may 16 and 17th, 2017 at the quality inn, clarkston, wa. fall chinook... · may 16 and 17th, 2017...

Snake River Fall Chinook Symposium May 16 and 17th, 2017 at the Quality Inn, Clarkston, WA.

Tuesday May 16th- Full Day

8:30am – Welcome - Scope and Purpose of Symposium

Presenter: Rod Engle – U.S. Fish and Wildlife Service LSRCP

9:00am - Overview of fall Chinook hatchery program and ongoing natural monitoring

Presenter: Joe Bumgarner – Washington Department of Fish and Wildlife.

9:45am - Update on fall Chinook recovery planning and goals.

Presenter: Patty Dornbusch – National Marine Fisheries Service

10:15am - Break

10:30am - Brief presentation followed by Q&A discussion

• Total escapement upstream of Lower Granite Dam (LGR). • Abundance of natural-origin returns (NOR). • Abundance of hatchery-origin returns. • Percentage of the hatchery-origin fish in the naturally spawning population (pHOS) at

the population scale. • Percentage of the natural-origin fish used in hatchery broodstocks (pNOB). • Genetic diversity and effective population size of natural and hatchery-origin population

segments. • Age-at-return for natural and hatchery-origin fish.

Presenter: Bill Young – Nez Perce Tribe

11:15am - Brief presentation followed by Q&A discussion

• Relative performance of various hatchery life stage at release and release type strategies.

Hatchery-origin harvest levels in mainstem and terminal area fisheries.

Presenter: Stuart Rosenberger – Idaho Power

12:00pm – Lunch (Provided to registered attendees)

1:00pm – Brief presentation followed by Q&A discussion

• Spawning distribution.

Presenter: Billy Connor – U.S. Fish and Wildlife Service

2:00pm - Brief presentation followed by Q&A discussion

• Previous Yearling Fidelity (Garcia et al. summary) • Fidelity of subyearling hatchery production groups to release site areas. • Spawning distribution upstream and downstream of LGR of Lyons Ferry Hatchery on-

station released fish (yearling and subyearlings).

Presenter: Peter Cleary – Nez Perce Tribe

2:45pm - Break


• Productivity of the naturally spawning population -The juvenile abundance component of the Snake River basin fall Chinook salmon life Cycle Model.

Presenters: Russel Perry – U.S. Geological Survey, Tom Cooney – National Marine Fisheries Service

3:30pm – Brief presentation followed by Q&A discussion.

• Relationship of juvenile fish growth and survival to food availability and abundance. • Smallmouth Bass Predation Study

.Presenters: Ken Tiffan and John Erhardt – U.S. Geological Survey Biological Resources Division and Billy Connor – U.S. Fish and Wildlife Service


• Distinguishing natural- and hatchery origin individuals using multiple chemical signatures in otoliths.

• Determining natal rearing locations and ocean entry timing of untagged adult fish based on otolith microchemistry.

Presenters: Jens Hegg and Brian Kennedy - University of Idaho

4:30pm – Adjourn for the day.

Wednesday May 17th – Half Day

8:30am – Brief presentation followed by Q&A discussion

• Lessons learned from juvenile survival studies through hydrosystem (COE Consensus Study).

• Heritability of age-at-emigration.

Presenters: Jay Hesse – Nez Perce Tribe and Dean Holocek – U.S. Army Corps of Engineers. Robin Waples – National Marine Fisheries Service.


• Fall Chinook Recovery Planning and the Natural Production Emphasis Area

Presenter – Tom Cooney National Marine Fisheries Service

10:30am – Break


• Management assumption status assessment and update. • Identification and update of critical uncertainties.

Presenter: Jay Hesse – Nez Perce Tribe

11:30pm – “Parking Lot” Discussion (If needed)

12:00pm – Adjourn Symposium

Snake River Fall Chinook Salmon Symposium - Scope and Purpose

Rod Engle – U.S. Fish and Wildlife Service, Lower Snake River Compensation Plan Office

In 2012, the National Marine Fisheries Service (NMFS) issued two Endangered Species Act (ESA) Section 10 Permits (16607 and 16615) specific to operation, monitoring and evaluation of four ESA-listed Snake River fall Chinook artificial production programs. These interconnected and highly coordinated programs are:

1) Lower Snake River Compensation Plan Program (LSRCP) of fall Chinook at Lyons Ferry Hatchery

2) The fall Chinook Acclimation Program (FCAP) 3) The Idaho Power Company Program (Lyons Ferry, and Irrigon Hatchery facilities), 4) The Nez Perce Tribal Hatchery Complex.

Snake River fall Chinook salmon production is a highly coordinated effort involving multiple facilities, complex logistics and co-manager coordination. Brood stock to complete these hatchery programs are primarily collected at the Lower Granite Dam trap facility, with additional collections from Lyons Ferry or Nez Perce Tribal hatcheries as needed. Early rearing occurs at the previously mentioned hatcheries with juvenile releases at 10 different locations across the Snake River, Clearwater and Grande Ronde Rivers. Both subyearling (4,600,000) and yearling smolts (900,000) release groups are representatively marked with annual releases occurring through both direct stream releases (1,400,000) and acclimation sites (4,100,000). Monitoring and evaluation of these programs covers the entire Snake Basin distribution of both hatchery and natural returns and involves Federal, State, Tribal, University as well as private entities. Artificial production of Snake River fall Chinook salmon meet several mitigation, compensation, conservation and harvest objectives as well as legal mandates. The hatchery releases associated with these four Snake River fall Chinook salmon programs are identified within the U.S. v. Oregon 2008-2017 Management Agreement Production Tables and have consensus agreement by the Tribal, Non-Tribal and Federal Parties.

The Symposium is a special condition of both Section 10 permits issued by NMFS for program operations and will communicate new technical information and inform adaptive management decisions associated with the Snake River fall Chinook salmon population. Summary documents and presentations are grouped by topics identified within the Research, Monitoring and Evaluation sections of the Section 10 permits and specifically identified topics within Hatchery and Genetic Management Plans and other submitted materials for ESA consultation on the programs.

Relevant Literature or Information

US v. Oregon. 2009. 2008-2017 US v. Oregon Management Agreement. Portland, Oregon.

Available: https://www.fws.gov/pacific/fisheries/hatcheryreview/Reports/snakeriver/SR--079.revised.2008-17USvOR_Mngmt_Agrmt.pdf

Nez Perce Tribe. 2011. Snake River stock fall Chinook – Nez Perce Tribal Hatchery. Hatchery

and Genetic Management Plan. Available: http://www.westcoast.fisheries.noaa.gov/publications/hatchery/decisions/hgmp/hgmp-srfc-bia.pdf

Washington Department of Fish and Wildlife. 2011. Snake River stock fall Chinook – Lyons

Ferry Hatchery, Fall Chinook Acclimation Program, and Idaho Power Company. Hatchery and Genetic Management Plan. Available: https://www.fws.gov/lsnakecomplan/Reports/HGMPs/Snake%20River%20Fall%20Chinook.pdf

Nez Perce Tribe and Washington Department of Fish and Wildlife. 2011. Addendum to Snake

River Fall Chinook HGMPs for Lyons Ferry Hatchery, Fall Chinook Acclimation Project, Idaho Power Company, and Nez Perce Tribal Hatchery. Available: http://www.westcoast.fisheries.noaa.gov/publications/hatchery/decisions/hgmp/hgmp-srfc-add.pdf

Independent Scientific Review Panel. 2013. Review of the Lower Snake River Compensation

Plan Fall Chinook Salmon Program. Available: https://www.nwcouncil.org/fw/isrp/isrp2014-4

Lower Snake River Compensation Plan. 2013. Presentations and summary documents for the

Snake River Fall Chinook Program Review for the Independent Scientific Review Panel. Available: https://www.fws.gov/lsnakecomplan/Meetings/2013SnakeRiverFallChinookProgramReviewSymposium.html

https://www.fws.gov/pacific/fisheries/hatcheryreview/Reports/snakeriver/SR--079.revised.2008-17USvOR_Mngmt_Agrmt.pdf

https://www.fws.gov/pacific/fisheries/hatcheryreview/Reports/snakeriver/SR--079.revised.2008-17USvOR_Mngmt_Agrmt.pdf

http://www.westcoast.fisheries.noaa.gov/publications/hatchery/decisions/hgmp/hgmp-srfc-bia.pdf

http://www.westcoast.fisheries.noaa.gov/publications/hatchery/decisions/hgmp/hgmp-srfc-bia.pdf

https://www.fws.gov/lsnakecomplan/Reports/HGMPs/Snake%20River%20Fall%20Chinook.pdf

https://www.fws.gov/lsnakecomplan/Reports/HGMPs/Snake%20River%20Fall%20Chinook.pdf

http://www.westcoast.fisheries.noaa.gov/publications/hatchery/decisions/hgmp/hgmp-srfc-add.pdf

http://www.westcoast.fisheries.noaa.gov/publications/hatchery/decisions/hgmp/hgmp-srfc-add.pdf

https://www.nwcouncil.org/fw/isrp/isrp2014-4

https://www.fws.gov/lsnakecomplan/Meetings/2013SnakeRiverFallChinookProgramReviewSymposium.html

https://www.fws.gov/lsnakecomplan/Meetings/2013SnakeRiverFallChinookProgramReviewSymposium.html

Snake River Fall Chinook Hatchery Program and RME Overview

Jay Hesse, Billy Connor, Joe Bumgarner

Historical fall Chinook salmon abundance in the Snake River has been estimated at high as 500,000 pre-1940’s, but quickly declined over the following decades due to a number of anthropogenic and natural factors. Snake River fall Chinook historically utilized habitats in the mainstem Snake River Basin from the mouth to as far upstream as Shoshone Falls, in the Clearwater River Basin upstream to Selway Falls, in the Salmon River Basin upstream to the East Fork South Fork Salmon River, and in the lower reaches of many other smaller rivers/streams to these basins. Current distribution is limited to mainstem Snake River below Hells Canyon Dam, the Clearwater River basin, and the lower portions of other major tributaries (Salmon, Grande Ronde, Imnaha, and Tucannon). Contemporary abundance has increased significantly in recent years to around 50,000 (returns include both natural origin fish and hatchery mitigation fish), but remains well below historical levels. However, the role of Snake River fall Chinook continue to contribute to downriver and ocean fisheries, and are an important part of this program.

Currently, three hatchery mitigation programs are involved with Snake River fall Chinook production (Lower Snake River Compensation Plan, Idaho Power Company Hells Canyon Settlement Agreement, and Columbia Basin Fish and Wildlife Program). Production of the 5.5 million fall Chinook from these programs occur at three facilities (Lyons Ferry, Nez Perce Tribal, and Irrigon hatcheries). The hatchery program complex, with adult broodstock being collected at multiple locations (primarily at Lower Granite Dam), spawning occurring at two facilities, eggs and juveniles transferred to different hatcheries for rearing, fish are released at 10 locations, as both subyearling (4,600,000) and yearling smolts (900,000), with some direct released (1,400,000), while most are acclimated (4,100,000). In addition, research and monitoring needs and the associated marking and tagging schemes for all of these aspects require a great deal of coordination.

Over the years, the Snake River fall Chinook salmon hatchery program implementation and research/monitoring has been an integrated and collaborative effort by multiple agencies. Successful management of Snake River fall Chinook salmon has required changes to hatchery programs in response to scientific information on the abundance, productivity, distribution and diversity integrated with non-technical factors and risks, including but not limited to: legally authorized and mandated mitigation obligations, tribal treaty-reserved fishing rights, logistical

challenges and infrastructure constraints, and funding and operating budgets for implementing the changes and monitoring their effectiveness. Adaptive management actions to hatchery operations, hydro-system, habitat, and harvest facets have all occurred to some degree to improve returns of fall Chinook to the Snake River basin. Hydro-system operations have been change to include summer spill and transportation. Habitat has been improved via summer flow augmentation to reduce water temperatures, flow stabilization during and post spawning, and reduced power peaking during summer rearing periods. Harvest management has moved to an abundance based scale and utilizes both nonselective and selective techniques. Within the hatchery setting adaptive management has been responsible for changes to broodstock trapping sites, release locations, proportions of yearling/subyearling, selective mating protocols reducing the use of age three males and increasing the percentage natural-origin fish used in broodstock, added protective netting over rearing vessels, altered rearing vessel cleaning protocols, and representative marking. Adaptive management, based on recent and future research/monitoring, will be utilized to adaptively manage the hatchery production programs. Eight management objectives, and 28 associated management assumptions have been established for the Snake River fall Chinook hatchery programs; providing context for research, monitoring and evaluations. .

1) Natural production - Maintain and enhance natural production in supplemented and un-supplemented areas.

2) Life History - Maintain life history characteristics and genetic diversity in supplemented and un-supplemented areas.

3) Genetic Diversity - Operate hatchery programs so that life history characteristics and genetic diversity of hatchery fish mimic natural fish.

4) Non-Target Populations - Effects of hatchery programs on non-target (same species) populations remain within acceptable limits.

5) Fisheries - Restore and maintain treaty-reserved tribal and non-treaty fisheries. 6) Hatchery Operations - Operate hatchery programs to achieve optimal production

effectiveness while meeting priority management objectives for natural production enhancement, diversity, harvest, impacts to non-target populations.

7) Status and Trends - Understand the current status and trends of natural-origin population and their habitats.

8) Communications - Coordinate monitoring/evaluation and hatchery activities and communicate program findings to resource managers.

1 Snake River Hatchery Fall Chinook Salmon Abundance and Composition

Snake River Hatchery and Natural Fall Chinook Salmon Escapement and Population Composition above Lower

Granite Dam

William Young, Nez Perce Tribe Deborah Milks, Washington Department of Fish and Wildlife

Stuart Rosenberger, Idaho Power Company John Powel, PSMFC/IDFG

Matt Campbell, IDFG Daniel Hasselman, CRITFC

Shawn Narum, CRITFC

May, 2017


Introduction

This report provides results of the abundance and genetic diversity performance metrics

identified in the Addendum to Snake River fall Chinook salmon Hatchery Genetic Mangement Plan (HGMP) critical for understanding the effect of the Snake River fall Chinook hatchery program on Viable Salmonid Population (VSP) criteria. The addendum was developed in response to specific questions about abundance, productivity, spatial structure and diversity of the Snake River fall Chinook population, with detailed research, monitoring and evaluation studies identified to specifically address uncertainties identified at that time. Completion of this research directly addressed these uncertainties and provided information for adaptive management of Snake River fall Chinook salmon hatchery program. This information will also be critical for regional recovery planning, hatchery management, and harvest planning and management.

Metrics presented here represent measures of abundance and diversity at the population level for hatchery and natural Snake River fall Chinook salmon including 1) hatchery and natural abundance to Lower Granite Dam (LGR); 2) hatchery and natural escapement above LGR; 3) proportion of hatchery-origin spawners (pHOS) at the population level; 4) proportion of natural-origin spawners (pNOB) in the hatchery broodstock; 5) age composition at adult return and 6) genetic diversity.

Methods Performance metrics and a brief description of methods are presented in Table 1.

Table 1. Performance metrics, scale, method and summary description of methods for Snake River hatchery and natural fall Chinook for results presented in this document. LGR – Lower Granite Dam.

Metric Population/Scale Method Summary description of methods

Total abundance, adult and jack

Snake River Hatchery and Natural to LGR

Sampling at LGR and run reconstruction Young et al. 2012

Adult escapement Snake River Hatchery and Natural, escapement above LGR

Sampling at LGR and run reconstruction

Adult abundance to LGR minus fallback, brood stock removals, harvest and hatchery volunteers.

Adult proportion of hatchery-origin spawner (pHOS)

Snake River population above LGR

Adult escapement and run Reconstruction

Estimate hatchery fraction in the naturally-spawning population above LGR. Compare results, use PBT to validate long-term run reconstruction-based estimates

Adult escapement and Parentage-based tagging (PBT)

Proportion natural- Lyons Ferry Hatchery, Run Reconstruction Estimate the proportion


origin brood stock (pNOB)

Nez Perce Tribal Hatchery and total

and hatchery spawning records

of natural brood stock. Compare results, use PBT to validate long-term run reconstruction-based estimates

Parentage-based tagging (PBT) and hatchery spawning records

Age-at-return Snake River Hatchery and Natural populations at LGR

Sampling at LGR and run reconstruction

Use scale pattern analysis and coded wire tag recovery data to estimate age composition

Genetic Diversity Snake River Hatchery and Natural populations at LGR

Genetic diversity, effective population size (Ne)

Review of genetic diversity information

Results

Abundance to Lower Granite Dam Estimated abundance fall Chinook salmon returning to LGR (Figure 1) demonstrated

significant increases in abundance of hatchery and natural components beginning in the late 1990’s. Natural-origin Snake River fall Chinook salmon adults have increased from a low of approximately 100 fish in 1990 to an estimated 20,829 in 2013. Hatchery returns demonstrated similar trends, with a peak return of 46,642 adults in 2014.


Figure 1. Estimated hatchery (blue) and natural (red) escapement above Lower Granite Dam from 1975 - 2016.

Escapement above Lower Granite Dam

Total escapement of hatchery and natural fall Chinook salmon above LGR from 1994 - 2016 was estimated by subtracting fish removed for brood stock at the LGR trap from total escapement to LGR estimated using run reconstruction methods (Young et al. 2012), then applying an estimated or average fall-back rate determined using PIT tags (Figure 2). Recent ten-year geometric mean escapement totaled 19,369 hatchery and 7,782 natural adult fall Chinook salmon.

0

20,000

40,000

60,000

80,000

1975

1977

1979

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

2001

2003

2005

2007

2009

2011

2013

2015

Fall

Chi

nook

Sal

mon

Ret

urn

to L

ower

G

rani

te D

am

Year

Hatchery

Natural


199419

9519

9619

9719

9819

9920

0020

0120

0220

0320

0420

0520

0620

0720

0820

0920

1020

1120

1220

1320

1420

1520

16

Adu

lt Es

capm

ent A

bove

Lo

wer

Gra

nite

Dam

0

10000

20000

30000

40000

Figure 2. Estimated escapement of hatchery-origin (blue) and natural-origin (red) adult fall Chinook salmon above Lower Granite Dam. The blue and red hashed lines represents the ten-year geometric mean for hatchery (18,566) and natural (7,730) fish, respectively.

