The intensive development and exploitation of coastal zones have made these environments highly sensitive, with extreme marine dynamic processes, such as tropical cyclones, exacerbating the vulnerability of marine ecosystems. Marine Protected Areas (MPAs), as critical tools for preserving marine habitats and biodiversity, play a vital role in enhancing resilience to extreme climate events. Strengthening the adaptive capacity of MPAs is key to achieving adaptive governance in complex socio-ocean systems. This study, using the Chinese White Dolphin National Nature Reserve in the Pearl River Estuary, Guangdong Province, China, as a case study, aims to clarify the mechanisms by which extreme climate change affects the resilience of socio-ocean systems within MPAs and to identify key driving factors.
Building on the SETs (Social-Ecological-Technical) framework, this research employs system dynamics modeling, agent-based modeling (ABM), and machine learning methods to establish an assessment model for the resilience processes of socio-ocean systems under the impact of extreme climate change. Specifically, system dynamics models are used to analyze the complex relationships between ecological system dynamics and socioeconomic elements under climate disturbance within the reserve. The study focuses on the behavioral changes of different stakeholders, such as residents, managers, and tourists within the reserve, using ABM to simulate individual decision-making in response to climate change shocks and assess their impact on overall system resilience. Additionally, the LASSO regression algorithm is applied to evaluate key factors influencing resilience and predict future resilience trends.
The findings reveal the impacts of sea level rise, extreme temperatures, and storm surges on dolphin populations, fishery resources, and reserve management strategies, providing scientifically informed management recommendations for adapting to climate change.
 

machine learning,climate change,system dynamic,Resilience of SES,MPAs