Study on the Construction of the Ecological Security Pattern of the Lancang River Basin (Yunnan Section) Based on InVEST-MSPA-Circuit Theory

Abstract

Optimizing the pattern of land space development and protection as well as improving regional ecological security are significant tasks in the construction of an ecological civilization. The Lancang River Basin is a significant ecological barrier in Southwest China, and its ecological security is critical to the Natural World Heritage Site of “Three Parallel Rivers” and even to Southeast Asia. In this study, the InVEST model and MSPA method were used for the identification of the ecological sources; the resistance surface was corrected by the night light coefficient; and the ecological linkages, pinch points, obstacles, and ecological breakpoints were extracted using the circuit theory to construct the ecological security pattern of the Lancang River Basin. The major study results indicated that: (1) A total of 23 ecological sources were identified in the watershed, covering an area of 9019.56 km², mostly large-scale irregular patches. (2) A total of 39 ecological corridors were identified in the study area; 15 key corridors, 17 important corridors, and 7 general corridors were extracted based on the gravity model, which together formed a ecological security network for the watershed. (3) Sixteen ecological pinch points, one primary ecological improvement area, twelve secondary improvement areas, and twenty-nine ecological breakpoints were identified in this study. The ecological protection and restoration of different important ecological areas are conducive to the protection of biodiversity and the construction of the ecological security pattern in the Lancang River Basin.

1. Introduction

Optimizing the patterns of land space development and protection, as well as enhancing regional ecological security, are important tasks in ecological civilization construction [1,2]. The 15th meeting of the Conference of the Parties to the Convention on Biological Diversity (CBD COP15) was held in Yunnan in October 2021 to promote the conservation of biodiversity and the sustainable utilization of its components. In May 2022, the Yunnan provincial government released the document “The Pioneer Plan of Ecological Civilization Construction in Yunnan Province (2021–2025)”, which proposed the construction of the southwest ecological security barrier. Yunnan Province is in a critical phase of social and economic development, and the construction of ecological civilization can alleviate the dual pressures of economic development and ecological protection [3,4]. Constructing an ecological security pattern (ESP) is an important strategy to guarantee regional ecological security [2]. Effective connection of fragmented habitat patches by building an ecological network can improve the quality of the ecosystem and conserve biodiversity [5,6].

Many scholars have conducted a lot of research on the assessment of habitat quality. Early research on habitat quality mainly obtained relevant parameter indicators through field surveys of single or a few types of habitats. This method has advantages in terms of accuracy, but field surveys are time-consuming and labor-intensive and are only suitable for small-scale sample point surveys, which are difficult to achieve at the watershed and regional scales and cannot effectively obtain long-term monitoring [7]. With the development of geographic information technology, the InVEST model, based on remote sensing information, has been widely used in habitat quality assessment, which provides an opportunity for watershed- and regional-scale habitat quality assessment and ecological source identification [8]. For example, Wang Lirong and Hao Yue used the InVEST model to assess habitat quality and determine ecological sources based on land use types [9,10]. In summary, the InVEST model has high accuracy and convenience in evaluating habitat quality at different scales and can comprehensively reflect the impact of land use change on habitat quality.

MSPA is an image processing method to measure, identify, and segment the spatial patterns of raster images based on mathematical morphology principles such as erosion, expansion, opening operation, and closing operation. It can more accurately distinguish the type and structure of the landscape [11]. Its working principle is to divide the binary raster data into foreground and background, and then use a series of image processing methods to divide the foreground into seven non-overlapping categories according to morphology, so as to identify the landscape types that are important for maintaining connectivity. The core area is the larger habitat patch in the foreground pixel. It can provide a larger habitat for species, which is of great significance to the protection of biodiversity, and increase the scientific selection of ecological sources [12].

At present, the construction of ESPs is a well-defined research topic, and ESP construction is currently carried out through the basic paradigm of “identifying ecological sources-constructing resistance surfaces-extracting corridors” [2,13]. Some researchers have also considered “ecological pinch points” and “ecological breakpoints” for the construction of ESPs [14,15,16]. There are mainly two kinds of identification of ecological sources in existing studies: one is to directly take scenic forest or nature reserves as ecological sources; the second is to identify ecological sources mainly from the quality of a single habitat quality or landscape connectivity. However, comprehensive consideration of habitat quality and landscape structure plays a pivotal role in the effective identification of ecological sources. To connect ecological sources, an ecological network is formed by the cross-connection of nodes and corridors. The energy, material, and information flows among different landscape elements are carried out through the network [2]. The minimum cumulative resistance model is widely used for the extraction of ecological corridors. This model can instantaneously identify and map the optimal path of ecological flow but cannot identify the scope and key nodes of the corridors. McRae integrated circuit theory into landscape ecology and predicted the uncertainty of species flows with the help of the random walk property of currents across a surface providing resistance [17]. At present, circuit theory is mostly used in studies on ecological corridor extraction and ESP because it can quantify the movement in the corridor system and estimate the relative importance of nodes using the current intensity between sources. Moreover, it can identify the specific locations of key nodes and the width of the ecological corridors, thereby providing a scientific basis for the construction of ecological corridors and their habitat suitability [18]. For example, scholar Zhou Lang constructed the ecological corridor of a mountain city to ensure its ecological sustainable development [19]. Li Jiulin used circuit theory to identify the ecological network and important areas of Anqing City and put forward targeted ecological restoration strategies [20]. Graph theory, gravity models, ant colony models, centrality mappers, and other models are used to evaluate the importance of ecological corridors. Most studies on the construction of ESPs are based on administrative units [21,22], which makes it difficult to maintain a completely natural geographical pattern, resulting in an incomplete and one-sided ESP [2].

