Spatial–Temporal Evolution of Ecological Network Structure During 1967–2021 in Yongding River Floodplain

Abstract

Constructing a rational ecological network is crucial for balancing regional development with environmental protection. However, existing research typically emphasizes the analysis of overall patterns, lacking an in-depth exploration of the dynamic changes in key elements and the interactions between different components. Using the Yongding River floodplain as a case study, this study applied morphological spatial pattern analysis, landscape connectivity metrics, and biodiversity assessments to identify core ecological source areas. Circuit theory was used to delineate ecological corridors and analyze network evolution across four key years, while graph theory facilitated an in-depth analysis of network structural characteristics. Furthermore, key areas for ecological restoration were identified within the floodplain. We found that the number of ecological source patches in the study area has remained relatively stable, though their total area has shown a fluctuating decline, accounting for approximately 10% of the floodplain. Additionally, ecological corridors have decreased significantly from 1967 to 2021, with a marked reduction in major corridors, leading to increased resistance to material and energy flow and a corresponding decline in network connectivity and stability. More importantly, current ecological pinch points are primarily distributed in a bead-like pattern along the Yongding River channel, while ecological barriers are concentrated in the northern and eastern floodplain, often at intersections of dense road networks and ecological corridors. These critical areas of fragmentation within the ecological network are prioritized for targeted ecological protection and restoration efforts. Overall, this study advances our understanding of the spatial distribution and composition of key ecological elements within river corridor networks and offers a framework for evaluating these networks through a multidimensional optimization approach for ecological source patches. At the same time, we conducted an in-depth analysis of key fragmentation areas in the Yongding River floodplain, providing valuable guidance for future ecological protection and restoration initiatives in river corridors.

1. Introduction

Floodplain, an area periodically inundated by rivers during floods, serves as a crucial habitat supporting high biodiversity and plays a significant role in maintaining ecological balance and promoting human well-being [1,2]. However, many riparian areas have experienced substantial degradation due to natural events such as flooding and wildfires, as well as human-driven factors including overdevelopment. This degradation has led to severe ecosystem fragmentation and imbalance, prompting increased focus on ecological protection and restoration efforts [3,4]. Within this context, developing an ecological network system using river corridors has emerged as a critical approach to achieving ecosystem integrity in floodplain environments.

In recent years, there has been growing recognition of ecosystem complexity and stability, emphasizing the importance of studying ecological network structure and function to optimize resource utilization, enhance ecosystem integrity, and improve resilience to disturbances—all essential for sustainable ecological development [5,6]. Ecological networks comprise habitat patches and their connections. By forming corridors through fragmented landscapes, effective conservation of species diversity can be promoted, thereby maximizing regional ecological equilibrium and enhancing ecosystem service functions through rational spatial planning and management [7,8]. Previous studies indicate that network resilience depends substantially on the importance, number, and connectivity of nodes, which play a pivotal role in the network’s ability to withstand and recover from disruptions [9]. However, frequent natural and human-induced disturbances can disrupt these networks, compromising node functionality and resulting in unpredictable ecological consequences [10,11]. To address this issue, optimizing ecological networks primarily involves enhancing network connectivity by improving the structure of corridors and nodes. Scholars have progressively revealed the significance of ecological network optimization in preserving ecosystem stability and improving ecosystem service functions by establishing theoretical frameworks and employing methods such as model simulations and complex network analyses [12,13].

In this context, ecological corridors are considered essential linkages connecting core habitat patches, mitigating species isolation, and maintaining gene flow [14]. To identify and optimize ecological corridors, researchers commonly employ cost–distance models, least-cost path analyses, and graph-theoretical methods, combined with remote sensing and Geographic Information Systems (GIS), to quantitatively evaluate corridor contributions to overall network connectivity and biodiversity conservation [15,16]. Nonetheless, given the escalating complexity of ecological systems and the multifaceted pressures they face, contemporary studies still encounter limitations in precisely delineating corridors and evaluating their comprehensive benefits [17]. These shortcomings hinder timely prioritization efforts and underutilize potential synergistic effects during corridor planning. The incorporation of complex network theory provides a more systematic approach for exploring the spatiotemporal evolution of ecological networks within contexts of rapid urbanization and climate change. Network analytical metrics enable researchers to characterize dynamic structural changes and identify critical nodes or corridors that significantly influence network connectivity and stability [18]. Recent findings indicate that ecological networks often exhibit nonlinear evolutionary patterns, especially under intensive anthropogenic disturbances, where abrupt structural reorganizations may occur [19]. More recent work emphasizes the need to incorporate high-resolution remote sensing, big-data methodologies, and machine learning to further refine corridor identification, assess multiscale connectivity dynamics, and advance real-time decision-making processes [20,21]. Such comprehensive approaches extend beyond traditional morphological or connectivity analyses, moving toward multidimensional system quantification. Despite these strides, existing frameworks often lack longitudinal analyses that capture how corridor configurations respond to long-term shifts in climate patterns and demographic pressures. Certain ecosystem types or regions may face more pronounced challenges. For example, riverine ecosystems often exhibit fluctuating flow regimes and seasonally variable habitats—factors complicating corridor identification [22]. These limitations underscore the necessity of developing more holistic and adaptable ecological network models, informed by recent methodological innovations and enriched data sources.

