A partnership between Oregon State University, Real Time Research, and the USGS - Oregon Cooperative Fish and Wildlife Research Unit

FEATURE STORIES

Avian Predation in the Columbia River Basin, 2021Avian Predation in the Columbia River Basin, 2021

Three separate management plans have been developed and are now being implemented to reduce the impacts of avian predation on the survival of juvenile salmonids (Oncorhynchus spp.) in the Columbia River basin (CRB). The targets of management include the Caspian tern (Hydroprogne caspia; hereafter referred to as “tern” or “terns”) and the double-crested cormorant (Nannopterum auritum; hereafter referred to as “cormorant” or “cormorants”) breeding colonies on East Sand Island in the Columbia River estuary (CRE) and the tern colonies on Crescent Island (McNary Reservoir) and on Goose Island (Potholes Reservoir) in the Columbia Plateau region (CPR). The primary goal of these management initiatives is to reduce predation rates (percentage of available smolts consumed) by piscivorous waterbirds on Endangered Species Act (ESA)-listed anadromous salmonid populations (referred to as Evolutionarily Significant Units [ESUs] or Distinct Population Segments [DPSs]) through reductions in the size of or elimination of breeding colonies at each of these four sites. Management initiatives implemented at these four colony sites have been primarily non-lethal strategies for terns (i.e. passive and active nest dissuasion) and a combination of lethal (i.e. culling and egg-oiling) and non-lethal (i.e. nesting habitat management) strategies for cormorants. As part of the management plans for terns, the U.S. Army Corps of Engineers (USACE) created or enhanced alternative nesting habitat for terns at locations outside the CRB (i.e. northeastern California, southern Oregon, and south San Francisco Bay) to compensate for reductions in tern nesting habitat at East Sand Island in the CRE and for near elimination of tern nesting habitat at Crescent and Goose islands in the CPR.

In 2021, implementation of management plans was conducted by the USACE for the tern and cormorant colonies on East Sand Island and by the Bureau of Reclamation (BOR) at the tern colony on Goose Island; active tern nest dissuasion was not conducted on Crescent Island in 2021. The primary objectives of this study were to evaluate the efficacy of management to reduce avian predation on ESA-listed juvenile salmonids in the CRB, and to assess the magnitude of predation on smolts by other piscivorous waterbirds nesting at unmanaged colonies, namely those of California gulls (Larus californicus) and ring-billed gulls (L. delawarensis; hereafter referred to as “gull” or “gulls”) and American white pelicans (Pelecanus erythrorhynchos; hereafter referred to as “pelican” or “pelicans”). Specifically, we sought to (1) locate and estimate peak colony size for piscivorous waterbird species (terns, cormorants, gulls, pelicans) and colonies within foraging range of juvenile salmonids in the middle Columbia River, lower Snake River, lower Columbia River, and CRE; (2) estimate colony-specific predation rates on smolts by piscivorous colonial waterbirds; and (3) measure the cumulative effects of predation by piscivorous waterbirds from multiple breeding colonies on salmonid survival, including an investigation into the additive effects of avian predation on salmonid mortality. The continued monitoring of colony sizes, locations, and predation rates on salmonids by piscivorous waterbirds nesting at both managed and un-managed colonies in the basin will help ensure that the intended benefits of management efforts are achieved and sustained, and that the accrued benefits from colony management are not offset by the emergence of new and growing colonies at other locations within the basin. In short, this research will guide managers in developing and monitoring long-term management initiatives for avian predators that are science-based, defensible, cost-effective, and have a high probability of success.