The adult proportion of hatchery-origin spawners (pHOS) in the spawning population

above LGR varied from a low of 2% in 1997 to a high of 81% in 2010 (Figure 3). From 1994 to 2016 the adult pHOS averaged 58%. The adult pHOS rate has been relatively consistent for the last ten years, averaging 70%.


199419

9519

9619

9719

9819

9920

0020

0120

0220

0320

0420

0520

0620

0720

0820

0920

1020

1120

1220

1320

1420

1520

16

Adu

lt Es

capm

ent A

bove

Lo

wer

Gra

nite

Dam

0

10000

20000

30000

40000

Prop

ortio

n of

Adu

lt H

atch

ery-

origi

n Sp

awne

rs (p

HO

S)

0.0

0.2

0.4

0.6

0.8

1.0

Figure 3. Estimated escapement of hatchery-origin (blue) and natural-origin (red) adult fall Chinook salmon and proportion of hatchery-origin adult spawners (black line) above Lower Granite Dam.

Complete brood year representation for PBT analysis was complete in 2016 enabling pHOS comparison using run reconstruction and PBT methods (Figure 4). Because PBT can accurately identify hatchery of origin, it didn’t rely on tag expansion and other estimation techniques required by the run reconstruction procedure. Consequently, PBT is assumed to be the “true” estimate of pHOS that will be used to verify the accuracy of the run reconstruction estimate. Results demonstrated that pHOS estimates generated by the two methods were nearly identical, varying by less than one-half a percentage point. Results from the first year were promising, however, additional PBT estimates of pHOS will be calculated over the next 5 years to validate and understand the error associated with the run reconstruction estimate.


Figure 4. Proportion of hatchery-origin spawners (pHOS) estimated using run reconstruction and parentage-based-tagging (PBT; top) and relative difference in estimates (bottom). Proportion of Natural-Origin Broodstock

Snake River fall Chinook broodstock were captured primarily at the LGR trap and hauled to Lyons Ferry Hatchery and Nez Perce Tribal Hatchery in order to better represent the run at large and promote the incorporation of natural-origin fish in the hatchery spawning population. The HGMP specified program goals specify a minimum pNOB of 0.1 with a target of 0.3. Estimated annual pNOB rates have varied across the years (Figure 5), averaging 22.3%. Estimates of pNOB were higher at NPTH, likely resulting from collection timing. NPTH collects a larger proportion of broodstock from early in the run compared to LFH.

0.754 0.748

0.831

0.769 0.763 0.753

0.743

0.849

0.774 0.766

0.700

0.750

0.800

0.850

0.900

Females Males Jacks<52 cm

All Males(male+jack)

All Fish

Prop

orito

n of

Hat

cher

y-or

igin

Sp

awne

rs (p

HOS)

PBT Run Reconstruction

-3.0%

-2.0%

-1.0%

0.0%

1.0%

Rela

tive

Diffe

renc

e in

pH

OS


Figure 5. Proportion of natural-origin broodstock from 2007 – 2016 for Snake River fall Chinook Salmon spawned at Lyons Ferry Hatchery and Nez Perce Tribal Hatchery.

Identification of origin of individual fish is not possible for the run reconstruction procedure. An estimated number of natural-origin fish is determined by subtracting the expanded number of unmarked/untagged hatchery fish (based on the tag proportion at release) from the entire pool of unmarked/untagged fish in the sample (Young et al. 2012). The resulting natural:hatchery ratio is a proportion that was used to calculate the number of natural fish in the pool of unmarked/untagged brood stock which is then used to estimate pNOB as shown in Figure 5. In contrast, PBT marking of hatchery fish accurately identifies origin of individual fish that is used in the calculation of an unbiased estimate of pNOB.

Results from 2016 demonstrated significantly higher pNOB estimated using PBT compared to the run reconstruction (Figure 6). Similar pHOS estimates generated by the two methods (Figure 4) suggested that the run reconstruction process of applying an estimated natural-origin proportion of works well for the randomly captured sample. However, brood stock selection is not random. Selecting the larger, older aged fish and fish trapped outside the systematic sampling design at the LGR trap may have biased the brood stock sample altered the results when applying the natural-origin proportion used to estimate pNOB. Although additional years comparing PBT and run reconstruction estimates of pNOB will be required to confirm these results, initial data suggests that the run reconstruction underestimates pNOB.

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

2007 2008 2009 2010 2011 2012 2013 2014 2015 2016

# na

tura

ls in

bro

odst

ock/

tota

l bro

odst

ock)

Spawn Year

WDFW pNOB NPTH pNOB Total pNOB


Figure 6. Proportion of natural-origin broodstock in 2016 estimated using run reconstruction and parentage-based-tagging (PBT). NPTH – Nez Perce Tribal Hatchery; LFH – Lyons Ferry Hatchery; all – all brood stock combined.

Age-at-Return

Snake River fall Chinook age-at-return was estimated from fish trapped at LGR using scale pattern analysis and CWT recoveries. Ocean age was used because of the different juvenile life histories exhibited by Snake River fall Chinook salmon, significant numbers of subyearling and yearling juveniles entering the ocean in different years, and large numbers of yearling hatchery fish released in the basin. Comparisons of hatchery and natural population ocean age compositions revealed fewer younger age and a greater proportion of older fish in the natural compared to the hatchery population (Figure 7).

0.36 0.37 0.36

0.28 0.28 0.28

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

NPTH LFH All

Prop

ortio

n of

Nat

ural

-orig

in

Bro

od S

tock

(pN

OB

)


Figure 7. Average age-at-return from 2003 – 2016 for hatchery (blue) and natural (red) Snake River fall Chinook Salmon (top) and relative difference by age. Oc – Ocean. Trends in ocean age composition from 2003 - 2016 revealed increasing proportions of ocean age 2, 3 and 4 in the hatchery population, with decreases in ocean age 0 (minijacks) and 1 (jacks). In the natural population, only ocean age 2 proportions increased, with slight decreases in the other age classes (Figure 8).

0.00

0.10

0.20

0.30

0.40

0.50

Oc Age 0 Oc Age 1 Oc Age 2 Oc Age 3 Oc Age 4

Prop

ortio

n

-80.0%

-40.0%

0.0%

40.0%

80.0%

Rel

ativ

e O

cean

A

ge D

iffer

ence


Figure 8. Trends in ocean age composition for hatchery (top) and natural (bottom) Snake River fall Chinook salmon returning to Lower Granite Dam from 2003 - 2016. Genetic Diversity

Relatively little information about the genetic diversity or structure of the Snake River fall Chinook salmon population has been collected in recent years. Previous studies demonstrated that the Snake River population, as represented by the Lyons Ferry Hatchery stock and natural juveniles captured in the Snake and Clearwater Rivers, demonstrated relatively high levels of genetic diversity (Marshall et al. 2011; Narum et al. 2007). Other studies revealed that the Snake River population was unique compared to Columbia River populations (Utter et al. 1982; Marshall et al. 2011) and ocean- and stream-type Chinook salmon populations from the Snake River were highly divergent (Narum et al. 2011). The more recent studies that are available revealed a lack of differentiation between Umatilla Hatchery (stray) and unmarked/untagged (presumably wild) fall chinook captured at Lower Granite Dam in 2002 and 2003 (Kassler et al. 2004). This suggested that the high presence of strays observed in the

0

0.1

0.2

0.3

0.4

0.5Hatchery, 2003 - 2016

0

0.1

0.2

0.3

0.4

0.5Natural, 2003 - 2016

Oc Age 0 Oc Age 1 Oc Age 2 Oc Age 3 Oc Age 4


hatchery broodstock during the mid- to late-1990’s may have influence the Snake River population. In a recent survey of environmental adaptation across the range of Chinook salmon, Hecht et al. (2015) observed little differentiation between a collection of Lyons Ferry Hatchery origin and Clearwater River natural origin fall Chinook salmon using 19,703 SNP loci.

A preliminary analysis of population structure between hatchery and natural fish identified by PBT from the 2016 LGR sample revealed no genetic differentiation (Fst = 0.0003) and supported the hypothesis of a single genetic cluster. With full PBT marking of hatchery fish a more comprehensive analysis of genetic diversity and population structure between Snake River hatchery and natural populations and among Columbia River stocks will be completed in the future.

Citations

Hecht, BC, AP Matala, JE Hess, and SR Narum. 2015. Environmental adaptation in Chinook salmon (Oncorhynchus tshawytscha) throughout their North American range. Molecular Ecology 24:5573-5595.

Kassler, TW., D. Milks and M. Schuck. 2004. A Microsatellite DNA Analysis of Snake River fall-run Chinook (2002 & 2003). Washington Department of Fish and Wildlife Report. https://www.researchgate.net/publication/268290549.

Narum, SR., WD Arnsberg, AJ Talbot and MS Powell. 2007 Reproductive Isolation Following Reintroduction of Chinook Salmon with Alternative Life Histories. Conserv. Genet.

Regional Mark Information System (RMIS). Regional Mark Processing Center, Pacific States Marine Fisheries Commission, 205 SE Spokane Street, Suite 100 Portland, Oregon 97202. http://www.rmpc.org/

Utter, FM., WJ Ebel, GB Milner and DJ Teel. 1982. Population Structures of Fall Chinook Salmon of the Mid-Columbia and Snake Rivers. Northwest and Alaska Fisheries Center, National Marine Fisheries Service, U.S. Department of Commerce. Report 82-10.

Young, W.P., S. Rosenberger and D. Milks. 2012. Snake River Fall Chinook Salmon Run Reconstruction at Lower Granite Dam; Methods for Retrospective Analysis. Nez Perce Tribe, Department of Fisheries Resources Management. http://www.nptfisheries.org/PublicationLibrary.aspx

https://www.researchgate.net/publication/268290549

http://www.rmpc.org/

Snake River Hatchery Fall Chinook Salmon

Age-at-Release Performance Evaluation

White Paper

Stuart Rosenberger, Idaho Power Company

William Young, Nez Perce Tribe

Deborah Milks, Washington Department of Fish and Wildlife

Bill Arnsberg, Nez Perce Tribe

Drew Wickard, Nez Perce Tribe

January, 2017

Recommended Citation:

Rosenberger, S., W. Young, D. Milks, B. Arnsberg and D. Wickard. 2017. Snake River Hatchery Fall

Chinook Salmon Age-at-Release Performance Evaluation White Paper.

http://www.nptfisheries.org/PublicationLibrary.aspx

1 Snake River Hatchery Fall Chinook Salmon Age-at-Release Performance Evaluation

Background

Hatchery production is an integral part of Snake River fall Chinook Salmon management,

with multiple parties involved. A highly coordinated monitoring and evaluation effort is in place

to generate information critical for future regional recovery planning, hatchery production,

harvest planning and management. Due to the differing management perspectives/priorities

among the managing entities of the Snake River fall Chinook hatchery program, this document is

structured as a white paper meant to inform decisions. Recommendations based upon the

presented data are being deferred to other management forums.

This white paper provides a summary of Snake River fall Chinook Salmon survival and

harvest, with an emphasis on the comparison among hatchery yearling and subyearling release

groups released downstream and upstream of Lower Granite Dam (LGR). Since being listed as

threatened under the Endangered Species Act (ESA) in 1992 Snake River fall Chinook Salmon

returning to LGR have made an astonishing rebound, increasing from less than a thousand fish in

the 1990’s to over seventy-five thousand fish in 2013 (Figure 1). The increased abundance of fall

Chinook Salmon in the Snake River Basin was achieved in large part by the use of a hatchery

program consisting of diverse release locations and rearing strategies. Releases began

downstream of LGR at Washington Department of Fish and Wildlife’s Lyons Ferry Hatchery in

1984, and have since expanded to include releases upstream of LGR from the Nez Perce Tribe’s

Fall Chinook Acclimation Program (FCAP) in 1996, an Idaho Power Company hatchery

program in 2001, and most recently, the development of the Nez Perce Tribe Hatchery Complex

in 2003.

Figure 1. Estimated returns of hatchery and natural Snake River fall Chinook Salmon to LGR

(1975 – 2015).

Beyond providing a boost in returns to LGR and to the natural spawning population,

Snake River fall Chinook Salmon hatchery programs have provided significant out-of-basin

0

10,000

20,000

30,000

40,000

50,000

60,000

70,000

80,000

Fall

Ch

inook

Salm

on

Ret

urn

to

Low

er G

ran

ite

Dam

Year

Hatchery

Natural


benefits. In recent years Snake River fall Chinook Salmon returns have provided a valuable

boost to tribal and non-tribal Ocean and Columbia River fisheries as well as emerging Snake

River sport and Tribal fisheries (Figure 2; Young et al. 2012; Regional Mark Information System)

Figure 2. Yearly estimate of the number of hatchery and natural Snake River Fall Chinook that

contribute to harvest and returns to LGR.

Snake River fall Chinook Salmon hatchery production, release sites, marking, tagging,

and egg distribution have been directed by the U.S. vs Oregon Management Agreement.

Production table B4B within the current 2008 – 2017 Agreement not only identifies and lists

release sites and strategies, but also prioritizes the hatchery programs should egg collections fall

short of that needed for the entire Snake River Basin hatchery program (Appendix Table A1).

This table is the basis for the metrics presented in this report.

This report provides multiple performance metrics for Snake River fall Chinook Salmon

that were released as juveniles into the Snake River Basin from 2006 through 2011. This report

compares performance of subyearling and yearling fall Chinook Salmon by release location and

rearing strategy (subyearling or yearling) for 1) smolt to adult return rates (SAR), 2) harvest

rates, 3) smolt to adult survival rates (SAS), 4) age composition at adult return, 5) size

composition at return, and 6) return timing at Bonneville Dam. These metrics provide

information that is critical for regional recovery planning, hatchery management, and harvest

planning and management.

The hatchery programs evaluated in this white paper are solely funded by the Bonneville

Power Administration, Idaho Power Company, and the Lower Snake River Compensation Plan.

0

25,000

50,000

75,000

100,000

125,000

150,000

175,000

2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015

Fall

Ch

ino

ok

Ha

rves

t a

nd

Ret

urn

Ab

un

da

nce

Year

Harvest, Natural

Harvest, Hatchery

Rtn to LGR, Natural

Rtn to LGR, Hatchery


Methods

In an attempt to present the data in a comparable and consistent manner, a series of

definitions/criteria were developed to promote the greatest amount of utility when comparing and

contrasting the different release sites and strategies. These definitions/criteria include:

1. Data presented are annual averages of fish harvested and fish returning to LGR from

release years 2006 – 2011, with the exception of Passive Integrated Transponder (PIT)

tag based adult run timing data which is a pooled average of 2011 – 2015 return years.

2. All comparisons made for Smolt to Adult Returns (SAR), Smolt to Adult Survival (SAS),

and Harvest were only done between groups that have both an Adipose Clip and a Coded

Wire Tag (ADCWT). A summary of release groups used in the comparison are presented

in Table 1. (Note: SAR calculations differ from Comparative Survival Study (CSS) PIT

tagged based methods which use number smolts passing Lower Granite Dam instead of

number of fish released).

3. Harvest data through return year 2014 is considered complete, however data from the

2015 return year may not be complete due to reporting deadlines stipulated by the

Regional Mark Information System (RMIS). Harvest data was downloaded from

RMPC.org on 12/15/2015.

4. Harvest data incorporates sport and commercial tribal and nontribal harvest in the ocean

and Columbia River. Due to the paucity of ADCWT recoveries (<1-2% of total

recoveries) in the Snake River and Columbia River tributary fisheries (i.e. Deschutes

River), harvest data from those two areas were excluded. Harvest data was also excluded

from sport and tribal fisheries conducted upstream of Lower Granite Dam.

5. Passage timing graphs incorporate all PIT tags associated with an individual release,

regardless of mark type.

6. Jack and Adult calculations were based on ocean age, not fork length. Jacks are defined

as one ocean age fish regardless of age at release (subyearling or yearling) or size.

Minijacks (0 ocean), averaging 27% of total yearling returns to LGR (unpublished data),

were excluded from all calculations. Minijacks are rarely observed in returns from

subyearling releases.

7. To accurately evaluate differences between yearling and subyearling release groups,

release years, rather than brood years, were compared in the analyses.

8. Due to an egg shortage in brood year 2006, eggs were not available for multiple releases

in 2007. To ensure all groups were compared equally, release year 2007 was excluded

from all analyses.

Methods and calculations specific to each metric are discussed in the results section of

this report. Where appropriate, additional tables and box plots that accompany the figures are

provided in an Appendix.

For the following analyses, multiple data sources were used and includes data

downloaded from the Pacific States Marine Fisheries Commission PTAGIS and RMIS databases,

as well as annual Snake River fall Chinook Salmon Lower Granite Dam run reconstruction

estimates (Young et al. 2012).


Table 1. Snake River fall Chinook Salmon AD/CWT release groups used in the analysis.

US v OR

Priority1 Age2 Release Group/Site

Juvenile

Rearing Site3

Release

Subbasin

Release

Strategy

1 1+ Lyons Ferry Onstation LFH Snake/Below LGR Acclimation

2 1+ Pittsburg Landing LFH Snake Acclimation

3 1+ Big Canyon LFH Clearwater Acclimation

4 1+ Captain John Rapids LFH Snake Acclimation

5 0+ Lyons Ferry Onstation LFH Snake/Below LGR Acclimation

6 0+ Captain John Rapids LFH Snake Acclimation

7 0+ Big Canyon LFH Clearwater Acclimation

8 &10 0+ Pittsburg Landing LFH Snake Acclimation

9 0+ Hells Canyon Dam/OFH OFH Snake Direct

11 0+ Couse Creek LFH Snake Direct

134 0+ Grande Ronde River LFH/IRR Grande Ronde Direct

15 0+ Hells Canyon Dam/Uma/IRR UFH/IRR Snake Direct

164 0+ Grande Ronde River LFH/IRR Grande Ronde Direct

17 0+ Hells Canyon Dam/Uma/IRR UFH/IRR Snake Direct 1 Release Groups 12 and 14 were associated with the transportation study and are not considered a long term release

strategy. 2Age at release: Yearling - 1+; Subyearling - 0+ 3Lyons Ferry Hatchery (LFH); Oxbow Fish Hatchery (OFH); Umatilla Fish Hatchery (UFH); Irrigon Fish Hatchery

(IRR) 4Grande Ronde River: Due to an egg shortage, an ADCWT group was not released in 2008.

