Research on Ecosystem Security and Restoration Pattern of Urban Agglomeration in the Yellow River Basin

Abstract

Reasonable identification of the ecosystem security space and pattern restoration for the Yellow River Basin will be significant for facilitating future ecological protection and restoration projects. This study employs ecosystem services and landscape connectivity to discern ecological source areas and conducts an optimization evaluation method of resistance surfaces and the minimum cumulative resistance (MCR) model to identify the ecosystem security pattern of urban agglomeration. Then, restoration measures are proposed. The results indicate that: (1) There are 46 ecological source areas in the ecosystem security patterns of the Lanxi urban agglomeration, with a total area of 8199.249 km². Moreover, in terms of spatial distribution, ecological source areas are sparse in the east while crowded in the west of the Lanxi urban agglomeration and are mainly composed of natural reserves, forest parks, and farmlands. (2) The ecosystem security patterns contain 914 ecological corridors with a total length of 62,970.181 km, most of which are short-distance corridors, being dense in the part of the northwest with a webbed feature. (3) The study adds 10 ecological source areas to restore the ecosystem security pattern and ecological corridors to improve the rationality of the ecosystem security pattern. Meanwhile, this study proposes restoration measures to protect the ecological environment by defining the levels of ecological security in order to provide a reference for the government to formulate policies and regulations.

1. Introduction

As threatening issues facing ecological and environmental security, ecosystem degradation and environmental pollution imperil the development of the national economy and society [1]. In the process of urbanization, individuals only concentrate on the production function of the ecosystem while ignoring the ecological function, triggering the landscape fragmentation of the ecological environment [2]. The regional ecology thereby faces tough challenges concerning the reduction of natural ecological space and the deterioration of ecological functions such as climate regulation and water conservation [3]. Along with the joint development and special integration of resources, ecology, and the economy, ecosystem security calls for improving the urban ecological environmental security and then promoting regional sustainable development [4]. Ecological security is defined as a comprehensive framework to mitigate ecological risk and maintain ecological health [5,6]. This means that the restoration of the regional ecological security pattern has an indispensable and irreplaceable role in dealing with ecological problems and in transforming how society and the economy develop. Ecological security serves to solve ecological landscape fragmentation, boost the species survival, and enhance the ecosystem service function, thereby tackling ecosystem security issues in development decisions [7,8].

As for restoration and environmental governance, the core is the comprehensive governance of the river basins, more importantly, ecological protection [2]. The river basin is a complex system composed of various composite functions of nature, society, and the economy, with a hierarchical structure and overall functions. Meanwhile, the river basin ecosystem brings about a heated discussion in the field of global ecological protection and restoration and even in earth system science. Previous research shows that the ecological environment of river basins faces challenges such as water pollution [9], natural forest vegetation reduction, soil pollution, soil erosion, biodiversity degradation, etc. In the future, to ensure the integrity of basins and maintain the characteristics of their ecological functions, it is essential to study the ecological issues in the target river basins, discern their ecological security patterns, and take corresponding measures to effectively handle the environmental issues. By doing so, we will promote the sustainable development of the Chinese river basins.

At present, a research model of ecosystem security has gradually formed which turns out to be a canonical paradigm that is “identifying source areas, constructing the resistance surfaces, extracting the corridors” [10]. Researchers use two main methods to identify source areas. The first is to directly select single patches, such as nature reserves, scenic resorts, or long-term, stable forests and cultivated lands, as ecological sources [11]. The other is to construct a comprehensive evaluation system to identify ecological source areas [12,13] and to feature the functional value of target ecological patches. The methods also include the construction of a resistance surface [14] that mostly assigns resistance coefficients in terms of natural factors such as elevation and altitude. However, it is difficult to reach consensus on the various types of resistance coefficients. Accordingly, the research methods mainly focus on quantitatively identifying ecological spaces by using circuit theory [15], MCR model [5], morphological spatial pattern analysis (MSPA) [16], etc. In particular, the MCR model is relatively mature in practical application, intuitively expressing the intrinsic connections of the energy flow of ecology patches in a river basin. From the perspective of administrative units, the research scale is at the level of provinces [17], cities [18], and counties [19], paying more attention to Karst mountainous areas [20], waterfront cities [21], and agriculture and husbandry [22] interconnected areas, etc. Little research has approached the issues from the perspective of urban agglomeration. Urban agglomeration is the pivotal issue in protecting ecological environments in river basins and is interconnected with the internal ecological elements within a city. Therefore, it is important to research the function of ecosystem security patterns from the perspective of urban agglomeration in the same way used to explore the western river basins from the perspective of administrative units. That is, lacking forceful empirical illustrations, studies of the ecological source areas need to be based on the features of the patterns themselves by using the interactions of multiple ecosystem service functions. By the same token, few studies combining the principles of landscape ecology aim to restore the ecosystem security patterns and develop corresponding measures facilitating their development.