Size of Piscivorous Waterbird Colonies
Columbia Plateau Region: The primary objective of the eighth year of implementation of the Inland Avian Predation Management Plan (IAPMP) was to limit the numbers of Caspian terns breeding on Goose Island and other islands in Potholes Reservoir, and on Crescent Island in McNary Reservoir, to less than 40 breeding pairs at each site and to less than 200 breeding pairs within the CPR to reduce impacts from tern predation on ESA-listed juvenile salmonids. To accomplish this task, the suitable tern nesting habitat at these sites was nearly eliminated by installing a variety of passive nest dissuasion materials on Goose Island prior to the 2021 breeding season, and by planting vegetation (in 2016) on Crescent Island. On both Goose and Crescent islands, passive nest dissuasion materials and/or vegetation covered all areas where terns had previously nested, as well as all areas of open, sparsely vegetated habitat that might be used by ground-nesting terns or gulls. Once terns arrived to initiate nesting in 2021, active nest dissuasion (i.e. human hazing) was used to dissuade terns from nesting on Goose Island and on other islands in Potholes Reservoir; these activities are described in detail in a separate report prepared by the BOR’s contractor (U.S. Department of Agriculture – Wildlife Services). Active hazing to prevent Caspian terns from nesting on Crescent Island was not necessary during 2015 to 2020; consequently, active nest dissuasion was not implemented at that site in 2021.

Despite the use of passive and active nest dissuasion techniques on Goose Island in each of the previous seven years, some terns continued to display high fidelity to the island as a nesting site in 2021. We estimated that 22 breeding pairs of terns were successful at raising young on Goose Island in 2021, up from 6 breeding pairs in 2020, and up from zero breeding pairs during 2016–2019. Successful tern nesting on Goose Island in 2021 was likely due, at least in part, to changes in the passive nest dissuasion array installed on the island prior to the 2021 nesting season. To create a passive nest dissuasion array that would require less in-season and between season maintenance, the braided rope that connected the stakes was replaced with wire prior to the 2021 breeding season. The flagging, a major component of the nest dissuasion array, was observed to slide down the wires during windstorms, leaving large areas of suitable tern nesting habitat without flagging during the 2021 breeding season. Efforts were made by the contractor to fix this problem during the 2021 field season; however, despite these efforts, terns were successful in establishing a colony and rearing young at a site on top of the island near the former tern colony site. The use of braided rope as part of the nest dissuasion array prior to 2021, although perhaps requiring more maintenance than wire, prevented the movement of flagging along the rope because the flagging was installed between the strands of rope. Similar to other years during the management period, terns were persistent in trying to nest along the shoreline of Goose Island in 2021, areas that became exposed during the nesting season as reservoir levels receded. These nesting attempts were not successful, however, likely due to active measures (i.e. human hazing) implemented by the contractor to prevent nesting in these areas.

A management effort that planted willows and other native vegetation on Crescent Island in 2016 eliminating most open, bare-ground habitat that terns prefer for nesting. Prior to management, there was also a large gull colony on Crescent Island. During the first five years of tern management on Crescent Island (2015–2019), nest dissuasion activities prevented both terns and gulls from nesting on the island. Beginning in 2020, and again in 2021, a gull colony became re-established on Crescent Island in and amongst the vegetation on the island. The recent die-back of vegetation on Crescent Island, perhaps caused by nesting gulls, beaver herbivory, and/or weather-related events (e.g., windstorms, drought), has created small patches of open, unvegetated habitat that is suitable for tern nesting. In 2021, one pair of Caspian terns was successful in rearing young on Crescent Island, the first year of successful tern nesting at that site since management commenced early in the 2015 nesting season. Without the implementation of mitigative measures to prevent further die-back of vegetation on Crescent Island, we expect that there will be an increasing number of terns nesting on Crescent Island in the future.

Aerial, ground, and boat-based surveys were conducted in the CPR to determine where terns displaced from the managed colonies in Potholes Reservoir and at Crescent Island might attempt to nest. In 2021, terns successfully nested at five different colony sites in the CPR: at both managed sites (Goose Island [22 breeding pairs] and Crescent Island [one breeding pair]) and at three unmanaged sites (Badger Island in McNary Reservoir [231 breeding pairs], Harper Island in Sprague Lake [85 breeding pairs], and Shoal Island in Lenore Lake [61 breeding pairs]). In 2021, a managed increase in the water level of John Day Reservoir during the 2021 Caspian tern breeding season eliminated all upland nesting habitat previously used by terns in the Blalock Islands. As a result, a tern colony was not established at that site in 2021, the first year that no tern colony formed in the Blalock Islands since the colony first became established in 2005. Most of the terns displaced from nesting at the Blalock Islands in 2021 apparently relocated to Badger Island to nest, where a total of 231 tern breeding pairs were successful in nesting at two sites on Badger Island, one near the shoreline and the other in the island’s interior amongst nesting gulls and pelicans. Prior to 2021, terns had not successfully nested on Badger Island since 2018, when a small colony of 8 breeding pairs nested there. We expect that terns will continue to nest on Badger Island in the future if there is no suitable tern nesting habitat available in the Blalock Islands and suitable unvegetated nesting habitat persists on Badger Island.

