Informing age-friendly cultural landscape management through older adults' perception: An interpretable multi-source...

Landscape perception is increasingly recognized as an important basis for environmental management, particularly in everyday urban landscapes used by socially diverse groups. In the context of population aging, cultural landscapes not only support daily outdoor activity and emotional attachment, but also provide perceptual and experiential values that can inform age-friendly landscape management. However, existing studies on older adults' landscape evaluation have mainly focused on single visual or natural elements, while the combined associations among surface thermal background, spatial morphology, greenness-related visual attributes, and perceived scene qualities remain insufficiently understood. To address this gap, this study develops a multi-source and interpretable analytical framework to examine older adults' landscape evaluation along the Nanjing Yangtze River cultural landscape corridor. A node-linked dataset was constructed from 58 landscape nodes, integrating 22,924 street-view images, remote-sensing indicators, and 174 questionnaire responses nested within these nodes. Human-scale environmental features were extracted from street-view imagery using computer vision, locally calibrated perception proxies were derived from a Place Pulse 2.0-based model, and six machine learning models were compared for exploratory classification. Extra Trees showed comparatively balanced exploratory performance and was selected for SHAP-based interpretation. The results indicate that land surface temperature, spatial openness, street enclosure, green view ratio, and perceived safety had relatively large model contributions. These variables displayed different model-derived response patterns, including candidate inflection, diminishing-return, and monotonic patterns. SHAP-based clustering further revealed spatially differentiated response typologies along the corridor. These findings provide an exploratory, perception-informed basis for differentiated environmental management, urban greening, and age-friendly cultural landscape planning.