Content uploaded by Geoffrey A. McMichael
Author content
All content in this area was uploaded by Geoffrey A. McMichael on Feb 12, 2017
Content may be subject to copyright.
PNNL-SA-23719
Survival of Wild Hanford Reach
and Priest Rapids Hatchery Fall
Chinook Salmon Juveniles in
the Columbia River: Predation
Implications
October 2014
RA Harnish H Li
ED Green B Rayamajhi
KA Deters KW Jung
KD Ham GA McMichael
Z Deng
PNNL-SA-23719
Survival of Wild Hanford Reach and
Priest Rapids Hatchery Fall Chinook
Salmon Juveniles in the Columbia
River: Predation Implications
RA Harnish H Li
ED Green B Rayamajhi
KA Deters KW Jung
KD Ham GA McMichael1
Z Deng
October 2014
Prepared for
the Pacific Salmon Commission
under DOE Contract DE-AC05-76RL01830
Pacific Northwest National Laboratory
Richland, Washington 99352
1Mainstem Fish Research, 65 Park St., Richland WA 99354
iii
Abstract
The population of fall Chinook salmon that inhabits the Hanford Reach comprises the majority of the
Columbia Upriver Bright (URB) stock and is one of the most productive Chinook salmon stocks in the
Pacific Northwest. Recent studies indicated that much of the high productivity of the population may be
attributed to very high survival during early freshwater life stages within the Hanford Reach. However,
some evidence suggests significant mortality of smolts occurs over a short period of time and distance as
they migrate from the Hanford Reach to McNary Dam. Large populations of piscivorous fishes and birds
inhabit the Columbia River and may be responsible for this mortality. We implanted 200 wild Hanford
Reach and 200 Priest Rapids Hatchery (PRH) URB fall Chinook salmon with acoustic transmitters and
estimated their survival through multiple reaches of the Columbia River to identify mortality “hot spots”
and to help classify the putative source(s) of mortality.
Acoustic-tagged wild Hanford Reach fall Chinook salmon had an estimated survival probability of
0.50 from release to McNary Dam. This estimate is considerably higher than was observed in 2014 for
the group of wild Hanford Reach fall Chinook salmon juveniles implanted with passive integrated
transponders (PIT-only; S = 0.34). The large discrepancy between survival estimates derived from
acoustic-tagged versus PIT-only groups is likely a result of the difference in fish size between groups.
We attempted to minimize the effect of the transmitter on the performance of implanted fish by only
tagging fish that measured ≥80 mm FL; whereas, fish as small as 60 mm FL were implanted with PIT
tags. As we demonstrated, survival of these fish is strongly, positively correlated with fish length.
Therefore, we expect that the survival of the overall population of juvenile wild Hanford Reach fall
Chinook salmon through the study area was substantially lower than it was for acoustic-tagged fish.
However, we believe that the relative losses of tagged fish by reach were representative of the overall
population.
Acoustic-tagged PRH smolts also had an estimated survival probability of 0.50 from release to
McNary Dam; albeit over a longer reach than was traversed by the wild group. This estimate is
substantially lower than what was observed for PIT-only PRH smolts in 2014 (S = 0.66). The difference
in survival between groups of acoustic-tagged and PIT-only PRH fall Chinook salmon juveniles may
have been the result of a reduction in performance of acoustic-tagged fish caused by the tagging
procedure or presence of the tag, and/or a result of acoustic transmitter loss (i.e., tag shedding). Although
results from a 60-day laboratory study conducted at PNNL found a very high rate of fish survival (99.2%)
and tag retention (100%) of 126 fish implanted with the same transmitter and surgical technique, we
observed relatively high post-tagging, pre-release mortality for the group of PRH fall Chinook salmon we
implanted with acoustic transmitters for the in-river survival evaluation described in this report.
Because reaches differed in length, survival is better compared among reaches on a per-kilometer
basis to identify potential mortality “hot spots”. Survival-per-kilometer (Skm) was generally lower in the
transition area between the Hanford Reach and McNary Reservoir, within McNary Reservoir, and in the
upper half of John Day Reservoir (down to Crow Butte) than in reaches located downstream of Crow
Butte. The lowest Skm was observed in the immediate forebay of McNary Dam for both wild and hatchery
fish. As expected, travel rates were fastest in flowing reaches (i.e., Hanford Reach and dam tailraces) and
slowest through reservoirs. We observed a significant, positive relationship between the probability of
survival to McNary Dam and fish length.
iv
Data from this study and others indicate much of the mortality incurred by URB fall Chinook salmon
juveniles between Priest Rapids and Bonneville dams can likely be attributed to predation from resident
piscivorous fish. Analyzing 8 years of data, we observed no significant relationship between the survival
of PIT-only wild Hanford Reach fall Chinook salmon to McNary Dam and the size of the primary avian
predator nesting colonies located in McNary Reservoir. We also did not observe mortality “hot spots” in
the reaches of the Columbia River that contain the largest colonies of predaceous waterbirds. Instead, we
observed relatively consistent mortality rates between release and Crow Butte, which is more indicative
of predation from piscivorous fish, which are more widely distributed than avian predators. In addition,
results of studies conducted to assess avian predation rates have consistently estimated very low predation
rates (<2%) on subyearling fall Chinook salmon upstream of Bonneville Dam. Alternatively, predation
rates estimated for piscivorous fish suggest they may be consuming 17% of the juvenile salmon that enter
John Day Reservoir during June, July, and August, when most salmon smolts entering the reservoir are
subyearling fall Chinook salmon.
Our study confirmed that the loss rates of juvenile URB fall Chinook salmon from the Hanford Reach
were high in areas where habitat has been influenced by hydropower development and native and non-
native predatory fish species. Whereas our study had some limitations due to 1) the size of fish we were
able to tag, 2) the potential for a tag or tagging effect on fish performance, and 3) possible tag loss, we
believe that the relative loss rates are representative for the wild Hanford Reach and Priest Rapids
Hatchery portions of the URB stock. Much of the mortality appears to be concentrated in the
river/reservoir transition area where large predator-rich tributaries enter as well as in the immediate dam
forebays where travel rates of outmigrating smolts are slowed. Additional work to document how the
predation rates we observed in the larger size classes of juvenile URB fall Chinook salmon relate to the
overall population, as well as efforts to determine the potential effectiveness of management actions
intended to reduce the populations and/or productivity of piscivorous fish species will provide the
information necessary to enable managers to design and implement strategies to improve the freshwater
survival of this important stock.
v
Acknowledgments
We sincerely appreciate the cooperation of Jeff Fryer and the CRITFC crew, including Bobby Begay,
for providing us with wild Hanford Reach fall Chinook salmon to tag and for sharing their holding tanks
and tagging location with us. Similarly, we would like to thank Glen Pearson and Mike Lewis of WDFW
and the Grant County Public Utility District (Grant PUD) for providing us with PRH smolts to tag and
access to the outflow channel for installation of the cabled JSATS array. We thank Zachary Boot, Sam
Cartmell, Tao Fu, Xinya Li, Bo Liu, Terence Lozano, Jun Lu, Jayson Martinez, Jason Reynolds, Spencer
Sandquist, Jie Xiao, and Yong Yuan, as well as John Stephenson for transmitter development, tag life
results, and/or the installation and monitoring of the cabled JSATS array at PRH. We thank James
Hughes and the North Bonneville PNNL crew, including Mark Weiland and others who conducted the
JSATS studies at McNary and John Day dams for the Corps of Engineers in 2014. These people
deployed, serviced, maintained, and downloaded autonomous and cabled acoustic telemetry receiver
arrays downstream of rkm 524. Thanks to Jina Kim for assistance with data management, processing,
and validation. We would also like to thank Ricardo Walker, Megan Nims, Bryan Jones, and Stephanie
Liss for assisting with transmitter implantation and Scott Titzler, Bob Mueller, Brian Bellgraph, and Kyle
Larson for servicing autonomous acoustic telemetry receivers. We thank John Clark (ADFG, PSC) for
providing study review, insight, and support. Finally, we would like to thank the Pacific Salmon
Commission and Grant PUD for funding this effort and the U.S. Army Corps of Engineers for funding the
performance standard evaluations at McNary and John Day dams, which were conducted in parallel to
this study and provided us with much additional data.
vi
Acronyms and Abbreviations
AABM Aggregate Abundance Based Management
ATLAS Acoustic Tag Life Adjusted Survival
CJS Cormack-Jolly-Seber
CR Columbia River
CRITFC Columbia River Inter-Tribal Fish Commission
DOE Department of Energy
FCRPS Federal Columbia River Power System
FL fork length
HR Hanford Reach
HRFCPPA Hanford Reach Fall Chinook Protection Program Agreement
JBS juvenile bypass systems
JDA John Day
JSATS Juvenile Salmon Acoustic Telemetry System
MCN McNary
NPM northern pikeminnow
ODFW Oregon Department of Fish and Wildlife
PIT passive integrated transponder
PNNL Pacific Northwest National Laboratory
PRD Priest Rapids Dam
PRH Priest Rapids Hatchery
PRI pulse rate interval
SMB smallmouth bass
TDA The Dalles
URB Upriver Bright
VBSA Vernita Bar Settlement Agreement
WAL walleye
WDFW Washington State Department of Fish and Wildlife
vii
Contents
Abstract .............................................................................................................................................. iii
Acknowledgments ............................................................................................................................... v
Acronyms and Abbreviations ............................................................................................................. vi
1.0 Introduction .............................................................................................................................. 1.1
2.0 Materials and Methods ............................................................................................................. 2.1
2.1 Fish Collection, Tagging, and Release ............................................................................ 2.1
2.1.1 Fish Collection and Holding ................................................................................. 2.1
2.1.2 Transmitter Specifications .................................................................................... 2.3
2.1.3 Tagging Procedure ................................................................................................ 2.3
2.1.4 Recovery, Holding, and Release .......................................................................... 2.4
2.2 Site Description and Array Locations ............................................................................. 2.4
2.2.1 Site Description .................................................................................................... 2.4
2.2.2 Acoustic Receiver Locations ................................................................................ 2.5
2.3 Autonomous Receiver Data Processing and Validation .................................................. 2.8
2.3.1 Time Correction .................................................................................................... 2.9
2.3.2 Filtering ................................................................................................................ 2.9
2.4 Tag-Life ........................................................................................................................... 2.9
2.5 Survival Estimation ......................................................................................................... 2.9
2.6 Travel Time and Travel Rate ......................................................................................... 2.11
3.0 Results ...................................................................................................................................... 3.1
3.1 Environmental Conditions ............................................................................................... 3.1
3.2 Size of Tagged Fish ......................................................................................................... 3.4
3.3 JSATS Performance ........................................................................................................ 3.6
3.3.1 Tag-Life ................................................................................................................ 3.6
3.3.2 Array Detection Probability ................................................................................. 3.8
3.4 Survival Probability ......................................................................................................... 3.9
3.4.1 Wild Hanford Reach Fall Chinook Salmon .......................................................... 3.9
3.4.2 Priest Rapids Hatchery Fall Chinook Salmon .................................................... 3.13
3.5 Travel Time and Travel Rate ......................................................................................... 3.17
3.5.1 Wild Hanford Reach Fall Chinook Salmon ........................................................ 3.17
3.5.2 Priest Rapids Hatchery Fall Chinook Salmon .................................................... 3.19
4.0 Discussion ................................................................................................................................ 4.1
5.0 References ................................................................................................................................ 5.1
viii
Figures
2.1. Map of the Columbia River from Priest Rapids Dam to McNary Dam ................................... 2.2
2.2. Photo of the Juvenile Salmon Acoustic Telemetry System (JSATS) acoustic transmitter
implanted in juvenile fall Chinook salmon in the Hanford Reach and at Priest Rapids
Hatchery in 2014 ...................................................................................................................... 2.3
2.3. Map of the Columbia River from McNary Dam to Bonneville Dam ...................................... 2.6
2.4. Google Earth image of Priest Rapids Dam that displays the channel pond in which acoustic-
tagged fish were held following surgery and the cabled JSATS array and PIT array that were
located in the outflow channel to detect acoustic- and PIT-tagged fish as they migrated from
the channel pond to the Columbia River. ................................................................................. 2.8
3.1. Priest Rapids Dam discharge from May 15 through August 7, 2014 versus the 10-year
average ..................................................................................................................................... 3.1
3.2. McNary Dam discharge from May 15 through August 7, 2014 versus the 10-year average .. 3.2
3.3. Water temperature, as measured at Priest Rapids Dam, from May 15 through August 7, 2014
versus the 10-year average ....................................................................................................... 3.3
3.4. Water temperature, as measured at McNary Dam, from May 15 through August 7, 2014
versus the 10-year average ....................................................................................................... 3.4
3.5. Length frequency distributions for acoustic-tagged Priest Rapids Hatchery fall Chinook
salmon smolts, wild Hanford Reach fall Chinook salmon juveniles captured via seining that
were implanted with acoustic transmitters, and wild Hanford Reach fall Chinook salmon
juveniles captured via seining that were randomly selected for length measurement in 2014 3.5
3.6. Fitted three-parameter Weibull model tag-life survivorship curve and the arrival-time
distributions of acoustic-tagged wild Hanford Reach and Priest Rapids Hatchery fall
Chinook salmon juveniles at the McNary Dam cabled array .................................................. 3.6
3.7. Fitted three-parameter Weibull model tag-life survivorship curve and the arrival-time
distributions of acoustic-tagged wild Hanford Reach and Priest Rapids Hatchery fall
Chinook salmon juveniles at the autonomous array located near Bingen, Washington .......... 3.7
3.8. Observed failure times of tag-life acoustic transmitters and the fitted three-parameter
Weibull model survivorship curve used to adjust survival estimates for tag-life .................... 3.8
3.9. Overall cumulative survival probability estimates for acoustic-tagged wild Hanford Reach
fall Chinook salmon from release in the Hanford Reach to downstream acoustic telemetry
receiver arrays ........................................................................................................................ 3.10
3.10. Survival probability-per-kilometer estimates for acoustic-tagged wild Hanford Reach fall
Chinook salmon through reaches of the Columbia River, 2014 ............................................ 3.12
3.11. Covariate analysis results displaying nonparametric and modeled survival probabilities of
acoustic-tagged wild Hanford Reach fall Chinook salmon from release in the Hanford Reach
to McNary Dam in relation to fork length ............................................................................. 3.13
3.12. Overall cumulative survival probability estimates for acoustic-tagged Priest Rapids Hatchery
fall Chinook salmon from acoustic detection in the PRH outflow channel to downstream
acoustic telemetry receiver arrays .......................................................................................... 3.14
3.13. Survival probability-per-kilometer estimates for acoustic-tagged Priest Rapids Hatchery fall
Chinook salmon through reaches of the Columbia River, 2014 ............................................ 3.16
ix
3.14. Covariate analysis results displaying nonparametric and modeled survival probabilities of
acoustic-tagged Priest Rapids Hatchery fall Chinook salmon from Priest Rapids Hatchery to
McNary Dam in relation to fork length .................................................................................. 3.17
3.15. Travel time of acoustic-tagged wild Hanford Reach fall Chinook salmon juveniles in each
reach of the Columbia River studied in 2014 ......................................................................... 3.18
3.16. Travel rate of acoustic-tagged wild Hanford Reach fall Chinook salmon juveniles in each
reach of the Columbia River studied in 2014 ......................................................................... 3.19
3.17. Travel time of acoustic-tagged Priest Rapids Hatchery fall Chinook salmon juveniles in each
reach of the Columbia River studied in 2014 ......................................................................... 3.20
3.18. Travel rate of acoustic-tagged Priest Rapids Hatchery fall Chinook salmon juveniles in each
reach of the Columbia River studied in 2014 ......................................................................... 3.21
4.1. Total numbers of northern pikeminnow, smallmouth bass, and walleye captured during
Oregon Department of Fish and Wildlife electrofishing surveys conducted annually from
1993–2010 between McNary and Priest Rapids dams. ............................................................ 4.4
4.2. Relationship between annual survival probability of PIT-tagged wild Hanford Reach fall
Chinook salmon and the number of Caspian tern breeding pairs counted on colonies of the
Columbia Plateau ..................................................................................................................... 4.6
x
Tab le s
2.1. Locations of cabled and autonomous acoustic telemetry receiver arrays deployed in the
Columbia River to detected JSATS acoustic transmitters during the spring and summer of
2014. ......................................................................................................................................... 2.7
3.1. Number, fork length, tag burden, and release dates for acoustic-tagged Priest Rapids
Hatchery upriver bright fall Chinook salmon juveniles and wild Hanford Reach upriver
bright fall Chinook salmon juveniles released into the Columbia River at Priest Rapids
Hatchery or in the Hanford Reach in 2014 .............................................................................. 3.5
3.2. Probability of detecting acoustic-tagged Priest Rapids Hatchery and wild Hanford Reach fall
Chinook salmon at autonomous and cabled JSATS acoustic telemetry receiver arrays
deployed in the mid and lower Columbia River in 2014 ......................................................... 3.9
3.3. Reach-specific survival probability estimates for acoustic-tagged wild Hanford Reach fall
Chinook salmon juveniles through each river reach studied in 2014 from release at rkm 595
to CR275 ................................................................................................................................ 3.11
3.4. Reach-specific survival probability estimates for acoustic-tagged Priest Rapids Hatchery fall
Chinook salmon juveniles through each river reach studied in 2014 from virtual release at
rkm 633 to CR275 .................................................................................................................. 3.15
1.1
1.0 Introduction
The population of fall Chinook salmon Oncorhynchus tshawytscha that inhabits the Hanford Reach
comprises the majority of the Columbia River Upriver Bright (URB) stock and is one of the most
productive Chinook salmon stocks in the Pacific Northwest (Peters et al. 1999; Langness and Reidinger
2003; Harnish et al. 2012, 2013). As such, it is able to sustain high rates of harvest and therefore has
great economic and cultural importance to native peoples and commercial and recreational fishers. The
URB stock is a far north-migrating stock and is an important contributor to all three Aggregate
Abundance Based Management (AABM) fisheries and a primary contributor to Columbia River fisheries.