A total of 400 breeding pairs of Caspian terns nested at five different colonies in the CPR in 2021, up from 209 breeding pairs in 2020. This is the first increase in the regional breeding population of terns since 2017. The cause(s) of the increase in the regional breeding population of terns is unknown but may be related to the poor nesting season for terns in the CRE and elsewhere within their breeding range.

In addition to Caspian terns, we also estimated the colony size of other piscivorous waterbirds breeding in the CPR during 2021. Because visits to these colonies were less frequent (1–2 times per month) as compared to our monitoring of tern colonies in the region (2–4 times per month), there is more uncertainty in our estimates of peak colony size at these other piscivorous waterbird colonies (i.e. actual peak colony size could be somewhat higher). As was the case in 2020, there was a total of nine active gull colonies in the CPR during the 2021 breeding season, ranging in size from about 1,200 breeding individuals (Lenore Lake) to over 14,000 breeding individuals (Goose Island in Potholes Reservoir). The biggest change in gull colony size in 2021 relative to 2020 was at Crescent Island, where colony size grew more than 5-fold; a total of 400 breeding individuals were counted in 2020 and 2,084 breeding individuals in 2021. Double-crested cormorants were confirmed nesting at four colony sites in the CPR, with colony sizes ranging from 147 breeding pairs (Shoal Island in Lenore Lake) to 520 breeding pairs (Harper Island in Sprague Lake). Finally, the size of the Badger Island pelican colony was estimated to be 3,624 breeding individuals in 2021, larger than the estimated colony size in 2020 (3,165 breeding individuals) and one of the largest estimates since counts of the Badger Island pelican colony were first conducted in 2005.

Columbia River Estuary: In 2021, we estimated the size of the double-crested cormorant colony on the Astoria-Megler Bridge to be 4,151 breeding pairs, down from 5,081 breeding pairs in 2020 (J. Lawonn, ODFW, pers. comm.). Estimates of colony size at the Astoria-Megler Bridge should be considered a minimum estimate in that cormorants are now nesting in areas of the bridge (inside beams and girders) not visible from the boat-based and aerial vantages used to count nesting cormorants on the bridge. Prior to 2021, the cormorant colony on the Astoria-Megler Bridge had increased in size every year since 2011, with the recent exponential growth in size of the bridge colony corresponding with the precipitous decline in the size of the East Sand Island colony; no egg-laying by double-crested cormorants was observed on East Sand Island in either of the last two years (Brandtner and Tidwell 2021).

The USACE estimated that 2,050 breeding pairs of Caspian terns nested on the prepared 1-acre main colony site on East Sand Island in 2021 (Brandtner and Tidwell 2021). As was the case in 2020, the size of the Caspian tern colony on East Sand Island was significantly less than the target colony size of 3,125–4,375 breeding pairs stipulated in the Caspian Tern Management Plan for the Columbia River Estuary. Large numbers of terns (hundreds to thousands of adults per week) once again attempted to nest along the southeast and east beaches of East Sand Island in 2021, despite ongoing nest dissuasion efforts to prevent terns from nesting outside of the prepared 1-acre colony area. None of the tern eggs laid outside the 1-acre colony area apparently hatched, and none of the tern chicks hatched on the 1-acre colony area survived to fledging, with gull depredation of eggs and chicks being the biggest factor in tern nest failure on the East Sand Island colony in 2021. In 2020, the East Sand Island colony of Caspian terns produced very few, if any, young. As was the case in 2020, terns attempted to nest but were unsuccessful in establishing a colony on Rice Island or elsewhere in the upper Columbia River estuary in 2021 (K. Tidwell, USACE, pers. comm).

