Designing an Ecological Network in Yichang Central City in China Based on Habitat Quality Assessment

Abstract

Habitat fragmentation seriously threatens urban biodiversity conservation and ecosystem integrity. Constructing an ecological network and improving the connection level between habitat patches can effectively alleviate the general ecological environmental problems of rapid urban development. In this paper, three focal animal species were selected in the central urban area of Yichang City in China. Based on the habitat quality assessment results of the InVEST model, the ecological network of the three focal species was designed by combining morphological spatial pattern analysis and least-cost path models, and a multi-species comprehensive ecological network of the study area was designed. The consensus identified 31 ecological sources, 64 ecological corridors and 151 ecological nodes. The results can provide support for biodiversity conservation and green space planning in the study area, and also provide a reference for the construction and optimization of ecological networks for biodiversity conservation in urbanized areas.

1. Introduction

Increasing urban development worldwide and the subsequent decrease of biodiversity has caused widespread concern [1,2]. In urban areas, habitat loss and fragmentation are often considered as the main threat to biodiversity and one of the causes of the current species extinction crisis [3,4]. Habitat loss directly leads to species decline [5], while habitat fragmentation leads to changes in environmental conditions within the habitat that do not meet the needs of species for habitat and movement space. The changes can intercept species’ normal dispersal, migration, and colony building behaviors, forcing them to live in habitats that may not be sufficient to sustain populations, reducing or eliminating their potential to achieve genetic variation, and leading to increased mortality and a decrease in species numbers [6]. Numerous studies have shown that constructing ecological networks can cascade fragmented habitats through ecological corridors, restore connectivity between patches of islanded habitats, reduce habitat fragmentation [7], protect landscape integrity and ensure the transmission of ecological functions, and is one of the most important tools for biodiversity conservation in cities [8,9].

Several methods have been applied to ecological network research both in China and internationally. In general, there are two technical paths of existing ecological network planning methods from different dimensions: one is based on the principle of structural optimization; the other is based on the identification of functions [10]. The design method based on the principle of spatial structure optimization identifies ecological sources and ecological corridors based on landscape pattern analysis [11], landscape morphology analysis [12,13], and network analysis [14], which focuses more on the spatial distribution and pattern of ecological networks and less on the ecological service functions and values of ecological spaces. The design method of identifying function-first spaces includes spatially based focal species conservation suitability assessment [15,16], landscape ecological performance [17], and ecosystem service values [18,19,20], but most approaches neglect the structural nature of landscape ecosystems [10]. In this study, based on the availability of data, we aim to combine these two technical paths organically, to increase the scientific and rational nature of ecological network planning. For example, morphological spatial pattern analysis (MSPA) and landscape connectivity evaluation help guide ecological source site identification from structural attributes, and ecological network structure analysis based on complex networks is also evaluated from a spatial structural perspective. Comprehensive evaluation of ecosystem services with the habitat quality assessment module of the Integrated Valuation of Ecosystem Services and Trade-offs (InVEST) model can quantitatively assess habitat quality [21,22,23,24], providing source site identification based on functional attributes and resistance surface information of the least-cost path (LCP) model. Habitat quality reflects the ability of an area to provide species survival conditions, and habitat quality assessment has been used to guide the construction of urban green space, becoming a popular research topic in recent years [25]. The model evaluation method is an important method for habitat quality assessment, among which the InVEST-HQ model is considered one of the most mature habitat quality assessment models because of its visual image presentation, intuitive clarity and easy interpretation [26,27]. Targeting the conservation of focal species is one of the patterns of network construction that identifies functional priority patches [10]. The focal species approach to biodiversity conservation was proposed by Lambeck in 1997 [28]. Multiple focal species characterise different aspects of the habitat in which all species are found, and these species are considered as a focal community [29]. By restoring, protecting and managing the habitat needed for this focal community, the aim is to conserve most species, and indeed biodiversity as a whole. Based on this, this paper explores the ecological network design method based on habitat quality evaluation, and integrates the optimization of ecological network spatial structure and function.

Yichang City in China’s Hubei Province is famous for its hydropower and is an important node city of the Yangtze River Economic Belt. The city is located in a fragile ecological zone in the central canyons and mountains of China, with a prominent ecological status. The rapid development of urbanization in Yichang has led to a series of urban problems such as increased habitat fragmentation, reduced biodiversity, and ecological damage. At present, the protection and optimization of ecological networks in Yichang are still imperfect, and there is a lack of accurate identification and scientific assessment of important patches and corridors, which is not conducive to the conservation of urban biodiversity and habitats, and also to the function of ecosystem services of urban green infrastructure. Thus, this study aims to design a potential ecological network that provides habitats and migration paths for species in the central city of Yichang based on the focal species approach, the InVEST model, the MSPA and the LCP model, the integrated spatial morphological and functional attributes, and using ecological spaces such as woodland, grassland, wetland, cropland, and urban green space as the basis. The aim is to provide a scientific basis for urban ecological construction and biodiversity conservation in Yichang.