Results

Smolt to Adult Return Rates

Smolt to adult return rates (SAR) were calculated as: (Total number of ADCWT fish

returning to LGR / Total number of ADCWT fish released) x 100. The number of fish returning

to LGR was calculated using current run reconstruction methods developed by Young et al.

(2012). Comparisons of SAR’s were made among release sites (Figure 3, Table A2, Figure A1)

and between yearling and subyearling release strategies (Figure 4, Table A3, Figure A1).


Figure 3. Average SAR and age composition of returning adults from Snake River fall Chinook

Salmon hatchery yearling (1+) and subyearling (0+) release groups (Release Years 2006, 2008-

2011).

Figure 4. Comparison of average SAR and age composition of returning adults by common release

site from Snake River fall Chinook Salmon yearling (1+) and subyearling (0+) hatchery releases

(Release Years 2006, 2008-2011).

0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40S

AR

Jacks

Adults

0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

SA

R

Jacks

Adults


Harvest Rates

Harvest data were obtained from the RMIS database and include harvest reported

primarily through 2014, although some 2015 data was reported. Harvest rates for each of the

hatchery release groups were calculated as: (Total number of estimated ADCWT fish harvested)

/ (Total number of ADCWT fish released) x 100. Harvest rates presented here includes all tribal

and non-tribal sport and commercial harvest in the ocean and Columbia River, but does not

include harvest of fish in the tributaries of the Columbia River and fish harvested in the Snake

River upstream and downstream of LGR due to their small contributions (<1-2%). Comparisons

of harvest rates were made between release sites (Figure 5, Table A4, Figure A2) and between

yearling and subyearling release strategies (Figure 6, Table A5, Figure A2).

Figure 5. Average harvest rate and age composition of Snake River fall Chinook Salmon yearling

(1+) and subyearling (0+) hatchery release groups (Release Years 2006, 2008-2011).

0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

Harv

est

Rate

Jacks

Adults


Figure 6. Comparison of average harvest rate and age composition by common release site of

Snake River fall Chinook Salmon yearling (1+) and subyearling (0+) hatchery releases (Release

Years 2006, 2008-2011).

Smolt to Adult Survival Rates

Smolt to adult survival rates (SAS) were calculated as: (Total number of ADCWT estimated

in harvest + number of ADCWT fish estimated at LGR) / (Total number of ADCWT fish

released) x 100. Comparisons of SAS were made between release sites (Figure 7, Table A6,

Figure A3) and between yearling and subyearling release strategies (Figure 8, Table A7, Figure

A3).

0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40H

arv

est

Rate

Jacks

Adults


Figure 7. Average SAS and age composition of Snake River fall Chinook Salmon yearling (1+) and

subyearling (0+) hatchery release groups (Release Years 2006, 2008-2011).

Figure 8. Comparison of average SAS and age composition by common release site of Snake River

fall Chinook Salmon yearling (1+) and subyearling (0+) releases (Release Years 2006, 2008-2011).

Size of Snake River fall Chinook within Harvest and Return to Lower Granite Dam

Harvest length data were obtained from fish measured as part of fishery sampling and

reported to the RMIS database. Length data presented from LGR were obtained as part of the

0.00

0.50

1.00

1.50

2.00

2.50

3.00

SA

SJacks

Adults

0.00

0.50

1.00

1.50

2.00

2.50

3.00

SA

S

Jacks

Adults


directed sampling effort at the LGR adult trap. If larger than 30 centimeters (cm), mini-jacks are

included in the LGR metrics. Fish with known lengths were separated into four categories: 30 –

52 cm; 53 – 69 cm, 70 – 79 cm and ≥ 80 cm and the percentage of each category represented in

the harvest and return to LGR samples were compared for yearlings and subyearlings (Figure 9).

Figure 9. Comparison of the average length (centimeters) distribution of adult returns from

yearling (1+) and Subyearling (0+) releases of fall Chinook Salmon that were intercepted in a

fishery and those that were not intercepted in a fishery and returned to LGR (Release Years 2006-

2011). Numbers within each bar graph represent the percentages by length category.

Run Timing Comparisons

Average run timing of PIT tagged Snake River fall Chinook Salmon adults at Bonneville

Dam for return years 2011 - 2015 were compared among release sites for yearling (Figure 10)

and subyearling (Figure 11) release groups and between all yearling and subyearling groups

(Figure 12).

8

37

6

45

44

40

40

23

39

14

38

24

9 916

8

0

20

40

60

80

100

Per

cen

tage

≥80

70-79

53-69

30-52


Figure 10. Average run timing at Bonneville Dam of adults returning from Yearling (1+) releases

of fall Chinook Salmon (2011-2015).

Figure 11. Average run timing at Bonneville Dam of adult returns from subyearling (0+) releases of

fall Chinook Salmon (2011-2015).

0.0%

25.0%

50.0%

75.0%

100.0%P

erce

nt

of

Ru

n C

om

ple

te

Date

BCCAP 1+

CJRAP 1+

LYFE 1+

PLAP 1+

0.0%

25.0%

50.0%

75.0%

100.0%

Per

cen

t of

Ru

n C

om

ple

te

Date

BCCAP 0+

CJRAP 0+

LYFE 0+

PLAP 0+


Figure 12. Average run timing at Bonneville Dam of adult returns from yearling and subyearling

releases of fall Chinook Salmon (2011-2015).

0%

25%

50%

75%

100%

Per

cen

t of

Ru

n C

om

ple

te

Date

Yearling

Subyearling


Citations

Regional Mark Information System (RMIS). Regional Mark Processing Center, Pacific States Marine

Fisheries Commission, 205 SE Spokane Street, Suite 100 Portland, Oregon 97202.

http://www.rmpc.org/

Young, W.P., S. Rosenberger and D. Milks. 2012. Snake River Fall Chinook Salmon Run Reconstruction

at Lower Granite Dam; Methods for Retrospective Analysis. Nez Perce Tribe, Department of

Fisheries Resources Management. http://www.nptfisheries.org/PublicationLibrary.aspx


Appendix

Table A1. Revised production table listing Snake River fall Chinook Salmon production priorities

for Lyons Ferry Hatchery per the US v OR Management Agreement, Table B4B, for brood years

2008-2017.


Table A2. Average SAR and age composition of returning adults from Snake River fall Chinook

Salmon hatchery yearling (1+) and subyearling (0+) release groups (Release Years 2006, 2008-

2011).

SAR %

Priority Age Release Site Adults by age Rank J by age Rank sum A+J Rank

1 1+ Lyons Ferry Onstation 0.50 5 0.76 1 1.26 1

2 1+ Pittsburg Landing 0.25 11 0.52 3 0.77 7

3 1+ Big Canyon 0.27 10 0.43 5 0.70 10

4 1+ Captain John Rapids 0.38 8 0.59 2 0.97 3



7 0+ Big Canyon 0.61 3 0.33 6 0.94 5


9 0+ Hells Canyon Dam/Ox 0.40 7 0.31 7 0.70 9

11 0+ Couse Creek 0.51 4 0.29 10 0.79 6

13 0+ Grande Ronde River 0.25 12 0.09 12 0.34 12

15 0+ Hells Canyon Dam/Uma 0.38 9 0.29 11 0.66 11

Table A3. Comparison of average SAR and age composition of returning adults by common

release site from Snake River fall Chinook Salmon yearling (1+) and subyearling (0+) hatchery

releases (Release Years 2006, 2008-2011).

SAR %

Release site Age at Release Adults by age J by age Sum A+J

Lyons Ferry Onstation 1+ Yearling 0.50 0.76 1.26

Lyons Ferry Onstation 0+ Subyearling 0.64 0.44 1.08

Captain John Rapids 1+ Yearling 0.38 0.59 0.97

Captain John Rapids 0+ Subyearling 0.65 0.30 0.95

Big Canyon 1+ Yearling 0.27 0.43 0.70

Big Canyon 0+ Subyearling 0.61 0.33 0.94

Pittsburg Landing 1+ Yearling 0.25 0.52 0.70

Pittsburg Landing 0+ Subyearling 0.43 0.29 0.72


Table A4. Average harvest rate and age composition of Snake River fall Chinook Salmon yearling

(1+) and subyearling (0+) hatchery release groups (Release Years 2006, 2008-2011).

Harvest Rate

# Recovered/# Released

Priority Age Release Site Adults Jacks A+J Rank

1 1+ Lyons Ferry Onstation 0.81 0.43 1.24 1

2 1+ Pittsburg Landing 0.46 0.18 0.65 3

3 1+ Big Canyon 0.45 0.18 0.63 4

4 1+ Captain John Rapids 0.59 0.24 0.83 2

5 0+ Lyons Ferry Onstation 0.46 0.03 0.49 7

6 0+ Captain John Rapids 0.57 0.02 0.59 5

7 0+ Big Canyon 0.55 0.01 0.56 6

8 0+ Pittsburg Landing 0.47 0.01 0.48 8

9 0+ Hells Canyon Dam/Ox 0.43 0.03 0.46 10

11 0+ Couse Creek 0.44 0.03 0.48 9

13 0+ Grande Ronde River 0.26 0.02 0.28 12

15 0+ Hells Canyon Dam/Uma 0.37 0.03 0.40 11

Table A5. Comparison of average harvest rate and age composition by common release site of

Snake River fall Chinook Salmon yearling (1+) and subyearling (0+) hatchery releases (Release

Years 2006, 2008-2011).

Harvest Rate

# Recovered/# Released

Release Site Adults Jacks A+J

Lyons Ferry Onstation 1+ 0.81 0.43 1.24

Lyons Ferry Onstation 0+ 0.46 0.03 0.49

Captain John Rapids 1+ 0.59 0.24 0.83

Captain John Rapids 0+ 0.57 0.02 0.59

Big Canyon 1+ 0.45 0.18 0.63

Big Canyon 0+ 0.55 0.01 0.56

Pittsburg Landing 1+ 0.46 0.18 0.65

Pittsburg Landing 0+ 0.47 0.01 0.48


Table A6. Average SAS and age composition of Snake River fall Chinook Salmon yearling (1+) and

subyearling (0+) hatchery release groups (Release Years 2006, 2008-2011).

SAS %

Priority Age Release Site Adults by age Rank J by age Rank sum A+J Rank



3 1+ Big Canyon 0.72 10 0.61 4 1.33 7




7 0+ Big Canyon 1.16 3 0.34 7 1.49 5


9 0+ Hells Canyon Dam/Ox 0.83 8 0.34 6 1.16 10

11 0+ Couse Creek 0.95 6 0.32 9 1.27 8

13 0+ Grande Ronde River 0.51 12 0.11 12 0.62 12

15 0+ Hells Canyon Dam/Uma 0.75 9 0.31 10 1.06 11

Table A7. Comparison of average SAS and age composition by common release site of Snake River

fall Chinook Salmon yearling (1+) and subyearling (0+) releases (Release Years 2006, 2008-2011).

SAS %

Release site Age at Release Adults by age J by age Sum A+J

Lyons Ferry Onstation 1+ Yearling 1.31 1.19 2.50

Lyons Ferry Onstation 0+ Subyearling 1.10 0.47 1.57

Captain John Rapids 1+ Yearling 0.97 0.83 1.80

Captain John Rapids 0+ Subyearling 1.22 0.32 1.55

Big Canyon 1+ Yearling 0.72 0.61 1.33

Big Canyon 0+ Subyearling 1.16 0.34 1.49

Pittsburg Landing 1+ Yearling 0.71 0.70 1.41

Pittsburg Landing 0+ Subyearling 0.90 0.30 1.20


Figure A1. Box plots and annual estimates (dots) of percent deviation from the mean smolt to adult

return (SAR) rate to Lower Granite Dam for Snake River adult, jack and total subyearling (0+)

and yearling (1+) hatchery release groups (left) and pairwise release location comparisons (right).

The mean SAR is represented by the dashed line. Horizontal solid lines in each box is the median

SAR from release year 2006, 2008 – 2011 and the boxes represent the 25th and 75th percent

deviation from the median. LFH – Lyons Ferry Hatchery; CJA – Captain Johns Rapids; PLA –

Pittsburg Landing; BCA – Big Canyon; CC – Couse Creek; GR – Grande Ronde; HCD OX – Hells

Canyon Dam, Oxbow; HCD UM – Hells Canyon Dam, Umatilla.

SAR, Jacks

LFH 1

+

CJA

1+

PLA 1

+

BCA 1

+

LFH 0

+

CC 0

+

GR 0

+

HCD O

X 0

+

HCD U

M 0

+

CJA

0+

PLA 0

+

BCA 0

+

Per

cent

Dev

iatio

n

from

the

Mea

n

-200

-100

0

100

200

300

400

SAR, Adults

LFH 1

+

CJA

1+

PLA 1

+

BCA 1

+

LFH 0

+

CC 0

+

GR 0

+

HCD O

X 0

+

HCD U

M 0

+

CJA

0+

PLA 0

+

BCA 0

+

Per

cent

Dev

iatio

n

from

the

Mea

n

-150-100

-50

050

100

150200

SAR, Total

LFH 1

+

CJA

1+

PLA 1

+

BCA 1

+

LFH 0

+

CC 0

+

GR 0

+

HCD O

X 0

+

HCD U

M 0

+

CJA

0+

PLA 0

+

BCA 0

+

Per

cent

Dev

iatio

n

from

the

Mea

n

-100

0

100

200

LFH 1

+

LFH 0

+

CJA

1+

CJA

0+

PLA 1

+

PLA 0

+

BCA 1

+

BCA 0

+

-100

-50

0

50

100

150

200

SAR, Total

LFH 1

+

LFH 0

+

CJA

1+

CJA

0+

PLA

1+

PLA

0+

BCA 1

+

BCA 0

+

-100

0

100

200

300

SAR, Jack

LFH 1

+

LFH 0

+

CJA

1+

CJA

0+

PLA 1

+

PLA 0

+

BCA 1

+

BCA 0

+

-200

-100

0

100

200

300

400

SAR, Adults


Figure A2. Box plots and annual estimates (dots) of percent deviation from the mean harvest rate

for Snake River adult, jack and total subyearling (0+) and yearling (1+) hatchery release groups

(left) and pairwise release location comparisons (right). The mean harvest rate is represented by

the dashed line. Horizontal solid lines in each box is the median harvest rate from release year 2006,

2008 – 2011 and the boxes represent the 25th and 75th percent deviation from the median. LFH –

Lyons Ferry Hatchery; CJA – Captain Johns Rapids; PLA – Pittsburg Landing; BCA – Big

Canyon; CC – Couse Creek; GR – Grande Ronde; HCD OX – Hells Canyon Dam, Oxbow; HCD

UM – Hells Canyon Dam, Umatilla.

Harvest, Adults

LFH 1

+

CJA

1+

PLA

1+

BCA 1

+

LFH 0

+

CC 0

+

GR 0

+

HCD O

X 0

+

HCD U

M 0

+

CJA

0+

PLA

0+

BCA 0

+

Per

cent

Dev

iatio

n

from

the

Mea

n

-150

-100

-50

0

50

100

150

200

Harvest, Jacks

LFH 1

+

CJA

1+

PLA 1

+

BCA 1

+

LFH 0

+

CC 0

+

GR 0

+

HCD O

X 0

+

HCD U

M 0

+

CJA

0+

PLA 0

+

BCA 0

+

Per

cent

Dev

iatio

n

from

the

Mea

n

0

200

400

600

800

Harvest, Total

LFH 1

+

CJA

1+

PLA 1

+

BCA 1

+

LFH 0

+

CC 0

+

GR 0

+

HCD O

X 0

+

HCD U

M 0

+

CJA

0+

PLA 0

+

BCA 0

+

Per

cent

Dev

iatio

n

from

the

Mea

n

-200

-100

0

100

200

300

Harvest, Adults

LFH 1

+

LFH 0

+

CJA

1+

CJA

0+

PLA 1

+

PLA 0

+

BCA 1

+

BCA 0

+

-100

-50

0

50

100

150

200

Harvest, Jacks

LFH 1

+

LFH 0

+

CJA

1+

CJA

0+

PLA 1

+

PLA 0

+

BCA 1

+

BCA 0

+

-200

0

200

400

600

800

Harvest, Total

LFH 1

+

LFH 0

+

CJA

1+

CJA

0+

PLA 1

+

PLA 0

+

BCA 1

+

BCA 0

+

-200

-100

0

100

200

300


Figure A3. Box plots and annual estimates (dots) of percent deviation from the mean smolt-to-

adult survival (SAS) rate for Snake River adult, jack and total subyearling (0+) and yearling (1+)

hatchery release groups (left) and pairwise release location comparisons (right). The mean SAS is

represented by the dashed line. Horizontal solid lines in each box is the median SAS of release years

2006, 2008 – 2011 and the boxes represent the 25th and 75th percent deviation from the median.

LFH – Lyons Ferry Hatchery; CJA – Captain Johns Rapids; PLA – Pittsburg Landing; BCA – Big

Canyon; CC – Couse Creek; GR – Grande Ronde; HCD OX – Hells Canyon Dam, Oxbow; HCD

UM – Hells Canyon Dam, Umatilla.

