Liana removal alters canopy chemistry more than structure in tropical seasonal forests: Insights from UAV-borne hyperspectral and LiDAR data

The abundance of lianas (woody vines) has increased in tropical forests due to climate change and fragmentation, altering important ecosystem functions related to both carbon and water cycles. The quantification and mapping of lianas’ effects on the forest canopy are therefore needed, yet they remain unresolved. This study investigates the effects of liana removal on the chemical and structural properties of the canopy in a tropical seasonal semideciduous forest in São Paulo, Brazil, using multitemporal UAV-borne hyperspectral (400–1000 nm) and LiDAR data. The study leveraged 18 plots (each 45 × 44 m; 0.198 ha), comprising 12 liana removal plots and 6 unmanaged controls, surveyed in both dry and wet seasons. In the dry season, pigment-sensitive vegetation indices (e.g., PARS, GMI1, SR3) showed medium to large negative responses to removal, reflecting reduced canopy greenness as evergreen lianas, sustained by deep roots, dominate amid tree defoliation. Conversely, wet-season indices (e.g., DD, OSAVI2, MSAVI) exhibited negligible effects, with resource abundance masking differences. LiDAR metrics revealed small to medium reductions in leaf area index (LAI), especially understory LAI, and increased gap fraction variability post-removal, with stronger contrasts during the dry season. These findings highlight UAV remote sensing’s potential for detailed, seasonal monitoring of liana influences on canopy chemistry and structure, informing targeted management to enhance forest resilience in varying tropical environments.