The Lanxi urban agglomeration located in the upper reaches of the Yellow River is now facing ecological issues, such as a weak ecological environment and soil erosion, etc. Its ecosystem security pattern not only directly affects the sustainable development of the local social economy but also plays a vital role in the construction of ecological barriers in the Yellow River Basin. Thus, it is urgent to identify the ecological system security pattern and to take related restoration measures in the Lanxi agglomeration. This study, based on multisource data, superimposes four ecological functions including water source conservation, water and soil conservation, windbreak and sand fixation, and biodiversity to acquire ecosystem service values in order to approximate the levels of ecosystem security service in the Lanxi urban agglomeration. Accordingly, this study constructs the ecological resistance surfaces in terms of the realities of the study area, reflects ecological security conditions of the Lanxi urban agglomeration by using the MCR model and the gravity model, and proposes relevant restoration measures and strategies about the ecosystem security pattern. It is expected that the results of the study will provide related references for the local ecological environment bureau and other relevant government departments to formulate policies, guidelines, standards, and regulations for improving the ecosystem service level of the Lanxi urban agglomeration, ensuring the construction of ecosystem security and sustainable development in the Yellow River Basin.

2. Study Area and Data Sources

2.1. Study Area

The Lanxi urban agglomeration, along the Yellow River Basin, stretches across Gansu Province and Qinghai Province, which borders the Qilan Mountains in the north, includes the Hehuang Valley in the middle, and enjoys the benefits of the grassland in the south, playing a vital strategic role in protecting “China’s water tower” and preventing the desertification area in the west from spreading eastward. In the Lanxi urban agglomeration, the area of Gansu Province is long and narrow and surrounded by mountains, with various and complex landforms mainly containing mountains, plateaus, deserts, etc., whose terrain slopes from southwest to northeast. In addition, for the area of Qinghai Province, the terrain is mainly composed of lofty mountains, basins, and grasslands, abounding in a stable natural ecosystem and relatively comprehensive ecological service. By the same token, Qilian Mountain Nature Reserve is an ecological security barrier for the Lanxi urban agglomeration, supporting wind protection and sand fixation, maintaining water resources and biological diversity, and conserving soil and water in the ecological space formed by the Yellow River, the Huangshui River, and other rivers. According to the Development Plan of Lanzhou-Xining Urban Agglomeration, the study area incorporates 39 regions such as Chengguan District of Lanzhou, with a total area of 97,500 km² (see Figure 1).

2.2. Data Sources

The original data in the study are from land use, remote sensing images, digital elevation model (DEM), soil, meteorology, economy, society, etc., including processing and calculation. The detailed information is in

Table 1. Data resources.

Data Name	Data Accuracy	Data Sources	Time
The data set of net primary productivity of vegetation (NPP)	30 m	The results of the remote sensing survey and assessment of ecological status in China	2000–2021
The data sets of meteorology	—	The Data Center for Environmental Sciences, Chinese Academy of Sciences (http://www.resdc.cn/DOI).	2010–2020
The data of net primary productivity of visible infrared imaging radiometer suite (NPP-VIIRS)	500 m	The National Aeronautics and Space Administration (https://www.nasa.gov/)	2020
The raster data of land use	30 m	The Data Center for Environmental Sciences, Chinese Academy of Sciences (http://www.resdc.cn/DOI)	2020
The data of normalized difference vegetation index (NDVI)	30 m	The Data Center for Environmental Sciences, Chinese Academy of Sciences (http://www.resdc.cn/DOI)	2018
The data set of soil	1:106	Soil Science Database, China (http://www.tpdc.ac.cn/zh-hans/DOI)	2017
Population density	—	The Data Center for Environmental Sciences, Chinese Academy of Sciences (http://www.resdc.cn/DOI)	2019
DEM, slope	30 m	The geographical space data from Computer Network Information Center, Chinese Academy of Science The website of cloud image (http://www.gscloud.cn)	2020
The distances to the Yellow River and roads	—	Euclidean distance analysis and calculation on ArcGIS platform	2020

, and the following multisources data were uniformly processed into 250 m resolution raster data by Geography Information Science 10.6 (GIS 10.6) was invented by Environmental Systems Research Institution (Esri) and renewed in 2017 in Redlands, CA, USA, and it can be obtained from https://www.esri.com/en-us/home.

3. Methods

3.1. Identify the Importance of Ecosystem Services

In this study, four factors as functional evaluation indicators were selected to identify the spatial pattern of ecosystem services in the target area, including water resource conservation, soil and water conservation, windbreak and sand fixation, and biodiversity. The relative principles and calculation methods can be seen in

Table 2. The principles and calculation methods of the four evaluation factors.

Ecosystem Services	Principles and Methods	Formula
Water resource conservation [23]	The ability to regulate runoff is assessed by indicators such as vegetation net primary productivity and soil seepage.	$WR=NP{P}_{mean}\times F_{sic}\times F_{pre}\times \left(1-F_{slo}\right)$ $WR$ is the service capacity index of the ecosystem water conservation; $NP{P}_{mean}$ means the average net primary productivity of vegetation for many years; $F_{sic}$ represents the soil seepage factor, $F_{pre}$ indicates the annual average precipitation factor, and $F_{slo}$ is the slope factor.
Soil and water conservation [10]	The amount of water and soil conservation is calculated by the difference between the potential soil erosion and the actual soil erosion.	$\mathrm{A}=\mathrm{R}\times \mathrm{K}\times \mathrm{L}\times \mathrm{S}\times \left(1-\mathrm{C}\right)$ A is the amount of soil and water conservation; R is the rainfall erosion factor; K refers to the soil erosion factor; L is the slope length factor; S is the slope factor, and C presents the vegetation factor.
Windbreak and sand fixation [23]	The capacity of preventing land desertification and protecting farmland from wind and sand is calculated by the net primary productivity of vegetation.	$S_{ws}=NP{P}_{mean}\times K\times F_q\times D$ $S_{ws}$ is the index of service capacity index of windbreak and sand fixation; $NP{P}_{mean}$ means the average net primary productivity of vegetation for many years; K is the soil erosion factor; $F_q$ is the multiyear average climatic erosion, and D is the surface roughness factor.
Biodiversity [24]	According to the biodiversity service equivalent weight, various protected areas are divided into important areas, and then combined with biological and other factors, the biodiversity service equivalent is calculated.	$D_i=\frac{NDV{I}_i}{NDV{I}_t}\times D_t$ $D_i$ is the biodiversity service equivalent weight based on NDVI correction; $NDV{I}_i$ means the NDVI mean value of the grid i, $NDV{I}_t$ is the NDVI mean value of all grids in landscape type t where the grid i located in; $D_t$ is the biodiversity service equivalent weight of landscape type t corresponding to the grid i.

