Life history theory predictions for hydrologic filtering of fish assemblages are rarely tested with historical time series data. We retrospectively analyzed flow regime and fish assemblage data from the Sabine River, USA, to test relationships between life history strategies and hydrologic variability altered by impoundment construction. Downstream flow variability, but not magnitude, was altered by completion of Toledo Bend Reservoir (TBR) in 1966. Consistent with life history theory, occurrence of opportunistic strategists declined while equilibrium strategists increased as the fish assemblage was transformed between periods immediately after (1967–1973) and approximately one decade after (1979–1982) completion of TBR. Assemblage transformation was related to decline of opportunistic strategists throughout 250 km of river downstream of TBR. Temporal trajectories for opportunistic and intermediate strategist richness modelled as a function of flow variability converged 12 years postimpoundment. The spatiotemporal scaling of our study is novel among tests of life history theory, and results suggest impoundment-induced alteration to natural hydrologic filtering of fish assemblages can operate on the scale of hundreds of stream kilometres and manifest within approximately one decade.

Movement of fishes through space and time is critical for population regulation and community structuring, but the dispersal of many benthic stream fishes remains unstudied. We used passive integrated transponders to track the short‐term dispersal of 51 banded sculpin Cottus carolinae throughout a 600 m reach of Little Creek in central Tennessee during April and May. Our objectives were to assess the efficacy of recently developed dispersal models, evaluate temporal variability in movement and determine whether individuals switched between stationary and mobile movement behaviours. Observed movement distances did not differ from modelled leptokurtic dispersal kernels estimated using the fishmove package in the R Statistical Environment for 12 of 13 recapture occasions. Leptokurtic dispersal kernel parameters including the mobile component (σmob) and shared stationary component (p) were temporally dynamic and differed from static median values reported for fishes in fishmove, while the more abundant stationary component (σstat) showed agreement with fishmove. The recapture occasion during which model predictions were not validated was associated with a large flow pulse that stimulated increased movement at the population scale. At the individual scale, 28 of 51 fish switched between stationary and mobile dispersal behaviour and the frequency distribution of switches was leptokurtic. Collectively, our findings reveal an emergent property characterised by consistent upstream movement of banded sculpin despite variability in population‐scale responses to flow and individual‐scale switches in movement behaviour. This paradox represents the march of the sculpin, in which fish diffusively spread upstream at a constant rate despite multiscale variability in movement behaviours.

The river continuum concept (RCC) provides a framework for processes structuring lotic ecosystems by synthesizing sources and transport of C in streams. Considerable attention, refinement, and testing of the RCC has occurred since its inception >35 y ago, but few investigators have tested its predictions by explicitly linking consumer groups. We assessed insect assemblage structure in the diet of a broadly distributed insectivorous fish (Cottus carolinae) in the Roaring River continuum of Tennessee to test 3 predictions from the RCC: 1) longitudinal change in relative biomass of insect functional feeding groups (FFGs) including decrease for shredders, increase for collectors, intermediate maximum for grazers, and consistency for predators; 2) maximum taxonomic diversity at stream orders 3 to 5; and 3) temporal turnover in taxonomic composition across 1 y. We found that relative biomass of insect FFGs consumed by C. carolinae broadly matched predictions from the RCC. Maximum taxonomic diversity assessed at the family rank occurred at stream order 4 where diel and annual water temperature fluctuations were greatest, and monthly prey assemblages followed a sequence of turnover and a return to starting conditions across 1 y. Our novel approach illustrates proof of concept that RCC tenets are integrated into the diet of at least 1 higher-level consumer and, therefore, transcend assemblage boundaries in regulating the longitudinal (up-to downstream) and vertical (multiple consumer groups) flow of C in streams.

Stream fish diversity is threatened by anthropogenic environmental alterations to landscapes, and successful conservation requires knowledge of the processes that degrade diversity. A primary step in identifying diversity losses is the comparison between historical and contemporary states of landscapes and fish assemblages, but uncertainty remains regarding the appropriate spatial scales of investigation. Historical data collected in 1976 were paired with two years of contemporary replication (2015, 2016) to assess fish diversity change at 10 sites in Blackburn Fork, TN, USA. Analyses focused on a nested hierarchy of spatial scales, including sampling sites (fine scale), nested within stream orders (intermediate scale), nested within the entire catchment (broad scale). Diversity change between 1976 and 2015–16 was assessed using traditional diversity metrics (site scale) and rarefaction (stream‐order scale), whereas spatial variation in contemporary diversity (2015–16) was assessed with nonmetric multidimensional scaling (catchment scale). At the site scale, locations on the east side of Blackburn Fork and in close proximity to developed land experienced diversity loss. At the stream‐order scale, the effective number of species declined in first‐order streams where land development was concentrated, but no consistent species losses occurred in other stream orders. At the catchment scale, assemblages responded significantly to stream size but not land use, perhaps because diversity was already homogenized by 2015–16. Mapping 40 years of land‐use change across the catchment underscored a pattern of spatial alignment between developed lands and stream fish diversity loss. This study highlights the benefits of considering multiple spatial scales when assessing historical change in stream fish assemblages, and highlights stronger inference derived from historical comparisons relative to contemporary space‐for‐time substitutions. This framework combines recent analytical advances in rarefaction with a riverscape perspective, and can be applied to conserve streams, and their biota, in riverscapes around the world.

No abstract