Note: Provided here is the Executive Summary from our Draft 2007 Season Summary (click link below for complete draft report).  The Draft 2007 Season Summary has been prepared for the Bonneville Power Administration and the U.S. Army Corps of Engineers for the purpose of assessing project accomplishments.  This report is not for citation in the published literature without the permission of the authors. Please contact clientservices@realtimeresearch.org for this authorization.

Date Posted : 3/17/2008

[Click here for the Draft 2007 Season Summary (pdf)]

This study investigates predation by piscivorous colonial waterbirds on juvenile salmonids (Oncorhynchus spp.) from throughout the Columbia River Basin. The study objectives for the Columbia River estuary in 2007, work funded by the Bonneville Power Administration, were to (1) monitor and evaluate previous management initiatives to reduce Caspian tern (Hydroprogne caspia) predation on juvenile salmonids (smolts); (2) measure the impact of double-crested cormorant (Phalacrocorax auritus) predation on smolt survival; (3) assess potential management options to reduce cormorant predation; and (4) monitor large colonies of other piscivorous waterbirds in the estuary (i.e., glaucous-winged/western gulls [Larus glaucescens/occidentalis]) to determine potential impacts on smolt survival. The study objectives for the middle Columbia River in 2007, work funded by the Walla Walla District of the U.S. Army Corps of Engineers, were to (1) measure the impact of predation by Caspian terns and double-crested cormorants on smolt survival in the mid-Columbia River; and (2) monitor large nesting colonies of other piscivorous waterbirds (i.e., California gulls (L. californicus], ring-billed gulls [L. delawarensis], and American white pelicans [Pelecanus erythrorhynchos]) on the mid-Columbia River to determine the potential for significant impacts on smolt survival.

Our previous studies to evaluate system-wide losses of juvenile salmonids to avian predation indicated that Caspian terns and double-crested cormorants nesting in the Columbia River estuary were responsible for the vast majority of smolt losses to avian predators in the Columbia Basin.  Again in 2007, East Sand Island in the Columbia River estuary supported the largest known breeding colonies of Caspian terns and double-crested cormorants in the world. The Caspian tern colony on East Sand Island consisted of ca. 9,900 breeding pairs in 2007, not significantly different than in 2006 (ca. 9,200 pairs). The size of the Caspian tern colony at East Sand Island has remained nearly stable since 2000. Tern nesting success averaged 0.64 fledglings per breeding pair in 2007, similar to 2006 (0.72 fledglings per breeding pair). Nesting success during 2005-2007 has been lower than during 2001-2004, when nesting success averaged 1.12 fledglings per breeding pair.

The proportion of juvenile salmonids in the diet of East Sand Island Caspian terns during the 2007 nesting season averaged 30% of prey items, similar to 2006 (31% of prey items), but higher than in 2004 (17% of prey items) or 2005 (23% of prey items). Consumption of juvenile salmonids by terns nesting at the East Sand Island colony in 2007 was approximately 5.5 million smolts (95% c.i. = 4.8 – 6.2 million), similar to smolt consumption the previous year (best estimate = 5.4 million smolts; 95% c.i. = 4.6 – 6.1 million).  This is less than half the annual consumption of juvenile salmonids by Caspian terns in the estuary prior to 2000, when their breeding colony was located on Rice Island in the upper estuary.  Caspian terns nesting on East Sand Island continued to rely primarily on marine forage fishes (i.e., northern anchovy, shiner perch, Pacific herring) as a food supply. Based on smolt PIT tag recoveries on the East Sand Island Caspian tern colony, predation rates on steelhead smolts were particularly high during 2007, at about 12.5% for in-river migrant smolts and 7.7% for transported smolts. Predation rates on steelhead were 2-12 times higher than those for other salmonid species and run types.

In 2008, the U.S. Army Corps of Engineers will begin implementing the plan “Caspian Tern Management to Reduce Predation of Juvenile Salmonids in the Columbia River Estuary,” outlined in the Final Environmental Impact Statement (FEIS) and Records of Decision (RODs) signed in November 2006.  This management plan seeks to redistribute a portion of the East Sand Island tern colony to alternative colony sites in Oregon and California by 2015.  The plan calls for the creation of up to 7 acres of new or enhanced tern nesting habitat in interior Oregon (i.e., Fern Ridge Lake, Crump Lake, and Summer Lake) and coastal California (i.e., the San Francisco Bay Area) and to actively attract terns to nest there.  As alternative tern nesting habitat is created or enhanced, the available tern nesting habitat on East Sand Island will be reduced from its current size (6 acres) to 1.5 – 2 acres.  Habitat enhancement at alternative sites will be accomplished in stages and the reduction of tern nesting habitat at East Sand Island will occur at a ratio of one acre reduced for each 2 acres of habitat created elsewhere. Once fully implemented, the plan would reduce the East Sand Island Caspian tern colony from its current size (approximately 9,500 nesting pairs) to approximately 3,100 – 4,400 nesting pairs.  This reduction in the size of the East Sand Island tern colony is intended to reduce tern predation on smolts in the Columbia River estuary by an estimated 2.4 – 3.1 million smolts annually.

