Recolonization of Anadromous Fish in the Cedar River above Landsburg Diversion Dam: A Ten-Year Evaluation P. Kiffney, T. Buehrens,† G. Pess, S. Naman, and T. Bennett Report of research by Fish Ecology Division, Mukilteo Biological Field Station Northwest Fisheries Science Center, National Marine Fisheries Service, National Oceanic and Atmospheric Administration 2725 Montlake Boulevard East Seattle, Washington 98112-2097 and †School of Aquatic and Fishery Sciences University of Washington Box 355020 Seattle, WA 98195 for Seattle Public Utilities 700 Fifth Avenue, Suite 4900 PO Box 34018 Seattle, WA 98124-4018 Contract No. DA2008-15 May 2011 ii EXECUTIVE SUMMARY As part of the City of Seattle Habitat Conservation Plan, a fish ladder was constructed at the Landsburg Diversion Dam to provide passage of Pacific salmon Oncorhynchus spp. to ~33 km of available habitat. In September 2003, adult salmon passed above Landsburg Dam for the first time since 1900. This report describes changes in fish populations, communities, and ecosystem attributes in the Cedar River above the dam before (2000 and 2001) and after (2004-2009) installation of the Landsburg Fish Passage Facility. The objectives of these collective studies were as follows: 1) Determine the efficacy of the fish passage facility in restoring anadromous salmon above Landsburg Dam 2) Document the colonization dynamics of salmon following habitat reconnection, 3) Determine whether salmon had measurable ecological effects on water chemistry, trophic relationships, or resident fish species 4) Evaluate the movement, growth, and survival of coho O. kisutch and rainbow trout O. mykiss and identify environmental covariates associated with variation in these individual traits; 5) Evalute habitat-fish associations to identify specific habitat or reach types that support high fish density, growth, and survival In each August from 2005 to 2009, we surveyed sections (ca. 200-800 m long) of main stem Cedar River habitat within 10 reaches established in 2000. Sections surveyed started above the diversion pool upstream from Landsburg Dam (reach 1 or CR1) and extended to Cedar Falls (CR10), a natural barrier to Pacific salmon. In 2004, reaches CR1-CR6 were surveyed. During these surveys, physical habitat was quantified and fish were identified and counted using snorkeling. Similar habitat and fish surveys were conducted in the main stem during winters 2007-2009. We also conducted seasonal (spring, summer, fall) snorkel surveys of multiple pools to monitor colonization of Rock Creek by juvenile coho. Beginning in 2005 we tagged and recaptured coho and trout in Rock Creek and the main stem Cedar above Landsburg to monitor fish movement, growth, and survival. To quantify the effects of salmon recolonization on ecosystem processes, tissue samples were collected from riparian vegetation and the stream food web (periphyton, invertebrates, sculpin Cottus spp. rainbow trout, and coho salmon) in September 2000-2001, 2004, and 2007-2009 (only fish were sampled in 2008) to measure concentrations of carbon and nitrogen isotopes. Fish collected below Landsburg Dam by iii King County were also analyzed, and provided a useful benchmark to examine whether ecosystem processes (nutrient flux from the ocean to freshwater) affected by salmon had changed with recolonization. Stomach contents were also analyzed from many of the fish sampled for isotopes. These samples allowed us to determine the importance of fish in the diets of resident trout and provided additional insight into trophic relationships in the Cedar River and its tributaries. Juvenile salmon, especially coho, have rapidly dispersed and colonized multiple habitats within the main stem and Rock Creek. Juvenile coho are now the most abundant salmonid in lower 5 km of Rock Creek and the lower 15 km of the main stem. Coho densities appear to have stabilized in reaches closest to the dam, but this pattern can certainly change, as it is still early in the recolonization process. Upstream expansion in Rock Creek appeared to be limited by the culvert at the 41 road; however, beginning in spring 2009, coho salmon young-of-the-year were observed in the first reach above this culvert, indicating spawning by coho the previous fall/winter. The reconnection of Walsh Creek with Rock Creek during winter 2009 provided an additional opportunity for spatial expansion of salmon into the Walsh Lake subbasin. Young-of-year coho were observed in sections immediately above the diversion gate in spring 2010, with adult coho spawning in Walsh and Webster Creeks during fall/winter 2009-2010. Low-velocity habitat with abundant cover was the preferred habitat for juvenile coho in both summer and winter. The lower 2 km of Rock Creek provides an abundance of these habitat types, which may be limited in the relatively confined main stem Cedar River. Beaver ponds, and similar wetland and lacustrine habitats with abundant edge habitat, are typically more productive overwintering habitats for coho than tributaries and mainstem areas of confined