The power of social media text: Mapping the distribution and drivers of cultural ecosystem services in Taihu Lake wetland

Inland wetlands are vital urban spaces that integrate ecological conservation with cultural experiences, offering diverse cultural ecosystem services (CES). However, existing CES assessment studies often fail to fully utilize social media text data, which may bias the identification of different service types and compromise the accuracy of spatial modeling. This study takes the Taihu Lake wetland as a case study, using user reviews from multiple social media platforms and applying in-depth semantic analysis to identify six types of CES: mental and psychological, cultural heritage, educational, inspiration, aesthetic, and recreational services. The MaxEnt model is used to map their spatial distribution and explore influencing factors. Results show that: (1) in-depth semantic analysis of multi-source social media texts enhances the accuracy of CES assessments, especially for subjective types like aesthetic and mental and psychological services; (2) CES in the Taihu Lake wetland show low values across the lake and a clear east–west decline on surrounding land. Mental and psychological services show the highest intensity, reflecting visitors’ reliance on urban ecological spaces for emotional well-being; and (3) the spatial distribution of CES is mainly shaped by socio-economic factors, with human–environment interactions as the foundation and transportation accessibility as a key catalyst. Natural factors provide ecological support but have limited explanatory power. This study advances the assessment of intangible CES, clarifies their spatial distribution, and provides theoretical and practical support for human–environment coordination and regional governance in inland wetlands.