2. Study Area and Data Sources

2.1. Overview of the Study Area

Yichang is located at the junction of the middle and upper reaches of the Yangtze River, in the transition zone between the Qinba Mountains and the Wuling Mountains to the Jianghan Plain. The city has a complex topography and a wide range of heights, with an average altitude difference of nearly 1000 m, and a humid subtropical monsoon climate.

The study area is the central urban area of Yichang City, Hubei Province (Figure 1), covering the administrative districts of Wujiagang District, Xiling District, Xiaoting District, Dianjun District, Sanxia Dam region, Xiaoxita Street, Sandouping Town, Taipingxi Town, Yaqueling Town, Anfusi Town, Gujiadian Town, Letianxi Town, Longquan Town, Baiyang Town, Gaobazhou Town and Honghuatao Town, with a total area of about 2832 km².

2.2. Data Sources

Data sources used in this paper include land use data, DEM data, NDVI data, nighttime lighting data, and road data of the central urban area of Yichang in 2020 (

Table 1. Data sources.

Data	Source	Spatial Resolution
Land use data for 2020	Resource and Environmental Science Data Service Center of the Chinese Academy of Sciences (http://www.resdc.cn (accessed on 5 February 2022)), and further corrected the data based on Google Earth HD historical images with visual interpretation	30 m
Digital Elevation Model (DEM)	Geospatial Data Cloud Platform of Chinese Academy of Sciences (http://wwwgscloud.cn/ (accessed on 4 January 2022))	30 m
eMODIS normalized vegetation index (NDVI) data	U.S. Geological Survey website (http://www.usgs.gov/ (accessed on 4 January 2022))	250 m
NPP/VIIRS night lighting data	National Oceanic and Atmospheric Administration website (http://www.noaa.gov/ (accessed on 6 February 2022))	500 m
Road data	OpenStreetMap	-

).

3. Research Methodology

The research method consists of four parts: screening of ecological source sites and comprehensive resistance surface construction, comprehensive identification of ecological corridors, comprehensive identification of ecological nodes, and ecological network structure analysis. The specific research framework is shown in Figure 2.

3.1. Focal Species Selection

Monitoring and managing each species in biodiversity conservation studies is unrealistic, so it is necessary to consider specific species as targets and analyze landscapes using their habitat requirements as the criteria [30]. The focal species approach is an efficient and feasible approach in situations where there is a relative lack of ecological data and where species and habitats are under threat and regional landscape ecological management plans need to be developed as soon as possible [30,31]. Based on the focal species theory and the research in related literature [28,32,33,34,35], this paper focuses on the following factors to select focal species: (1) representing different habitat types in the study area; (2) being widely distributed in the study area and facing the threat of urbanization; (3) being able to attract public attention; and (4) being biologically representative and typical.

According to the above selection criteria, after expert consultation and review of relevant literature, three species were selected as focal species for the design of ecological networks in this study: yellow-bellied tit (Parus venustulus), black-spotted side-pleated frog (Pelophylax nigromaculatus), and grass hare (Lepus capensis) (

Table 2. Analysis of three focal species.

Species	Yellow-Bellied Tit (Parus venustulus)	Black-Spotted Side-Pleated Frog (Pelophylax nigromaculatus)	Grass Hare (Lepus capensis)
Animal category	Birds, climbing birds	Amphibians	Small mammals
Habitat characteristics	It mainly inhabits various forests in the mountains below 2000 m in altitude and descends to secondary forests, plantation forests and forest edge sparse scrub areas in the low mountains and foot plains in winter [40]	It lives in coastal plains to hills and mountains at an altitude of about 2000 m. It is commonly found in paddy fields, ponds, lakes and ditches near still water or slow flowing rivers [41]	Most of the subspecies in the Yangtze River valley are distributed in the mountains of the red loam belt and live mostly in winter around villages, narrow valleys, ravines, scrubs and grasses at the edge of fields. In summer and autumn, they tend to dig holes in the forest or scrub on ventilated and cool hills [42]
Protection level	China’s special birds; List of Terrestrial Wild Animals of National Protection Beneficial or of Economic or Scientific Research Value (China); Red List of Threatened Species (IUCN): 2016–Least Concern (LC)	Red List of Threatened Species (IUCN) ver 3.1: 2004–Near Threatened (NT); Red List of China’s Biodiversity—Vertebrates Volume (Amphibians)—Near Threatened (NT)	List of Terrestrial Wild Animals of National Protection Beneficial or of Economic or Scientific Research Value (China); Red List of Mammalia rabbits (IUCN) ver 3.1: 2008–Low Risk (LC)

). These species are widely distributed and threatened in the study area, representing the main habitat types in the central urban area of Yichang. They are not only culturally and educationally significant, but also popular among the public [36,37,38,39].

