Survival Analyses in Support of NOAA’s Draft Biological Opinion on

California WaterFix

Russell W. Perry, USGS, Western Fisheries Research CenterA. C. Pope, J. G. Romine, A. Blake, and J. Burau, USGS

P, Brandes, USFWS, A. Ammann and C. Michel, NMFS

23 January 2017U.S. Department of the InteriorU.S. Geological Survey

NRDC-18

Overview• New Bayesian Mark-Recapture Model

– Overview of methods and results– Forms basis for BiOp Analyses

• Using these models for Cal WaterFix– Simulating survival under NAA and PA– Evaluating NDD Bypass Rules– Shaping operations with survival criteria

NRDC-18

Acoustic Telemetry Datafor Analysis

• Data from 2 Acoustic Telemetry Studies– NOAA (CALFED) and USFWS (Delta Action 8)– Late-fall Chinook salmon– Vemco acoustic telemetry– 2,170 Acoustic tagged fish– 5 Years (2007 – 2011)– 17 unique release groups– Migrated between late Nov. and early March– Sacramento River Flows at Freeport

• ~6,000 – 77,000 ft3/s

NRDC-18

12 miles

12 km

Survival Reaches

1

23

45 67

8

NRDC-18

MultistateModel

Schematic3 types of parameters:p = Detection probabilityΨ = Routing probabilityS = Survival probability

NRDC-18

Estimation Framework• Time-varying individual covariates

– Covariate values based on date of reach entry

• Date of entry unknown for undetected fish– Need to integrate likelihood over missing data– Requires a model for missing data

• Model for reach-specific travel times– Estimate parameters from observed travel times– Impute missing travel times

NRDC-18

Strength of Bayesian Framework• Time-varying individual covariates

– Previous approaches used average values

• Single integrated model– Survival and travel time

• MCMC to integrate over missing data

• Random effects– Quantify “extra” variation among release groups

NRDC-18

Model for Travel Times• Assume travel times (ti, j) distributed lognormally

µj = mean of log(ti,j)σj = standard deviation of log(ti,j)exp(µj) = median travel time

• Goal is to estimate µ and σ for each reach

NRDC-18

Effect of Discharge on Travel Times• Relate median travel times to Delta inflows at

Freeport

αj = reach-specific slope parameters

Qi,j,d = Freeport discharge on day d when ith fish entered jth reach

I(DCCi,j,d = open) = binary indicator for reaches downstream of DCC

εg,j = deviation of gth release group, ~Normal(0, ξ)

( ), 0, 1, , , 2, , , , jDCC = openi j j j i j d j i j d gQ Iµ α α α ε= + + +

NRDC-18

Effect of Discharge on Survival• Relate survival to Delta inflows at Freeport

βj = reach-specific slope parameters

Qi,j,d = Freeport discharge on day d when ith fish entered jth reach

I(DCCi,j,d = open) = binary indicator for reaches downstream of DCC

εg,j = deviation of gth release group, ~Normal(0, ξ)

( ) ( ), 0, 1, , , 2, , , ,logit DCC = openi j j j i j d j i j d g jS Q Iβ β β ε= + + +

NRDC-18

R

R

R

R

R

R

R

R

-2 0 1 2

1Flow-0.6 -0.2 0.2

2DCC

Parameter Estimates:Travel Time

• Negative slopes for all reaches

• Except DCC (Reach 6)

• Travel time decreases with inflow in all reaches

• DCC effects less certain• Except Rio Vista – Chipps

(Reach 7)

NRDC-18

Parameter Estimates:Survival

-2 0 2 4

Reach 8

Reach 7

Reach 6

Reach 5

Reach 4

Reach 3

Reach 2

Reach 1

1Flow-1.5 -0.5 0.5

2DCC 12 miles

12 km

1

23

4

5 67

8

NRDC-18

Flow-Survival Relations

12 miles

12 km

1

2

0.0

0.4

0.8

10 30 50 70

0.0

0.4

0.8

10 30 50 70

Surv

ival

pro

babi

lity

Reach 1

Reach 2

Discharge (ft3/s x 1000)

NRDC-18

0.0

0.4

0.8

10 30 50 70

0.0

0.4

0.8

10 30 50 70

4

12 miles

12 km

3

5

Surv

ival

pro

babi

lity

Reach 3

Reach 4 Reach 50.0

0.4

0.8

10 30 50 70

Flow-Survival Relations

Reach 5

Discharge (ft3/s x 1000)

NRDC-18

0.0

0.4

0.8

10 30 50 70

Parameter Estimates:Survival

12 miles

12 km

7

8Surv

ival

pro

babi

lity

0.0

0.4

0.8

10 30 50 70

Reach 7

Reach 8

NRDC-18

Route-Specific Survival

Sacramento River Discharge at Freeport (ft3/s x 1000)

0.0

0.2

0.4

0.6

0.8

1.0

10 20 30 40 50 60 70 80

Surv

ival

pro

babi

lity

0.0

0.2

0.4

0.6

0.8

1.0

10 20 30 40 50 60 70 800.

