QUANTIFYING CONNECTIVITY: A RECOVERY NETWORK OF AMERICAN MARTENS (MARTES AMERICANA) IN THE UPPER GREAT LAKES REGION

Abstract

The reintroduction of species back into their former ranges is increasingly occurring in landscapes that differ dramatically from historical conditions. These novel landscapes often feature unknown connectivity, leading to uncertainty on the recovery and long-term viability of reintroduced populations. As one of the most frequently reintroduced taxa, carnivores are especially sensitive to landscape fragmentation. We studied a recovery network of American martens (Martes americana) in the upper Great Lakes region, where martens have been reintroduced >15 times across multiple sites, resulting in at least 8 extant populations. Many of these individual reintroduced populations possess uncertain viabilities and appear to be declining. Critically, the functional connectivity of subpopulations within this recovery network is unknown; however, recent observations of martens outside of reintroduction sites suggest that natural dispersal is occurring. Using a databank of >1,100 genotyped individual American martens collected from 8 subpopulations from 2000-2023, we quantified the degree of connectivity and identified corridors across the region. We hypothesized the populations within the recovery network are connected via dispersers, especially from the source populations in Michigan and Minnesota, and predicted that the recovery network is principally structured by isolation-by-resistance (IBR). We identified two spatial genetic structures in the recovery network: a continuous cline of genetic relatedness that originated in the eastern Upper Peninsula of Michigan and extended westward into Wisconsin, and a genetically distinct population of martens in Minnesota that were principally related to martens in the rest of the recovery network due to historical human translocations. Through our quantification of a resistance landscape

parameterized with a resource selection function, we found landscape resistance limits connectivity, and that martens across the Great Lakes recovery network were primarily structured by IBR when high-quality habitat is limited. Notably, habitat fragmentation and pine barrens created high landscape resistance in Wisconsin and reduced connectivity amongst reintroduced populations in Wisconsin and with those in Minnesota. In general, identified corridors within the recovery network were broadly protected; however, we identified large gaps in corridor protection within Minnesota. These areas could represent important opportunities for management to facilitate natural dispersal between the large marten population in Minnesota with the rest of the recovery network.