The double-crested cormorant colony on East Sand Island consisted of about 13,770 breeding pairs in 2007, similar to the estimate of colony size last year (13,740 pairs). Since our monitoring began in 1997, this cormorant colony has increased by about 275%. Nesting success in 2007 (2.78 fledglings per breeding pair) was the highest ever recorded for this colony, and up considerably from 2006 (1.92 fledglings per breeding pair). As in previous years, salmonids made up a small portion (9%) of the cormorant diet in 2007, while marine forage fish (i.e., northern anchovy) and estuarine resident fish (i.e., sculpin, flounder) made up over 50% of the diet. Despite the lower reliance on salmonids as a food source by cormorants compared to terns, total smolt consumption by cormorants was similar to or greater than that by terns. This is because double-crested cormorants are about four times larger than Caspian terns and the cormorant colony consists of about 40% more nesting pairs than the tern colony. In 2006, cormorants nesting on East Sand Island consumed an estimated 10.3 million juvenile salmonids (95% c.i. = 4.7 – 15.9 million), compared to an estimated 5.4 million juvenile salmonids (95% c.i. = 4.6 – 6.1 million) consumed by terns nesting on East Sand Island (estimates of cormorant consumption of salmonid smolts in 2007 are pending further analyses).

An analysis of salmonid PIT tags detected at the double-crested cormorant colony on East Sand Island indicated that all species of anadromous salmonids (i.e., Chinook salmon, coho salmon, sockeye salmon, steelhead, and sea-run cutthroat trout) from all run-types (fall, winter, summer, and spring), and from all tagged ESUs were susceptible to cormorant predation in 2007.  The numbers of PIT tags from the various salmonid species and run-types that were recovered on the cormorant colony were roughly proportional to the relative availability of PIT-tagged salmonids released in the basin, suggesting that cormorant predation on salmonid smolts in the estuary was less selective than tern predation. In contrast, PIT tag recoveries on the East Sand Island tern colony indicated that steelhead were far more vulnerable to Caspian tern predation as compared to other salmonid species.  An analysis of salmonid predation rates, based on the proportion of available PIT-tagged fish subsequently deposited on the cormorant colony, indicated that both hatchery- and wild-origin smolts were consumed, with rates averaging between 2 and 5% for most species and run-types of PIT-tagged fish originating upstream of Bonneville Dam.  Predation rates in excess of 20% were observed for some groups of hatchery-origin fall Chinook salmon released in or near the estuary.

If the cormorant breeding colony on East Sand Island continues to expand and/or the proportion of salmonids in cormorant diets increases, cormorant predation rates on juvenile salmonids may far exceed those of Caspian terns nesting in the estuary. The discrepancy in predation rates for the two colonies will be even greater if the Caspian tern colony is reduced in size by >50% by 2015, as intended under the management plan now being implemented. Resource management agencies have not decided whether management of the large and expanding colony of double-crested cormorants on East Sand Island is warranted. Elsewhere in North America, management of double-crested cormorants has consisted primarily of lethal control (i.e., shooting of adults, oiling of eggs, and destruction of nests in trees). Non-lethal management approaches, such as relocating a portion of the colony to alternative colony sites along the coast of Oregon and Washington, seem more appropriate in the context of the cormorant colony on East Sand Island, which constitutes nearly 50% of the entire breeding population of the Pacific Coast subspecies P. auritus albociliatus. Studies designed to test the feasibility of employing habitat enhancement and social attraction (i.e., old tires with nest material, decoys, audio playback systems) to relocate nesting cormorants have shown some promise; cormorants were previously attracted to nest and nested successfully (raised young to fledging) on Miller Sands Spit and Rice Island, two islands in the upper estuary where no successful cormorant nesting attempts have been recorded recently. In 2007, habitat enhancement and social attraction were retained at Miller Sands Spit, but removed from Rice Island; the cormorant colony on Miller Sands Spit was again successful in raising young, while there was no cormorant nesting on Rice Island.

