Analysis of Marine Protected Area Networks and Spatial Prioritization in China: Integrating Ecological Connectivity with Anthropogenic Pressure

Junjing Cui^1*Minghao Wu¹Qiaoyun Sun¹Congcong Huo²

• ¹School of Architecture and Urban Planning, Shenzhen University, Shenzhen, China
• ²Guangdong Vocational Institute Of Public Administration, Guangzhou, 510812, China, Guangzhou, China

In recent years, the combined pressures of biodiversity loss, climate change, and human activities have become as major challenges facing global marine ecosystems (Wheeler et al., 2013). Marine systems are experiencing substantial, widespread, and dynamic cumulative impacts from human activities, profoundly affecting the structure, function, and stability of marine ecosystems. Some regions even face the risk of systemic degradation (Halpern et al., 2014;Rocha et al., 2015;Wu et al., 2024). Rapid economic development, coastal construction, and intensive resource exploitation have placed substantial stress on marine ecosystems along China's coastal areas, highlighting the urgent need to optimize the spatial configuration of marine conservation at the national scale (Bohorquez et al., 2021;Zeng et al., 2022).Marine Protected Areas (MPAs) have the potential to protect and restore marine ecosystems, enhance resilience to climate change, and generate socio-economic benefits through healthy marine environments (Leary et al., 2018;Brander et al., 2020;Pascal et al., 2018). The establishment and effective management of MPAs are widely regarded as a key approach to address marine ecosystem degradation, strengthen ecosystem resilience, and promote sustainable marine resources use (Pimm et al., 2018;Thomas et al., 2015;Yang et al., 2020). China's marine conservation efforts have spanned decades and been further integrated recent, resulting in a multi-type and multitiered MPA system comprising marine nature reserves, special marine protected areas, marine parks, and fishery resource reserves (Miao & Chen, 2025;Zhao et al., 2022). Coastal provinces have widely established MPAs to achieve diverse ecological, cultural, and economic objectives. However, existing studies indicate that increased protected area coverage do not necessarily translate into improvements in overall conservation effectiveness (Bohorquez et al., 2021;Zeng et al., 2022). China's MPAs continue to exhibit spatial challenges, including regional imbalance, fragmentation, and insufficient ecological connectivity (Hu et al., 2020;Bohorquez et al., 2021). International databases remain incomplete and delayed in recording and updating Chinese MPAs, which to some extent constrains the assessment of their overall effectiveness and international comparability (Day et al., 2012). In practice, conservation planning has often prioritized area-based targets, while comparatively less attention has been given to ecological representativeness, spatial configuration quality, and cross-regional connectivity. This results in inadequate coverage of key habitats and biogeographic units (Ma et al., 2013;Li et al., 2017;Bohorquez et al., 2019).Ecological connectivity is widely recognized as one of the core attributes determining the conservation effectiveness and long-term stability of MPAs (John et al., 2021). Connectivity not only supports species dispersal and genetic exchange, but also provides a critical foundation for ecosystems to respond to climate change and cumulative human disturbances (John et al., 2021;Chen et al., 2024;Gownaris et al., 2019;Fischer et al., 2019;Hilty et al., 2020). Compared with protected area extent alone, connectivity is more sensitive to external pressures and tends to decline more rapidly, thereby weakening the overall functioning of conservation systems (Grande et al., 2020;Liang et al., 2021). Consequently, research perspectives have gradually shifted from single-site quality assessments toward an emphasis on cross-patch and cross-regional ecological network (EN) structures and their implications for system robustness (Chen & Xu, 2024a). International policy frameworks, including successive targets under the Convention on Biological Diversity, the "30×30" initiative of the Kunming-Montreal Global Biodiversity Framework, and the United Nations Sustainable Development Goal 14 (SDG 14), emphasize expanding and optimizing MPA networks to achieve long-term protection and sustainable use of marine resources (CBD, 2010;Toonen et al., 2013).Methodologically, techniques for identifying EN components and potential corridors are well-established, including the minimum cumulative resistance (MCR) model, gravity models, ant colony optimization algorithms, and circuit theory (Yang et al., 2024;Peng et al., 2019;Qian et al., 2023). However, existing studies have largely focused on network construction itself, while systematic analyses of the relationships among network structural characteristics, functional performance, and responses to disturbances remain limited. In particular, quantitative