Recent studies have indicated that much of the high productivity of the population may be attributed
to very high survival during early freshwater life stages. In fact, results from a cohort reconstruction
indicated that nearly two-thirds (65%) of the broods from 1975 through 2004 that displayed above-
average egg-to-presmolt survival also had above-average adult/spawner production. Thus, Hanford
Reach fall Chinook salmon brood year strength appears to be largely determined by interannual variation
in freshwater survival, indicating the importance of the freshwater life phase to the overall productivity of
the population. Enactment of operational constraints to limit discharge fluctuations downstream of Priest
Rapids Dam have resulted in increased productivity and egg-to-pre-smolt survival rates. Harnish et al.
(2014) observed a 217% increase in egg-to-pre-smolt productivity (Ricker α) that corresponded with
constraints enacted by the Vernita Bar Settlement Agreement (VBSA), which limited redd dewatering,
and an additional 130% increase that coincided with enactment of the interim Hanford Reach Fall
Chinook Protection Program Agreement (HRFCPPA) in 1999, which limited stranding and entrapment of
juveniles. Additionally, the average egg-to-pre-smolt survival probability estimate increased from 0.30
during the pre-VBSA period (brood years [BY] 1975–1983) to 0.36 during the period of the VBSA (BY
1984–1998) to 0.42 during the HRFCPPA period (BY 1999–20041). In addition, a study conducted in
2012 estimated the egg-to-fry survival of fall Chinook salmon to be 71% in the Hanford Reach
(Oldenburg et al. 2012). The survival rates discovered during these studies for the Hanford Reach
population are much higher than those reported for other populations of Chinook salmon. From 215
published and unpublished estimates for wild or naturally rearing populations of Chinook salmon, Quinn
(2005) calculated a mean egg-to-fry survival of 38% and a mean egg-to-smolt survival of 10%.
Although egg-to-pre-smolt survival has been found to be very high for the Hanford Reach fall
Chinook salmon population, survival from pre-smolt to age-3 adult equivalent averaged just 0.29% for
BY 1986–2004. Some evidence suggests significant mortality of smolts occurs over a short period of
time and distance as they emigrate from the Hanford Reach to McNary Dam. Survival from release in the
Hanford Reach to McNary Dam has averaged just 37% since 1995 for PIT-tagged wild fall Chinook
salmon juveniles (Fish Passage Center 2013). Annual losses of this magnitude represent an obvious
bottleneck to production.
Large populations of piscivorous fishes, such as smallmouth bass Micropterus dolomieu, northern
pikeminnow Ptychocheilus oregonensis, walleye Sander vitreus, and channel catfish Ictalurus punctatus
inhabit the Columbia River along with nesting colonies of avian predators, such as terns, cormorants,
1 The Priest Rapids Project is currently operated under the HRFCPPA. The productivity analysis conducted by
Harnish et al. (2012, 2013) included BY 1975–2004.
1.2
gulls, and pelicans. The objective of this study was to estimate survival of acoustic-tagged Hanford
Reach fall Chinook salmon juveniles through multiple reaches of the Columbia River to identify mortality
“hot spots” and help to classify the putative source(s) of mortality (i.e., fish or birds).
2.1
2.0 Materials and Methods
2.1 Fish Collection, Tagging, and Release
2.1.1 Fish Collection and Holding
Wild Hanford Reach fall Chinook salmon
Wild fall Chinook salmon juveniles were collected from multiple locations in the Hanford Reach
during the first week of June 2014 by Columbia River Inter-Tribal Fish Commission (CRITFC) personnel
using stick and beach seines (4.8 mm mesh size). Seining was conducted in sections of the river with
moderate velocity and 0.3 m to 1.4 m depth that were primarily located upstream of the tagging site at the
Hanford town site boat ramp (river kilometer [rkm] 582; as measured from the mouth of the Columbia
River) to reduce the likelihood of re-capturing previously tagged fish (Figure 2.1).
Captured wild fall Chinook salmon juveniles were temporarily placed into 19-L plastic buckets before
being transferred to the oxygen-aerated holding tank of the boat. Once a full load of approximately
10,000 fish had been captured, or about three hours had passed, the fish were transported to the tagging
site located at the Hanford town site boat ramp. Fish were then transported from the boat in the 19-L
plastic buckets to a 0.9 m × 0.9 m × 4.9 m fiberglass tank equipped with a pump to provide a continuous
flow of river water. Fish that measured >80 mm fork length (FL) were held separately from smaller fish
in four partially-perforated 76-L buckets within the fiberglass tanks. Captive fish were not directly fed;
however, they did have access to organisms present in the river water. On June 5, 2014, surgical
candidates were netted from the perforated holding buckets in batches of 20 to 50 and transferred in a 38-
L bucket to the mobile tagging trailer.
2.2
Figure 2.1. Map of the Columbia River from Priest Rapids Dam to McNary Dam. Locations of cabled
and autonomous acoustic telemetry receiver arrays deployed in 2014 are shown as a
concatenation of “CR” and the river kilometer (as measured from the mouth of the Columbia
River) at which they were deployed. Other locations of interest are also shown: these include
the tagging and release location for acoustic-tagged wild Hanford Reach fall Chinook salmon
juveniles; and the islands that host avian predator nesting colonies.
Priest Rapids Hatchery fall Chinook salmon
Hatchery URB fall Chinook salmon juveniles were reared at the Priest Rapids Hatchery (PRH) from
the time of spawning until release to the river in June 2014. The hatchery is located along the bank of the
Columbia River immediately downstream of Priest Rapids Dam and is operated by Washington State
Department of Fish and Wildlife (WDFW) and owned by the Public Utility District No. 2 of Grant
County, Washington. Prior to tagging, juvenile hatchery fall Chinook salmon were held at PRH in a
concrete raceway supplied with a continuous flow of river water. Food was withheld for 24 h prior to
tagging. On May 28, 2014, surgical candidates were netted from the raceway and transferred to the
mobile tagging trailer in a 38-L plastic bucket in batches of 20 to 50 fish.
2.3
2.1.2 Transmitter Specifications
All Chinook salmon were implanted with an acoustic transmitter and a passive integrated transponder
(PIT). The mean dimensions of the downsized Juvenile Salmon Acoustic Telemetry System (JSATS)
acoustic transmitter (developed by the U.S. Army Corps of Engineers and Pacific Northwest National
Laboratory [PNNL]; Chen et al. 2014) were 15.0 mm long by 3.3 mm in diameter (Figure 2.2).
Transmitters had a mean weight in air of 0.22 g, a mean weight in water of 0.11 g, and a mean volume of
0.11 mL. The transmitters had a nominal pulse rate interval (PRI) of one complete transmission every 3
seconds with a source level of 155–156 dB. The nominal transmitter life was expected to be about 60
days. The PIT tag (Model HPT12, Biomark, Inc., Boise, Idaho) was 12.5 mm long, 2 mm wide, and
weighed 0.10 g in air (0.06 g in water; 0.04 mL volume; 134.2 kHz). The combined weight of the tags
gave each implanted fish an added burden of 0.32 g in air.
Figure 2.2. Photo of the Juvenile Salmon Acoustic Telemetry System (JSATS) acoustic transmitter
implanted in juvenile fall Chinook salmon in the Hanford Reach and at Priest Rapids
Hatchery in 2014. The transmitter is shown next to a metric ruler to display the size of the
transmitter.
2.1.3 Tagging Procedure
After surgical candidates were delivered to the mobile tagging trailer, they were anesthetized in
batches of 2-3 fish. A dose of 80 mg/L of tricaine methanesulfonate (MS-222; buffered with 80 mg/L of
sodium bicarbonate) was used to sedate juvenile Chinook salmon to stage 4 anesthesia (as described by
Summerfelt and Smith 1990). The FL and weight of sedated fish were obtained as the acoustic
transmitter and PIT tag codes were assigned. Only fish that measured ≥80 mm FL were selected for this
study based on results from a 60-day laboratory study conducted at PNNL that found a very high rate of
fish survival (99.2%) and tag retention (100%) of fish implanted with both a PIT tag and downsized
JSATS transmitter (n = 126 fish) using the same surgical technique described below.
An anesthetized fish and tags were delivered to one of two surgeons. The fish was placed on its left
side in a small pool of water on a foam pad lubricated with PolyAqua, a water conditioner. The surgeon
then made a shallow incision 3 mm in length with a sterile #11 surgical blade approximately 3-5 mm
away from the linea alba and beneath the distal end of the pectoral fin. The PIT tag was then inserted
through the incision into the peritoneal cavity. With the subsequent insertion of the acoustic transmitter,
there was an attempt by the surgeon to get the two tags side-by-side (not end-to-end), by slightly changing
the angle of insertion, with the intent to reduce the likelihood of tag expulsion through the open wound.
In addition, the wound was gently massaged posteriorly, to ensure the tags were completely inside the
peritoneal cavity and to move them away from the incision opening. Immediately, the tagged fish was
2.4
placed in a recovery bucket filled with aerated river water. The entire surgical process took
approximately 20 to 30 s per fish. The same two surgeons tagged all fish in the wild Hanford Reach and
Priest Rapids Hatchery groups; each surgeon tagged half of each group.
2.1.4 Recovery, Holding, and Release
Wild Hanford Reach fall Chinook salmon
Following implantation, the 200 acoustic-tagged wild Hanford Reach fall Chinook salmon juveniles
were held at densities of less than 5 g/L in four partially-perforated 76-L buckets that were placed into a
0.9 m × 0.9 m × 4.9 m fiberglass holding tank equipped with a pump to provide a continuous flow of river
water. The buckets remained in the holding tank for about 24 hours until the day of release (June 6) when
they were loaded into a trailered boat and transported with continuous aeration to the White Bluffs boat
launch (rkm 595; Figure 2.1). The dissolved oxygen and water temperature in the buckets were measured
with an YSI meter before and during transport to ensure these metrics stayed within acceptable limits.
Once at the ramp, the boat was launched and maneuvered downstream 0.5 km to the release location.
Equal numbers of fish were released at four locations along a line transect across the river. Before
release, buckets were checked for dead fish and dropped tags then each bucket was submerged in the
water so that fish could swim out on their own volition.
Priest Rapids Hatchery fall Chinook salmon
After each group of PRH fall Chinook salmon juveniles were implanted with transmitters and had
recovered from surgery, they were placed in one of four partially-perforated 76-L buckets that were
suspended in the concrete raceway. The 200 tagged fish recovered at densities less than 5 g/L for
approximately 24 hours, at which time holding buckets were removed from the raceway and inspected for
dead fish and dropped tags. The buckets were taken to the adjacent channel pond and submerged in the
water so that fish could swim out on their own volition. Tagged hatchery Chinook salmon resided in the
channel pond for a full two weeks before releases of all fish in the channel ponds began with the removal
of the pond gates on June 12, 2014. However, acoustic-tagged juveniles were detected on PIT and
acoustic receiver arrays migrating from the outflow channel between June 12 and 21, 2014.