In their study of ecological security patterns, scholars Zhang Mufeng and Zheng Qunming directly took nature reserves and national parks as ecological sources to construct ecological networks, which is highly subjective [23,24]. On the research scale, it mainly focuses on the province, city, and county. There is little research on the ecological security patterns of river basins. These studies are primarily focused on the Yangtze River [25] and the Yellow River Basin [26], and not much attention is paid to the river basins in Southwest China.

A watershed is a typical complex ecosystem that forms an integral part of natural ecosystems. Taking the river basin as the research area is more in line with the distribution characteristics of natural geographical ecological patterns and avoids the incompleteness of natural geographical patterns caused by taking the administrative unit as the research unit [2]. The Lancang River Basin contains several nature reserves, which are essential for biodiversity protection, soil and water conservation, and other functions. In this study, the Lancang River Basin (Yunnan section) was considered as the research area of interest, and the InVEST model and morphological spatial pattern analysis (MSPA) were used to identify the ecological sources. The analytic hierarchy process (AHP) was used to comprehensively consider the natural environment, human society, and the patterns of landscape elements to construct the resistance surface, and the night light data were used to correct it. The ecological corridor and key nodes were extracted using circuit theory, and the importance of the potential corridor was quantified by a gravity model. The research results provide a scientific basis for ecological protection and related planning of the Lancang River Basin (Yunnan section). This method can provide a reference for the construction of regional ESPs.

2. Overview and Data Sources of the Research Area

2.1. Overview of the Research Area

The Lancang River Basin, located at 94–102° E and 21–34° N, is the longest north-south river in China, with a total length of 2162 km. Generally, the terrain of the basin is high in the northwest and low in the southeast and distributed in bands from north to south. The major stream of the Lancang River in Yunnan Province is 1247 km long. The basin has diverse climate types, numerous nature reserves, and abundant animal and plant resources [27], all of which play essential roles in the construction of ecological security barriers in Southwest China. The Lancang River in China is known as the Mekong River outside China and is an important link between China and Southeast Asia.

2.2. Data Sources

The land use data for 2020 was obtained from the Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences. The data on roads and residential areas was obtained from the national basic geographic information database of the national geographic information center at a scale of 1:1 million (https://www.webmap.cn/commres.do?method=result100W, accessed on 13 October 2022). The spatial distribution data of nature reserves were obtained from the National Earth System Science Data Center (http://www.geodata.cn/, accessed on 26 April 2022). The elevation data were obtained from EARTHDATA (https://search.asf.alaska.edu/#/, accessed on 26 April 2022). The night light data were obtained from the National Qinghai-Tibet Plateau Scientific Data Center (http://data.tpdc.ac.cn, accessed on 13 October 2022), and the water system data was obtained from the Data Registration and Publishing System of Resources and Environmental Sciences (http://www.resdc.cn/DOI, accessed on 19 October 2022).

3. Research Methods

3.1. Research Framework

As shown in Figure 1, the step-by-step framework for the construction of the ecological network in this study is as follows: (1) The target species were determined, followed by extraction of the land use types in the study area. Furthermore, the data were analyzed based on MSPA using InVEST model and Guidos 3.1 software (European Commission Joint Research Centre (JRC), Brussels, Belgium, https://forest.jrc.ec.europa.eu/en/activities/lpa/gtb/, accessed on 9 October 2022) to extract 7 model classes, determine a certain amount of core patches at the source area, and analyze their connectivity to determine the final ecological source. (2) The natural, social, and landscape pattern elements were comprehensively to construct the resistance surface, and the night light data were applied for correction. (3) The potential ecological corridor was first extracted according to circuit theory. Then, the cumulative current density values were mapped and the pinch points, obstacle points, and ecological fault points were identified based on the potential ecological corridor. (4) The gravity model was used to evaluate the importance of ecological corridors and construct an ecological safety network to provide a reference for land spatial planning, the construction of the ESP, and the ecological restoration of the Lancang River Basin (Yunnan section).

