Assessment of Ecosystem Service Quality and Its Correlation with Landscape Patterns in Haidian District, Beijing

Landscape architecture with urban green space as the main research object is an evidence-based science. It is an important issue to optimize green space systems from the point of view of ecosystem services. In this paper, high-resolution (1.5 m resolution) remote sensing images are combined with data-processing software, such as ENVI, ArcGIS, and Fragstats, to evaluate ecosystem service quality and compute the landscape pattern in the Haidian District (Beijing, China), so that the relationship between the ecosystem service quality and landscape pattern can be quantitatively studied and a strategy can be provided for green space optimization in cities. The following conclusions are drawn: (1) for the evaluated quality of 14 ecosystem services in Haidian District (refer to Section Analysis of the Association of the Percentage of Patches (PLAND) Index of Forest Land and Quality of Ecosystem Service in Haidian District). Forest land is the main provider of the ecosystem service in Haidian District, while construction land only provides cultural services; (2) on the whole, the spatial distribution of the ecosystem services in Haidian District gradually decreases from the west to the east, which basically matches with the spatial distribution of the forest land. The regulating service and supporting service are matched with the distribution of the urban green space. The cultural service is closely associated with history resource points; and (3) the analysis results of the association between landscape pattern and ecosystem service quality show that the percentage of patches (PLAND) index for forest land has a significant logarithmic relationship with the regulating service and supporting service. The critical value of the PLAND index is 30. Besides the Xishan area with the most coverage of forest land, the landscape shape index (LSI) of the brushwood has a logarithmic relationship with the ecosystem service quality. The critical value of the LSI value is 50. Finally, this paper proposes an area optimization strategy of green space in Haidian District from the view of the ecosystem system service. The Xishan area is classified into the ecosystem red line to control city expansion. The regulating and supporting services can be enhanced in the north flat area by improving the patch shape index. The ecosystem service capabilities can be improved by adding the forest land in the existing green space for the southeast urban areas.


Introduction
Ecosystem services (ES) are considered to be the benefits obtained by human beings from an ecosystem. ES are generally divided into four categories: supply services, regulating services, grime on the leaf surface have better effect [56]. Urban green spaces intercept rainfall, suppress evaporation, reduce surface runoff, and increase the underground runoff to conserve contents under common action of plants, soil, and weather, and the water conservation capabilities can be quantified by regulating water capacity [57].
The supporting service is an elementary function required to ensure other ecosystem service functions, and it can influence human beings in an indirect or long-term action, e.g., soil formation, nutrient circulation, and soil conservation. The soil conservation is based on the following principle: The living mulch and withered and fallen substance layer will intercept rainfall to reduce the flushing of water drops on surface soils and corrosion of surface runoff. The plant root system can conserve soils, avoid the collapse and effusion of soils, reduce soil fertility loss, and improve soil structure [58].
Cultural service indicates intangible interests, entertainment, spiritual feeling, and aesthetic experience acquired by people from the ecosystem. Cultural services are very subjective and are difficult to evaluate quantitatively. With the development of technology and the generation of big data, partial experts and scholars evaluate ecosystem cultural services by using the point of interest (POI) of photos.

Haidian District Overview
The Haidian District is located in the northwest area of the central area of Beijing City ( Figure 1) and has an area of 430.8 km 2 . The Haidian District has a semi-humid continental monsoon climate of a warm temperate zone. The average annual rainfall is 585 mm. The district is located at the crossing point between the edge of the northern North China Plain and the Taihang Mountains and includes the mountainous areas in the west with over 100 m elevation and the east plain. Haidian District includes about 10 large and small rivers, and the green coverage is 52.2%. The Haidian District is located at the combination belt of a shallow mountainous area, plain area, and urban constructed are, and has rich green space and a diversified landscape pattern. This district has historical cultural resources, with "Three Hills and Five Parks" as the core, and has very high ecological, cultural, and social value.

Method
A variety of technical methods and means were adopted in the study, and the specific technical recording route is shown in Figure 2.

Method
A variety of technical methods and means were adopted in the study, and the specific technical recording route is shown in Figure 2.