00.

20.

40.

60.

81.

0

10 20 30 40 50 60 70 80

0.0

0.2

0.4

0.6

0.8

1.0

10 20 30 40 50 60 70 80

Sacramento R.

Georgiana Slough

Sutter and Steamboat S

NRDC-18

0 10 20 30 40 50 60

0.0

0.1

0.2

0.3

0.4

70,000 cfs

Travel time (d)

Pro

babi

lity

dens

ity

Sacramento R. (2.1 d)Sutter and Steamboat S. (2.5 d)Georgiana S. (4.5 d)Delta Cross Channel (NA)

Route-Specific Travel Times

0 10 20 30 40 50 60

0.0

0.1

0.2

0.3

0.4

25,000 cfs

Travel time (d)

Pro

babi

lity

dens

ity

Sacramento R. (6.6 d)Sutter and Steamboat S. (7.3 d)Georgiana S. (11.6 d)Delta Cross Channel (14.4 d)

0 10 20 30 40 50 60

0.0

0.1

0.2

0.3

0.4

5,000 cfs

Travel time (d)

Pro

babi

lity

dens

ity

Sacramento R. (11.3 d)Sutter and Steamboat S. (11.9 d)Georgiana S. (17.9 d)Delta Cross Channel (19.7 d)

Travel time (d)

NRDC-18

Summary

• Inflows affect travel times in all reaches

• Inflows affect survival in some reaches– Upper reaches: high survival at all flows– Transition reaches: strongest flow-survival

relations– Tidal reaches

• no evidence of flow effect• imposes upper limit on route-specific survival

NRDC-18

Simulating Survival, Travel Time, and Routing for NAA and PA

1. “Release” 10,000 fish at Freeport each day.2. Reach 1 survival same for all fish.3. Draw reach 1 travel times as f(flow)

– NAA: flow = Freeport discharge– PA: flow = Bypass discharge

4. At junction of Sutter/Steamboat and Sac, draw route as f(flow).

5. Reach-specific survival f(flow) at arrival time.6. Repeat for all subsequent reaches.

NRDC-18

5000

1000

020

000

5000

0

DCC Closed

DCC OpenByp

assd

isch

arge

ft3

s

NAAPA

0.1

0.2

0.3

0.4

0.5

0.6

0.7

Mea

n su

rviv

al th

roug

h D

elta

-0.2

-0.1

0.0

0.1

0.2

Oct Nov Dec Jan Feb Mar Apr May Jun Jul

Diff

eren

ce in

sur

viva

l (P

A -

NA

A)

1943 (WY type = W)

Outputs for Each Year: SurvivalNRDC-18

Outputs for Each Year: Travel Time

5000

1000

020

000

5000

0

DCC Closed

DCC OpenByp

assd

isch

arge

ft3

s

NAAPA

24

68

1012

14

Med

ian

trave

l tim

e th

roug

h D

elta

(d)

-4-2

02

4

Oct Nov Dec Jan Feb Mar Apr May Jun Jul

Diff

eren

ce in

trav

el ti

me

(PA

- N

AA

)

1943 (WY type = W)

NRDC-18

Outputs for Each Year: Routing

5000

1000

020

000

5000

0DCC Closed

DCC OpenB

ypas

sd

isch

arge

ft3s

NAAPA

0.0

0.2

0.4

0.6

0.8

1.0

0.0

0.2

0.4

0.6

0.8

1.0

Cum

ulat

ive

fract

ion

usin

g ea

ch ro

ute

(PA

)

DCCGeorgiana S.Sutter & Steamboat S.Sacramento R.

Oct 01 Nov 01 Dec 01 Jan 01 Feb 01 Mar 01 Apr 01 May 01 Jun 01

0.0

0.2

0.4

0.6

0.8

1.0

0.0

0.2

0.4

0.6

0.8

1.0

Cum

ulat

ive

fract

ion

usin

g ea

ch ro

ute

(NA

A)

1943 (WY type = W)

NRDC-18

Summarizing Survival Differences

-0.2

0-0

.10

0.00

0.05

Diff

eren

ce in

sur

viva

l (P

A -

-0.2

0-0

.15

-0.1

0-0

.05

0.00

0.05

Diff

eren

ce in

sur

viva

l (L1

- N

AA

)-0

.10

-0.0

50.