3.2. Target Species Selection

The target species used to design the habitat network should have certain regional mobility, observability, and high habitat quality, and improvement in their habitat should meet all the habitat conditions the species along the food chain need to survive. Therefore, birds, fish, amphibians, reptiles, and small mammals can be selected as target species [28,29]. In this study, birds were selected as the target species because of their high mobility, high dispersal ability, and ability to simulate a larger network. It is important to note that birds are often used as indicators for monitoring biodiversity [30] and habitat restoration [31], thereby making ecological networks more universal.

3.3. Methods

3.3.1. Identification of the Ecological Sources

Evaluation of Habitat Quality Based on the InVEST Model

Habitat quality highlights the quality of natural resources that regional ecosystems can provide for the survival and continuation of animals and plants in the ecosystem [26], and the biodiversity functions are evaluated according to the quality of habitats [32,33]. The habitat quality module of the InVEST model was used to quantitatively evaluate the river habitat quality, and the plate functional attribute was used to guide the source identification. This study combined InVEST 3.12.0 (The Natural Capital Project, Palo Alto, CA, USA, https://naturalcapitalproject.stanford.edu/, accessed on 30 September 2022) and the MSPA for the identification of ecological sources. The model parameters were set based on the InVEST model user manual, relevant literature, and the present situation of the study area. The Equation for calculating the habitat quality index in the model is provided below:

$$Q_{xj}=H_j\left[1-\left(\frac{D_{xj}^z}{D_{xj}^z+k^z}\right)\right]$$

(1)

where Q_xj is the habitat quality index of grid x in j land type; H_j is the habitat suitability of j land type, falling within the range of 0–1; k is a semi-saturation constant, which is considered as 0.5; and z is the default parameter of the model. D_xj is the degree of habitat degradation of grid x in j land type, which indicates the degree of habitat degradation after stress. D_xj is given by [31]:

$$D_{xj}={\displaystyle \sum }_{r=1}^R{\displaystyle \sum }_{y=1}^{Y_r}\left({\omega }_r/{\displaystyle \sum }_{r=1}^R{\omega }_r\right)r_y{i}_{rxy}{\beta }_x{S}_{jr}$$

(2)

where R is the number of stress factors; Y_r represents the total number of grids of stress factors; ω_r is the weight; r_y is the number of stress factors on the grid unit; β_x represents the accessibility level of grid x; S_jr represents the sensitivity of land type j to stress factors and has a value between 0 and 1; and i_rxy is the influence distance of stress factors, which is calculated according to the linear and exponential decline models, as given below:

$$i_{rxy}=1-\left(\frac{d_{xy}}{d_{rmax}}\right)$$

(3)

$$i_{rxy}=\exp \left(-\left(\frac{2.99}{d_{rmax}}\right)d_{xy}\right)$$

(4)

where d_xy is the direct linear distance between grid x and y and d_rmax represents the maximum value of threat factor R.

The operation of the model requires three steps: obtaining habitat stress source data, determining the habitat sensitivity matrix, and generating habitat quality maps. Habitat stress source data include determining threat source, maximum threat distance, weight, and spatial attenuation type. Then the habitat type and ecological sensitivity are determined according to the land use type to form a habitat sensitivity matrix. Finally, the above relevant data are input into the InVEST model for calculation to obtain the habitat quality results of the study area. The major input parameters to the model include land use maps, habitat threat factors, threat factor weights, influence distances, and the relative sensitivity of land use types to threats [26,29,34,35,36,37]. The model parameters in this study were set according to the InVEST mode user manual, related literature, and the present situation of the study area (

Table 1. Attributes of threat factor.

Threat Factors	Maximum Influence	Weight	Attenuation Mode
Drylands	8	0.7	Linear
Paddy fields	8	0.7	Linear
Rural residential areas	5	0.6	Index
Urban land uses	10	1	Index
Other construction lands	5	0.5	Index

and

Table 2. Habitat suitability and its sensitivity to threat factors.