The Yongding River serves as an important ecological barrier in China, and its ecological restoration is integral to the national strategy for Beijing–Tianjin–Hebei cooperative development. Over recent decades, however, the ecological landscape within the river basin has become increasingly scarce and fragmented. To achieve the restoration goal of a “flowing river”, the Yongding River has been subjected to ecological water transfer and diversion projects, including water replenishment from the Yellow River. Since 2017, these efforts have intensified, and, by spring 2020, continuous water replenishment enabled full-basin flow, effectively ending the year-long situation of water disruption [23]. While ecological water replenishment has reshaped the river channel, expanded surface water areas, and facilitated the revival of aquatic plants, it has also altered the existing ecological network within the basin. Therefore, under ongoing ecological recharge, further research is needed to evaluate differences between the existing ecological network and the river’s natural state prior to fragmentation. Understanding the natural state of the river corridor and studying the historical connectivity and structure of the Yongding River’s ecological network will support the identification of restoration targets, which are crucial for maintaining species diversity and ecosystem stability [24]. However, current research on Yongding River Basin’s ecological network primarily focuses on contemporary conditions [25], with some studies examining changes in channel morphology [26], which no longer fully address the needs of comprehensive ecological restoration. Critically, previous research largely neglects the historical, pre-fragmentation state of the river corridor. Specifically, quantitative analyses documenting the Yongding River floodplain’s historical ecological conditions are lacking. Most existing studies emphasize assessments conducted after the river experienced substantial drying or interruption. This oversight limits the identification of restoration targets and impedes the strategic formulation of restoration measures based on historical natural benchmarks.

Given the complex ecological environment of floodplain river corridors, this study aims to refine methods for identifying and delineating ecological networks by integrating a wider range of key indicators. Specifically, it seeks to optimize network construction and assessment by considering multiple factors such as ecological patch area, locational importance, and habitat quality. Utilizing satellite imagery from 1967, 1980, 2004, and 2021, this study examines the regional ecological environment and river conditions under typical historical hydrological states. It investigates ecological source areas, resistance surfaces, and corridors, focusing particularly on temporal ecological network evolution associated with periods of natural state before the dry flow and disruption. The comprehensive historical analysis provides essential baseline data and methodological insights, thereby directly supporting the optimization of floodplain ecological functions, informing targeted ecological water replenishment strategies, and significantly improving future ecological network assessments and management strategies.

2. Materials and Methods

2.1. Study Area and Data

2.1.1. Study Area

The Yongding River floodplain is located in the lower reaches of the Yongding River, within a plain region shaped by historical shifts in the river’s course. This floodplain includes the “Lianggezhuang-Qujiadian” river segment, extending approximately 67 km and covering approximately 581 km² across Beijing, Tianjin, and Heibei (Figure 1). In recent years, ecological challenges—such as surface water depletion, riverbed desertification, and vegetation degradation—have significantly reduced the floodplain’s ecological service functions. To address these issues, the Ministry of Water Resources launched a series of ecological water replenishment projects beginning in 2019 [27]. These efforts have effectively mitigated water shortages within the basin and improved the local ecological environment.

2.1.2. Data Sources

This study utilizes remote sensing imagery from the wet seasons (June to September) of 1967, 1980, 2004, and 2021 as the primary data source. Given the extensive time span of the research, the satellite images were captured by different satellite platforms: KH-4A, KH-9, SPOT-5, and GF1. All images have a resolution of 2–5 m, ensuring consistency in spatial detail and minimizing errors across the dataset. According to the actual characteristics of the river corridor in the study area and the relevant principles of landscape ecology, the landscape unit types of the Yongding River countryside corridor were visually interpreted in ArcGIS and classified into seven categories: arable land, forestland, grassland, bottomland, water, construction land, and unused land. Prior to visual classification, comprehensive preprocessing procedures were conducted on all satellite images, including geometric correction, radiometric calibration, and image enhancement techniques to ensure data quality. The visual interpretation criteria for each landscape category strictly adhered to China’s national standard for current land use classification (GB/T 21010-2017) [28], with emphasis on distinct remote sensing characteristics such as shape, tone, and texture. To enhance classification accuracy and reliability, multiple interpreters independently classified the imagery, followed by rigorous cross-validation. Finally, vector-based land-use data was uniformly transformed into 10 m × 10 m raster data.

The topographic data come from the geospatial data cloud platform (http://www.gscloud.cn, accessed on 31 July 2024) with an accuracy of 30 m; the spatial road data come from the road vector data crawled by the Map World (https://www.tianditu.gov.cn/, accessed on 31 July 2024).

2.2. Methods

This study was conducted in three phases (Figure 2). Firstly, core ecological patches were identified using Morphological Spatial Pattern Analysis (MSPA), landscape connectivity metrics, and habitat quality assessments. Subsequently, the ecological resistance surface was constructed based on circuit theory to delineate ecological corridors and identify key ecological nodes. Furthermore, graph theory methods were employed to evaluate changes in the ecological network structure at both the overall and node levels, while also identifying key areas for ecological restoration.

