| Summary: | Abstract Invasive species have led to precipitous declines in biodiversity, especially in island systems. Brown ( Rattus norvegicus ) and black rats ( R. rattus ) are among the most invasive animals on the planet, with eradication being the primary tool for established island populations. The need for increased research for defining eradication units and monitoring outcomes has been highlighted as a means to maximize success. Haida Gwaii is an archipelago ~100 km off the northern coast of British Columbia, Canada that hosts globally significant breeding populations of seabirds that are at risk due to invasive rats. Here, we paired sampling of brown (n=287) and black (n=291) rats across the Haida Gwaii archipelago (British Columbia, Canada) with genotyping-by-sequencing (10,770-27,686 SNPs) to investigate patterns of population connectivity and infer levels/direction of gene flow among invasive rat populations in Haida Gwaii. We reconstructed three regional clusters for both species (north, central, south), with proximate populations within regions being largely more related than those that were more distant, consistent with predictions from island biogeography theory. Population assignment of recently detected individuals post-eradication on Faraday, Murchison, and the Bischof Islands revealed all were re-invaders from Lyell Island, rather than being on-island survivors. Based on these results, we identified six eradication units constituting single or clusters of islands that would limit the potential for re-invasion, some of which will need to be combined with biosecurity measures. Overall, our results highlight the importance of targeted research prior to conducting eradications and demonstrates a framework for applying population genomics for guiding invasive species management in island systems. Methods rano_final.vcf: genotypic data for 297 brown rats (Rattus norvegicus) at 27,686 SNPs generated via ddRAD rara_final.vcf: genotypic data for 242 black rats (Rattus rattus) at 10,770 SNPs generated via ddRAD
|
|---|