rivers. For this reason, it is expected that further improving access to the upper portions of the Rock and Walsh Creek basins through planned restoration actions will increase the productivity of coho populations above Landsburg Dam. The reconnection of this junction has also provided an opportunity for invasive species to move out of Walsh Lake into habitat that was previously free of non-native fish. Two largemouth bass Micropterus salmoides were captured in summer 2009 during electrofishing surveys of Rock Creek, while one bass was observed during a snorkel survey of Rock Creek in 2010. Juvenile Chinook has also increased in abundance and distribution; we found juvenile Chinook up to reach CR9 in the main stem Cedar River, which is about 17 km upstream from the dam. However, Chinook summer abundance was orders of magnitude lower than that of coho, and Chinook were found primarily in the main stem. Lower summer abundance of Chinook relative to coho was not surprising iv given that Chinook primarily exhibits an ocean-type life history. During summer surveys in 2009 and 2010, two fish taxa (Pacific lamprey Lampetra tridentata and steelhead trout O. mykiss) were observed for the first time since our study began, and a third species (bull trout) observed for only the second time. Analyses of habitat-fish associations showed that juvenile and larger size classes of salmonids were positively associated with cover, including undercut banks, overhanging riparian vegetation, and especially cover provided by wood. We also observed that food resources were positively associated with total salmonid density and growth of juvenile coho. Thus, both cover and prey availability were positively related to fish populations. The experimental study clearly demonstrated that increased prey availability causes increased growth and body size of juvenile salmonids. There was contradictory evidence to suggest that the return of Pacific salmon has affected ecosystem processes (i.e., energy flow). For example, in reach CR1 of the main stem Cedar River, C13 levels in resident fish were higher after than before the ladder was installed; this reach received the highest inputs of salmon-derived nutrients as determined by abundance of salmon nests. Although some patterns in isotope levels above the dam were suggestive of positive salmon-nutrient feedback, levels of N15 in fish collected below the dam in 2008 were twice as high as those in fish above the dam. These data corroborated samples collected from periphyton and invertebrates from Rock Creek below Landsburg Dam in previous years, which showed N15 levels higher in tissue from this tributary below compared above the dam. In terms of trophic relationships, diet analysis (n = 1,263 diet samples analyzed) revealed that piscivory was relatively low, ranging from 0 to 4%. We found that fish became piscivorous when they exceed 100 mm in fork length. The proportion of fish exhibiting piscivory was similar before vs. after the ladder was installed. Our results show a significant shift in fish populations and communities in the Cedar River and Rock Creek as a result of Pacific salmon recolonization. The rapid colonization by salmon of the Cedar River above Landsburg Dam emphasized their innate ability to colonize newly available habitat. We observed, however, that expansion can be limited by small, artificial barriers (e.g., culverts, diversion gates) and species. Specifically, the coho population has generally increased over time, whereas Chinook population dynamics were more variable, with no clear time trend. This difference may reflect life history differences, with coho populations increasing because of their shorter life span and longer freshwater residence. Despite these differences, salmon now occur in most of the accessible habitat except for reach CR10 of the main stem, which lies above numerous cascades and a bedrock canyon, and Rock Creek above the major wetland complex. In contrast to the v expanding Pacific salmon populations, especially coho, trout populations have remained relatively stable over time. In 2009 juvenile coho density surpassed that of trout in Rock Creek for the first time; yet consistent with previous years, there has been no observable negative association between coho density and trout population dynamics or individual performance. Although salmon populations are generally increasing above the Landsburg Diversion Dam, we found limited evidence to support the hypothesis that salmon recolonization has affected nutrient dynamics as determined by carbon and nitrogen isotopes. Isotope values from sites below the dam, which have experienced continual exposure to anadromy, were almost twofold higher than those from sites above the dam. This comparison indicated little or no evidence that a positive feedback effect of adult salmon nourishing future generations has returned to Cedar River above Landsburg Dam. Our experimental study showed that such a feedback would occur at a biomass loading of 0.6 kg/m2. By comparison, cumulative biomass loading