Aquatic Invertebrate Responses to Riparian Restoration and Flow Extremes in Three Degraded Intermittent Streams: an Eight‐year Field Experiment

1. Worldwide, riparian zones of many intermittent rivers and ephemeral streams (IRES) have been severely degraded. However, compared with perennially flowing waters, there have been few restoration efforts (e.g., revegetation, livestock removal) to halt or reverse this degradation. Even rarer are assessments of whether desired responses occurred and, if so, their underlying mechanisms, especially where flow extremes such as drying and flooding in IRES might alter or delay ecological responses to restoration. 2. In order to assess how aquatic invertebrate communities responded to riparian replanting and livestock removal in three degraded lowland IRES in southeastern Australia, we conducted an eight-year experiment (2004-2013), sampling paired control (unrestored) and treatment (restored, between 2005 and 2007 across creeks) sites twice yearly before and after restoration. To evaluate likely mechanisms underlying invertebrate responses to riparian restoration, we used a range of structural (e.g., richness) and functional (e.g., trait-based) metrics. We also monitored environmental variables predicted to be likely drivers of aquatic invertebrate responses such as macrophyte and canopy cover, water quality, in-stream organic matter and instream wood. The study coincided with the last few years of a decade-long drought that broke with severe floods (in 2010-2012), so we opportunistically examined how these hydrological extremes affected invertebrate responses to riparian restoration. 3. At all three sites, macrophyte and canopy cover, water quality, organic matter and instream wood were comparable at treatment and control sites throughout the eight-year experiment. There were also no significant differences in structural or functional metrics to indicate aquatic invertebrate responses to restoration after 6-8 years. This lack of predicted invertebrate responses apparently reflected how severely the drought impeded the riparian restoration, delaying even short-term responses. Pervasive effects of catchment-scale degradation may have further diminished the predicted benefits of the reach-scale restoration efforts. 4. By contrast, invertebrate metrics responded strongly to flow extremes, especially the brief floods that occurred as drought broke. These responses were evident for functional groups and individual families (especially rheophilic ones), but were not detected by a commonly used integrative measure of stream condition. This highlights the importance of selecting appropriate indicators in restoration monitoring programs and recognising their variability in responsiveness to restoration versus flow extremes. 5. IRES are abundant and, in many parts of the world, increasingly common. Understanding how their inherent hydrological extremes may delay, modify and even overwhelm predicted responses to riparian restoration is crucial to setting realistic goals for IRES in the Anthropocene. Furthermore, assessing the success of riparian restoration in IRES requires a long-term view given the extended time frames over which responses may occur, especially when expectations are often based on results of riparian restoration of perennial streams and rivers.