Hydrologic and environmental thresholds in stream fish assemblage structure across flow regimes

The characteristic pattern of variation in flow magnitude, frequency, duration, timing, and rate of change defines the flow regime of rivers and streams and is a key driver of ecosystem processes in fluvial ecosystems. Understanding how freshwater biotic assemblages change across gradients of hydrology and anthropogenic-source disturbance in different streamflow regimes is crucial to managing for sustainable environmental flows and watershed conservation. We compiled long-term (1916–2016) occurrence records for fishes collected in the Ouachita-Ozark Interior Highlands and West Gulf Coastal Plain streams, together with hydrologic metrics calculated from daily streamflow data measured at USGS stream gauging stations (n = 111), to examine important drivers and thresholds for fish assemblage turnover in groundwater (GW), runoff (RO), and intermittent (INT) flow regimes. We also examined the importance of spatial gradients (latitude, longitude, elevation, drainage area) and anthropogenic-source stressors (Hydrologic Disturbance Index; HDI) for fish assemblage turnover using a gradient forest modeling approach. Watershed fragmentation was of high importance for fish assemblage turnover in RO and INT streams, while changes in dam storage were more important for fishes in GW streams. Hydrologic metrics describing seasonal and stochastic properties of daily streamflow (Mag6) were most important for fish assemblage turnover in INT streams. Timing of high flow events had significantly higher importance compared to flow magnitude, duration, and frequency metrics, especially for fish assemblages in GW and INT streams. The frequency and timing of low flow events had high importance for fish assemblage turnover across all stream flow classes, while the magnitude of low flows and the magnitude and rate of change of average flows was most important for INT stream fish assemblages. In addition to benefiting multi-species conservation and management actions through identification of local and regional flow-ecology relationships generalized across different flow regimes, the results of this study provide a better understanding of complex nonlinear threshold effects, which is critical to anticipating changes in aquatic ecosystems and communities.