Limitations of hypothesis-testing in defining management units for continuously distributed species

Résumé

Estimating the risk to wildlife populations resulting from human-induced mortality relies on adequately defining population structure. For marine populations, including cetaceans, identifying population boundaries is difficult because most species have large continuous distributions with no obvious barriers to dispersal. For many species, the extreme ends of the range differ in morphology, indicating that population structure exists. However, the lack of distributional hiatuses often makes this structure difficult to detect. A common method of defining structure in such situations is to use genetic differentiation as a proxy for limited movement between areas. Genetic analyses of population structure usually take the form of hypothesis testing, which requires the a priori definition of hypothesised units and testing for significant genetic differentiation between them. Simulations are used to examine the performance of hypothesis testing to correctly define population structure. Results show that hypothesis testing is likely to lead the researcher to define fewer management units than are necessary to adequately protect local populations from over-exploitation. The need for the development of new methods of defining management units and for rigorous performance testing of all methods applied in a management context is highlighted.