Comparative Survival Study Annual Meeting - FPC … from the 2013 CSS Annua… · Comparative...
Transcript of Comparative Survival Study Annual Meeting - FPC … from the 2013 CSS Annua… · Comparative...
Comparative Survival Study Annual Meeting Date / time: April 30th 2013 8:30 AM to 1:30 PM Place: Water Resources Center
4600 SE Columbia Way Vancouver, WA 98661
We ask that you please hold your questions until either of the two question and answer timeslots. Each presentation will have slide numbers for referencing back.
Time (Minutes) Presenter
CSS 2012 Annual Report 08:30 Introduction to CSS (30) Jack Tuomikoski 09:00 Upper Columbia River Chinook and Steelhead (15) Robin Ehlke 09:15 Snake River Fall Chinook (20) Jerry McCann 09:35 Questions / Discussion (15)
09:50 Break (15)
CSS 2013 Workshop
10:05 Introduction and Background (20) Howard Schaller
10:25 Scenario Development and Prospective Models (45) Margaret Filardo & Charlie Petrosky
11:10 Experimental Management Choices (20) Steve Haeseker 11:30 Questions / Discussion
To end at approximately 1:30 PM
Introduction to Comparative Survival Study
Jack TuomikoskiCSS Annual Meeting 30th Apr, 2013
2
2012 Annual Report
Workshop (March 7th-8th, 2013)
• GOALS:Review a draft design for a management experiment to
increase the amount of voluntary spill at FCRPS projects
Provide recommendations to strengthen the proposed experiment
Opportunity for leading researchers and professionals involved in the assessment of dam operations on various components of salmon and steelhead survival rates to share and compare recent results.
20 attendees from agencies and universities
• Facilitated by ESSA Technologies Ltd.
2012-2013 Overview
3
Initiated in 1996 by states, tribes & USFWS to estimate survival rates at various life stages
• Designed to assess hydrosystem operations on state, tribal, and federal fish hatcheries and LSRCP
• PATH – “can transportation . . . compensate for the effect of the hydrosystem?”
• NPCC has established the need to collect annual migration characteristics including survival
• NOAA biological opinions require research, monitoring and evaluation
Management-oriented large scale monitoring• Observational study• Aligned with basin wide monitoring needs (RME)
Background
4
Background
GOALS 1.Quantify the efficacy of transportationDevelop a more representative control group
2.Compare survival rates within and across species
3.Establish long term data set
5
Background
CSS data is derived from PIT tags
• Tagged specifically for CSS
• Cooperative marking between CSS and other research studies reduce costs/handling, eliminate duplication
• Groups marked for other studies
6
Background
Collaborative scientific process was implemented for study design and to perform analyses
CSS project independently reviewed and modified a number of times• Draft report typically posted – Aug 31st• ISAB, ISRP and other entities
7
History of ISAB/ISRP Reviews of CSS
1997 – ISAB First review
1998 – ISAB Extend to other species & life history types (Steelhead)nonparametric bootstrap approach
2002 – ISRP Additional evaluate bootstrap, compare with likelihood methods, Monte Carlo simulator evaluation
8
History of ISAB/ISRP Reviews of CSS
2003 – ISAB Review of flow augmentation“understanding of the relation betweenreach survival, instantaneous mortality,migration speed, and flow”
2006 – ISAB Review of 2005 CSS report1) “finer scale analyses of the relationships
between survival and specific operationalactions or environmental features”
2) Develop a ten year summary report
9
History of ISAB/ISRP Reviews of CSS
2007 – ISAB/ISRP Review CSS “10-year” report1) continue coordination
cost savings/ avoid redundancy
2) Evaluate if PIT tag SARs are less than run reconstruction SARs
2009 – ISAB Tagging ReportCompare CSS SARs with Run Reconstruction SARs
>2009 ISAB annually reviews CSS reports
10
DESIGN• WDFW, CRITFC, USFWS, ODFW, IDFG
IMPLEMENTATION & TAGGING• FPC: Logistics, coordination• PTAGIS: Raw Data; FPC: Reports, Estimates
DATA PREPARATION & ANALYSIS• CSS Oversight Committee• Fish Passage Center
REGIONAL REVIEW• Draft on BPA & FPC websites• Regional Public Review; ISAB, ISRP, FPAC, NMFS,
etc.