2.2 Site Description and Array Locations
The area of the Columbia River between Priest Rapids and McNary dams defines the primary study
area. However, data collected opportunistically from reaches of the Columbia River located between
McNary and The Dalles dams are also presented. The array locations used in this study were chosen to
differentiate survival among important reaches of the river and were selected because the associated river
characteristics allow for good detection of acoustic tags. This section provides details about where
detection arrays were deployed.
2.2.1 Site Description
The Hanford Reach, an 80-km stretch of river located between Priest Rapids Dam (river kilometer
[rkm] 639) and the head of McNary Reservoir (rkm 557) near the town of Richland, Washington, is the
last segment of the Columbia River that has not been inundated, dredged, or channelized (Whidden 1996)
2.5
and is available to anadromous salmonids (Figure 2.1). As such, the Hanford Reach contains the only
remaining substantial mainstem spawning area for fall Chinook salmon in the Columbia River
(Bauersfeld 1978; Chapman et al. 1986; Dauble and Watson 1997).
Three major tributaries, the Yakima, Snake, and Walla Walla rivers flow into the Columbia River in
McNary Reservoir. The Yakima River flows into McNary Reservoir at rkm 538, near the town of
Richland, Washington. The Yakima River has been identified as a major spawning area for smallmouth
bass of the Columbia River (Fritts and Pearsons 2004). The Snake River flows into McNary Reservoir at
rkm 522 and the Walla Walla River enters McNary Reservoir at rkm 505. Between the mouths of these
two rivers are three islands used as nesting and roosting sites by multiple piscivorous water bird species.
These include a nesting colony of cormorants on Foundation Island at rkm 518, a nesting colony of
pelicans on Badger Island at rkm 510, and nesting colonies of terns, gulls, and herons on Crescent Island,
a manmade island constructed of dredge spoils, at rkm 508 (Antolos et al. 2004, 2005; Evans et al. 2012).
2.2.2 Acoustic Receiver Locations
Acoustic transmissions from tagged fish were decoded by stationary JSATS autonomous receivers
(Model N201, Sonic Concepts, Inc., Bothell, Washington; and SR5000 Trident, Advanced Telemetry
Systems, Inc., Isanti, Minnesota), which were deployed via the methods described by Titzler et al. (2010).
In total, autonomous acoustic telemetry receivers were deployed at 48 locations between the head of
McNary Reservoir (near the town of Richland, Washington) and McNary Dam and at 94 locations
between McNary and Bonneville dams during the outmigration of fall Chinook salmon juveniles (June 6
to August 1). The majority of these receivers (n=133) were deployed for studies funded by the U.S.
Army Corps of Engineers, but data were made available for analyses for this predation loss assessment.
Receivers were deployed in lines, referred to as arrays, which ran approximately perpendicular to the
shore. Based on their effective detection range, receivers were spaced about 100 to 200 m apart.
A total of six autonomous receiver arrays were deployed upstream of McNary Dam and an additional
12 arrays were deployed downstream of McNary Dam (Figure 2.1 and Figure 2.3; Table 2.1). JSATS
acoustic transmissions were detected and decoded by these receiver arrays and used to estimate survival
and travel times of acoustic-tagged natural-origin Hanford Reach and PRH fall Chinook salmon juveniles
in the reaches located between the arrays.
In addition to the autonomous receivers deployed to detect acoustic-tagged fish in the Columbia
River, cabled JSATS receiver systems (Weiland et al. 2011) were deployed in the PRH outflow channel
and on the face of McNary and John Day dams for dam passage survival studies conducted by PNNL for
the U.S. Army Corps of Engineers. The deployment of hydrophones along the dam faces generally
followed the design and methodology described by Deng et al. (2011). Prior to field deployment, all
autonomous and cabled receivers were calibrated in an acoustic tank located at the PNNL Bio-Acoustics
and Flow Laboratory, which is accredited by the American Association for Laboratory Accreditation
(Deng et al. 2010). Detections of acoustic-tagged fall Chinook salmon juveniles on autonomous and dam
face systems were used for estimation of reach survival and travel times.
2.6
Figure 2.3. Map of the Columbia River from McNary Dam to Bonneville Dam. Locations of cabled and
autonomous acoustic telemetry receiver arrays deployed in 2014 are shown as a
concatenation of “CR” and the river kilometer (as measured from the mouth of the Columbia
River) at which they were deployed.
2.7
Table 2.1. Locations of cabled and autonomous acoustic telemetry receiver arrays deployed in the
Columbia River to detected JSATS acoustic transmitters during the spring and summer of 2014.
Array
Location
Latitude
Longitude
Autonomous or
Cabled
# of
Hydrophones
CR633
Priest Rapids Hatchery
46.637033
-119.878798
Cabled
4
CR552
Richland, WA
46.352325
-119.2610833
Autonomous
6
CR524
Snake River
46.19887143
-119.051
Autonomous
7
CR498
Port Kelly, WA
45.99767042
-118.9906672
Autonomous
8
CR489
Van Skinner Island
45.95559753
-119.0678685
Autonomous
9
CR480
Hat Rock
45.92761134
-119.1772737
Autonomous
10
CR472
McNary Dam forebay
45.9393581
-119.2732181
Autonomous
8
CR470
McNary Dam
45.93569206
-119.2974027
Cabled
89
CR455
Irrigon, OR
45.90591741
-119.4938649
Autonomous
6
CR449
Paterson, WA
45.92042962
-119.5584028
Autonomous
6
CR439
Boardman, OR
45.88502679
-119.6585717
Autonomous
10
CR422
Crow Butte
45.83856001
-119.8555434
Autonomous
7
CR412
Willow Lake
45.82475567
-119.9559078
Autonomous
8
CR381
Sundale, WA
45.71256584
-120.3204116
Autonomous
6
CR351
John Day Dam forebay
45.72255187
-120.6810192
Autonomous
8
CR349
John Day Dam
45.71583597
-120.6929465
Cabled
85
CR325
Wishram, WA
45.65323093
-120.9653195
Autonomous
18
CR311
The Dalles Dam forebay
45.62699729
-121.1126313
Autonomous
15
CR275
Bingen, WA
45.70758426
-121.472588
Autonomous
6
CR236
Bonneville forebay
45.64968612
-121.9202764
Autonomous
4
The cabled system deployed in the PRH outflow channel consisted of four hydrophones located in the
deepest pool of the channel, which was located about 1.5 km downstream from the holding ponds and 1.1
km upstream from the mouth of the channel (Figure 2.4). A PIT array consisting of multiple antennas
was also present in the outflow channel about 200 m upstream from the mouth of the channel. Detections
of the double-tagged (acoustic + PIT) fall Chinook salmon juveniles on the cabled JSATS and PIT arrays
were used to evaluate post-tagging/pre-release mortality and tag loss/failure and to estimate the number of
tagged fish that actually left the hatchery with an active acoustic transmitter. Detections of double-tagged
fall Chinook salmon in the juvenile bypass systems (JBS) of McNary and John Day dams were also used
to evaluate acoustic tag loss/failure.
2.8
Figure 2.4. Google Earth image of Priest Rapids Dam that displays the channel pond in which acoustic-
tagged fish were held following surgery and the cabled JSATS array (CR633) and PIT array
that were located in the outflow channel to detect acoustic- and PIT-tagged fish as they
migrated from the channel pond to the Columbia River.
2.3 Autonomous Receiver Data Processing and Validation
Signals received by JSATS receivers were processed and filtered to validate the presence of a tagged
fish within the vicinity of a receiver at a specific time. Receivers recorded receptions of possible tag
signals along with a timestamp for each reception. Raw files from autonomous receivers were time-
corrected and files from both autonomous and cabled receivers were filtered to remove spurious
receptions. The time series of validated locations for individual fish were then used to estimate survival
rates and travel times. A laboratory study of tag-life was conducted to allow estimates to be corrected for
early tag failures if necessary.
2.9
2.3.1 Time Correction
Some of the autonomous receivers used in this study were subject to clock errors that resulted in
timestamps being incorrect at unpredictable times throughout the file. Raw files were processed through
a time correction application to repair incorrect timestamps based upon correct timestamps that preceded
it. In many cases, the algorithm precisely identified a correction that was accurate to the second, whereas
in others, the correction resulted in a difference of a few seconds for the block of data being corrected.
After time correction, the files are referred to as time-corrected files, whether or not a correction was
needed and applied.
2.3.2 Filtering
Because JSATS autonomous receivers are configured to detect tag signals just above the acoustic
noise floor, raw files often include spurious receptions that arise from noise in addition to valid tag
signals (Ingraham et al. 2014). To filter out detections that did not meet criteria (false detections), a post-
processing program was used (McMichael et al. 2010). This program comprised a sequence of steps that
included comparing each detection to a list of tags that were released (only detections of tags that were
released were kept), then comparing the detection date to the release data (only tags detected after they
were released were kept). Then, a minimum of four detections in 60 seconds was required, and the time
spacing between these detections had to match the PRI of the tag or be a multiple of the PRI for the
detections to be kept in the valid detection file. This final filter takes advantage of the fact that spurious
receptions do not exhibit the temporal consistency among pulses that is characteristic of an actively
transmitting JSATS tag.
2.4 Tag-Life
For the tag-life study, 32 tags (3-s PRI) were randomly chosen over time from the manufacturing line
of tags used in this study. All tag-life tags were enclosed in water-filled plastic bags and suspended from
a rotating foam ring within a 2-m (diameter) fiberglass tank. Two 90o × 180o hydrophones were
positioned 90o apart in the bottom of the tank and angled upward at approximately 60o to maximize
coverage for detecting acoustic signals. Hydrophones were cabled to a quad-channel receiver that
amplified all acoustic signals, which were then saved, decoded, and post-processed. Post-processing
software calculated the number of hourly decodes for each acoustic tag, allowing tag failure times to be
determined within + 1 hour.
2.5 Survival Estimation
Survival estimates were derived from conventional statistical models for mark-recapture data
(Cormack 1964; Jolly 1964; Seber 1965; Skalski et al. 1998). This model is known by various names,
including Cormack-Jolly-Seber (CJS), Single-Release, or Single-Release-Recapture Model. For survival
(Si) and detection (pi) probability estimation of mark-recapture data, detection data are summarized as the
“detection history” for each marked fish. With only two opportunities for detection, the possible
detection histories for tagged fish are:
00 = never detected;
10 = detected by the upstream (primary) array but not the downstream (secondary) array(s);
2.10
01 = detected by the downstream (secondary) array(s) but not by the upstream (primary) array;
and
11 = detected by the upstream (primary) array and the downstream (secondary) array(s).
To estimate survival to the primary array for a release group of tagged fish, the number of fish in the
group with each detection history is determined, denoted n00, n10, n01, and n11, along with the total number
of fish released (R). The proportion of fish detected on the primary array [(n10 + n11)/R] is an estimate of
the joint probability that a fish survived from release to the primary array and that the fish was detected
given that it survived. The joint probability of both events occurring is the simple product of the two
probabilities.
To separate the two probabilities in the product requires a method to estimate either of the
probabilities individually. The remaining probability can then be estimated by dividing the joint estimate
by the estimate of the first. Detection probability of the primary array can be estimated independently by
assuming that fish that survived to the secondary array and were detected there (n11 + n01) represent a
random sample of all fish from the group that were alive as they passed the primary array. Detection
probability of the primary array is then estimated as the proportion of the sample detected by the primary
array (i.e., n11/[n11 + n01]).
The program ATLAS (Acoustic Tag Life Adjusted Survival; version 1.5.3;
http://www.cbr.washington.edu/analysis/apps/atlas) and the methods described by Townsend et al.
(2006) were used to adjust CJS survival estimates for the probability of premature tag failure.
Preliminary tag-life data were fit with the two- and three-parameter Weibull models and the vitality
model of Li and Anderson (2009). The model that provided the best fit to the tag-life data was used to
adjust survival estimates by the conditional probability of a tag being active at each detection array.
Cumulative survival of acoustic-tagged PRH and wild Hanford Reach fall Chinook salmon was
estimated from release to the each downstream detection array. For PRH fish, only those fish that were
detected by the cabled JSATS receiver array located in the outflow channel were included in the
estimate1. Survival was also estimated for each river reach located between receiver arrays by forming a
“virtual release” of fish detected by the upstream (primary) array.
Because the distance between receiver arrays was not equal, it was desirable to have a measure of
reach survival that was independent of the distance over which it was estimated. Therefore, survival per
river kilometer was estimated from each reach survival estimate by:
Skm = Sreach
1/L
where
Skm is the estimate of survival per river kilometer,
Sreach is the reach survival estimate, and
L is the reach length in river kilometers.
We assessed the effect of fish length on the probability of survival to McNary Dam for acoustic-
tagged fall Chinook salmon in 2014 using program SURPH (SURvival under Proportional Hazards;
version 3.5.2), whereby survival probabilities were modeled as a function of FL as an individual-based
covariate using the hazard link (Skalski et al. 1993; Smith et al. 1994). A nonparametric survival curve
1 Survival from tagging to acoustic tag detection in the outflow channel was estimated separately.
2.11
that did not depend on the parameters of any particular model was also plotted. The nonparametric curve
can be thought of as a “moving average” survival as the selected individual covariate (i.e., fish length)
increases across the range of observed values. The size of the “window” for which the moving average
survival probability was calculated ranged from a minimum of eight individuals up to 20% of the entire
number at risk in the selected interval (Smith et al. 1994). The effect of fish length on survival
probability was evaluated using the likelihood ratio test to compare the fish length covariate model to a
model of no covariate effect.
2.6 Tra v e l Time and Travel Rate
Travel time was calculated for acoustic-tagged wild Hanford Reach and PRH fall Chinook salmon in
each river reach studied in 2014. Travel time was calculated for each fish detected at both the upstream
and downstream arrays by subtracting the date and time of first detection (or release) at the upstream
array from the date and time of the first detection at the downstream array. Travel rate was calculated
from each travel time by dividing the travel time by the distance between arrays. Because calculation of
travel time requires detection at both the upstream and downstream arrays, estimates of travel time and
travel rate within each reach only consider fish that successfully migrated through the entire reach and
were detected at both arrays.
3.1
3.0 Results
The results section includes a brief summary of the environmental conditions in the mid-Columbia
River during the study period to provide context for the detailed results of the estimated survival and
travel time of acoustic-tagged fish in this study. The tag-life and detection probability information for the
JSATS used in this study are presented to provide the necessary background information on system
performance.
3.1 Environmental Conditions
A sharp decline in total daily discharge, as measured at Priest Rapids (Figure 3.1) and McNary
(Figure 3.2) dams, coincided with the release and early migration period of acoustic-tagged wild Hanford
Reach fall Chinook salmon, which were released on June 6. Discharge from Priest Rapids Dam declined
from about 6,000 m3/s on June 6 to about 3,500 m3/s on June 15. This reduction in discharge was part of
normal spring river management to allow for the refilling of reservoirs by June 30. The volitional release
of acoustic-tagged fall Chinook salmon from PRH began on June 12; thus, discharge was increasing
throughout much of their early migration period before stabilizing around the 10-year average.