Predation Rates on Juvenile Salmonids
To investigate the impacts of predation by piscivorous colonial waterbirds on the survival of juvenile salmonids (smolts), and to determine the efficacy of on-going management actions to reduce predation, we estimated salmonid population (ESU/DPS)-specific predation rates based on recoveries of smolt PIT tags on bird colonies following the 2021 nesting season. Estimates were generated using previously published, standardized methods, providing a means to compare predation rates across avian predator species and colonies, salmonid species and ESUs/DPSs, and years. A total of 7,236 steelhead smolts were captured, PIT-tagged, and released into the tailrace of Rock Island Dam (RIS) in the middle Columbia River as part of this study in order to ensure that a sufficient number of ESA-listed Upper Columbia River (UCR) steelhead, a population that is highly susceptible to avian predation and therefore a suitable population to evaluate the efficacy of management actions, were tagged and available for predation analyses in 2021. Tagging at RIS commenced in 2008, resulting in a long-term dataset (2008–2021) with which to evaluate the impacts of piscivorous waterbirds on the survival of ESA-listed steelhead and to evaluate relative changes in predation and survival associated with management actions.

Efficacy of Avian Management Plans: Two primary objectives of the IAPMP were to reduce predation rates by Caspian terns on ESA-listed juvenile salmonids to less than 2% per salmonid ESU/DPS, per colony, and to less than 5% per salmonid ESU/DPS by terns from all colonies in the CPR combined. Recoveries of smolt PIT tags in 2021 indicated that predation rates exceeded this threshold for UCR steelhead, whereby an estimated 3.9% (95% credible interval = 2.4–6.8%) of available smolts were consumed by terns on Goose Island in Potholes Reservoir and an estimated 6.3% (4.5–9.3%) were consumed by terns from all colonies in the CPR combined. Predation rates by terns that nested on Crescent Island in 2021 were < 0.1% per ESU/DPS due to the small size (1 nesting pair) of that colony. Predation rates at the unmanaged tern colony on Lenore Lake (North Rock) were < 1.0% per ESU/DPS, with the highest rate observed on UCR steelhead at 0.9% (0.6–1.5%). Due to higher reservoir levels, no terns nested on the Blalock Islands in 2021, formerly the largest colony in the CPR and a location where predation rates often exceeded the 2% threshold for several ESUs/DPSs in previous years (2015–2020). Coincident with the elimination of the Blalock Islands colony, terns nested at an unmanaged colony site on Badger Island in 2021 and predation rate estimates of 1.4% (1.0–2.2%) and 1.5% (0.8–2.9%) were observed on UCR steelhead and Snake River (SR) steelhead, respectively. Estimates of predation rates on salmon populations by terns nesting on Badger Island, however, were < 1.0% per ESU/DPS, with the highest predation rate observed on SR spring/summer Chinook salmon at 0.8% (0.2–2.4%).

Coincident with increases in the number of terns nesting at colonies in the CPR in 2021, tern predation rates on juvenile salmonids were greater than those observed in 2020 for most of the salmonid ESUs/DPSs evaluated. Over the course of the entire study period, however, comparisons of tern predation rates on juvenile salmonids prior to (2007–2013) and during (2014–2021) implementation of the IAPMP indicate that there have been survival benefits to several salmonid ESUs/DPSs, especially UCR steelhead, from management under the IAPMP. For example, average annual predation rates by terns nesting on Goose Island and elsewhere in Potholes Reservoir have been reduced from an estimated 15.7% (14.1–18.9%) prior to management (2007–2013) to 1.7% (1.2–2.5%) during the management period (2014–2021). There was also evidence that survival of UCR steelhead smolts, on average, has increased significantly in the river reach where most terns from the Goose Island and Crescent Island colonies foraged (RIS to McNary Dam) following implementation of the IAPMP; increases in steelhead survival rates were commensurate with reductions in tern predation rates. In 2020, tern predation rates were at record low levels, coincident with record high survival rates for UCR steelhead smolts. In 2021, the increase in tern predation rates was commensurate with a decrease in steelhead survival (see also Additive Effects of Predation below). Due to increases in predation on juvenile salmonids by terns nesting on Goose Island and the re-establishment of tern colonies on Crescent Island and on Badger Island in 2021, adaptive management actions will likely be necessary to achieve the over-all goals of the IAPMP in the future.

