Abstract: Public health crises have presented evolving challenges and opportunities for tourism management. By integrating ground survey data with social media big data using Bayesian Network modeling, we analyzed how the pandemic changes in visitor engagement with forest environments in national park from 2019 to 2023 based on 45,007 unique records. Results show increased sensitivity of nature and forest-based attributes and decreased sensitivity of infrastructure to overall satisfaction in 2020 and 2021. Educational level, income, and age were key demographic factors associated with satisfaction. Four scenario analyses explored outcomes of hypothetical visitor shifts and management interventions, while backpropagation analyses identified efficient pathways to optimal satisfaction, with park infrastructure and hospitality yielding the greatest marginal benefits. This study supports data-driven strategies to increase park ecological and operational resilience, enhance visitor experience and loyalty, and inform adaptive park management for the new normal and future public health crises.

Key words: National park planning and management, Public health crises, Forest resources management, Human-nature interactions, Big data and AI analytics, Bayesian network modelling