Butterfly dispersal in inhospitable matrix: rare, risky, but long-distance

Piotr Nowicki,Vladimir Vrabec,Thomas Hovestadt,Barbara Zakšek,Josef Settele,Birgit Binzenhöfer,Johann Feil

doi:10.1007/s10980-013-9971-0

Abstract

Metapopulation models typically assume that suitable habitats occupied by local populations and unsuitable matrix separating them form a 'black- and-white' landscape mosaic, in which dispersal is primarily determined by the spatial configuration of habitat patches. In reality, however, the matrix com- position is also likely to influence dispersal. Using intensive mark-recapture surveys we investigated inter-patch movements in Maculinea (Phengaris) nausithous and M. teleius occurring sympatrically in six metapopulations. Three of these metapopulations had the matrix dominated by forest, an inhospitable environment for grassland butterflies, whereas in the remaining three the matrix was mostly composed of open environments. Dispersal parameters derived with the Virtual Migration model revealed significant differences between both groups of metapopulations. Both species had a lower propensity to emigrate from their natal habitat patches, and they suffered substan- tially higher dispersal mortality in the metapopula- tions with forest matrix. On the other hand, mean dispersal distances were roughly an order of magni- tude longer in forest matrix as compared with open landscapes (ca. 500-1,500 vs. 100-200 m). Our results suggest that inhospitable forest matrix induces strong selection against dispersal, leading to a reduced emigration rate. At the same time, the selection may promote emigrants with good dispersal abilities, which are able to perform long-distance movements. Thus, while it is generally believed that a matrix

Highlights

Dispersal is a key process for species survival in fragmented landscapes (Fahrig 2003; Bowne and Bowers 2004; Reed 2004)
Metapopulation models typically assume that suitable habitats occupied by local populations and unsuitable matrix separating them form a ‘blackand-white’ landscape mosaic, in which dispersal is primarily determined by the spatial configuration of habitat patches
It is noteworthy that the estimates of the three main dispersal parameters of interest, i.e. emigration rate, dispersal distance, and dispersal-related mortality, were highly repeatable among the three metapopulations with open-land matrix as well as among the three metapopulations with predominantly forest matrix

Summary

Introduction

Dispersal is a key process for species survival in fragmented landscapes (Fahrig 2003; Bowne and Bowers 2004; Reed 2004). It enables gene flow, preventing local populations from genetic variability loss due to genetic drift and inbreeding (O’Grady et al 2006; Hanski 2009; Lowe and Allendorf 2010). It provides rescue effects for declining populations, and it makes possible colonisation of vacant habitat patches, ensuring metapopulation persistence despite occasional extinctions of local populations (Hanski et al 1996; Gonzalez et al 1998; Poethke et al 2003). Differences in dispersal patterns may depend on whether environments forming the matrix are structurally similar to the habitats of a species (hereafter termed hospitable matrix for the sake of brevity) or clearly distinct from them (inhospitable matrix)

Methods

Results

Conclusion