Figure 3.1. Priest Rapids Dam (PRD) discharge from May 15 through August 7, 2014 versus the 10-year
(2004–2013) average. Dotted lines indicate the approximate time period in which acoustic-
tagged fish were affected by PRD discharge. This period included the time between the
release of wild Hanford Reach fall Chinook salmon (June 6) and the last detection at McNary
Dam (CR470; July 7).
3.2
Figure 3.2. McNary Dam (MCN) discharge from May 15 through August 7, 2014 versus the 10-year
(2004–2013) average. Dotted lines indicate the approximate time period in which acoustic-
tagged fish were affected by MCN discharge. This period included the time between the
first detection of acoustic-tagged fish at MCN (CR470; June 10) and the last detection at
CR275 (July 15).
The temperature of the mid-Columbia River, as measured at Priest Rapids Dam, was slightly above-
average during most of the study period (Figure 3.3). Colder than average water temperatures in the
Snake River resulted in near-average temperatures at McNary Dam during the period of interest (Figure
3.4).
3.3
Figure 3.3. Water temperature, as measured at Priest Rapids Dam, from May 15 through August 7, 2014
versus the 10-year (2004–2013) average. Dotted lines indicate the approximate time period
in which acoustic-tagged fish were affected by the temperature in this area of the Columbia
River. This period included the time between the release of wild Hanford Reach fall
Chinook salmon (June 6) and the last detection at McNary Dam (CR470; July 7).
3.4
Figure 3.4. Water temperature, as measured at McNary Dam (MCN), from May 15 through August 7,
2014 versus the 10-year (2004–2013) average. Dotted lines indicate the approximate time
period in which acoustic-tagged fish were affected by water temperatures near MCN. This
period included the time between the first detection of acoustic-tagged fish at MCN (CR470;
June 10) and the last detection at CR275 (July 15).
3.2 Size of Tagged Fish
The length distributions of acoustic-tagged PRH and wild Hanford Reach fall Chinook salmon were
similar at the time of tagging (Figure 3.5; Table 3.1). However, fish tagged at PRH were held in the
channel ponds and fed for an additional two weeks after tagging, whereas wild Hanford Reach fall
Chinook were released the day after tagging. Based on the water temperature of the Columbia River
during this time (~12 oC) and the temperature-growth relationship of PRH fall Chinook salmon from a
laboratory study, we would expect these fish to grow an additional 9 mm between tagging and release.
Thus, we suspect the acoustic-tagged PRH fish were significantly larger, on average, than the acoustic-
tagged wild Hanford Reach at the time they entered the river in the Hanford Reach. Both tagged groups
were substantially larger than the random subsample of wild Hanford Reach fall Chinook salmon that
were captured in seines and measured for length by the CRITFC (Figure 3.5). We attempted to minimize
tag burden (tag weight expressed as a percentage of fish weight) and any potential tag or tagging effects
by only implanting tags into fish that measured 80 mm FL or greater. Therefore, the size distribution of
wild Hanford Reach fall Chinook salmon implanted with transmitters differed significantly from the size
distribution of the general population.
3.5
Figure 3.5. Length frequency distributions for acoustic-tagged Priest Rapids Hatchery fall Chinook
salmon smolts (PRH), wild Hanford Reach fall Chinook salmon juveniles captured via
seining that were implanted with acoustic transmitters (HR), and wild Hanford Reach fall
Chinook salmon juveniles captured via seining that were randomly selected for length
measurement in 2014 (Seined).
Table 3.1. Number, fork length, tag burden, (acoustic + PIT tag weight expressed as a percentage of fish
body weight), and release dates for acoustic-tagged Priest Rapids Hatchery upriver bright fall
Chinook salmon juveniles (H-URB) and wild Hanford Reach upriver bright fall Chinook
salmon juveniles (W-URB) released into the Columbia River at Priest Rapids Hatchery or in
the Hanford Reach in 2014 (rkm = river kilometer; min = minimum; max = maximum)
Release
Release
Rearing
Release
Fork length (mm)
Tag burden (%)
location
rkm
type
date
n
Min
Max
Mean
Min
Max
Mean
Priest
Rapids
Hatchery
633
H-URB
June 12
200
80
103
87
2.6
6.8
4.6
Hanford
Reach
595
W-URB
June 6
198
80
100
87
3.1
6.8
4.9
3.6
3.3 JSATS Performance
3.3.1 Tag-Life
Although tag-life expectancy was 60 days for acoustic tags in this study, 30 of the 32 (93.8%) tag-life
transmitters lasted longer than 60 days. In fact, the average transmitter life was 101.5 days at the time of
this report (September 10, 2014). However, 25 of the tags were still transmitting at the time of this report,
having been activated 97 to 131 days ago. Therefore, the actual average life of tag-life transmitters is
greater than 101.5 days. The first tag-life transmitter expired after 47.6 days. Because greater than 99%
of the fish we tagged migrated through the study area before the time at which any tag failure was
observed during the tag-life study (Figure 3.6 and Figure 3.7), only a relatively small adjustment for tag
failure was required. The three-parameter Weibull model (Figure 3.8) fit the preliminary tag-life data
better than either the two-parameter Weibull model or the four-parameter vitality model of Li and
Anderson (2009). Therefore, this tag-life survivorship model was subsequently used to estimate the
probabilities of tag failure and provide tag-life adjusted estimates of juvenile fall Chinook salmon
survival.
Figure 3.6. Fitted three-parameter Weibull model tag-life survivorship curve (red line) and the arrival-
time distributions of acoustic-tagged wild Hanford Reach (green line) and Priest Rapids
Hatchery (blue line) fall Chinook salmon juveniles at the McNary Dam cabled array
(CR470).
3.7
Figure 3.7. Fitted three-parameter Weibull model tag-life survivorship curve (red line) and the arrival-
time distributions of acoustic-tagged wild Hanford Reach (green line) and Priest Rapids
Hatchery (blue line) fall Chinook salmon juveniles at the autonomous array located near
Bingen, Washington (CR275).
3.8
Figure 3.8. Observed failure times of tag-life acoustic transmitters (+) and the fitted three-parameter
Weibull model survivorship curve used to adjust survival estimates for tag-life. The average
tag-life at the time of this report was 101.5 days. Bold crosses (+) indicate transmitters that
were still transmitting at the time of this report (September 10, 2014); thus, the model does
not fit the data particularly well and tag-life is likely underestimated.
3.3.2 Array Detection Probability
Detection probability was quite high at all arrays for both PRH and wild Hanford Reach fall Chinook
salmon (Table 3.2). The probability of detecting acoustic-tagged fish was ≥0.945 for PRH fish and
≥0.959 for wild fish at all arrays and equaled 1.000 at most arrays.
3.9
Table 3.2. Probability of detecting acoustic-tagged Priest Rapids Hatchery and wild Hanford Reach fall
Chinook salmon at autonomous and cabled JSATS acoustic telemetry receiver arrays
deployed in the mid and lower Columbia River in 2014.
Array
Wild Hanford Reach
Priest Rapids Hatchery
CR633
N/A
1.000 (0.000)
CR552
1.000 (0.000)
1.000 (0.000)
CR524
1.000 (0.000)
1.000 (0.000)
CR498
1.000 (0.000)
1.000 (0.000)
CR489
0.990 (0.010)
1.000 (0.000)
CR480
1.000 (0.000)
1.000 (0.000)
CR472
1.000 (0.000)
1.000 (0.000)
CR470
0.967 (0.019)
0.945 (0.027)
CR455
1.000 (0.000)
1.000 (0.000)
CR449
1.000 (0.000)
1.000 (0.000)
CR439
0.959 (0.023)
0.984 (0.016)
CR422
1.000 (0.000)
1.000 (0.000)
CR412
1.000 (0.000)
0.983 (0.017)
CR381
1.000 (0.000)
1.000 (0.000)
CR351
1.000 (0.000)
1.000 (0.000)
CR349
1.000 (0.000)
1.000 (0.000)
CR325
1.000 (0.000)
1.000 (0.000)
CR311
1.000 (0.000)
1.000 (0.000)
CR275
1.000 (0.000)
1.000 (0.000)
3.4 Survival Probability
Survival is an important metric for identifying when or where unfavorable conditions may exist for
juvenile fall Chinook salmon. Evaluating survival on a per-kilometer basis can put the reach survival
estimates into a relative context for comparisons between reaches. This section provides reach survival
probabilities and Skm estimates for each river reach examined in this study. Cumulative survival
probabilities, as estimated from release to each downstream detection array, are also presented.
3.4.1 Wild Hanford Reach Fall Chinook Salmon
The probability of acoustic-tagged wild fall Chinook salmon surviving migration through the lower
half of the Hanford Reach (from release at rkm 595 to CR552) was estimated to be 0.824 (SE = 0.027)
and the probability of surviving from release to McNary Dam was 0.497 (0.036) in 2014 (Figure 3.9).
Survival probability from release to the most downstream array, located in the reservoir of Bonneville
Dam at rkm 275 (CR275), was 0.278 (0.032).
3.10
Figure 3.9. Overall cumulative survival probability estimates for acoustic-tagged wild Hanford Reach
fall Chinook salmon from release in the Hanford Reach (rkm 595) to downstream acoustic
telemetry receiver arrays. Error bars denote standard errors.
Survival of acoustic-tagged wild Hanford Reach fall Chinook salmon varied among reaches, from
0.824 (SE=0.027) between release and CR552 to 1.00 (multiple reaches; Table 3.3). Because reaches
differed in length, survival is better compared among reaches using Skm estimates.
3.11
Table 3.3. Reach-specific survival probability estimates (S, and associated SE) for acoustic-tagged wild
Hanford Reach fall Chinook salmon juveniles through each river reach studied in 2014 from
release at rkm 595 to CR275. Survival-per-kilometer (Skm) estimates are also shown.
Reach
S (SE)
Skm
Release to CR552
0.824 (0.027)
0.9937
CR552 to CR524
0.847 (0.028)
0.9942
CR524 to CR498
0.855 (0.030)
0.9943
CR498 to CR489
0.950 (0.020)
0.9936
CR489 to CR480
0.928 (0.025)
0.9919
CR480 to CR472
0.971 (0.017)
0.9962
CR472 to CR470
0.973 (0.017)
0.9865
CR470 to CR455
0.926 (0.027)
0.9952
CR455 to CR449
1.000 (0.003)
1.0000
CR449 to CR439
0.928 (0.028)
0.9924
CR439 to CR422
0.864 (0.038)
0.9917
CR422 to CR412
0.986 (0.014)
0.9983
CR412 to CR381
0.945 (0.027)
0.9983
CR381 to CR351
0.971 (0.021)
0.9991
CR351 to CR349
1.000 (0.004)
1.0000
CR349 to CR325
0.909 (0.036)
0.9961
CR325 to CR311
1.000 (0.004)
1.0000
CR311 to CR275
0.917 (0.036)
0.9976
Upstream of McNary Dam, Skm was considerably lower in the immediate forebay of McNary Dam
(Skm = 0.9865; CR472 to CR470) compared to all other reaches (Figure 3.10). The other reach upstream
of McNary with a Skm estimate that was notably low was also near McNary Dam between CR489 and
CR480 (Skm = 0.9919). Anomalously, the reach located between these two reaches (CR480 to CR472)
had the highest Skm of all reaches upstream of McNary Dam for acoustic-tagged wild Hanford Reach fall
Chinook salmon. Survival-per-kilometer estimates were generally similar among all reaches located
between release and CR489, ranging from 0.9936 to 0.9943.
Downstream of McNary Reservoir, two reaches had Skm estimates that were considerably lower than
all others for acoustic-tagged wild Hanford Reach fall Chinook salmon. These included the reach located
between Boardman, OR (CR439) and Crow Butte (CR422; Skm = 0.9917) and the next upstream reach,
located between Paterson, WA (CR449) and Boardman, OR (CR439; Skm = 0.9924).
3.12
Figure 3.10. Survival probability-per-kilometer estimates for acoustic-tagged wild Hanford Reach fall
Chinook salmon through reaches of the Columbia River, 2014. Dashed vertical lines
indicate the locations of McNary (MCN), John Day (JDA), and The Dalles (TDA) dams.
We observed a significant, positive relationship between the probability of survival to McNary Dam
and fish length for wild Hanford Reach fall Chinook salmon (χ = 7.486; p = 0.006; Figure 3.11). The
difference in survival was rather large across the length range of tagged fish. Those at the upper end of
the length distribution (100 mm FL) were about twice as likely to survive to McNary Dam as fish at the
lower end of the distribution (80 mm FL).
3.13
Figure 3.11. Covariate analysis results displaying nonparametric (black line) and modeled (blue line)
survival probabilities of acoustic-tagged wild Hanford Reach fall Chinook salmon from
release in the Hanford Reach (rkm 595) to McNary Dam (rkm 470) in relation to fork length.
The frequency histogram displays the number of tagged fish in each 1-mm fork length bin.
3.4.2 Priest Rapids Hatchery Fall Chinook Salmon
The probability of acoustic-tagged PRH fall Chinook salmon surviving migration through the
Hanford Reach (from CR633 to CR552) was estimated to be 0.659 (SE = 0.037) and the probability of
surviving to McNary Dam was 0.498 (0.039) in 2014 (Figure 3.12). Survival probability from CR633 to
the most downstream array, located in the reservoir of Bonneville Dam at rkm 275 (CR275), was 0.281
(0.035).
3.14
Figure 3.12. Overall cumulative survival probability estimates for acoustic-tagged Priest Rapids Hatchery
fall Chinook salmon from acoustic detection in the PRH outflow channel (CR633) to
downstream acoustic telemetry receiver arrays. Error bars denote standard errors.
Survival of acoustic-tagged PRH fall Chinook salmon varied widely among reaches, from 0.659
(SE=0.037) between CR633 and CR552 to 1.00 (multiple reaches; Table 3.4). Because reaches differed
in length, survival is better compared among reaches using Skm estimates.
3.15
Table 3.4. Reach-specific survival probability estimates (S, and associated SE) for acoustic-tagged Priest
Rapids Hatchery fall Chinook salmon juveniles through each river reach studied in 2014 from
virtual release (detection in the hatchery outflow channel) at rkm 633 to CR275. Survival
from tagging to virtual release (Release to CR633) and survival-per-kilometer (Skm) estimates
are also shown.
