Predicting habitat suitability for Little Owls in Switzerland on different spatial scales

Accurate predictions of habitat suitability depend on habitat analyses on multiple spatial scales. Broad-scale distribution and composition of land cover plays an important role for the choice of breeding locations and occurrence probabilities; home-range scale structural composition of habitats reflects resource availability influencing reproduction and survival rates. However, habitat suitability studies incorporating parameters on landscape and home-range spatial scale are scarce, which impairs comprehensive planning of conservational measures. Land-cover data are rarely sufficient to predict local habitat suitability without including fitness relevant resources. Vice versa, high definition structural habitat data are normally missing when trying to analyze landscape suitability on wide spatial scales. This study combined both spatial scales in a two-step approach for endangered little owls Athene noctua in Switzerland. In a first step, habitat suitability was modeled based on landscape scale parameters and breeding occurrence data for Switzerland and Baden-Württemberg (Southern Germany), resulting in habitat suitability hotspot areas. Suitable habitat hotspots were then analyzed by field survey quantifying regional differences in habitat composition, structural richness, prey accessibility, and predation risk. A comparison on both levels between Southern Germany and Swiss regions identified potential factors responsible for the virtual absence of little owls in Switzerland. Suitable habitats for little owls on landscape scale were fruit orchard rich regions, distant to forests and urban areas. Switzerland showed less and highly scattered suitable habitat (4.23% of total area, Southern Germany: 15.59%) due to patchy forests and highly urbanized lowlands. On home-range scale Swiss suitable habitats were shown to be distinguished from Southern German sites by intensive meadow management, lower structural richness, and lower availability and complexity of tree cavities. The pattern was detected in all Swiss suitability hotspot regions, though in Western and Northern Switzerland habitat characteristics were more similar to Southern Germany. Little owl presence in Switzerland is therefore likely to be restricted by the missing structural diversity in available suitable habitats, rather than by a general lack of suitable habitats. Based on these results region specific conservation measures can be planned. Local and regional differences on both spatial scales demonstrate the importance of a multi-scale evaluation to predict suitable habitats. Our twostep approach was most adequate and efficient to detect multi-scale suitable habitats for little owls in Switzerland.

Keywords: Conservation, habitat suitability, multi-scale, predictive model, little owl, GIS-modeling, species-habitat relationship, resource availability