Stream habitat modelling for conserving a threatened headwater fish in the upper cumberland river, Kentucky

The conservation of stream biodiversity requires more explicit knowledge on the distribution of aquatic species within the context of their specific environmental settings and stresses. Although species distribution models (SDMs) have been widely used for organisms occupying contiguous spatial extents, the implementation of SDMs in relatively complex and segmented riverine networks is still at its early stage. In this study, we explicitly modelled the headwater stream habitat for the threatened blackside dace (Phoxinus cumberlandensis) endemic to the upper Cumberland River, Kentucky, USA. An occurrence record data set, along with variables describing stream properties and land use impacts, was used to predict the fish habitat suitability at the stream segment level. An approach combining geographic information systems and the maximum entropy species distribution modelling (MaxEnt) was adopted. Results demonstrated that natural conditions and land use disturbances, respectively, form the primary and secondary environmental constraints on the species' habitat. We generated regional-scale management-friendly maps showing subwatershed habitat suitability and locations of the clustered suitable habitats (hotspots) and thus set an example for spatially explicit management of threatened and endangered riverine species. This study demonstrates the usefulness of SDMs for stream network-based environments in the facilitation of biogeographic conservation efforts and studies.