assessments of network robustness under random and targeted disturbance scenarios are still scarce (Chen et al., 2024;Sun et al., 2024;Gao et al., 2024). Meanwhile, recent policy and governance frameworks emphasize the need to systematically identify and mitigate cumulative human pressures within marine spatial planning and the development of protected area systems. The combined effects of fishing activities, climate change, and ocean-based and land-based pressures play a significant role in shaping resistance patterns along ecological corridors, thereby strongly influencing the realized connectivity of MPA networks (Zeng et al., 2022;Chen et al., 2023). Nevertheless, integrated national-scale quantitative assessments of the relative contributions of these pressures and their impacts on ecological network structure and stability across different sea regions remain lacking.At the intersection of these research gaps and management needs, network-level evidence on China's MPA networks remains limited. This scarcity is particularly evident in the following aspects: (1) regional differences in MPA network structure and spatial heterogeneity across sea regions; (2) the vulnerability of network structures; (3) the dominant roles of different types of human pressures in shaping resistance patterns and bottleneck segments along ecological corridors; and (4) spatial prioritization strategies aligned with policy objectives. To address these issues, this study constructs a national-scale MPA EN in China's seas by integrating network structural analysis, robustness scenario simulations, and cumulative human pressure decomposition, aiming to address the following questions and derive actionable management insights: Q1: What regional differences exist in China's MPA EN in terms of scale, clustering/modularity, and hub-bridge structural attributes? Q2: How sensitive are network connectivity and efficiency under targeted and random disturbance scenarios, and where are the key nodes and potential breakpoints located? Q3: How do fishing, climate change, and ocean-based and land-based sources pressures shape corridor resistance patterns, and what are the spatial characteristics of per-unit-length resistance and bottleneck segments? By integrating network structure, system robustness, and cumulative human pressure analyses at the national scale, this study aims to provide quantifiable and reproducible evidence to support the spatial optimization and prioritization of China's MPA networks, while also offering a Chinese case study and methodological reference for connectivity-centered marine spatial conservation and adaptive governance. This study includes 3 components: (1) identification of an EN based on MPAs; (2) evaluation of the EN's structural characteristics using complex network theory; and (3) assessment of EN robustness through attack simulations, with comparisons of differences across sea regions. China manages marine special protected areas in its internal waters, territorial seas, contiguous zones, exclusive economic zones, continental shelves, and other jurisdictional waters, implementing effective protection measures and scientific development approaches for regions with unique geographical conditions, ecosystems, biological and non-biological resources, or special marine utilization needs. This study extracted 257 marine-related protected areas, including 110 national-level and 147 local-level reserves, encompassing coastal geological parks, nature reserves, marine parks, and wetland parks (Fig. 1). For regional comparisons, these areas were categorized into four seas based on their locations: the Bohai Sea(B), Yellow Sea(Y), East China Sea(EC), and South China Sea(SC), to analyze differences in EN structure and stability. The resistance surface quantifies the difficulty of organism or ecological flow movement (Beier et al., 2011) or the relative cost of traversing a gridded map (Belote et al., 2016). Many studies construct resistance surfaces based on human impact intensity, naturalness, or similar indicators (Dickson et al., 2017;Correa et al., 2017). In this study, we developed a resistance surface using 14 anthropogenic stressors from four primary categories: fishing, climate change, ocean-based, and land-based. These include:• Fishing: commercial demersal destructive, commercial demersal nondestructive high bycatch, commercial demersal nondestructive low bycatch, pelagic high bycatch, pelagic low bycatch, artisanal fishing.• Climate change: sea surface temperature, sea level rise.• Ocean-based: commercial shipping, invasive species, ocean-based pollution.