FINAL REPORT• Posted on BPA & FPC websites
The CSS is a joint project of the state & tribal fishery managers and the USFWS
TEMPORAL COVERAGE
11
Snake RiverLonger Time SeriesMore groups developed
~20 stocks sp/su/fall Chinook,
steelhead, sockeye
Middle Columbia RiverBegin in 2000 [BOA detect] ~ 9 stocks
Upper Columbia RiverBegin in 2000 [BOA detect] ~8 stocks
12
Spatial Coverage: Hatchery Chinook
13
Spatial Coverage: Wild Chinook
14
Spatial Coverage: Snake River Hatchery Steelhead
15
Spatial Coverage: Wild steelhead
SR
LGR
MCN
BON
FRESHWATER
ESTUARY
OCEAN
Smolt Survival Rearing Habitat Actions
Hydro-system Actions
SLGR-MCN
SMCN-BON
16
LGSLMNIHR
JDA
LGR
BON
FRESHWATER
ESTUARY
OCEAN
Adult Success
Hydro-system Actions
Wild
Hatchery
Har
vest
Man
agem
ent
Estuary Habitat Actions
Hydro-system ActionsTransportation
Effects
17
MCNIHR
LGR
BON
FRESHWATER
ESTUARY
OCEAN
Hydro-system Actions
Har
vest
Man
agem
ent
Estuary Habitat Actions
Hydro-system ActionsTransportation
Or Bypass
effects
SNAKE RIVER SARS
18
BON
FRESHWATER
ESTUARY
OCEAN
Mid and Upper Columbia R. SARS
Regional Monitoring &
EvaluationJDAMCN
19
20
Long term consistent information collaboratively designed and implemented
Information easily accessible and transparent• CSS PIT-tags accessed by any PTAGIS users, including fisheries
managers, researchers, and academics.
Long term indices (identify bottlenecks) :• Travel Times• In-river Survival Rates• In-river SARs by route of passage• Transport SARs• Adult success, conversion
Comparisons of SARs• Transport to In-River• NPCC Regional SAR goal• By geographic location• By hatchery group• Hatchery to Wild• Chinook to Steelhead
Management questions: hydropower operations, hatchery evaluations, habitat evaluations
What does CSS provide for the region?
21
Snake River Chinook SARs Hatchery Chinook
5 spring 3 summer
Summer > Spring
Highly correlated 1999, 2000, 2008 > 2%
22
Snake River Chinook SARs Hatchery Chinook
5 spring 3 summer
Summer > Spring
Highly correlated 1999, 2000, 2008 > 2%
Wild Chinook Similar results
23
11.5% SARs > 2 77% SARs < 2 11.5% SARs = NS
CSS Results: Chinook SARs
24
Less correlated than Chinook stocks
2008* Highest in time series
Snake River Steelhead SARs
25
Snake River Steelhead SARs
15% SARs > 2 50 % SARs < 2 35% SARs = NS
26
Used to evaluate transportation program (SR stocks)
Ratio of Transported ÷ InriverSARs
Snake River TIR
FRESHWATER
OCEAN
27
TIR is related to in-river survival
As in-river survival increases, TIR decreases
When in-river surv ~ 57%, transport will not be beneficial (for wild stocks)
TIR vs. in-river surv.
28
The success rate for transported was 90% of that for their in-river counter parts (on average)
Transported steelhead strayed about 4.5% and in-river strayed at 0.4% (11:1) Deschutes and John DayThis is a large out-of-basin population as compared to
total natural spawner abundance◦ Hatchery strays identified as limiting factor to recovery of John Day
and Deschutes River stocks (NOAA 2009 Mid C. St. Recovery Plan)
Transported hatchery Chinook strayed about 0.7% and in-river strayed 0.03% (23:1)Columbia above SR confluence
Snake River Adult Success and Straying
29
Developed as monitoring tool and to inform harvest managementUpdate in 2012 report, 7 additional stocks (16 total)
and one more year of data
Age at maturity and jack percentage of Chinook in Snake River and Middle/Upper Columbia were influenced by both stock and year factorsA common sibling model across all stocks may not
perform well
No strong association between age at maturity and transport history for Snake River stocks
Age at Maturity for Chinook
1
SARs and Juvenile Metrics of Upper Columbia Stocks
Robin Ehlke
CSS Annual Meeting Apr 30th 2013
CSS ObjectivesUpper Columbia
Establish long term survival estimates over the full life-cycle of upper Columbia stocksDevelop Smolt to Adult Return rates
(SARs) from the upper-most damDevelop estimates of ocean survival ratesUse additional mark groups as they come
available
2
Upper Columbia Mark GroupsFive Basin-Specific Groups
• Wenatchee BasinHatchery spring Chinook (Leavenworth)Wild ChinookSteelhead (hatchery/wild Cross)
• Entiat-Methow aggregateWild Chinook
• Wenatchee-Entiat-Methow aggregateWild Steelhead
Three Groups marked at Rock Island Dam• Yearling Chinook, subyearling Chinook,
steelhead• All three are hatchery/wild aggregates
3
Upper Columbia Juvenile and Adult Metrics
Juvenile passage metrics, travel time, instantaneous mortality and survival from Rock Island to McNary DamSmolt to Adult Return ratesIncorporated detection capability at Rocky
Reach DamReport analyses of passage metrics and
SARs relative to environmental variables
5
Smolt to Adult Return
Upper Columbia Smolts from McNary Dam to Bonneville Dam• MCN to BON SARs do not include or account for
juvenile mortality occurring through the Upper Columbia to McNary Dam
• For this reason the MCN to BON reported SARs are biased high
• As an example, for Wenatchee the SARs would be ~ 58% of reported if RIS to MCN juvenile survival were taken into account
6
Steelhead survive slightly better than Chinook
Typically both species’ survival is less than 60%.