3.2. Habitat Quality Assessment Based on the InVEST Model

The InVEST model allows a quantitative assessment of habitat quality through its habitat quality module. The core of this method is to calculate the degree of habitat degradation by calculating the negative impact of threat sources on the habitat, and then calculating the habitat quality in relation to the habitat suitability [43]. To run the habitat quality module, data such as land use data, threat sources, sensitivity to threat sources and other parameters are needed to calculate the habitat quality index. The formula of the InVEST model is as follows.

$${\mathrm{Q}}_{\mathrm{xj}}={\mathrm{H}}_{\mathrm{j}}\left[1-\left(\frac{{\mathrm{D}}_{\mathrm{xy}}^{\mathrm{Z}}}{{\mathrm{D}}_{\mathrm{xy}}^{\mathrm{Z}}+{\mathrm{k}}^{\mathrm{z}}}\right)\right],$$

(1)

$${\mathrm{D}}_{\mathrm{xj}}={{\displaystyle \sum }}_{\mathrm{r}=1}^{\mathrm{R}}{{\displaystyle \sum }}_{\mathrm{y}=1}^{{\mathrm{Y}}_{\mathrm{r}}}\left(\frac{{\mathrm{W}}_{\mathrm{r}}}{{{\displaystyle \sum }}_{\mathrm{r}=1}^{\mathrm{R}}{\mathrm{W}}_{\mathrm{r}}}\right){\mathrm{r}}_{\mathrm{y}}{\mathrm{i}}_{\mathrm{rxy}}{\mathsf{\beta }}_{\mathrm{x}}{\mathrm{S}}_{\mathrm{jr}},$$

(2)

In Equations (1) and (2)$, {\mathrm{Q}}_{\mathrm{xj}}$ is the habitat quality ${\mathrm{H}}_{\mathrm{j}}$ is each land use type assigned a habitat score from 0 to 1 (1 = highest habitat suitability, 0 = no habitat). $\mathrm{z}$ = 2.5 and $\mathrm{k}$ are scaling parameters (or constants). The $\mathrm{k}$ constant is 0.5. $\mathrm{r}$ is the land use type threat, with $\mathrm{r}$ = 1,2,… n; $\mathrm{R}$ is the index of all the modeled degradation sources; $\mathrm{y}$ is the index of all grid cells on $\mathrm{r}$’s raster map; ${\mathrm{Y}}_{\mathrm{r}}$ indicates the set of grid cells on $\mathrm{r}$’s raster map; ${\mathrm{W}}_{\mathrm{r}}$ are the weight parameters; ${\mathsf{\beta }}_{\mathrm{x}}$ is the accessibility factor.

$${\mathrm{i}}_{\mathrm{rxy}}=1-\left(\frac{{\mathrm{d}}_{\mathrm{xy}}}{{\mathrm{d}}_{\mathrm{r} \max }}\right) (\mathrm{Linear} \mathrm{decay}),$$

(3)

$${\mathrm{i}}_{\mathrm{rxy}}=\exp \left[-\left(\frac{2.99}{{\mathrm{d}}_{\mathrm{r} \max }}\right){\mathrm{d}}_{\mathrm{xy}}\right] (\mathrm{Exponential} \mathrm{decay}),$$

(4)

In Equations (3) and (4)$, {\mathrm{d}}_{\mathrm{xy}}$ is the linear distance between grid cells $\mathrm{x}$ and $\mathrm{y}$. ${\mathrm{d}}_{\mathrm{r} \max }$ is the maximum effective distance of threat $\mathrm{r}$’s reach in space.

Since the environmental conditions of cultivated and construction land are vulnerable to anthropogenic disturbance, such as anthropogenic tillage disturbance, anthropogenic predation, extensive use of pesticides and fertilizers, and environmental pollution of farmland can affect the species and numbers of amphibians and other animals, and endanger their survival and reproduction [44,45,46,47], water fields, dryland, urban construction land, rural settlements, and other construction land were defined as threat sources. With reference to the recommended values of the InVEST model manual, relevant literature [23,37,48,49,50], habitat requirements of focal species [36,40,41,42] and the results of expert consultation, the impact range and weights of threat sources (Table S1 in Supplementary Information), the sensitivity parameters of habitat types to threats and habitat suitability parameters (Table S2 in Supplementary Information) were determined.

3.3. Ecological Source Site Identification Based on MSPA and Landscape Connectivity Index

Morphological spatial pattern analysis(MSPA)is an image processing method based on mathematical morphological principles to identify, measure, and segment the spatial pattern of raster images [51]. This can show the role of connectivity at the level of spatial structure and precisely identify the type and structure of the landscape [52,53]. Based on the Euclidean distance threshold between raster cells, the binary raster image can be segmented into seven elements: core, edge, islet, bridge, perforation, branch, and loop. Landscape connectivity evaluation can intuitively reflect the strength of connectivity between various patches. We use the Patch Importance Index (dPC) to screen ecological sources, where dPC represents the importance of the elements. The larger the dPC value, the more important the elements are. The equations were as follows.