000.

050.

10

Oct Nov Dec Jan Feb Mar Apr May Jun Jul

Diff

eren

ce in

sur

viva

l (L1

- P

A)

NRDC-18

Evaluating NDD Bypass Rules

• Apply rule sets under “equilibrium” conditions– Assume constant inflows and operations for

cohort

• Calculate survival with and without diversion

• Evaluate survival differences for each rule set

NRDC-18

Oct.-Nov. Bypass Rules

Dis

char

ge f

t3s

1000

Bypass flowDiversion flow

010

2030

4050

600.

10.

20.

30.

40.

50.

60.

7

SD

elta w

ith D

CC

clo

sed

7

0.00

0.05

0.10

0.15

Diff

eren

ce in

sur

vi

Oct. - Nov. Bypass Rules

5 15 25 35 45 55

5 15 25 35 45 55

Sacramento River discharge at Freeport ft3 s 1000

0.00

0.05

0.10

0.15

0.20

10 20 30 40 50 60

Without diversionWith diversion

NRDC-18

Constant Low-Level Pumping

5 15 25 35 45 55

5 15 25 35 45 55

Sacramento River discharge at Freeport ft3 s 1000

Dis

char

ge f

t3s

1000

Bypass flowDiversion flow

010

2030

4050

600.

10.

20.

30.

40.

50.

60.

7

SD

elta w

ith D

CC

clo

sed

0.00

0.05

0.10

0.15

Diff

eren

ce in

sur

vi

Constant Low-Level Pumping (Dec-Jun)

0.00

0.05

0.10

0.15

0.20

10 20 30 40 50 60

Without diversionWith diversion

NRDC-18

Level 1 Post-Pulse Operations

5 15 25 35 45 55

5 15 25 35 45 55

Sacramento River discharge at Freeport ft3 s 1000

Dis

char

ge f

t3s

1000

Bypass flowDiversion flow

010

2030

4050

600.

10.

20.

30.

40.

50.

60.

7

SD

elta w

ith D

CC

clo

sed

0.00

0.05

0.10

0.15

Diff

eren

ce in

sur

vi

Level 1 Post-Pulse Operations (Dec-Apr)

0.00

0.05

0.10

0.15

0.20

10 20 30 40 50 60

Without diversionWith diversion

NRDC-18

Summary: NDD Bypass Rules• Some large survival differences

– Depends on • Bypass flows• Rule set

• In CalSim simulations– Highlights why larger differences in Oct., Nov.,

Jun.

• How else might operations be structured?

NRDC-18

Determining Operations based on Maximum Allowable “Take”

• Example criteria– No more than a 0.03 decrease in mean survival– 90% probability that survival is decreased by no

more than 0.03

• Use survival model to identify diversions that satisfy criteria– Find by optimization routine

NRDC-18

Diversions Based on Median Survival

0 20000 40000 60000 80000

020

0040

0060

0080

0010

000

Direct NDD mortality = 0

DischargeatFreeportft3 s

North

Del

taD

ivers

ionf

t3s

Median survival reduction0.0050.010.020.030.040.05

DCC closedDCC open

NRDC-18

Diversions Based on Full Posterior Distributionfor 0.03 survival reduction

Median

90% probabilityof <0.03 difference

NRDC-18

Survival Difference Based on10th percentile of NDD flowsfor 0.03 survival reduction

Median

90th Percentile

NRDC-18

Summary• Survival model can help identify operations

that meet specific survival criteria• Variability in survival can explicitly play a

role in setting criteria• New set of operations can be assessed with

other models– CVLCM, DPM, etc.– More robust inferences

NRDC-18

Acknowledgments

Delta Stewardship CouncilNOAADWRUC Davis

NRDC-18

Important Assumptions• Extending inferences:

– Late Fall Chinook = Winter Run?– Nov. – Mar. = Apr. – Jun.?– Hatchery = Naturally produced?– Current system state = future system state?– Predicting outside range of observed data?

• Relative vs. Absolute comparisons– Relative more robust

• NAA vs. PA• Shape of driving relationships similar

NRDC-18

Diversions Based on Median Survival

0 20000 40000 60000 80000

020

0040

0060

0080

0010

000

Direct NDD mortality = 0.03

DischargeatFreeportft3 s

North

Del

taD

ivers

ionf

t3s

Median survival reduction0.0050.010.020.030.040.05

DCC closedDCC open

NRDC-18

0 20000 40000 60000 800000.

00.

10.

20.

30.

40.

5

DischargeatFreeportft3 s

Ent

rain

men

t pro

babi

lity

Georgiana Slough, DCC closedGeorgiana Slough, DCC openDCC

NRDC-18