A large component of life-cycle is not represented in MCN to BON SARs
7
Rock Island to McNary Juvenile Survival
Juvenile Passage RIS-MCN Metrics/Environmental conditions Fish Travel Time
• Faster with higher flow and with Julian date Instantaneous Mortality
• Decreased for Chinook as spill levels increased at Wanapum and Priest Rapids
• Increased for steelhead with increase in Julian date
Reach Survival• Increased with higher flow and spill
8
Wild and Hatchery Chinook SAR MCN to BON
Entiat/Methow R. Wild Chinook SARs averaged 1.35% (0.5%-3%) Enter Columbia River upstream of Rocky Reach Dam
Wenatchee River Wild Chinook SARs averaged 1.62% (0.8%-3%)Enter the Columbia River upstream of Rock Island Dam
Leavenworth Hatchery Chinook SARs averaged 0.58%Enter the Columbia River upstream of Rock Island Dam
All GroupsExceeded 2% in 2008Do not include Upper Columbia reach2010 data does not include 3-salt fish
9
Wild and Hatchery Chinook SAR MCN to BON
Entiat/Methow R. Wild Chinook SARs averaged 1.35% (0.5%-3%) SARs less when calculate from Rocky Reach back to Bonneville Dam
Wenatchee River Wild Chinook SARs averaged 1.62% (0.8%-3%)Enter the Columbia River upstream of Rock Island Dam
Leavenworth Hatchery Chinook SARs averaged 0.58%Enter the Columbia River upstream of Rock Island Dam
All GroupsExceeded 2% in 2008Do not include Upper Columbia reach2010 data does not include 3-salt fish
10
Wild and Hatchery steelhead SAR MCN to BON
Wild SteelheadSARs averaged 3.97% Aggregate mark group -Wenatchee, Entiat and Methow stocksUpper Columbia reach not included
Hatchery SteelheadSARs averaged 2.16%Wenatchee basin Upper Columbia reach not included
11
Wild and Hatchery steelhead SAR MCN to BON
Wild SteelheadSARs averaged 3.97%SARs less when calculate from Rocky Reach back to Bonneville Dam (Entiat and Methow stocks)
Hatchery SteelheadSARs averaged 2.16%Wenatchee basin Upper Columbia reach not included
12
Chinook and Steelhead SAR RIS to BON
Hatchery/Wild Yearling Chinook SARs averaged 0.3% Bypass inoperable during spring of 2003
– no data
Hatchery/Wild Subyearling Chinook SARs averaged 0.6%
Hatchery/Wild Steelhead SARs averaged 1.17% Bypass inoperable during spring of 2003
– no data
13
14
15
16
Reach Survival Comparison of Juvenile Salmon:Snake River to Upper Columbia Stocks (FPC report)
Yearling Chinook
Steelhead
17
Average Percent Spill at Wanapum and Priest Rapids Dams
Smolt Migration Year
Avg.
Per
cent
Spi
ll
18
ConclusionThe Overall Upper Columbia MCN-BON SARs for
2000-2010 Chinook were highly correlated with spring Chinook SARs from the Middle Columbia and with spring/summer Chinook SARs from the Snake River
Indication that upper Columbia stocks have similar responses to shared in-river and ocean life-cycle experiences.
Upper Columbia stocks also showed similar patterns of response to environmental variables when compared to mid Columbia and Snake River Stocks
ConclusionCollaboration and coordination with other Upper
Columbia specific marking efforts increases cost effectiveness and the benefits to the region
Monitoring the effect of hydro system passage on Upper Columbia groups from existing marking is value added for managers
Increase in the number of mark groups/tags and the number of detection sites would strengthen the data.
Recent increase in USFWS marked hatchery steelhead and Chinook will be available for future years (Winthrop, Entiat) 19
The End
20
Snake River Fall Chinook SARs
from PIT groups with low or no Holdover Bias
Presenter: Jerry McCann
CSS Annual Meeting Apr 2013
22
Background CSS was requested to develop estimates of
subyearling fall Chinook SARs.
CSS approach compares SARs for transported fish to fish undetected in-river (C0) at transport dams.
Holdover fish can bias estimates of C0 SARs.
3
Subyearling Fall Chinook PIT-tag Releases
WildResearchProduction
What is a holdover?