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

(5)

$$\mathrm{dPC}=100\times \frac{\mathrm{PC}-{\mathrm{PC}}_{\mathrm{remove}}}{\mathrm{PC}},$$

(6)

In Equations (5) and (6)$, \mathrm{PC}$ is the possible connectivity index, 0 < $\mathrm{PC}$ < 1; n is the total number of patches. ${\mathrm{a}}_{\mathrm{i}}$ and ${\mathrm{a}}_{\mathrm{j}}$ are the patches $\mathrm{i}$ and patches $\mathrm{j}$ the area of the patch. ${\mathrm{P}}_{\mathrm{ij}}^{*}$ is the number of species in patch $\mathrm{i}$ and patch $\mathrm{j}$ the maximum of the product of all path probabilities between patch $\mathrm{i}$ and patch $\mathrm{j}$. ${\mathrm{A}}_{\mathrm{L}}$ is the total area of the landscape. $\mathrm{dPC}$ is the patch importance index.${\mathrm{PC}}_{\mathrm{remove}}$ is the index of possible connectivity of the remaining patches after removing a patch.

Based on the Guidos Toolbox software platform, this study used the habitat quality evaluation results to generate binary images with high quality habitats (0.8~1) as foreground elements and medium to low quality habitats (0~0.79) as background elements. It used the eight-neighborhood analysis method to identify seven types of landscape elements based on MSPA for the study area. Considering the landscape structure and patch area comprehensively, for the core area patches for yellow-bellied tits and grass hares with habitat area larger than 200 ha and black-spotted side-pleated frogs with habitat area larger than 100 ha, landscape connectivity analysis was conducted using Conefor2.6 software. The natural breakpoint method was used to classify the dPC values into seven levels, and the three levels with high dPC values were extracted as ecological source sites. According to the dispersal ability of different species, the patch connectivity distance thresholds for the yellow-bellied tit, black-spotted side-pleated frog, and grass hare were set to 15 km, 1500 m, and 3000 m, respectively [54,55,56,57].

3.4. Resistance Surface Construction Based on Hierarchical Analysis

Landscape resistance indicates the ease of species migration between different landscape units. The spatial distribution of resistance values in all patches of the landscape constitutes the landscape resistance consumption surface [58]. According to the relevant literature [59,60] and the actual situation of the study area, six factors were selected as resistance factors: habitat quality, normalized difference vegetation index, elevation, slope, night lighting, and roads. Considering the influence of the same land use type on the landscape resistance value under different development and construction intensity, the night lighting data can better characterize the intensity of human activities and spatial distribution [61], and thus night lighting data was selected as one of the resistance factors.

Hierarchical analysis (AHP) is a better method to determine the weights of the influencing factors [48]. In this study, the cost surface was used as the target layer and the resistance categories (ecological resistance factor and social resistance factor) as the scheme layer, and the resistance factors were compared two by two and assigned importance scores to construct the judgment matrix of the resistance factors. The judgment matrix was tested for consistency to determine the weights of each resistance factor. ArcGIS was used to standardize the resistance values into five levels, and the six resistance factors were weighted and superimposed to obtain the comprehensive resistance surface of each species. The consistency test equation is as follows.

$$\mathrm{CR}=\frac{\mathrm{CI}}{\mathrm{RI}},$$

(7)

$$\mathrm{CI}=\frac{{\mathsf{\gamma }}_{\max }-\mathrm{n}}{\mathrm{n}-1},$$

(8)

In Equations (7) and (8)$, \mathrm{CR}$ is the stochastic consistency ratio of the judgment matrix. $\mathrm{CI}$ is the general consistency index of the judgment matrix. $\mathrm{RI}$ is the average stochastic consistency index of the judgment matrix, whose value depends on the size of the two-comparison matrix. ${\mathsf{\gamma }}_{\max }$ is the maximum eigenvalue; and $\mathrm{n}$ is the number of evaluation factors.

3.5. Simulation of Ecological Corridors Based on Least-Cost Paths

The least-cost pathway (LCP) model identifies the least-consumption pathway between sources and targets, simulates the optimal pathway for species migration dispersal, and is a common method for designing ecological networks [62]. The Linkage Mapper tool was used in ArcGIS platform to import integrated resistance surfaces and ecological source sites to identify the least-cost paths for species migration dispersal and obtain the least-cost distance corridors, i.e., potential ecological corridors.

3.6. Ecological Node Identification Based on Circuit Theory

Circuit theory draws on the physics of electrons traveling randomly in a circuit to model the migratory movement of biological flows through a heterogeneous landscape [63]. Pinchpoints are locations in ecological corridors with high current density, representing a higher probability of species passage. The Pinchpoint Mapper module in the Linkage Mapper plugin was used to select the “all to one” mode for iterative operations. The current density was divided into seven categories according to the natural breakpoint method, and the highest category was extracted as the ecological pinchpoint. Barrier points represent the areas with high resistance to species movement between sources. The Barrier Mapper tool was used to calculate the model without checking the improvement score relative to the minimum cost path percentage, and the model was set to the “Maximum” calculation mode, and the iteration radius was set to 200 m [61]. Ecological pinch points and barrier points are ecological nodes that play a key role in the ecological processes between patches in the ecological corridor.