• Land-based: nutrient pollution, organic chemical pollution, direct human impacts.The datasets for these 14 stressors estimate cumulative human impacts on oceans, with values ranging from 0 to 1 at a 1 km spatial resolution. Global marine data inherently involve uncertainties, representing the best available data on human impacts in global seas (Halpern et al., 2019). We used the graph theory-based Graphab software to generate the EN and its topological structure, which is widely applied for identifying ENs and assessing connectivity (Foltete et al., 2014). Nodes represent individual MPAs, and edges represent potential ecological corridors. Based on the identified EN, we employed NetworkX to compute the topological structure, selecting six indicators reflecting connectivity and centrality to evaluate the EN's structural features (Gao et al., 2024;Ji et al., 2024). The meanings and applications of these indicators are described below. Connectivity describes the linkage states between network nodes. We adopted 2 connectivity indicators: average degree (reflecting local connectivity and overall network density or sparsity) and average clustering coefficient (indicating the likelihood of nodes forming local clusters and serving as a proxy for local robustness).Centrality measures node importance and influence. We selected four centrality indicators: degree centrality (measuring the number of direct connections, reflecting direct interaction capability); closeness centrality (assessing average shortest path distances to other nodes, describing overall accessibility); betweenness centrality (quantifying intermediary roles in paths between other nodes, indicating ecological bridging functions); and harmonic centrality (measuring the harmonic mean distance to all other nodes).We conducted categorical statistical analyses on these indicators to assess the influence of different stressor categories on network structural characteristics and reveal differences. Analysis of variance (ANOVA) was used to compare the four centrality indicators across categories. For connectivity, contributions to overall network density and local clustering were calculated using average degree and average clustering coefficient, respectively, reflecting global transmission efficiency. Corridor resistance values were also used to quantify resistance in EN corridors. Connectivity robustness refers to a network's ability to maintain connections under node disturbances or attacks, while vulnerability robustness indicates an ecosystem's capacity for self-regulation and recovery of ecological flows post-disturbance. Under various attack scenarios, the EN's average degree and global network efficiency were used as key metrics for these aspects (Men et al., 2024;Yang et al., 2024).Global network efficiency, representing the average shortest corridor length between all nodes, was selected for vulnerability assessment. It reflects overall information transmission efficiency and connection strength, revealing the network's ability to sustain functionality under partial or complete damage.Contributions to connectivity and vulnerability robustness across sea regions were calculated as follows:□□ □□ = ∑ | (□□ □□ -□□ □□-1 ) □□ 0 × 100%| □□ □□=1where □□ □□ is the cumulative percentage change in indicator □□ (average degree for connectivity robustness or global efficiency for vulnerability robustness) after removing the □□-th node; □□ is the total number of nodes in the EN; □□ □□ is the indicator value after removing the □□-th node;□□ □□-1 is the value after removing the (□□ -1)-th node; and □□ 0 is the initial indicator value before any node removal.In a Python 3.12 environment using NetworkX, we simulated network stability under random attacks (random node removal) and targeted attacks (node removal in descending order of degree). The constructed MPA EN identified 309 nodes and 562 corridors (Fig. 2), with a total node area of 39,233.73 km² (Table S1). Among the four seas, the Bohai Sea had the largest total node area, while the South China Sea had the most nodes (Fig. 3). Specifically, the Bohai Sea contains 45 nodes covering 17,924.88 km² , accounting for 45.69% of the total area with an average node size of 398.33 km² . The Yellow Sea has 74 nodes spanning 7,689.15 km² , representing 19.60% of the total area with an average node size of 103.91 km² . The East China Sea has 57 nodes covering 6,843.82 km² , accounting for 17.44% of the total area, with an average node size of 120.07 km² . The South China Sea has 133 nodes covering 6,775.88 km² , accounting for 17.27% of the total area, with an average node size of 50.95 km² . The total corridor length was 19,622.94 km, with the South China Sea having the longest total length and the most corridors (Table S2). Specifically, the Bohai Sea has 78 corridors with a total length of 2,681.80 km, accounting for 13.67% of the total, with an average length of 34.38 km. The Yellow Sea has 129 corridors with a total length of 4,845.11 km, accounting for 24.69% of the total, with an average length of 37.56 km.The East China Sea has 101 corridors with a total length of 4,203.67 km, accounting for 21.42% of the total, with an average length of 41.62 km. The South China Sea has 240 corridors with a total length of 7,168.21 km, accounting for 36.53% of the total, with