Subyearling Migration - fish migrate past dams in year of release/emergence
Holdover – or yearling migration - fish migrate past some dams as yearlings – the year after release/emergence
40
50
60
70
80
90
100
4/7 5/17 6/26 8/5
Avg
Leng
th (m
m)
PROD_SNAKE
PROD_CLWR
40
50
60
70
80
90
100
4/7 5/17 6/26 8/5
Avg
Leng
th (m
m)
PROD_SNAKE
PROD_CLWR
SUR_SNAKE
WILD_SNAKE
5
Release date and length at release for PIT tagged Ch0 in 2009
Later releases at shorter lengths = Higher Holdover Probability
40
50
60
70
80
90
100
4/7 5/17 6/26 8/5
Avg
Leng
th (m
m)
PROD_SNAKE
PROD_CLWR
SUR_SNAKE
WILD_SNAKE
SUR_CLWR
WILD_CLWR
6
Proportion Holdovers by release date for PIT tagged Ch0 in 2009
Later releases at shorter lengths = Higher Holdover Probability
SARs on Fall Chinook
Present SAR data• Present SARs estimates by study category for release
groups from 2006 to 2009 with low or no holdovers
Results of Simulations to calculate potential bias to SARs due to holdover fish• C0 population may be biased low by unmonitored
passage during winter or holdover passage
7
Approach to bias calculation
Used holdover detections exiting hydro-system at Bonneville Dam to calculate holdover population.• Determined the number of Bonneville detections as holdovers
(i.e. possibly passed LGR during shutdown or as yearlings). • Expanded those yearling detections based on detection
probability at Bonneville Dam to a Population.• Expressed Bonneville population in LGR equivalents using a
range of survival estimates (0.25 to 0.75)
Add proportion unmonitored winter passage (i.e. passing LGR to BON during shutdown).• Use late season travel times to IHR to simulate potential
unmonitored winter passage LGR to BON.
8
Results of Simulations – projected potential C0 SAR bias due to holdovers
9
Migration year 2006 Migration year 2007
Migration year 2008 Migration year 2009
Conclusions from Simulations
Most groups particularly production releases have very low potential for holdover bias and are suitable for CSS methodology.When large numbers of late season
migrants and early season holdovers are observed at LGR, CSS approach not presently appropriate due to potential bias.
10
Results of SAR (with jacks) estimates for migr years 2006 to 2009
11
Comparison of Transport vs Inriver SARs (no jacks) for migr years 2006 to 2009
12
Big Canyon Creek Prod
Captain John’s Rapid Prod
Snake River Surrogates
Snake River Wild
Transport Inriver SARs ratios (TIRs) for migr years 2006 and 2008
13
Conclusions
Bias due to holdovers is not a concern for most subyearling production releases and Snake River wild PIT-groups in years analyzed.
SAR estimates can be reliably made using CSS methodology
SARs by study category (Transport and In-river) can be reliably estimated and be compared.
14
1515
Howard Schaller
CSS Annual MeetingApril 30, 2013
Introduction & Background for Comparative Survival Study Workshops and Outcomes
1
CSS Workshops• Workshop 2004
– Direct & delayed effects of the hydrosystem on salmon & steelhead survival
Workshop July 2011 – Factors influencing marine survival for Snake River
populations– Build tools to evaluate FCRPS operations
Workshop March 2013 – Management experiment to evaluate increased
voluntary spill
2
3 4 5 6 7 8 dams
Decline in Snake R. Chinook & steelhead associated with dams…
3
Dramatic changes in outmigration conditions
with dams…
marine conditions were not static…
Need to account for both
4
Challenges and management objectives
Multiple factors operating at same time• Capitalize on temporal patterns of variation• Capitalize on spatial patterns of variation• Consistency from multiple lines of evidence
NPCC SAR objectives of 4% average, 2% minimum
5
ApproachWeight of evidence Multiple lines of evidence for relative importance of major factors influencing survival rates
CSS SARs (Chin & Sthd)
SARs (run rec. ‐ Snake Chin & Sthd)
Spawner:recruit (Snake & John Day Chin)
3 4 5 6 7 8dams
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010
Environmental Contrast
Life cycle
Life stage
Precision & Spe
cificity
6
Estimate hydrosystem and ocean effects using reference populations
Estimate hydrosystem and ocean effects without using reference populations
Examine multiple lines of evidence –consistent results?
7
Spatial contrast
Spawner-recruit & SAR data
Snake R. Chinook populations survived ¼ to 1/3 as well as reference populations since FCRPS completion
0.0
2.0
4.0
6.0
8.0
10.0
12.0
14.0
2000 2002 2004 2006 2008 2010
% SAR
Chinook SARs
John Day
Snake
3 dams vs. 8 dams
8
‐5
‐4
‐3
‐2
‐1
0
1
2
1950 1960 1970 1980 1990 2000 2010
S‐R residuals
John Day
Snake
Temporal Analyses• Influence of river & ocean conditions on survival rates
– Employ long time series:• Pre & post Snake River dam completion• Survival rates for different life stages (S-R residuals, SAR,
marine survival)• Variables for ocean conditions• Variables for river conditions
• For different life stages & species – contrast the set of ocean & river conditions that explain variation in survival rates (temporal)
• Temporal/spatial contrasts to estimate FCRPS impacts – mortality at later life-stages (delayed hydrosystem mortality)
9
California Current brings cold polar water from the north keeping coastal temperatures cool
Broad scale:• Pacific Decadal Oscillation
Near shore:•Coastal Upwelling•Spring Transition•Near shore Temp.