Estimates of predation rates on multiple salmonid ESUs/DPSs by terns nesting on East Sand Island in the CRE during 2021 were not available, estimates that were generated with funding from the USACE in previous years. Smolt PIT tags, however, were recovered on the East Sand Island tern colony by the USACE following the 2021 breeding season and thus predation rates could be retrospectively analyzed in the future. In 2020, estimates of predation rates were the lowest recorded since terns began nesting on East Sand Island in 1999, with estimates ranging from 0.4% (0.2–0.9%) on UCR spring Chinook salmon to 5.9% (4.5–8.1%) on SR steelhead. Results from 2020 indicated that predation by terns nesting on East Sand Island had been reduced by 65% to 76% for steelhead DPSs, reductions that met or exceeded those anticipated in the Caspian Tern Management Plan for the Columbia River Estuary. Although predation rates by terns that nested on East Sand Island in 2021 are currently unknown, impacts may be greater than those in 2020 based on the number of smolt PIT tags recovered in 2021 (n = 5,081) compared with 2020 (n = 4,595). Large numbers of terns also continued to nest outside of the designated 1-acre nesting areas on East Sand Island in 2021, as well as nesting attempts by terns on Rice Island in the upper Columbia River estuary. Thus, continued implementation of nest dissuasion activities, continued monitoring of tern nesting sites throughout the estuary, and empirically derived estimates of ESU/DPS-specific predation rates will be necessary to ensure that predation impacts in the future do not exceed the levels stipulated in the management plan.

For the second consecutive year there was no evidence of an established (long-lived) nesting colony of cormorants on East Sand Island in the CRE during 2021. Several hundred pairs of cormorants did, however, attempt to nest on East Sand Island in 2021; however, no estimates of predation rates on juvenile salmonids by cormorants attempting to nest on East Sand Island are available, estimates that have been generated with funding from the USACE in previous years. Unlike the tern colony on East Sand Island, predation rates on salmonid smolts by cormorants that attempted but failed to nest on East Sand Island in 2021 were presumably low based on the small number of smolt PIT tags (n = 117) recovered following the breeding season.

Astoria-Megler Bridge Cormorant Predation Assessment: An unintended consequence of management actions at the double-crested cormorant colony on East Sand Island during 2015–2019 was the complete abandonment of the colony site and the subsequent rapid expansion of the cormorant colony on the Astoria-Megler Bridge (AMB). The AMB is located upstream of East Sand Island in the freshwater mixing zone of the CRE, an aquatic environment where fewer alternative prey fish are likely available and juvenile salmonids may be consumed in greater proportion by cormorants relative to foraging areas around the former cormorant colony site on East Sand Island which is in the marine zone of the estuary. To estimate predation rates on salmonid smolts by cormorants breeding on the AMB in 2021, we recovered smolt PIT tags deposited by cormorants nesting in five discrete plots on a concrete footing of the bridge, areas where smolt PIT tags could be recovered by researchers following the breeding season. Estimates of per capita (per breeding pair) predation rates derived from cormorants nesting in plots (n = 159 breeding pairs) were then extrapolated to account for all double-crested cormorants nesting elsewhere on the bridge (n = 3,992 breeding pairs) to generate colony-wide estimates of predation rates. Results indicated that per capita predation rates of cormorants nesting on the AMB ranged from 0.0009% (95% credible interval = 0.0005–0.0016%) on yearling Chinook salmon smolts to 0.0040% (0.0028–0.0069%) on sub-yearling Chinook salmon smolts originating from populations upstream of Bonneville Dam. Per capita predation rates were 0.0013% (0.0004–0.0025%) on steelhead smolts and 0.0015% (0.0004–0.0031%) on coho salmon smolts originating from populations downstream of Bonneville Dam. Colony-wide estimates of predation rates ranged from 3.9% (2.0–6.8%) on yearling Chinook salmon to 16.8% (7.3–28.4%) on sub-yearling Chinook salmon and from 5.5% (1.8–10.5%) on steelhead to 6.1% (1.6–12.7%) on coho salmon. Estimates of per capita predation rates on salmonid smolts by cormorants nesting on the AMB in 2021 were approximately 2 to 4 times greater (depending on salmonid species/age-class) than average annual per capita predation rates by cormorants that nested on East Sand Island prior to implementation of management actions (2003–2014). Estimates of colony-wide predation probabilities by cormorants nesting on the AMB in 2021 were similar to those of cormorants breeding on East Sand Island during 2003–2014, despite the much smaller size of the cormorant colony on the AMB in 2021 (4,151 pairs) compared with the cormorant colony on East Sand Island (12,787 pairs on average during 2003–2014). Collectively, results suggest that colony location is closely related to the impact of cormorant predation on survival of salmonid smolts in the CRE, and that these impacts have likely remained unchanged since management actions on East Sand Island were first implemented in 2015.