Major Land Types	Ground Class Code	Land-Use Types	Habitat Suitability	Paddy Fields	Drylands	Rural Settlements	Urban Lands	Other Construction Lands
Plough	11	Paddy field	0.4	0.3	0.3	0.6	0.7	0.5
	12	Dryland	0.3	0.3	0.3	0.6	0.6	0.4
Woodland	21	Forested land	1	0.6	0.6	0.8	0.9	0.8
	22	Spinney	0.8	0.6	0.6	0.6	0.8	0.7
	23	Sparse woodland	0.7	0.5	0.5	0.7	0.7	0.4
	24	Other woodlands	0.6	0.5	0.5	0.7	0.7	0.8
Lawn	31	High coverage grassland	0.7	0.6	0.6	0.7	0.7	0.7
	32	Medium coverage grassland	0.6	0.5	0.5	0.7	0.7	0.7
	33	Low coverage grassland	0.5	0.5	0.5	0.7	0.7	0.7
Water	41	Rivers and canals	1	0.6	0.6	0.7	0.9	0.8
	42	Lake	1	0.6	0.6	0.7	0.9	0.8
	43	Reservoir pit pond	0.8	0.5	0.5	0.6	0.9	0.8
	44	Permanent glacier and snow	0	0	0	0	0	0
	46	Beachland	0.5	0.5	0.5	0.6	0.9	0.8
Construction lands	51	Urban land	0	0	0	0	0	0.2
	52	Rural settlements	0	0	0	0	0	0.7
	53	Other construction lands	0	0	0	0.6	0.7	0
Unused land	64	Marshland	0.5	0.4	0.4	0.4	0.4	0.3
	65	Barren earth	0	0	0	0	0	0
	66	Bare rock texture	0	0	0	0	0	0

).

Ecological Source Identification Based on MSPA and Landscape Connectivity Index

MSPA is a method that emphasizes structural connectivity which can identify seven landscape types that are important for maintaining connectivity at the pixel level and is less affected by the spatial scale. Based on the habitat quality result generated by the InVEST model, we used the natural breaks method to select a high-habitat-quality area. A habitat quality index greater than 0.8 was selected as the alternative ecological source and the basis for MSPA analysis. The data obtained were converted into TIFF format binary raster files, and seven non-overlapping landscape types were obtained by eight neighborhood analysis methods using the Guidos 3.1 software (European Commission Joint Research Centre (JRC), Brussels, Belgium, https://forest.jrc.ec.europa.eu/en/activities/lpa/gtb/, accessed on 9 October 2022). The core area can provide habitats for species, thereby playing a crucial role in the protection of biodiversity. The study of Hinsley and Uezu pointed out that the minimum forest area satisfying bird breeding requirements is 25 hm², and only habitats with an area of no less than 60 hm² can ensure the stable existence of species [29]. Larger areas with better habitat quality are more suitable as ecological sources. Therefore, a core area with more than 100 km² was extracted for the landscape connectivity analysis.

The level of landscape connectivity can quantitatively characterize whether a particular landscape type is conducive to species migration within the source patches, and the probability of connectivity (PC, Equation (5)) is widely used for the evaluation of landscape connectivity [38,39]. Based on the previous literature, the landscape connectivity index (d_pc) of the core area was calculated by selecting the landscape PC using Conefor 2.6 software (Developed by Santiago Saura and Josep Tornéat at Universidad Politécnica de Madrid, Madrid, Spain, http://www.conefor.org, accessed on 22 October 2022), and the patches of the core area with d_pc > 0.5 were selected as the final ecological sources:

$$d_{pc}=\frac{{{\displaystyle \sum }}_{i=1}^n{{\displaystyle \sum }}_{j=1}^n{P}_{ij}^{\ast }·a_i·a_j}{A_L^2}$$

(5)

where d_pc is the landscape connectivity index; A_L is the total area of the regional landscape; n represents the total number of patches in the landscape; a_i and a_j are the areas of patches i and j, respectively; and P_ij* is the maximum probability of species spreading directly in patches i and j.

3.3.2. Construction of the Resistance Surfaces

In this study, resistance surface refers to different landscape types assigned relatively low or high resistance values based on whether they promote species migration and diffusion [13]. As indicated in the previous studies [40,41,42], the present study adopted AHP and comprehensively considered the natural environmental, social, and landscape pattern elements for constructing the resistance surface (

Table 3. Ecological resistance factors and weights.

Level	Evaluation Factor	Grading Criteria	Drag Coefficient	Weight
Natural environmental factors	Slope	0–5	1	0.0834
		5–15	100
		15–25	200
		25–35	300
		>35	500
	Elevation	<1500	1	0.0357
		1500–2000	100
		2000–2500	200
		2500–3000	300
		>3000	500
	Distance from the water system	0–500	1	0.0225
		500–1500	100
		1500–2500	200
		2500–3500	300
		>3500	500
Social and economic elements	Distance from road	0–500	800	0.0834
		500–1000	500
		1000–1500	300
		1500–2000	200
		>2000	1
	Distance from the residential area	0–1000	1,000	0.2503
		1000–2000	500
		2000–3000	200
		3000–4000	100
		>4000	1
Ecological pattern elements	Land use type	Plough	100	0.3498
		Woodland	1
		Lawn	100
		Waters	10
		Land for construction	1000
		Unused land	300
	Habitat quality index	0	1000	0.1749
		0–0.4	800
		0.4–0.7	500
		0.7–0.8	200
		0.8–1	1

). As human interference is one of the important resistances to species migration and diffusion, this study selected the night light index to correct the ecological resistance coefficient [40], as given by Equation [43] below:

$$R^{\prime }=\frac{TL{I}_i}{TL{I}_a}\times R$$

(6)

where R’ is the ecological resistance coefficient of the grid after correction based on the night light index; TLI_i represents the night light coefficient of grid i; TLI_a is the average night light value of grid i; and R is the basic resistance value of grid i.

