Unveiling nature-based tourism destinations on the Qinghai-Tibet plateau: A journey-based semantic embedding framework

Identifying tourism destinations is a prerequisite for governing them, yet prevailing static paradigms, grounded in spatial concentrations of stopover behavior, systematically exclude the destinations that matter most for conservation planning. This study operationalizes a flow-based view of destinations through a journey-based semantic embedding framework and applies it to the Qinghai-Tibet Plateau (QTP), a region of exceptional ecological significance. Using GPS trajectory data from an outdoor recreation platform, the framework identifies nature-based tourism destinations (NBTDs) by detecting collective travel patterns rather than point-level density concentrations. Three findings advance both theory and policy. First, the resulting destination system is broader and more typologically diverse than official recognition systems suggest, encompassing not only numerous niche sites but also linear destinations, such as the Sichuan-Tibet Highway and the Yunnan-Tibet Route, where the journey itself constitutes the primary tourism experience. Second, temporal analysis reveals a shift from expansion toward competitive consolidation, with a peak in newly emerging destinations in 2017 followed by accelerating decline, and newer destinations entering decline significantly earlier than older ones. Third, among NBTDs overlapping national nature reserves, 63.2% fall outside formal management systems, and evidence from Sanjiangyuan and Qomolangma suggests that administrative closures may redirect rather than eliminate visitor flows toward adjacent, less-regulated areas. These findings demonstrate that the choice of destination identification method is not a technical detail but a governance precondition, and that making the full destination system visible is the first necessary step toward managing nature-based tourism in ecologically fragile regions.