Revealing the alterations of social-ecological system networks by China's national parks: Evidence from the Qinghai-Tibet Plateau, China

National parks are pivotal instruments for advancing biodiversity conservation and fostering regional sustainable development. Yet, against the backdrop of intensifying climate change and growing anthropogenic pressures, the mechanisms by which national park governance reshapes coupled social-ecological systems (SES) remain insufficiently elucidated. Taking Qinghai Province, China, as a case study, this research integrates multi-source data to systematically investigate county-level ecological dynamics from 2005 to 2023. We employed a mixed-methods framework combining social network analysis (SNA), machine learning, and partial least squares structural equation modeling to quantify spatiotemporal variations in ecological importance, trace the dynamic evolution of SES linkage networks, and disentangle the direct and indirect pathways through which natural and anthropogenic factors drive SES dynamics. The results indicate that the evolution of SES shows a general trend of growing connectivity and stronger coupling. The linkage between economic and ecological components is positively shaped by a topography-soil-vegetation cascade, with standardized path coefficients of 0.72, 0.73, and 0.90, respectively. Compared with non-park areas, national park areas showed enhanced network structural cohesion, with policy-induced increases of 40% in network density, 35.16% in average clustering coefficient, and 70.89% in average weighted degree. Meanwhile, natural processes contribute to a more notable rise in coupling intensity. These findings advance the theoretical understanding of the socio-ecological impacts of protected area governance and provide a rigorous scientific basis for guiding the sustainable transition toward integrated ecological protection and socio-economic development at the regional scale.