3.7. Analysis of Ecological Network Structure Based on Complex Networks

The ecological network is coupled with the complex network for analysis, and the ecological source points and ecological corridor numbers of the ecological network are substituted into the nodes and edges in the complex network by Gephi software to construct the network. The degree, average path length and clustering coefficient are calculated to explore the characteristic laws of the ecological network.

Degree refers to the number of edges connected to a node in a network. In an ecological network, the degree of the corresponding node represents the number of potential ecological corridors connected to that ecological node, and the greater the degree of the ecological node, the higher its importance in the ecological network is proved [64]. The formula is as follows.

$${\mathrm{k}}_{\mathrm{i}}={{\displaystyle \sum }}_{\mathrm{j}\in \mathrm{N}}{\mathrm{A}}_{\mathrm{ij}},$$

(9)

In Equation (9), ${\mathrm{A}}_{\mathrm{ij}}$ is the correlation matrix of network indicators, when node $\mathrm{i}$ is correlated with node $\mathrm{j}$, ${\mathrm{A}}_{\mathrm{ij}}$ = 1, otherwise ${\mathrm{A}}_{\mathrm{ij}}$ = 0.

The average path length is the average number of other ecological nodes spaced on the shortest path between any two ecological nodes, reflecting the association relationship between network nodes and the closeness of the network, and representing the accessibility of the entire ecological network energy flow [64]. The formula is as follows.

$$\mathrm{L}=\frac{1}{\mathrm{N}\left(\mathrm{N}-1\right)}{{\displaystyle \sum }}_{\mathrm{i}\ne \mathrm{j}}{\mathrm{d}}_{\mathrm{ij}},$$

(10)

In Equation (10), ${\mathrm{d}}_{\mathrm{ij}}$ is the shortest path through nodes $\mathrm{i}$ and $\mathrm{j}$. The clustering coefficient is the ratio of the actual number of edge connections existing between neighboring nodes of a node in the network to the number of edge connections existing in theory. It represents the clustering characteristics of ecological nodes in the ecological network and reflects the degree of connection between the nodes and their neighbors. When the average clustering coefficient is larger, the more frequent the connections between the nodes and the more even and dispersed the distribution of ecological nodes in the network [65].

$${\mathrm{C}}_{\mathrm{i}}=\frac{{\mathrm{E}}_{\mathrm{i}}}{{\mathrm{k}}_{\mathrm{i}}\left({\mathrm{k}}_{\mathrm{i}}-1\right)/2},$$

(11)

In Equation (11), ${\mathrm{C}}_{\mathrm{i}}$ is the the clustering coefficient of node $\mathrm{i}$. ${\mathrm{E}}_{\mathrm{i}}$ is the actual number of edge connections between adjacent nodes of node $\mathrm{i}$,${ \mathrm{k}}_{\mathrm{i}}\left({\mathrm{k}}_{\mathrm{i}}-1\right)/2$ is the theoretical number of edge connections between adjacent nodes of node $\mathrm{i}$.

4. Results and Analysis

4.1. Habitat Quality Assessment

The distribution of habitat quality indices for the three species was quantified using the habitat quality module in the InVEST model, and the habitats in the study area were classified into three categories (Figure 3): high quality habitats (0.8 $<$ HQ $\le$ 1), medium quality habitats (0.5 $<$ HQ $\le$ 0.8) and low quality habitats (0 $\le$ HQ $\le$ 0.5), and the area and proportion of different quality habitats were counted (

Table 3. Area and proportion of habitat quality classes for each focal species.

Habitat Quality	Yellow-Bellied Tit		Black-Spotted Side-Pleated Frog		Grass Hare
	Area (ha)	Proportion	Area (ha)	Proportion	Area (ha)	Proportion
Low(0~0.5)	109,834.2	38.78%	214,167.15	75.62%	111,075.21	39.22%
Medium (0.5~0.8)	50,321.07	17.77%	216.54	0.08%	120,743.19	42.63%
High (0.8~1)	123,055.65	43.45%	68,827.23	24.30%	51,392.52	18.15%
Total	283,210.92	100.00%	283,210.92	100.00%	283,210.92	100.00%

).

In terms of the area of different quality habitats, the area of high quality habitats of the yellow-bellied tit is higher than that of the black-spotted side-pleated frog and the grass hare. This indicates the study area is rich in habitats such as open forests and scrublands, which are more suitable for woodland species such as the yellow-bellied tit. The area of medium quality habitats of the black-spotted side-pleated frog is only 0.08%, as the habitat quality is polarized and is either good or poor. The area of low quality habitats of the frog is the largest of the three species, and needs to be improved. The habitat quality of the grass hare was not optimistic, and there is an urgent need to protect high quality habitats and restore low and medium quality habitats.

In terms of spatial distribution, high quality habitats of the yellow-bellied tit were more evenly distributed in the study area, high quality habitats of the black-spotted side-pleated frog were mainly distributed in the southeast, and high quality habitats of the grass hare were scattered, mainly in the northwestern and central parts of the study area.

