Dryland stream fishes are adapted to highly connected habitats with unpredictable hydrologic conditions, including frequent low flows and sometimes extreme drought.The low flow recruitment hypothesis predicts that some fishes spawn in main channel habitats during low flows when water temperatures and prey densities are high.However, extreme low flows during drought periods might be disruptive even among fishes whose life histories otherwise benefit from lower flows. We studied recruitment dynamics of six fishes (family Cyprinidae) at 15 sites in a fragmented Great Plains riverscape in Kansas, USA, during 2 years of extreme drought. We tracked the chronology of gonadal development and age 0 recruitment to test the hypothesis that recruitment by fishes that broadcast spawn planktonic ova would be less successful compared with fishes that spawn demersal or adhesive ova. We found all six taxa were reproductively active but recruitment was evident for only four.The two species for which recruitment was not evident, peppered chub (Macrhybopsis tetranema) and silver chub (Macrhybopsis storeriana), are confirmed or suspected pelagic-broadcast spawning fishes that have declined in other fragmented and dewatered Great Plains riverscapes. Our data highlight the potential for extreme low flows within isolated stream fragments to cause complete or near-complete recruitment failure for pelagic-broadcast spawning fishes, especially those with small population sizes. Failed recruitment during extreme drought may be related to spawning mode, ova characteristics, or both. Our work informs management of fish diversity in drought-prone riverscapes by establishing mechanistic linkages among extreme drought, fish recruitment, and assemblage structure.

No abstract

Reservoirs and associated river fragments are novel ecosystems not experienced by fishes in their evolutionary history, yet they are now commonplace across the globe. Understanding how fishes use these novel habitats is vital to conservation efforts in contemporary riverscapes. Movement patterns of the endangered razorback sucker (Xyrauchen texanus) synthesized from tagging efforts in the upper Colorado River basin, USA, illustrate the applications of tagging technology and data sharing by multiple agencies to better understand the spatial ecology of large river fishes. Tagging studies between 2014 and 2018 in Lake Powell and its two main tributary rivers, the Colorado (unfragmented) and San Juan (waterfall‐fragmented), were used to quantify movement of razorback sucker within this river–reservoir habitat complex. In addition, facilitated translocations of fish upstream of a waterfall barrier in the San Juan River were assessed in 2016–2017. Extensive movement of fish occurred within and across river and reservoir habitats. Of 722 fish captured in the Colorado River arm of Lake Powell, 36% of re‐encounters occurred upstream in the Colorado or Green rivers, or fish dispersed through the reservoir and were detected in the San Juan River arm. Fourteen fish moved more than 600 km. In the San Juan arm of the reservoir, 29% and 20% of fish in 2017 and 2018, respectively, had moved ~30–40 km upstream below the waterfall in the San Juan River within a year. In 2016–2017, 303 fish were translocated upstream of the waterfall into the San Juan River, but 80% were re‐encountered downstream of the waterfall within a year. Long‐distance movements by razorback sucker were common within and among rivers and reservoirs illustrating how large river fish, in general, might maintain population connectivity in highly altered ecosystems.

Disturbance regimes of arid regions throughout the world are changing along with global warming. Severity of drought associated with decreased stream discharge and increased frequency and intensity of wildfires are increasing in many of these systems. Combined, these factors can have potentially devastating effects on stream fish communities. To quantify the response of fish communities to annual variation in climate, hydrology, and wildfire, we used long‐term and spatially extensive datasets from the Gila River drainage, NM, U.S.A. Additionally, we tracked the recovery of multiple fish communities within the study area following extreme drought and associated wildfires that occurred between 2011 and 2013. Twenty‐nine years of monitoring at six sites, beginning in 1988, indicated that fish community richness and density were negatively associated with drought, as quantified by the Palmer Drought Severity Index. Ordination of fish species densities across these sites further suggested that native species were more sensitive to inter‐annual variation in climate than non‐native fishes. The resistance and resilience of fishes to drought and ash flows following the most severe series of wildfires on record between 2011 and 2013 were highly variable across sites and species. The percent of common species that were able to resist a significant decline ranged from 0 to 70%, depending on location. As of 2017, all but two species (one native and one non‐native) had recovered at one or more sites. Although we hypothesised that the response to wildfire and drought that occurred between 2011 and 2013 should vary with body size and species provenance (native or non‐native), this was not supported. Rather, resistance of fish communities to these events appeared to be driven primarily by spatial factors, such as intensity of disturbance, and resilience was driven by species ability to rapidly rebound from severely depressed numbers or immigrate from nearby refuge populations. Current climate trends suggest an increase in the intensity of disturbance associated with wildfire and drought; thus, maintenance of native biodiversity should focus on preservation of habitat integrity and connectivity among habitats that allows for recolonisation of disturbed habitats.

Increasing trends in fragmentation and dewatering of streams warrants research on how populations and communities respond to varying water levels and barriers to movement. Although these responses are complicated by many spatial and temporal processes, long‐term datasets might help reveal complex patterns and processes driving variability in species abundances. The objective of this study was to develop a predictive framework for fish community and population responses to varying levels of water availability across six sites in two intermittent stream networks sampled >10 years. We predicted that fishes would emigrate into intermittent reaches during wet conditions; thus, overall abundances within perennial source locations will decline. Accordingly, when intermittent reaches dry, fishes will contract to wetted habitats resulting in high abundance. Observed fish community abundances were highly variable within and among study sites, but four of six sites matched our predictions. A tagging study confirmed these results and demonstrated a substantial proportion of individuals moved away from perennial reaches and into newly wetted intermittent reaches. However, site and species‐specific relationships were variable and likely depended on the habitat, metacommunity dynamics, and life history strategies. Findings suggest that species dispersal dynamics, in addition to recruitment and mortality, should be carefully considered when interpreting species responses to varying water levels, particularly in intermittent stream networks where access to habitat can change drastically with water availability.