Land use structures fish assemblages in reservoirs of the Tennessee River L. E. Miranda A,C , J. M. Bies B and D. A. Hann B A US Geological Survey, Mississippi Cooperative Fish and Wildlife Research Unit, PO Box 9691, Mississippi State, MS 39762, USA. B Mississippi State University, Department of Wildlife, Fisheries, and Aquaculture, Mail Stop 9690, Mississippi State, MS 39762, USA. C Corresponding author. Email: smiranda@usgs.gov Abstract. Inputs of nutrients, sediments and detritus from catchments can promote selected components of reservoir fish assemblages, while hindering others. However, investigations linking these catchment subsidies to fish assemblages have generally focussed on one or a handful of species. Considering this paucity of community-level awareness, we sought to explore the association between land use and fish assemblage composition in reservoirs. To this end, we compared fish assemblages in reservoirs of two sub-basins of the Tennessee River representing differing intensities of agricultural development, and hypothesised that fish assemblage structure indicated by species percentage composition would differ among reservoirs in the two sub-basins. Using multivariate statistical analysis, we documented inter-basin differences in land use, reservoir productivity and fish assemblages, but no differences in reservoir morphometry or water regime. Basins were separated along a gradient of forested and non-forested catchment land cover, which was directly related to total nitrogen, total phosphorous and chlorophyll-a concentrations. Considering the extensive body of knowledge linking land use to aquatic systems, it is reasonable to postulate a hierarchical model in which productivity has direct links to terrestrial inputs, and fish assemblages have direct links to both land use and productivity. We observed a shift from an invertivore- based fish assemblage in forested catchments to a detritivore-based fish assemblage in agricultural catchments that may be a widespread pattern among reservoirs and other aquatic ecosystems. Received 1 July 2014, accepted 17 September 2014, published online 30 January 2015 Introduction Aquatic ecosystems are commonly subsidised by inputs from adjacent terrestrial systems (Pace et al. 2004). Inputs such as detritus, dissolved nutrients and nutrients embedded in sus- pended sediments are transported into streams and other water bodies by runoff. Depending on the extent of subsidies, inputs can have important influences over primary productivity in aquatic systems (Carpenter et al. 1998). Such influences are particularly notable in reservoirs. Because reservoirs have large catchment-to-reservoir area ratios, and are generally positioned towards the end of a large catchment, they may receive sub- stantial inputs of nutrients, sediments and detritus from streams that drain their catchments (Likens and Bormann 1974). These inputs can be large in catchments modified to support agricul- ture because of systematic deforestation, soil erosion and the applications of fertilisers (Arbuckle and Downing 2001). Subsidies from catchments can promote selected compo- nents of reservoir fish assemblages. When nutrient subsidies are large, they stimulate phytoplankton and zooplankton production and in turn production of planktivorous fish (Vanni et al. 2005). Similarly, when detritus subsidies are large, they stimulate production of detritivores (Gonza ´lez et al. 2010). Fish assem- blages in reservoirs of agricultural regions of eastern North America, where nutrient and detritus subsidies from catchments are large, are often dominated by gizzard shad Dorosoma cepedianum (Stein et al. 1995), a clupeid that relies on small plankton as larvae and can consume detritus in later stages (Yako et al. 1996; Miranda and Gu 1998). When at elevated densities, gizzard shad influence many functions of reservoir ecosystems, including nutrient cycling, primary production and composition and structure of the entire fish assemblage (Power et al. 2004). Conversely, when nutrient and detritus subsidies are small, phytoplankton production is reduced, water clarity increases and zooplankton production is shifted towards grazing zooplankters such as Daphnia (Kirk and Gilbert 1990; Mazum- der 1994). Lacking meaningful levels of detritus and increased water transparency, species composition shifts towards taxa that rely on visual selection of zooplankton prey or other inverte- brates through some or all of their life stages (Power et al. 2004). In eastern North America, those fish assemblages are often dominated by centrarchid species (Near and Koppelman 2009). Investigations linking catchment subsidies and fish assem- blages have generally focussed on one or a handful of species, albeit often species found in high densities within the reservoir (Vanni et al. 2005; Gonza ´lez et al. 2010). Considering this paucity of community-level awareness, we sought to explore the CSIRO PUBLISHING Marine and Freshwater Research http://dx.doi.org/10.1071/MF14188 Journal compilation Ó CSIRO 2015 www.publish.csiro.au/journals/mfr