2.2.1. Identification of Core Ecological Patches

Core ecological patches, or source sites, are typically large habitats or nature reserves with high landscape connectivity. This study identifies ecological sources through a three-dimensional approach based on morphological spatial pattern analysis (MSPA) results. First, patches were assessed by area, and those larger than 100 ha were classified as suitable core patches. Second, landscape connectivity was evaluated for each patch, selecting those with a connectivity value greater than 1 as suitable. Third, habitat quality was considered by calculating integrated habitat quality scores using ArcGIS regional statistical analysis, with patches scoring an average value above 60 deemed suitable. Each criterion was categorized into three levels, as shown in

Table 1. Classification criteria for patch area, connectivity, and integrated habitat suitability levels.

Name	Classification Criteria	Suitability Level
Core area (A)	A ≥ 100 ha	Suitable
	100 ha > A ≥ 30 ha	Moderately suitable
	A < 30 ha	Unsuitable
Location importance of the core (P)	dPC ≥ 1	Suitable
	1 > dPC ≥ 0.1	Moderately suitable
	dPC < 0.1	Unsuitable
Integrated habitat quality (Q)	Q ≥ 60	Suitable
	60 > Q ≥ 45	Moderately suitable
	Q < 45	Unsuitable

. Only patches that met at least two suitability levels and showed no unsuitability levels were ultimately designated as core ecological patches. Furthermore, patches satisfying all three suitability criteria were classified as important patches, while those meeting two suitability criteria were classified as general patches.

Compared with single-method approaches, this three-dimensional framework offers a more robust evaluation. While alternative methods have been proposed in landscape ecology for ecological source identification, most single approaches present limitations. For instance, methods relying solely on MSPA or vegetation indices often overlook variations in habitat quality or underlying functional connections between patches [29,30]. Additionally, methods such as species distribution heavily rely on species occurrence data, which may suffer from data scarcity [31]. By integrating patch size, connectivity, and habitat quality within a unified analytical framework, this approach mitigates these methodological gaps and better captures the multifaceted nature of ecological source identification.

(1) Analysis of ecological spatial patterns (MSPA)

MSPA, developed by Vogt et al., is a landscape connectivity analysis method that accurately differentiates landscape types and structural elements at the pixel level. According to the actual situation of the study area, this study takes forestland, water, bottomland, and grassland as foreground elements and other landscape unit types such as construction land and arable land as background elements. In this study, we utilized Guidos Toolbox 3.3 software to classify seven non-overlapping landscape elements, including core areas and isolated patches. Core areas, which are typically larger and less isolated ecological patches in the foreground image, possess high ecological service value and contribute significantly to the overall stability of the ecosystem [29].

(2) Landscape connectivity

Landscape connectivity measures the ease with which organisms can migrate between habitat patches, which is crucial for maintaining ecological patch function and facilitating ecological flows [32]. In this study, core areas identified from MSPA output data were analyzed for connectivity using the Conefor plug-in in ArcGIS. The connectivity index or probability of connectivity (PC) was applied to evaluate the degree and importance of connectivity between patches [33] as it provides an accurate reflection of connectivity dynamics. The specific formula is as follows

$$PC={\displaystyle \frac{\sum _{i=1}^n\sum _{j=1}^n{a}_i{{ \times a}_j\times p \ast }_{ij}}{A}}$$

(1)

$$dI(dPC)={\displaystyle \frac{I-I\prime }{I}}\times 100$$

(2)

where n is the number of plaques in the study range; $a_i$ and $a_j$ refer to the area of patch i and j; ${p \mathrm{\ast }}_{ij}$ is the maximum probability of species dispersal in patches i and j; dI represents the importance of the removed element; I is the connectivity calculation result; I′ represents the result of the connectivity calculation after removing an element. Based on previous research and the scale of the study area [34], a connectivity distance threshold of 1000 m and a PC of 0.5 were set for the analysis.

(3) Integrated habitat quality

Integrated habitat quality (referred to as Q) encompasses both ecosystem service values (Q1) and habitat quality (Q2) as derived from the InVEST model [35]. The integrated habitat quality for each core area patch was obtained through grid multiplication operations and raster averaging.

The first component, Q1, represents the annual average value of biodiversity support services per unit area. This measure is based on related research findings [36], as well as economic factors specific to the Yongding River floodplain, such as average prices of major food crops and economic yields [37], which contribute to ecosystem service value coefficients (

Table 2. Ecosystem service equivalent value per unit area for each ecological space type.

Land-Use Type	Biodiversity Support Services (Value per Unit Area)
Arable land	14.66
Forestland	67.32
Grassland	22.51
Bottomland	51.63
Water	51.42
Construction land	0
Unused land	0.41

). The second component, Q2, refers to habitat quality as calculated using the InVEST model. This module evaluates habitat quality by incorporating various threat factors that impact habitat integrity [38,39]. Among them, the setting of threat factors mainly considered the land-use type and the degree of interference of human activities on the habitat. Based on the basic theory of landscape ecology and the field situation of Yongding River floodplain, arable land, construction land, bare land, railways, and highways were identified as the threat factors, and the habitat sensitivity parameters of each threat factor were scientifically set according to the relevant research results [40,41] (

Table 3. Habitat threats and its maximum impact distance, weight, and attenuation type.

Habitat Threats	Maximum Stress Distance/km	Weight	Type of Spatial Decay
Construction land	10	0.9	Exponential
Arable land	8	0.6	Linear
Unused land	6	0.8	Linear
Railway	10	1	Linear
Highway	6	0.6	Linear

and

Table 4. Habitat suitability and threat sensitivity parameters with different land-use types.