Other Piscivorous Colonial Waterbirds: An investigation of predation by other, unmanaged piscivorous colonial waterbird species indicated that smolt consumption rates by California and ring-billed gulls nesting at certain colonies in the CPR were consistently higher than those of nearby managed and unmanaged Caspian tern colonies in 2021. Unlike terns, gulls are scavengers and are known to consume dead or moribund fish and to steal (kleptoparasitize) fish from other waterbirds, like terns. Consequently, we use the term “gull consumption rates” rather than “gull predation rates”. Estimates of consumption rates as high as 12.4% (6.2–23.1%) and 13.9% (9.1–21.8%) on UCR and SR steelhead, respectively, were documented for gulls nesting at Miller Rocks in The Dalles Reservoir. Consumption rates as high as 2.8% (1.1–6.3%) on SR Fall Chinook salmon by gulls nesting at Miller Rocks were among the highest estimates of consumption rate observed for a salmon ESU in 2021. Estimates of consumption rates by gulls nesting at several other colonies in the CPR were in excess of 2% and as high as 8.6% (3.6–14.7%) for steelhead DPSs, including gulls nesting on Island 20 in the middle Columbia River, Badger and Crescent islands in McNary Reservoir, and the Blalock Islands in John Day Reservoir. Estimates of consumption rates by gulls nesting at colonies in the CPR during 2021 were similar to or greater than those observed in previous years (2007–2020). Cormorant predation rates on salmonid smolts originating from Upper Columbia River ESUs/DPSs were low (< 0.2% per ESU/DPS) but were significantly higher on smolts from SR ESUs/DPSs, with predation on SR steelhead and SR Spring/Summer Chinook salmon at 3.0% (0.8–9.3%) and 4.4% (0.8–15.6%), respectively.

Despite the high levels of consumption/predation on smolts observed at some piscivorous waterbird colonies in 2021, impacts from birds nesting at other colonies, particularly impacts on salmon ESUs, were often low. For example, estimates of predation rates by cormorants nesting on Hanford Island in the middle Columbia River were < 0.4% per salmonid ESU/DPS, estimates of smolt predation rates for pelicans nesting at Badger Island in McNary Reservoir were < 1.0% per ESUs/DPSs, and estimates of predation rates by pelicans nesting on Miller Sands Spit in the Columbia River estuary were < 0.1% per ESU/DPS, among the lowest of any colony evaluated in 2021. Significantly higher predation impacts on non-ESA-listed juvenile salmonid stocks (e.g., subyearling Fall Chinook salmon from the Upper River Bright stock) and, possibly, adult salmonids (e.g., adult sockeye salmon) by pelicans have been documented in other studies and warrant additional investigation. Nevertheless, results of this study indicate that pelicans nesting on Badger Island in the CPR and, especially, at the Miller Sands Spit colony in the CRE, posed only a minor predation threat to survival of actively migrating, ESA-listed UCR and SR juvenile salmonids in 2021.