.

3.2. Ecological Network Construction and Restoration

3.2.1. Discerning Ecological Source Areas

Landscape connectivity servers discern the ecosystem stability between two ecological patches, regulating the flow of ecological processes. The calculation formula is as follows [25].

$$\mathrm{PC}=\frac{{{\displaystyle \sum }}_{\mathrm{i}=1}^{\mathrm{n}}{{\displaystyle \sum }}_{\mathrm{j}=1}^{\mathrm{n}}{\mathrm{a}}_{\mathrm{i}}{\mathrm{a}}_{\mathrm{j}}{\mathrm{p}}_{\mathrm{ij}}^{*}}{{\mathrm{A}}_{\mathrm{L}}^2}$$

(1)

$$\mathrm{dPC}=100\%\times \left(\mathrm{PC}-{\mathrm{PC}}_{\mathrm{remove}}\right)/\mathrm{PC}$$

(2)

In the formula, PC is the possible connectivity index of the landscape; PCremove is the overall connectivity index of the landscape; dPC evaluates the landscape connectivity of ecological patches; n is the total number of ecological patches; $a_{i }$ and $a_j$ correspondingly represent the areas of ecological patches i and j; and ${\mathrm{p}}_{\mathrm{ij}}^{*}$ is the maximum probability of species distribution in ecological patches i and j. The dPC is calculated by Confor to demonstrate the smoothness or obstruction of landscape types to ecological flow diffusion. The distance of the probability index was 2500 m, and the probability was 0.5.

In order to ensure the integrity of ecological source units and maintain the connectivity of ecological flows, this study comprehensively selected multifunctional ecological patches as ecological source units whose patch areas were more than 5 km² and landscape connectivity indexes were more than 0.2.

3.2.2. Constructing Resistance Surfaces

The study selected 9 ecological resistance factors from the natural environment, ecological resources, and economic society, aiming to construct the resistance surface. Through conducting the analytic hierarchy process (AHP) and consulting experts, the 9 factors were evaluated and weighted by comparing their relative importance (see

Table 3. Ecological resistance factors and resistance coefficients in Lanxi.

Resistance Classification	Resistance Factors	Weights	Units	Resistance Coefficient
				Level 1	Level 2	Level 3	Level 4	Level 5
Natural environment	DEM	0.071	m	<1300	1300–2000	2000–2500	2500–3000	>3000
	Slope	0.113	(°)	<0.5	0.5–1.5	1.5–2.5	2.5–4.5	>4.5
	Soil erosion	0.149	—	General	Mild	Moderate	High	Extremely high
Ecological resources	Vegetation cover	0.145	%	<20	20–40	40–60	60–80	>80
	Land use	0.243	—	Wetland, water body, tundra	Shrubs, grassland, forest	cultivated land	Artificial land, Bare land	The glacier
	Distance to the Yellow River	0.105	km	<5	5–20	20–50	50–100	>200
Social economy	Distance to the road	0.061	m	<500	500–1000	1000–2000	2000–5000	>5000
	Night lights	0.027	—	<20	20–80	80–150	150–200	>200
	Population density	0.085	people/km2	<1000	1000–4000	4000–10,000	10,000–30,000	>30,000

and Figure 2). Moreover, by using Yaahp software, the study was conducted to build a judgment matrix, and by comparing the factors of each row and each column score, and normalized eigenvector of the largest eigenvalue of the judgment matrix, the weight value of each factor was obtained which was checked through a consistency test to ensure that the tested coefficient was less than 0.1 (CR = 0 < 0.1) [26] which was in complete consistency. Then, we were able to determine the weight of the ecological resistance factor of the Lanxi urban agglomeration. The results are shown in Table 3 and Figure 2.

The spatial distribution of ecological resistance in the study area was explored by using exploratory spatial data analysis (ESDA) [27]. The grid of total resistance values was calculated on the ArcGIS platform, and the results were produced with an accuracy of a 500 m × 500 m grid map. Moreover, the local indicators of spatial association (LISA) aggregation map of ecological resistance were generated in terms of Moran’s I index of ecological security calculated by Geoda.