In order to reduce cormorant predation on juvenile salmonids from the Columbia Basin, however, it will be necessary to relocate nesting cormorants to suitable habitat outside the Columbia River estuary. In 2007, we conducted a pilot study to test the feasibility of attracting double-crested cormorants to nest at a site remote from the Columbia River estuary and where cormorants had not previously been known to nest. We placed old tires with nest material, cormorant decoys, and audio playbacks of cormorant colony sounds on a floating platform in the Fern Ridge Wildlife Area, adjacent to Fern Ridge Lake near Eugene, Oregon. While double-crested cormorants were repeatedly seen in the area, no cormorants were seen on the platform and no nesting attempts occurred there. This pilot study will be repeated in 2008. While studies of the use of habitat enhancement and social attraction in the Columbia River estuary have been promising, results to date indicate that double-crested cormorants are not as responsive to these techniques as Caspian terns.

As was the case with Caspian tern management in the Columbia River estuary, any management of double-crested cormorants to reduce smolt losses will likely require additional research and analysis under NEPA, including assessments of (1) the population status of the Pacific Coast subspecies of double-crested cormorant, (2) the availability of suitable nesting habitat for the subspecies outside the Columbia River estuary, and (3) the potential enhancement of salmonid recovery rates in the Columbia River Basin due to management of cormorants in the estuary. These and other related studies are planned for 2008 and beyond.

The Caspian tern colony on Crescent Island in the mid-Columbia River has received comparatively little attention from salmon management agencies because of its relatively small size (ca. 500 nesting pairs, ca. 1/20th the size of the Caspian tern colony in the estuary) and low annual consumption of juvenile salmonids (ca. 0.5 million smolts, ca. 1/10th the consumption of the tern colony in the estuary). In 2007, there were two breeding colonies of Caspian terns on the mid-Columbia River; about 355 pairs nested on Crescent Island in the McNary Pool and about 40 pairs nested at a relatively new colony site on Rock Island in the John Day Pool. The Crescent Island tern colony declined by 21% from 2006, when 448 breeding pairs nested at the colony; this colony is now smaller than in any year since 1997. It is still the largest Caspian tern colony on the Columbia Plateau, however, and the third largest Caspian tern colony in the Pacific Northwest. The Rock Island Caspian tern colony in 2007 was substantially smaller than in 2006, when 110 breeding pairs attempted to nest there, but was larger than in 2005 (6 breeding pairs), the first year that Caspian terns were known to nest on Rock Island. Nesting success at the Crescent Island tern colony was 0.68 young fledged per breeding pair, up 58% from 2006 (0.43 young fledged per breeding pair).  Tern productivity at the Crescent Island colony in 2006 was the lowest recorded at this colony since monitoring began in 2000. In 2007, the Rock Island Caspian tern colony failed to produce any young, apparently due to avian predation on all tern eggs and chicks. In 2006, the Rock Island Caspian tern colony also failed, apparently due to mink predation.

At Crescent Island in 2007, salmonid smolts represented 69% of prey items in tern diets, up slightly from 2006 (63%). We estimated that Caspian terns nesting on Crescent Island in 2007 consumed 360,000 juvenile salmonids (95% c.i. = 250,000 – 460,000), a ca. 10% decline in smolt consumption compared to 2006 (best estimate = 402,000, 95% c.i. = 310,000–500,000). Steelhead comprised an estimated 20.5% of the identifiable salmonid smolts, or roughly 74,000 fish, an increase over the previous year (56,000 fish).  Per capita smolt consumption by Crescent Island terns in 2007 (507 smolts per nesting tern across the breeding season) was also greater compared to 2006 (446 smolts per nesting tern). Although no data on diet composition were collected at the Rock Island tern colony, we estimate that 677 smolt PIT tags were deposited on the colony during the 2007 nesting season, indicating that salmonids were a significant part of the diet before the colony failed. A comparison of smolt PIT tags recovered from the Crescent Island and Rock Island tern colonies suggests that Rock Island terns consumed about 1/8th as many PIT-tagged salmonid smolts as did Crescent Island terns, or roughly 45,000 smolts.