4.2. Ecological Source Site Identification

Based on the results of the InVEST model habitat quality assessment and connectivity calculation (

Table 4. Importance index and area of ecological source site patches.

Yellow-Bellied Tit	dPC	Area/ha	Black-Spotted Side-Pleated Frog	dPC	Area/ha	Grass Hare	dPC	Area/ha
a1	15.41	6467.58	b1	8.46	1695.60	c1	6.33	272.79
a2	25.91	9875.70	b2	28.37	5611.50	c2	11.67	591.21
a3	9.41	3803.13	b3	39.06	368.82	c3	7.41	594.18
a4	11.48	5748.66	b4	58.79	5661.90	c4	21.98	1185.48
a5	5.93	2743.65	b5	43.44	6300.81	c5	7.77	397.62
a6	42.40	15,645.33	b6	49.31	10,145.79	c6	28.42	735.39
a7	3.74	1310.49	b7	6.20	186.30	c7	14.09	1064.61
a8	37.41	12,533.49	b8	14.23	3513.24	c8	16.81	535.95
a9	11.79	3347.64				c9	28.01	1586.70
a10	4.01	1771.74				c10	19.09	1965.60
						c11	8.14	519.03
						c12	12.52	1702.08
						c13	7.06	310.50

), a total of 31 high-quality core area patches (Figure 4) were selected as ecological source sites. There were 10 source sites for the yellow-bellied tit, 8 for the black-spotted side-pleated frog and 13 for the grass hare, accounting for 21.01%, 11.13% and 3.81% of the study area respectively. Among them, the a6 source of yellow-bellied tit is the largest and its patch importance is the highest, followed by the a8 source. The black-spotted side-pleated frog has the largest b6 source and the highest b4 source patch importance. The c10 source of the grass hare is the largest, while the c6 source patch is the most important.

The ecological source sites were generally in the peripheral space of the built-up area, and the overall distribution was relatively balanced, but there were clustering characteristics in the distribution of different species of source sites. The source sites of the yellow-bellied tit were distributed in the northwestern mountain area of the study area, the source sites of the black-spotted side-pleated frog were mainly in the southeastern plain area and the Yangtze River, and the source sites of the grass hare were mainly in the central northward hilly area.

4.3. Resistance Surface Construction and Ecological Corridor Simulation

The weights of resistance factors for the three species were determined by hierarchical analysis (Tables S3–S5 in Supplementary Information), and the combined resistance surface for each focal species was simulated (Figure 5). For the yellow-bellied tit and grass hare, the areas with high resistance values were concentrated in the urban built-up area, followed by the southeastern agricultural area of the study area, while for the black-spotted side-pleated frog, resistance values were high in all areas except the southeastern agricultural area, indicating that the area where black-spotted side-pleated frogs can migrate is limited.

Based on the LCP model, the potential ecological corridors of the three species were simulated (Figure 6 and

Table 5. Potential ecological corridor length and cumulative resistance values.

Species	Serial Number	Ecological Corridor	Length (km)	Cumulative Resistance Value
Yellow-bellied tit	1	Corridor 1–2	17.473	1480
	2	Corridor 1–6	29.381	3826
	3	Corridor 2–3	13.281	1561
	4	Corridor 2–6	20.56	3204
	5	Corridor 3–4	40.495	8029
	6	Corridor 3–5	39.286	8537
	7	Corridor 3–6	20.131	1688
	8	Corridor 3–8	30.674	4791
	9	Corridor 4–5	8.539	1366
	10	Corridor 4–7	30.97	4349
	11	Corridor 4–8	45.642	10,958
	12	Corridor 5–7	24.62	3684
	13	Corridor 5–8	47.545	5914
	14	Corridor 6–8	19.953	2479
	15	Corridor 7–8	29.422	6243
	16	Corridor 7–9	14.779	4577
	17	Corridor 7–10	27.434	6784
	18	Corridor 8–9	19.326	2445
	19	Corridor 8–10	32.202	3058
	20	Corridor 9–10	13.983	1838
Black-spotted side-pleated frog	1	Corridor 1–2	11.519	6380
	2	Corridor 1–4	18.849	8876
	3	Corridor 1–6	32.836	31,698
	4	Corridor 2–4	12.531	6216
	5	Corridor 2–5	15.137	3406
	6	Corridor 3–4	14.147	11,827
	7	Corridor 3–5	12.466	5209
	8	Corridor 3–6	26.991	15,373
	9	Corridor 3–7	12.853	5543
	10	Corridor 3–8	22.145	10,329
	11	Corridor 4–5	10.752	3239
	12	Corridor 4–6	26.703	16,720
	13	Corridor 5–7	10.381	5094
	14	Corridor 7–8	9.551	5354
Grass hare	1	Corridor 1–2	6.376	1366
	2	Corridor 1–4	12.623	3208
	3	Corridor 1–7	14.753	3120
	4	Corridor 2–4	7.378	1570
	5	Corridor 2–5	11.817	2284
	6	Corridor 2–6	10.738	2769
	7	Corridor 2–7	11.851	2334
	8	Corridor 3–4	7.288	1887
	9	Corridor 3–5	7.367	1185
	10	Corridor 3–12	32.9	10,667
	11	Corridor 4–5	5.781	1079
	12	Corridor 4–6	12.307	2862
	13	Corridor 5–6	10.383	2365
	14	Corridor 5–8	5.917	1358
	15	Corridor 5–9	13.287	3542
	16	Corridor 5–12	30.516	9928
	17	Corridor 6–7	8.578	2167
	18	Corridor 6–8	5.12	1120
	19	Corridor 6–9	9.875	2619
	20	Corridor 6–10	13.828	1903
	21	Corridor 6–11	15.993	3425
	22	Corridor 7–10	11.102	2125
	23	Corridor 8–9	7.255	3049
	24	Corridor 9–10	17.032	3435
	25	Corridor 9–11	11.159	2575
	26	Corridor 9–12	26.034	6709
	27	Corridor 9–13	20.43	4610
	28	Corridor 10–11	8.383	1592
	29	Corridor 11–13	24.711	4926
	30	Corridor 12–13	11.286	3099