Candidate Ocean Variables
10
Candidate River VariablesWater travel time (Lewiston – BON Dam):
•2 days pre-dam•10-40 days (19 day ave.) post-dam
In-river migrants now pass through:
•up to 8 powerhouses•depending on spill & surface weirs
Collection and transport:•25% - 99% of smolts transported , 1977
11
0.0
0.2
0.4
0.6
0.8
1.0
1960 1970 1980 1990 2000 2010
Prop
ortio
n tran
sported
Migration Year
Multiple Regression – long-term data setsSpawner-recruit residuals, SARs & marine survival
showed similar results
Reduced survivalis associated with:
• Warm PDO• Reduced upwelling
• Increased # powerhouse passages (reduced spill)
• Slow WTT• Increased
proportion transported
12
0.0
0.2
0.4
0.6
0.8
1.0
Model Scenario Development
Presenter: Margaret FilardoCSS Annual Meeting
April 30, 2013
Workshop Tasks
• Develop reasonable estimates of the amount of water that could be spilled (spill caps) at each of the hydroprojects on the Lower Snake and Columbia rivers– using present project configuration; and, – existing TDG waivers/rule changes.
• Develop scenarios for alternate spill programs for modeling exercises.
2
3
Spill Benefits
• Historic data has consistently shown a juvenile survival advantage.
• Spill is the only mitigation measure provided that can be provided in every flow year.
• Spill can be provided without impact to reservoir elevations.
4
Entrainment of Air
5
6
Risk Based Spill Program
Survival benefits of spill > potential TDG related
mortality
7
Objectives
• Identify any reasonable physical or biological limitations at each project.
• Determine the spill caps that could be used in the estimation of spill under the various total dissolved gas scenarios.
• Identify representative water years and estimates of spill to develop input for spill variables used in the prospective modeling.
8
Objective 1 Physical Limitations
• Adult Passage Restrictions:– Little Goose - 30% spill limitation– Ice Harbor – 45 Kcfs daytime spill limitation – Bonneville Dam – 110 kcfs spill limit – Adult passage success at various spill levels
• Minimum generation requirements.– Incorporated
9
Objective 2 Determination of Gas Caps
• Source of Gas Cap Spill Levels
– Information developed by the COE for the real time distribution of involuntary spill
10
Objective 3Representative Water Years
• Model the spill scenarios under a range of flow conditions: – low, medium and high.
• Used hourly flow and spill data from 2009, 2010 and 2011.
11
0
10
20
30
40
50
60M
AF
Lower Granite Dam Runoff Volume 1929-2012
Jan-July
12
0
10
20
30
40
50
60M
AF
Lower Granite Dam Runoff Volume 1929-2012
Jan-July2011
12
0
10
20
30
40
50
60M
AF
Lower Granite Dam Runoff Volume 1929-2012
Jan-July2011
2009
12
0
10
20
30
40
50
60M
AF
Lower Granite Dam Runoff Volume 1929-2012
Jan-July2011
2009
12
13
0
10
20
30
40
50
60M
AF
2010
Lower Granite Dam Runoff Volume 1929-2012
Jan-July2011
2009
14
0
20
40
60
80
100
120
140
160
180
MA
F
The Dalles Dam Runoff Volume 1929-2012
Jan-July
15
0
20
40
60
80
100
120
140
160
180
MA
F
The Dalles DamRunoff Volume 1929-2012
Jan-July 2011
15
0
20
40
60
80
100
120
140
160
180
MA
F
The Dalles DamRunoff Volume 1929-2012
Jan-July 2011
2009
15
0
20
40
60
80
100
120
140
160
180
MA
F
The Dalles DamRunoff Volume 1929-2012
Jan-July 2011
2009
15
0
20
40
60
80
100
120
140
160
180
MA
F
2010
The Dalles DamRunoff Volume 1929-2012
Jan-July 2011
2009
15
Development of Scenarios Common Assumptions and Applications
• Used actual hourly flows. • Minimum generation flows as specified in the 2008
Biological Opinion.• Hydraulic capacity of each project equal to flow with all
units operating at the upper end of the 1% efficiency range based on the 2012 FPP.
• No excess generation (lack of market) spill included. • Includes uncontrolled spill in excess of (HC+Planned spill)• No adult passage constraints on daytime spill amounts,
except for Bonneville.• Spill Caps as determined by the COE for the distribution of
Uncontrolled Spill.16
Scenarios• Started with Actual Hourly Flow and
spill for three representative flow years and applied:– Biological Opinion Spill Levels– Spill to the present TDG standards during
fish migration season of 115% FB and120% TR
– Spill to 120% TDG at the TR monitors– Spill to 125% TDG at the TR monitors
17
Resulting Spring Spill Proportions
18
Application of Scenarios
• Provided average proportion spill per project/time period as needed by the models for prospective analyses.
• Spill scenarios were input to the PIT Tag SAR Model and the Run Reconstruction SAR Model.
• Applied the actual spill and 4 spill scenarios to 3 different flow years.