Based on smolt PIT tag recoveries on the Caspian tern colony at Crescent Island, the predation rate on in-river migrants from the Snake River (all species and run types) was about 1.1% in 2007, down substantially from 7.5% and 3.8% in 2005 and 2006, respectively.  These predation rates have been corrected for both the detection efficiency of PIT tags on-colony and the proportion of PIT tags ingested by terns that were subsequently deposited on-colony. Although predation rates were dramatically down in 2007, the numbers of Snake River smolts available to terns foraging in McNary Pool were substantially up, as fewer fish were collected for transportation at Snake River dams.  As in previous years, predation rates on PIT-tagged steelhead smolts were greater than for other salmonid species.  In 2007, ca. 4.9% of the hatchery-origin and 4.8% the wild-origin steelhead smolts from the Snake River were consumed by Crescent Island terns (these predation rates are based on the proportion of PIT-tagged fish interrogated passing Lower Monumental Dam between 1 April and 31 July that were subsequently detected on the Crescent Island tern colony).  Because fewer Snake River steelhead were transported around McNary Pool in 2007 compared to 2006, a larger proportion of the Snake River steelhead population was susceptible to predation from Crescent Island terns in 2007. Consequently, the total predation rate by Crescent Island terns on the Snake River steelhead ESU in 2007 was the highest observed since 2004.  Predation rates on wild-origin steelhead vs. hatchery-origin steelhead from the Snake River were similar and not statistically different when pooled over the entire 2007 out-migration; this finding differs from results during 2004 – 2006, when predation rates on hatchery-origin smolts were consistently higher than on wild-origin smolts.

In 2007, the double-crested cormorant colony on Foundation Island in the mid-Columbia River consisted of at least 330 nesting pairs, and was somewhat smaller than in 2006. The largest cormorant colony on the Columbia Plateau in 2007 was again on Potholes Reservoir, where about 1,015 pairs nested in trees at the north end of the reservoir. The size of this colony was also somewhat lower than in 2006. The limited diet data for Foundation Island cormorants suggest that juvenile salmonids represented 16-18% of the diet.  For the first time since this research was initiated in 2004, smolt PIT tag recoveries, and in some cases reach and stock-specific salmonid predation rates, were higher for the Foundation Island cormorant colony than for the Crescent Island tern colony.  In fact, of all the piscivorous waterbird colonies studied on the Columbia River in 2007, the Foundation Island cormorant colony had the highest per capita consumption rate of PIT-tagged juvenile salmonids (ca. 11.3 PIT-tagged fish per breeding adult), followed by the Rock Island tern colony (7.87) and the Crescent Island tern colony (7.24).  These results suggest that predation rates on salmonid smolts by Foundation Island cormorants are increasing and may now be similar to or greater than that of Caspian terns nesting on nearby Crescent Island.  Similar to predation by Crescent Island terns, steelhead were particular vulnerability to predation by Foundation Island cormorants in 2007.  Unlike terns, however, Foundation Island cormorants also keyed in on groups of Chinook salmon (both yearlings and sub-yearlings) migrating through McNary Pool.  In contrast to the Foundation Island cormorant colony, there is little evidence to suggest that cormorants nesting at the larger colony on Potholes Reservoir are affecting the survival of juvenile salmonids from the Columbia and Snake rivers during the nesting season, based on the paucity of PIT tags from Columbia Basin salmonid smolts recovered at the colony in 2007 (n = 6 smolt PIT tags).

Unlike Caspian terns, which depart the Columbia Basin during the non-breeding season, some double-crested cormorants over-winter on the Columbia and Snake rivers. Over-wintering cormorants could potentially affect the survival of hold-over fall Chinook salmon smolts, particularly near Snake River dams. A pilot study to investigate this potential impact suggested that small numbers of cormorants (< 100) over-winter near two lower Snake River dams (Little Goose and Lower Granite) and that salmonids make up a significant, although not predominant, proportion of their diet.  Based on identifiable fish tissue in fore-gut samples, juvenile salmonids comprised 11.8% by mass of the diet of double-crested cormorants foraging at Little Goose and Lower Granite dams in 2007 (n = 40 fore-gut samples).  Juvenile shad were the most abundant fish found in fore-gut contents, representing 47.7% of prey biomass, followed by centrarchids (22.0%).  It should be noted, however, that these diet composition results are based on a small sample size and the counts of cormorants at two dams on the Snake River tell us little about the system-wide abundance and distribution of over-wintering cormorants on the Snake River and their potential impact on survival of juvenile salmonids.  In 2008, we plan to conduct more comprehensive surveys of the distribution and abundance of over-wintering cormorants along the Snake River from the confluence with the Columbia River to Lewiston, Idaho.  Additionally, we will increase our sampling efforts to measure diet composition in order to better assess the impacts on ESA-listed salmonid stocks, particularly hold-over fall Chinook salmon smolts, of double-crested cormorants over-wintering along the lower Snake River.
 
