecological corridors were identified using a MCR model based on sources and resistance surface modified through human disturbance factors. The results show that water production, soil and water conservation, and carbon fixation sources were mainly distributed in mountainous regions of the study area, recreational sources were mostly distributed in the plains, and the extracted ecological sources improve the recognition of ecological corridors. The modification of resistance surface makes the length of ecological corridors in Tianjin, Tangshan, Cangzhou, and Beijing change significantly, while this modified resistance surface improves the recognition of ecological corridors. Based on the importance of the divided ecological sources and the identified ecological corridors, this study proposes some planning suggestions for the existing ecological security pattern with the goal of forming a reasonable spatial layout and improving the ecological service function of the ecological security pattern of this agglomeration. Future research of the proposed methodology contains three aspects: (a) a more analytical ecological source identification is that comprehensive consideration of the trade-off and synergy of ecological services to identify the level of ecological source. (b) In terms of resistance surface correction, the nighttime light index is used to directly correct the resistance surface based on land use, the soil erosion sensitivity index is used to correct the resistance surface based on land use, and the resistance surface based on land use is corrected by geological disaster sensitivity. The length of the ecological corridor and the coincidence length of the actual ecological corridor, explore the degree of influence of different human interference factors on the resistance surface. (c) In the aspect of ecological corridor extraction, the identified ecological source level is introduced into the gravity model, and the traditional gravity model is improved to make the identified ecological corridor more scientific and reasonable.