Mapping emotional patterns in linear recreational landscapes: Spatial structure, positive-negative co-occurrence patterns, and nonlinear drivers

As landscape planning shifts toward experience-oriented design, emotion mapping has become critical to understanding how spatial environments shape human perception. However existing approaches are largely grounded in site-based or area-based spatial units, limiting their ability to represent emotional variation in directional and continuous linear recreational landscapes. To address this gap, this study develops an integrated framework combining panoramic visual capture, VR-based emotion elicitation experiments, and MaxEnt spatial prediction modeling to quantify and map emotional fluctuations along linear landscapes. Using the No. 1 Scenic Road of Wuyi Mountain National Park as a case study, we find that: (1) emotional fluctuations exhibit a hierarchical spatial structure, clustering along the main road and gradually diffusing into surrounding buffer zones; (2) positive and negative emotions significantly co-occur in core scenic segments, challenging the traditional assumption of emotional mutual exclusivity; (3) emotional distributions are jointly influenced by multiple landscape attributes, with key drivers demonstrating pronounced nonlinear threshold effects. This study advances emotional mapping from static area-based settings to continuous linear recreational systems and provides evidence-based support for refined landscape planning and management.