Resilient ecological security patterns: A multi-layer network approach to support assessment under cascading failures in karst regions

In the ecologically fragile karst region of Hechi City, Guangxi, unique landforms and human activities have increased demand for multiple ecosystem services (ES), degrading ecological functions, weakening ecological network resilience, and highlighting the need to assess regional ecological security patterns (ESP). This study developed a multi-layer ecological network framework—integrating carbon sequestration, forest resources, and water conservation—and applied a cascading failure model to evaluate and improve ESP through the lens of spatial resilience in multi-layer ES. Ecological sources were identified via sensitivity, service importance, habitat quality, and morphological spatial pattern analysis, totaling 236 patches (12,604.28 km2, accounting for 37.57% of the land area of Hechi City), primarily located in northwestern Tian'e and Nandan. Multi-layer resistance surfaces, based on land use, topography, NDVI, and human activity, showed higher resistance in the northwest. Minimum cumulative resistance models extracted 618–629 corridors (among which 340 were corridors with common ecological intersection points), densely distributed along the Fengshan-Donglan-Bama borders. Resilience under node/edge attacks and cascading failure simulations revealed vulnerability to targeted disruptions, with 15 critical nodes (e.g., IDs 102, 230) reducing the largest connected component (LCC) below 0.2. Edge-based restoration targeting pinch-points and critical corridors effectively restored LCC >0.8 using ∼76 edges. Proposed Restoration Strategy 3, based on shared ecological pinch-point corridors, effectively balances cost-effectiveness and efficiency, with an average LCC increase of 0.0012 per 100 million yuan invested and 0.0067 per restored corridor edge. Five ecological restoration areas were proposed in Hechi City. This framework, by constructing a multi-layer ecological network as the structural representation of the ESP, provides a resilience evaluation method to support ESP design, promote ecosystem service synergies, and guide sustainable karst restoration.