19
Spill Study Implementation Constraints
• To Implement Spill to 120%
– WA DOE would need to make a modification to the Washington State Rule to eliminate the forebay monitors from compliance monitoring.
– OR DEQ already eliminated the forebay monitors from their waivers.
20
• To Implement Spill to 125%:
– WA DOE would need to make a modification to the Washington State Rule to increase the TDG to 125% and use TR monitors only.
– OR DEQ already eliminated the forebay monitors from their waivers, but would need to issue a waiver to the 125% TDG in the tailrace.
21
Summary of GBT Samples (1995-2012) as a function of TDG
22
23
Prospective Model Development
Presenter: Charlie Petrosky
CSS Annual MeetingApril 30, 2013
In‐river Passage RoutesNon‐powerhouse = Spill (traditional or surface spillway weirs)
Powerhouse = Turbine or juvenile collection/bypass
Submersible traveling screen
Collection channel
(3) Turbine
Forebay
Tailrace
(1) Spillway Reservoir
(2) Juvenile Bypass Systems
Gatewell
Direct survival:spill > bypass > turbine
Direct & indirect survival(delayed mortality):
spill > bypassspill > turbine
2
Spill metrics
• Average spill proportion
• N_Powerhouse – assume fish pass in spill proportionally to amount of water spilled
• N_PHPIT – estimate the proportion of fish passing in spill as a function of spill, flow and surface passage structures
3
Spill metrics
• Average spill proportion – commonly used metric (FPC, CSS, NOAA analyses, Haeseker et al. 2012)
• N_Powerhouse– Expected number of powerhouse passages (turbine + collection/bypass) for defined time period across N‐dams (Petrosky & Schaller 2010)
– N_dams changed over historic period ‐ Snake River fish faced 2 dams in 1956 to 8 dams by 1975
– Can account for variation during development (2 through 8 dams) and post‐development (8 dams)
4
New spill metric
• N_PHPIT– recommendation from 2011 CSS Workshop– expected number of powerhouse passages based on PIT tags– accounts for spill passage efficiency
• spillway passage proportion differs by dam, flow, spill proportion, surface collectors, and species
5
Fish Guidance Efficiency (FGE)
• Proportion of powerhouse fish that are “guided” to collection/bypass systems via screens
• FGE = Bypass / Powerhouse• Powerhouse = Bypass / FGE
• If you have bypass and FGE estimates, then you can estimate powerhouse passage
6
New spill metric
N_PHPIT approach:• Summarize CJS detection probabilities• Develop models for detection probabilities as function of spill, flow, surface passage
• Apply best‐fit model and FGE estimate to estimate spillway passage
• Estimates validated using telemetry data
7
Spill metricsLower Monumental Dam steelhead
0.0 0.2 0.4 0.6 0.8 1.0
0.0
0.2
0.4
0.6
0.8
1.0
Prop.Spill
Det
ectio
n pr
obab
ility
0.0 0.2 0.4 0.6 0.8 1.0
0.0
0.2
0.4
0.6
0.8
1.0
Observed
Pre
dict
ed
8
Spill metricsLower Monumental Dam steelhead
0.0 0.2 0.4 0.6 0.8 1.0
0.0
0.2
0.4
0.6
0.8
1.0
Prop.Spill
Det
ectio
n pr
obab
ility
0.0 0.2 0.4 0.6 0.8 1.0
0.0
0.2
0.4
0.6
0.8
1.0
Observed
Pre
dict
ed
FGE = 0.62
9
Spill metricsLower Monumental Dam steelhead
0.0 0.2 0.4 0.6 0.8 1.0
0.0
0.2
0.4
0.6
0.8
1.0
Prop.Spill
Det
ectio
n pr
obab
ility
0.0 0.2 0.4 0.6 0.8 1.0
0.0
0.2
0.4
0.6
0.8
1.0
Observed
Pre
dict
ed
FGE = 0.62
Detection prob. = f (spill, flow, spillway weir)
10
Spill metricsLower Monumental Dam steelhead
0.0 0.2 0.4 0.6 0.8 1.0
0.0
0.2
0.4
0.6
0.8
1.0
Prop.Spill
Det
ectio
n pr
obab
ility
0.0 0.2 0.4 0.6 0.8 1.0
0.0
0.2
0.4
0.6
0.8
1.0
Observed
Pre
dict
ed
0.00.10.20.30.40.50.60.70.80.91.0
0 0.2 0.4 0.6 0.8 1
50K flow80K flow160K flowPetrosky‐SchallerTel. 06, 07, 08, 09
Prop. Spill
Spillway passage
11
Spill metricsLower Monumental Dam steelhead
0.0 0.2 0.4 0.6 0.8 1.0
0.0
0.2
0.4
0.6
0.8
1.0
Prop.Spill
Det
ectio
n pr
obab
ility
0.0 0.2 0.4 0.6 0.8 1.0
0.0
0.2
0.4
0.6
0.8
1.0
Observed
Pre
dict
ed
0.00.10.20.30.40.50.60.70.80.91.0
0 0.2 0.4 0.6 0.8 1
50K flow80K flow160K flowPetrosky‐SchallerTel. 06, 07, 08, 09
Prop. Spill
Spillway passage
12
Spill metricsLower Monumental Dam steelhead
0.0 0.2 0.4 0.6 0.8 1.0
0.0
0.2
0.4
0.6
0.8
1.0
Prop.Spill