Land-Use Type	Habitat Suitability	Threat Sensitivity Parameters
		Construction Land	Unused Land	Arable Land	Highway	Railway
Forestland	0.95	0.9	0.5	0.7	0.5	0.8
Grassland	0.7	0.75	0.7	0.75	0.4	0.7
Bottomland	0.8	0.8	0.5	0.6	0.4	0.75
Water	1	0.85	0.6	0.7	0.4	0.65
Construction land	0	0	0	0	0	0
Unused land	0.1	0.4	0	0.2	0.1	0.1
Arable land	0.4	0.7	0.4	0	0.4	0.6

).

Integrated habitat quality (referred to as Q) encompasses both ecosystem service values (Q1) and habitat quality (Q2) as derived from the InVEST model [33]. The integrated habitat quality for each core area patch was obtained through grid multiplication operations and raster averaging. The natural breakpoint approach divided the calculated values of Q1, Q2, and Q into four categories (1–4) from lowest to highest. The classification results are presented in

Table 5. Classification criteria for ecosystem service, habitat quality, and integrated habitat quality levels.

Level	Ecosystem Service (Q1)	Habitat Quality (Q2)	Integrated Habitat Quality (Q)
1	Q1 < 0.41	Q2 < 0.15	Q < 15.55
2	0.41 ≤ Q1 < 22.51	0.15 ≤ Q2 < 0.50	15.55 ≤ Q < 41.13
3	22.51 ≤ Q1 < 51.63	0.50 ≤ Q2 < 0.91	41.13 ≤ Q < 51.41
4	Q1 ≥ 51.63	Q2 ≥ 0.91	Q ≥ 51.41

.

2.2.2. Ecological Resistance Surface Construction

Ecological resistance refers to the barriers that hinder species exchange and energy flow, and the magnitude of resistance is affected by different landscape types and anthropogenic activities [42]. Due to the geographical environment and the nature of the land use [43], topography, land-use type, and distance from water, highways, railroads, and construction land are selected as resistance factors in this paper. Different element types and levels of resistance factors are assigned values (

Table 6. Resistance factors and their weights.

Resistance Factors	Weight	Resistance Value
		1	2	3	4
Elevation	0.08	<10	10–30	30–50	>50
Slope	0.11	<3	3–7	7–15	>15
Land-use type	0.36	Forestland, water	Grassland, bottomland	Arable land	Construction land, unused land
Distance from water	0.14	<100 m	100–500 m	500–1000 m	>1000 m
Distance from construction land	0.12	>1000 m	500–1000 m	100–500 m	<100 m
Distance from railway	0.09	>1000 m	500–1000 m	100–500 m	<100 m
Distance from highway	0.10	>1000 m	500–1000 m	100–500 m	<100 m

). Higher resistance values for each factor indicate lower suitability for ecological corridor formation. A comprehensive resistance surface was subsequently generated through overlay analysis of these resistance values.

2.2.3. Identification of Ecological Corridors and Critical Fragmentation Areas

Ecological corridors are essential for species migration and serve as the foundational structure for maintaining regional ecological security [44]. Circuit theory uses ecological source sites as nodes and ecological resistance surfaces as conductive surfaces and determines the optimal ecological corridor by minimizing the cumulative resistance between habitat patches [45]. In this study, ecological corridors were generated using Linkage Mapper software (LM 3.0.0) based on circuit theory and categorized into “important corridors” and “general corridors” by calculating the ratio of cost distance to corridor length.

Ecological pinch points represent high-probability crossing areas for migrating organisms, where ecological risks are elevated, making these priority areas for conservation within ecological networks [46]. Ecological barriers, on the other hand, are zones that impede organism migration; optimizing these barrier areas can help reduce migration resistance and improve landscape connectivity [47]. To identify critical fragmentation areas within the ecological network, Pinchpoint Mapper and Barrier Mapper tools within Linkage Mapper software were employed to generate corridor current densities and cumulative restoration current densities. The Natural Breaks (Jenks) Method was then applied to identify the highest-priority zones, classifying them as ecological pinch points and obstacle points while excluding crushed and fine nodes within these areas.

2.2.4. Eco-Network Analysis

Network analysis serves as the primary research method for assessing ecological benefits and exploring the structural characteristics of ecological networks [48]. At the global scale, the analysis focused on several quantitative metrics, including the network closure index, dot rate, and connectivity index, to evaluate the degree of connectivity and connectivity rate of ecological networks across different chronological panhandles. At the node scale, each ecological source was regarded as a node within the network. Three key measures from graph theory—degree centrality, betweenness centrality, and closeness centrality—were employed to quantitatively assess the importance of each node within the network. These metrics reflect both the structural and functional characteristics of the network, providing insights into the role and significance of individual nodes in maintaining overall ecological connectivity. The meanings and calculation methods of the indicators are shown in

Table 7. Formulas for indices and their description.