Data Sources and Processing
A high-resolution (1.5 m resolution) remote sensing image of the GF-2 satellite taken on 12 September 2015 was used as the data source (images purchased from the China Resources Satellite Application Center were processed by the Beijing Changdi Friends Mapping Technology Center). Geometric correction, radiometric calibration, and atmospheric correction were performed in pre-processing ( Figure 3). The supervision classification in the ENVI 5.3 software (Harris Geospatial Solutions, Broomfield, CO, USA) is used to explain green space in Haidian District. With Classification of Land Utilization Conditions as the classification reference, and by combining actual conditions in the research area, the green space in Haidian District can be divided into forest land, brush, grassland, water area, construction land, and other lands. The explanation standards are shown in Table 1.

Data Sources and Processing
A high-resolution (1.5 m resolution) remote sensing image of the GF-2 satellite taken on 12 September 2015 was used as the data source (images purchased from the China Resources Satellite Application Center were processed by the Beijing Changdi Friends Mapping Technology Center). Geometric correction, radiometric calibration, and atmospheric correction were performed in preprocessing ( Figure 3). The supervision classification in the ENVI 5.3 software (Harris Geospatial Solutions, Broomfield, CO, USA) is used to explain green space in Haidian District. With Classification of Land Utilization Conditions as the classification reference, and by combining actual conditions in the research area, the green space in Haidian District can be divided into forest land, brush, grassland, water area, construction land, and other lands. The explanation standards are shown in Table 1.

Type of Land Coverage Determination Mark Picture
Forest land Located in mountainous areas. The image is deep red, grainy, and has even texture.

Brush
Mainly located at the foot of hills and inside the city. This image includes deep red and shiny red and is grainy.

Grasslands
Its area is small inside the city. The image is shiny red.

Construction land
Distributed in the constructed urban area. The image is cyan and has distinctive geometric features.

Water area
The image is deep cyan and has a distinctive boundary and fine texture.

Other lands
Mainly located outside the city. The image is light gray and white.

Brush
Mainly located at the foot of hills and inside the city. This image includes deep red and shiny red and is grainy.

Type of Land Coverage Determination Mark Picture
Forest land Located in mountainous areas. The image is deep red, grainy, and has even texture.

Brush
Mainly located at the foot of hills and inside the city. This image includes deep red and shiny red and is grainy.

Grasslands
Its area is small inside the city. The image is shiny red.

Construction land
Distributed in the constructed urban area. The image is cyan and has distinctive geometric features.

Water area
The image is deep cyan and has a distinctive boundary and fine texture.

Other lands
Mainly located outside the city. The image is light gray and white.

Grasslands
Its area is small inside the city. The image is shiny red.

Type of Land Coverage Determination Mark Picture
Forest land Located in mountainous areas. The image is deep red, grainy, and has even texture.

Brush
Mainly located at the foot of hills and inside the city. This image includes deep red and shiny red and is grainy.

Grasslands
Its area is small inside the city. The image is shiny red.

Construction land
Distributed in the constructed urban area. The image is cyan and has distinctive geometric features.

Water area
The image is deep cyan and has a distinctive boundary and fine texture.

Other lands
Mainly located outside the city. The image is light gray and white.

Construction land
Distributed in the constructed urban area. The image is cyan and has distinctive geometric features.

Type of Land Coverage Determination Mark Picture
Forest land Located in mountainous areas. The image is deep red, grainy, and has even texture.

Brush
Mainly located at the foot of hills and inside the city. This image includes deep red and shiny red and is grainy.

Grasslands
Its area is small inside the city. The image is shiny red.

Construction land
Distributed in the constructed urban area. The image is cyan and has distinctive geometric features.

Water area
The image is deep cyan and has a distinctive boundary and fine texture.

Other lands
Mainly located outside the city. The image is light gray and white.

Water area
The image is deep cyan and has a distinctive boundary and fine texture. Table 1. Land coverage type determination mark of picture.

Type of Land Coverage Determination Mark Picture
Forest land Located in mountainous areas. The image is deep red, grainy, and has even texture.

Brush
Mainly located at the foot of hills and inside the city. This image includes deep red and shiny red and is grainy.

Grasslands
Its area is small inside the city. The image is shiny red.

Construction land
Distributed in the constructed urban area. The image is cyan and has distinctive geometric features.

Water area
The image is deep cyan and has a distinctive boundary and fine texture.

Other lands
Mainly located outside the city. The image is light gray and white.
Other lands Mainly located outside the city. The image is light gray and white.

Type of Land Coverage Determination Mark Picture
Forest land Located in mountainous areas. The image is deep red, grainy, and has even texture.