an average length of 29.87 km. Additionally, there are 14 corridors spanning the four major sea areas, totalling 724.15 km in length, accounting for 3.69% of the total, with an average length of 51.73 km (Fig. 4).MPA EN exhibited significant regional variations. The Bohai Sea has fewer nodes but larger average areas, while the South China Sea has the most nodes but the smallest average areas. The South China Sea leads in both the number and total length of corridors, though its average corridor length is shorter. The Bohai Sea has fewer corridors, but their average length is relatively balanced. This study analyses structural differences in nodes and corridors across four major marine regions to thoroughly examine the structural characteristics of MPA EN. Six key indicators were calculated for the overall network and each sea to assess node roles in the EN, evaluating differences in MPA connectivity and structure and identifying influential nodes (Table S3).In terms of connectivity, the overall network had a degree of 3.638 and average degree of 1.270. the East China Sea (EC) nodes exhibit the highest degree (3.667) and average degree (1.547). The Bohai Sea (B) and Yellow Sea (Y) share the lowest degree (3.622), while the South China Sea (SC) has the lowest average degree (1.087). , indicating superior local connectivity in the East China Sea (Fig. 5).The overall clustering coefficient was 0.487. Among the nodes, Bohai Sea (B) had the highest value (0.562), followed by South China Sea (SC) (0.514) and Yellow Sea (Y) (0.445), while East China Sea (EC) had the lowest (0.420). This indicates that nodes in the Bohai Sea exhibit a clustered characteristic (Fig. 5). For centrality, degree centrality was uniform at 0.012 across the overall network and seas. The East China Sea had the highest closeness centrality (0.061) and betweenness centrality (0.109), exceeding the overall values (0.047), highlighting its dominant intermediary role. the Bohai Sea (B) had the lowest closeness centrality (0.039) and betweenness centrality (0.010). The South China Sea (SC) exhibited the highest harmonic centrality (29.688), reflecting relatively close distances between its nodes (Fig. 6 and7). To validate regional differences and identify core nodes, we ranked the top 10 nodes for four centrality indicators (Table 1). Node 261 (SC) exhibited the highest degree centrality, followed by nodes 47 and 10 (B). The top 10 nodes were distributed across the South China Sea (5 nodes: 261, 207, 265, 189, 296), Bohai Sea (3 nodes: 47, 10, 65), and Yellow Sea (2 nodes: 48, 42). The South China Sea had the most nodes, indicating its advantage in direct connectivity.The nodes with the highest closeness centrality are nodes 163 and 153 (EC), followed by nodes 164 and 165 (EC). All top 10 nodes are located in the East China Sea. The node with the highest betweenness centrality is 189 (SC), followed by 153 and 152 (EC). The top 10 nodes are primarily distributed across the East China Sea (EC,9 nodes: 153,152,168,173,176,175,149,142,137) and the South China Sea (SC, 1 node: 189). This reflects the concentrated advantage of East China Sea nodes in overall network accessibility and their mediating role within the network.The node with the highest harmonic centrality is node 47 (B), followed by 261 (SC) and 48 (Y). The top 10 nodes are distributed across the Bohai Sea (B, 4 nodes: 47, 10, 44, 52), South China Sea (SC,3 nodes: 261,246,277),and Yellow Sea (Y,3 nodes: 48,42,58). This indicates that nodes in the Bohai and Yellow Seas are relatively close in distance, locally clustered, and tightly connected.Overall, the East China Sea excelled in average degree, closeness and betweenness centrality; the Bohai Sea in clustering coefficient; and the South China Sea in harmonic centrality. These patterns were mirrored in top node rankings, underscoring regional heterogeneity. Corridor resistance values were calculated to quantify connectivity barriers in the MPA EN, with further decomposition into four stressor categories (fishing, climate change, ocean-based, land-based) to reveal heterogeneity and sources. To account for length variations, per-unit-length resistance values were computed for better comparison (Table S4).Total corridor resistance was 19,179.76, with an average of 34.13. The total resistance values of the Yellow Sea and East China Sea corridors were at 5,195.42 and 4,416.61, respectively, but their average values were higher at 40.27 and 43.73, indicating greater resistance between nodes. The Bohai Sea had the lowest total resistance value at 2,567.40. Inter-sea corridors had a total of 636.76 but the highest average (45.48), indicating significant cross-sea resistance (Fig. 8). By stressor source, ocean-based pressures dominated (40.54%), followed by climate change (31.50%), fishing (24.66%), and land-based (3.30%). Fishing and climate change caused the highest per-unit-length resistance in the Yellow Sea (0.29 and 0.35). Ocean-based and land-based