Indices	Description [49]	Scale	Formulas [19,50]
Degree centrality	The number of other nodes connected to the node	node	$DC(i)={\displaystyle \frac{k_i}{N-1}}$
Betweenness centrality	The sum of the shortest paths through a node	node	$BC(i)={\sum }_{i\ne s\ne t}{\displaystyle \frac{{\mathrm{n}}_{st}^i}{{\mathrm{g}}_{st}}}$
Closeness centrality	The sum of the distances between node i and all other nodes in the network	node	$CC(i)={\displaystyle \frac{{\sum }_{y}d(y,i)}{N-1}}$
Network closure index	The extent to which the network is looped	global	$\alpha ={\displaystyle \frac{L-V+1}{2V-5}}$
Dot rate	Number of connections for each node	global	$\beta ={\displaystyle \frac{L}{V}}$
Network connectivity index	The degree of connectivity of all nodes in the network	global	$\gamma ={\displaystyle \frac{L}{3(V-2)}}$

where $k_i$ refers to the degree of the ith node; N refers to the total number of nodes; ${\mathrm{g}}_{st}$ refers to the number of shortest paths between nodes s and t; ${\mathrm{n}}_{st}^i$ refers to the number of shortest paths between nodes s and t passing through node i; $d(y,i)$ refers to the cost distance between nodes y and i; L refers to the number of potentially important ecological corridors; V refers to the number of ecological sources.

.

3. Results

3.1. Spatial–Temporal Evolution of Landscape Pattern

The MSPA model delineated seven landscape types across the Yongding River floodplain (Figure 3), with the core area occupying the largest proportion (

Table 8. The results of MSPA landscape classification.

Landscape Type	1967		1980		2004		2021
	Area (km2)	Area Proportion	Area (km2)	Area Proportion	Area (km2)	Area Proportion	Area (km2)	Area Proportion
Core	138.18	23.77%	114.64	19.72%	92.33	15.88%	112.64	19.38%
Islet	6.10	1.05%	7.34	1.26%	8.42	1.45%	7.81	1.34%
Perforation	2.61	0.45%	1.05	0.18%	1.38	0.24%	2.66	0.46%
Edge	63.77	10.97%	56.79	9.77%	46.93	8.07%	54.15	9.31%
Bridge	27.62	4.75%	1.81	0.31%	2.20	0.38%	31.37	5.40%
Loop	1.69	0.29%	24.53	4.22%	23.52	4.05%	3.83	0.66%
Branch	16.63	2.86%	14.13	2.43%	16.38	2.82%	18.87	3.25%

). In 1967, core landscapes were concentrated along the river channel, largely due to the river’s substantial water surface area before the Yongding River dried up and the extensive, continuous vegetation cover along both banks. This continuous vegetation corridor provided habitat stability. By 2021, however, the core area had decreased by 4.39%, and its distribution became more fragmented, reflecting a shift towards a less contiguous spatial arrangement. The edge area comprised approximately 10% of the total floodplain area, ranking second in terms of area among landscape types. This stability in edge area size indicates that patches in the core area are relatively stable and exhibit resistance to external disturbances, though certain extent of patch fragmentation is evident. The area of connecting bridges increased from 4.75% in 1967 to 5.40% in 2021, underscoring enhanced landscape connectivity and improved conditions for species dispersal within the study area. Spurs, similarly essential for landscape connectivity, also grew in area share, supporting the development of a more robust ecological network. Roundabouts, isolated habitat patches with an area increase of 1.71 km² and a relatively fragmented distribution, may serve as potential stepping stones in future ecological planning, facilitating connectivity between ecological patches and alleviating the effects of habitat fragmentation on species migration. Traffic circles, which function as expedited migration routes [19], remained limited in area (<1%), reflecting higher resource consumption for migrating species and challenges to maintaining biodiversity across fragmented habitats. In summary, between 1967 and 2021, the Yongding River floodplain experienced significant shifts in landscape functions: the core area became smaller and more dispersed, fringe areas maintained stability but showed signs of patch fragmentation, and the area share of both connecting bridges and spurs expanded, enhancing connectivity. The observed changes in isolated habitat patches and traffic circles highlight ongoing challenges to species migration and ecosystem connectivity.

3.2. Evolution and Partial Recovery of Integrated Habitat Quality

Overall, the integrated habitat quality began at a moderate level but declined noticeably over time before partially recovering in later years. A similar trajectory emerged in the habitat quality and ecosystem service (Figure 4). For instance, the proportion of the highest category (Level 4) of habitat quality declined from 40.25% in 1967 to 29.35% in 2004, followed by a recovery to 36.31% in 2021. Ecosystem service exhibited a comparable pattern, with its highest category dropping from 37.52% (1967) to 27.44% (2004), then rising to 33.13% (2021). These changes are mirrored in their respective mean values, each undergoing an overall decrease until the early 2000s, then improving in response to conservation strategies.

Spatially, from 1967 to 1980, high-quality habitat areas were concentrated near the river channel (Figure 5), reflecting high biodiversity and a relatively stable ecological environment before the river was disrupted. Between 1980 and 2004, habitat quality around the river channel substantially declined, with a significant reduction in high-quality habitat area, attributed to environmental pressures from agricultural expansion and other anthropogenic activities. Between 2004 and 2021, habitat quality around the river channel showed signs of recovery, with an expansion of high-quality habitats particularly noticeable in the central and southern floodplain regions. This improvement indicates the positive impact of recent river corridor restoration and ecological recharge initiatives. Although the current habitat quality remains below levels observed before the river disruption, it has significantly improved compared to that in 2004. Spatial distribution mapping of habitat quality reveals a pattern of initial decline followed by partial recovery within the river channel. Concurrently, the increased area of low-quality habitats outside the river channel is closely associated with the expansion of construction land and intensified agricultural activity.