Brush
Mainly located at the foot of hills and inside the city. This image includes deep red and shiny red and is grainy.

Grasslands
Its area is small inside the city. The image is shiny red.

Construction land
Distributed in the constructed urban area. The image is cyan and has distinctive geometric features.

Water area
The image is deep cyan and has a distinctive boundary and fine texture.

Other lands
Mainly located outside the city. The image is light gray and white.

Division of Research Units
In order to study the relationship between the ecosystem service quality and landscape pattern in Haidian District, the picture was evenly divided into 4000 × 4000 m fishnet grids in the ArcGIS software (ESRI, Redlands, CA, USA), and a total of 42 research units were obtained (Figure 4) to study the relationship between the ecosystem service quality and landscape pattern. Compared to the administrative area division, terrain, landform, and urbanization, the grid division can eliminate influences such as man-made management and natural features, and data results are not objective.

Computing Method of Ecosystem Service Quality
The service quality evaluation method is a quantitative method that is frequently used for the analysis of ecosystem service, and which can objectively show structures, functions, and ecology of the ecosystem. In this paper, the carbon fixation and oxygen release, cooling and humidifying, air purification, water conservation, soil fixation and fertility conservation functions are selected, according to forest ecosystem service function evaluation specification (LY/T1721-2008) issued by the State Forestry Bureau, by combining related research work of Wu [59]. The service quality evaluation index system of the ecosystem service in Haidian District is constructed according to actual data from Beijing City.
The cultural services are expressed by using the number of POI photo with coordinate information. Flickr (SmugMug, San Francisco, CA, USA) is the largest image sharing website in the world and includes detailed geological information and text labels. Wang [60] analyzed the attraction degree of the landscape based on the Flickr website. Luo identified the scenes based on the Flickr website [61]. Xue [62] evaluated the terrain of a tourist destination based on the Flickr website. For the photographic data analyzed in this paper, the photo number is used to express the cultural service capabilities. The specific steps are as follows: download images with geological coordinate information by using the Flickr Application Programming Interface (API); select labels such as "outdoor" and "landscape"; remove unrelated photos; and, finally, obtain 3363 scene photos from

Division of Research Units
In order to study the relationship between the ecosystem service quality and landscape pattern in Haidian District, the picture was evenly divided into 4000 × 4000 m fishnet grids in the ArcGIS software (ESRI, Redlands, CA, USA), and a total of 42 research units were obtained (Figure 4) to study the relationship between the ecosystem service quality and landscape pattern. Compared to the administrative area division, terrain, landform, and urbanization, the grid division can eliminate influences such as man-made management and natural features, and data results are not objective.

Computing Method of Ecosystem Service Quality
The service quality evaluation method is a quantitative method that is frequently used for the analysis of ecosystem service, and which can objectively show structures, functions, and ecology of the ecosystem. In this paper, the carbon fixation and oxygen release, cooling and humidifying, air purification, water conservation, soil fixation and fertility conservation functions are selected, according to forest ecosystem service function evaluation specification (LY/T1721-2008) issued by the State Forestry Bureau, by combining related research work of Wu [59]. The service quality evaluation index system of the ecosystem service in Haidian District is constructed according to actual data from Beijing City.
The cultural services are expressed by using the number of POI photo with coordinate information. Flickr (SmugMug, San Francisco, CA, USA) is the largest image sharing website in the world and includes detailed geological information and text labels. Wang [60] analyzed the attraction degree of the landscape based on the Flickr website. Luo identified the scenes based on the Flickr website [61]. Xue [62] evaluated the terrain of a tourist destination based on the Flickr website. For the photographic data analyzed in this paper, the photo number is used to express the cultural service capabilities. The specific steps are as follows: download images with geological coordinate information by using the Flickr Application Programming Interface (API); select labels such as "outdoor" and "landscape"; remove unrelated photos; and, finally, obtain 3363 scene photos from Haidian District, Beijing City, and express the cultural service capabilities by using the number of photos.

Computing Method of Landscape Pattern
The Fragstats 4.2 software (Oregon State University, Corvallis, OR, USA) was used to analyze the landscape pattern based on the interpreted data. According to the research results of Cen [63], three landform indices of the patch type level are selected for analysis, including percentage of landscape (PLAND) index, interspersion and juxtaposition index (IJI), and landscape shape index (LSI), for forest land. The PLAND index indicates the percentage of the patch in the landscape and describes the quality characteristics of the research area. The IJI index is used to express the general distribution of the landscape. A higher value indicates that the alternate occurrence law of different patches is more significant, and the types are distributed. The LSI shows the complexity of the shape.