3.3.3. Extraction and Evaluation of the Ecological Network

Circuit theory links the “random walk theory” with behavioral ecology in the connectivity evaluation model, and this makes it effective in the prediction of the path of the random migration of species and population diffusion probability [38]. The connectivity evaluation model is expressed as follows:

$$I=\frac{V}{R}$$

(7)

where I is the current, V is the voltage, and R is the effective resistance. The corresponding landscape ecological meaning [38] is presented in

Table 4. Circuit theory and corresponding ecological terms and ecological meanings.

Circuit Factors	Corresponding Ecological Terms	Ecological Meaning
Power Supply	Ecological source	Habitat in the region with high quality and suitable for species survival
Electric current	Source connectivity	The net number of times the specified species (wanderers) in the habitat leave the nest area for evacuation and migration and pass through the landscape nodes before reaching the target nest area is proportional to the probability of net migration through the corresponding landscape nodes; a high-current-density area forms a “radiation path” to a certain extent
Electric conductance	Habitat suitability	Habitat suitability refers to the capacity of the habitat and is directly proportional to the diffusion ability of the species
Electric resistance	Landscape resistance	The impedance encountered by species migration or diffusion in habitats is directly proportional to the challenges in migration
Voltage	Source connectivity probability	The net migration probability of any species (random walkers) in the habitat leaving the nest for migration and successfully reaching the designated landscape node or nest source is the basis for the selection of the possible connectivity index

.

The ecological network of the Lancang River Basin (Yunnan section) was constructed based on the circuit theory using Circuitscape software to obtain the current density map representing the degree of landscape connectivity. The Linkage Pathways Tool of the Linkage Mapper Toolbox (The Nature Conservancy, Arlington, VA, USA, https://linkagemapper.org/, accessed on 14 October 2022) was used to generate a low-resistance ecological corridor with an ecological source directly transferring ecological flows and energy flows, thereby forming a potential ecological network. The gravity model can reflect the interaction intensity between ecological sources [43]. Therefore, the gravity model was used to quantify the potential value of the ecological corridor to the ecological network and reduce the ecological resistance to species migration and diffusion. The Equation for the interaction intensity is given as follows:

$$G_{ab}=\frac{N_a{N}_b}{D_{ab}^2}=\frac{L_{max}^2\ln S_a\ln S_b}{L_{ab}^2{P}_a{P}_b}$$

(8)

where G_ab is the direct interaction intensity between the ecological sources a and b (the importance of ecological corridor); N_a and N_b represent the weight values of sources a and b, respectively; D_ab is the standardized value of potential corridor resistance between sources a and b; P_a and P_b represent the average resistance values of sources a and b, respectively; S_a and S_b are the areas of sources a and b, respectively; L_ab represents the corridor resistance value between sources a and b; and L_max is the maximum value of the minimum cumulative resistance in the area.

3.3.4. Identification of Important Nodes of the Ecological Network

The current density maps generated by Circuitscape 4.0.5 software (Brad McRae, https://circuitscape.org/, accessed on 14 October 2022) indicated the possibility and selection frequency of species passing through this area in landscape ecology. A higher current density value indicates a greater probability of species passing along this area. Therefore, current density maps were used to identify important areas in the ecological network [38]. Using the Pinchpoint Mapper function of the Linkage Mapper Toolbox (The Nature Conservancy, Arlington, VA, USA, https://linkagemapper.org/, accessed on 14 October 2022), each source node was given a current value of 1 A, the cost-weighted distance of the corridor was set to its default value, and, finally, the “all to one” mode was selected for iterative calculations. The plaques with high current density and strong irreplaceability located in corridor bottleneck points and narrow points were identified as “pinch points”. An obstacle indicates the area with high resistance to the movement of species between the source areas. The Barrier Mapper tool was used to determine the barriers affecting the quality of the existing corridors. The improved scoring model was used for the identification of obstacles and calculation of the recovery value of the accumulated current after obstacle removal. A larger recovery value indicates a significant improvement in the landscape connectivity after obstacle removal or repair. This model was primarily used to identify the areas affecting the ecological flow. From the ecological point of view, the existing traffic lines in the study area greatly influenced or hindered the migration and flow of species. In this study, the spatial superposition analysis of railways, national highways, provincial highways, and ecological corridors was performed to identify the ecological breakpoints in the watershed ecological network for ecological restoration.

