Habitat highs and lows: Using terrestrial and UAV LiDAR for modelling avian species richness and abundance in a restored woodland

Vegetation structure influences landscape use and habitat quality for many bird species. Owing to the difficulties associated with collecting structural data from traditional field measurements, numerous studies have investigated the utility of Light detection and ranging (LiDAR) for providing landscape-scale structural information that may be useful for exploring animal-habitat associations. Notably, almost all of these studies have involved the use of LiDAR from airborne rather than terrestrial platforms. However, vegetation metrics that might be important for explaining bird species occurrence and diversity, such as understory vegetation complexity and overall vegetation volume, may be partially obscured from airborne sensors by tree canopy cover. These challenges might be overcome by terrestrial and UAV LiDAR sensors that can provide detailed information of understory forest strata. For the first time, we collected terrestrial LiDAR (TLS) and unoccupied aerial vehicle LiDAR (ULS) data in a woodland landscape to compare the ability of both sensors to identify relationships among vegetation structural metrics and bird species richness and abundance. Overall, TLS and ULS models provided similar results based on the sampling methodology we used for LiDAR data collection in an open woodland landscape. Canopy roughness, ground vegetation vertical complexity, total vegetation volume and canopy height derived from these sensors were among the most common significant variables in explaining avian diversity and individual species abundance. Individual species abundance models provided better prediction power (up to R2 = 0.82 (TLS) and R2 = 0.83 (ULS)) than bird community abundance by functional guilds (up to R2 = 0.40 (TLS), R2 = 0.41 (ULS)) and overall bird abundance (R2 = 0.10 (TLS), R2 = 0.16 (ULS)), species richness (R2 = 0.14 (TLS), R2 = 0.14 (ULS)) and diversity (R2 = 0.17 (TLS), R2 = 0.16 (ULS)). Additionally, we found that several vulnerable bird species are strongly associated with LiDAR structural variables, which may assist with habitat assessment and conservation management.