Total Service Quality of Ecosystem Service
The land coverage data for Haidian District in 2015 ( Figure 5), whose interpretation precision is over 80%, shows that the total area of the forest land, brush, grassland, water, and other lands is 23.7 km 2 , which is 56% of the total area of Haidian District. The area of forest land is 12 km 2 , which is half the total area of green space. The area of brush and grassland is basically the same, and the percentage of water is smaller. On the whole, the total green space in Haidian District is rich and mainly includes forest land.
After calculation, the total quality of the ecosystem service in Haidian District in 2015 is described as follows: For regulating service, the carbon fixation capacity and oxygen release capacity are 92, 339, 108 kg, and 247, 206, 505 kg, respectively. The cooling capacity and humidifying capacity are 1, 178, 705, 379, and 339 KJ, and 417, 718, 842, and 3 kg, respectively. The sulfur dioxide, oxynitride, and hydrogen fluoride capacity of the purified air are 2, 368, 389, and 597 kg, 129, 312, and 450 kg, and 4, 19, and 51 kg, respectively. The stagnant dust capacity is 313, 202, 416, and 986 kg. The regulation water capacity is 14, 544, 419, and 439 kg. For the supporting service, the soil fixation capacity is 614, 729, and 832 kg. The nitrogen, phosphor and potassium retention capacity in the soil conservation are 101, 430, and 422 kg, 7, 991, and 488 kg, and 88, 521, and 96 kg, respectively. For the cultural service, total photo is 3, 363. The quality for the ecosystem services of different land coverage types are shown in Table 2. Forest land is the main provider of the ecosystem service in Haidian District, and the quality of its services is over half. The construction land only provides cultural service. Int. J. Environ. Res. Public Health 2019, 16, x 9 of 34 After calculation, the total quality of the ecosystem service in Haidian District in 2015 is described as follows: For regulating service, the carbon fixation capacity and oxygen release capacity are 92, 339, 108 kg, and 247, 206, 505 kg, respectively. The cooling capacity and humidifying capacity are 1, 178, 705, 379, and 339 KJ, and 417, 718, 842, and 3 kg, respectively. The sulfur dioxide, oxynitride, and hydrogen fluoride capacity of the purified air are 2, 368, 389, and 597 kg, 129, 312, and 450 kg, and 4, 19, and 51 kg, respectively. The stagnant dust capacity is 313, 202, 416, and 986 kg. The regulation water capacity is 14, 544, 419, and 439 kg. For the supporting service, the soil fixation capacity is 614, 729, and 832 kg. The nitrogen, phosphor and potassium retention capacity in the soil conservation are 101, 430, and 422 kg, 7, 991, and 488 kg, and 88, 521, and 96 kg, respectively. For the cultural service, total photo is 3, 363. The quality for the ecosystem services of different land coverage types are shown in Table 2. Forest land is the main provider of the ecosystem service in Haidian District, and the quality of its services is over half. The construction land only provides cultural service.

