Get access

Where movement happens: scale-dependent landscape effects on genetic differentiation in the European tree frog


  • Sonia Angelone,

  • Felix Kienast,

  • Rolf Holderegger

S. Angelone, WSL Swiss Federal Research Inst., Zürcherstrasse 111, CH-8903 Birmensdorf, Switzerland and Inst. of Zoology, Univ. of Zurich, Winterthurerstrasse 190, CH-8057 Zurich, Switzerland. Present address of SA: EAWAG, Seestrasse 79, CH-6047 Kastanienbaum, Switzerland. – F. Kienast and R. Holderegger (, WSL Swiss Federal Research Inst., Zürcherstrasse 111, CH-8903 Birmensdorf, Switzerland.


Functional connectivity among fragmented populations depends on the landscape matrix between occupied habitat patches and its effect on the frequency of animal movement and gene flow. The quantification of landscape effects on gene flow should therefore be scale-dependent. Here, we explored the impact of different spatial scales in a landscape genetic analysis of the European tree frog Hyla arborea in a fragmented landscape in Switzerland. We examined the effects of landscape elements and geographic distance on genetic differentiation at three distance classes reflecting varying frequencies of tree frog movement. We calculated pairwise FST-values and assembled 16 landscape elements within 1 km wide corridors between all pairs of tree frog breeding sites. Per distance class, we computed a multiple regression model with stepwise backward elimination and permutation testing. At distances of<2 km, only a larger river acted as a barrier to gene flow. At distances>2 km, geographic distance had a negative effect on gene flow as had landscape elements such as forests and roads. In general, hedgerows and various structure-rich landscape elements positively affected gene flow. As we found distinct scale-dependent landscape effects on gene flow, future landscape genetic studies should analyse the effects of landscape variables at different spatial dimensions relevant for the movement and dispersal of the study organisms. Corresponding studies should also carefully consider relevant correlations among the landscape elements tested and should preferentially replicate their analysis at the landscape-level in order to avoid idiosyncratic results owing to the particular scale and landscape studied.