pressures were highest in the East China Sea (0.47 and 0.05) (Fig. 9). The EN was primarily influenced by ocean-based pressures and climate change. The South China Sea had the most corridors and highest total resistance, while the Yellow and East China Seas showed higher average resistance, driven by fishing/climate change and ocean-/land-based stressors, respectively. To evaluate the MPA EN robustness, we simulated changes in connectivity and network efficiency under targeted attacks and random attacks, identifying breakpoints where connectivity and network efficiency sharply declined. EN connectivity robustness was assessed via changes in average degree under targeted and random attacks. Targeted attacks had a more pronounced impact: after removing 10% of nodes, average degree decreased by 32.95% (from 3.638 to 2.439); after 50%, it dropped by 82.97% (to 0.619). Random attacks were milder: 10% removal led to a 7.64% decline (to 3.360), and 50% to 51.40% (to 1.770) (Fig. 10). Similarly, EN efficiency robustness showed greater sensitivity to targeted attacks. Under targeted attacks, 10% node removal reduced efficiency by 71.95% (from 0.095 to 0.027), and 50% by 95.02% (to 0.005). Random attacks caused a 5.40% drop after 10% removal (to 0.090) and 72.66% after 50% (to 0.026) (Fig. 11). Network efficiency exhibited breakpoints-nodes whose removal caused abrupt efficiency drops-under both attack types, highlighting vulnerabilities (Table 2). For targeted attacks, the South China Sea had 4 breakpoints (219,261,277,189), while the Bohai Sea and Yellow Sea each had 3 (B: 65, 10, 47; Y: 42, 105, 104). For random attacks, the South China Sea had 4 (264,195,266,245),Yellow Sea 3 (97,116,71),East China Sea 2 (172,135), and Bohai Sea 1 (11). China's MPA EN exhibits pronounced regional heterogeneity in both structure and function. This finding is highly consistent with recent studies highlighting spatial fragmentation and regional imbalance in China's MPA system (Hu et al., 2020;Bohorquez et al., 2021;Zeng et al., 2022). From the perspective of network scale and structural metrics, the MPA network as a whole demonstrates modular and small-world characteristics (Barabá si, 2009), while individual sea regions exhibit differentiated balances between local clustering and cross-regional connectivity.The northern seas (the Bohai Sea and the Yellow Sea) are characterized by relatively high clustering coefficients (with the Bohai Sea reaching a maximum of 0.562), but lower degree/average degree, closeness centrality, and betweenness centrality, forming a "high-clustering-low-bridging" structural pattern. This configuration indicates strong internal connectivity and stability at the local scale, but limited capacity for transmitting ecological processes across regions. MPAs in northern China are densely distributed along the coast and are strongly influenced by the enclosed or semi-enclosed geographic settings; the network results of this study provide structural evidence supporting this spatial pattern (Hu et al., 2020).In contrast, the East China Sea exhibits the highest average degree, closeness centrality, and betweenness centrality, but a relatively low clustering coefficient, reflecting a typical "low-clustering-high-bridging" structure. This configuration enables the East China Sea to function as a critical hub and bridge within the national MPA network, exerting a decisive influence on ecological connectivity among different sea regions. This result emphasized the key role of the East China Sea in China's MPA spatial coverage and biogeographic transitions, and it also underscores its strategic importance in management and spatial planning (Bohorquez et al., 2021).The South China Sea contains a large number of nodes, short corridors, and pronounced local clustering, but exhibits the lowest overall average degree, forming a network characterized by multiple nodes with strong local clustering but overall sparsity. While such a structure may enhance redundancy for local ecological processes to some extent, it may also create potential bottlenecks at broader, cross-regional scales. Zeng et al. (2022) reported that MPAs in the South China Sea are numerous but highly diverse in type and governed by multiple authorities; the complexity of this spatial structure is quantitatively captured by the network analysis in this study.Overall, substantial differences exist among sea regions in terms of clustering, bridging capacity, and node importance. The Bohai-Yellow Sea region exhibits high clustering and low bridging, the East China Sea shows low clustering and high bridging, and the South China Sea contains the largest number of nodes with strong local clustering effects. Network metrics themselves should not be interpreted as inherently "good" or "bad"; rather, they reflect structural trade-offs between local robustness and cross-regional connectivity in different regions. By revealing these trade-offs at the network-structural