Reach
S (SE)
Skm
Release to CR633
0.821 (0.027)
N/A
CR633 to CR552
0.659 (0.037)
0.9951
CR552 to CR524
0.898 (0.029)
0.9962
CR524 to CR498
0.897 (0.031)
0.9960
CR498 to CR489
0.977 (0.016)
0.9971
CR489 to CR480
0.988 (0.012)
0.9987
CR480 to CR472
1.000 (0.003)
1.0000
CR472 to CR470
0.970 (0.021)
0.9850
CR470 to CR455
0.896 (0.035)
0.9932
CR455 to CR449
1.000 (0.003)
1.0000
CR449 to CR439
1.002 (0.004)
1.0002
CR439 to CR422
0.875 (0.039)
0.9924
CR422 to CR412
0.969 (0.022)
0.9961
CR412 to CR381
0.967 (0.023)
0.9990
CR381 to CR351
0.934 (0.032)
0.9978
CR351 to CR349
0.982 (0.018)
0.9857
CR349 to CR325
0.928 (0.035)
0.9970
CR325 to CR311
1.000 (0.005)
1.0000
CR311 to CR275
0.902 (0.042)
0.9971
Similar to the results observed for wild tagged fish, Skm of PRH fall Chinook salmon was
considerably lower in the immediate forebay of McNary Dam (Skm = 0.9850; CR472 to CR470) compared
to all other reaches upstream of McNary Dam (Figure 3.13). With the exception of this reach, Skm
generally increased from upstream to downstream between CR633 and CR472 for acoustic-tagged PRH
fall Chinook salmon.
Downstream of McNary Reservoir, the Skm of PRH fall Chinook salmon was considerably lower in
the immediate forebay of John Day Dam (Skm = 0.9857; CR351 to CR349) than all other reaches. The
reach that included McNary Dam (CR470 to CR455) and the reach located between Boardman, OR and
Crow Butte (CR439 to CR422) also had relatively low Skm estimates for PRH fall Chinook salmon
(0.9932 and 0.9924, respectively).
3.16
Figure 3.13. Survival probability-per-kilometer estimates for acoustic-tagged Priest Rapids Hatchery fall
Chinook salmon through reaches of the Columbia River, 2014. Dashed vertical lines
indicate the locations of McNary (MCN), John Day (JDA), and The Dalles (TDA) dams.
Similar to the relationship found for wild Hanford Reach fall Chinook salmon, we observed an even
stronger, positive relationship between survival probability to McNary Dam and fish length for PRH fall
Chinook salmon (χ = 14.164; p < 0.001; Figure 3.14). Again, there was a large difference in survival
across the length range of tagged fish. Those at the upper end of the length distribution (~100 mm FL)
were more than twice as likely to survive to McNary Dam as fish at the lower end of the distribution (80
mm FL).
3.17
Figure 3.14. Covariate analysis results displaying nonparametric (black line) and modeled (blue line)
survival probabilities of acoustic-tagged Priest Rapids Hatchery fall Chinook salmon from
Priest Rapids Hatchery (CR633) to McNary Dam (CR470) in relation to fork length. The
frequency histogram displays the number of tagged fish in each 1-mm fork length bin.
3.5 Tra v e l Time and Travel Rate
The amount of time fish spend in a particular river reach and the speed at which they travel is often
linked to survival probability. This section describes the travel times and rates of acoustic-tagged wild
Hanford Reach and PRH fall Chinook salmon through reaches of the mid and lower Columbia River in
2014.
3.5.1 Wild Hanford Reach Fall Chinook Salmon
The median travel time was less than 2 days for acoustic-tagged wild Hanford Reach fall Chinook
salmon in each river reach examined in 2014 (Figure 3.15). We observed relatively little variability in
travel times within each reach except in the first reach (release to CR552) where the median travel time
was 1.4 days but over 25% of the fish took >6 d and 25% took <13 h to traverse the reach. The median
travel time of wild Hanford Reach fall Chinook salmon detected at McNary Dam was 10.7 d (25th
percentile = 6.7 d; 75th percentile = 16.2 d).
3.18
Figure 3.15. Travel time (days) of acoustic-tagged wild Hanford Reach fall Chinook salmon juveniles in
each reach of the Columbia River studied in 2014. Solid lines within the boxes are median,
the box boundary represents the 25th and 75th percentiles, whiskers indicate the 10th and 90th
percentiles, and dots indicate the 5th and 95th percentiles. Dashed vertical lines indicate the
locations of McNary (MCN), John Day (JDA), and The Dalles (TDA) dams.
Acoustic-tagged wild Hanford Reach fall Chinook salmon generally migrated most quickly through
the free-flowing Hanford Reach (release to CR552), and through the tailraces of Federal Columbia River
Power System (FCRPS) dams (CR470 to CR455; CR349 to CR325; CR311 to CR275; Figure 3.16). We
also observed the greatest variability in travel rate within these reaches. For example, wild Hanford
Reach fall Chinook salmon had a median travel rate of 30 km/d from release to CR552; however, 25% of
the fish had travel rates <10 km/d and 25% had rates >80 km/d. Conversely, travel rates were slowest,
with the least amount of variability in reservoir reaches (all reaches between CR552 and CR470, between
CR449 and CR349, and from CR325 to CR311). For example, median travel rates were generally around
10 km/d for acoustic-tagged wild Hanford Reach fall Chinook salmon in reaches of McNary Reservoir
(part of CR552 to CR524, and all reaches between CR524 and CR470).
3.19
Figure 3.16. Travel rate (km/d) of acoustic-tagged wild Hanford Reach fall Chinook salmon juveniles in
each reach of the Columbia River studied in 2014. Solid lines within the boxes are median,
the box boundary represents the 25th and 75th percentiles, whiskers indicate the 10th and 90th
percentiles, and dots indicate the 5th and 95th percentiles. Dashed vertical lines indicate the
locations of McNary (MCN), John Day (JDA), and The Dalles (TDA) dams.
3.5.2 Priest Rapids Hatchery Fall Chinook Salmon
Similar to the trends observed for wild Hanford Reach fall Chinook salmon, acoustic-tagged fall
Chinook salmon from PRH migrated through most river reaches in less than 2 d (Figure 3.17). The one
exception was the Hanford Reach (release at PRH to CR552) where the median travel time was 3.7 d.
PRH fall Chinook salmon had a longer travel time through the Hanford Reach than wild fish because they
had a longer distance to travel to CR552 (81 km versus 43 km). Also similar to the trend observed for
wild fish, we found the variability in travel times was greatest for acoustic-tagged PRH fall Chinook
salmon in the Hanford Reach where 25% of the fish had travel times <1.5 d and 25% of the fish took >7.0
d to migrate through the reach. The median travel time of PRH fall Chinook salmon detected at McNary
Dam was 11.6 d (25th percentile = 9.1 d; 75th percentile = 14.1 d).
3.20
Figure 3.17. Travel time (days) of acoustic-tagged Priest Rapids Hatchery fall Chinook salmon juveniles
in each reach of the Columbia River studied in 2014. Solid lines within the boxes are
median, the box boundary represents the 25th and 75th percentiles, whiskers indicate the 10th
and 90th percentiles, and dots indicate the 5th and 95th percentiles. Dashed vertical lines
indicate the locations of McNary (MCN), John Day (JDA), and The Dalles (TDA) dams.
Similar to the trends observed for wild Hanford Reach fall Chinook salmon, acoustic-tagged PRH fall
Chinook salmon migrated most quickly, with the greatest variability, through flowing reaches,
particularly those downstream from FCRPS dams (Figure 3.18). Again, the slowest travel rates were
observed in McNary Reservoir where median travel rates were around 15 km/d. PRH fall Chinook
salmon had higher median travel rates than wild fall Chinook salmon through all reaches examined in
2014, except in the two most upstream reaches (release to CR552 and CR552 to CR524).
3.21
Figure 3.18. Travel rate (km/d) of acoustic-tagged Priest Rapids Hatchery fall Chinook salmon juveniles
in each reach of the Columbia River studied in 2014. Solid lines within the boxes are
median, the box boundary represents the 25th and 75th percentiles, whiskers indicate the 10th
and 90th percentiles, and dots indicate the 5th and 95th percentiles. Dashed vertical lines
indicate the locations of McNary (MCN), John Day (JDA), and The Dalles (TDA) dams.
4.1
4.0 Discussion
This study was the first to attempt to partition mortality of wild Hanford Reach and PRH fall Chinook
salmon into specific river reaches to identify potential sources of mortality. We identified river reaches in
which survival was low, relative to the length of the reach. These data, combined with existing
knowledge from previous studies, provided us with the information necessary to make inferences about
the causes of the observed mortality.
We found groups of acoustic-tagged wild Hanford Reach and PRH fall Chinook salmon had a 0.50
probability of surviving to McNary Dam. Whereas this estimate is considerably higher than has been
previously found for wild Hanford Reach fall Chinook salmon juveniles, it is substantially lower than
what is typical for PRH smolts.
Survival of wild Hanford Reach fall Chinook salmon juveniles to McNary Dam has been estimated
since 1995 from annual releases of ~3,000 to ~23,000 PIT-tagged fish (Fish Passage Center 2013).
Survival of these groups to McNary Dam has ranged from 0.27 to 0.62 with an average survival
probability of 0.37 (SE = 0.02). Similarly, the 9,940 wild Hanford Reach fall Chinook salmon juveniles
that were implanted with PIT tags (PIT only) and released in 2014 had a survival probability of 0.34 (SE
= 0.02) to McNary Dam. The large discrepancy between survival estimates derived from acoustic-tagged
versus PIT-only groups is likely a result of the difference in fish size between groups. For comparison,
PIT-only fish that measured <80 mm FL had a 0.31 (SE = 0.02) survival probability from release to
McNary Dam in 2014 compared to 0.72 (SE = 0.12) for PIT-only fish that measured ≥80 mm FL. As
previously mentioned, we attempted to minimize the effect of the transmitter on the performance of
implanted fish by only tagging fish that measured ≥80 mm FL; whereas, fish as small as 60 mm FL were
implanted with PIT tags. As we demonstrated, survival of these fish is strongly, positively correlated
with fish length. Therefore, we expect that the survival of the overall population of juvenile wild Hanford
Reach fall Chinook salmon through the study area was substantially lower than it was for the fish we
tagged. However, we believe that the relative losses of tagged fish by reach were representative of the
overall population.
Survival of PRH fall Chinook salmon juveniles to McNary Dam has been estimated since 1997 from
annual releases of PIT-tagged fish (Richards et al. 2013). Survival of these groups to McNary Dam has
ranged from 0.50 to 0.84 with an average of 0.68 (SE = 0.02). In 2014, the 31,980 PRH fall Chinook
salmon juveniles that were implanted with PIT tags (PIT-only) had a 0.66 (SE = 0.02) probability of
surviving to McNary Dam. The difference in survival between groups of acoustic-tagged and PIT-only
PRH fall Chinook salmon juveniles observed in 2014 may have been the result of a reduction in
performance of acoustic-tagged fish caused by the tagging procedure or presence of the tag, and/or a
result of acoustic transmitter failure or loss.
A laboratory study was conducted at PNNL in 2013 to determine the minimum size fish that could be
implanted with the downsized acoustic transmitter without affecting fish performance or survival.
Results from this study found only 1 of 126 (0.8%) fall Chinook salmon (80–104 mm FL) surgically
implanted (no suture; same method as used in this study) with a PIT tag and downsized acoustic
transmitter died over a 60-day examination period and no fish dropped either tag during the study. Based
4.2
on the results of this study, we felt confident in using this method during a field trial. However, we
observed relatively high post-tagging, pre-release mortality for the group of PRH fall Chinook salmon we
implanted with acoustic transmitters for the in-river survival evaluation described in this report.
Acoustic-tagged PRH fall Chinook salmon juveniles had an estimated probability of surviving from
tagging to acoustic detection in the outfall channel of 0.82 (SE = 0.03). Although several great blue
herons Ardea herodias were frequently observed foraging in the outfall channel, it is unlikely heron
predation accounted for all the mortality we observed in the channel pond and outflow channel since we
did not observe the same level of mortality for the PIT-only group. The group of 31,980 PIT-only PRH
fall Chinook salmon, which were implanted on May 29 (the day after acoustic tagging), had an estimated
survival probability of 0.97 (SE < 0.01) from tagging to PIT detection in the outfall channel. Thus, it
appears the acoustic-tagged group may have suffered some tag- or tagging-related mortality.
It is also apparent that some level of acoustic tag loss or failure occurred between tagging and
volitional release to the river for the PRH group. Of the 167 PRH juveniles implanted with acoustic
transmitters and PIT tags that were detected by the PIT array in the outflow channel, only 159 (95.2%)
were also detected by the cabled acoustic array located in the outflow channel. Because the acoustic array
in the outflow channel had a detection probability of 1.0, these results suggest an acoustic tag loss or
failure rate of 4.8% occurred between tagging and detection in the outflow channel. The first tag in the
tag-life study that died did so after 47.6 days, with over 75% of the tags lasting >100 days. Therefore, it
is likely tag loss and not tag failure accounted for the 5% non-detection rate observed during the first
couple of weeks between tagging and detection in the outflow channel.
Because we estimated survival of acoustic-tagged PRH juveniles by forming a virtual release of only
those fish detected by the cabled acoustic array located in the outflow channel (CR633), fish that died or
dropped their tag prior to volitional release into the river were not included in the estimate. However, it is
possible that some tag- or tagging-related mortality continued to occur once fish left the PRH outflow
channel and entered the Columbia River. We also have evidence that tag loss continued after fish entered
the river. Twenty-one acoustic-tagged hatchery fall Chinook salmon that had an active transmitter when
they left PRH (i.e., they were detected at CR633) were detected by the PIT array in the JBS of McNary
Dam. Of those, two (9.5%) were not detected by any adjacent acoustic receiver arrays, suggesting the
fish were still alive but no longer had acoustic transmitters. Two of 27 (7.4%) acoustic-tagged wild
Hanford Reach fall Chinook salmon that were detected by the McNary Dam JBS PIT array appeared to
have dropped their acoustic tags (i.e., they were not detected by adjacent acoustic arrays). Three of 23
(13.0%) acoustic-tagged fish detected by the PIT array in the JBS of John Day Dam were not detected by
adjacent acoustic receiver arrays.
Existing evidence suggests that fish routed through the JBS at hydroelectric dams of the FCRPS may
be smaller or weaker, on average, than fish that pass the dams using other routes (Zabel et al. 2005). Fish
that expelled their transmitter may be expected to have complications that could potentially inhibit their
performance, making them more likely to pass through the JBS at FCRPS dams. Thus, the tag loss
percentages presented above may be biased high and represent an absolute worst-case scenario.
