Assessing tourism pressure on wildlife in a protected area: A social big data approach using photo-sharing data and species distribution modeling

Protected areas face growing challenges in balancing biodiversity conservation with the pressures of nature-based tourism. This study develops a geospatially explicit framework integrating social big data analytics and species distribution modeling (SDM) to assess tourism-induced ecological stress in Shei-Pa National Park, Taiwan. We used datasets on the potential distribution range of 49 protected terrestrial vertebrate species. We selected 16 protected species on the IUCN Red List whose potential distributions overlapped with the park. We analyzed 2914 spatial grids (500 × 500 m) using photo-user-days (PUD) derived from geotagged photographs uploaded to the Flickr platform (2008–2017). The monthly PUD values were strongly correlated with official visitor statistics ( r = 0.818, p = 0.001), validating their reliability for seasonal analysis. Kernel density estimation and regression analyses revealed that infrastructure variables—particularly visitor centers (estimate = 2.0160) and hotels (1.6550)—were the strongest predictors of tourism pressure (Adjusted R 2 = 0.3568), while rivers showed no effect. Spatial overlay analysis revealed that high PUD values strongly overlapped with the predicted habitats of Level 1 protected species (e.g., Passer rutilans, Prionailurus bengalensis chinensis ), especially in the eastern region of the park, including a Generally Protected Area and a Conservation Area. Seasonal circular statistics revealed a bimodal tourism pattern, peaking in February and in November–December (Rayleigh Z = 5.72, p < 0.01), with minimal visitation in May. The PUD index ranged from 0 (minimum) to 35.2 (maximum), with a median value of approximately 0.16. This seasonal trend highlights the potential value of implementing visitor access controls during off-peak months. Tourism pressure was lowest in remote, high-altitude conservation zones, which often harbor sensitive species. We identified four conservation risk zones-LL, LH, HL, and HH-based on spatial overlays of habitat suitability and tourism pressure. Note that HH zones, indicating high ecological value and high tourism pressure, were concentrated in the eastern region of the park and accounted for approximately 6 % of its total area. These areas represent the most critical zones for biodiversity conflict and inform targeted zoning, seasonal access control, and visitor redistribution strategies. This study operationalizes open-access social media data and ecological modeling to support the adaptive management of biodiversity-rich protected areas. • This study integrates social big data (Flickr-derived Photo-User-Days) with species distribution modeling to assess tourism pressure in Shei-Pa National Park. • Tourism pressure is highest in infrastructure-accessible zones such as visitor centers, hotels, and trailheads. • Overlay analysis reveals that several endangered and nationally protected species experience high spatial overlap with tourism hotspots. • Remote, high-altitude areas show low visitor intensity and function as ecological refugia for conservation-priority species. • The study proposes evidence-based management strategies, including spatial zoning, seasonal access regulation, and real-time monitoring, to mitigate tourism impacts.