Cooperative Fish and Wildlife Research Units Program: Oregon
Education, Research and Technical Assistance for Managing Our Natural Resources

Peterson, J. T., C. R. Jackson, C. P. Shea, and G. Li. The development and evaluation of a stream channel classification for estimating the response of fishes to changing streamflow. Transactions of the American Fisheries Society


Current approaches to assessing the potential effects of river regulation and water use on stream fish communities are based on physical habitat simulation and are not feasible for estimating these effects over large spatial scales. We developed a channel classification for streams in the lower Flint River basin, Georgia, based on gross channel morphology and geology and evaluated its usefulness at 23 study sites representing the four channel types in the basin. Our channel classification separated stream channel types based on dominant geology (upland residuum versus Ocala limestone) and channel form (confined versus unconfined). Fish were sampled and habitat measured at the study sites in spring, summer, and winter from 2001 to 2004, a period that included some of the lowest and highest seasonal flows ever recorded. The channel types differed with respect to diurnal variability in temperature and dissolved oxygen concentration as well as habitat characteristics (substrate, large wood). Statistical habitat modeling indicated that channel unit (e.g., pool, riffle) availability could be predicted accurately for each channel type using streamflow, with errors rates that averaged 0.7% across channel unit types. Comparisons between the accuracy of species presence and fish abundance models fit using habitat characteristics to those fit using geomorphic channel characteristics indicated similar model accuracy with slightly less bias (6% on average) for the geomorphic channel characteristics models. Our results indicate that the channel types were useful for predicting changes in habitats and stream fish presence and abundance in response to changing streamflow. Channel classification is potentially a useful and cost-effective approach to quantifying changes in the physical environment in response to varying streamflow. Additionally, channel classification could be combined with dynamic fish population models to evaluate the effects of river regulation and water use across large spatial scales.