4. Results and Analysis

4.1. Construction of the Ecological Security Pattern of the River Basin

4.1.1. Identification of the Ecological Sources

Figure 2 shows that the Lancang River Basin is mostly mountainous terrain with a large area covered by forest land. The overall habitat quality of the basin is good, but the degree of fragmentation is high, and the areas with high habitat quality are scattered. The MSPA results (

Table 5. Landscape classification based on MSPA.

Landscape Types	Area (km2)	Proportion of Foreground Elements (%)	Proportion of Total Area (%)
Core	21,972.60	83.84	23.80
Islet	5.69	0.02	0.01
Perforation	454.56	1.73	0.49
Edge	3422.63	13.06	3.71
Bridge	76.82	0.29	0.08
Loop	14.11	0.05	0.02
Branch	259.96	0.99	0.28

) indicated that the core area was 21,972.6 km², accounting for 23.80% of the total area and 83.84% of the seven landscape types. Furthermore, the core areas greater than 100 km² were analyzed for landscape connectivity, and patches with d_pc greater than 0.5 were selected as the final ecological sources (

Table 6. Statistics of ecological connectivity.

Serial Number	Source Number	dpc *	Serial Number	Source Number	dpc *
1	10	72.08984	13	7	2.723802
2	13	19.42792	14	21	1.590649
3	6	11.40909	15	28	1.554688
4	9	7.514814	16	42	1.211476
5	23	5.423434	17	20	1.211075
6	32	4.78267	18	40	1.125012
7	4	4.769856	19	41	0.978304
8	22	4.641909	20	34	0.685949
9	5	4.442316	21	26	0.510728
10	43	3.716952	22	3	0.509328
11	24	3.612249	23	19	0.501599
12	11	3.018604

* Only patches of ecological with d_pc greater than 0.5 are shown in the table.

). A total of 23 ecological sources with a total area of 9019.56 km² and a maximum patch area of 2478.22 km² were obtained, mostly large-scale irregular patches, including Baima Snow Mountain, Yunling, Yunlong Tianchi, Wuliang Mountain, Cangshan Erhai, Xishuangbanna, and several other nature reserves. The land use analysis revealed that the area of ecological lands such as forestland, grassland, and open water accounted for 96.62% of the land area, while cultivated land, construction land, and unused land accounted for only 3.38%, indicating the high ecological value of the land.

4.1.2. Evaluation of Corridor Importance and Construction of the Ecological Network

The results of the resistance assessment indicated that the basic resistance values in the study area were evenly distributed between high resistance values and ground resistance values. After night light coefficient correction was applied to the resistance values, the maximum comprehensive resistance value in the study area reached 363,746,875, and the areas with high resistance values showed a clumped distribution, which is consistent with the distribution of areas with active human activities (Figure 3). A total of 39 ecological corridors were extracted in the study area (Figure 4), with lengths ranging from 0.48–264.38 km and a total area of 1873.17 km². Fewer corridors in the south of the study area had longer lengths, while many corridors in the middle and north of the study area had shorter lengths. The average corridor length of the basin was 48.03 km, and short-distance corridors below the average length accounted for 66.67% of the area, mostly distributed in the central and northern areas of the study area. That is because the central and northern parts are mostly tall mountains, with large forest and grassland coverage, good habitat quality, proximity to the source areas, low resistance, and strong connectivity. The terrain in the south area is relatively flat, with high human activity intensity and high resistance values. Currently, the southern sources are far away from each other, and no habitat is suitable for the species to inhabit or travel through the middle. Therefore, the ecological corridors identified in this study are few and large. In this study, the gravity model was used to quantify the importance of the corridor by analyzing the interaction strength between the sources. In total, 15 key corridors (G_ij > 3000), 17 important corridors (1 < G_ij < 3000), and 7 general corridors (0 < G_ij < 1) were extracted to form an ecological corridor network, which can increase the probability of species migration and diffusion.

4.1.3. Identification of the Ecological Restoration Areas

Identification of the Ecological Pinch Points and Obstacles

A pinch point is the most active area between ecological sources and plays an important role in maintaining the connectivity between these sources. The current intensity was divided into four grades, and the area with the highest current density was considered the ecological “pinch point”. Thus, 16 ecological pinch points were identified (Figure 5) with an area of 76.58 km², mainly distributed as strips. The largest “pinch point” was in the north–south direction, located in the west of Ning’er County, with an area of 15.19 km². This area connects Ning’er County with Pu’er City in Simao District, connects the No. 18 source area with Laiyang River Nature Reserve (No.19 source area), and coincides with the important No. 34 ecological corridor. In addition, four key corridors coincide with this pinch point. These areas are important for ESP construction and have extremely high ecological value. The other 15 pinch points are mainly distributed in Lanping County (4), Yunxian County (3), Jinghong City (2), and Yunlong County (2). Thus, several pinch points are observed in Lanping County, and the ecological flow in this region is found to be large. Therefore, particular attention should be paid to the safety of the ecological network by considering the system as a whole.

