Effects of landscape structure on genetic diversity of Geum urbanum L. populations in agricultural landscapes

Plant species in fragmented populations are affected by landscape structure because persistence within and migration among inhabited patches may be influenced by the identity and configuration of surrounding habitat elements. This may also be true for species of the semi-natural vegetation in agricultural landscapes. To determine the effect of landscape elements we analyzed Wood Avens (Geum urbanum L.) populations within three 4×4 km2 agricultural landscapes in Germany, Switzerland and Estonia, which differ in levels of land use intensity and habitat fragmentation. Genetic variation was determined in 15 randomly selected populations in each landscape using 10 microsatellite loci. The landscape structure was assessed at two circles around each population, with radii defined by the range limits of spatial genetic autocorrelation. Multiple regression analysis was used to determine the influence of landscape structure variables for inter- and intrapopulation genetic diversity. Gene diversity was equally high in Germany (He=0.27) and Switzerland (He=0.26) but lower in Estonia (He=0.16). A high overall inbreeding coefficient (FIS=0.89) was found, as expected for a selfing breeding system in G. urbanum. Genetic differentiation among populations was high (overall FST=0.43, 0.48, and 0.45 in Estonia, Switzerland and Germany, respectively), and did not differ among the three landscapes. Only a moderate influence of individual land use types on genetic diversity within and among populations was found with some idiosyncratic relationships. Genetic variation within populations was correlated to the amount of hedgerows positively in Estonia but negatively in Switzerland. The study demonstrates that the distribution of individual land use types affects the genetic pattern of a common plant species. However, different variables were identified to influence the genetic structure in three different landscapes. This indicates a major influence of landscape-specific land use history and stochastic processes determining gene flow and plant population structur