Deposited sediment influences occurrence of functional traits of stream fishes

Abstract

To better understand stream-fish sensitivity to fine sediment, we documented assemblage-wide responses by selected traits along a sedimentation gradient. We then discuss the management implications of these ‘dose–response’ relations in the contexts of biotic assessments and conservation of sediment-sensitive species. We identified a spatial gradient in sediment deposition among streams within the upper Piedmont of the Roanoke River basin in North Carolina and Virginia. We assessed fine-sediment sensitivity of 81 species based on eight species traits stratified by four attributes: food preference, feeding location, spawning substrate and spawning behaviour. We then ranked each trait and scored each species with respect to its sediment sensitivity. Using data from electrofishing surveys during 2018–2019, we calculated proportional abundances of traits observed at 30 sites throughout the study area and grouped species by their aggregate sensitivity scores. We assessed relations between embeddedness and silt cover and occurrences of species and traits using a combination of regression and ordination approaches. All traits tested responded to embeddedness or silt cover, or both. Feeding traits exhibited the strongest responses to embeddedness, while reproductive traits exhibited the strongest responses to silt cover. Our findings indicate that negative responses of the probability of presence for high-sensitivity traits to embeddedness and silt cover were linear, with no apparent thresholds. Additionally, proportional abundances of species with multiple high-sensitivity traits were inversely related to embeddedness and silt cover. Overall, our findings regarding population-level responses to sedimentation were consistent with our findings for trait-specific responses. Our analysis of species sensitivity to fine sediment corroborated the patterns we saw in our trait-specific analyses, indicating that population responses to sedimentation can be predicted from combinations of species traits. The ‘dose–response’ relations we documented may be applicable to managing sediment impacts on fishes, especially in the contexts of biotic assessments and conservation of sediment-sensitive species.