Multi-scale and multi-site resampling of study area in spatial genetics: implications for flying insect species

The use of multiple sampling areas in landscape genetic analysis has been recognized as a useful way to generalize the patterns of environmental effects on gene flow. It allows reducing the variability of inference, accounting for multiple scales and locations of study areas. Although several reviews have stressed the importance of this point, few studies have considered multiple sampling areas in analysis and formally tested their effects on inference. In this study, we present a method for resampling of study areas at multiple scales and multiple locations (sliding windows) to track the variation of inference in spatial genetics. We explored the effects of environmental features on gene flow of a flying longhorned beetle (Monochamus galloprovincialis) in 3*10 study areas ranging in scale from 220 to 1000 km and spread over 132 locations among the Iberian Peninsula. We show that there were no general or recurrent effects of environmental features detected among scales and locations, independent of variation in environmental features. Detection of environmental features on gene flow generally increased with an increasing scale of study, and was variable between locations. The resampling method presented here provides the opportunity to explore the effects of environmental features on gene flow of organisms in their whole extent and to conclude about general landscape effects on the dispersal of organisms, while keeping sampling effort to a reasonable level.