in CR1, which has experienced the highest numbers of salmon redds, was ca. 0.014 kg/m2. This biomass load remains an order of magnitude lower than streams in Alaska that have robust salmon populations, or in Griffin Creek, a productive coho tributary to the Snoqualmie River. In summary, the Landsburg fish passage facility has been successful in reconnecting the Cedar River above and below the dam for Pacific salmon. Pacific salmon rapidly colonized and established populations above Landsburg Dam. They now occur from above Landsburg Dam to CR9, which is about 20 km upstream. Salmon have also colonized Rock Creek, and are now observed in the lower portions of the wetland complex about 3 km upstream from the confluence with the main stem. Coho appear to be more successful in establishing a self-sustaining population relative to Chinook, but given the longer life cycle of Chinook, positive population growth trends may take longer to manifest. As mentioned in previous reports, increasing the complexity of edge habitat in the main stem Cedar River may accelerate population growth of Chinook. Both winter and summer surveys have demonstrated that this edge habitat supports some of the highest densities of juvenile trout, coho and Chinook, especially when associated with wood cover. Our surveys indicated that wood additions would most benefit juvenile salmon populations in CR1, CR2, CR4, CR6 and CR8, all relatively low-gradient reaches that support relatively high densities of juvenile salmon. The partial reconnection of Walsh Creek to Rock Creek allowed adult coho salmon to spawn in Walsh and Webster creeks in winter 2009-2010, and coho fry were observed rearing there in spring 2010. Juvenile coho densities in Rock Creek are now similar to those of other streams in the Pacific Northwest that have had uninterrupted vi anadromy, while juvenile Chinook densities are lower than those found in similar-sized Puget Sound Rivers (K. Bartz, unpublished data). We anticipate coho populations will continue to increase, given the recent colonization of Walsh Lake subbasin and the anticipated reconnection of Walsh and Rock Creek subbasins. Diet and isotope analysis showed little evidence that resident fishes are directly or indirectly benefiting from salmon-derived nutrients or energy. Energy and nutrients (N and P) provided by adult salmon have increased over time, but they are extremely low relative to those found in healthy salmon streams, with little evidence that these amounts have affected nutrient flux. Thus, salmon have re-established populations, but they remain orders of magnitude lower than necessary to achieve a positive feedback effect of adult salmon on future generations. vii CONTENTS EXECUTIVE SUMMARY ............................................................................................... iii INTRODUCTION .............................................................................................................. 1 1.0. HABITAT CHARACTERIZATION.......................................................................... 3 Methods................................................................................................................... 3 Results ..................................................................................................................... 5 2.0. JUVENILE SALMONID RECOLONIZATION ....................................................... 6 2.1. Mark-Recapture Study .................................................................................... 6 Methods....................................................................................................... 6 Results ......................................................................................................... 7 Fish Capture and Tagging ............................................................... 7 Recapture Rates ............................................................................ 10 Patterns of Fish Movement ........................................................... 12 Growth .......................................................................................... 14 Trout Genetic Markers .................................................................. 16 2.2. Snorkel Surveys ............................................................................................ 18 Methods..................................................................................................... 18 Results ....................................................................................................... 19 Density and Species Composition in the Main Stem Cedar River 19 Density and Species Composition in Rock Creek ........................ 30 Density in Williams Creek ............................................................ 34 Effect of Habitat on Abundance ................................................... 35 Abundance vs. Prey Availability .................................................. 37 3.0 ECOSYSTEM STUDIES .......................................................................................... 