), resulting in 20 corridors for the yellow-bellied tit, 14 corridors for the black-spotted side-pleated frog, and 30 corridors for the grass hare, with average corridor lengths of 26.3 km, 16.9 km, and 13.1 km, respectively. The yellow-bellied tit corridors were widely distributed, spanning the central and northwestern parts of the study area; the black-spotted side-pleated frog corridors were mainly distributed in the southeast; and the grass hare corridors were the most densely populated. The highest density of grass hare corridors was concentrated in the central north area.

4.4. Ecological Node Identification

Ecological pinch points and ecological barrier points were mostly distributed among ecological source sites and near the edges of ecological source sites (Figure 7). The ecological network of the yellow-bellied tit identified 32 ecological pinch points and 20 ecological barrier points, and the ecological nodes in Corridor 1–6, Corridor 3–5, Corridor 3–8, and Corridor 4–8 were denser. The ecological network of the black-spotted side-pleated frog identified 19 ecological pinch points and 15 ecological barrier points, and the ecological nodes in corridors No. 3 and No. 12 were the most numerous, and protection and restoration efforts should be strengthened. The ecological network of the grass hare had 42 ecological pinch points and 23 ecological obstacle points, with a large number of ecological nodes and relatively uniform spatial distribution.

4.5. Ecological Network Structure

The degrees of the ecological source sites of the three ecological networks, as well as the average path length and the average clustering coefficient, were calculated and counted by constructing a complex network (Figure 8 and Figure 9). By calculating the degrees of each node in the network, it can be seen that the ecological source sites of yellow-bellied tit 1, black-spotted side-pleated frog 8, and grass hare 1, 3, 8 and 13 are of the highest importance and should be built and protected. The average path length of the grass hare ecological network is the highest, which indicates that the accessibility of its ecological network energy flow is higher. The mean clustering coefficient of the ecological network of the yellow-bellied tit is the highest, indicating that the ecological sources are strongly connected and the distribution of the sources is more even.

5. Discussion

Urban ecological networks are usually designed without selecting specific species and using single factors such as land use types for resistance analysis to simulate corridors [66,67]. However, this approach is not applicable to urban environments with complex geography and diverse habitat types [37,68]. The central city of Yichang is located in the transition zone of different topographic areas, and the differences in natural background leads to heterogeneity of ecological environments and species habitats. A single-function ecological network has many problems in practice construction, which can lead to uneven distribution of ecological networks and ecological disconnection between regions. Therefore, this paper addresses the specificity of cities with complex ecological environments, and optimizes the ecological network design method applicable to this type of city based on the focal species approach and using habitat quality assessment. This is the key to promoting a balanced layout of biodiversity conservation and protect the integrity of urban ecosystems.

In this study, unlike other studies in natural areas where endangered animals or large mammals are often selected as focal species [69,70], three more common and small species were selected to form the target species system, considering that the study area is located in a rapidly urbanizing area with strong anthropogenic disturbance, and the main animal taxa are very different compared to natural areas, as well as considering the concerns of people and the species’ role in cultural education. In previous studies related to habitat quality evaluation, generic species were selected, where the parameter settings were not guided by specific proxy species, but mainly based on the recommended data from the manual of model use [71,72]. The ecological network of the three focal species was superimposed to obtain a multi-species integrated ecological network (Figure 10), which generated different ecological corridors with little overlap. This was similar to the results of Hepcan et al., who selected four target species for ecological network analysis in Izmir province, Turkey [73], indicating that the selection of suitable habitat patches for different species and the establishment of a reasonable resistance surface are crucial for the design of ecological networks. In this paper, we also discussed the methods of ecological network construction and explored new methods in the selection of source sites by calculating the landscape connectivity. Previous studies usually delineated a single value as a criterion [71,72]. In this study, the habitat patches of three species needed to be selected, and their dPC values differed greatly, so we chose the method of separate ranking to select patches as source sites in the simulation of barrier point setting. The iteration radius was determined by referring to previous study [61] and according to the extent of this study area in the city.