level, this study complements existing research on China's MPAs that has largely focused on management effectiveness and coverage (Zhao et al., 2022;Chen et al., 2023), and underscores the necessity of developing region-specific conservation strategies tailored to local structural characteristics. Network structural heterogeneity is further reflected in the different responses of the MPA network to disturbances. Robustness analysis indicates that network connectivity declines significantly faster under targeted disturbances than under random disturbances, suggesting that network functionality is highly sensitive to a small number of key nodes. This characteristic has been widely discussed in recent studies on the robustness of complex systems and conservation networks (Chen et al., 2024).Under targeted disturbance scenarios, several network breakpoints overlap with high-centrality nodes (top 10), indicating that damage to key hub or bridge nodes can rapidly reduce network efficiency and overall connectivity. This pattern is particularly pronounced in the East China Sea, where a high-bridging structure enhances overall network efficiency but simultaneously increases vulnerability to targeted disturbances. This finding argues that "high-connectivity regions are often associated with elevated systemic risk" (Chen & Xu, 2024b).In contrast, under random disturbance scenarios, breakpoints do not coincide with high-centrality nodes, and network degradation is driven more by cumulative structural damage than by the failure of individual nodes. This result highlights that different disturbance types correspond to distinct risk mechanisms: targeted disturbances reveal the critical importance of key nodes, whereas random disturbances reflect deficiencies in overall structural redundancy and balance.By synthesizing the spatial distributions of high-centrality nodes and breakpoints (see Table S5), areas of high management priority can be identified. The East China Sea contains the largest number of key nodes (15), underscoring its hub status within the national network and its associated systemic risks. The South China Sea follows with 12 key nodes, reflecting the presence of several bottlenecks within its multi-node structure. Although the Bohai Sea and the Yellow Sea rely more heavily on strong local clustering, they nonetheless exhibit a degree of vulnerability under targeted disturbances. These regional differences further support the governance recommendations, which advocate a shift in China's MPA management from "average improvement" toward a "key-node-first" strategy (Zhao et al., 2022;Chen et al., 2023).Overall, by integrating network structural analysis with disturbance scenario simulations, this study identifies key nodes and potential vulnerable areas within China's marine MPA network at the national scale. This network-level evidence complements previous research that has primarily focused on management institutions and single-site evaluations, and provides quantitative support for optimizing MPA spatial configurations and enhancing system resilience. The drivers affecting MPA ecological networks are characterized by multiple sources and pronounced regional unevenness. Ocean-based pressures dominate corridor resistance across all sea regions (approximately 35.84%-48.51%), followed by climate change and fishing pressures, while land-based pressures generally contribute a relatively small proportion (typically below 5%). These differences indicate that pressures associated with marine activities exert particularly strong constraints on ecological connectivity at both nearshore and offshore scales (Zeng et al., 2022;Chen et al., 2023). After normalizing corridor resistance by unit length, clearer structural contrasts emerge among sea regions. The South China Sea bears the highest total resistance burden due to its large number of ecological corridors and extensive network scale. In contrast, the Yellow Sea and the East China Sea exhibit relatively high average resistance per unit length, driven primarily by fishing and climate pressures in the Yellow Sea, and by combined ocean-based and land-based pressures in the East China Sea. The Bohai Sea shows comparatively low overall resistance, but with a disproportionately high contribution from ocean-based pressures. Cross-regional corridors exhibit the highest resistance per unit length and are predominantly driven by ocean-based pressures. These patterns reflect substantial differences among sea regions in resource use practices, development intensity, and governance contexts (Hu et al., 2020;Zhao et al., 2022).Integrating these pressure factors with ecological network structural characteristics provides further insight into the core constraints faced by China's MPA network in different regions, thereby enabling the formulation of region-specific optimization strategies:• For the Bohai-Yellow Sea region, the network structure is characterized by a "high-clustering-low-bridging" pattern, with strong local connectivity but limited cross-regional linkage capacity. While this structure helps maintain local ecological stability, it constrains the continuity of ecological processes across sea regions. Given the relatively high contribution of ocean-based pressures in this region, this study suggests maintaining internal network integrity while selectively strengthening functional connections with the East China Sea, and reducing ocean-based pressures along key corridors through refined spatial management measures (Hu et al., 2020;Bohorquez et al., 2021). • As a critical bridging region within the national MPA network, the East China Sea exhibits a "low-clustering-high-bridging" structure, contributing disproportionately to overall network connectivity while remaining highly sensitive to disturbances. Corridor resistance in this region is mainly driven by the combined effects of ocean-based and land-based pressures, reflecting high human activity intensity and concentrated spatial conflicts (Chen et al., 2023). Accordingly, conservation strategies should prioritize nodes and corridors with high betweenness and closeness centrality, reduce potential breakage risks by increasing cross-regional redundancy, and alleviate compounded coastal pressures (Zhao et al., 2022). • Cross-regional ecological corridors play a critical role in facilitating inter-regional connectivity within the national MPA network, yet they exhibit the highest resistance per unit length and low redundancy, indicating pronounced vulnerability. As their resistance is primarily driven by ocean-based pressures, this study recommends prioritizing cross-regional corridors for coordinated inter-regional governance, strengthening low-redundancy segments through joint planning and targeted reinforcement to improve overall network connectivity. This strategy provides an important complement to existing conservation practices that are largely structured around administrative boundaries (Zhao et al., 2022;Chen et al., 2024).Overall, by coupling network structure with cumulative pressure analysis, this study provides quantitative evidence supporting the transition of China's MPA system from an emphasis on area expansion toward structural optimization and priority-based conservation, thereby enriching recent research frameworks on marine spatial governance in China. Although this study systematically constructed and analyzed China's MPA EN at the national scale, several limitations should be acknowledged. Uncertainty is unavoidable in the construction of resistance for ecological corridors, as the weighting of different pressure factors may influence estimates of dispersal probability and least-cost distances. While previous studies have extensively discussed approaches to resistance surface construction (Hu et al., 2020;Chen et al., 2024), no unified standard has yet been established for MPA network assessments, which to some extent constrains the comparability of results. Existing research suggests that incorporating policy scenarios and management adjustments into spatial analyses can enhance the practical applicability of MPA network optimization strategies (Zhao et al., 2022;Chen et al., 2023). Future studies could therefore introduce multi-scenario simulations to evaluate changes in network structure and function under alternative development pathways. In addition, integrating monitoring data with higher temporal resolution to dynamically update fishing intensity, marine resource exploitation, and climate-related variables would further improve the accuracy of network robustness analyses and spatial prioritization assessments. Such efforts would provide more forward-looking scientific support for the long-term adaptive management of China's MPA system. This study, through EN analysis of MPAs in China, elucidates overall network performance and regional differences in nodes and corridors, offering insights for future ecological conservation and restoration. The constructed EN identified 309 nodes and 562 corridors, with Bohai MPAs showing strong local clustering, Yellow Sea MPAs exhibiting moderate connectivity, East China Sea MPAs serving as key bridges, and South China Sea MPAs characterized by numerous but sparse linkages. Notably, the Bohai Sea led in clustering coefficient (0.562), reflecting strong local aggregation; the East China Sea demonstrated high average degree (1.547), closeness centrality (0.061), and betweenness centrality (0.109), indicating robust bridging functions; and the South China Sea had elevated harmonic centrality (29.688) but low average degree (1.087). Robustness simulations under random and targeted attacks underscored network vulnerabilities, providing a basis for prioritizing corridor enhancements and node protections. These findings support targeted MPA expansions to achieve global biodiversity framework goals, emphasizing regional adaptability for sustainable marine conservation in China.