3.3. Evolution Characteristics of Ecological Source Areas

The evolution of core patches indicates that ecological source areas within the Yongding River floodplain have primarily consisted of water bodies and forests. With urbanization, however, city expansion and the increase in cultivated land have led to a marked decline in these areas. As shown in

Table 9. Number and area of ecological sources during 1967–2018.

Year	Core Ecological Patches		Important Patches
	Number	Area (km2)	Number	Area (km2)
1967	30	84.67	8	27.05
1980	31	60.16	8	14.79
2004	29	51.67	6	8.74
2021	30	65.41	9	14.55

, the total area of ecological source areas decreased from 84.67 km² in 1967 to 61.16 km² in 1980, further declining to 51.67 km² in 2004, before partially recovering to 65.41 km² in 2021, representing an overall reduction of 19.25 km². The largest single ecological source area significantly decreased from 32.57 km² to 11.71 km². This highlights substantial alterations in spatial configuration, indicating the fragmentation or spatial redistribution of these critical ecological areas. Parallel trends were observed for important patches, defined as patches meeting all three suitability criteria. The number of these important patches remained stable at 8 between 1967 and 1980, though their total area notably declined from 27.05 km² to 14.79 km². A further decline was observed in both the number and area of important patches in 2004, suggesting increasing pressures on these vital areas. By 2021, the number slightly increased, and their collective area partially recovered to 14.55 km², indicating some degree of ecological improvement or conservation success.

Spatially (Figure 6), the distribution of ecological source patches transitioned from being concentrated along the river channels to a more dispersed pattern. As river channel widths narrowed, the proportion of water bodies gradually declined, while the proportion of forests increased.

Analysis of changes in patch suitability—measured across area, connectivity, and habitat quality (Figure 7)—reveals a declining trend in the number and area proportion of suitable patches, indicating a gradual reduction in ecologically robust regions. Conversely, the area proportion of suitable patches in terms of connectivity increased, while the number and area proportion of moderately suitable and unsuitable patches decreased, suggesting potential improvements in ecological connectivity. Nevertheless, moderately suitable and unsuitable patches exhibited a rising trend in area proportion in terms of habitat quality and patch area. This trend was particularly pronounced in the habitat quality dimension, highlighting a significant intensification in habitat quality degradation. These findings underscore an urgent need for targeted conservation and restoration efforts to protect high-quality ecological patches and sustain ecosystem health.

3.4. Variation Characteristics of Ecological Resistance

Using land-use types and related geographic data, we calculated the minimum cumulative resistance across the floodplain from 1967 to 2021, reclassifying the resistance surface to create a comprehensive ecological resistance map. Figure 8 and Figure 9 illustrate that low-resistance value areas are primarily distributed within the river channel, where water bodies and beaches, characterized by strong ecological connectivity, exert minimal impact on species migration and energy flow. Protecting high-value natural woodlands in these areas by establishing protective forests would enhance ecosystem resilience. Medium-resistance value areas mainly consist of cropland and grassland, which serve as essential ecological buffer zones supporting biodiversity [46]. In contrast, high-resistance value areas primarily consist of construction land, which is spatially scattered and fragmented. To mitigate ecological fragmentation in the central floodplain, strategies such as selective building demolition and the establishment of ecological corridors can be implemented. From 1967 to 2021, the overall resistance value increased from 2.03 to 2.30, marking a 13.30% increase. This shift reflects a reduction in the area of the first-level resistance zone and a notable expansion of the fifth-level resistance zone, highlighting increased ecological stress and degradation risks within the Yongding River floodplain. High-resistance value areas with notable increases are concentrated near Gu’an County and the disrupted areas of the original tributaries, where urban growth and transportation infrastructure have intensified ecological resistance. Although resistance value within the Yongding River’s main channel has also risen, the increase remains relatively minor, and this area has consistently maintained a low-resistance value since 1967. To effectively safeguard and restore degraded ecosystems within the floodplain, stricter urban planning and ecological protection measures in high-resistance value areas are essential. These strategies will enhance ecological connectivity and support biodiversity conservation across the landscape.

3.5. Evolution and Restoration Requirements of Ecological Corridors

From 1967 to 2021, the number of ecological corridors within the Yongding River floodplain displayed a decreasing trend (

Table 10. Number of ecological corridors during 1967–2018.

Year	Corridor	Important Corridor	General Corridor
1967	63	37	26
1980	64	26	38
2004	57	14	43
2021	57	10	47

). Notably, the number of important corridors declined significantly from 37 in 1967 to 10 in 2021, indicating an increase in resistance to material–energy flow across the landscape. Spatially (Figure 10), after 1980, corridor distribution became uneven, with corridors concentrated primarily in the central and northeastern floodplain regions, reflecting relatively high forest and grass cover in these areas. In other regions, however, corridor coverage was sparse, resulting in reduced ecological connectivity. By 2004, most of the key ecological corridors within the river’s core area had nearly disappeared, a consequence of both river fragmentation and intensified human activity, contributing to severe ecosystem degradation during this period. In 2021, following the implementation of the ecological water replenishment project, the ecological environment improved slightly, leading to an increase in the number of critical ecological corridors within the river’s core area. This has partially restored ecosystem function and connectivity; however, overall connectivity remains limited, with the number of critical corridors still low. These findings underscore the need for continued ecological restoration and protection measures.