Quality of Ecosystem Service in Unit Area of Research Units
The quality of the ecosystem service of unit area of 42 cells was calculated, and compared with the quality per unit area in Haidian District. The results are shown in Table 3. For the regulating service, the quality of the ecosystem service of units Z10, Z16, Z17, Z18, Z21, Z22, Z23, Z29, and Z31 are more than the average in Haidian District. Except for the regulating water capacity, the quality of units Z18 and Z23 is more than the average. Except for the humidifying capacity, the quality of unit Z30 is more than the whole average. The sulfur dioxide, hydrogen fluoride, and stagnant dust capacity absorbed by units Z11 and Z39 are lower than the average. The water regulation capacity of units Z24 and Z38 is more than the average, which indicates stronger water conservation capability. The quality of the supporting services of units Z10, Z11, Z16, Z17, Z18, Z21, Z22, Z23, Z29, Z30, Z31, Z32, and Z39 is better than the average in Haidian District. The "Three Hills and Five Parks", important historical cultural resources in Haidian District, are located in units Z11, Z17, Z18, and Z19, and the cultural service capabilities of these four units are far stronger than those of other areas; therefore, the cultural service is closely associated with the historical resource points The ratio of the quality of the units to the average is used for data standardization in research. The total quality of the ecosystem of the area of units is shown in Figure 6. The quality of units Z10, Z11, Z16, Z17, Z18, Z19, Z21, Z22, Z29, and Z31 is higher than the average. The quality of unit Z30 is close to the average of Haidian District. The percentages of the forest land, brush, grassland, water area, construction land, and other lands are 26%, 17%, 12%, 1%, 42%, and 3%, respectively. The quality of the ecosystem service of units Z16 and Z21 is maximal, and the percentage of the forest land in these units is 83% and 61%, respectively. The quality of the ecosystem service of unit area of 42 cells was calculated, and compared with the quality per unit area in Haidian District. The results are shown in Table 3. For the regulating service, the quality of the ecosystem service of units Z10, Z16, Z17, Z18, Z21, Z22, Z23, Z29, and Z31 are more than the average in Haidian District. Except for the regulating water capacity, the quality of units Z18 and Z23 is more than the average. Except for the humidifying capacity, the quality of unit Z30 is more than the whole average. The sulfur dioxide, hydrogen fluoride, and stagnant dust capacity absorbed by units Z11 and Z39 are lower than the average. The water regulation capacity of units Z24 and Z38 is more than the average, which indicates stronger water conservation capability. The quality of the supporting services of units Z10, Z11, Z16, Z17, Z18, Z21, Z22, Z23, Z29, Z30, Z31, Z32, and Z39 is better than the average in Haidian District. The "Three Hills and Five Parks", important historical cultural resources in Haidian District, are located in units Z11, Z17, Z18, and Z19, and the cultural service capabilities of these four units are far stronger than those of other areas; therefore, the cultural service is closely associated with the historical resource points The ratio of the quality of the units to the average is used for data standardization in research. The total quality of the ecosystem of the area of units is shown in Figure 6. The quality of units Z10, Z11, Z16, Z17, Z18, Z19, Z21, Z22, Z29, and Z31 is higher than the average. The quality of unit Z30 is close to the average of Haidian District. The percentages of the forest land, brush, grassland, water area, construction land, and other lands are 26%, 17%, 12%, 1%, 42%, and 3%, respectively. The quality of the ecosystem service of units Z16 and Z21 is maximal, and the percentage of the forest land in these units is 83% and 61%, respectively.   (Figure 7), the ecosystem service of Haidian District gradually decreases from the west to the east, and the values of the northwestern units Z16, Z17, Z18, Z21, and Z29 are maximum, which match with the distribution of forest land. The quality of the northern plain (units Z22, Z23, Z24, Z30, Z31, Z32, and Z39) and the southwestern area ahead of hills (units Z10, Z11, and Z19) is followed. The whole share of the green space in these areas is higher. The quality of the ecosystem in the southeastern urban areas is minimal, which is closely related with the share of construction lands, i.e., the urban construction strength. From these data, forest land can be seen to be the significant positive influencing factor for the quality of the ecosystem service in Haidian District, and construction land to be the main negative    (Figure 7), the ecosystem service of Haidian District gradually decreases from the west to the east, and the values of the northwestern units Z16, Z17, Z18, Z21, and Z29 are maximum, which match with the distribution of forest land. The quality of the northern plain (units Z22, Z23, Z24, Z30, Z31, Z32, and Z39) and the southwestern area ahead of hills (units Z10, Z11, and Z19) is followed. The whole share of the green space in these areas is higher. The quality of the ecosystem in the southeastern urban areas is minimal, which is closely related with the share of construction lands, i.e., the urban construction strength. From these data, forest land can be seen to be the significant positive influencing factor for the quality of the ecosystem service in Haidian District, and construction land to be the main negative influencing factor.