3.2.3. Ecological Corridor Extraction and Ecological Node Determination

Ecological corridors connect scattered ecological sources, being regarded as the low-resistance channels for the transfer and exchange of ecosystem energy [17]. The relative formula is as follows:

$$\mathrm{MCR}={\mathrm{f}}_{\min }{\displaystyle \sum }_{\mathrm{j}=\mathrm{n}}^{\mathrm{i}=\mathrm{m}}{\mathrm{D}}_{\mathrm{ij}}\times {\mathrm{R}}_{\mathrm{i}}$$

(3)

where D_ij represents the spatial distance from ecological source j to landscape unit I, Rj is the resistance coefficient of landscape unit i to biological movement, and f symbolizes the positive correlation between ecological process and MCR.

The gravity model is conducted to discern the importance of ecological corridors. The formula is as follows [5]:

$$G_{ab}=\frac{N_a{N}_b}{D_{ab}{}^2}=\left[\left(\frac{1}{p_a}\times \ln S_a\right)\left(\frac{1}{p_b}\times \ln S_b\right)\right]/{\left(\frac{L_{ab}}{l_{max}}\right)}^2$$

(4)

where G_ab is the interaction force between ecological source areas a and b, D_ab means the standardized value of ecological corridor resistance between two ecological source areas, N_a (N_b), p_a (p_b), S_a (S_b), respectively, represent the weights, resistance values, and areas of ecological source a (b), L_ab is the cumulative resistance value of the potential ecological corridor from the area a to b, and L_max is the maximum resistance value.

The ecological nodes discerned by the simulation results of ecological corridors serve for extracting the intersections between ecological corridors and those of the minimum cumulative cost distance of ecological sources to diffuse to the high-resistance surface [28].

3.2.4. Network Structure Analysis

The rationality of the ecological security pattern was verified by taking advantage of the network structure analysis. The formulas are as follows [29]:

$$\mathsf{\alpha }=\frac{\mathrm{L}-\mathrm{V}+1}{2\mathrm{V}-5}$$

(5)

$$\mathsf{\beta }=\frac{\mathrm{L}}{\mathrm{V}}$$

(6)

$$\mathsf{\gamma }=\frac{\mathrm{L}}{3\left(\mathrm{V}-2\right)}$$

(7)

$$\mathrm{Cos}\mathrm{tRotio}=1-\frac{\mathrm{L}}{\mathrm{C}}$$

(8)

As for the formulas, V indicates the number of ecological source areas, L and C correspondingly present the number and length of ecological corridors, α reflects the accessibility of the network, β reflects the connectivity proportion of each node, γ reflects that of all nodes, and CostRotio represents the relationship of input or output which implies that the smaller its value is, the more conducive to the construction of ecosystem security.

4. Results

4.1. Identification of the Importance of Ecosystem Services in the Urban Agglomeration

According to Figure 2, the indexes of ecosystem services are calculated to present the ecosystem values. In addition, in Figure 3, for the Lanxi urban agglomeration, the value of water sources conservation is between 0 and 0.653, the value of water and soil conservation is between 0 and 0.907, the value of biodiversity is between 0 and 0.299, and the value of windbreak and sand fixation is between 0 and 0.074. In line with the Standards for Ecological Redlines Delineation [30], this study employs the reclassification model in the GIS platform to calculate the values of the ecological services and discern the top 20% as the high-value areas. The high-value areas of water sources conservation are scattered around Shuangshimen, Weiyuan county, and Tongren and Haiyan county, with grasslands as the main type of land use. The grassland soil and grassroots hold the infiltration and storage of regional rainfall in the areas, protecting snow, delaying snow melting, regulating the surface runoff of snow water, and playing a role in water conservation. The distribution of high-value areas of water and soil conservation presents “more in the west and less in the east” which are mainly dotted in the 19 cities of Qinghai Province such as Minhe county, with cultivated land as the main type of land use, improving the ground coverage rate and soil erosion resistance, and promoting water and soil conservation. Meanwhile, the high-value areas of biodiversity depend on the Xinglong Mountain Nature Reserve, Dadunxia Resort, and Shihai scenic area, distributed in Minghe, Jishi, Yuzhong, and other counties, boasting various and multifarious ecosystems, plants, and animal resources that facilitate biodiversity protection. Furthermore, the high-value areas of windbreak and sand fixation are supported by Beishan National Forest Park, which is located in Huzhu county. At the same time, the natural forest areas form a barrier for maintaining the diversity of the forest landscape, which can reduce wind speed and fix quicksand in that the soil particles are immobilized by the roots of trees, shrubs, and grasses, preventing the process of desertification and then functioning as wind protection and sand fixation.

Furthermore, by employing the GIS platform, the study calculated and found the overlapping areas between the high-value regions and the ecological functional complexity was measured by the ratio of the overlapping areas to the high-value areas of ecosystem services, as shown in

Table 4. The proportions of areas with overlapping areas of high ecosystem service (%).

Indicators	Water Resource Conservation	Soil and Water Conservation	Windbreak and Sand Fixation	Biodiversity
Water resource conservation	100	0.320	0	2.526
Soil and water conservation	0.320	100	0.001	0.001
Windbreak and sand fixation	0.000	0.001	100	0
Biodiversity	2.526	0.001	0	100

. It can be seen that the ratio of biodiversity to water sources conservation is the largest, accounting for 2.526%, and the ratios of wind break and water sources conservation to biodiversity, account for 0.

4.2. Construction of Urban Agglomeration Ecological Network

4.2.1. Identifying Ecological Source Areas

The ecological source areas, as protected areas, protect the natural ecological space and maintain ecological security; meanwhile, all human development and construction activities are banned in such areas. The study area, as shown in

Table 5. The classification of ecological patch.