SAS, Adult

LFH 1

+

CJA

1+

PLA 1

+

BCA 1

+

LFH 0

+

CC 0

+

GR 0

+

HCD O

X 0

+

HCD U

M 0

+

CJA

0+

PLA 0

+

BCA 0

+

Per

cent

Dev

iatio

n

fro

m t

he M

ean

-100

-50

0

50

100

150

SAS, Jacks

LFH 1

+

CJA

1+

PLA 1

+

BCA 1

+

LFH 0

+

CC 0

+

GR 0

+

HCD O

X 0

+

HCD U

M 0

+

CJA

0+

PLA 0

+

BCA 0

+

Per

cent

Dev

iatio

n

fro

m t

he M

ean

-100

0

100

200

300

400

500

SAS, Total

LFH 1

+

CJA

1+

PLA

1+

BCA 1

+

LFH 0

+

CC 0

+

GR 0

+

HCD O

X 0

+

HCD U

M 0

+

CJA

0+

PLA

0+

BCA 0

+

Per

cent

Dev

iatio

n

fro

m t

he M

ean

-100

-50

0

50

100

150

200

SAS, Adult

LFH 1

+

LFH 0

+

CJA

1+

CJA

0+

PLA 1

+

PLA 0

+

BCA 1

+

BCA 0

+

-100

-50

0

50

100

150

SAS, Jacks

LFH 1

+

LFH 0

+

CJA

1+

CJA

0+

PLA 1

+

PLA 0

+

BCA 1

+

BCA 0

+

-200

-100

0

100

200

300

400

500

SAS, Total

LFH 1

+

LFH 0

+

CJA

1+

CJA

0+

PLA 1

+

PLA 0

+

BCA 1

+

BCA 0

+

-100

-50

0

50

100

150

200

1

Spatial Distribution of Redds During Periods of Low and High Abundance

Billy Connor, Bill Arnsberg, Jim Chandler, Brad Alcorn, and Debbie Milks

USFWS, NPTDFRM, IPC, and WDFW

Orofino Idaho, Boise Idaho, Dayton Washington

INTRODUCTION

Redd surveys have a long history in the management of Snake River basin fall Chinook salmon (see Connor et al. 2016 for details). The first redd surveys were conducted along the Middle Snake River from fixed wing aircraft in 1947 to document spawning distribution prior to the construction of Brownlee, Oxbow, and Hells Canyon dams (Figure 1). Fish traps were installed at both Brownlee and Oxbow dams to collect adult fall Chinook salmon that were destined for upstream spawning areas. After being trapped, the adults were hauled upstream of Brownlee Reservoir and released to spawn. Redd surveys were conducted in association with trapping and hauling to confirm that the fish did spawn. As the population of fall Chinook salmon in the Middle Snake River collapsed in the 1960s, biologists redirected redd survey effort to investigate spawning in the Lower Snake River and some of its tributaries (Figure 1).

Beginning in 1991, the number and spatial coverage of redd surveys increased in anticipation of listing of Snake River basin fall Chinook salmon under the Endangered Species Act. Technological advances were made to count redds that were too deep to see from the air (Groves and Garcia 1998), and to eliminate the risk associated with manned-flights (Groves et al. 2016). The data collected during the contemporary redd surveys have been used to define suitable spawning habitat (Groves and Chandler 1999), predict spawning habitat carrying capacity (Connor et al. 2001), validate the predictions of spawning habitat carrying capacity (Groves et al. 2013), and describe the changes in the spatial distribution of redds over time.

Connor et al. (2016) used redd count data collected starting in 1991 to delineate the contemporary spawning areas, and rank them according to their production potential. Spawning areas that could support at least 500 females were classified as primary production areas. Spawning areas that could support at least 200 females were classified as secondary production areas. If there was evidence for limited spawning with no consecutive lapses in occupancy during years when redds surveys were conducted, a spawning area was classified as a tertiary spawning area. The Clearwater spawning aggregate includes the Clearwater River lower reach, Clearwater upper reach, SF Clearwater lower reach, MF Clearwater, and the Selway lower reach (Table 1; Figure 2). The Lower Hells Canyon spawning aggregate includes the Lower Hells Canyon spawning area, Imnaha River lower reach, Salmon River lower reach, and Grande Ronde River lower reach (Table 1; Figure 2). A single spawning aggregate occupies Upper Hells Canyon (Table 1; Figure 2).

2

Figure 1.—The study area including the contemporary distribution of the spawning areas.

3

Table 1. —Snake River basin fall Chinook salmon spawning areas located upstream of Lower Granite Dam. Accessible channel

length is given in kms. Channel length (kms) were determined from river charts or redd locations documented by the authors except

for the Lower Snake upper and lower reaches that are from Dauble et al. (2003). Abbreviations: rkm, river km (river mouth = rkm 0);

R., River; D., Dam; Res., Reservoir; S. F., South Fork; M.F., Middle Fork.

Accessible Production

Spawning area1 Reach Upper end2 Lower end3 length ranking

Upper Hells Canyon Contiguous Hells Canyon Dam rkm 398.7 Salmon River rkm 302.9 95.8 Primary

Lower Hells Canyon2 Contiguous Salmon River rkm 302.9 Lower Granite Reservior rkm 234.0 68.9 Primary

Grande Ronde Lower rkm 116.0 rkm 0.0 116.0 Secondary

Imnaha Lower rkm 33.0 rkm 0.0 33.0 Tertiary

Salmon Lower rkm 176.9 rkm 0.0 176.9 Tertiary

Clearwater Lower rkm 65.0 Lower Granite Reservoir rkm 6.0 59.0 Primary

Upper rkm 120.0 rkm 65.0 55.0 Tertiary

S. F. Clearwater Lower rkm 30.9 rkm 0.0 30.9 Tertiary

M. F. Clearwater Contiguous Selway River. rkm 157.0 S. F. Clearwater River rkm 120.0 37.0 Tertiary

Selway Lower rkm 31.0 rkm 0.0 31.0 Tertiary

1The Tucannon River is a secondary spawning area that is is located downstream of Lower Granite Dam (Figure 1).

2Land marks are sometimes given (e.g., M.F. Clearwater extends from the Selway mouth to the S. F. Clearwater mouth; Figure 1). 3Presented as one reach for simplicity, but can also be divided into two reaches using the Grande Ronde River as the dividing point.

4

Figure 2.—The contemporary fall Chinook salmon spawning aggregates located upstream of

Lower Granite Dam.

Acclimation and direct release sites for hatchery-origin subyearling and yearling fall Chinook salmon smolts were selected within the primary, secondary, and tertiary spawning areas with the intent to return large portions of the returning adults to the spawning area of release (i.e., fidelity). It was suspected in advance, however, that portions of the adults that had been acclimated or directly released as juveniles at a particular site would distribute elsewhere in the basin (i.e., dispersal). The location of the acclimation and release sites combined with the number of hatchery-origin smolts released at those sites was intended to increase the proportion of spawning within the Upper Hells Canyon and Clearwater spawning aggregates, while maintaining a large portion of spawning within the Lower Hells Canyon spawning aggregate.

An increase in the proportion of spawning within the Upper Hells Canyon spawning aggregate was desired because the contemporary juveniles produced in that area have phenotypic characteristics that are most similar to the juveniles that were produced by the Middle Snake River population prior to the elimination of that population (see Connor et al. 2002, 2016 for details). In 1927 Lewiston Dam was constructed 7.4 kms upstream of the Clearwater River mouth to create a mill pond and power facility for a large timber mill that received timber from a

5

large log drive originating along the North Fork Clearwater River. The fish ladder at Lewiston Dam did not pass fish from mid-summer through winter until 1938. Lack of passage during that 11-year period functionally extirpated fall Chinook salmon in the Clearwater River drainage including upstream production in the Selway River lower reach. Managers hoped to increase the phenotypic diversity of the contemporary population by increasing spawning within the Clearwater aggregate, and by restoring spawning as far upstream as the Selway lower reach.

The objective of this presentation is to compare the spatial distribution of redds during periods of low and high abundance that were largely affected by the return of hatchery-origin adults from acclimated and direct releases of hatchery-origin smolts.

METHODS

Redd survey methods are described by Groves et al. (2013, 2016), and Arnsberg and Kellar (2014). In some years, flights scheduled during the mid-to-latter portion of the spawning period were not completed due to fall freshets and high levels of turbidity. Data collected during years when temporal coverage was complete were used to calculate the cumulative proportion of the total annual redd count for each survey date expressed as day of year (January 1 = 1). Those results were then used to fit spawning area specific regression equations to estimate the cumulative proportion of the total annual redd count (Pcumulative) from day of year. The regression equations were then applied as described in the following example to expand total annual redd counts during years the survey coverage was incomplete. In 2012, the last survey along the Clearwater River lower reach was conducted on November 8th (day of year 313) roughly one month before the end of the spawning season. The total redd count on day of year 313 was 1,084. The regression equation fitted to the Clearwater lower reach data collected during the five years when the surveys had complete temporal coverage predicted that Pcumulative on day of year 313 was 0.645 ± 0.27. Dividing 1,084 by 0.645 provided an expanded total annual redd count of 1,677 for 2012.

RESULTS

Total annual redd counts before and after expansion for all three spawning aggregates combined are given in Table 2.

Trends in the total annual redd counts made among the spawning areas upstream of Lower Granite Reservoir confirmed a general increase in spawning during 1991–2016, while supporting the delineation of study years into low (1991‒2000) and high (2001–2016) abundance periods (Figure 3). Annual redd counts generally increased from a low of 66 in 1991 to a high of 9,378 in 2015. Total annual redd counts during the last year of the low abundance period and first year of the high abundance period were 538 and 1,330, respectively.

6

Table 2.—Unexpanded redd counts (all three spawning aggregates combined) compared to redd that were expanded to account for incomplete temporal coverage of the surveys, 1991–2016. The data are provisional as the method is still being refined.

Year Unexpanded Expanded Difference

1991 64 64 0 1992 83 83 0 1993 219 220 1 1994 111 120 9 1995 109 145 36 1996 195 228 33 1997 175 175 0 1998 303 352 49 1999 578 584 6 2000 536 538 2 2001 1,278 1,330 52 2002 1,851 1,864 13 2003 2,241 2,288 47 2004 2,555 2,570 15 2005 2,121 2,144 23 2006 1,368 1,670 302 2007 1,950 1,959 9 2008 3,035 3,134 99 2009 3,412 3,420 8 2010 4,933 4,933 0 2011 4,668 4,693 25 2012 3,435 4,434 999 2013 5,934 6,533 599 2014 6,407 8,583 2,176 2015 8,826 9,378 552 2016 4,687 5,985 1,298

7

Figure 3.—Total annual redd counts (all three spawning aggregates combined; expanded) during the low and high abundance periods.

During the low abundance period, the largest proportion of the redds was counted on average within the Lower Hells Canyon spawning aggregate, followed by Clearwater spawning aggregate, and then the Upper Hells Canyon spawning aggregate (Figure 4). The inter-annual mean (± SE) proportion of redds counted within the Upper Hells Canyon spawning aggregate increased from 29.2 ± 4.3% during the low abundance period to 32.9 ± 1.3% during the high abundance period. The inter-annual mean (± SE) proportion of redds counted within the Lower Hells Canyon spawning aggregate decreased from 43.7 ± 5.3% during the low abundance period to 31.5 ± 1.8% during the high abundance period. The inter-annual mean (± SE) proportion of redds counted within the Clearwater spawning aggregate increased from 27.1 ± 3.1% during the low abundance period to 35.6 ± 2.7% during the high abundance period.

19911993

19951997

19992001

20032005

20072009

20112013

2015

Year

0

1

2

3

4

5

6

7

8

9

10Number of redds counted (thousands)

Low abundance (1991‐2000) N = 10 years

High abundance (2001‐2016) N = 16 years

8

Figure 4.—The Inter-annual mean (± SE) proportion of redds counted in each of the three spawning aggregates during the low and high abundance periods.

TAKE HOME MESSAGE

The trends in the spatial distribution of redds between the low and high abundance periods support the conclusion that the location of the acclimation and release sites combined with the number of hatchery-origin smolts released at those sites increased the proportion of spawning within the Upper Hells Canyon and Clearwater spawning aggregates, while maintaining a large portion of spawning within the Lower Hells Canyon spawning aggregate.

Abundance period

Low High0.0

0.2

0.4

0.6

0.8

1.0

Upper Hells Canyon Lower Hells Canyon Clearwater

aggregate aggregate aggregate

Inter‐an

nual m

ean (+ SE) proportion of redds

9

REFERENCES

Arnsberg, B.D. and D.S. Kellar. 2014. Nez Perce Tribal Hatchery monitoring and evaluation project; Fall Chinook salmon (Oncorhynchus tshawytscha) supplementation in the Clearwater River Subbasin. 2013 Annual Report to the U.S. Department of Energy, Bonneville Power Administration, Project No. 1983-350-003. Connor, W.P., A.P. Garcia, A.H. Connor, E.O. Garton, P.A. Groves, and J.A. Chandler. 2001. Estimating the carrying capacity of the Snake River for fall Chinook salmon redds. Northwest Science 75:363–370. https://research.libraries.wsu.edu/xmlui/bitstream/handle/2376/1051/v75%20p363%20Connor%20et%20al.PDF?sequence=1 (Available May 6, 2017) Connor, W.P., H.L. Burge, R. Waitt, and T.C. Bjornn. 2002. Juvenile life history of wild fall Chinook salmon in the Snake and Clearwater rivers. North American Journal of Fisheries 22:703–712. http://afs.tandfonline.com/doi/abs/10.1577/1548-8675(2002)022%3C0703%3AJLHOWF%3E2.0.CO%3B2 (Available May 6, 2017) Connor, W.P., B.D. Arnsberg, J.A. Chandler, T.D. Cooney, P.A. Groves, J.A. Hesse, G.W. Mendel, D.J. Milks, D.W. Rondorf, S.J. Rosenberger, M.L. Schuck, K.F. Tiffan, R.S. Waples, and W. Young. 2016. A Retrospective (circa 1800–2015) on abundance, spatial distribution, and management of Snake River Basin fall Chinook salmon. Draft 2 Parts I, II, and III. http://www.streamnetlibrary.org/?page_id=1357 (Available May 13, 2016). Groves, P.A., and A.P. Garcia. 1998. Two carriers used to suspend an underwater video camera from a boat. North American Journal of Fisheries Management 18:1004–1007.Groves et al. 2013 http://www.tandfonline.com/doi/abs/10.1577/1548-8675(1998)018%3C1004%3ATCUTSA%3E2.0.CO%3B2 (Available May 13, 2016). Groves, P.A., and J.A. Chandler. 1999. Spawning habitat used by fall Chinook salmon in the Snake River. North American Journal of Fisheries Management 19:912–922.

10

http://afs.tandfonline.com/doi/abs/10.1577/1548-8675(1999)019%3C0912%3ASHUBFC%3E2.0.CO%3B2 (Available May 13, 2016). Groves, P.A., J.A. Chandler, B. Alcorn, T.J. Richter, W.P. Connor, A.P. Garcia, and S.M. Bradbury. 2013. Evaluating salmon spawning habitat capacity using redd survey data. North American Journal of Fisheries Management 33:707–716. Groves, P. A, B. Alcorn, B., M. M. Wiest, J. M., Maselko, and W. P. Connor. 2016. Testing unmanned aircraft systems (UASs) for salmon spawning surveys. FACETS. DOI: 10.1139/facets-2016-0019. http://www.facetsjournal.com/article/facets-2016-0019 (Available November 2016).

Fidelity, Dispersal, and Fallback of Snake River Fall Chinook Salmon Yearling and Subyearling Adult Returns

Peter Cleary, Nez Perce Tribe Debbie Milks, Washington Department of Fish and Wildlife

Afton Oakerman, Washington Department of Fish and Wildlife Billy Connor, United States Fish and Wildlife Service

Kenneth Tiffan, United States Geological Service

May 8, 2017

2

Background

The need to determine fidelity and fallback rates was proposed as a result of consultation with co-managers for the Snake River Fall Chinook Hatchery and Genetic Management Plan (HGMP) Addendum in association with the subsequent issuance of a Section 10(a)(l)(A) ESA take permit. The National Oceanic Atmosphere Administration (NOAA) Fisheries indicated that this research would contribute to satisfying the requirements for the Snake River Fall Chinook portion of the Federal Columbia River Power System (FCRPS) Biological Opinion Reasonable Prudent Action (RPA); specifically RPA 64 and 65.

The HGMP hatchery programs were developed to mitigate for the effects of the FCRPS and Idaho Power Company dams. These programs were developed by the US v. Oregon Parties who agreed to implement them according to the 2008-2017 US vs. Oregon Management Agreement. The hatchery programs are intended to supplement the natural spawning component of Snake River fall Chinook and provide for continued fisheries within and outside of the Snake River basin. Uncertainty exists regarding the success or impacts of supplementation on the natural population. Recent scientific evidence suggests (Araki et al 2007; Chilcote 2011) that a high percentage of hatchery origin spawners (pHOS) in salmon and steelhead populations may result in reduced natural productivity. NOAA staff recommended reducing the proportion of hatchery origin fall Chinook spawning naturally to reduce the potential risk to productivity. During consultation with NOAA Fisheries for ESA coverage for the programs, several critical data gaps were identified; including the feasibly of altering pHOS via modified hatchery release locations. An agreement was reached to design and implement studies to answer these questions as a special condition of Section 10 Permits 16607 and 16615. One of these studies sought to determine fidelity and fallback rates of hatchery fall Chinook for the HGMP proposed during the consultation. Prior to consultation, Garcia et al. (2004) tracked a total of 515 radio tagged fall Chinook from 1997 to 2001 to spawning areas in the Snake River Basin. They calculated dispersal and fidelity rates and assigned the radio tagged fish to one of three groups of adult radio tagged Snake River fall Chinook salmon; the Lower Snake River, Upper Snake River, and Lower Clearwater River.

The Interior Columbia Technical Recovery Team (ICTRT) stated in a draft 2003 identification of independent populations of the Evolutionarily Significant Units (ESUs) that the distribution and abundance of Snake River fall Chinook salmon suggested it was a single population. However, they noted there is finer-scale differentiation between spawning areas (ICTRT 2003). Core spawning areas identified by ICTRT were “a 32 km section of the mainstem Snake River starting approximately 10 km above the Asotin Creek confluence; lower portions of the Salmon, Clearwater, and Grande Ronde rivers and Snake River rkm 343 to rkm 353” (ICTRT 2003). In addition to core spawning areas, NPT has documented fall Chinook redds in the lower Imnaha River (Arnsberg et al. 2014) and as a result acclimated and released fall Chinook salmon in the South Fork Clearwater and Selway rivers.

