Associations between ecological indicators and public satisfaction in urban parks: Evidence from a multi-indicator assessment in xi’ an, China

Although ecological indicators and public satisfaction have both been widely used to evaluate urban parks, they are still commonly examined in separate analytical frameworks. City-scale evidence remains limited on how multidimensional ecological indicators are associated with perception-based public satisfaction, particularly under heterogeneous urban contexts. Using 206 urban parks in Xi’ an, China, this study assessed park ecological value through twelve indicators grouped into intrinsic ecological attributes and spatially extrinsic attributes. The indicators covered vegetation condition, habitat quality, carbon storage, landscape pattern, accessibility, surrounding urban vitality, and service provision. A park-level ecological value index was generated by principal component analysis, while public satisfaction was estimated from sentiment scores extracted from user-generated posts on Weibo and Xiaohongshu. District-level random forest models were then used to examine how the ecological indicators were associated with satisfaction in different urban contexts. The park system showed marked ecological differentiation: only 11.65% of parks fell into the high-value class, whereas more than 70% were classified into low-value classes. Social-media sentiment was positive overall, but satisfaction still varied considerably across parks and districts. In most districts, satisfaction was more closely associated with spatially extrinsic indicators—particularly point-of-interest density, landscape pattern, and accessibility—than with intrinsic ecological attributes. By contrast, vegetation condition and carbon storage showed weaker and less direct links with user evaluations. The influential indicators also shifted noticeably from one district to another, pointing to strong contextual variation in park-user relationships. These findings suggest that ecological assessment based only on biophysical performance cannot fully represent how parks are experienced in practice. Integrating ecological indicators with perception-based social data can therefore support more context-sensitive park evaluation, planning, and management. • Twelve indicators assessed ecological value across 206 parks in Xi'an. • Public satisfaction was estimated from 47,000 social media posts. • Ecological value showed pronounced spatial heterogeneity across parks. • Satisfaction was linked more to accessibility and urban context than ecology. • Indicator–satisfaction relationships varied markedly across districts.