Application of lidar and camera traps to model wildlife habitat

Abstract

Many wildlife studies develop habitat models based on spatially limited, ground-based surveys. These surveys cannot adequately measure forest canopy features or characterize features at landscape scales, nor can these models be applied at larger extents, making them less useful for conservation. Lidar offers the opportunity to objectively measure habitat features across landscape-level extents, making it possible to predict landscape suitability at the level of entire states. We measured structural diversity of forest vegetation at 179 plots across Indiana, USA, at 3 radii (50 m, 100 m, and 300 m) to explore species associations with forest characteristics at different spatial scales. We developed occupancy models at each scale to predict species presence by relating detection–non-detection of wildlife species from camera traps (2019–2020) to indices of forest structure derived from aerial lidar (2016–2020). We quantified the relationship between the presence of 7 wildlife species and the 3 forest structure metrics. The effect sizes for each metric varied depending on the plot extent and species, often displaying opposite trends at different radii. Deep gap fraction had a strong positive association with species like eastern cottontail (Sylvilagus floridanus) at the 100-m and 300-m extents (95% credible intervals did not cross zero), but weaker relationships at 50 m. Some species also used areas with higher vegetation diversity, such as coyote (Canis latrans), which was positively associated with higher values for Gini foliage diversity at the 300-m extents. Habitat generalists did not show strong evidence of selection for specific metrics. Our study demonstrates that lidar shows great promise for habitat modeling based on its ability to characterize forest structure at landscape scales across large geographic extents. These models identify patterns of space use that are difficult and expensive to capture through traditional survey methods and have clear conservation and management applications owing to their extensive spatial coverage.