Spawning aggregates within the Snake River Basin are not isolated and although the fish released from the Fall Chinook Acclimation Project (FCAP) facilities have relatively strong fidelity to their release location there is a measurable amount of interchange of fish. Garcia et al. (2004) determined that 32% of the supplementation fish overall spawned in areas other than their

3

release reaches. For example, they estimated that 11% of the males and 4% of the females returning from Captain Johns Rapids acclimation pond (Lower Snake River Reach) yearling releases spawned in the Upper Snake River Reach. Since the early 2000’s, increased releases of subyearling smolts have occurred throughout the basin from FCAP acclimation facilities and as direct releases from hatcheries in Washington, Oregon and Idaho. Fidelity of adults to release site areas was unknown for these subyearling release groups. The study proposed during consultation attempts to understand spawner distribution associated with subyearling releases by replicating much of the Garcia et al. (2004) study using returning adults that were Passive Integrated Transponder (PIT) tagged as juveniles under the U.S. Army Corp of Engineers’ (COE) Evaluating the Responses of Snake and Columbia River Basin Fall Chinook Salmon to Dam Passage Strategies and Experiences study (Consensus Study) returning from 2013 to 2017.

Additionally, the study attempts to understand the behavior of adults released as subyearlings and yearlings below Lower Granite Dam (LGD) at Lyons Ferry Hatchery (LFH), that immigrate upstream of LFH to the natural production areas upstream of LGD. Subyearling and yearling releases at LFH are referred to as “On-station” releases and have been documented passing LGD through systematic sampling at the adult trap facility at the dam and from detections of PIT tags passing the ladder. Alternatively, subyearlings and yearlings reared at LFH, but released at sites upstream of Lower Granite Dam are referred to as “Off-station” releases. During the planning of the study NOAA was concerned that On-station adults contribute a significant number of hatchery fish to the spawning population upstream of LGR Dam. If these LFH On-station released fish do intermix with wild fish in significant numbers, they could be depressing development of within population diversity that may arise from spawning in diverse habitats. Mendel and Milks (1997) documented upstream migration, passage at dams and the ultimate spawning location of fall Chinook released from Lyons Ferry Hatchery returning as adults in 1992 and 1993. They found that a significant number of those fish fell back over or through the dam and thus did not remain in upstream spawning reaches. Radio tagging of subyearlings and yearlings from 2013 to 2016 was done to determine if subyearlings behave similarly to yearling releases. The intent of the study from 2013 to 2016 was to determine if: 1) homing fidelity from subyearling releases was sufficient to manage individual spawning aggregates, 2) homing fidelity was sufficient to develop localized brood stocks programs that will promote population substructure, 3) there were differences in fall back rates between LFH on-station and LFH off-station releases, and 4) fallback rates were consistent across years.

4

Methods

Returning fall Chinook were trapped at Lower Granite Dam and radio tagged with Lotek VFH radio tags. Adult yearling fall Chinook salmon tagged by Garcia et al. returned from 1997 to 2001. Returning yearling Snake River Fall Chinook salmon radio tagged by Garcia et al. (2004) were identified by the presence of visual implant elastomer (VIE) tags posterior to the eye, and a small number (2-6%) of PIT tags. Colors of VIE tags were specific to release juvenile sites upstream of Lower Granite Dam.

Subyearling and yearling fall Chinook salmon from brood years 2007 to 2013 PIT tagged as juveniles were identified as they returned over dams downstream of Lower Granite Dam from 2013 to 2016. PIT tag codes were compiled weekly and submitted to PTAGIS and NOAA Fisheries for identification at the Lower Granite Dam adult trap. When these PIT tag codes were detected they were diverted into the Lower Granite Dam adult trap and radio tagged. Subyearling and yearling fall Chinook salmon were tracked using both stationary radio telemetry receivers and mobile receivers operating from automobiles and boats. The locations of radio telemetry receivers used to track adults during from 2013 to 16 were stationed at the locations described by Garcia et al. (2004) except a receiver was not placed at rkm 240.7 along the Lower Snake River, and a receivers were placed at Little Goose Dam and Lyons Ferry Hatchery. Stationary and mobile telemetry receivers used to track returning adult subyearling and yearling fall Chinook salmon from 2013 to 2016 operated in the same areas as shown in Figure 1.

The following groups were targeted for radio tagging from 2013 to 2017: 1) Lyons Ferry subyearlings(n = 125) 2) Lyons Ferry yearlings (n = 192) 3) Couse Creek subyearlings direct releases subyearlings (n = 110) 4) Captain John Rapids Acclimation subyearlings (n = 110) 5) Grande Ronde direct release subyearlings (n = 110) 6) Pittsburg Landing Acclimation subyearlings (n = 110) 7) Hells Canyon direct release subyearlings (n = 110) 8) Big Canyon Creek Acclimation, North Lapwai Valley Acclimation, and Nez Perce Tribal

Hatchery Acclimation subyearlings (n = 110) 9) Luke Gulch and Cedar Flats Acclimation subyearlings (n = 110)

Adult returns from LFH represented On-station releases below Lower Granite Dam. Snake

River releases near Couse Creek, Captain John Rapids, and Grande Ronde direct releases represented adult returns from the Lower Hells Canyon area. Pittsburg Landing Acclimation and Hells Canyon releases, represented adult returns from the Upper Hells Canyon area. Big Canyon Creek Acclimation, North Lapwai Valley Acclimation, and Nez Perce Tribal Hatchery Acclimation releases represented returns to the Lower Clearwater area. Luke Gulch and Cedar Flats Acclimation releases represented adult returns from the South Fork Clearwater and Selway rivers, or Upper Clearwater area.

5

Radio tags were tracked into the last reach entered. Detection history was reviewed if the last reach entered differed from the furthest upstream reach entered or when multiple detections occurred at a stationary receiver located between reaches. Radio tags detected were considered to have either become non-functional, regurgitated, or harvested if only detected in Lower Granite Reservoir and therefore eliminated from further analysis. Radio tags last detected downstream of Lower Granite Dam were referred to as fallbacks. Radio tags last detected in Lower Granite Reservoir after entering an upstream spawning reach were thought to have spawned in next section above the reservoir and drifted downstream into the reservoir or could have been prespawning mortalities. Radio tags last detected in a reach upstream of Lower Granite Reservoir were referred to as spawners. Fidelity was defined as the number of radio-tagged fish that returned to their release area, or reach, divided by the number of fish tagged from that release that retained functional tags and were not harvested. Dispersal was defined as the number of radio-tagged fish that did not return to their release reach divided by the number of fish tagged from that release that retained functional tags and were not harvested. Appendix Tables A-1 and A-2 show the number of fish radio tagged and tracked for the two studies conducted from 1997 to 2001, and from 2013 to 2016.

Figure 1. Spawning areas and reach boundaries for the radio telemetry studies tracking fall Chinook salmon in the Snake River from 1997 to 2001 (Garcia et al. 2004) and from 2013 to 2016. Juvenile release locations are shown by the fish symbol. Stationary receivers are shown with a solid black circle.

6

Results Inter-annual mean fidelity rates for the off-station releases of yearlings that returned as adults during 1997–2001 were highest for fish from Big Canyon Creek acclimation facility, followed by fish from Pittsburg Landing acclimation facility, and then by fish from the Captain John Rapids acclimation facility (Table 1; Figure 2). The inter-annual mean dispersal rates into the Upper Hells Canyon spawning area (i.e., the proposed natural production emphasis area) were much lower for the releases of yearlings releases from Big Canyon Creek acclimation facility compared to yearling releases from Captain John Rapids acclimation facility (Table 1; Figure 3).

Inter-annual mean fidelity rates for off-station releases of subyearlings that returned as adults during 2013–2016 from Captain John Rapids, Pittsburg Landing, and Big Canyon Creek acclimation facilities were higher compared to corresponding rates observed for the off-station releases of yearlings during 1997–2001 (Table 1; Figures 2 and 4). Inter-annual mean dispersal rates into the Upper Hells Canyon spawning area were similar between releases of Captain John Rapids yearling and subyearlings when viewed in terms of the SEs and point estimates of the two rates (Table 1; Figures 3 and 5). Along the same lines, the inter-annual mean dispersal rates of the Big Canyon Creek acclimation facility releases of yearlings and subyearlings were also similar (Table 1; Figures 3 and 5).

Lyons Ferry Hatchery subyearlings had the highest inter-annual mean dispersal rate into

the Upper Hells Canyon natural production emphasis area (Table 1; Figure 5). However, inter-annual mean dispersal for the Lyons Ferry subyearling adult returns includes 2013; represented by only 3 Lyons Ferry adults. If the 2013 Lyons Ferry subyearling dispersal rate is discarded the inter-annual mean dispersal rate becomes 8.6% ± SE of 4.4%. Dispersal of other release groups into the Upper Hells Canyon spawning area was lower by comparison to the Lyons Ferry Hatchery on-station releases of yearlings (Table 1; Figure 5).

Inter-annual mean fallback rates of Lyons Ferry Hatchery on-station releases of yearlings

and subyearlings were much higher compared to the fallback rates of off-station releases of yearlings and subyearlings (Table 1; Figure 6).

Take Home Message

This presentation was intended to demonstrate the progress we have made to understand fidelity, dispersal, and fallback of hatchery subyearlings and yearlings released both on-station at LFH and off-station upstream of Lower Granite Dam. The final year of data collection will be 2017. Future analysis of fidelity and dispersion will determine if sex and brood year influenced fidelity and dispersion rates. If sex and brood year does not influence fidelity or dispersion within release groups annual results will be grouped together in the final analysis. Once the final analyses are completed, they will be useful for evaluating the potential for dispersal of hatchery adults into the proposed natural emphasis area, as well for distributing hatchery adults among spawning areas for the purpose of life cycle modeling.

7

Table 1. Inter-annual mean fidelity, dispersal, and fallback rates for yearlings from 1997 to 2001 and yearlings subyearlings from 2013 to 2016. Fidelity and fallback rates are highlighted in yellow and grey, respectively.

Release Group

Below Lower Granite

Dam

Lower Granite

Reservoir

Lower Hells

Canyon

Upper Hells

Canyon Salmon River

Clearwater River

Yearlings (1997-2001) Captain John Rapids 2.5% 7.5% 49.9% 10.4% 0.4% 29.3% Pittsburg Landing 6.4% 5.8% 22.7% 64.4% 0.0% 0.7% Big Canyon Creek 4.0% 6.7% 13.0% 2.3% 0.0% 74.1% Subyearlings and Yearlings (2013-2016)

Lyons Ferry Subyearlings 44.0% 9.9% 10.5% 17.9% 0.0% 17.7% Lyons Ferry Yearlings 50.3% 11.5% 9.3% 2.8% 0.0% 26.1%

Couse Creek 15.0% 0.0% 52.5% 2.5% 0.0% 30.0%

Captain Johns 3.1% 2.8% 63.2% 15.2% 5.0% 10.8% Grande Ronde 0.0% 1.3% 85.3% 3.8% 0.0% 9.7%

Pittsburg Landing 1.1% 1.6% 9.6% 86.3% 0.5% 1.1% Hells Canyon 3.8% 0.0% 11.6% 79.4% 0.6% 4.6%

Clearwater 1.2% 5.9% 1.2% 5.3% 0.0% 86.4% Upper Clearwater 1.9% 7.7% 0.5% 2.8% 0.0% 88.7%

8

Figure 2. Fidelity rates of yearling adult returns from 1997 to 2001 calculated by reanalyzing data collected by Garcia et al. (2004) to standardize the boundaries of release reaches and account pre-spawning mortalities.

Figure 3. Dispersion rates of yearling adult returns from 1997 to 2001 into the Upper Hells Canyon spawning area calculated by reanalyzing data collected by Garcia et al. (2004) to standardize the boundaries of release reaches and account pre-spawning mortalities.

Big Canyon Creek Captain John Rapids Pittsburg Landing

Release group

30

40

50

60

70

80

90

Inter‐an

nual m

ean

(+ SE) fidelity rate (%)

0

5

10

15

20

Inter‐an

nual m

ean

(+ SE) dispersal rate (%

)

Big Canyon Creek Captain John Rapids

Release group

9

Figure 4. Fidelity rates of subyearling adult returns from 2013 to 2016.

Figure 5. Dispersion rates of subyearling and yearling adult returns into the Upper Hells Canyon spawning area from 2013 to 2016.

10

Figure 6. Fallback rates of On-station subyearlings and yearlings are shown with fallback rates of Lyons Ferry Off-station subyearling releases at Couse Cr., Captain John Rapids, Pittsburg Landing, and Clearwater release groups. Also shown are fallback rates for subyearlings produced at the Nez Perce Tribal Hatchery and released at Luke Gulch and Cedar Flats acclimation sites on the South Fork Clearwater and Selway rivers, and subyearlings produced at Irrigon Fish Hatchery.

11

Citations Arnsberg, B., P. Groves, F. Mullins, D. Milks, M. Allen. 2014. 2013 Snake River Fall Chinook

Salmon Spawning Summary. Online (May 8, 2017): http://www.fpc.org/documents/fachin_planningteam/2013CooperativeFallChReddSummary.pdf

Araki. H, B. Cooper, and M.S. Blouin. 2007. Genetic effects of captive breeding cause a rapid,

cumulative fitness decline in the wild. Science. 318:100-103. Chilcote M.W., K.W. Goodson, M.R. Falcy. 2011. Reduced recruitment performance in natural

populations of anadromous salmonids associated with hatchery-reared fish. Canadian Journal of Fisheries and Aquatic Sciences. 68: 511-522.

Garcia, A.P, W.P Connor, D. J. Milks, S. J. Rocklage and R.K. Steinhorst. 2004. Movement

and Spawner Distribution of Hatchery Fall Chinook Salmon Adults Acclimated and Released as Yearlings at Three Locations in the Snake River Basin. NAJFM 24:1134-1144.

ICTRT (Interior Columbia Basin Technical Recovery Team). 2003. Independent Populations of

Chinook, Steelhead, and Sockeye for Listed Evolutionarily Significant Units Within the Interior Columbia River Domain. Online (May 8, 2017): https://www.nwfsc.noaa.gov/research/divisions/cb/genetics/trt/col_docs/independentpopchinsteelsock.pdf

Mendel, G. and D. Milks. 1997. Chapter 1 in: Upstream Passage, Spawning, and Stock

Identification of Fall Chinook Salmon in the Snake River, 1992 and 1993. Editors H.L. Blankenship and G. W. Mendel. Final report to the Bonneville Power Administration. Project Number 92-046. DOE/BP-60415-2.

12

Appendix A-1. Adult returns of yearling Snake River fall Chinook salmon radio tagged and tracked from 1997 to 2001 (data from Garcia et al. 2004).

Release Group Return Year Tagged Unknown

Tag loss Fallback

Pre-spawning Mortality Spawner

Big Canyon Cr. 1998 12 1 2 1 8 1999 28 4 2 2 20 2000 48 3 4 2 3 36 2001 84 2 5 2 75 Captain John Rapids 1999 7 7 2000 34 1 2 2 7 22 2001 118 7 18 1 92 Pittsburg Landing 1997 9 1 2 1 5 1998 30 8 2 2 18 1999 19 2 1 16 2000 43 6 3 34

13

Appendix A-2. Adult returns of subyearling and yearling Snake River fall Chinook salmon radio tagged and tracked from 2013 to 2016.

Release Group Return Year Tagged Unknown

Tag loss Fallback

Pre-spawning Mortality Spawner

Lyons Ferry Subyearlings

2013 3 0 0 2 0 1 2014 25 11 0 6 1 7 2015 32 5 1 10 7 9 2016 34 15 1 5 1 12

Lyons Ferry Yearlings

2013 12 3 2 3 0 4 2014 85 46 2 17 6 14 2015 46 7 4 16 6 13 2016 32 8 0 16 3 5

Couse Creek 2013 12 1 1 0 0 10 2014 12 2 0 1 0 9 2015 4 0 0 0 0 4 2016 2 0 0 1 0 1

Captain John Rapids

2013 14 3 3 0 0 8 2014 73 18 2 4 6 43 2015 22 0 1 1 0 20 2016 5 0 0 0 0 5

Grande Ronde 2013 11 1 0 0 0 10 2014 47 6 2 0 2 37 2015 17 1 0 0 0 16 2016 3 0 0 0 0 3

Pittsburg Landing

2013 7 2 1 0 0 4 2014 60 10 2 2 3 43 2015 18 0 0 0 0 18 2016 1 0 0 0 0 1

Hells Canyon 2013 13 2 1 1 0 9 2014 52 11 1 2 0 38 2015 21 5 0 0 0 16 2016 3 2 0 0 0 1

Clearwater 2013 20 1 1 0 2 16 2014 48 6 1 2 1 38 2015 13 3 0 0 1 9 2016 10 1 0 0 0 9

Upper Clearwater

2013 10 2 0 0 0 8 2014 56 3 1 4 3 45 2015 21 0 0 0 0 21

2016 4 0 0 0 1 3

14

Appendix B-1. They number of yearling fall Chinook salmon detected per reach as a fallback, pre-spawning mortality, or spawner from 1997 to 2001.

Release Group Year

Below Lower Granite

Dam

Lower Granite

Reservoir

Lower Snake River

Upper Snake River

Salmon River

Clearwater River

Captain John Rapids

1997 1998 1999 4 3 2000 2 7 18 2 2 2001 1 32 23 1 36

Pittsburg Landing

1997 1 3 2 1998 2 2 3 15 1999 1 16 2000 3 10 23 1 2001

Big Canyon Creek

1997 1998 1 1 7 1999 2 2 5 1 14 2000 2 3 5 2 29

2001 2 6 69

15

Appendix B-2. They number of subyearling and yearling fall Chinook salmon detected per reach as a fallback, pre-spawning mortality, or spawner from 2013 to 2016.

