Last changed on: 04.10.2019

2016 Holderried, Philip: Using LiDAR derived vegetation metrics for multiscale species distribution mapping of Eurasian Pygmy owl (Glaucidium passerinum) in the Black Forest National Park



Successful conservation management should always be knowledge-driven and thus requires information on species-environment interactions and species distribution. Yet, species distribution is, even with expert knowledge, often hard to assess due to limited data on species occurrence. Presence-only species distribution models can help overcoming this hurdle by utilizing species records that, in fact were not systematically sampled, but are often more widely

In this study occurrence records of Eurasian Pygmy owl (Glaucidium passerinum) in the Northern Black Forest in Germany were analysed in presence-only software MaxEnt. 31 forest structural measures were derived from small footprint LiDAR data and used as predictors in combination with a dataset of 76 species records. Sampling bias was reduced by introducing a bias file. Effects of different bias files and model complexity penalties were tested before
ecologically relevant scales of predictor variables were determined. The resulting multiscale habitat model was simplified in a reduction process until 12 significant variables remained.

Edge density in vegetation layers between 5 and 35 m had the biggest effect on species distribution, followed by vegetation density and measures for vertical height distribution. Forests with a normal height distribution and sparse ground vegetation were positively related to Pygmy owl occurrence, as were areas with moderate exposure to wind and high values for topographic wetness. Most covariates had their best performing scale at 450 and 1050m. Small scale variability only played a role in average vegetation height and edge density between 20 and 35m above ground, both of which performed best on a 50m scale.