38 3.1. Carcass Analog and Small Wood Experiments ............................................ 38 Methods..................................................................................................... 38 Results ....................................................................................................... 41 3.2. Adult Salmon Biomass Input ........................................................................ 48 3.3. Diet Content Analyses .................................................................................. 49 Methods..................................................................................................... 49 Results ....................................................................................................... 49 3.4. Carbon and Nitrogen Isotopes ...................................................................... 52 DISCUSSION ................................................................................................................... 55 Recolonization by Juvenile Pacific Salmon .......................................................... 55 Density and Habitat Use ........................................................................... 55 Trends in Community Structure................................................................ 57 Ecosystem Studies ................................................................................................ 61 Carcass Analog and Small Wood Experiments ........................................ 61 Adult Salmon Biomass Input .................................................................... 62 Diet Content Analyses .............................................................................. 62 Carbon and Nitrogen Isotopes .................................................................. 62 viii ACKNOWLEDGEMENTS .............................................................................................. 65 REFERENCES ................................................................................................................. 66 APPENDIX: Study Summary .......................................................................................... 72 ix INTRODUCTION Fishing, habitat loss and degradation, poor hatchery practices, climate change, and non-native species are the main causes of decline for Atlantic Salmo salar and Pacific Oncorhynchus spp. salmon populations (NRC 1996; Montgomery 2003). Some of the primary culprits in habitat loss are barriers to fish migration, such as road crossings, levees, and dams, which block access to upstream and floodplain habitats. Obstruction to fish passage is a problem that has been documented throughout North America (USGAO 2001; Langill and Zomora 2002) and Europe (Yanes et al. 1995; Glen 2002). In Washington State, over 7,700 km of historical salmon habitat are inaccessible to migratory fishes because of impassable culverts or road crossings, despite state regulations requiring fish passage (Roni et al. 2002). In the United States, many salmon occupying truncated river systems have precipitously low population levels, and several have recently been listed as either threatened or endangered under the United States Endangered Species Act (NRC 1996; Montgomery 2003). Removal of a blockage, whether it is a small culvert or a series of dams in a large watershed, is considered a key restoration action to aid in the recovery of listed salmon. These actions are currently being implemented across North America and will likely become more prevalent in the next 5-10 years (Roni et al. 2002). Although much effort has been made to remove blockages to salmon passage, surprisingly little is known about why salmon colonize new habitats and what occurs after a barrier is removed. For example, What are the key environmental factors that determine salmon colonization success? What restoration actions might promote colonization success? As part of the City of Seattle Habitat Conservation Plan (HCP) for the Cedar River Watershed, a fish ladder was opened at the Landsburg Diversion Dam located on the Cedar River main stem in September 2003. For over 100 years, this diversion had blocked access to approximately 33 km of main stem and tributary habitat, potentially contributing to population declines of a number of anadromous fish species and resulting in the loss of important food resources for a variety of other species. An additional 10 km of lake and river habitat was made accessible with reconnection of the Walsh Lake subbasin to Rock Creek during winter 2009. It has been shown in other studies that salmon carcasses provide important nutrient subsidies to their natal streams and the surrounding terrestrial ecosystem (Bilby et al. 1996; Willson et al. 1998; Chaloner and Wipfli 2002). In addition, resident fishes above Landsburg Dam have been isolated from anadromous salmon for a number of generations; thus there are likely to be ecological effects (e.g., competition, predation) on these resident fishes resulting from the return of anadromous species.