Release Group Year

Below Lower Granite Dam

Lower Granite

Reservoir

Lower Snake River

Upper Snake River

Salmon River

Clearwater River

Lyons Ferry Subyearlings

2013 2 1 2014 6 1 3 3 1 2015 10 7 1 3 5 2016 5 1 3 1 8

Lyons Ferry Yearlings

2013 3 4 2014 17 6 7 7

2015 16 6 5 1 7 2016 16 3 1 2 2 Couse Creek 2013 5 5 2014 1 6 1 2 2015 2 1 1 2016 1 1 Captain John Rapids

2013 6 2 2014 4 6 27 9 7

2015 1 14 5 1 2016 3 1 1 Grande Ronde

2013 6 1 3 2014 2 34 2 1

2015 15 1 2016 3 Pittsburg Landing

2013 4 2014 2 3 5 35 1 2

2015 5 13 2016 1 Hells Canyon 2013 1 1 7 1 2014 2 7 29 1 1 2015 3 12 1 2016 1 Clearwater 2013 2 16 2014 2 1 2 36 2015 1 1 8 2016 1 8 Upper Clearwater

2013 8 2014 4 3 1 44 2015 1 20

2016 1 3

1

The Juvenile Abundance Component of the Snake River Basin Fall Chinook Salmon Life Cycle Model

Russell Perry, John Plumb, Billy Connor, Ken Tiffan, Tom Cooney, and Bill Young

USGS, USFWS, NPTDFRM, NOAA

Cook WA, Orofino ID, Portland OR, McCall ID

INTRODUCTION

The historical population of Snake River basin fall Chinook salmon was abundant, diverse, productive, and spatially distributed. After a nearly 100-year period of habitat loss, harvest, and climatic variation that began in the late 19th century, these measures of status declined until the population was listed for protection under the U.S. Endangered Species Act (NMFS 1992). Since listing, much management effort has been expended to increase the number of the adults that spawn in the wild and the survival of their natural offspring (e.g., reduced harvest, stable minimum spawning flows, summer flow augmentation, predator control; increased hatchery production and supplementation; improved dam passage structures, summer spill operations). However, understanding how past management actions have contributed to the dynamics and self-regulation of the population is a requirement for managing this at-risk population.

Multistage life cycle models provide a powerful framework for understating how each life stage of a population contributes to the population growth rate by models allowing the effects of density dependence at different life stages to be estimated in the context of management actions. For example, the importance of habitat restoration to population recovery of Chinook salmon may depend on the mechanisms affecting particular life stages. Multistage models may also be used as an analytical framework to explicitly estimate demographic parameters of a population model. This approach has an advantage over single-stage stock-recruitment models by allowing population growth rates to be partitioned among life stages rather than aggregated over an entire life cycle. Such partitioning allows for estimating 1) stage-specific density dependence, and 2) stage-specific effects of environmental factors or management actions. Although multistage models have been applied in the context of a population viability analyses for spring/summer Chinook salmon in the Snake River, such an approach has yet to be applied to fall Chinook salmon in the Snake River basin. The model being developed for Snake River basin fall Chinook salmon includes both juvenile and adults stages. The fundamental components of the adult stage include age, sex, and

2

abundance of adults that escape to the spawning grounds upstream of Lower Granite Dam. Those components are calculated during adult run reconstruction—a process that has been evolving since the 1980s—and in the past several years has become well refined. Efforts to reconstruct the juvenile run and estimate the abundance of natural-origin and hatchery-origin smolts began in 2011. The objective of this presentation is to provide an update on the current state of the two-stage life cycle model emphasizing the progress that has been made on estimating juvenile abundance at Lower Granite Dam.

METHODS Abundance of Early Migrants (March through October of year t)

The abundance of juvenile fall Chinook salmon passing LGR was estimated using a Bayesian hierarchical model that is informed from both daily PIT-tag detections of hatchery-origin smolts and daily counts of marked and unmarked fish in the sample tank at the LGR juvenile fish facility (JFB). Although it is beyond the scope of this write up to describe the model in full, the following provides a brief description of how the annual abundances of natural juveniles passing the LGR were calculated from the sample tank and PIT-tag detection data. Specifically, the annual abundance of natural juvenile fall Chinook salmon ( NN̂ ) was calculated as:

(1) N ,N1

ˆ ˆD

dd

N n

where D is the total number of days the fish were being sampled from the juvenile fish bypass, from March through October. The daily abundance of natural-origin fall Chinook salmon passing the Lower Granite Dam, ,Ndn , was estimated as:

(2) ,N,N

ˆˆ d

dd

bn

C

where dC is the daily probability of collection into the sample tank at the dam. A new method has been developed and soon be implemented to estimate collection probabilities and juvenile abundance. This new method should alleviate critical assumptions and potential sources of bias in past methods to estimate collection probability (Sandford and Smith 2002; Plumb et al. 2012). The parameter ,Nd̂b is the estimated number of natural-origin juveniles that are collected into the

sample, which was estimated as:

3

(3) ,N,N

ˆˆ dd

d

sb

r .

Here ,Nˆds is the daily number of natural-origin juveniles in the sample tank, and dr is the daily

sample rate of fish that have entered the JFB. The daily number of natural-origin juveniles in the sample tank was estimated as: (4) ,N ,Hˆ ˆ(1 )d d ds s p

where ds is the total number of fish in the sample tank on day d, and ,Hˆdp is the estimated

fraction of ds that is hatchery-origin. The proportion of the fish in the sample tank that is

hatchery-origin ( ,Hˆdp ) is estimated using Bayes Theorem:

(5) ,H|M ,M,M|H

,H

d dd

d

p pp

p

where ,M|Hdp = the daily probability that a fish in the sample tank is marked (M), given that it is

hatchery (H) origin (hereafter, the daily mark rate). Hatchery-origin subyearlings can be either adipose clipped (AD) or Coded Wire-Tagged (CWT), and so ,M|Hdp is the daily fraction of

marked hatchery fish that are either AD or CWT. The parameter ,H|Mdp = the daily probability

that the fish is hatchery-origin, given the fish is marked. If a fish has been marked with either AD or CWT then it is known to be a fish of hatchery origin, and so ,H|Mdp = 1. The parameter ,Mdp =

the daily proportion of fish in the sample tank that are marked. Setting ,H|Mdp = 1 and solving for

,Hdp yields:

(6) ,M,H

,M|H

ˆˆ

ˆd

dd

pp

p

The daily fraction of marked fish in the sample ( ,Mˆdp ) is straightforward to estimate from the

number of marked and unmarked fish counted in the daily sample. However, the fraction of hatchery fish that are marked ( ,M|Hˆ dp ) is more difficult to estimate because marking rates and

release dates vary among hatchery release groups, which causes ,M|Hˆ dp to vary over the

migration season. Given information provided by release group-specific PIT-tagging rates and

4

the passage distribution of each group, ,M|Hˆ dp was estimated using the PIT-tag detection rates

over the release groups and their subsequent detections at the dam to obtain the abundance and relative contribution of the marked and unmarked hatchery groups to the passage of juveniles at the dam.

Abundance of Late Migrants (November of year t through March of year t +1, and April through June year t +1)

If the juvenile fish bypass system and Smolt Monitoring Program at Lower Granite Dam were operated from November of year t through March of year t +1, the method outlined in the previous section of this write up could be applied to the data to estimate smolt abundance at the dam during that period. That was not the case because the juvenile fish bypass was routinely dewatered from November through late March during 1992–2005 and 2007. In 2006, and after 2007, the U.S. Army Corps of Engineers extended the period of bypass water up to include the entire month of November, and a varying portion of the month of December. Efforts were also made to water the bypass up as early as possible in March. The PIT-tag detection system has been operated during those periods of water up, but the sample tank has not been operated or staffed. The present method, which is being refined, starts with the calculation of the PIT-tagging rate of natural-origin subyearlings each year: (7) PIT-tagging rate = n^PIT-tagged natura Oct-Sep/ n^Natural Oct-Sep) where n^PIT-tagged natural Oct-Sep = the number of natural-origin subyearlings that were PIT-tagged upstream of Lower Granite Dam that were subsequently detected during September and October at the dam expanded by collection probability, and n^Natural Oct-Sep = is the estimated abundance of natural-origin subyearlings at Lower Granite Dam during October and September taken from Eqs. 1—6. Passage abundance during November was estimated by dividing the estimated passage abundance of PIT-tagged natural-origin subyearlings in November (number detected divided by collection probability) by the tagging rate from Eq. 7. The same calculation was made to estimate December and March passage abundance, but the estimates for those months includes a step to expand based on the percentage of each of the months the juvenile fish bypass was watered up. Summing the abundance estimates for November, December, and March provides an abundance estimate for late migrants (n^natural-origin Nov–March). Those estimates can be further refined by interpolating passage abundance estimates in January and February, although existing radio-telemetry data indicate that relatively few fish pass Lower Granite Dam during those winter months. Again, the juvenile fish bypass was not operated during November, December, January, February, or early March from 1992 to 2005, or in 2007. In those cases, a regression model was

5

fitted from data collected in 2006, and 2008-2013 to predict n^natural-origin Nov–March from n^Natural

Oct-Sep. Finally, passage abundance for natural-origin subyearlings from April through June of year t +1 was estimated by dividing the estimated number of natural-origin fish PIT-tagged upstream of Lower Granite Dam that passed the dam (number detected divided by collection probability) by the PIT-tagging rate calculated with Eq. 7.

State-Space Life Cycle Model

We formulated a two-stage state-space life cycle model for Snake River fall Chinook salmon by building upon the single-stage state-space framework provided by Fleischman et al. (2013) and Scheuerell et al. (In press). The state-space model consists of two parts: (1) a process model for the underlying state dynamics, and (2) an observation model that links the data to the true underlying state. A key feature of the state-space model is that it separates observation uncertainty (measurement error in the data used to fit the model) and process uncertainty (variation in true underlying population dynamics). The state-space model may also be thought of as a hierarchical model where the state (abundance) evolves according to a population dynamics process model (e.g., a Ricker model) with some process error, and observations on the state (“the data”) are made conditional on the true but unobservable state. Accounting for observation error in a life cycle model for naturally produced Snake River Fall Chinook salmon is absolutely critical because abundance is estimated with considerable observation uncertainty and may vary among years. For example, juvenile abundance estimates at Lower Granite Dam exhibit a coefficient of variation (CV) that ranges from 2-42% among years. Not accounting for this uncertainty could lead to biased parameter estimates and an overestimate of process uncertainty, which can influence the outcomes of population viability analyses.

The State Model

For the adult to juvenile transition, we used the Ricker model to express the number of juveniles passing LGR as a density-dependent function of the number of female adults passing LGR in brood year y (Ricker 1954):

(8) J, F, F, P,J,ln( ) ln( ) ln( )y y y yR E E ,

where J,yR is the true but unobserved number of juvenile recruits produced by female adult

escapement F,yE in brood year y, is the productivity parameter estimating the slope at the

6

origin (juvenile recruits per female spawner), is the density dependence parameter where 1/ estimates the spawner level producing the maximum recruitment (Smax), and J, y is a normally

distributed process error with standard deviation P,J .

For the juvenile to adult transition, we model the number of adult returns as a lognormal function of a density independent smolt-to-adult return rate (SAR): (9) A, J, P,A,ln( ) ln( ) ln(SAR)y y yR R

where A,yR is the number of adult recruits (male and female) produced from the J, yR juveniles

passing LGR that arose from female spawners in brood year y and P,A, y is a normally

distributed.

RESULTS The present method illustrates the following. Early migrants were numerically dominant as was to be expected (Figure 1; panel a).

The proportion of late migrants varied across years as was to be expected (Figure 1; panel b). In all years, hatchery-origin smolts were numerically dominant (Figure 2). There is evidence for a density-dependent relationship between female spawners and natural-origin smolts (Figure 3). That relationship might change as additional years are added to the model.

7

Figure 1.—Preliminary estimates of the abundance of early- and late-migrating, natural-origin smolts at Lower Granite Dam (panel a), and the estimated proportions of each migrant type (panel b), 1992–2013.

8

Figure 2.–Preliminary estimates of the abundance of natural- and hatchery-origin smolts at Lower Granite Dam, 1992–2014.

9

Figure 3.—Preliminary evidence for a density-dependent relationship between female spawners and natural-origin smolts.

10

TAKE HOME MESSAGE

Fitting a two-stage life cycle model was called for in the recovery plan. Such a model will refine the estimates of productivity for informing status reviews and conducting prospective analyses. The presentation of preliminary results made herein showed that much progress has been made with regard to life cycle modeling, especially given the difficulty of estimating abundance of late migrants. Operation of the juvenile fish bypass system at Lower Granite Dam was, and is, the key to estimating abundance of late migrants. In all cases, refined estimates of collection probability will increase the numbers of both natural-origin and hatchery-origin smolts but the patterns observed are not expected to change. After those estimates are refined, the next step will be to develop spawning aggregate specific (i.e., Upper Hells Canyon, Lower Hells Canyon, Clearwater) life cycle models.

REFERENCES

Fleischman, S.J., M.J. Catalano, R.A. Clark, and D.R. Bernard. 2013. An age-structured state-

space stock–recruit model for Pacific salmon (Oncorhynchus spp.). Canadian Journal of Fisheries and Aquatic Sciences 70:401-414.

National Marine Fisheries Service. 1992. Threatened status for Snake River spring/summer

Chinook salmon, threatened status for Snake River fall Chinook salmon. [Docket 910847-2043 22 April 1992] 57(78):14536-14663.

Plumb, J.P., C.M. Moffitt, W.P. Connor, K.F. Tiffan, R.W. Perry, and N.S. Adams. 2012.

Estimating and predicting collection probability of fish at dams using multistate modeling. Transactions of the American Fisheries Society 141:1364-1373.

Ricker, W.E. 1954. Stock and recruitment. Journal of the Fisheries Board of Canada. 11:559-

623. Sandford, B.P., and S.G. Smith. 2002. Estimation of smolt-to-adult return percentages for

Snake River basin anadromous salmonids, 1990–1997. Journal of Agricultural, Biological, and Environmental Statistics, 7:243-263.

Scheuerell, M. D., C. P. Ruff, J, H. Anderson, E. M. Beamer. In press. Estimating density-

dependent population dynamics in a variable environment with imperfect data. Journal of Applied Ecology.

1

Relationship of Juvenile Fall Chinook Salmon Growth and Survival to Food Availability and Abundance

Ken Tiffan and Billy Connor USGS and USFWS

Cook, WA and Ahsahka, ID Connor et al. (2013) and Connor and Tiffan (2012) analyzed density-dependent changes in parr growth between periods of low and high fish abundance and explored growth as a factor affecting parr to smolt survival. Parr growth in riverine rearing areas declined from the low abundance period to the high abundance period suggesting that density dependence was influencing growth. Growth in the upper reach of Hells Canyon declined from a mean±SE of 0.29±0.02 g/d during the low abundance period to 0.18±0.02 g/d during the high abundance period. Growth in the lower reach of Hells Canyon declined from 0.19±0.03 g/d during the low abundance period to 0.15±0.01 g/d during the high abundance period. Smolt growth (calculated at Lower Granite Dam) also declined from the low to high abundance periods and the reduction was even more pronounced than in the river. Smolt growth during the low abundance period declined from 0.6±0.1 g/d to 0.2±0.1 g/d in the high abundance period. Because these growth rates potentially included some riverine growth, Tiffan et al. (2014) compared riverine to reservoir growth of PIT-tagged fish tagged and recaptured exclusively within each area. They also found growth to be lower in the reservoir than in the river. Growth in length was about 0.3 mm/d lower in the reservoir than in the river and growth in standardized mass was about 0.02 g/g/d lower in the reservoir. There are a number of mechanisms that may be affecting parr and smolt growth. For parr, the increase in the juvenile population may be imposing density-dependent competition for food and space in rearing areas, but this has not been investigated to date. For smolts growing in the reservoir, Connor et al. (2013) showed positive relationships between smolt growth and fish weight at capture, days at large, and reservoir temperature. However, there was a negative relationship between smolt growth and basin-wide smolt abundance. Again, this suggests that competition in the reservoir could be affecting growth. This is supported by recent stable isotope information collected by the USGS. Stable isotopes of nitrogen can be used to infer trophic position and the potential for competition. Fish occupying the same trophic position have a greater likelihood of competing for food than those that do not. In riverine habitats, only juvenile mountain whitefish are at a similar trophic position to juvenile fall Chinook salmon, but mountain whitefish are not very abundant. In contrast, in Lower Granite Reservoir, pumpkinseeds, bluegills, peamouth, juvenile northern pikeminnow, juvenile smallmouth bass, and sculpins are very abundant and are all at a similar trophic position to juvenile fall Chinook salmon. Thus, there is greater potential for competition with resident fishes, in addition to conspecifics, in the reservoir. In addition, Tiffan et al. (2014) explored food availability and quality as mechanisms of differential growth in riverine and reservoir habitats. They found that the biomass of invertebrates available to fall Chinook salmon in the drift was lower in the

2

reservoir than in the river. They also found that the energy content of prey items and the proportion of terrestrial insects in diets were higher in the river than in the reservoir suggesting better prey resources existed in riverine habitat. Interestingly, the non-native opossum shrimp Neomysis mercedis comprised the majority of the epibenthic biomass in the reservoir. Paradoxically, this should be a valuable prey for juvenile fall Chinook salmon, and indeed Neomysis constitute the majority of the diet at times, but their abundance does not translate to higher fish growth. The effect of biological and environmental variables on fall Chinook salmon smolt survival was first demonstrated by Connor and Tiffan (2012). They showed a positive, linear relationship between smolt survival and parr growth. Thus, if density dependence reduces parr growth, then it is probable that smolt survival would be reduced as well. They also showed that a positive, linear relationship between smolt survival and water velocity in Lower Granite Reservoir. This supports current management actions to increase river flows, and hence, water velocity. A more thorough presentation of methods, results, and discussion the topics discussed above can be found in the following peer-reviewed papers. References Connor, W.P., and K.F. Tiffan. 2012. Evidence for parr growth as a factor affecting parr-to-

smolt survival. Transactions of the American Fisheries Society 141:1207-1218. Connor, W.P., K.F. Tiffan, J.M. Plumb, and C.M. Moffitt. 2013. Evidence for density-

dependent changes in growth, downstream movement, and size of Chinook salmon subyearlings in a large-river landscape. Transactions of the American Fisheries Society 142:1453-1468.

Tiffan, K.F., J.M. Erhardt, and S.J. St. John. 2014. Prey availability, consumption, and quality

contribute to variation in growth of subyearling Chinook Salmon rearing in riverine and reservoir habitats. Transactions of the American Fisheries Society 143:219-229.

1

Smallmouth Bass Predation on Subyearling Fall Chinook Salmon Ken Tiffan and John Erhardt

U.S. Geological Survey, Cook, WA

Studies of smallmouth bass predation on subyearling Snake River fall Chinook salmon were initiated in 2012 to reevaluate the magnitude of predation since recovery measures were implemented. Previous studies of smallmouth bass predation in Hells Canyon (Nelle 1999) and in Lower Granite Reservoir (Anglea 1997; Naughton et al. 2004) were conducted in the mid-1990s when subyearling abundance was low following ESA listing. These studies showed that subyearling loss to bass predation was low at that time. Since then, the population of subyearlings in the Snake River basin has increased substantially due to a variety of recovery measures. We duplicated the studies of the aforementioned investigators following peer-reviewed methods to make our contemporary results as comparable as possible to these studies. Results are summarized here at a high level and it is important to recognize that the data include a great deal of complexity owing to a wide range of spatial and temporal variation in our sampling and the biological metrics we investigated.