Category	Number	Number Proportion (%)	Area (km2)	Area Proportion (%)
Small patch	8185	95.419	917.020	8.438
Medium patch	201	2.343	324.148	2.983
Large patch	73	0.851	252.127	2.320
Oversized patch	119	1.387	9373.857	86.259

, the Lanxi urban agglomeration, incorporates 8578 multifunctional ecological patches with a total area of 10,867.152 km². Referring to the relative research [19], the ecological patches can be divided into four categories, namely, the small patch whose area is smaller than 1 km², the medium patch between 1 and 2 km², the large patch between 2.5 and 5 km², and the oversized patch more than 5 km², as listed in Table 5. The ecological patches are mainly composed of small patches, of which the number of the Lanxi urban agglomeration comes in about 95.419%. Such a result demonstrates the high fragmentation of the landscape of the ecological source in the study area.

The ecological patches collected in different cities are distinguished by the fragmental distribution, as shown in Figure 4. That is, in the ecological patches of the Lanxi urban agglomeration, the oversized patches are chiefly scattered in Guinan, Weiyuan, and Huzhu county, both the proportion of the number and area more than 10% which are the same as those of the medium patches in Ledu District. By the same token, the area with the oversized patches accounting for less than 2% is distributed in 12 counties or districts, including Anding, Chengbei, and Chengzhong District. In these districts, multiple patches with a high degree of landscape fragmentation are contiguously and spatially distributed. Therefore, it is urgent to reinforce the protection of patches with similar geographical characteristics and similar ecological attributes.

The landscape connectivity was calculated by formula 1 and 2 on Confor to measure the importance of the ecological source areas. Meanwhile, in order to ensure the integrity of source areas and the connectivity of ecological flow, the patch, in line with the comprehensive results, whose dPC is more than 0.2 and whose area is more than 5 km², was selected as the ecological source area. In Figure 5, there are a total of 46 ecological source areas in the Lanxi urban agglomeration, which are spatially “sparse in the east and dense in the west”. The total area is 8199.249 km², accounting for about 8.237% of that of the study area. Grounded on the three types of ecological source areas divided by natural breaking points in the GIS platform, the ecological source areas can be described as the primary source area, whose dPC is more than 15.929 (dPC > 15.92), the secondary source area, whose dPC is less than or equal to 15.929 but more than 5.697 (5.697 < dPC < 15.929), and the tertiary source areas, whose dPC is less than or equal to 5.697 (dPC > 5.697) (see

Table 6. The connectivity of ecological source landscape.

Serial Number	Area (km2)	Area Proportion (%)	dPC	Serial Number	Area (km2)	Area Proportion (%)	dPC
1	46.715	0.570	0.218	24	5.991	0.073	0.431
2	50.979	0.622	0.233	25	449.565	5.483	2.857
3	3.761	0.046	0.201	26	1036.200	12.638	30.689
4	7.912	0.096	0.201	27	473.678	5.777	32.907
5	115.299	1.406	0.812	28	7.971	0.097	0.525
6	57.304	0.699	0.278	29	289.733	3.534	0.902
7	64.794	0.790	0.504	30	83.893	1.023	0.670
8	7.771	0.095	0.360	31	184.486	2.250	0.453
9	661.733	8.071	13.750	32	54.374	0.663	0.413
10	8.173	0.100	0.410	33	845.912	10.317	5.698
11	234.594	2.861	1.477	34	183.454	2.237	0.452
12	11.063	0.135	1.111	35	8.883	0.108	0.433
13	7.115	0.087	0.491	36	184.867	2.255	1.083
14	11.313	0.138	1.119	37	154.181	1.880	0.691
15	34.552	0.421	1.346	38	8.614	0.105	0.412
16	54.849	0.669	1.471	39	13.394	0.163	0.529
17	29.656	0.362	1.254	40	861.640	10.509	15.930
18	5.165	0.063	0.411	41	16.651	0.203	0.609
19	4.213	0.051	0.351	42	21.699	0.265	0.732
20	768.890	9.378	41.970	43	12.985	0.158	0.519
21	1029.290	12.553	34.738	44	13.451	0.164	0.530
22	8.229	0.100	0.455	45	25.823	0.315	0.205
23	10.817	0.132	0.573	46	27.619	0.337	0.278

