Environmental flow rules are developed to provide a flow regime necessary to maintain healthy river and floodplain ecosystems in rivers regulated for human uses. However, few studies have experimentally assessed potential ecological mechanisms causing declines in the health and productivity of freshwater fish assemblages in regulated rivers to inform the development of appropriate environmental flows.We tested whether an experimental flow release in a regulated tributary of the Hunter River, Australia, altered the diet of two widely distributed fish species (Australian smelt 'Retropinna semoni' and Cox's gudgeon 'Gobiomorphus coxii') compared with data from unregulated reference and regulated control tributaries. Neither species had significant differences in the number of prey taxa ingested, gut fullness or composition of gut contents due to the environmental flow release (EFR). The diet of 'R. semoni' did not differ significantly between regulated and unregulated tributaries in either catchment. However, the diet of 'G. coxii' differed in only one of the two pairs of rivers consistently across all sample times. Assuming the EFR was sufficient to alter the composition of prey available for consumption by the fish species studied, our findings imply that functional indicators, such as the diet of generalist higher-order consumers, may be more suitable indicators of long-term flow regime change rather than short-term flow events.

Defining appropriate environmental flow regimes and criteria for the use of environmental water allocations requires experimental data on the ecological impacts of flow regime change and responses to environmental water allocation. Fish assemblages in one regulated and one unregulated tributary paired in each of two sub-catchments of the Hunter River, coastal New South Wales, Australia, were sampled monthly between August 2006 and June 2007. It was predicted that altered flow regime due to flow regulation would reduce species richness and abundance of native fish, and assemblage composition would differ between paired regulated and unregulated tributaries. Despite significant changes in richness, abundance and assemblage composition through time, differences between regulated and unregulated tributaries were not consistent. In February 2007, an environmental flow release ('artificial flood') of 1400 ML was experimentally released down the regulated tributary of one of the two catchments over 6 days. The flow release resulted in no significant changes in fish species abundances or assemblage composition when compared to nearby unregulated and regulated tributaries. Flow regulation in this region has reduced flow variability and eliminated natural low-flow periods, although large floods occurred at similar frequencies between regulated-unregulated tributaries prior to and during 2006-2007, resulting in only moderate changes to regulated flow regimes. Barriers to dispersal within catchments also compound the effects of flow regulation, and findings from this study indicate that the location of migratory barriers potentially confounded detection of the effects of flow regime change. Further experimental comparisons of fish assemblages in regulated rivers will refine river-specific response thresholds to flow regime change and facilitate the sustainable use of water in coastal rivers.