Predation loss The loss of subyearlings to bass predation is determined by multiplying empirical consumption rates by estimates of bass abundance over a sampling interval. Loss estimates are sensitive to abundance estimates which can be difficult to obtain for an open population in a large study area. Within Hells Canyon during 2013-2014, subyearling losses in the upper reach (annual mean = 248,135) were two and half times higher than in the middle reach (annual mean = 106,367) and about three times higher than in the lower reach (annual mean = 79,246; Table 1). By comparison, Nelle’s (1999) loss estimates (annual mean = 947–2,941) were about two orders of magnitude lower in the mid-1990s. Nelle’s lower reach was synonymous with our middle and lower reaches combined. Loss was estimated in three sections of the upper end of Lower Granite Reservoir. Two of them were the same sections studied by Naughton et al. (2004). In the SRTZ (Snake River transition zone) where the Snake River transitions from free-flowing to an impounded state, mean annual loss of subyearlings was 47,777 fish (Table 1). By comparison, Naughton estimated a mean annual loss of 2,152 fish. In the CLTZ (Clearwater River transition zone) where the Clearwater River transitions from free-flowing to an impounded state, mean annual loss of subyearlings was 7,674 fish (Table 1). By comparison, Naughton estimated a loss of 1,834 fish, and all this occurred on one sampling event. In addition, some of these fish may have been yearling Chinook salmon as he did not distinguish between yearlings and subyearlings. We also estimated a mean loss of 73,564 subyearlings in the CON reach which extended from the confluence downstream 7 km to the Port of Wilma (Table 1). This reach was not sampled by Naughton et al. (2004).

2

Table 1. Estimated mean (range) annual losses of subyearling Snake River fall Chinook salmon to smallmouth bass predation during 2012–2016 and 1994–1997 in different study reaches of the Snake River.

Area and reach Contemporary loss Historical loss

Hells Canyon 2013–2014 1996–1997 Upper reach 248,135 (240,385–255,886) 2,941 (714–5,798)

Middle reach 106,367 (65,600–147,134) 947 (0–2,651)

Lower reach 79,246 (33,013–125,480)

Lower Granite Reservoir 2012–2015 1996–1997

SRTZ 47,777 (19,647–80,158) 2,152 (1,200–3,104)

CLTZ 7,674 (4,115–11,233) 1,834 (all loss on one event)

CON 73,564 (64,119–88,951) not sampled

Lower Granite Reservoir 2016 1994–1995 Upper reservoir 71,408 (2,818–43,980)a 33,611 (19,588–47,635)

Middle reservoir 71,688 (542–56,282)a 17,375 (13,944–20,807)

Lower reservoir 129,437 (11,747–54,719)a 13,211 (8,261–18,162) a means and ranges are for biweekly sampling intervals

In 2016, we began sampling the remainder of Lower Granite Reservoir downstream from the Port of Wilma. At this time, the abundance estimates used to estimate loss are crude and results should be considered preliminary. Subyearling loss increased in a downstream direction from the upper (71,408 fish) to lower (129,437 fish) reservoir (Table 1). Anglea (1997) found the opposite trend in which loss decreased from the upper (33,611 fish) to lower (13,211 fish) reservoir (Table 1). In Lower Granite Reservoir, we found that most subyearlings were consumed by bass <250 mm TL. In all reaches we examined, subyearling loss to smallmouth bass predation increased substantially since the mid-1990s. Since loss is estimated from both bass abundance and consumption rate, changes in these two factors may explain increases in loss. Bass abundance In Hells Canyon during 2013–2014, absolute smallmouth bass abundance was highest in the upper reach (annual mean = 62,839) followed by the middle (annual mean = 25,418) and lower (annual mean = 21,532) reaches (Table 2). However, on a river kilometer (rkm) basis,

3

Table 2. Estimated mean (range) abundance of smallmouth bass in different study reaches of the Snake River during 2012–2016 and 1994–1997.

Area and reach Contemporary abundance Bass/rkm Historical abundance Bass/rkm

Hells Canyon 2013–2014 1996–1997 Upper reach 62,839 (29,729–114,329) 458 17,458 (266–32,912) 406

Middle reach 25,418 (10,685–66,125) 819 29,040 (412–55,007) 440

Lower reach 21,532 (7,757–78,485) 566


SRTZ 3,050 (1,004–7,059) 305 11,877 (8,818–16,002) 990

CLTZ 531 (143–2,170) 115 3,820 (2,328–6,283) 637

CON 3,726 (1,858–8,219) 532 not sampled

Lower Granite Reservoir 2016 1994–1995

Upper 11,091 (7,417–14,793)a 591

20,911 406 Middle 12,520 (7,724–19,554)a 822

Lower 25,057 (18,832–27,728)a 1,569 a means and ranges are for biweekly sampling intervals abundance was highest in the middle reach. By comparison, Nelle (1999) estimated 17,458 bass in the upper reach and 29,040 bass in the combined middle and lower reaches (Table 2). It is important to note that we estimated bass abundance concurrent with our diet sampling in the spring whereas Nelle estimated abundance in the fall after sampling diets in June and July, which may explain some of the differences between the two periods (i.e., 1996–1997 and 2013–2014). In Lower Granite Reservoir, bass abundance in the SRTZ during 2012–2015 (annual mean = 3,050 fish) was about one-fourth of that estimated by Naughton et al. (2004) (11,877 fish) during 1996–1997 (Table 2). In contrast, we only estimated an annual mean of 531 bass in the CRTZ whereas Naughton estimated 3,820 bass. Slightly warmer water temperatures in the CRTZ during the Naughton study could explain the difference. Within the lower portion of Lower Granite Reservoir, we estimated at total of 48,668 bass (950 bass/rkm) in 2016 whereas Anglea (1997) estimated a total of 20,911 bass (406 bass/rkm) during 1994–1995 (Table 2). Part of this difference is due to the fact that we estimated abundance for bass ≥150 mm TL whereas Anglea’s estimate was for bass ≥175 mm TL. Smallmouth bass abundance has increased in some study areas but not others since the mid-

4

1990s. Hells Canyon and the lower portion of Lower Granite Reservoir support the greatest numbers of smallmouth bass. Bass consumption rates The most likely factor responsible for the increase in subyearling loss to predation is the increase in bass consumption rates in response to greater prey availability. In Hells Canyon during 2012–2016, mean annual consumption rates ranged from 0.05 to 0.08 Chinook/bass/day with peak consumption ranging from 0.13 to 0.23 Chinook/bass/day (Table 3). In contrast, Nelle (1999) reported mean consumption rates that ranged from 0.0015 to 0.003 Chinook/bass/day and a peak rate of 0.027 Chinook/bass/day. In the upper portion of Lower Granite Reservoir, we found consumption rates (0.08–0.16 Chinook/bass/day) were over an order of magnitude higher than the 0.002–0.004 Chinook/bass/day reported by Naughton et al. (2004; Table 3). The highest consumption rates we observed were for bass sampled in the lower portion of Lower Granite Reservoir, which ranged from 0.21 to 0.35 Chinook/bass/day. These are substantially higher than those reported by Anglea (1997), which ranged from 0.007 to 0.017 salmonids/bass/day (Table 3) that included mainly Chinook salmon, but potentially some steelhead as well. Table 3. Estimated mean (peak) consumption rates (Chinook/bass/day) of smallmouth bass in different study reaches of the Snake River during 2012–2016 and 1994–1997.

Area and reach Contemporary consumption Historical consumption

Hells Canyon 2013–2014 1996–1997 Upper reach 0.07 (0.18) 0.003 (0.027)

Middle reach 0.08 (0.23) 0.0015 (0.009)

Lower reach 0.05 (0.13)


SRTZ 0.08 (0.30) 0.004 (0.02)

CLTZ 0.16 (0.62) 0.002 (0.002)

CON 0.05 (0.25) not sampled

Lower Granite Reservoir 2016 1994–1995 Upper reservoir 0.35 (0.88)a 0.010 (0.032)b

Middle reservoir 0.21 (0.57)a 0.007 (0.018)b

Lower reservoir 0.25 (0.44)a 0.017 (0.027)b a Means and ranges are for biweekly sampling intervals. b Consumption rates were for salmonids which included mainly Chinook salmon but potentially some steelhead as well.

5

Hatchery releases In 2016, we evaluated smallmouth bass predation before and after releases of hatchery subyearlings to correct potential bias in earlier loss estimates that did not account for variable consumption rates within our biweekly sampling intervals. In the upper reach of Hells Canyon, consumption rates increased 9–10-fold the day after releases from Pittsburg Landing and Hells Canyon Dam, but consumption rates returned to prerelease levels in 1–2 days. Smallmouth bass consumed and estimated 3.6% of the Hells Canyon Dam release and 14.8% of the Pittsburg Landing release in 2016. Bass consumption rates increased 6-fold following the subyearling release at Captain John acclimation facility and returned to prerelease levels in 5 days. We estimate that 6% of this release was lost to predation in 2016. References Anglea, S.M. 1997. Food habits, abundance, and salmonid fish consumption of smallmouth

bass and distribution of crayfish in Lower Granite Reservoir, Idaho-Washington. Master’s thesis. University of Idaho, Moscow.

Naughton, G.P., D.H. Bennett, and K.B. Newman. 2004. Predation on juvenile salmonids by

smallmouth bass in the Lower Granite Reservoir system, Snake River. North American Journal of Fisheries Management 24:534–544.

Nelle, R.D. 1999. Smallmouth Bass predation on juvenile fall Chinook Salmon in the Hells

Canyon Reach of the Snake River, Idaho. Master’s Thesis, University of Idaho, Moscow.

1

Determining natal rearing locations and ocean entry timing of untagged adult fish based on otolith microchemistry

Jens Hegg1, Brian Kennedy1, Billy Connor2

University of Idaho1, USFWS2

Moscow, ID; Orofino, ID

INTRODUCTION

Juvenile life-history decisions can have major impacts on the survival of juvenile salmonids and consequently can influence population dynamics. Juvenile Snake River fall Chinook salmon exhibit multiple in-river movement timings, culminating in sub-yearling and yearling outmigration strategies (Connor et al. 2003, 2005). These strategies appear to be spatially explicit within spawning areas in the basin and related to temperature variation across the basin, setting the stage for selection based on life-history traits (Connor et al. 2002, Hegg et al. 2013). Evidence indicates that these life histories are under strong selection and at least partially heritable, such that their representation in the population may not be simply a plastic response, but may be the result of recent evolution (Williams et al. 2008, Waples et al. 2017). Thus, understanding the representation of these strategies, both in juveniles and in adults surviving to spawn, is important to understanding and managing the population as a whole.

Otolith microchemistry has been shown to be a useful method of determining location and movement in juvenile Snake River fall Chinook Salmon in the past (Hegg et al. 2013). The daily increments laid down in the otolith record trace amounts of multiple chemical tracers which are indicative of the river in which the fish resides. Our work has shown that fish originating in the Snake River basin can be classified to six distinct regions; the Clearwater River (CW), the free-flowing Snake River (SK), Lower Granite pool (LGR), Lyons Ferry Hatchery (LFH), Nez Perce Tribal Hatchery (NPTH), and a group including the Grande Ronde, Imnaha and Tucannon Rivers (GR).

Otolith microchemistry is also able to distinguish when a juvenile fish enters the ocean, due to the large increase in strontium concentration in ocean water around the globe in relation to fresh water (Walther and Limburg 2012). We have previously shown that ocean entry can be determined in Snake River fall Chinook salmon using the sharp increase in strontium intensity in the otolith when the fish enters the ocean (Hegg et al. 2013). In combination with our ability to determine natal location using multiple otolith tracers, this allows us to examine the differences in the relative size of fish from different areas when they enter the ocean.

2

METHODS

We used the multivariate discriminant function developed from known origin juveniles (detailed in the previous section) to predict the natal location of a large dataset of returning adult fish (n=591). The dataset was made up of unmarked, untagged fish used as broodstock at Lyons Ferry hatchery from 2006 to 2013. Thus, the sample should be a roughly random subsample of the returning adult population above the length cutoff for jacks.

For each fish a mean was taken for each of five microchemical tracers between 300-400m from the core of the otolith. This period represents the earliest stable signature on the otolith and should represent the natal location in most fish based on results from the previous section. These means were then used as predictors in the previously developed discriminant function, outputting predicted natal locations for each fish.

The location of ocean entry on each otolith was determined using a CUSUM changepoint algorithm on the strontium intensity measured on the MC-ICP-MS during 87Sr/86Sr analysis. This determination was then checked visually and manually adjusted to correspond to the point during the increase in strontium intensity at which 87Sr/86Sr converged to the global ocean signature of 0.70918 (Faure and Mensing 2004).

The proportion of returning adult fish originating from each natal location was plotted, both by hatch year and return year. Finally, the location of ocean entry was plotted for fish in each hatch year, as well as for combined trends across the entire dataset.

RESULTS

Applying the discriminate function to unknown fish showed several interesting results (Fig. 1). First, the grouping of LFH and NPTH fish are very tight, indicating that classification of hatchery fish appears to be fairly specific. High dispersion in the classification to CW origin is likely an effect of the fact that the variables are unweighted, despite the higher specificity of 87Sr/86Sr to location. Dispersion in the classification to the Snake River may indicate that some natal averages were captured before they had equilibrated, with fish moving from the Grande Ronde into the Snake, as well as overlap with the LGR group.

3

Figure. 1 - Classification of unknown origin, returning adult fish using a multivariate discriminate function trained on known-origin juvenile fish. Low variation in the LFH and NPTH hatchery groups indicates good classification accuracy. High variance in the CW and SK groups may be a result of the weighting of predictors, as well as fish whose signatures are not equilibrated after recent movement.

The predicted natal locations of fish returning to spawn indicates an increase in hatchery representation, as well as a decrease in returning adults from the Clearwater River (FIg. 2). Sample sizes for early hatch years are low, limiting the accuracy of these proportions. The proportion of fish classified to each natal location was also plotted for each return year (FIg. 3).

4

Figure. 2 - The proportion of returning adults classified to each natal location for the hatch years included in the dataset. Sample sizes for 2002, 2003, and 2011 are low, limiting the accuracy of their proportions. Sample size for each year is shown at the top of the plot.

5

Figure. 3 - The proportion of returning adults classified to each natal location for return years included in the study. Sample size for each year is shown at the top of the plot.

After removing fish classified to either LFH or NPTH hatchery, the data shows year-to-year variation in the representation of adults returning to spawn in comparison to the relatively stable proportion of redds across the same years (Fig. 4). Of particular interest is the variation in early-moving juveniles classified to Lower Granite pool, and a general decrease in representation from the Clearwater River.

6

Figure 4 - The proportion of returning adult fish from each natal location, with hatchery-origin fish excluded (right-hand plot) shown much more variation than the proportion of redds over the same period (left-hand plot). Sample size for each year of otolith data is shown at the top of the plot. Redd data from Idaho Power (https://www.idahopower.com/AboutUs/Sustainability/Stewardship/reddcount.cfm).

Ocean entry was also highly variable over the hatch years represented in the study (Fig. 5). In years fish originating in the Snake River, hatchery fish, and early-moving juveniles left at smaller otolith radius’ than fish from the Clearwater River. The 2007 hatch year (Fig. 5)is one example of significant variation in the relative size at which fish entered the ocean.

7

Figure 5 - A stacked histogram shows otolith radius at which fish entered the ocean. The results show variation across hatch years, with Clearwater fish leaving later in general, but variation both within natal location and between hatch years.

Across all hatch years fish originating in the Clearwater exhibited uch larger otolith radii at ocean entry (Fig. 6). It should be stressed that otolith radius is related to size, but masks any differences in hatch date or specifics of growth rate between locations. Thus, otolith radius should not be interpreted as ocean entry timing, but rather a rough approximation of size at ocean entry. Future work will work to correct otolith radius for differences in growth and hatch date to arrive at an estimate of relative ocean entry timing across natal locations.

8

Figure 6 - Panels show the relative size at which returning adult fish entered the ocean as juveniles from each natal location across all hatch years. Estimated density curves (lines) show peaks in ocean entry, histograms represent the underlying data (colored bars).

SUMMARY AND IMPLICATIONS

The ability of otolith microchemistry to identify the juvenile location and ocean entry timing of returning adult fall Chinook salmon provides important information on life-history in naturally produced fish. The low dispersion within the hatchery groups indicates that multi-tracer otolith

9

chemistry appears capable of identifying unmarked/untagged fish when applied to unknown returning adult fish. Preliminary analysis indicates that the proportion of returning adult fish from each hatchery and major spawning reach varies year-to-year, with a general trend toward a larger proportion of hatchery fish since hatch year 2008 and return year 2011. In comparison to relatively stable proportions of redds in the basin otolith analysis shows variable representation from the Clearwater, Snake, and early-moving juveniles classified to Lower Granite pool. Ocean entry analysis is preliminary, and does not correct for growth or hatch date to provide true ocean entry timing, however, otolith radius is directly linked to fish size. Ocean entry timing shows interesting year-to-year variation, with Clearwater fish most often entering the ocean at a larger size than fish originating in the Snake River, hatchery, or early-moving juveniles classified to lower granite pool.

REFERENCES

Connor, W. P., H. L. Burge, R. Waitt, and T. C. Bjornn. 2002. Juvenile life history of wild fall Chinook salmon in the Snake and Clearwater rivers. North American Journal of Fisheries Management 22:703–712.

Connor, W. P., J. G. Sneva, K. F. Tiffan, R. K. Steinhorst, and D. Ross. 2005. Two alternative juvenile life history types for fall Chinook salmon in the Snake River basin. Transactions of the American Fisheries Society 134:291–304.

Connor, W. P., R. K. Steinhorst, and H. L. Burge. 2003. Migrational behavior and seaward movement of wild subyearling fall Chinook salmon in the Snake River. North American Journal of Fisheries Management 23:414–430.

