Habitat loss and subdivision are additive mechanisms of fish extinction in fragmented rivers

Nevertheless, authors have increasingly emphasized fragment length as a critical metric for conservation (Table S1). Some have discounted habitat loss altogether, attributing effects of fragmentation to subdivision alone. They propose that the main significance of fragment length is decreased propagule retention in shorter river fragments (Wilde & Urbanczyk, 2013; Worthington et al., 2014). This hypothesis applies only to minnows that produce pelagic-broadcast eggs, so it does not explain why fishes with other spawning modes also succumb to fragmentation (e.g. Brown, 1986). Further, the ‘drift distance’ required by pelagic-broadcast spawning minnows is undetermined. Habitat conditions determine locations of populations within a fragment (Welker & Scarnecchia, 2004; Hoagstrom et al., 2006), so drift distance available to propagules does not necessarily equal fragment length. Potential drift distance depends on the location of spawning aggregations and nurseries (Alò & Turner, 2005), which are rarely, if ever, known. Also, propagule retention is reduced in narrowed river channels (Dudley & Platania, 2007) and dams that fragment rivers on the plains cause up to 91% reductions in active-channel area (Graf, 2006). One can only imagine the retentive capacity lost in these cases. Hence, not only is empirical evidence that drift distance alone can cause population declines lacking but also habitat loss is intimately associated with impoundments that fragment rivers and increase drift distance.

Lack of evidence for drift-related extinction could reflect difficulty in documenting this mechanism. Some historic extinctions may have been due to drift truncation and it may threaten some existing populations (Osborne et al., 2005). Regardless, habitat loss remains as a salient factor. Impounded rivers of the plains are ‘shrunken, simplified versions of former unregulated rivers’ (Graf, 2006) and all studies that estimated minimum fragment length for pelagic-broadcast spawning minnows also found that flashy flow regimes (Dieterman & Galat, 2004; Dudley & Platania, 2007) or persistent base flows (Perkin & Gido, 2011; Perkin et al., 2013; Wilde & Urbanczyk, 2013; Worthington et al., 2014) were associated with relict populations. This supports observations that high-quality river-fragments on the plains invariably sustain a consistent flow regime (Hoagstrom et al., 2011). Flow variability bolsters geomorphic diversity (Graf, 2006), facilitates pelagic-broadcast spawning (Hoagstrom & Turner, 2013), generates expansion-contraction cycles that amplify ecosystem complexity (Stanley et al., 1997), and fosters propagule retention (Dudley & Platania, 2007). Hydraulic retention supports many ecosystem functions, such as creating fish nurseries (Schiemer & Hein, 2007). Persistent base flows provide fluvial habitats required by juveniles and adults (Hoagstrom et al., 2008; Wild & Durham, 2013).

Several factors increase extinction risk in shorter reaches: (i) more direct impacts of dams lessen habitat suitability and retentive capacity (Dudley & Platania, 2007; Hoagstrom et al., 2008); (ii) retentive capacity is less due to fragment length (Wilde & Urbanczyk, 2013; Worthington et al., 2014); (iii) less habitat in shorter fragments supports smaller populations more susceptible to flux-related extinction (Hubbell, 2001); and (iv) nonnative species are more prevalent near impoundments (Quist et al., 2004). Importance of each factor is undetermined because no extinctions were directly observed, but impacts of habitat change and faunal flux (1, 3, and 4) are well documented in field studies (Table S2).

River fragments are dynamic landscape features. Resident faunas face continued, long-term degradation for many decades, or longer (Petts & Gurnell, 2005). Extinction debts (Hoagstrom et al., 2011) may exist due to chronic habitat degradation, ongoing resource depletion, reduced capacity to withstand climatic variability, or genetic drift (Table S2). Conservation efforts must identify habitats necessary for all life-history stages and assess their integrity (Rosenfeld & Hatfield, 2006). Place-based efforts focused on fragment-specific threats are most likely to be effective (Palmer et al., 2009) because threats are caused by local controls, such as operating rules for dams (Graf, 2006) or disturbances including urbanization, agriculture, and pollution (Stanford & Ward, 2001; Hoagstrom, 2009).

Longer fragments deserve conservation attention because they are most likely to shelter pseudo-natural reaches. Relictual riverscapes persist within intact landscapes where unregulated tributaries restore native flow regimes and channel dynamics (Johnson, 2002; Sabo et al., 2012). These de facto river refuges are natural laboratories where mechanisms of population persistence and extinction threat can be studied. But, they require immediate protection from further degradation (Hoagstrom et al., 2011), from appropriation of water resources (Gleeson et al., 2012), and from effects of climate change (Perry et al., 2012). This can only be done if threats are accurately identified.