Structural vulnerability of an urban tourism flow network during a major holiday: a case study of Fuzhou, China

Major holidays can concentrate visitor mobility around a few attractions and transport hubs, increasing the potential sensitivity of urban tourism systems. This study examines Fuzhou during the 2023 National Day holiday (September 30–October 6) and assesses spatial concentration, sensitivity to hypothetical node loss, and conditional spatial associations. We analyse 1,911,577 qualified stay points from 376,780 anonymous visitor IDs using 1-km grid measures, an observed 45-node tourism-flow network, node-removal simulations, and spatial Durbin models. The top 5% of occupied grids contained 54.53% of stay points, indicating a concentrated but polycentric pattern. In the 95% cumulative-flow backbone, removing the four highest-betweenness nodes reduced the largest-connected-component ratio from 0.944 to 0.500 and normalised efficiency from 1.000 to 0.354; random removal produced smaller mean declines. Spatial models showed conditional associations between tourism intensity and supporting facilities, with a sample-specific threshold of 4,360.2 mean daily stay points. These findings describe observed concentration, counterfactual topological sensitivity, and regime-dependent associations, not actual operational failure or causal mechanisms. The conclusions are limited to one city and a seven-day window.