Cumulative Predation and Smolt Survival: To investigate the cumulative effects of avian predation (predation/consumption by all colonial waterbird species nesting at all colonies combined) and to determine what proportion of all sources of smolt mortality (1-survival) were associated with avian predation/consumption, we conducted a mark-recapture-recovery analysis on UCR steelhead smolts that were PIT-tagged and released at RIS in 2021. We used previously published methods to jointly estimate predation/consumption and survival probabilities during smolt passage through multiple river reaches and we compared results from 2021 to those from previous years (2008–2020). In 2021, the cumulative effects of avian predation/consumption on mortality of UCR steelhead during smolt passage from RIS to the Pacific Ocean were estimated at 42.4% (35.4–50.6%) of available smolts, with the highest levels of mortality associated with predation/consumption by gulls at 29.1% (22.8–36.3%), followed by predation by terns at 8.6% (6.7–11.8%), predation by cormorants at 2.9% (0.1–7.0%), and predation by pelicans at 0.4% (0.1–2.4%). Results indicate that despite reductions in average annual mortality rates of UCR steelhead smolts due to tern predation in the CPR and the CRE, the cumulative effects of avian predation/consumption remain a substantial source of steelhead smolt mortality during out-migration to the Pacific Ocean. Estimates of the cumulative effects of predation/consumption on mortality of UCR steelhead in 2021 were the third highest recorded since studies of system-wide predation begin in 2008, with estimates in 2021 only slightly lower than those observed in 2009 and 2015. Increases in tern predation probabilities in the CPR, gull consumption probabilities in the CPR, and the inclusion of predation probabilities from cormorants nesting on the AMB all contributed to higher estimates of cumulative predation/consumption of UCR steelhead smolts in 2021 relative to most, but not all, previous years.

Comparisons of total mortality (1-survival) to mortality associated with avian predation/consumption, indicated that avian predation/consumption in 2021 was associated with more mortality of UCR steelhead during smolt out-migration from RIS to Bonneville Dam than any other direct mortality factor, with avian predation/consumption associated with 61.6% (48.8–76.8%) of all smolt mortality in this river reach. Results from 2021 were consistent with those from most previous study years, with avian predation/consumption upstream of Bonneville Dam associated with more than 50% of all steelhead mortality during 11 of the previous 13 years (2008–2020). Even after passage through the impounded sections of the middle and lower Columbia River upstream of Bonneville Dam, the impact of predation by piscivorous colonial waterbirds on the survival of steelhead smolts in the CRE were substantial, with terns and cormorants breeding on East Sand Island and cormorants breeding on the AMB collectively depredating an estimated 13.5% (6.6–25.1%) of available UCR steelhead smolts in 2021.

Additive Effects of Predation: To investigate to what degree avian predation/consumption of UCR steelhead smolts limited fish survival, we used a previously published model to explicitly measure the strength, magnitude, and direction of the relationship between avian predation/consumption and fish survival. Data from 2021 were used to update a long-term dataset (2008–2020) to evaluate more recent trends in avian predation and smolt survival. Results indicated that the higher levels of tern predation on UCR steelhead in 2021 were associated with lower levels of smolt survival from RIS to Bonneville Dam compared with 2020, a year when record low levels of tern predation were associated with record high levels of UCR steelhead smolt survival. An investigation of weekly and annual estimates of tern predation and smolt survival probabilities suggested that a greater proportion of UCR steelhead smolts would have survived out-migration to Bonneville Dam in the absence of tern predation upstream of Bonneville Dam, with the estimated average annual difference in observed survival versus baseline survival (i.e. survival in the absence of tern predation) of 0.174 (0.106–0.234) during 2008–2021. Due to low levels of cormorant predation on UCR steelhead smolts upstream of Bonneville Dam (0.01 or 1%), only small increases in survival of UCR steelhead smolts to Bonneville Dam would be possible in the absence of cormorant predation upstream of Bonneville Dam. Although there was some evidence of a relationship between consumption probabilities by gulls and survival probabilities of UCR steelhead smolts, results were not statistically significant when considered across all years. The statistical power to accurately determine to what degree smolt consumption by gulls influenced smolt survival was limited by a truncated time series, coupled with high levels of uncertainty in both consumption and survival and a lack of weekly variation in estimates of consumption probabilities. Gulls are also known to consume dead fish and to kleptoparasitize fish from other piscivorous waterbirds, like terns, so it is likely that consumption of UCR steelhead smolts by gulls was a more compensatory source of mortality compared to predation by terns or cormorants.