Spatial Distribution Features of Ecosystem Service in Haidian District
The quality of the regulating service ( Figure 8) and supporting service (Figure 9) is maximum in the western mountainous area, followed by the northern plain. The quality of the southeastern urban area is minimal, and is basically matched with the distribution of the urban green space. This indicates that the green space area (including forest land, brush, grassland, and water area) is the main factor affecting the regulating service and support service. The cultural service (Figure 10) of "Three Hills and Five Parks" (Z18) is optimal. The cultural service of the southeastern urban area is better than that of the western mountainous area. This indicates that the cultural service is closely associated with the quality and features of the green space. The very high cultural service capability of unit Z18 can significantly improve the quality of the whole ecosystem service. influencing factor. The quality of the regulating service ( Figure 8) and supporting service (Figure 9) is maximum in the western mountainous area, followed by the northern plain. The quality of the southeastern urban area is minimal, and is basically matched with the distribution of the urban green space. This indicates that the green space area (including forest land, brush, grassland, and water area) is the main factor affecting the regulating service and support service. The cultural service ( Figure 10) of "Three Hills and Five Parks" (Z18) is optimal. The cultural service of the southeastern urban area is better than that of the western mountainous area. This indicates that the cultural service is closely associated with the quality and features of the green space. The very high cultural service capability of unit Z18 can significantly improve the quality of the whole ecosystem service.   influencing factor. The quality of the regulating service ( Figure 8) and supporting service (Figure 9) is maximum in the western mountainous area, followed by the northern plain. The quality of the southeastern urban area is minimal, and is basically matched with the distribution of the urban green space. This indicates that the green space area (including forest land, brush, grassland, and water area) is the main factor affecting the regulating service and support service. The cultural service ( Figure 10) of "Three Hills and Five Parks" (Z18) is optimal. The cultural service of the southeastern urban area is better than that of the western mountainous area. This indicates that the cultural service is closely associated with the quality and features of the green space. The very high cultural service capability of unit Z18 can significantly improve the quality of the whole ecosystem service.

Analysis of the Association Between the Landscape Pattern and Quality of the Ecosystem Service in Haidian District
The Fragstats software was used to calculate the landscape indices of 42 research units, and the calculated results are shown in Table 4.

Analysis of the Association Between the Landscape Pattern and Quality of the Ecosystem Service in Haidian District
The Fragstats software was used to calculate the landscape indices of 42 research units, and the calculated results are shown in Table 4.

Analysis of the Association between the Landscape Pattern and Quality of the Ecosystem Service in Haidian District
The Fragstats software was used to calculate the landscape indices of 42 research units, and the calculated results are shown in Table 4.       1 The PLAND index indicates the percentage of the patch in the landscape and describes the quality characteristics of the research area. See document Fragstats Help [64] for more details of the algorithm. 2 The interspersion and juxtaposition index (IJI)index is used to express the general distribution of the landscape. See document Fragstats Help [64] for more details of the algorithm. 3 The Landscape shape index (LSI) index measures the perimeter-to-area ratio for the landscape as a whole. See document Fragstats Help [64] for more details of the algorithm.
The association between the landscape and quality of ecosystem service in different units was studied. The results show that the PLAND index of the forest land and the LSI of the brush are significantly associated with the quality of the ecosystem service. This paper mainly studies two associations.

