Machine learning-based construction and network resilience assessment of wind-resistant tree species ecological...

Despite growing attention on urban ecological networks, limited research has focused on the spatial structure and resilience of networks centered on wind-resistant tree species, especially in typhoon-prone coastal cities. The construction of Wind-Resistant Tree Species Ecological Networks (WRTSEN) is essential for enhancing urban disaster resilience and maintaining ecosystem stability. Habitat patches of wind-resistant tree species were identified using a MaxEnt model based on potential suitability distributions. An ecological resistance surface for wind-resistant species in the Xiamen-Zhangzhou-Quanzhou (XZQ) region was generated using the XGBoost algorithm. Based on circuit theory, the WRTSEN was constructed, and its resilience was evaluated using complex network analysis. The results indicate that: (1) A total of 87 habitat patches were identified, covering 10.13% of the total area, primarily located in the eastern and southern coastal zones of the XZQ region, while sources in the central and western areas were more fragmented. (2) The resistance surface constructed using XGBoost demonstrated high accuracy and strong ecological consistency with observed spatial patterns. (3) Ecological corridors exhibited marked spatial heterogeneity: primary corridors were concentrated in central and northern Zhangzhou, secondary corridors were evenly distributed, forming a dense network, while coastal urban centers showed sparse and isolated connections. (4) Resilience assessments revealed that targeted attacks caused significantly faster degradation of network structure and function than random attacks. These findings underscore the need to prioritize natural reserves, water systems, and critical coastal habitat patches of wind-resistant species in urban planning. By integrating machine learning with ecological network construction, this research provides data support for optimizing regional ecosystem functions and informing sustainable development strategies in coastal regions.