However, even at these rates, the effect of tag loss on survival estimation is relatively small. For
example, 101 of the 198 (51.0%) acoustic-tagged wild Hanford Reach fall Chinook salmon were detected
at CR472. Because this array had a detection probability of 1.0, the probability of survival from release
4.3
to CR472 is 0.51 (SE = 0.04). If we assume 7.4% of the fish that were not detected at CR472 were living
fish that had expelled their transmitter, the survival estimate becomes 0.55, which is within the 95%
confidence interval of the original estimate.
The greatest bias associated with the survival estimates for the group of wild Hanford Reach fall
Chinook salmon may be the presence of a tag or tagging effect, which we would expect to manifest itself
soon after implantation, as we observed for the PRH fish. Because wild Hanford Reach fall Chinook
salmon were released just 24-h after tagging, they were not afforded the time to exhibit the tag or tagging
effect prior to release. Thus, survival of the wild Hanford Reach group was likely underestimated in
reaches located near the release site if they exhibited a tag or tagging effect similar to that experienced by
the PRH group.
Reach survival of wild Hanford Reach fall Chinook salmon, estimated on a per-kilometer basis, was
lower in all reaches located between release (rkm 595) and CR422 compared to those located downstream
of CR422. We observed relatively low and similar estimates of Skm among the three most upstream
reaches we studied. As mentioned previously, the presence of a tag or tagging effect may have
contributed to relatively low survival of acoustic-tagged wild fall Chinook salmon in the Hanford Reach
between release and CR552. However, the potential for predation from piscivorous birds and fishes
exists within the Hanford Reach.
Each spring (May and June), the Oregon Department of Fish and Wildlife (ODFW) conducts
electrofishing surveys for predators in the Columbia River. The focus of the electrofishing effort is to
capture and tag as many pikeminnow as possible for estimation of sport-reward fishery exploitation rates.
Therefore, capture priorities have focused on northern pikeminnow with other predators (particularly
smallmouth bass, walleye, and channel catfish) sampled less consistently. However, these data provide
empirical information of the distribution of piscivorous fish predators in the Columbia River.
Electrofishing catches indicate northern pikeminnow and walleye are more abundant in the Hanford
Reach than in McNary Reservoir (Peter McHugh, [ODFW], unpublished data; Figure 4.1). Using
recoveries of marked fish at the sport reward stations and the Cormack-Jolly-Seber model for open
populations (Seber 1982; Hayes et al. 2007), we estimated the annual (2001–2009) population abundance
for northern pikeminnow ≥228 mm FL that inhabit the Columbia River between the mouth of the Yakima
River and Priest Rapids Dam. Excluding two years that were obvious outliers due to low numbers of
recaptures, population abundance averaged 37,392 (SE = 6,843) northern pikeminnow.
Northern pikeminnow have been identified as a major predator of juvenile salmonids in the Columbia
River (Poe et al. 1991; Rieman et al. 1991; Vigg et al. 1991; Zimmerman 1999). Poe et al. (1991) and
Zimmerman (1999) estimated juvenile salmonids accounted for 67% and >84%, respectively, of northern
pikeminnow diets in reservoirs of the Columbia River. Although to a lesser extent, these same studies
also identified walleye as a predator of juvenile salmonids. For example, Poe et al. (1991) found juvenile
salmonids made up 14% of the diet of walleye. The presence of these predators has the potential to
reduce survival of upriver bright fall Chinook salmon juveniles migrating through the lower Hanford
Reach.
4.4
Figure 4.1. Total numbers of northern pikeminnow (NPM), smallmouth bass (SMB), and walleye
(WAL) captured during Oregon Department of Fish and Wildlife electrofishing surveys
conducted annually from 1993–2010 between McNary and Priest Rapids dams.
Upriver bright fall Chinook salmon are also susceptible to predation from Caspian terns Hydroprogne
caspia that nest on Goose Island on Potholes Reservoir, which is located about 33 km north-northeast of
the Hanford Reach. GPS-tagged terns from this colony have been recorded making foraging trips to the
Hanford Reach. Avian predation rates, estimated as the proportion of PIT tags recovered (i.e., detected
by mobile PIT antennas) on Goose Island that were previously detected by the PIT array at Rock Island
Dam, averaged 0.2% for this colony on upper Columbia River summer/fall Chinook salmon between
2009–2012 (Roby et al. 2013). In 2014, the nesting colony consisted of 340 breeding pairs (Bird
Research Northwest 2014).
Acoustic-tagged wild Hanford Reach fall Chinook salmon also experienced relatively low survival in
the reach located between CR552 and CR524. Results from ODFW electrofishing surveys reveal an
abundance of both northern pikeminnow and smallmouth bass within this reach (Figure 4.1). This reach
contains the mouth of the Yakima River, which has been identified as a major spawning tributary for
Columbia River smallmouth bass. From 1998 to 2001, Fritts and Pearsons (2004) observed an increase in
the abundance of smallmouth bass >150 mm in the Yakima River from an annual average of about 3,000
bass in mid-March to almost 20,000 bass in mid-June. The authors attributed the increase primarily to
immigration of fish from the Columbia River and estimated that an average of just over 200,000
4.5
salmonids, most of which were fall Chinook salmon, were consumed annually by smallmouth bass in the
Yakima River during the spring. It is likely that high rates of smallmouth predation on fall Chinook
salmon occur in the Columbia River during this time as well. A study conducted by Tabor et al. (1993) in
a 6-km stretch of the Columbia River near Richland, WA found juvenile salmonids, primarily subyearling
fall Chinook salmon, made up 59% of smallmouth bass diet by weight. The authors attributed the high
predation rates to the abundance of subyearling fall Chinook salmon juveniles of suitable forage size
emigrating from the Hanford Reach and the overlap of habitats of the two species. Others have identified
the vulnerability of wild subyearling fall Chinook salmon juveniles to predation by smallmouth bass due
to habitat overlap in low velocity nearshore areas (Curet 1993) and the small size of wild fall Chinook
salmon juveniles at the time of emigration (Zimmerman 1999).
Although difficult to quantify, the Yakima River also seems to contain a rather sizeable population of
channel catfish, which appear capable of consuming large numbers of juvenile salmonids (Pearsons et al.
2001). A naturally reproducing population of channel catfish also inhabits the Columbia River where
they have been found to consume large numbers of juvenile salmonids (Poe et al. 1991). The presence of
large populations of predatory fish, combined with the reduction in water particle travel rate as the river
transitions from free-flowing to reservoir-influenced, makes juvenile fall Chinook salmon vulnerable to
predation within this reach (CR552 to CR524).
The risk of avian predation in this reach (CR552 to CR524) remains relatively unknown. Large
nesting colonies of California gulls Larus californicus and ring-billed gulls Larus delawarensis inhabit
Island 20 near the town of Richland, Washington at rkm 545 (Figure 2.1). In 2014, 12,500 nesting gulls
were observed on the island (Bird Research Northwest 2014), which has only ever been partially scanned
for PIT tags (Roby et al. 2013). Thus, reliable predation rate estimates do not exist for these colonies.
However, diet analyses of gulls from colonies located upstream of McNary Dam indicated these birds
consume very small amounts of salmonids (Roby et al. 2013).
The next downstream reach, CR524 to CR498, contains the mouth of the Snake River, a large
backwater slough, several islands that host colonies of piscivorous birds, and the mouth of the Walla
Walla River. The ODFW electrofishing survey data indicate the abundance of northern pikeminnow and
walleye are relatively low in this reach. However, walleye are frequently the target of recreational fishers
in this section of the Columbia River, suggesting they are present. Electrofishing catches indicate a rather
sizeable smallmouth bass population is present in this reach as well (Figure 3.18). In addition, the Snake
and Walla Walla rivers are two of the few rivers in Washington that contain naturally reproducing
populations of channel catfish (Lower Columbia Fish Recovery Board 2004). Thus, there is no shortage
of piscivorous fishes in this reach of the Columbia River that may contribute to the below-average
survival estimate we observed for acoustic-tagged wild Hanford Reach fall Chinook salmon.
As mentioned, the reach located between CR524 and CR498 is also home to several nesting colonies
of piscivorous waterbirds. These include populations of double-crested cormorants Phalacrocorax
auritus on Foundation Island, American white pelicans Pelecanus erythrorhynchos on Badger Island, and
California gulls and ring-billed gulls and Caspian terns on Crescent Island (Evans et al. 2012). Bird
Research Northwest conducted waterbird surveys of the islands during the spring and summer of 2014
and counted 390 nesting pairs of double-crested cormorants on Foundation Island, 273 American white
pelicans on Badger Island, and 395 nesting pairs of Caspian terns and 6,200 California gulls on Crescent
4.6
Island (Bird Research Northwest 2013). Several other bird species, including great blue herons, great
egrets Ardea alba, black-crowned night-herons Nycticorax nycticorax, and ring-billed gulls, were
frequently observed on the islands in relatively small numbers.
The outmigration timing of upriver bright fall Chinook salmon coincides with the chick rearing
period (May and June) for the majority of birds on these colonies. Thus, juvenile fall Chinook salmon
from the Hanford Reach are migrating through this reach during the period of highest energy demand for
these predatory birds. Roby et al. (2012) found salmonids accounted for almost 70% of tern prey items at
the Crescent Island colony over a 12-year period between 2000 and 2011, representing an average of
about 500,000 salmonids consumed annually. However, this estimate includes steelhead, coho, sockeye,
spring Chinook, and Snake River fall Chinook in addition to URB fall Chinook salmon. During the
period of URB fall Chinook salmon outmigration, salmonids, which would be primarily fall Chinook
salmon at this time, still make up about 60–70% of the Crescent Island tern diet (Roby et al. 2013). We
observed a negative relationship between survival to McNary Dam as estimated for PIT-tagged wild
Hanford Reach fall Chinook salmon and the number of Caspian tern breeding pairs counted on colonies
of the Columbia Plateau (primarily Crescent and Goose islands; Figure 4.2). However, the relationship
was not significant (p = 0.210; R2 = 0.248) but should continue to be evaluated into the future to
determine whether a significant trend develops. It is unlikely cormorants of the Foundation Island colony
substantially affect survival rates of URB fall Chinook salmon in McNary Reservoir. Roby et al. (2013)
found salmonids accounted for only 10% of the prey biomass in the diet of Foundation Island cormorants
during the outmigration period of URB fall Chinook salmon juveniles.
Figure 4.2. Relationship between annual survival probability of PIT-tagged wild Hanford Reach fall
Chinook salmon and the number of Caspian tern breeding pairs counted on colonies of the
Columbia Plateau (2005, 2007–2013).
4.7
Although the estimated number of smolts consumed by the Crescent Island tern colony is relatively
large, it may not represent a significant percentage of the population of salmonid smolts that migrate past
these islands. Avian predation rates, estimated as the proportion of tags recovered (i.e., detected by
mobile PIT antennas) on the islands that were previously detected by PIT arrays at upstream dams, have
been consistently low for subyearling fall Chinook salmon juveniles at these colonies. In a study to
estimate avian predation rates on Endangered Species Act-listed salmonid evolutionary significant units
of the Columbia River basin between 2007 and 2010, Evans et al. (2012) found that all colonies in this
reach combined to consume an annual average of 1.6% of the Snake River fall Chinook salmon that were
last detected at Lower Monumental Dam. Although this should be viewed as a minimum estimate due to
the large distance between the colonies and Lower Monumental Dam (76 km) and uncertainty regarding
the off-colony deposition of tags (Roby et al. 2013), it indicates the actual predation rate on juvenile
Snake River fall Chinook salmon may be quite low. We would expect the predation rate of URB fall
Chinook salmon to be similarly low.
The reaches with the lowest Skm estimates were those located near McNary Dam, with the lowest
being observed in the immediate forebay. An evaluation of predation by resident piscivorous fish on
juvenile salmonids between McNary and John Day dams revealed predation was most intense in areas
near the dams (Poe et al. 1988). The authors attributed this finding to the delay and disorientation of
salmonids associated with dam passage and the increased densities of piscivorous fish species in slack
water areas near dams. Indeed, we observed the slowest travel rates of acoustic-tagged fish in reaches of
McNary Reservoir, indicating their migration was slowed by presence of the dam, thereby subjecting
them to predation for a longer period of time. Electrofishing catches indicate the forebay of McNary Dam
may contain a rather sizeable smallmouth bass population (Figure 4.1).
In addition to attracting predaceous fishes, feeding aggregations of piscivorous waterbirds are also
frequently observed near dams of the Columbia River. In addition to terns and cormorants, even gulls
find success preying on salmonid smolts near dams of the Snake and Columbia rivers. Low survival of
acoustic-tagged juvenile salmonids in the tailrace of McNary Dam in 2012 was attributed to high rates of
avian predation by ring-billed gulls (Hughes et al. 2013). Juvenile salmonids, disoriented after dam
passage, are particularly susceptible to avian predation in the immediate tailrace of FCRPS dams
(Williams 2006). For example, gull predation rates of 6% and 11% were observed in the tailrace of The
Dalles Dam for radio-tagged subyearling and yearling Chinook salmon, respectively (Collis et al. 2002).
High rates of avian predation at FCRPS dams has led to bird hazing and installation of wires stretched
across the river to discourage birds from entering the tailrace.
The tailrace of McNary Dam has also been identified as an area of high salmonid predation by
piscivorous fish. Poe et al. (1991) found that about 80% of northern pikeminnow and 60% of channel
catfish diets (by weight) were composed of juvenile salmonids in the immediate tailrace of McNary Dam.
Salmonids made up a smaller percentage of the diets of walleye (~15%) and smallmouth bass (<5%) in
McNary tailrace. Rieman et al. (1991) estimated an average of 2.7 million juvenile salmonids were lost
annually (for the period 1983–1986) to predation by piscivorous fish (northern pikeminnow, walleye,
smallmouth bass) between McNary and John Day dams, which represented about 9% to 19% of all
salmonids that entered the reach. Much of the loss (21%) was estimated to have occurred in the
immediate tailrace of McNary Dam where northern pikeminnow and channel catfish were abundant (Poe
et al. 1991; Rieman et al. 1991). Thus, the reported estimates would likely have been higher had
4.8
predation by channel catfish been included. Of the species that were included, northern pikeminnow
accounted for 78% of the total salmonid loss, walleyes accounted for 13%, and smallmouth bass for 9%.
However, the contribution of walleyes and smallmouth bass to the total mortality increased in July and
August when mortality rates were highest and the majority of salmonids consumed were subyearling fall
Chinook salmon.