The setting of width thresholds for ecological corridors has a significant effect on the construction of ecological networks, particularly influencing the locations of pinch points within corridors [45]. Through iterative calculations of width thresholds, it was observed that as the cost-weighted corridor width increased, the corridors widened, while the locations of pinch points remained relatively stable. As shown in Figure 11, a corridor width of 500 m was identified as optimal, revealing distinct ecological pinch points, which were then used as the basis for final pinch point extraction. Obstacle points—primarily located within a 100 m search radius in the core area—were found to be spatially stable. As shown in Figure 12, a total of 13 ecological pinch points were identified across the Yongding River floodplain, covering a total area of 6.23 km². These pinch points are predominantly concentrated along the Yongding River channel, arranged in a bead-like pattern. Concentrations of pinch points are especially notable in the central region, where corridor lengths and migration resistance are higher, thereby functioning as essential “stepping stones.” Additional pinch points are dispersed along the Xinlong River tributary and Yuejin Canal. The degradation or loss of these pinch point areas can impair the connectivity of the entire ecological network [51], highlighting the need for their strict protection along the ecological corridors. The analysis also identified ecological obstacle points, totaling an area of 22.36 km², primarily distributed in the northern and eastern parts of the floodplain, often where road networks intersect with ecological corridors. These areas, with high resistance values, indicate an urgent need for ecological restoration efforts in urban areas with dense road networks. Overlay analysis with the comprehensive resistance surface reveals that obstacle points largely coincide with higher-resistance value areas, further challenging landscape connectivity and material flow across the floodplain. The overlap of ecological pinch points and obstacle points suggests that while these areas facilitate frequent material and energy flows, their high resistance values make them critical priorities for ecological restoration. By identifying and restoring these key areas within the Yongding River floodplain, the stability of ecological spaces can be reinforced, enhancing the capacity of river corridors to provide ecosystem services.

3.6. Evolution of Ecological Network Structure

Using graph theory, the structural evolution of the ecological network in the Yongding River floodplain from 1967 to 2021 was analyzed through multiple network indices, as shown in Figure 13. Over this period, the degree centrality index showed an upward trend, with a continuous increase in the number of significant nodes. This growth indicates a strengthening of the node control within the network and a more pronounced mutual influence among nodes. However, while the node influence grew, the control exerted by the core nodes over the network declined. The distribution of significant nodes gradually transitioned from initially dispersed to more concentrated, achieving a relatively balanced spatial distribution by 2021. The betweenness centrality index exhibited a fluctuating upward trend, signifying an increased frequency of nodes acting as bridges for information and resource flow, thereby improving opportunities for species supplementation from distant patches and enhancing network interconnectivity. Key nodes were primarily concentrated along river channels, underscoring the critical role of river corridors within the ecological network of the floodplain. The steady rise in the closeness centrality index indicated reduced efficiency and convenience in the flow of material, energy, and information across the network. This finding, alongside the decline in the number of key nodes, suggests a contraction in the network’s capacity for information dissemination. Additionally, the calculated network closure index (α) showed a continuous decrease, with an overall decline of 17.65%, pointing to fewer closed loops within the network and weaker internal connectivity. The dot rate (β) highlighted a high initial complexity within the ecological network, although this complexity gradually diminished over time, leading to a decrease in network stability. Similarly, the consistent decline in the network connectivity index (γ) further illustrated a reduction in the degree of ecological network connectivity. Collectively, these indices reveal that the Yongding River floodplain’s ecological network has become increasingly tight and centralized over time, accompanied by a decline in connectivity and structural stability. These observed shifts are largely driven by anthropogenic factors—such as urban expansion, agricultural development, and altered hydrological regimes—which influence habitat fragmentation and the spatial distribution of key corridors. This trend underscores the necessity of reshaping ecological source and corridor dynamics. Strengthening the protection and restoration of river ecosystems is essential for formulating robust ecosystem management strategies and sustaining the core ecological functions of the floodplain.

4. Discussion

4.1. Ecological Network Evolution and Associated Ecological Impacts

The ecological network in the Yongding River floodplain underwent substantial transformations from 1967 to 2021, driven by a combination of hydrological engineering, land-use changes, and subsequent ecological restoration efforts. Over this period, ecological source areas—primarily water and forests—have substantially decreased, leading to narrower river channels and increased landscape fragmentation. This contraction of ecological sources was particularly evident in the floodplain’s central regions, such as the Gu’an County area, where intensive urbanization and agricultural expansion have significantly reduced forest coverage and disrupted natural wetland habitats, which served as critical migratory pathways and habitats for local bird species (e.g., Egretta garzetta and Ardea cinerea) and other wildlife. As a result, corridor connectivity notably diminished, and multiple river segments became isolated due to reduced flows and poorly connected irrigation canal tributaries. Early reservoir construction and flood diversion projects significantly curtailed downstream water availability and contributed to drying out much of the floodplain [52]. This decline notably affected biodiversity, leading to the disappearance or drastic reduction in certain species. By around 1995, populations of previously abundant fish and migratory birds had completely vanished.