). As shown in Figure 6, there are 4 primary ecological source areas, including Areas 20, 21, 27, and 26. The first three source areas are all located in five counties, such as Datong County, etc., in the Hehuang Valley, with cultivated lands as the main type of land use, which are intensively used and prevent farmland. Area 26, with grasslands as the main type of land use, is in Huangyuan County and Haiyan County. The source areas of Huangyuan County form its ecological reserves by depending on cultivated lands of Hehuang Valley to develop its farmlands, while the source areas of Haiyan County form Jinshawan Nature Reserve by grasslands of Qinghai Lake. Furthermore, there are 2 secondary ecological source areas, situated in Weiyuan, Linzhao, and Longxi County of Gansu Province, involving area 40, and 9, with cultivated lands and forests as the main types of land use. Moreover, there are 40 tertiary ecological source areas. Areas 6 and 46, located in Yuzhong County, have Xinglong Mountain Nature Reserve, with grasslands and forests as the main types of land use. Areas 8, 35, 41, 42, 43, and 44, located in Weiyuan County, rely on Shuangshimen Natural Resort, boasting high habitat quality, with forests as the main type of land use. Areas 38 and 39 are in Longxi County, including Juetou Mountain Resort, and the main type of land use is grasslands. Area 45 is located in Yongdeng County, possessing Tulugou National Forest Park and Liancheng National Forest Ecological Nature Reserve, with forests as the main type of land use. Area 2 distributed in Huzhu Tu Autonomous County, has Huzhu Beishan National Forest Park and Zalonggou Resort, with grasslands and forests as the main types of land use. Areas 5, 30, and 36, distributed in Minhe County, include Huangcao Mountain, Gachang Mountain, and Hou Mountain, with cultivated lands and forests as the main types of land use. Areas 13, 15, 17, 18, 19, 22, 23, 24, 28, 29, 31, and 33 are distributed along the Hehuang Valley and located in seven counties, such as Hualong County, with cultivated lands as the main type of land use, playing a role of protecting food security. Area 1 is located in Guide County, having Guide Huangheqing National Wetland Park. Areas 11, 12, 14, 16, and 25 are located in Gonghe County, having Qinghai Lake Natural Resort and farmland reserves, with grasslands and cultivated lands as the main types of land use. Areas 3 and 4 are located in Jianzha County, including Cambra National Geopark, with grasslands as the main type of land use. Area 10, located in Datong County, possesses Mengda National Nature Reserve, with grasslands as the main type of land use. Area 34 is located in Tongren County and is abundant in grassland resources, with grasslands as the main type of land use. Area 7, located in the Jishishan and Xunhua County, includes Dadunxia Resort and Mengda National Nature Reserve, with forests and grasslands as the main types of land use. Areas 32 and 37 are scattered in Jishishan County, with cultivated lands and forests as the main types of land use. Still, there are a small number of artificial surfaces and bare lands in ecological source areas, mainly distributed in Guide County, Ping’an District, Tongren County, and other regions, which urgently need ecological restorations.

As shown in Figure 7, Guinan and Weiyuan counties are the cities whose ecological source area accounts for more than 10% of the total ecological source areas. Meanwhile, the cities whose proportion is less than 1% are distributed in seven cities including Guide County, Jianzha County, the Ping’an district, etc. The cities whose ecological source area accounts for more than 20% of the total administrative area are Weiyuan County, Huangzhong District, and Jishishan County. Among them, Weiyuan County accounts for the highest proportion of the total administrative area, more than 50%. Cities with ecological source areas accounting for less than 1% of the total administrative area are distributed in four cities, including Jianzha County, Qilihe District, Xunhua County, etc.

4.2.2. Constructing of Ecological Resistance Surfaces

This study employs ESDA to explore the spatial distribution of ecological resistance in the study area. By conducting binary local autocorrelation analysis on Geoda software, the ecological resistance of the spatial correlation between landscape pattern index and dPC was obtained, and the figure of the LISA distribution was acquired, so as to distinguish the high-high cluster and the low-low cluster with statistics significant (p = 0.05) and establish a grid of total resistance values and make the results into a grid diagram with an accuracy of 500 m × 500 m. Moreover, the results show that the maximum value of the ecological resistance factor is 4.123, while the minimum value is 1.07 (see Figure 8). According to the distributions of the ecological resistance factors, the resistance values decrease from the periphery to the center. The distribution of the low-value areas of the resistance surface is consistent with that of the ecological source areas, which indicates that in the area, the closer to the ecological source, the more complete the ecological function, the higher the service value, and the smaller the resistance value, and the ecosystem service flows move in multiple directions. The high-value areas of resistance are mainly concentrated in urban areas such as Chengguan District. The resistance values are rising under the control of human activities because of the dense road network and high economic development level.

4.2.3. Extracting and Determining the Ecological Corridors and Ecological Nodes

The ecological corridors, as the pivotal characters in improving the ecological service value of multiple ecological source areas, ensure interactions and diffusions of the ecological source areas. Grounded on the ecological resistance, the study targeted the minimum resistance value of each ecological source area and then extracted 914 ecological corridors with a total length of 62,970.181 km. Through the ecological source matrix conducted by the gravity model, the study further measured the interaction strength of 46 numbered ecological corridors and thereby set the interaction value of 3878.001 as a threshold to determine the importance of ecological corridors. In this sense, they can be divided into 17 important ecological corridors and 897 general ecological corridors with the former placed in priority (see Figure 9). As for the spatial distribution of the corridors, they are dense in the central and southeast and sparse in the west and south. The important ecological corridors spatially situated around Areas 3, 5, 6, 7, 9, 13 15, 17, 18, 22, 24, 30, 32, 38, and 40 into a gathering center composed of 12 cities such as Weiyuan County in 6 clusters. Similarly, the general corridors also formed 6 clusters in series, connecting the ecological source areas in the west and south. In line with the simulation results of the ecological corridors, 62 ecological nodes were determined.

4.3. Restoration of the Ecological Pattern of Urban Agglomerations

Regarding the concept of landscape ecology, namely, “patches–corridors–matrixes”, the ambitions of the study were to superimpose the components of the ecological landscape and then to provide relative measures to restore the ecosystem security patterns in the urban agglomerations, relying on the core elements encompassing mountains, forests, farmlands, grasslands, and lakes.