Det
ectio
n pr
obab
ility
0.0 0.2 0.4 0.6 0.8 1.0
0.0
0.2
0.4
0.6
0.8
1.0
Observed
Pre
dict
ed
0.00.10.20.30.40.50.60.70.80.91.0
0 0.2 0.4 0.6 0.8 1
50K flow80K flow160K flowPetrosky‐SchallerTel. 06, 07, 08, 09
Prop. Spill
Spillway passage
Powerhouse = 1 ‐ Spillway13
Factors influencing detection probabilities
Project Flow Spill % Flow * Spill % RSW/TSWLGRLGSLMN
Chinook IHRMCNJDATDA NABON NA
LGRLGSLMN
Steelhead IHRMCNJDATDA NABON NA
14
0.0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
1950 1960 1970 1980 1990 2000 2010Smolt year
Retrospective ‐ # powerhouse passages
N_Powerhouse
N_PH_PIT
15
Spill Metric Summary• N_Powerhouse and N_PHPIT spill metrics are highly correlated
– SPE > 1:1– More deviation between spill metrics in recent years
• Addressed 2011 CSS Workshop recommendations to account for spill efficiency at different flow, spill, & configuration based on detection efficiencies– Recent years, increased spill across flow conditions, including lower Q – Spill at lower Q effective to pull fish away from powerhouse– Presence of spillway weirs
• Prospective analyses of alternative spill scenarios used N_Powerhouse and N_PHPIT
16
Prospective tools• Generate distribution of SARs from retrospective models for range of river and ocean conditions for alternative spill scenarios
• Summarize distributions relative to desired goals
17
010203040506070
0% 1% 2% 3% 4% 5% 6%
BiOp
Alternative Spill
Prospective tools• Summarize distributions relative to desired goals (e.g., population viability)
18
0
10
20
30
40
50
60
70
0% 1% 2% 3% 4% 5% 6%
Undesirable
Prospective tools• Summarize distributions relative to desired goals (e.g., NPCC SAR goals, Recovery)
19
0
10
20
30
40
50
60
70
0% 1% 2% 3% 4% 5% 6%
Desirable
Prospective Models ‐ Long Time Series • Model Coefficients: best models
– Ocean variables• Capture range of conditions for predictions
– River Variables• Water Velocity (WTT) • N_Powerhouse ‐ with increased spill• Proportion Transported with increased spill
• Project SARs for Alternative Spill Scenarios – Range of flow conditions (High, Average, Low)– Range of ocean conditions (Good, Average, Poor)
21
0
0.1
0.2
0.3
0.4
0.5
Biop 115/120 120 125
Prob
ability of SA
Rs below
1%
Probability of Chinook SAR Predictions less than 1%
Avoid Undesirable SARs
22
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
Biop 115/120 120 125
Prob
ability of SA
Rs abo
ve 2%
Probability of Chinook SAR Predictions greater than 2%
Achieve Desirable SARs
Prospective models – long time series• Simulations show survival rate increases from BiOp to
alternative spill scenarios:– Chinook life‐cycle survival rates– Chinook and steelhead SARs – All flow levels, especially low flows
• Complimentary analyses to more detailed, CSS PIT tag data sets– Consistent response between short and long time series
• Help measure future spill responses
• Experimental Management focused primarily on CSS PIT tag data sets (next presentation)
23
Presenter: Steve Haeseker
CSS Annual MeetingApril 30, 2013
Experimental Spill Management Design
Plan for measuring response to a treatment
‐ Treatment = increase in spill for fish passage
‐ Response = change in survival
‐ Plan = implement CSS monitoring methods
What is experimental design?
2
‐ Large change (perturbation)
‐ High precision of measured response variable
‐ High degree of replication
‐ Minimize and account for confounding factors
Elements of “good” experimental design
3
Which response should we measure?
Response Precision Change
Juvenile travel time High Low
Juvenile survival Low Medium
Ocean survival Medium Medium‐High
Smolt‐to‐Adult Survival High High
4
SARs should be the primary response variable monitored (but not the only one)
Advantages:
‐ Highest precision (estimate is mainly counts)
‐ Highest expected change across spill levels
‐ Easiest to collect consistently
‐ Incorporates freshwater and delayed mortality
Disadvantages:
‐ Requires 2‐3 year lag for adult returns
‐ Influenced by both freshwater and ocean conditions5
What are the magnitudes of expected SARs under the alternative spill operations?
