An application of Manel's model: Detecting bobcat poaching in Michigan

The illegal harvest of natural resources (i.e., poaching) has the potential to threaten the persistence of many plant and animal species. In Michigan bobcats (Lynx rufus) are distributed throughout the Upper Peninsula (UP) and the northern half of the Lower Peninsula (LP) and are a biologically and economically important species. The popularity of bobcat hunting and trapping in Michigan, along with different harvest regulations between the 2 peninsulas, has created the need for a reliable method of identifying incidences of poaching. Because the bag limit is higher in the UP, we hypothesized that some bobcats harvested in the LP are being registered as originating from the UP. We used 8 polymorphic microsatellite markers and the statistical package STRUCTURE to assign individuals to the population in which they had the highest likelihood of occurrence based on their genotype. We evaluated the influence of using posterior probability threshold values from T > 0.9-0.999 on the number of animals classified as poached. Based on this range, STRUCTURE produced correct assignment rates of 53-82%. All instances of genetic re-assignment involved bobcats claimed as harvested in the UP but genetically assigned to the LP following the suspected method of bobcat poaching in Michigan. This approach provides a reliable method of determining the source population for bobcats harvested in the state and should provide enforcement agencies with a useful way of identifying potential poaching cases.