Building an ecological network is one of the most important methods for biodiversity conservation. Based on the above study, the following suggestions are made for the actual construction of an ecological network in the central city of Yichang. First of all, we should strengthen the protection of ecological source sites. At present, some ecological source sites have been established as protected areas and forest parks, such as Sanxia Dam region Wetland Nature Reserve (source site a2), Chexi Nature Protection Community (source site a6), Wenfoshan Nature Protection Community (source site a8), Songshan Forest Park (a10), Gaobazhou Reservoir Area Wetland Nature Reserve (source site b6), Xiaoting Egret Nature Protection Community (source site b4), Muntjacs Nature Protection Community (c12), Xilingxia Zhendan Reserve (source c6), and Huaniuyan Scenic Area (c7). However, other ecological sources have not been protected [74]. It is recommended that ecological boundaries be strictly controlled, the integrity of ecological sources protected, and ecological management and environmental protection strengthened. Ecological corridors can provide relatively safe paths for the migration movement of organisms, thus protecting the biodiversity in the region [75]. The actual construction of ecological corridors should consider the suitable migration environment for major species, build landscape types with different plants, and interweave the original river and plant corridors with potential corridors to improve the connectivity between ecological source sites, giving priority to potential corridors with more ecological nodes. As an important spatial node for biological diffusion, ecological pinch points should protect the existing natural landscape, repair the more damaged fracture sections, delineate the protection range and minimize human interference. Ecological obstacle points need to optimize their habitat quality to reduce landscape resistance. For water bodies and roads with high resistance values, biological migration channels such as animal bridges, biological culverts, and ecological roads should be considered.

There are still some shortcomings in this study. Due to the lack of detailed species observation data and detailed research data of related species in the study area, the parameters of species habitat quality evaluation, and the score and weight of resistance surface construction in this study are all from related literature and expert consultation, which are somewhat subjective. The distribution data of species in the study area and more detailed information of the biological ecological habits can further validate and improve the study results. The representativeness of focal species for the overall biotope and all habitats needs to be further studied. The Yangtze River is the main water body in Yichang, and is an important habitat for a large number of fish and waterbirds. However, since the migration of species that depend on the Yangtze River needs to be considered comprehensively at the watershed scale or even at a larger scale, fish and waterbirds are not included as focal species in this study. Different organisms respond differently to the same landscape type and have different needs for habitat patches, migration paths, etc. The ecological corridors constructed in this study may not be applicable to all species. In terms of key ecological nodes, in addition to considering spatial location, the influence of the area and morphology of the nodes on biological migration should be further explored to better meet the conservation needs of realistic spatial geographic conditions. Although the ecological network was established using the LCP model and other methods, the study still did not involve interaction among ecological sources, the hierarchical classification of ecological corridors and the relationship among the ecological networks of the three focal species. Further analysis and optimization are needed. The construction of multi-scale and multi-level ecological networks is more conducive to the systematic protection of urban landscapes, and this study is at the central city-wide landscape scale. The construction of more refined ecological networks in built-up areas and urban communities can make the regional ecological network an organic whole, and there are higher requirements for data and indicators. The establishment of ecological networks at different levels and the articulation between them are key points that need further depth and strengthening in future research [61]. We can also consider adding factors related to seasonal changes guided by seasons, and taking into account seasonal behaviors such as hibernation and migration of animals.

6. Conclusions

In this study, three focal animal species—the yellow-bellied tit, the black-spotted side-pleated frog and the grass hare—were selected for habitat quality evaluation using the InVEST model. Based on the evaluation results, ecological source sites were extracted using MSPA spatial pattern analysis and landscape connectivity index as auxiliary methods. Potential ecological corridors were simulated using the LCP model, and ecological pinch points and ecological barrier points were identified based on circuit theory. A multi-level, point-line, nodal urban ecological network of “source–corridor–node” was constructed. There are three findings.

(1)

The high-quality habitats in the central city of Yichang are mainly distributed outside the built-up area, with an overall scattered and locally aggregated distribution, and the habitat quality distribution of the three focal species varies greatly.

(2)

Based on the results of habitat quality evaluation, the ecological network of the central city of Yichang was designed, and 31 ecological source sites were extracted, with a total area of 108,192.51 ha, accounting for 28.26% of the study area. By spatial characteristics, the ecological source sites were distributed more in the southeast and northwest and less in the middle. A total of 64 ecological corridors were simulated, with a length of 1154.63 km, and corridor density was high in the north and low in the south. A total of 151 nodes were identified, of which 93 were ecological pinch points and 58 were ecological barrier points.

(3)

With a relative lack of biological information in urbanized areas, it is feasible to design ecological networks based on the focal species approach and the results of habitat quality evaluation. The results of the study have significance and practical value for the construction of ecological networks targeting biodiversity conservation in rapidly urbanizing areas, and can also provide references and lessons for the construction of ecological networks in other areas.