Recent interventions—such as ecological water replenishment from the Yellow River, targeted reforestation programs near Langfang and Gu’an, and levee reconstruction around the suburban Beijing sections—have partially alleviated habitat loss and improved connectivity. For instance, ecological water replenishment has increased aquatic habitat extent, partially restoring the seasonal migration of certain bird populations. In 2021, more than 20 fish species were detected in areas previously dry, and submerged plants such as Myriophyllum verticillatum L. were recorded alongside significant improvements in water quality [53]. However, the continuity and habitat quality necessary for sustaining stable wildlife populations remain compromised due to the fragmentation of remaining sources and inconsistent water replenishment cycles.

Graph theory analysis of the network structure reveals a trend towards denser but less stable connectivity within the ecological network, with declining overall network efficiency and stability. As a result, the fragmentation of ecological sources persists, restricting species migration and dispersal and ultimately weakening the capacity of floodplain ecosystems to provide vital services such as water regulation and biodiversity support. This shift toward smaller, more fragmented ecological source areas and weaker ecological corridors underscores the interplay between large-scale anthropogenic modifications—such as earlier reservoir operations and channelization—and more recent ecological compensation measures. Understanding the drivers and consequences of these network changes is essential for guiding future restoration policies, reinforcing ecological corridors, and maintaining the Yongding River floodplain’s ecological functionality over the long term.

Accurate identification of the current status of ecological networks in river corridors is fundamental for optimizing these networks. This study proposes an optimization method for ecological network identification and extraction based on MSPA and analyzes the factors affecting ecological sources from multiple perspectives, including morphology, area, positional importance, and habitat quality. This study offers a more scientific and comprehensive means of extraction compared to traditional single-method approaches, allowing for a better reflection of ecosystem composition and spatial configuration. By constructing and analyzing ecological corridors, we observed a notable reduction in overall network efficiency and the disappearance of critical ecological corridors. In addition, the ecological fracture zones between core source sites were identified, and the key defective areas within the ecological network were considered as priority areas for ecological restoration. By considering the geographical location of the river network prior to fragmentation, we emphasize active restoration of previously disappeared tributaries to enhance the connectivity of the existing water system and restore or reconstruct the floodplains. These efforts can create sufficient space for dynamic landscape development and establish a rich habitat encompassing the mainstream, tributaries, and shallow waters. Moreover, we also emphasize the establishment of a robust ecological source protection mechanism to effectively safeguard core ecological sources with significant areas and strong connectivity. This underscores the importance of ecological construction and environmental protection in the restoration efforts.

4.2. Cross-Regional Comparison and Future Perspectives on Ecological Management

Similar trends of ecological degradation and decreased connectivity have been observed in other urbanizing river basins, including sections of the Yangtze River and Yellow River in China, as well as Ganges and Brahmaputra River in Southeast Asia [54,55]. In these regions, expanding urban infrastructure has led to habitat fragmentation and reduced network efficiency, mirroring the ecological challenges observed in the Yongding River floodplain. The “Comprehensive Treatment and Ecological Restoration Master Plan for the Yongding River” emphasizes the restoration of the natural characteristics of the floodplain river segment as a priority. It advocates for measures such as river ecological restoration, beach management, and intensified efforts to convert farmland back to forests and grasslands. Despite these efforts, the restoration outcomes in the Yongding River have been relatively modest compared to other river basins, such as the Yangtze River and Yellow River basins, where more substantial recovery has been reported [55]. At present, the ecological functions of Yongding River are primarily sustained through human interventions [27,52]. This highlights both the feasibility and the limitations of large-scale ecological restoration efforts: while targeted management can mitigate anthropogenic impacts, sustained governmental support and integrated land-use planning remain critical for achieving long-lasting results.

Looking ahead, the priority lies in reinforcing important corridors and safeguarding high-quality patches to fortify the ecological network against further fragmentation. Given that the Yongding River floodplain is significantly impacted by human activities and exhibits characteristics such as system openness and ecological vulnerability, future studies should consider incorporating anthropogenic disturbances—such as population density—to enhance the ecosystem’s resistance factors. Effective corridor restoration requires not only the physical reconnection of river segments but also ongoing monitoring and adaptive management that can respond to dynamic land-use pressures. Additionally, the current longitudinal time-series analysis lays the groundwork for future land-use predictions and simulations based on scenario analysis. Future research could involve horizontal comparisons of different development scenarios to explore strategies for optimizing ecosystem structure and function within ecological spatial constraints.

5. Conclusions

River floodplains, essential for biodiversity and ecosystem services, are increasingly under threat from climate change and human activities, making them one of the most vulnerable ecosystems worldwide. The research introduces four temporal nodes to conduct an in-depth examination of ecological network evolution, presenting an innovative approach for identifying ecological sources. By integrating the MSPA method with landscape connectivity and habitat quality assessment, this approach optimizes the methodology for identifying and extracting for ecological networks. Focusing on the Yongding River floodplain, the study employed circuit theory to construct an ecological network, pinpointing critical fragmentation areas and examining ecological network evolution from 1967 to 2021. The findings indicate a fluctuating decline in ecological source area over this period, reaching their lowest in 2004, with partial recovery by 2021. The number and length of ecological corridors, especially first-level corridors, have also declined, resulting in decreased network connectivity and stability. As key nodes within the network, ecological pinch points and obstacle points are critical for preserving landscape structure and function; thus, these areas should be prioritized for restoration and protection to support the ecosystem’s sustainability. These results provide valuable insights for ecological planning and management, offering a reference for restoration and sustainable development strategies in the Yongding River floodplain.