Compared to Caspian terns and double-crested cormorants, other piscivorous colonial waterbirds that nest along the mid-Columbia River (i.e., California gulls, ring-billed gulls, American white pelicans) are having less impact on the survival of juvenile salmonids from the Columbia and Snake rivers.  One gull colony that may be having a significant impact on salmonid smolt survival, however, is the large California gull colony (ca. 3,500 nesting pairs) on Miller Rocks in The Dalles Pool, where 2,653 smolt PIT tags were recovered in 2007.  Previous research in 1997 and 1998 indicated that salmonid smolts, and fish in general, constituted a very small proportion of the diet of California and ring-billed gulls nesting at up-river colonies (Collis et al. 2002a). At the American white pelican colony on Badger Island 1,160 smolt PIT tags were recovered in 2007; this represents about 0.64 PIT-tagged smolts consumed per nesting adult at this growing colony.  In comparison, double-crested cormorants nesting at Foundation Island and Caspian terns nesting on Crescent Island consumed 11.3 and 7.2 PIT-tagged smolts per nesting adult, respectively.  The size of some up-river gull colonies (≥ 7,000 breeding pairs on several islands) and the Badger Island white pelican colony (> 900 pairs), however, exceeds that of the up-river tern and cormorant colonies and should be taken into account when evaluating overall impacts of avian predation on salmonid smolt survival on the Columbia Plateau.  Further research and monitoring is necessary to determine whether particular gull and pelican colonies might be having a significant effect on survival of juvenile salmonids in the lower and mid-Columbia River.

In contrast to the gull and pelican colonies on the Columbia Plateau, previous research on glaucous-winged/western gulls nesting in the Columbia River estuary indicated that these birds consumed significant numbers of juvenile salmonids (Collis et al. 2002a).  Gulls nesting on Rice Island (river km 34) ate mostly riverine fishes, including out-migrating salmonids, whereas gulls nesting on East Sand Island (river km 8) ate primarily marine fishes.  In 1997 and 1998, juvenile salmonids comprised 10.9% and 4.2% of the diet (by mass) of glaucous-winged/western gulls nesting on Rice Island/Miller Sands Spit and East Sand Island, respectively. PIT tag studies have not been conducted on these colonies, nor have diet data been collected since 1998. As such, the current impact on salmonid smolt survival of predation from gulls nesting at these estuary colonies is unknown.

In 2007 we conducted a pilot study to investigate how smolt morphology, condition, and origin might be related to differences in smolt vulnerability to avian predation. We hypothesized that the probability of smolt mortality due to avian predation increases with the declining physical condition of the fish.  We also hypothesized that river conditions and dam operational strategies may be associated with a smolt’s vulnerability to avian predators. As part of this pilot study, we scored the condition of 7,088 steelhead smolts that were PIT-tagged and released at Lower Monumental and Ice Harbor dams. Subsequent recovery of some of these PIT tags on piscivorous waterbird colonies downstream indicated that avian predation is partially condition-dependent, with diseased steelhead or steelhead with severe external damage more likely to be consumed by birds than fish with little or no external damage or disease.  For example, steelhead with severe external damage were 1.8 times more likely to be consumed by an avian predator than fish with no signs of external damage.  Similarly, there was a positive relationship between the extent of de-scaling of smolts and their vulnerability to avian predation, slight to severely de-scaled fish were 1.2 to 2.4 times more likely to fall prey to birds than smolts with little to no de-scaling.  Preliminary results indicate that at least some smolt mortality is compensatory, and that not all mortality due to avian predation is additive. 

A system-wide assessment of avian predation on juvenile salmonids based on recent available data indicates that the most significant impacts to smolt survival occur in the estuary, with Caspian terns and double-crested cormorants nesting on East Sand Island combined to consume ca. 7 – 16 million smolts annually during 2003-2006.  Although estimates of smolt consumption for East Sand Island cormorants in 2007 are not yet available, combined smolt losses to terns and cormorants nesting on East Sand Island in 2007 are expected to be within this range.  Estimated smolt losses to piscivorous birds that nest in the estuary are more than an order of magnitude greater than those observed on the mid-Columbia River. Additionally, when compared to the impact of avian predation on the mid-Columbia River, avian predation in the Columbia estuary affects juvenile salmonids that have survived freshwater migration to the ocean and presumably have a higher probability of survival to return as adults, compared to those fish that have yet to complete out-migration.  Finally, juvenile salmonids belonging to every ESA-listed stock from the Columbia River basin are susceptible to predation in the estuary because all surviving fish must migrate in-river through the estuary to reach the ocean.  For these reasons, management of Caspian tern and double-crested cormorant colonies on East Sand Island has the greatest potential to benefit ESA-listed salmonid populations from throughout the Columbia River basin, when compared to potential management of other colonies of other piscivorous waterbirds.  The Caspian tern colony on Crescent Island and the double-crested cormorant colony on Foundation Island may be exceptions to this rule; management of these small, up-river colonies may benefit certain salmonid stocks, particularly steelhead.