Analysis of the Association of the Percentage of Patches (PLAND) Index of Forest Land and Quality of Ecosystem Service in Haidian District
The association between the PLAND index of the forest land and the quality of the ecosystem service in the research units is shown in Figures 11-24. The PLAND index has a significant logarithmic relationship with the regulating service and supporting service, however has no significant association with the cultural service ( Figure 24). When the PLAND value is between 30 and 40, the regulating services, such as carbon fixation and oxygen release, cooling and humidifying, and air purification, Figures 11-19 result. When the PLAND value is more than 40, the growth trend will slow down. The regulation water capacity and soil fixation and fertility retention service (Figures 20-23) will result when the PLAND value is between 20 and 30. When the PLAND value is more than 30, the growth will slow down. It is concluded that the quality of the regulating service and supporting service grows quickly when the forest land area is 30% of the landscape area. When the proportion of the forest land is less than 30%, with growth of the proportion of the forest land, the quality of the regulating and supporting service grows quickly. When the proportion is over 30%, the quality growth will slow down. Health 2019, 16, x 20 of 34 and 40, the regulating services, such as carbon fixation and oxygen release, cooling and humidifying, and air purification, Figures 11-19 result. When the PLAND value is more than 40, the growth trend will slow down. The regulation water capacity and soil fixation and fertility retention service (Figures 20-23) will result when the PLAND value is between 20 and 30. When the PLAND value is more than 30, the growth will slow down. It is concluded that the quality of the regulating service and supporting service grows quickly when the forest land area is 30% of the landscape area. When the proportion of the forest land is less than 30%, with growth of the proportion of the forest land, the quality of the regulating and supporting service grows quickly. When the proportion is over 30%, the quality growth will slow down.  and 40, the regulating services, such as carbon fixation and oxygen release, cooling and humidifying, and air purification, Figures 11-19 result. When the PLAND value is more than 40, the growth trend will slow down. The regulation water capacity and soil fixation and fertility retention service (Figures 20-23) will result when the PLAND value is between 20 and 30. When the PLAND value is more than 30, the growth will slow down. It is concluded that the quality of the regulating service and supporting service grows quickly when the forest land area is 30% of the landscape area. When the proportion of the forest land is less than 30%, with growth of the proportion of the forest land, the quality of the regulating and supporting service grows quickly. When the proportion is over 30%, the quality growth will slow down.                               Figure 25 show the relationship between the LSI of brush of research units and the quality of ecosystem service in the units. The results show that the LSI of the brush and regulating service (Figures 25-33) and supporting service (Figures 34-37) show significant exponential growth trends. The cultural service (Figure 38) grows. When the LSI value is 50, the regulating service and supporting service start to grow quickly. It is concluded that the landscape is optimized with 50 as the critical LSI value. The values of units Z16, Z17, Z21, and Z29 in the curve do not satisfy the growth trend. By studying these four units, they are located in the western mountainous areas, and the proportion of the forest land is high (PLAND values are 52.36, 61.14, 54.28, and 54.58, respectively). The quality of the provided ecosystem service is dominant. At this time, it is not associated with the shape index of the brush patches. Therefore, the research results show that the shape index of the brush is exponentially associated with the quality of the ecosystem service, except in the Xishan area with most forest land in Haidian District, and 50 is also the critical value of the LSI.  Figure 25 show the relationship between the LSI of brush of research units and the quality of ecosystem service in the units. The results show that the LSI of the brush and regulating service (Figures 25-33) and supporting service (Figures 34-37) show significant exponential growth trends. The cultural service (Figure 38) grows. When the LSI value is 50, the regulating service and supporting service start to grow quickly. It is concluded that the landscape is optimized with 50 as the critical LSI value. The values of units Z16, Z17, Z21, and Z29 in the curve do not satisfy the growth trend. By studying these four units, they are located in the western mountainous areas, and the proportion of the forest land is high (PLAND values are 52.36, 61.14, 54.28, and 54.58, respectively). The quality of the provided ecosystem service is dominant. At this time, it is not associated with the shape index of the brush patches. Therefore, the research results show that the shape index of the brush is exponentially associated with the quality of the ecosystem service, except in the Xishan area with most forest land in Haidian District, and 50 is also the critical value of the LSI.   (Figures 34-37) show significant exponential growth trends. The cultural service (Figure 38) grows. When the LSI value is 50, the regulating service and supporting service start to grow quickly. It is concluded that the landscape is optimized with 50 as the critical LSI value. The values of units Z16, Z17, Z21, and Z29 in the curve do not satisfy the growth trend. By studying these four units, they are located in the western mountainous areas, and the proportion of the forest land is high (PLAND values are 52.36, 61.14, 54.28, and 54.58, respectively). The quality of the provided ecosystem service is dominant. At this time, it is not associated with the shape index of the brush patches. Therefore, the research results show that the shape index of the brush is exponentially associated with the quality of the ecosystem service, except in the Xishan area with most forest land in Haidian District, and 50 is also the critical value of the LSI.            . Figure 34. Association between brush LSI and soil fixation capacity.

Analysis of Landscape Shape Index (LSI) of Brush and Quality of Ecosystem Service in Haidian District
. Figure 35. Association between brush LSI and nitrogen retention capacity.

