Scaling Down Habitat Selection by Large River Fishes to Understand Patterns Relevant to Individuals

Modification and homogenization of habitat in large‐river ecosystems have led to the reduction of >50% of native fish species. Rehabilitating these complex ecosystems to recover fish populations requires an understanding of habitat availability and selection at multiple scales. Habitat selection by river fishes is typically assessed at the functional unit scale (100–10 000 m²). For example, in large, sand‐dominated rivers of the Central USA, alluvial islands are critical functional units for endangered sturgeon. Functional units, however, can be subdivided into mesohabitats (<100 m²), but very little is known about mesohabitat selection by large river fishes. We evaluated the mesohabitat selection of the federally endangered pallid sturgeon (Scaphirhynchus albus) and more abundant shovelnose sturgeon (Scaphirhynchus platorynchus) experimentally. We tested for selection among four common mesohabitat types that are nested within alluvial island complexes: (1) sand‐only substratum with no structure; (2) sand substratum with a sand dune structure; (3) sand substratum with simulated vegetation; and (4) a gravel‐only substratum. Sturgeon selected for the sand substratum, structureless mesohabitat, followed by the mesohabitat with a sand dune. Vegetated habitat retained less sturgeon than these two habitats but more than the gravel mesohabitat. Age‐0 pallid sturgeon and shovelnose sturgeon almost completely avoided gravel substrate, perhaps because of increased energetic costs associated with turbulent benthic flow. We posit that age‐0 sturgeon may prefer the sand and sand dune habitats over the vegetation and gravel habitats because flow may be more linear (or unidirectional) and predictable in these habitat types, whereas vegetation and gravel can create substantial benthic turbulence. Lastly, shovelnose sturgeon were on average denser in vegetated habitat than pallid sturgeon. Scaled to the population level, patterns revealed here could have implications for the macro‐distribution of both species. Restoration efforts may want to consider selection differences in the management of these two species and rehabilitation of riverine habitats. Copyright © 2015 John Wiley & Sons, Ltd.