Although Rieman et al. (1991) observed very high predation rates in the immediate tailrace of
McNary Dam, predation in the main body of John Day Reservoir represented the majority (79%) of the
total salmonid loss to piscivorous fish. The authors observed relatively low consumption rates by
northern pikeminnow in the main body of the reservoir but emphasized the effect a low consumption rate
can have when the abundance of predators is high, as appears to be the case in John Day Reservoir.
Rieman et al. (1991) estimated there to be 85,000 northern pikeminnow and 10,000 walleyes >250 mm
and 35,000 smallmouth bass >200 mm in the reservoir.
We observed low survival of acoustic-tagged wild Hanford Reach fall Chinook salmon between
CR449 and CR422. This reach contains Paterson Slough on the Washington shore, McCormack Slough
on the Oregon shore, a backwater area near Crow Butte, and many miles of heavily rip-rapped shorelines.
The three embayments (Paterson, McCormack, and Crow Butte), which cover about 1,700 acres (U.S.
Army Corps of Engineers 1995), have been identified as flow refugia and potential spawning areas for
nonnative piscivorous fish species (Nigro et al. 1985). Smallmouth bass and walleye are frequently
targeted by anglers in the area around Paterson Slough, McCormack Slough, and the Blalock Islands,
suggesting increased densities of these predators in those areas.
In addition to providing habitat to nonnative predaceous fishes, several islands in this reach, including
the Blalock Islands, are home to nesting colonies of multiple avian predators, including California and
ring-billed gulls and Caspian and Forster’s terns. Surveys conducted by Bird Research Northwest during
the spring and summer of 2014, revealed colonies of 199 terns (both Caspian and Forster’s terns) and
4,630 gulls (both California and ring-billed) on the island complex (Bird Research Northwest 2014).
Other birds, such as great egrets, black-crowned night-herons, great blue herons, and American white
pelicans were also observed on the island in smaller numbers. Minimum predation rates of Blalock
Island-nesting terns on Snake River fall Chinook salmon have been historically quite low, average <0.1%
from 2007–2010 (Evans et al. 2012). Again, we would expect the predation rate on URB fall Chinook
salmon to be similarly low. Predation rates from the Blalock Island complex gull colonies have not been
estimated to our knowledge.
Relative survival (Skm) was high for acoustic-tagged wild Hanford Reach fall Chinook salmon from
CR422 down to John Day Dam (CR349) before dipping slightly in the reaches that included passage
through John Day and The Dalles dams and their tailraces (CR349 to CR325 and CR311 to CR275). The
reach between CR349 and CR325 is home to a nesting colony of California gulls on Miller Rocks Island
that numbered 3,100 individuals in 2014 (Bird Research Northwest 2014). Evans et al. (2012) estimated
the average annual minimum predation rate of the Miller Rocks Island gulls to be 0.4% of the Snake
River fall Chinook salmon that passed McNary Dam. The rate is likely similarly low for URB fall
Chinook salmon juveniles.
4.9
Much of the mortality in the tailraces may be attributed to predation by resident fish, which is known
to be a substantial source of mortality in dam tailraces and outfall locations (Lower Columbia River Fish
Recovery Board 2004). The tailrace of John Day Dam has been identified as an area with relatively high
densities of walleye (Porter 2009). The Dalles Dam tailrace has a complex basin with a series of
downriver islands where predators reside, is relatively shallow with armored bedrock substrate, has an
adjacent slough-like habitat on the south side of the river, and riprap-lined banks. Petersen et al. (2001)
found relatively high numbers of smallmouth bass compared to northern pikeminnow in The Dalles Dam
tailrace. The authors estimated 1,000 to 2,000 smallmouth bass were present in the immediate tailrace of
The Dalles Dam, although this estimate was based on relatively few marked and recaptured fish.
Acoustic-tagged fall Chinook salmon from PRH survived at a higher rate than the wild group in most
reaches, particularly those located upstream of McNary Dam. The lower survival of wild Hanford Reach
fall Chinook salmon upstream of McNary Dam may have been a result of a tagging effect. As mentioned
previously, the group of acoustic-tagged PRH fish suffered high mortality, likely as a result of a tagging
effect during the two-week period between tagging and release. The wild group was released 24 hours
after tagging and therefore suffered any potential tagging effect in-river. Trends in Skm were generally
similar between groups of acoustic-tagged wild Hanford Reach and PRH fall Chinook salmon. The
primary differences included higher Skm rates for PRH fish in the forebay of McNary Dam between
CR498 and CR472 and lower Skm for PRH fish in the immediate forebay of John Day Dam.
Data from this study and others indicate much of the mortality incurred by URB fall Chinook salmon
juveniles between Priest Rapids and Bonneville dams can likely be attributed to predation from resident
piscivorous fish. We observed no significant relationship between the survival of PIT-tagged wild
Hanford Reach fall Chinook salmon to McNary Dam and the size of the primary avian predator nesting
colonies located in McNary Reservoir. We also did not observe mortality “hot spots” in the reaches of
the Columbia River that contain the largest colonies of predaceous waterbirds. Instead, we observed
relatively consistent mortality rates between release and CR422, which is more indicative of predation
from piscivorous fish, which are more widely distributed than avian predators. Additionally, it is likely
we “missed” much of the predation by piscivorous fish (thereby overestimating reach survivals) due to
the relatively large size of fish we were able to implant with acoustic transmitters. Avian predators, on
the other hand, appear to target larger individuals, as evidenced by their high predation rates on steelhead
smolts (Collis et al. 2001; Antolos et al. 2005); thus, it is unlikely we would have “missed” any mortality
“hot spots” due to avian predation. In addition, results of studies conducted to assess avian predation
rates have consistently estimated very low predation rates on subyearling fall Chinook salmon upstream
of Bonneville Dam (<2%; Evans et al. 2012; Roby et al. 2013). Alternatively, predation rates estimated
for piscivorous fish suggest they may be consuming 17% of the juvenile salmon that enter John Day
Reservoir during June, July, and August, when most salmon smolts entering the reservoir are subyearling
fall Chinook salmon (Rieman et al. 1991). Harnish et al. (2013) estimated about 43 million URB fall
Chinook salmon presmolts were produced annually in the Hanford Reach between BY 1984–2004.
Assuming a survival probability of 0.37 to McNary Dam (as estimated from annual releases of PIT-only
wild URB fall Chinook salmon in the Hanford Reach), about 16 million Hanford Reach URB juveniles
enter John Day Reservoir annually. Thus, if piscivorous fish consume 17% of the population, an
estimated 2.7 million URB fall Chinook salmon juveniles would be consumed annually in John Day
4.10
Reservoir. If predation rates are of similar magnitude in other reservoirs, predation by resident
piscivorous fish is clearly an important source of mortality.
The high rate of salmonid smolt predation observed by Rieman et al. (1991) for resident piscivorous
fish in John Day Reservoir led to development of the Northern Pikeminnow Management Program
(NPMP) in 1990–1991. The NPMP consists of a “sport-reward” fishery, which offers public anglers a
monetary incentive to catch northern pikeminnow, and “dam-angling”, whereby agency personnel are
hired to angle for northern pikeminnow at FCRPS dams. The program was founded on modeling
simulations that indicated a 10–20% exploitation rate on predator-sized northern pikeminnow would
reduce predation on juvenile salmonids by 50% (Rieman and Beamesderfer 1990). The program has
appeared effective at reducing the abundance of northern pikeminnow. The catch-per-unit-effort and
abundance index data have shown a continued and persistent decrease in the number of northern
pikeminnow ≥250 mm in the Snake and Columbia rivers since the NPMP was implemented (Gardner et
al. 2013; Barr et al. 2014).
Removal of northern pikeminnow will only improve survival of migrating juvenile salmonids if a
compensatory response by other predatory fishes does not offset the net benefit of removal. Although an
increase in the proportion of smallmouth bass diets containing juvenile salmonids has not been observed
from smallmouth bass captured annually during electrofishing and dam-angling efforts of the NPMP,
smallmouth bass abundance and predation index values have increased in recent years in some areas of
Snake and Columbia river reservoirs (Gardner et al. 2013; Barr et al. 2014). As noted by Carey et al.
(2011), smallmouth bass have become a large component of the fish community of the Snake and
Columbia rivers, largely due to the habitat created by human modifications (e.g., dams) of the landscape.
Juvenile salmonids continue to be a common item in the diets of Columbia River walleyes, which have
also shown an increase in abundance index in areas of John Day and The Dalles reservoirs (Gardner et al.
2013). Increases in the abundance index of these predators may be an early indication of a compensatory
response to the removal of northern pikeminnow from the system (Gardner et al. 2013; Barr et al. 2014).
If indeed a compensatory response develops, the NPMP may need to be expanded to include other
predatory species, such as smallmouth bass and walleye to achieve the same benefit to salmonid survival.
Whereas smallmouth bass and walleye represent a potential significant threat to the survival of salmonid
smolts in the Snake and Columbia rivers, options to manage these species are complicated because
fisheries agencies are simultaneously charged with enhancing fishing opportunities and controlling
predators of threatened and endangered salmon (Carey et al. 2011). However, if salmon survival and
conservation is to be prioritized, there is a clear need to identify and test potential management options
aimed at reducing predation from resident piscivorous fishes.
Altering dam operations is another potential management option that has been used successfully in
the past to improve survival of smolts through the FCRPS. For example, increases in the amount and
percentage of water that is routed through the spillways at dams has been attributed to increased survival
of salmonid smolts in the Snake and Columbia rivers (e.g., Adams et al. 2012). It may be possible to
manage reservoir levels in such a way as to disrupt the spawning activities or recruitment success of
predaceous fish species. Several studies have demonstrated that fluctuations in discharge can negatively
affect the reproductive success of smallmouth bass by flooding nests with cooler water, depositing silt,
4.11
driving away adult bass guarding nests, exposing eggs to desiccation, or stranding emerged fry
(Henderson and Foster 1957; Becker et al. 1981; Lukas and Orth 1995). A study of factors that influence
smallmouth bass production in the Hanford Reach indicated fluctuations in discharge from Priest Rapids
Dam reduced productivity (Montgomery et al. 1980). In order to be successful, disruptions to spawning
activities would need to occur throughout the major spawning areas for sufficient duration over multiple
years to cause year-class failures. Major spawning areas would need to be identified and a feasibility
study would be required to assess whether the operational flexibility exists at dams of the Columbia River
to implement the operations necessary to create the desired disruptions.
Our study confirmed that the loss rates of juvenile URB fall Chinook salmon from the Hanford Reach
were high in areas where habitat has been influenced by hydropower development and native and non-
native predatory fish species. Whereas our study had some limitations due to 1) the size of fish we were
able to tag, 2) the potential for a tag or tagging effect on fish performance, and 3) possible tag loss, we
believe that the relative loss rates are representative for the wild Hanford Reach and Priest Rapids
Hatchery portions of the URB stock. Most of the loss appears to be concentrated in the river/reservoir
transition area where large predator-rich tributaries enter as well as in the immediate dam forebays where
travel rates of outmigrating smolts are slowed. Additional work to document how the predation rates we
observed in the larger size classes of juvenile URB fall Chinook salmon relate to the overall population,
as well as efforts to determine the potential effectiveness of management actions intended to reduce the
populations and/or productivity of piscivorous fish species will provide the information necessary to
enable managers to design and implement strategies to improve the freshwater survival of this important
stock.
5.1
5.0 References
Adams, N.S., H.C. Hansel, R.W. Perry, and S.D. Evans. 2012. A multi-year analysis of spillway survival
for juvenile salmonids as a function of spill bay operations at McNary Dam, Washington and
Oregon, 2004–2009. U.S. Geological Survey open-file report 2012–1125, 68 p.
Antolos, M., D.D. Roby, and K. Collis. 2004. Breeding ecology of Caspian terns at colonies on the
Columbia Plateau. Northwest Science 78:303–312.
Antolos, M., D.D. Roby, D.E. Lyons, K. Collis, A.F. Evans, M. Hawbecker, and B.A. Ryan. 2005.
Caspian tern predation on juvenile salmonids in the mid-Columbia River. Transactions of the
American Fisheries Society 134:466–480.
Barr, C.M., E. Tinus, M. Gardner, C. Mallette, and A.J. Storch. 2014. System-wide predator control
program: fisheries and biological evaluation. Pages 46–93 in S. Williams, editor. Report on the
predation index, predator control fisheries, and program evaluation for the Columbia River basin
experimental northern pikeminnow management program. 2013 annual report. Prepared by
Pacific States Marine Fisheries Commission, Oregon Department of Fish and Wildlife, and
Washington Department of Fish and Wildlife for Bonneville Power Administration.
Bauersfeld, K. 1978. The effect of daily flow fluctuations on spawning fall Chinook in the Columbia
River. Technical Report No. 38. State of Washington, Department of Fisheries, Olympia,
Washington.
Becker, C.D., D.H. Fickeisen, and J.C. Montgomery. 1981. Assessment of impacts from water level
fluctuations on fish in the Hanford Reach, Columbia River. PNL-3813, Pacific Northwest
Laboratory, Richland, Washington.
Beeman, J.W., T.C. Robinson, P.V. Haner, S.P. Vanderkooi, and A.G. Maule. 1999. Gas bubble disease
monitoring and research of juvenile salmonids. U.S. Geological Survey, Biological Resources
Division, Columbia River Research Laboratory, Cook, Washington.
Bird Research Northwest. 2014. Research, monitoring, and evaluation of avian predation on salmonid
smolts in the lower and mid-Columbia River: 2012 draft annual report. Real Time Research,
Inc., Bend, Oregon. Available online at http://www.birdresearchnw.org.
Carey, M.P., B.L. Sanderson, T.A. Friesen, K.A. Barnas, and J.D. Olden. 2011. Smallmouth bass in the
Pacific Northwest: a threat to native species; a benefit for anglers. Reviews in Fisheries Science
19:305–315.
Chapman, D.W., D.E. Weitkamp, T.L. Welsh, M.B. Dell, and T.H. Schadt. 1986. Effects of river flow
on the distribution of Chinook salmon redds. Transactions of the American Fisheries Society
115:537–547.
Chen, H., S.S. Cartmell, Q. Wang, T.J. Lozano, Z. Deng, H. Li, X. Chen, Y. Yuan, M.E. Gross, T.J.
Carlson, and J. Xiao. 2014. Micro-battery development for Juvenile Salmon Acoustic Telemetry
System applications. Scientific Reports 4:Article Number 3790.