The analysis of obstacles in the ecological network indicated that the areas identified as inner obstacles in the watershed have scope for ecological improvement (Figure 6). These areas were divided into primary improvement areas and secondary improvement areas. Among them, one primary improvement area is 18.28 km², with an improved score of 10.78–17.73.

There are 12 secondary improvement areas, covering 161.19 km², of which 158.12 km² is located along the ecological corridor, indicating that the expansion of the ecological corridor is either blocked or that it is difficult to maintain ecological stability. Removing or improving obstacles is extremely important for regional landscape connectivity. There is an urgent need for ecological restoration in primary improvement areas, as significant ecological restoration measures can lead to substantial effects. For example, ecological restoration in this area can directly improve the overall efficiency of the ecological network connectivity of watersheds. The secondary improvement areas are widely distributed, and the effects of restoration measures on these areas are relatively slow, indicating that a long-term restoration strategy is required for these areas.

Identification and Restoration of the Ecological Breakpoints

The ecological breakpoint is the intermittent point of discontinuity on the corridor, which is mainly caused by the disruption of the ecological corridor due to traffic lines. The existing ecological breakpoints will significantly increase the difficulties in the migration of species between different ecological sources and obstruct the exchange and diffusion of species [44]. In this study, 29 areas where railways, national highways (including expressways), provincial highways, and ecological corridors intersect are identified as ecological breakpoints (Figure 7), with Yumo Railway identified as one breakpoint, national highways contributing to 14 breakpoints, and provincial highways contributing to the remaining 14 breakpoints. Railways, national highways, and provincial highways constitute important transportation facilities, which cannot be directly cleared. Therefore, wildlife passages for animals, such as culverts, under bridges, flyovers, and underground passages, should be set up in the ecological breakpoints based on the local conditions.

4.1.4. Protection and Restoration Strategies Based on the Key Ecological Elements

Protection and restoration measures are proposed based on the identification and spatial distribution of the ecological network, pinch points, obstacles, and ecological breakpoints in the Lancang River basin. Firstly, there is no doubt that for the protection of ecological patches, i.e., ecological source areas, including Baima Snow Mountain, Yunling, Yunlong Tianchi, Wuliang Mountain, Cangshan Erhai, Xishuangbanna, and several other nature reserves, ecological conservation measures must be strictly implemented, and ecological conservation projects, such as returning farmland to forests and closing off mountains for afforestation must be undertaken to improve forest coverage. It was observed that the key ecological corridors and important ecological corridors have good connectivity to the source area. There should be a focus on planning and protecting the corridors with better habitat suitability based on the ecological corridor identification results and terrain, vegetation, food resources, and other factors related to the ecological corridor. Trees should be planted on both sides of the roads in the primary improvement areas and ecological breakpoint areas for ecological restoration. Meanwhile, safe passages for wildlife should be established and monitored regularly. In addition, wildlife warning signs should be installed and maintained properly. The major land types in the secondary improvement areas are cultivated land and forest land. The area of cultivated land should be strictly controlled, and the conversion of farmland to forests should be carried out in a timely manner in areas that are unsuitable for cultivation. The ecological restoration of forest land can involve reasonable control over logging, strengthening forest management, and proper economic considerations in forest plantations, such as the cultivation of Chinese herbal medicine and mushrooms. Country parks can be set up in areas close to cities and towns, and the development of the forest fruit industry, vacation and leisure industry, and tourism can promote the economic development of the area while also maintaining ecological integrity.

