Remotely sensed forest habitat structures improve regional species conservation
- Journal Article
Rights / licenseCreative Commons Attribution-NonCommercial 4.0 International
Recent studies show that light detection and ranging (LiDAR) derived habitat variables significantly increase the performance and accuracy of species distribution models (SDMs). In particular, the structure of complex habitats such as forest can be accurately parametrized by an area-wide, LiDAR-based vegetation profile. However, evidence of specific applications of such models in real-world conservation management still remains sparse. Here, we developed a resource selection SDM for hazel grouse (Bonasa bonasia L.) in a Swiss nature park with the aim to map habitat suitability and to inform the park management about habitat improvement measures. We used remote sensing, particularly LiDAR to derive ecologically relevant forest vegetation characteristics at the local scale and used them as predictors in an ensemble SDM approach. The predicted habitat suitability was mainly affected by local, fine grained vegetation structure. Average vegetation height, shrub density and canopy height variation contributed most to the habitat quality for hazel grouse. This clearly shows how LiDAR provides the means to develop ecologically interpretable predictor variables of forest habitat structure and that these predictors can be used to reliably map local-scale habitat quality, indicated by high model performance scores (median AUC of 0.918). This improves spatial conservation planning, and at the same time, provides meaningful information to derive habitat improvement measures that can be implemented in the field by foresters. Hazel grouse occurrence in the park is restricted to a few highly suitable, disjunct habitat patches. Therefore, conservation management should increase the connectivity of suitable habitat with the aim to stimulate an increase and better exchange of individuals in the regional hazel grouse population. Habitat improvements can be achieved by forestry measures that regularly integrate early successional forest stages into production forests. They should contain stands with a shrub density of around 30% as well as heterogeneous stands in terms of vegetation height. Show more
Journal / seriesRemote Sensing in Ecology and Conservation
Pages / Article No.
SubjectConservation; forest; habitat suitability; hazel grouse; LiDAR; Species distribution modelling
MoreShow all metadata