Collis, K.D., D. Roby, D.P. Craig, B.A. Ryan, and R.D. Ledgerwood. 2001. Colonial waterbird
predation on juvenile salmonids tagged with passive integrated transponders in the Columbia
River estuary: vulnerability of different salmonid species, stocks, and rearing types.
Transactions of the American Fisheries Socieity 130:385–396.
Collis, K., D.D. Roby, D.P. Craig, S. Adamany, J.Y. Adkins, and D.E. Lyons. 2002. Colony size and
diet composition of piscivorous waterbirds on the lower Columbia River: Implications for losses
of juvenile salmonids to avian predation. Transactions of the American Fisheries Society
131:537–550.
Conover, W.J. 1980. Practical nonparametric statistics. Wiley & Sons, New York.
Cormack, R.M. 1964. Estimates of survival from the sightings of marked animals. Biometrika 51:429–
438.
5.2
Curet, T.S. 1993. Habitat use, food habits and the influence of predation on subyearling Chinook salmon
in Lower Granite and Little Goose reservoirs, Washington. Master’s thesis. University of Idaho,
Moscow.
Dauble, D.D., and D.G. Watson. 1997. Status of fall Chinook salmon populations in the mid-Columbia
River, 1948–1992. North American Journal of Fisheries Management 17:283–300.
Deng, Z., M.A. Weiland, T.J. Carlson, and M.B. Eppard. 2010. Design and instrumentation of a
measurement and calibration system for an acoustic telemetry system. Sensors 10:3090–3099.
Deng, Z.D., M.A. Weiland, T. Fu, T.A. Seim, B.L. LaMarche, E.Y. Choi, T.J. Carlson, and M.B. Eppard.
2011. A cabled acoustic telemetry system for detecting and tracking juvenile salmon: Part 2.
Three-dimensional tracking and passage outcomes. Sensors 11:5661–5676.
Evans, A.F., N.J. Hostetter, D.D. Roby, K. Collis, D.E. Lyons, B.P. Sandford, R.D. Ledgerwood, and S.
Sebring. 2012. Systemwide evaluation of avian predation on juvenile salmonids from the
Columbia River based on recoveries of passive integrated transponder tags. Transactions of the
American Fisheries Society 141:975–989.
Fish Passage Center. 2013. “Update: passage timing, travel times, and juvenile survival of Hanford
Reach sub-yearling Chinook to McNary Dam (1993–2012).” Memorandum to Washington
Department of Fish and Wildlife.
Fritts, A.L., and T.N. Pearsons. 2004. Smallmouth bass predation on hatchery and wild salmonids in the
Yakima River, Washington. Transactions of the American Fisheries Society 133:880–895.
Gardner, M., E. Tinus, M. Hughes-Weaver, C. Mallette, and E.S. Van Dyke. 2013. System-wide
predator control program: indexing and fisheries evaluation. Pages 52–77 in R. Porter, editor.
Report on the predation index, predator control fisheries, and program evaluation for the
Columbia River basin experimental northern pikeminnow management program. 2012 annual
report. Prepared by Pacific States Marine Fisheries Commission, Oregon Department of Fish and
Wildlife, and Washington Department of Fish and Wildlife for Bonneville Power Administration.
Harnish, R.A., R. Sharma, G.A. McMichael, R.B. Langshaw, T.N. Pearsons, and D.A. Bernard. 2012.
Effect of Priest Rapids Dam operations on Hanford Reach fall Chinook salmon productivity and
estimation of maximum sustainable yield, 1975–2004. PNWD-4339, Battelle, Pacific Northwest
Division, Richland, Washington.
Harnish, R.A., R. Sharma, G.A. McMichael, R.B. Langshaw, and T.N. Pearsons. 2013. Effect of
hydroelectric dam operations on the freshwater productivity of a Columbia River fall Chinook
salmon population. Canadian Journal of Fisheries and Aquatic Sciences 71:602–615.
Hayes, D.B., J.R. Bence, T.J. Kwak, and B.E. Thompson. 2007. Abundance, biomass, and production.
Pages 327 – 374 in C.S. Guy and M.L. Brown, editors. Analysis and interpretation of freshwater
fisheries data. American Fisheries Society, Bethesda, Maryland.
Henderson, C., and R.F. Foster. 1957. Studies of smallmouth bass (Micropterus dolomieu) in the
Columbia River near Richland, Washington. Transactions of the American Fisheries Society
86:112–127.
Hughes, J.S., M.A. Weiland, C.M. Woodley, G.R. Ploskey, S.M. Carpenter, M.J. Hennen, E.F. Fischer,
G.W. Batten III, T.J. Carlson, A.W. Cushing, Z. Deng, D.J. Etherington, T. Fu, M.J. Greiner, M.
Ingraham, J. Kim, X. Li, J. Martinez, T.D. Mitchell, B. Rayamajhi, A. Seaburg, J.R. Skalski, R.L.
Townsend, K.A. Wagner, and S.A. Zimmerman. 2013. Survival and passage of yearling and
subyearling Chinook salmon and steelhead at McNary Dam, 2012. PNNL-22788. Draft report
submitted by the Pacific Northwest National Laboratory to the U.S. Army Corps of Engineers,
Walla Walla, Washington.
Ingraham, J.M., Z. Deng, J.J. Martinez, B.A. Trumbo, R.P. Mueller, and M.A. Weiland. 2014.
Feasibility of tracking fish with acoustic transmitters in the Ice Harbor Dam tailrace. Scientific
Reports 4:Article Number 4090.
Jolly, G.M. 1965. Explicit estimates from capture-recapture data with both death and immigration –
stochastic model. Biometrika 52:225–247.
5.3
Langness, O.P., and K.F. Reidinger. 2003. Escapement goals for upriver bright (URB) fall Chinook
salmon stocks of the Columbia River. Washington Department of Fish and Wildlife, Olympia.
Li, T., and J.J. Anderson. 2009. The vitality model: A way to understand population survival and
demographic heterogeneity. Theoretical Population Biology 76:118–131.
Lower Columbia River Fish Recovery Board. 2004. Lower Columbia salmon and steelhead recovery
and subbasin plan. Prepared for the Northwest Power and Conservation Council.
Lukas, J.A., and D.J. Orth. 1995. Factors affecting nesting success of smallmouth bass in a regulated
Virginia stream. Transactions of the American Fisheries Society 124:726–735.
McMichael, G.A., M.B. Eppard, T.J. Carlson, J.A. Carter, B.D. Ebberts, R.S. Brown, M. Weiland, G.R.
Ploskey, R.A. Harnish, and Z.D. Deng. 2010. The Juvenile Salmon Acoustic Telemetry System:
a new tool. Fisheries 35:9–22.
Montgomery, J.C., D.H. Fickeisen, and C.D. Becker. 1980. Factors influencing smallmouth bass
production in the Hanford area, Columbia River. Northwest Science 54:296–305.
Nigro, A.A., R.C. Beamesderfer, J.C. Elliott, M.P. Faler, L.M. Miller, and B.L. Uremovich. 1985.
Abundance and distribution of walleye, northern squawfish, and smallmouth bass in John Day
Reservoir, 1985. Annual report (Contract DE-AI79-84BP35097) to Bonneville Power
Administration, Portland, Oregon.
Oldenburg, E.W., B.J. Goodman, and G.A. McMichael. 2012. Forms of production loss during the early
life history of fall Chinook salmon in the Hanford Reach of the Columbia River. PNWD-4314,
Battelle, Pacific Northwest Division, Richland, Washington.
Pearsons, T.N., A.L. Fritts, K.D. Ham, G.A. McMichael, E.L. Bartrand, G.M. Temple, and C.W. Hopley.
2001. Yakima River species interactions. Washington Department of Fish and Wildlife,
Olympia.
Peters, C.N., D.R. Marmorek, and I. Parnell (eds.). 1999. PATH decision analysis report for Snake River
fall Chinook. ESSA Technologies Ltd., Vancouver, British Columbia, Canada.
Petersen, J.H., C. Barfoot, and M. Sheer. 2001. Predation by northern pikeminnow on juvenile
salmonids in The Dalles Dam tailrace: Field, laboratory, and habitat modeling studies (FY2000).
Report by U.S. Geological Survey to U.S. Army Corps of Engineers, Portland, Oregon.
Poe, T.P., H.C. Hansel, S. Vigg, D.E. Palmer, and L.A. Prendergast. 1988. Predation by northern
squawfish, walleye, smallmouth bass, and channel catfish in a mainstem Columbia River
reservoir: I. Feeding ecology during the salmonid smolt out-migration. Pages 13 – 55 in
Predation by resident fish on juvenile salmonids in John Day reservoir, 1983 – 1986: Volume I –
Final report of research. Report prepared for the Bonneville Power Administration, Portland,
Oregon.
Poe, T.P., H.C. Hansel, S. Vigg, D.E. Palmer, and L.A. Predergast. 1991. Feeding of predaceous fishes
on out-migrating juvenile salmonids in John Day Reservoir, Columbia River. Transactions of the
American Fisheries Society 120:405–420.
Porter, R. 2009. Report on the predation index, predator control fisheries, and program evaluation for the
Columbia River basin experimental northern pikeminnow management program. Prepared by
Pacific States Marine Fisheries Commission in cooperation with Oregon Department of Fish and
Wildlife for Bonneville Power Administration.
Richards, S.P., P.A. Hoffarth, and T.N. Pearsons. 2013. Priest Rapids Hatchery monitoring and
evaluation annual report for 2012–2013. Public Utility District No. 2 of Grant County,
Washington.
Rieman, B.E., and R.C. Beamesderfer. 1990. Dynamics of a northern squawfish population and the
potential to reduce predation on juvenile salmonids in a Columbia River reservoir. North
American Journal of Fisheries Management 10:228–241.
Rieman, B.E., R.C. Beamesderfer, S. Vigg, and T.P. Poe. 1991. Estimated loss of juvenile salmonids to
predation by northern squawfish, walleyes, and smallmouth bass in John Day reservoir, Columbia
River. Transactions of the American Fisheries Society 120:448–458.
5.4
Roby, D.D., K. Collis, D.E. Lyons, J.Y. Adkins, Y. Suzuki, P. Loschl, T. Lawes, K. Bixler, A. Peck-
Richardson, E. Dykstra, J. Harms, W. Mashburn, J. Tennyson, N. Ventolini, A. Evans, B.
Cramer, M. Hawbecker, N. Hostetter, R.D. Ledgerwood, and S. Sebring. 2012. Research,
monitoring, and evaluation of avian predation on salmonid smolts of the lower and mid-Columbia
River. Final 2011 annaual report. Prepared for the Bonneville Power Administration and the
U.S. Army Corps of Engineers.
Seber, G.A.F. 1965. A note on the multiple recapture census. Biometrika 52:249–259.
Seber, G.A.F. 1982. The estimation of animal abundance and related parameters. Edward Arnold,
London.
Skalski, J.R., A. Hoffman, and S.G. Smith. 1993. Testing the significance of individual- and cohort-level
covariates in animal survival studies. Pages 9–28 in J.D. Lebreton and P.M. North, editors. The
use of marked individuals in the study of bird population dynamics: Models, methods, and
software. Birkhauser Verlag, Basel, Switzerland.
Skalski, J.R., S.G. Smith, R.N. Iwamoto, J.G. Williams, and A. Hoffman. 1998. Use of PIT-tags to
estimate survival of migrating juvenile salmonids in the Snake and Columbia rivers. Canadian
Journal of Fisheries and Aquatic Sciences 55:1484–1493.
Smith, S.G., J.R. Skalski, W. Wchlechte, A. Hoffman, and V. Cassen. 1994. SURPH.1 manual:
Statistical survival analysis for fish and wildlife tagging studies. Bonneville Power
Administration, Portland, Oregon.
Summerfelt, R.C., and L.S. Smith. 1990. Anesthesia, surgery, and related techniques. Pages 213–272 in
C. B. Schreck and P.B.Moyle, editors. Methods for fish biology. American Fisheries Society,
Bethesda, Maryland.
Tabor, R.A., R.S. Shively, and T.P. Poe. 1993. Predation on juvenile salmonids by smallmouth bass and
northern squawfish in the Columbia River near Richland, Washington. North American Journal
of Fisheries Management 13:831–838.
Titzler, P.S., G.A. McMichael, and J.A. Carter. 2010. Autonomous acoustic receiver deployment and
mooring techniques for use in large rivers and estuaries. North American Journal of Fisheries
Management 30:853–859.
Townsend, R.L., J.R. Skalski, P. Dillingham, and T.W. Steig. 2006. Correcting bias in survival
estimation resulting from tag failure in acoustic and radio telemetry studies. Journal of
Agricultural, Biological, and Environmental Statistics 11:183–196.
U.S. Army Corps of Engineers. 1995. Columbia River system operation review: Final environmental
impact statement. Portland, Oregon.
Vigg, S., T.P. Poe, L.A. Predergast, and H.C. Hansel. 1991. Rates of consumption of juvenile salmonids
and alternative prey fish by northern squawfish, walleyes, smallmouth bass, and channel catfish
in John Day Reservoir, Columbia River. Transactions of the American Fisheries Society
120:421–438.
Weiland, M.A., Z.D. Deng, T.A. Seim, B.L. LaMarche, E.Y. Choi, T. Fu, T.J. Carlson, A.I. Thronas, and
M.B. Eppard. 2011. A cabled acoustic telemetry system for detecting and tracking juvenile
salmon: Part 1. Engineering design and instrumentation. Sensors 11:5645–5660.
Whidden, S.M. 1996. The Hanford Reach: protecting the Columbia’s last safe haven for salmon.
Environmental Law 26:265–297.
Williams, R.N. 2006. Return to the river: Restoring salmon to the Columbia River. Elsevier Academic
Press.
Zabel, R.W., T. Wagner, J.L. Congleton, S.G. Smith, and J.G. Williams. 2005. Survival and selection of
migrating salmon from capture-recapture models with individual traits. Ecological Applications
15:1427–1439.
Zimmerman, M.P. 1999. Food habitats of smallmouth bass, walleyes, and northern pikeminnow in the
lower Columbia River basin during outmigration of juvenile anadromous salmonids.
Transactions of the American Fisheries Society 128:1036–1054.
PNNL-SA-23719
Distribution
No. of No. of
Copies Copies
Distr.1
# Name
Organization
Address
City, State and ZIP Code
# Organization
Address
City, State and ZIP Code
Name
Name
Name
Name
Name (#)
# Name
Organization
Address
City, State and ZIP Code
# Foreign Distribution
# Name
Organization
Address
Address line 2
COUNTRY
# Local Distribution
Pacific Northwest National Laboratory
Name Mailstop
Name Mailstop
Name Mailstop
Name Mailstop
Name (PDF)