5. Discussion

A proper ecological security pattern can improve the fragmentation of regional habitats, enhance the mobility of species between patches, and protect biodiversity. The overall habitat quality in the Lancang River basin is good, but the degree of fragmentation is high. The ecological network constructed and ecological nodes identified in this study can improve the overall connectivity of the ecological patches for overcoming the above shortcomings. In this paper, the habitat quality index and landscape connectivity are used as important bases for ecological source identification, and circuit theory is used to construct an ecological network. The research realizes the organic unity of the InVEST model, the MSPA model, and circuit theory, which is a new idea for the study of ecological security patterns. The ecological source is the basis for the construction of ecological security patterns. Previous studies have used wetlands, forest parks, and natural reserves as ecological sources, which is highly subjective. However, this study used the InVEST, MSAP, integrated habitat quality index, and landscape connectivity index to identify ecological sources, which makes this study more scientific. For example, scholar Wu Jiansheng integrated habitat quality assessment and landscape connectivity to construct an ecological source identification system in Shenzhen [45]. Zhou Jing integrated the InVEST and MSPA models to scientifically identify the ecological source of the Kaidu River Basin [16]. The research method of this paper is supported by existing studies. Through InVEST and MSPA models, high-quality ecological source areas can be screened. The resistance surface was constructed by comprehensively considering the natural environment, human society, and landscape pattern elements, and night light data were used to modify the resistance surface. The final resistance surface was the result of the combined action of the natural substrate and human influence. This study established a comprehensive resistive surface with objectivity and logic. The application of circuit theory to construct the ecological network has the following advantages: (1) simulating species movement based on current intensity, which makes it unnecessary to artificially eliminate the ecological corridor to avoid subjectivity, and (2) identification of the pinch points and obstacles based on the current density helps to explicitly identify the key areas for ecological protection and restoration.

Although the Lancang River Basin (Yunnan section) was selected as the research area, the research method in this paper was also applicable to other areas. Considering the scale of the study area, this study selected giant patches with more than 100 km² as ecological sources, and some smaller high-quality patches may be omitted. Other scholars should consider the scale of the study area in future research, and determine parameters according to different research scales. However, this study also has some limitations. Different species have different habitat suitability and migration capacity, which can affect the ecological network structure to a certain extent. Therefore, the model parameters should be set differently for different species. In addition, corridor width is also an important factor that should be considered. The width of the migration corridor varies with species, corridor structure, connectivity, and the matrix in which the corridor is located. For birds and small animals, the corridor width should be tens of meters to meet these species’ requirements. However, the migration of large mammals requires larger corridor widths, which can even reach thousands of meters or even tens of kilometers [6]. Therefore, future studies should be based on indicators, parameters, and the width of the ecological corridor for specific species.

6. Conclusions

In this study, the Lancang River basin (Yunnan section) was considered as the study area, and InVEST, MSPA, and circuit theory were comprehensively used to build the ecological security pattern of the basin. The following conclusions can be drawn from this study:

In this study, the InVEST model was used to evaluate the habitat quality of the study area. Based on the habitat quality index, MSPA and Guidos software were used to identify the core area as an alternative ecological source. Then, through the analysis of the landscape connectivity of the alternative ecological sources, 23 ecological sources with an area of 9019.56 km² were obtained. The main land use types in the ecological source area are woodland, grassland, and water areas, which all have high ecological value. Using the InVEST and MSPA models to identify ecological sources has the advantages of accuracy, objectivity and improved ecological connectivity. Based on the current density map generated by the circuit theory, the Linkage Pathways Tool was used to generate potential ecological corridors, and a total of 39 ecological corridors were obtained. The ecological corridors in the study area showed the characteristics of fewer corridors but longer lengths in the south and more corridors but shorter lengths in the central and northern parts. The main reason is that the central and northern regions are mostly mountainous areas with good habitats and good vegetation conditions, while the southern regions are flat, with frequent human activities and relatively high resistance. The importance of corridors was quantified by the gravity model, and 15 key corridors (G_ij > 3000), 17 important corridors (1 < G_ij < 3000), and 7 general corridors (0 < G_ij < 1) are finally extracted to form a watershed ecological safety network, thereby increasing the probability of species migration and diffusion. Among them, key corridors and important corridors are of great significance to improve the connectivity of ecological sources, and should be prioritized for planning and protection.

In this study, 16 ecological pinch points are identified, mainly distributed as strips, and Lanping County is found to be the region with the most pinch points. One primary ecological improvement area and 12 secondary improvement areas are identified using an analysis of obstacle factors. The primary improvement areas can significantly improve the connectivity of the watershed ecological network. Therefore, high priority should be given to this area for ecological restoration. The secondary improvement areas are widely distributed, and the improvement in these areas as a result of restoration can be relatively slow, indicating the need for a long-term restoration strategy in these areas. Moreover, 29 ecological breakpoints are formed when the ecological corridor intersects with the traffic line, and ecological protection and restoration should be carried out in the corresponding ecological breakpoint areas.

In conclusion, the construction of an ecological security pattern in the Lancang River Basin (Yunnan section) is of great significance for the protection of biodiversity and regional sustainable development. In the process of regional development and urban expansion, development on these important ecological sources, ecological corridors, and key nodes should be avoided as much as possible to maintain the stability of the regional ecology. This study can provide useful reference for future biodiversity conservation and urban development planning and also provide some methods for constructing ecological security patterns. In view of some deficiencies and limitations of this study, future research should be conducted on more specific species and target areas.