Association between Landscape Pattern and Ecosystem Service
Research on the ecosystem of urban green spaces can facilitate the protection of the urban ecological system and improve the residential environment. The quantitative evaluation of the ecosystem is the important prerequisite for the protection and improvement of the ecosystem environment. From the view of the ecosystem service, by studying the association between the landscape pattern and service quality, the quantitative ratio of the land coverage type is the important influencing factor for the quality of the ecosystem service. Tang, Shao et al. [65] studied the quality of the ecosystem service in the southern areas of Guizhou Province, China, and proposed that a higher forest land coverage will indicate better ecosystem service condition. Additionally, Zhang et al. [66] studied the karst region in the northwest of Guangxi Province, China, and found that the forest land and brush are the main providers of the ecosystem service, accounting for about 70% of the total service. Furthermore, Cui and Xu [54] studied different ecosystems in Beijing and found that forest land provides the maximal ecosystem service. This research conclusion is consistent with the results of the present study. A high-resolution (1.5 m resolution) remote sensing image is studied in this paper. The data precision is far higher than that in past research. Therefore, the relationships determined in the present quantitative research are more reliable than those of past research. The research results show that the forest land PLAND index has a significant logarithmic correlation with the regulating service and supporting service, with the critical PLAND index value being 30. When this value is less than 30, with growth of the forest land proportion, the quality of the regulating and supporting service will grow quickly. After 30 days, the quality growth will slow down. According to regulations in China's "Forest Law of the People's Republic of China", the forest coverage rate shall reach 30% in China, which matches the present research results. The present research results show that the more complicated landscape shape index will bring a higher quality of ecosystem service under certain conditions. With Ningbo City, China, as one example, Cen concluded that the complexity of the forest land patch will affect the water conservation and soil conservation. The present research results show that the shape index of the brush is exponentially associated with the quality of the ecosystem service, except in the Xishan area with most forest land in Haidian District, and 50 is also the critical value of the LSI value.

Optimization Strategy of Green Space in Haidian District
According to the present research results, from the view of the ecosystem service, this paper proposes an optimization strategy for the green space in Haidian District, Beijing City. The ecosystem service functions are better in the western mountainous areas with higher forest coverage. Therefore, these areas should be protected and an ecological red line should be drawn to strictly control city expansion. The regulating and supporting services are better in the northern plain area. Under the prerequisite of protecting the ecosystem, the ecosystem service is further enhanced by improving the patch shape index. The cultural service is better and the regulating and supporting services are worse in the southeast of the city. By referring to the research results, the forest land proportion should be added in the existing green space to improve the ecosystem service capability of the urban area and enhance the service capability of the urban ecosystem.

Conclusions
Using a high-resolution (1.5 m resolution) remote sensing image from the GF-2 satellite as the foundation, the quality of the ecosystem service in Haidian District, Beijing, China, was evaluated, and the research units were divided to study the association between the landscape pattern and ecosystem service. The following conclusions are drawn: (1) The research results show that forest land is the main provider of the ecosystem service in Haidian District, while construction land only provides cultural service. For the total quality of the unit area of the ecosystem service, the quality of units Z10, Z11, Z16, Z17, Z18, Z19, Z21, Z22, Z29, and Z31 is higher than the quality of unit Z30 approximates the whole average in Haidian District; (2) On the whole, the ecosystem service spaces of Haidian District are divided into the western mountainous area, the northern plain, and the southwestern urban area, and decrease from the west to the east. These are roughly matched with the spatial distribution of the forest land. The regulating service and supporting service are maximum in the western mountainous area, followed by the northern plain. The southwestern urban area has the minimal regulating service and supporting service. This is consistent with the distribution of the urban green space. The cultural service capabilities in the research unit where there are important historical cultural resources in Haidian District-namely, the "Three Hills and Five Parks"-are far stronger than those of other areas. Therefore, the cultural service is closely associated with the historical resource points; (3) The PLAND index of the forest land has a significant logarithmic relationship with the regulating service and supporting service in the analysis on association between the landscape pattern and ecosystem service quality. The critical PLAND index value is 30. When the PLAND value is smaller than 30, with growth of the proportion of the forest land, the quality of the regulating and supporting service will grow quickly. When the value is over 30, the quality growth will slow down. Additionally, the research results show that the shape index of the brush is exponentially associated with the quality of the ecosystem service, except in the Xishan area with most forest land in Haidian District (the quality of the ecosystem service of the forest land is dominant, and at this time it is not associated with the shape index of the brush spot). The critical LSI value is 50; (4) Finally, this paper proposes an area optimization strategy for green space in Haidian District, Beijing City, from the view of the ecosystem system service. The Xishan area is classified as the ecosystem red line to control city expansion. The regulating and supporting services can be enhanced in the northern flat area by improving the patch shape index. The ecosystem service capabilities can be improved by adding forest land in the existing green space for the southeastern urban areas.
This paper studies the quality of the ecosystem service, however, does not cover research on biological diversity conservation. This will be studied later. The research results show that the distribution of the regulating and supporting service is roughly the reverse of that of the cultural service. Therefore, it is concluded that the ecosystem services are coordinated and balanced with each other, and the relation mechanism shall be further studied to provide theoretical support for urban construction decisions.