4.3.1. Replenishing Ecological Source Sites

It is urgent to concentrate on the comprehensive management of river basins to reinforce ecological protection and governance to maintain the ecological security of the Lanxi urban agglomeration. Therefore, this study, in terms of the distributions in the Lanxi urban agglomeration, applied the similarity search module of the GIS platform to connect the fragmented patches with the largest similarity index to large landscape patches as a supplementary ecological source to improve the overall landscape connectivity of the study area.

From the spatial distributions of the ecosystem security patterns, the ecological source areas are mainly assembled in the west part of Qinghai Province and the south part of Gansu Province, while the other areas lack corridor connectivity in the Lanxi urban agglomeration. As shown in Figure 10, the 10 new ecological source areas effectively fill the gap in the northeast of the Lanxi urban agglomeration with a total area of 441.471 km². As for the 10 new source areas, the new Area 1 is located in Tongren County, providing the grass ecosystems and playing a role in water conservation; the new Areas 2, 3, and 4 are in Yuzhong County and Lintao County to support the construction of woodlands and grasslands such as the Xinglong Mountain Nature Reserve, Shifogou National Forest Park, and Guantango Resort, as an important part of the forest ecosystem. The new Area 5, located in Anding District, is to replenish cultivated lands construction and to promote the ecological protection of basic farmland, boasting the protection of the permanent basic farmland, and the construction of a typical hill and gully area of the Loess Plateau. The new Area 6 is situated in Jingtai County, supplementing Shulu Mountain Nature Reserve and strengthening the construction of primeval forests, and playing a role of water conservation, climate regulation, and air purification. The new source Area 7 is located in Pingchuan District, relying on the construction of Wushan Farm, strengthening the restoration of ecological vegetation, and serving for returning farmland to forests. The new Area 8 is in Jingyuan County, as a supplementary of Harth Hill Nature Reserve, with the capacities of constructing water conservation forests, conserving and maintaining the water resources, and water and soil, and safeguarding the clean water sources in the Yellow River Basin; The new Area 9 is located in Ledu District, as a supplementary of promoting forests and grasslands construction, and then improving the ecosystem values; The new Area 10, situated in Yongdeng County, is replenishing cultivated land construction to establish a cultivated land system with a high and stable yield of crops. Accordingly, on the whole, there are 56 ecological source areas in the restoration proposition with a total area of 8640.72 km².

4.3.2. Restoring Ecological Corridors

In order to improve the circulation efficiency of ecological corridors, this study restores ecological corridors on the basis of increasing ecological source areas. As shown in Figure 11, picture A represents a potential ecological network, including 56 ecological source areas, with multiple reachable overlapping roads including 1467 ecological corridors. Therefore, based on the network typology [31], an ecological network improvement scheme was generated. The principle of selecting ecological corridors is to preferentially use the areas with high values in dPC and mutual gravity value. The only corridors with relatively high gravity values will be chosen, while there are multiple parallel corridors around adjacent distance areas in the potential ecological network. Meanwhile, the interconnected corridor will be selected, while two corridors are similar in ecological values. Moreover, for each source area, there is at least one corridor connecting to another area. In terms of ecological network B, each ecological source area only crosses once. In order to restore the ecological corridors of the Lanxi urban agglomeration, 12 outer circle corridors, and 15 inner circle corridors were added to generate ecological network C, which finally formed 82 ecological corridors and 80 ecological nodes, with a total length of 2465.208 km. Therefore, according to the network structure analysis method, the study calculated the rationality of the ecosystem security patterns connection in the restoration proposition, and the results demonstrated that the values of α, β, and γ increased by 0.252, 0.167, and 0.482, respectively, and CostRotio decreased by 0.0004, which improved the rationality of the ecological corridor.

4.3.3. Determining the Ecological Security Levels

In this study, the ecosystem safety level was divided into three types by employing the Reclassify module of the GIS platform in line with the analysis method of the natural breaking points, namely, ecological buffer zone, ecological sensitive zone, and ecological restoration zone, with areas of 56,732.429 km², 21,621.806 km², and 10,991.317 km², respectively, and correspondingly accounting for about 63.50%, 24.20%, and 12.30% of the total area of the study area (see Figure 10). The ecological buffer zones are close to the ecological source areas, mainly distributed in Minhe County and other areas, which block the ecological damage caused by human activities and are backup space resources for protecting the ecological source areas. The ecological sensitive zones are mainly located in Haiyan, Gonghe County, etc., as the links between the ecological source areas and human activity regions. Moreover, they are transitions between the areas with high resistance values and the areas with low resistance values, being ideal for carrying out the farmland regulation, farmland returning to woodland, and grassland restoration. As for the ecological restoration zones, the areas are relatively low levels of ecological service and with dense populations, mainly distributed in Chengguan District, etc. They are, however, the areas with a prominent contradiction between human activity development and ecological environment restoration, and the restoration propositions they are to have the construction of urban green parks, the treatment of polluted water bodies, and the greening and restoration of mining mountains to improve the urban ecological environment.

