- Streams across the western United States are impaired from human alterations that have reduced freshwater habitat by simplifying channel complexity and disconnecting floodplains (Knox et al., 2022; Waples et al., 2008; Wohl, 2014). Climate change is likely to continue exacerbating these risks by warming summer surface stream temperatures (Crozier et al., 2019), reducing summer discharge (Isaak et al., 2017) and increasing flood frequency and intensity that can scour rearing habitat for juvenile salmon (Wainwright & Weitkamp, 2013). Scientists and practitioners have increasingly turned to process-based restoration to increase resilience of streams in the face of these challenges. This approach aims to restore watershed processes that generate desired habitat conditions, as opposed to efforts to directly restore desired conditions that are not maintained naturally (Beechie et al., 2010; Wheaton et al., 2019).
The role that beaver can have in facilitating process-based restoration has garnered much attention, particularly as they relate to juvenile salmon habitat (Beechie et al., 1994; Pollock et al., 2004), through the influences they exert by constructing dams in flowing waters (Jordan & Fairfax, 2022). North American beaver (Castor canadensis; hereafter beaver) occupy a wide range of aquatic habitats across the continent including lakes, wetlands and rivers (Baker & Hill, 2003). Beaver are perhaps best known for altering smaller, ‘wadable’, streams, principally through dam construction, which can create large bodies of water that they use for access to forage and refuge from predators (Allen, 1982; DeStefano et al., 2006). Studies evaluating the influence of beaver related-restoration (BRR) have shown improvements to aquatic ecosystems (Weber et al., 2017), though responses can be variable (Collen & Gibson, 2000; Ecke et al., 2017; Grudzinski et al., 2022) and uncertain due to a lack of monitoring (Pilliod et al., 2018).
This dissertation considers the role that BRR could have in improving stream habitats used by coho salmon in Oregon coastal streams based on evaluation of: 1) summer responses of water temperature and dissolved oxygen concentration to beaver dams, key water quality parameters for juvenile coho salmon; 2) distributions of beaver and beaver dams across a large watershed extent used by coho salmon, and; 3) the influence that beaver dams may have on the quality of juvenile coho salmon rearing habitats based on the combination of winter and summer habitat limitations.
We considered these objectives across the Umpqua River Basin (Umpqua), a large coastal drainage in southern Oregon (12,124 km2) with an extensive network of stream habitat used by coho salmon (Anlauf et al., 2009). Our evaluation of stream responses during the summer of 2019 in and below beaver dams showed that conditions for coho salmon declined and that streams warmed below beaver dams compared to baseline conditions but that the heating signal dissipated with downstream distances. We also found that dissolved oxygen conditions in the ponds that we monitored during late summer in 2019 were anoxic for much of the monitoring period and unsuitable for salmonids. Our surveys of beaver distributions in the Umpqua indicated that beaver were active across approximately half of the surveyed streams considered suitable for damming but that presence of dams was rare. We also evaluated damming behavior across a larger extent of coastal watersheds used by coho salmon from 17 years of stream surveys and found that dam presence was most strongly, and negatively, associated with high stream flow that likely caused these structures to fail.
Our last study considered the distribution of streams that were suitable for coho salmon and beaver damming and the relative influences of beaver dams on two limiting factors to habitat quality in streams available to juvenile coho salmon in the Umpqua: summer stream temperature and winter flooding. We found that the distribution of suitable coho salmon habitat occurred in just over one third of Umpqua streams but that the co-occurrence with streams that were also suitable for beaver damming was less than five percent of the Umpqua. This finding suggests that there may be limited opportunities to apply BRR methods in the Umpqua over a wide extent. We evaluated summer and winter stress responses to three scenarios in which 1) dams were not present (No Dams/Baseline) 2) dams warmed streams but reduced flood risk (Warm), and 3) dams cooled streams and reduced flood risk (Cool). We found that summer stress increased by a greater margin under the Warm scenario than it decreased under the Cool scenario, but that overall habitat quality increased by a greater margin in Cool scenario than it decreased in the Warm scenario.
To our knowledge this study is unique in considering the range of potential beaver behaviors and how they are related to landscape and river networks at a broad extent. Our results emphasize the numerous contingencies that drive outcomes from BRR actions, which begin with consideration of beaver itself (Nash et al., 2021). We found that successful application of BRR will consider the range of potential stream responses to beaver dams and the spatial extent that streams prioritized for restoration co-occur in locations that beaver or human constructed impoundments can persist.