6
Integrating across the four main sources of variability
Seasonality
Four cohorts
Flow
Three flow years
Spill
Four spill levels
Ocean
PDO level
Combinations influences the SAR response
7
Used SAR model to project SARs:
‐ Across four release cohorts
‐ Three flow levels (low, medium, high)
‐ Four spill operations (BIOP, 115/120, 120, 125)
‐ PDO conditions over 1900‐2012
8
Integrating across river and ocean conditions‐ steelhead
0
10
20
30
40
50
60
70
0% 1% 2% 3% 4% 5% 6%
BIOP_Low
Projected SAR
Frequency
9
Integrating across river and ocean conditions‐ steelhead
0
10
20
30
40
50
60
70
0% 1% 2% 3% 4% 5% 6%
BIOP_Low
BIOP_Medium
Projected SAR
Frequency
10
Integrating across river and ocean conditions‐ steelhead
0
10
20
30
40
50
60
70
0% 1% 2% 3% 4% 5% 6%
BIOP_Low
BIOP_Medium
BIOP_High
Projected SAR
Frequency
11
0
10
20
30
40
50
60
70
0% 1% 2% 3% 4% 5% 6%
BIOP_Low
Integrating across river and ocean conditions‐ steelhead
Projected SAR
Frequency
12
0
10
20
30
40
50
60
70
0% 1% 2% 3% 4% 5% 6%
BIOP_Low
115_Low
Integrating across river and ocean conditions‐ steelhead
Projected SAR
Frequency
13
0
10
20
30
40
50
60
70
0% 1% 2% 3% 4% 5% 6%
BIOP_Low
115_Low
120_Low
Integrating across river and ocean conditions‐ steelhead
Projected SAR
Frequency
14
0
10
20
30
40
50
60
70
0% 1% 2% 3% 4% 5% 6%
BIOP_Low
115_Low
120_Low
125_Low
Integrating across river and ocean conditions‐ steelhead
Projected SAR
Frequency
15
Summarized probability distributions for SARs at various levels
‐ Undesirable (< 1%): strong link to viability
‐ Desirable (> 2%): NPCC regional goal
16
0
10
20
30
40
50
60
70
0% 1% 2% 3% 4% 5% 6%
BIOP_Low
115_Low
120_Low
125_Low
Integrating across river and ocean conditions‐ steelhead
Projected SAR
Frequency
Undesirable
17
0
10
20
30
40
50
60
70
0% 1% 2% 3% 4% 5% 6%
BIOP_Low
115_Low
120_Low
125_Low
Integrating across river and ocean conditions‐ steelhead
Projected SAR
Frequency
Desirable
18
Probability
Chinook‐ Undesirable (< 1% SARS)
0%
25%
50%
75%
100%
125 120 115/120 BIOP
Spill Treatment 19
60%Since ‘98: 65%
Probability
Chinook‐ Desirable (> 2% SARS)
Spill Treatment
0%
25%
50%
75%
100%
125 120 115/120 BIOP
20
14%Since ‘98: 10%
How to measure whether change has occurred?
1) Test against expected frequency?
For Chinook, 60% of SARs should be in desirable range with 125% spill
21
2) Compare new SARs against prior SARs?
2a) Compare new SARs against “analog” years with similar flow, spill and/or PDO conditions?
22
How to measure whether change has occurred?
3) Compare new SARs against model predictions?
‐WTT, Spill and PDO known before adults return‐Within range of model uncertainty?
23
How to measure whether change has occurred?
4) Update and refine model parameters over time to determine whether associations are changing?
5) Use multiple sources of data (FTT, SH, SO, SAR, stock‐recruit residuals, run‐reconstruction SARs) to evaluate whether change has occurred?
24
How to measure whether change has occurred?
6) All of the above!!!
25
How to measure whether change has occurred?
1) SARs were most critical response to consider
2) Use all five SAR comparisons to assess overall strength of response
Workshop Comments and Recommendations
26
3) Assess how increased spill will affect detection efficiency and the precision of SARs
‐ New spillway detectors will increase detections
4) Improve communication of differences between spill scenarios and terminology for different audiences
Workshop Comments and Recommendations
27
5) Five years likely insufficient to achieve desired learning given variability in ocean and flow conditions
6) Consider linking duration to next Biological Opinion, implementing when learning will be maximized, developing stopping rules, and assess within‐season experiments
Workshop Comments and Recommendations
28
Summary:
• Experimental design = a work in progress– CSS Workshop final report with independent reviewer comments ~ Summer 2013
• Definition of spill scenarios for simulations based on what appears technically possible with current FCRPS configuration
• Not part of any existing implementation plan or current BiOp consultation
Summary:• Projections suggest spill to 115/120 or higher would:
– reduce risk of very low SARs (<1.0%)– increase likelihood of SARs >2% (the lower end of NPCC 2‐6% SAR goals)
• Expected benefits to Upper‐ & Mid‐Columbia stocks – These stocks provide for additional monitoring/learning
Summary:
• Simulations are encouraging in terms of:– expected response (conservation benefit)– likelihood of detecting response (learning)
BPA Project 19960200