Pacific Hake Management

upsetsubduedΔιαχείριση

9 Νοε 2013 (πριν από 3 χρόνια και 9 μήνες)

85 εμφανίσεις

Pacific Hake Management
Strategy Evaluation

Joint Technical Committee

Northwest Fisheries Science Center,
NOAA

Pacific Biological Station,
DFO

School of Resource and Environmental Management,
SFU


Outline



Introduction


Review the MSE
workplan

objectives


Methods


Example
simulations


General behavior of the existing management
strategy


Performance
metrics


Summary
figures


Discussion and Conclusion


Introduction

Stock Assessment

Data

Harvest control rule

Catch recommendation

Catch that comes out of water

-

structure

-
selectivity shape

-
obs
/process error

-
spatial


-
survey design

-
survey frequency

-
converting backscatter to index

-

mathematical form

-

target harvest rate

-

inflection points

-
h
edging

-
u
n
-
quantified uncertainties

-
o
bjectives/constraints


-
s
patial restrictions

-
i
ndividual quotas

-
other fishing opportunities


Examples of some decisions

M
a
n
a
g
e
m
e
n
t

S
t
r
a
t
e
g
y

MSE
Workplan

Objectives


Use the 2012 base case as the operating
model.


Two objectives


Evaluate the performance of the harvest control
rule


Evaluate the performance of annual, relative to
biennial surveys.

Operating

Model

*
Stock dynamics

*
Fishery dynamics

*
True population

Management

Strategy

*
Data choices

*
Stock Assessment

*
Harvest control rule

Catch

Data

Performance

Statistics

*
Conservation
objectives

*
Yield objectives

*
Stability objectives

Feedback

Loop

Overview of the MSE Process

Closed
-
Loop Simulations

Use the MPD (not posterior medians, or other
quantiles
) for applying the harvest control rule

1960

1970

1980

1990

2000

2010

2020

2030

0.0

0.5

1.0

1.5

2.0

2.5

3.0

Year

S

S

B

t


------
Conditioning

period
------


(2012 assessment)

MSE Simulations

Cases Considered


No fishing


Perfect Information Case


Annual Survey


Biennial Survey

No fishing case


Set catches to zero, no assessment model


Exists to provide the first reference case to
describe how the stock will behave in the
absence of fishing

Perfect Information Case


We created a reference, perfect information
case where we simulated data with no error


The purpose of the perfect information case
was as follows:


To separate observation
vs

process error i.e.
variable data don’t affect management procedure
performance



to provide a standard relative to which a
comparison of the test (biennial and annual) cases
could be made


Perfect information case


Every year operating model simulates
dynamics of the stock (i.e. recruitments, stock
size
etc
)


No assessment model is fit, simulated catches
come from the application of the control rule
to the true stock

Biennial Survey Case


Every year operating model simulates dynamics of the
stock (i.e. recruitments, stock size
etc
)


Every odd year operating model simulates and
assessment model fits:


catch


s
urvey age
-
composition data


c
ommercial age
-
composition data


s
urvey biomass


In even years operating model simulates and
assessment model fits



catch


c
ommercial age composition data


Biennial survey

Annual Survey Case


Every year operating model simulates
stock
dynamics (
i.e. recruitments,
numbers at age,
etc
)


Every
year
operating model
simulates the
following data:


catch


survey age composition data


commercial age composition data


survey biomass



The assessment
model
fits these data and
returns the catch given the harvest control rule
back to the operating model

But remember


starting points are
not the same for each MSE run

Annual Survey

Some lessons about the performance
of the current management strategy

The assessment chases the latest
survey observation

Assessment errors are frequent

Aggregate Performance


Choose metrics that capture the
tradeoffs

between
conservation, variability in catch and total yield for specific
time periods.


Define short, medium and long time periods as
Short=2013
-
2015, Medium=2016
-
2020, Long=2021
-
2030.


The main conservation metric is the proportion of years
depletion is below 10%


The main variability in catch metric is the Average Annual
Variability in catch for a given time period.


For yield we used the median average catch


We’ve chosen what we think are the top six. We’d like to
discuss if others are needed.



Statistics Break
-

Medians
vs

Means

Average Annual Variability in Catch
(illustration)

Comparisons of Depletion, Catch and
AAV for All Cases

Summary for long
-
term depletion

Summary for long term AAV

Summary for long
-
term catch

Key Performance Statistics





Short
Term





Medium
Term





Long
Term



Percentage of years:

Per

Ann

Bie

Per

Ann

Bie

Per

Ann

Bie

Depletion above 40%

34.30%

35.90%

35.64%

28.95%

31.29%

32.67%

27.07%

29.54%

31.06%

Depletion below 10%

4.44%

6.61%

6.87%

0.94%

7.17%

8.59%

0.39%

5.39%

7.04%

Depletion between 10 and 40%

61.26%

57.49%

57.49%

70.11%

61.54%

58.74%

72.54%

65.08%

61.90%

MS closes fishery

0.00%

4.70%

3.90%

0.00%

8.51%

8.21%

0.00%

10.11%

13.61%

Key Performance Statistics II





Short
Term





Medium
Term





Long
Term



Medians of:

Per

Ann

Bie

Per

Ann

Bie

Per

Ann


Bie

Average catch

251

284

273

216

226

217

230

217

218

Average depletion

31.7%

31.4%

31.6%

27.9%

26.9%

27.8%

27.6%

27.3%

28.0%

AAV in catch (%)

36.6%

35.5%

32.5%

23.1%

34.1%

34.7%

23.3%

32.5%

33.2%

Additional Analyses


on the general performance of the harvest
control rule as a function of the default target
harvest
rate


Characterize the conservation, yield and
variability
tradeoffs
.

Alternative target harvest rates

Analysis of alternative target harvest
rates


The
hake treaty doesn't specify a target
depletion level,
only
a target
harvest
rate (
F40%)
and a control rule (40
-
10
).


This
makes it difficult to evaluate the efficacy of the
control
rule (i.e. relative to what?)


One additional curiosity that we considered was what
would the target harvest rate have to be in order to achieve
a range of target depletion levels


The
MSE can be used to explore how changes to the target
harvest
rate
might
affect depletion, AAV, and
average
catch.


This
is an exploration of trade
-
offs, not a proposal to
change the
hake
treaty
.

Discussion and Conclusion


The current management strategy (assessment
model formulation and F40%
-
40:10 rule)
performs as follows:


Median average depletion on the 7
-
17 year time
horizon ~28%, mean average depletion ~37%


Benefits of annual survey marginal


Assessment design results in chasing most recent
data


Since the survey is itself variable, this produces a high
probability of assessment error


Future work


It’s not an MSE until objectives have been defined
and the performance of alternative management
strategies evaluated against them.


The definition of these objectives and the JMC’s
key interested will determine if we consider:


O
perating
models that consider more complicated
hake life
-
history (i.e.
movement, Canada and US
areas)


A
lternative
management procedures to damp
variability


Etc.


Extra Slides


Other available performance metrics


First quartile depletion


Third quartile depletion


Median final depletion


Median of lowest depletion


Median of lowest perceived depletion


First quartile of lowest depletion


Third quartile of lowest depletion


First quartile of AAV in catch


Third quartile of AAV in catch


First quartile of average catch


Third quartile of average catch


Median of lowest catch levels


First quartile of lowest catch levels


Third quartile of lowest catch levels


Proportion with any depletion below SB10%


Proportion perceived to have any depletion below SB10%