Abstract: With the intensification of human activities in recent years, the structure and function of wetlands have been strongly disturbed and damaged. Understanding the changes of structure of wetland ecosystems after disturbance is of great significance to the conservation, management, and restoration of wetlands. We construct a network model of wetland ecosystem based on complex network related theory and ecosystem characteristics, and propose an evolutionary model of wetland ecological network structure considering overload and underload states based on the Shelford’s law of tolerance of organisms and the self-regulation ability of the ecosystem. The evolutionary process of an attacked ecological network structure is described in terms of initial load, load capacity, load redundancy range, and load redistribution, while robustness indicators are used to reveal the impacts of species on the robustness of wetland ecosystem after disturbance. For the Teacapan wetland ecosystem, numerical simulations revealed that the extinction of keystone species at different trophic levels had different degrees of impacts on ecosystem structure and robustness. The loss of key species in the primary producers (mangrove debris and material in the water) led to a rapid collapse of the Teacapan wetland ecosystem network structure and a rapid decline in ecosystem robustness index to zero, while the loss of key species in the consumers had less impacts on the fluctuation of ecosystem network structure and system robustness, and the ecosystem could reach a secondary equilibrium within a certain period of time. At this time, wetland ecosystem still has some ability to resist external disturbances. The results demonstrate the importance of mangrove detritus as a major food source supporting food chain, and provide a scientific reference for the conservation and ecological reconstruction of wetland ecosystems.

Key words: wetland ecosystem, complex network, robustness, evolutionary model.