Local response to global risk: a case study of risk perception and communication on Chinese social media regarding Fukushima's treated radioactive water discharge.

This study examines localized risk perception and communication concerning Fukushima's treated radioactive water discharge through analysis of 111,660 Weibo posts using large language model sentiment analysis and hierarchical linear modeling. Three key findings emerge: (1) Temporal dynamics dominate public engagement, with attention peaking within two weeks post-discharge before gradually declining, while fear persists as the predominant emotion despite diminishing anger responses; (2) Coastal regions demonstrate heightened participation due to geographic proximity and seafood consumption patterns, though risk perception shows no consistent spatial gradient; (3) Socioeconomic factors - particularly education levels and seafood consumption - significantly predict communication intensity, whereas economic indicators like per capita GDP show limited influence. The study proposes a three-pronged risk governance framework: temporal targeting of critical communication windows, spatially adapted messaging for coastal populations, and emotion-focused guidance to address persistent public fears. These findings illuminate how global environmental risks become localized through China's unique sociocultural filters, offering empirical support for precision policy interventions in transnational risk management.