There was evidence that higher levels of tern predation on UCR steelhead smolts in the CRE were associated with lower returns of adult steelhead to Bonneville Dam, with increases in tern predation probabilities associated with statistically significant decreases in adult survival probabilities. Results suggest that in the absence of tern predation on UCR steelhead smolts, smolt-to-adult returns (SARs) for UCR steelhead would have nearly doubled, even though, in the absence of tern predation, the majority of smolts depredated by terns would have died from other causes before returning to Bonneville Dam as an adult (i.e., partial additivity). These results, along with those from a recently published, peer-reviewed study (Payton et al. 2020) indicate that tern predation on steelhead smolts is an additive source of mortality and a partially additive source of mortality to the adult life-stage. There was some evidence that higher levels of cormorant predation on UCR steelhead smolts in the estuary were associated with lower adult returns to Bonneville Dam; however, results were not statistically significant at an alpha level of 0.05 when considered across those years when adequate data were available (2008–2015). Collectively, our results indicate that efforts to reduce tern predation on UCR steelhead have enhanced steelhead smolt survival in the CRB, especially in those years when tern predation rates were dramatically reduced because of management actions at Goose and Crescent islands, as in 2020. More importantly from a steelhead conservation perspective, our results suggest that in the absence of tern predation on UCR steelhead smolts, significantly more adult steelhead would return to Bonneville Dam. Managers, regional stakeholders, and the public at-large must decide whether these potential increases in smolt survival and SARs for UCR steelhead warrant continued efforts to manage terns in the Columbia River basin. Additional research is needed, however, to evaluate to what degree cormorant predation and, especially, gull consumption of salmonids limits smolt survival and SARs in the CRB.

Biotic and Abiotic Factors: Previous research indicates that biotic and abiotic factors experienced by smolts during out-migration influence their survival, including, but not limited to, the probability of avian predation. As part of this study, and as recently recommended by the Northwest Power and Conservation Council’s Independent Scientific Advisory Board, we are investigating the influence of various covariates on smolt survival during out-migration. Covariates under consideration include biotic factors such as fish size, rearing-type (hatchery, wild), abundance (density), and run-timing, plus abiotic factors such as spill, discharge, and measures of river flow speed (e.g., water transit time). The goal of this analysis is to describe those factors that best explain variation in smolt survival and to identify potential “management relevant” variables, variables that resource managers may be able to control to some degree. Analyses of covariates are on-going and herein we provide an update of results from analyses of acoustic-tagged UCR steelhead smolts and a discussion of next steps. A more complete analysis will be presented as part of our 2022 Annual Report.

Smolt Survival to Bonneville Dam
Recoveries of smolt PIT tags on breeding colonies of piscivorous colonial waterbirds can be used to increase the precision and accuracy of smolt survival estimates by increasing the sample sizes of tagged fish used in mark-recapture survival models. To provide information for use in smolt survival estimates to Bonneville Dam in 2021, we recovered smolt PIT tags from several piscivorous waterbird nesting, loafing, and roosting sites in the CRE, sites that were not included in the original scope of work for this study, but where we suspected PIT tags were being deposited by birds. In total, scanning associated with this additional effort detected 1,932 current migration year (2021) PIT tags from juvenile salmonids in the CRE. This information, coupled with tags recovered by the USACE on the East Sand Island tern colony (n = 5,081) and tags of live fish detected at the National Marine Fisheries Service pair trawl net detection system (n = 4,537), were used by this project and other federally funded projects to estimate smolt survival to Bonneville Dam in 2021.

-Bird Research Northwest

Click here to see full report

Back to Feature Stories »

/columbia-bird-research-bg-2.jpg