Faure, G., and T. M. Mensing. 2004. Isotopes: principles and applications. John Wiley & Sons Inc.

Hegg, J. C., B. P. Kennedy, P. M. Chittaro, and R. W. Zabel. 2013. Spatial structuring of an evolving life-history strategy under altered environmental conditions. Oecologia 172:1017–1029.

Walther, B. D., and K. E. Limburg. 2012. The use of otolith chemistry to characterize diadromous migrations. Journal of Fish Biology 81:796–825.

Waples, R. S., A. Elz, B. D. Arnsberg, J. R. Faulkner, J. J. Hard, E. Timmins-Schiffman, and L. K. Park. 2017. Human-mediated evolution in a threatened species? Juvenile life-history changes in Snake River salmon. Evolutionary Applications.

Williams, J. G., R. W. Zabel, R. S. Waples, J. A. Hutchings, and W. P. Connor. 2008. Potential for anthropogenic disturbances to influence evolutionary change in the life history of a threatened salmonid. Evolutionary Applications 1:271–285.

1

Distinguishing natural- and hatchery-origin individuals using multiple chemical signatures in otoliths

Jens Hegg1, Brian Kennedy1, Billy Connor2

University of Idaho1, USFWS2

Moscow, ID; Seattle, WA; Orofino, ID; Cook, WA

INTRODUCTION

Understanding the origin of juvenile Snake River fall Chinook salmon is of particular importance for understanding outmigration life-history. Prior work has shown that the population has recently developed a novel, late-migrating, juvenile life-history strategy, presumably in response to anthropogenic changes to the river system (Connor et al. 2005, Williams et al. 2008). This later, stream-type, outmigration strategy has been linked to heterogeneous temperature conditions in the river, with the majority of late migrants originating from the Clearwater River (Connor et al. 2002, Hegg et al. 2013). Further, evidence indicates that growth rates and the propensity to migrate may be hereditary and evidence of evolutionary adaptation to anthropogenic change (Williams et al. 2008, Waples et al. 2017).

Understanding the factors driving these changes in juvenile life-history, and their impacts on the population, requires the ability to determine the natal location of juvenile fish as well as that of returning adults, both of hatchery origin as well as of important spawning areas within the basin. While adipose fin clips and coded wire tagging positively identifies the majority of hatchery fish, significant numbers of returning hatchery fish lack these marks and tags. In the past the hatchery or wild origin of unmarked, untagged adults was determined using scale analysis. Recent validation of scale analysis, however, has revealed an unacceptable level of error in the determination of hatchery or wild origin.

Our prior work has shown that the natal location and juvenile movement of fall Chinook salmon can be traced using otolith microchemistry (Hegg et al. 2013). These prior studies have relied solely on 87Sr/86Sr isotope ratios, a chemical signature with strong links to river location. We used a suite of four additional elemental tracers, as well as a larger, multi-year dataset of known origin juveniles and returning adults, to determine whether hatchery and wild origin could be successfully determined using otolith microchemistry. In addition, we evaluated the ability of multiple chemical tracers to increase the number of locations within the range of Snake River fall Chinook salmon that could be distinguished chemically.

This work required an in-depth analysis of the precise location on an otolith where maternally derived signatures predictably transition to chemistry derived from the natal river. Following the

2

determination of the location of stable natal signatures on the otolith, we used a large database of known-origin juvenile fish to create a discriminate function and tested its ability to correctly classify new samples.

METHODS

Water samples were collected from 2008 through 2016 throughout the spawning range of Snake River fall Chinook salmon to characterize the spatial and temporal variation in strontium isotopic chemistry within the basin (Fig. 1) as well as to characterize relative differences in elemental chemistry (unpublished data). Samples were collected during low flow periods in the fall in all years to capture the most representative signatures within each river. Starting in 2009, as resources permitted, samples were taken seasonally. Annual sampling began in the spring as soon as flows were safe to sample and included summer, and fall seasons to characterize the monthly stability of the signatures. Additionally, in 2010, samples were taken in the Clearwater and Salmon Rivers at three periods encompassing the peak of the hydrograph to characterize seasonal variation observed in prior studies (Hegg et al. 2015).

3

Figure 1 - Boxplot of water samples from throughout the basin show chemically distinguishable groups based on differences in 87Sr/86Sr isotope ratios.

Wild juvenile fish (n=430) were captured using beach seines from spawning areas between in the Snake and Clearwater Rivers as part of seasonal population survey efforts between April and August in years spanning 2007 to 2014 (Connor et al. 1998). Some of these fish (n=111) had been PIT tagged, released, and then recaptured and sacrificed at fish passage facilities at Lower Granite dam on the Snake River during their downstream out-migration, providing two known locations for these fish. Juvenile fish were also collected from the two hatcheries that produce fall Chinook salmon in the Snake River basin. Hatchery fish were collected from Lyons Ferry Hatchery in 2009 and 2011 (n=28), and from Nez Perce Tribal Hatchery in 2011 and 2012 (n=35).

Otoliths were analyzed for 87Sr/86Sr ratios and trace elemental concentrations of strontium (Sr), barium (Ba), Magnesium (Mg), and Manganese (Mn), each expressed as a ratio to Calcium (Ca). For each analysis method otoliths were analyzed using a transect moving from the edge of the otolith to the core.

To determine the location of a stable natal signature, transects were analyzed graphically to determine the location of a change between maternally derived and post-hatch chemistry. To determine which of the five chemical signatures were most indicative of the maternal/natal change, mean signatures 100µm before the selected location were compared to the mean signature 100µm after using a two-sided, paired t-test assuming unequal variance (α = 0.05) with Bonferonni correction for multiple comparisons within each river group.

Once the location of the stable natal signature was determined we took the mean of each chemical tracer in the 100µm section immediately following the onset of the stable natal signature on each otolith. For fish which were PIT tagged, released and recaptured at Lower Granite Dam we also took the mean of the final 50µm at the edge of the otolith, representing the period just prior to recapture. The values of these control fish were paired with their known location of each fish to provide a training set of juvenile natal signatures.

This dataset was further subset, with 80% of the samples being randomly assigned to a training-set used to develop a discriminate function, and the remaining 20% assigned to a test-set used to validate the model. The {mclust} package in R was used to develop a model-based, multivariate discriminate function which classified fish to each of the locations within the basin. The discriminate function was then applied to the test-set to determine the classification ability of the model for external data. The 10-fold cross-validation error rate was then calculated for the training set and the overall error rate was calculated for the test-set.

RESULTS

4

Graphical analysis indicated that 87Sr/86Sr signatures began to change at 150µm and reached a stable equilibrium at 250-300µm depending on the natal location of fish (Fig. 2).

Figure 2 - Strontium isotope ratio (87Sr/86Sr) changes across the otoliths of known-origin juvenile fish from the Clearwater River (above blue line), and Grande Ronde River (Below blue line), show the onset of change in the juvenile fish at ~150µm (Green bar) followed by equilibration to the chemistry of the natal stream (Blue bars) at 250µm and 300µm respectively.

Meanwhile, elemental signatures, Mn/Ca and Ba/Ca in particular, showed a statistically significant change at 225µm (Fig. 3).

5

Figure 3 - Comparison of elemental signatures before and after 225µm showed a significant change in elemental signature between maternal and natal signatures. * indicates significant difference (paired T-test, α=0.05 with Bonferroni correction).

Our results suggest that strontium isotope ratios in juvenile otoliths begin to change ~75µm earlier than do elemental signatures, a distance which corresponds to roughly 12 - 19 days based on the range of known Chinook growth rates in the basin (Zabel et al. 2010). It is reasonable, based on this difference in timing, to assume that the shift in elemental and 87Sr/86Sr ratios are synchronized with different ontogenetic changes in the juvenile fish. Prior research has shown that 87Sr/86Sr ratios in Chinook salmon in California’s Central Valley attain a stable, natal signature at 250µm, similar to the equilibrium values expressed in our results (Barnett-Johnson et al. 2008).

Since hatching represents the first time the egg is capable of a large degree of ion exchange with the surrounding water, the change in 87Sr/86Sr ratios at ~150µm likely represents hatching. This is supported by evidence that strontium and calcium uptake in juvenile fish begins to climb steadily after hatching, and experimental results indicate that it is possible to isotopically mark

6

non-feeding salmonid fry using water spiked with 84Sr (Hayes et al. 1946, Yamada and Mulligan 1987, De Braux et al. 2014).

Previous research indicates that the change in elemental ratios at 225µm may indicate the onset of exogenous feeding. This is supported by experimental evidence indicating that the rate of strontium intake increases to an even faster rate following first-feeding (Yamada and Mulligan 1987), and that magnesium concentration also increases 12-15 days after juveniles hatch (Hayes et al. 1946).

Taken together our data suggest that both elemental ratios and 87Sr/86Sr ratios provide information on the duration of maternally derived chemical influence on the otolith. However, the change from maternal to natal chemistry appear to correspond to different ontological stages. Changes in 87Sr/86Sr ratios appear to correspond to the hatching of the larval fish, with equilibration continuing until sometime after the onset of exogenous feeding. Elemental ratios of manganese, barium, and strontium appear to reflect the onset of exogenous feeding at ~225µm.

Figure 4 - Natal signatures (87Sr/86Sr ratios) of known origin fish, plotted by location, show that Upper and Lower Snake River juvenile groups (USK & LSK) overlap and difficult to distinguish.

7

Fish with signatures consistent with the Upper Snake River were captured in the lower section of the free-flowing Snake River as early as April, indicating early downstream movement.

Based on this analysis, we used a mean between 300-400µm as a measure of the natal signature for each sampled fish, as this is the first location at which all tracers, in all fish, had reached a stable signature. This revealed that fish from the Upper and Lower Snake were indistinguishable, with fish from the Lower Snake exhibiting signatures indicative of both the Upper and Lower sections of the free-flowing Snake River (Fig. 4). Therefore, we combined the Upper and Lower Snake River groups, resulting in a grouping including the Clearwater River (CW), free flowing Snake River (SK), Lyons Ferry Hatchery (LFH), New Perce Tribal Hatchery (NPTH), the Grand Ronde/Imnaha/Tucannon Rivers (GR), and Lower Granite Pool (LGR). A {mclust} discriminate function on the training-set resulted in a 10-fold cross-validation error rate of 15% +/-2% (Table). This model, applied to the test set, resulted in a 9% overall misclassification error rate.

Predicted

CW GR/Imnaha/

Tucannon

LFH LGR NPTH SK

Obs

erve

d

Obs

erve

d

CW 64 0 0 0 0 0

GR/Imnaha/

Tucannon 0 21 0 0 0 0

LFH 0 0 19 0 2 0

LGR 2 0 0 60 0 9

NPTH 2 0 2 0 21 5

SK 0 0 0 2 1 92

Table 1 - Classification accuracy of a training set of the data for known-origin juvenile fish resulted in 15% +/- 2% 10-fold cross-validation classification error rate (n=298). The test set

8

(n=72) resulted in a 9% overall error rate. Green cells indicate correct classifications, yellow cells highlight incorrect classifications.

SUMMARY AND IMPLICATIONS

The study allowed us to expand otolith microchemistry methods as a result of a large database of known origin fish and detailed analysis of the transition between maternal and natal chemical signatures on the otolith. Collectively, our approach has improved our ability to discriminate juvenile natal location. Analysis of the transition from maternal to natal chemical signatures has provided insight into efforts to determine when chemistry best reflects the natal river, rather than chemistry inherited from the mother. Using model-based multivariate discriminate analysis we are able discriminate juveniles originating from both Lyons Ferry Hatchery and Nez Perce Tribal Hatchery. In addition, wild juveniles originating in the Clearwater River, as well as the free-flowing Snake River can be distinguished chemically, and fish moving downstream into Lower Granite Reservoir also exhibit a distinguishable signature. Early downstream movements of juveniles, and the consequent incorporation of otoliths signatures from throughout the system, appear to limit the discrimination of the Snake River above and below the confluence with the Salmon River.

REFERENCES

Barnett-Johnson, R., T. E. Pearson, F. C. Ramos, C. B. Grimes, and R. B. MacFarlane. 2008. Tracking natal origins of salmon using isotopes, otoliths, and landscape geology. Limnology and Oceanography 53:1633–1642.

De Braux, E., F. Warren-Myers, T. Dempster, P. G. Fjelldal, T. Hansen, and S. E. Swearer. 2014. Osmotic induction improves batch marking of larval fish otoliths with enriched stable isotopes. ICES Journal of Marine Science 71:2530–2538.

Connor, W., H. Burge, and D. Bennett. 1998. Detection of PIT-tagged subyearling chinook salmon at a Snake River dam: implications for summer flow augmentation. North American Journal of … 18:530–536.

Connor, W. P., H. L. Burge, R. Waitt, and T. C. Bjornn. 2002. Juvenile life history of wild fall Chinook salmon in the Snake and Clearwater rivers. North American Journal of Fisheries Management 22:703–712.

Connor, W. P., J. G. Sneva, K. F. Tiffan, R. K. Steinhorst, and D. Ross. 2005. Two alternative juvenile life history types for fall Chinook salmon in the Snake River basin. Transactions of the American Fisheries Society 134:291–304.

Hayes, F. R., D. A. Darcy, and C. M. Sullivan. 1946. Changes in the Inorganic Contituents of Developing Salmon Eggs. Journal of Biological Chemistry 163:621–632.

9

Hegg, J. C., T. Giarrizzo, and B. P. Kennedy. 2015. Diverse Early Life-History Strategies in Migratory Amazonian Catfish: Implications for Conservation and Management. bioRxiv (in-review, PloS One) pre-print:1–37.

Hegg, J. C., B. P. Kennedy, P. M. Chittaro, and R. W. Zabel. 2013. Spatial structuring of an evolving life-history strategy under altered environmental conditions. Oecologia 172:1017–1029.

Waples, R. S., A. Elz, B. D. Arnsberg, J. R. Faulkner, J. J. Hard, E. Timmins-Schiffman, and L. K. Park. 2017. Human-mediated evolution in a threatened species? Juvenile life-history changes in Snake River salmon. Evolutionary Applications.

Williams, J. G., R. W. Zabel, R. S. Waples, J. A. Hutchings, and W. P. Connor. 2008. Potential for anthropogenic disturbances to influence evolutionary change in the life history of a threatened salmonid. Evolutionary Applications 1:271–285.

Yamada, S. B., and T. J. Mulligan. 1987. Marking Nonfeeding Salmonid Fry with Dissolved Strontium. Canadian Journal of Fisheries and Aquatic Sciences 44:1502–1506.

Zabel, R., K. Haught, and P. Chittaro. 2010. Variability in fish size/otolith radius relationships among populations of Chinook salmon. Environmental biology of fishes 89:267–278.

Human-mediated evolution in a threatened species? Juvenile life-history changes in Snake River salmon Robin S. Waples1, Anna Elz1, Billy D. Arnsberg2, James R. Faulkner1, Jeffrey J. Hard1, Emma Timmins-Schiffman1,3, and Linda K. Park1 1NoAA Fisheries, Seattle, WA, 98112 USA 2Nez Perce Tribe, Department of Fisheries Resources Management, Lapwai, Idaho 83540 USA 3Department of Genome Sciences, University of Washington, Seattle, WA, 98105 USA Evolutionary Applications (in press); accepted manuscript available online at: http://onlinelibrary.wiley.com/doi/10.1111/eva.12468/full Abstract Evaluations of human impacts on Earth’s ecosystems often ignore evolutionary changes in response to altered selective regimes. Freshwater habitats for Snake River fall Chinook salmon (SRFCS), a threatened species in the U.S., have been dramatically changed by hydropower development and other watershed modifications. Associated biological changes include a shift in juvenile life history: historically essentially 100% of juveniles migrated to sea as subyearlings, but a substantial fraction have migrated as yearlings in recent years. In contemplating future management actions for this species should major Snake River dams ever be removed (as many have proposed), it will be important to understand whether evolution is at least partially responsible for this life-history change. We hypothesized that if this trait is genetically based, parents who migrated to sea as subyearlings should produce faster-growing offspring that would be more likely to reach a size threshold to migrate to sea in their first year. We tested this with phenotypic data for over 2600 juvenile SRFCS that were genetically matched to parents of hatchery and natural origin. Three lines of evidence supported our hypothesis: 1) the animal model estimated substantial heritability for juvenile growth rate for three consecutive cohorts; 2) linear modeling showed an association between juvenile life history of parents and offspring growth rate; and 3) faster-growing juveniles migrated at greater speeds, as expected if they were more likely to be heading to sea. Surprisingly, we also found that parents reared a full year in a hatchery produced the fastest-growing offspring of all—apparently an example of cross-generational plasticity associated with artificial propagation. We suggest that SRFCS is an example of a potentially large class of species that can be considered to be “anthro-evolutionary”—signifying those whose evolutionary trajectories have been profoundly shaped by altered selective regimes in human-dominated landscapes.

http://onlinelibrary.wiley.com/doi/10.1111/eva.12468/full

Table A. Estimates of broad-sense heritability (H2) for growth rate of fall Chinook salmon from each of the best-fit models that considered the fixed effects of brood year (BY), female fork length (FFL), and rearing site. Model selection was based on AICc. The random effect of individual breeding value (‘animal’) is present in all models. H2 estimates are from animal models using a restricted maximum-likelihood algorithm; all estimates differ significantly (P < 0.05) from zero. Model Fixed Random H2 SE(H2) 95% CI BY2007 FFL animal 0.733 0.098 0.541-0.925 BY2008 FFL animal 0.804 0.078 0.651-0.957 BY2009 Site, FFL animal 0.744 0.093 0.562-0.926 All BYs FFL animal 0.775 0.047 0.683-0.867

Figure A. Effect of parental life history (S = subyearling smolt; Y = volitional yearling smolt; FY = forced yearling smolt) on predicted growth rate relative to the overall mean growth rate for each parent/sex. Results are for the best model by AIC. Bars represent 95% confidence intervals for predicted effects, and the horizontal dotted line is a point of reference for measuring the magnitude of each effect. Numbers in parentheses are sample sizes for each group.

Figure B. As in Figure A, but only showing results for pure crosses. Note the small sample size of YxY.

may 16 and 17th, 2017 at the quality inn, clarkston, wa. fall chinook... · may 16 and 17th, 2017...

Documents