Population Projections policy analysis

1
Population Projections(policy analysis)

• Fish 458 Lecture 21

2
Policy Evaluation-I

• It is often the objective for developing and
  fitting a model is to address what if
  questions. What is the impact of
• removal limits (quotas individual / Olympic)
• time / area closures
• gear restrictions (number of pots, traps,
  gillnets)
• bag limits
• minimum / maximum sizes and
• vessel numbers / size of vessels.

3
Policy Evaluation - II

• We are often not looking for optimal policies.
  Rather, we want to identify polices that are
  robust to
• Estimation error.
• Uncertainty regarding the true model.
• Implementation uncertainty.
• Environmental variability and environmental
  change.
• Optimal policies can often be found if we know
  the true model but these may perform poorly if
  applied to the wrong model.

4
Policy Evaluation-III(objectives and tactics)

• Policies are based on choosing tactics (quotas,
  minimum sizes, closed areas) to achieve
  management objectives / goals.
• Corollary - if we dont know the management
  objectives we cannot (sensibly) compare different
  policies.
• Problem often the decision makers have not
  agreed on any objectives (or are unwilling to
  state their actual objectives publicly).

5
Policy Evaluation-IV(objectives and tactics)

• We distinguish between high-level objectives
  (e.g. conserve the stock) and operational
  (quantitative) objectives (the probability of
  dropping below 0.1K should not be greater than
  0.1 over a 20-year period).
• Many decision makers confuse the tactics (what to
  do next year) with the objectives (why are we
  doing what we are doing next year).

6
Objectives for Fisheries Management(typical
high-level objectives)

• High level objectives arise from
• National legislation (MMPA, Magnusson-Stevens
  Act, ESA).
• International Agreements (CCAMLR, IWC, UN Fish
  Stocks Agreement).
• Court decisions.

7
Objectives for Fisheries Management(Objectives
for commercial whaling)

• Acceptable risk level that a stock not be
  depleted (at a certain level of probability)
  below some chosen level (e.g. some fraction of
  its carrying capacity), so that the risk of
  extinction of the stock is not seriously
  increased by exploitation
• Making possible the highest continuing yield from
  the stock and
• Stability of catch limits.
• The first objective was assigned highest priority
  but was not fully quantified.

8
Objectives for Fisheries Management(Australian
Fisheries Management Authority)

• Implementing efficient and cost-effective
  fisheries management on behalf of the
  Commonwealth
• Ensuring that the exploitation of fisheries
  resources and the carrying on of any related
  activities are conducted in a manner consistent
  with the principles of ecologically sustainable
  development and the exercise of the precautionary
  principle
• Maximising economic efficiency in the
  exploitation of fisheries resources
• Ensuring accountability to the fishing industry
  and to the Australian community and
• Achieving government targets in relation to the
  cost recovery.

9
Operational and High-level objectives

• Operational objectives describe the high-level
  objectives quantitatively.
• Preserve biodiversity (have at least 80 of all
  species protected in a system of reserves).
• Protect endangered species (have an 80
  probability that all currently endangered species
  are no longer endangered within 50 years).
• Protect ecosystem functioning (who knows what
  exactly what this means??)

10
Techniques for Policy Evaluation

• We can sometimes evaluate the implications of a
  policy analytically (e.g. the impact of changes
  in fishing intensity on yield-per-recruit).
• More commonly, we have to evaluate policy
  alternatives using Monte Carlo simulation
  methods.
• Specify the high-level management objectives.
• Specify the operational management objectives.
• Develop models of the system to be managed
  (including their uncertainty).
• Use simulation to determine the implications of
  each policy.
• Summarize the results.

11
Projecting Forward - I

• Define the state of the system in the first year
  of the projection.
• Calculate the catch limit based on the current
  state of the system.
• Project ahead one year (there may be
  implementation error at this stage) and update
  the dynamics.
• Repeat steps 2-3 for each future year.
• Repeat steps 1-4 many times.

12
The Simplest Decision Rules

• Constant catch (b0).
• Constant harvest rate (a0).
• Constant escapement (alt0).

13
The Simplest Decision Rules
(a10,b0)
(a0,b10)
(a-2.5,b12.5)
14
Evaluating the Simplest Rule

• Model of the state of the system (Schaefer
  model)
• This a deterministic model so we only have to do
  a single simulation as there is no uncertainty.

15
Average Catch / Population Sizevs. slope and
Intercept
500
Intercept
0
500
Intercept
0
16
Extending to a Stochastic Model

• Model of the state of the system (Schaefer
  model)
• This is now a stochastic model so we do 100
  simulations (?p0.1).

17
Catch and Population Size Trajectories
18
Average Catch / Population Size / CVvs. slope
and Intercept
Between simulation CV of average catch
19
Average catch vs. Population Size
20
CV of catch vs. Average Catch
21
Allowing for Errors in Stock Assessment

• We now allow for correlated errors when
  conducting assessments (if this years assessment
  is wrong, next years is also likely to be wrong)
  
• This approach to modeling assessment errors
  ignores biases in assessment results also
  assessment errors are unlikely to be log-normally
  distributed.

22
Allowing for Errors in Stock Assessment
Measuring the within-year variance in catches
No Stock Assessment Errors
With Stock Assessment Errors
23
Going Beyond the Simple Case

• Rather than assume assessment errors are
  log-normally distributed, simulate the process of
  conducting annual assessments (this is highly
  computationally intensive).
• Examine strategies designed to achieve specific
  management objectives (e.g. select catch limits
  so that the probability of recovery equals a
  desired level).

24
Readings

• Burgman et al. (1993) Chapter 3.
• Hilborn and Walters (1992) Chapters 15-18.
• Quinn and Deriso (1999) Chapter 11.