4.3.4. Providing Restoration Measures

Based on the results of exploring the ecological source areas, ecological corridors, and ecological security levels in the restoration, this study intends to provide the restoration measures for the ecological system security in the urban agglomeration of the Yellow River Basin and as the reference to the local ecological environment bureau and other government departments to formulate standards, regulations, and policies. Firstly, the ecological forest and grass protection areas should be formed around Dadunxia Resort, Mengda National Nature Reserve in Xunhua County and Jishishan County, and Huangcao Mountain in Minhe County, by carrying out restoration projects such as planting grass on degraded grassland, greening barren hills and slopes, and recovering natural vegetation. We should also promote the protection of natural forest and grass resources and build forest and grass protective systems. Secondly, it is necessary to form ecological wetland conservation areas and implement restoration projects such as biodiversity protection, ecological landscape design, microclimate regulation, bird monitoring stations, leisure facilities construction, etc., by utilizing Guide Huangheqing National Wetland Park and Qinghai Lake to form a wetland ecological system. Furthermore, ecological agricultural development zones should be constructed by taking advantage of cultivated lands of Anding District, Longxi County, and Lintao County. The cultivated lands resources consolidation and restoration measures should be carried out as restoration measures, such as establishing high-standard farmlands with drought and flood prevention. Meanwhile, by utilizing new modern agricultural technologies, the production capacity of cultivated lands will be greatly improved to promote green and efficient and cold and drought modern agriculture. Fourthly, ecological security barriers should be formed, by concentrating on the natural ecological areas. These areas should adhere to the principle of giving priority to environmental protection and focusing on natural recovery, serving for promoting the process of ecological protection and restoration projects for mountains, forests, fields, and lakes. Finally, it should connect the fragmented biological habitats and build spatial connecting channels for biological species to live, move, and diffuse around the ecological corridors under repair, which is conducive to strengthening the interactions and flows of species in ecological spaces, functions, and elements.

5. Conclusions and Discussion

5.1. Conclusions

This study, on the basis of ecosystem services and landscape connectivity, conducted an experiment to discern the ecological source areas and identified ecological corridors and ecological nodes by employing an optimization evaluation method of resistance surface and then proposed the restoring measures of ecological patterns. The results are as follows:

(1)

In terms of ecosystem service values, there are 10,867.152 km² of multifunctional ecological patches, presenting the characteristics of highly fragmented landscapes. Such patches are mainly distributed along Beishan National Forest Park, Xinglong Mountain Nature Reserve, Dadunxia Resort, and other areas. Among them, the proportion of overlapping patches of biodiversity to water sources conservation is the largest, while the proportion of wind break and water sources conservation and biodiversity account for 0.

(2)

The study area contains 46 ecological source areas with a total area of 8199.249 km², with a spatial distribution of “sparse in the east and dense in the west” and boasting high coverage of cultivated lands, grasslands, and forest lands, which are concentrated in Datong County, Huangyuan County, Haiyan County, etc. By the same token, there are 914 ecological corridors with a total length of 62,970.181 km. Most of them are short-distance corridors, which are relatively densely located in the northwest part of the study area with a webbed shape. Moreover, the important ecological corridors form six small clusters in space, while the general ecological corridors connect the clusters and other areas, accelerating the diffusion of ecological flows.

(3)

This study proposes restoration measures to restore the ecological system security pattern and integrate the fragmented ecological patches to form 10 new ecological source areas with nature reserves, forest parks, and farmland as the main ecological elements. Furthermore, this study calculates the rationality of ecosystem security patterns in restoration by the network structure analysis method, which can be employed to restore ecological corridors and construct ecological security barrier areas, ecological forests, grass protection areas, ecological wetland conservation areas, and ecological agriculture development areas. Meanwhile, with the regional management, precise measures should be taken to promote environmental protection and improve the ecological system security patterns of the Lanxi urban agglomeration.

5.2. Discussion

By constructing the ecosystem security patterns of the urban agglomeration in the basin, the study provides a method suitable for the administrative scale of the urban agglomeration, which is conducive to the effective allocation of ecological resources, enhancing the internal connectivity of the landscape, maintaining the ecological value, and ensuring the ecological functions required for ecological security. Regarding the selection of ecological source areas, the study discerned the source areas on the basis of the importance of ecological services, comprehensively considering the influences from the natural environment, economy, and society. The method employed in the study, compared with the method of directly selecting the fixed patch of the nature reserve as the ecological source, avoids the inadequacy of the single factor evaluation, for it preferably offers the reflections of the overall benefits of the ecosystem with a certain reference value. The results are basically consistent with the ecosystem security patterns of the Lanxi urban agglomeration. In terms of the construction of resistance surface, NPP, NPP-VIIRS, population density, and other data are introduced, which not only take the natural environment factors into consideration but also consider the impact of human activity intensity, providing a new treatment for the future research on the ecosystem security patterns. Based on expert consultation and field investigation, this study also proposes the restoration measures of ecological security patterns, scientifically guiding the spatial patterns restoration to a certain extent, solving the conflict between development and protection, and realizing sustainable ecological environment governance. It is a long-term project to govern the environment and ecosystem of the Yellow River Basin, and it has also turned out to be an important driving force for high-quality development under the new situation. Therefore, further research needs to consider the time factors, and it is indispensable to compare the ecosystem security patterns in different years and to have more accurate results by finding the discrepancies. Secondly, the ecological corridors selected in this study are a linear ecological element, which measures the connectivity strength without referring to the distribution of ecological sources, the ecological service value of ecological patches, and the impact of biological species on themselves. Hence, determining how a reasonable and feasible ecological corridor width is the focus of future research, which provides guidance for the accurate delineation of the ecosystem security patterns.