ESSA Technologies

1
How ESSA has successfully used Decision Analysis
to overcome challenges in multi-objective
resource management problems
Developed by ESSA Technologies Ltd.
General overview January 10 2002
David Marmorek, Calvin Peters, Ian Parnell,
Clint Alexander
2
Common challenges in resource management

• Getting stakeholder groups to agree on a course
  of action, given multiple values and objectives
• Getting scientists to agree on which
  uncertainties most critically affect management
  decisions, and what decisions are most robust to
  these uncertainties
• Evaluating the costs and benefits of adaptive
  management - is it worth it?

3
How decision analysis can help with these
challenges

• It provides a toolbox for handling multiple
  objectives / values, and analyzing tradeoffs
  among these objectives
• It systematically analyzes the impacts of
  uncertainties on decisions
• It can be used to evaluate the ability of
  Adaptive Management experiments to improve
  decisions
• It provides a helpful way to integrate many
  techniques employed by managers and scientists
  (i.e. models, interactive workshops, sensitivity
  analysis) into products that better clarify
  management decisions

4
Three examples

• Getting scientists to agree PATH
• Getting stakeholders to agree Cheakamus
• Evaluating adaptive management Keenleyside

5
PATH Decision Context
Multiple historical changes in Columbia and Snake
River ecosystems and fisheries management
practices Endangered species listings for Snake
River salmon populations Multiple hypotheses and
uncertainties held by different groups of
scientists Duelling models representing these
hypotheses and uncertainties Best management
policies for species recovery?
6
PATH Washington State, US
7
Decision Analysis 8 elements

• 1. List of alternative management actions
• 2. Management objectives composed of performance
  measures (to rank management actions)
• 3. Uncertain states of nature (different
  hypotheses)
• 4. Probabilities of those states (to account for
  uncertainty)
• 5. Model to calculate outcomes of each
  combination of management action and hypothesised
  state of nature
• 6. Decision tree
• 7. Rank actions based on expected value of the
  performance measures and,
• 8. Sensitivity analyses.

8
Decision Analysis Basic Elements
9
PATH Decision Tree
10
Benefits of decision analysis in PATH

• Allowed evaluation of multiple hypotheses for 14
  uncertainties - scientists did not have to agree!
• Only 3 of these turned out to make a difference
  to the decision - created a common focus for AM,
  research
• Preferred actions were those which were most
  robust to the critical uncertainties (drawdown
  A3)
• Sensitivity analyses defined how much belief you
  would have to have in a given hypothesis to
  change decision

11
Recent Publications on PATH

• Marmorek, David R. and Calvin Peters. 2001.
  Finding a PATH towards scientific collaboration
  insights from the Columbia River Basin.
  Conservation Ecology 5(2) 8. online URL
  lthttp//www.consecol.org/vol5/iss2/art8gt
• Deriso, R.B., Marmorek, D.R., and Parnell, I.J.
  2001. Retrospective Patterns of Differential
  Mortality and Common Year Effects Experienced by
  Spring Chinook of the Columbia River. Can. J.
  Fish. Aquat. Sci. 58(12) 2419-2430
  http//www.nrc.ca/cgi-bin/cisti/journals/rp/rp2_to
  cs_e?cjfas_cjfas12-01_58
• Peters, C.N. and Marmorek, D.R. 2001.
  Application of decision analysis to evaluate
  recovery actions for threatened Snake River
  spring and summer chinook salmon (Oncorhynchus
  tshawytscha). Can. J. Fish. Aquat. Sci.
  58(12)2431-2446. ltsame web site as abovegt
• Peters, C.N., Marmorek, D.R., and Deriso, R.B.
  2001. Application of decision analysis to
  evaluate recovery actions for threatened Snake
  River fall chinook salmon (Oncorhynchus
  tshawytscha). Can. J. Fish. Aquat. Sci.
  58(12)2447-2458. ltsame web site as abovegt

12
Cheakamus WUP Decision Context

• British Columbia Hydro, Water Use Planning
  Stakeholder driven multi-objective consultation /
  decision process.
• No formal incorporation of uncertainty as for
  PATH
• Emphasis values, objectives, performance
  measures, trade off analysis (DA steps 1, 2, 5
  and 7).
• Used PrOACT approach (Smart Choices, Hammond et
  al 1999)

13
Cheakamus WUP Process
WUP Steps
14
Cheakamus WUPDecision Problem

• Select operating alternatives for Daisy Lake Dam
  that
• 1) recognize multiple water uses in the Cheakamus
  and Squamish Rivers, and
• 2) achieve a balance between competing interests
  and needs.

15
Cheakamus WUPObjectives and PMs
Power
First Nations
Recreation
Flooding
Fish
Aquatic Ecosystem
16
Cheakamus WUP Alternatives

• Consultative Committee specifies operating
  alternatives for Hydro operations model (AMPL).
• Basic constraints minimum flow at Brackendale
  gauge, minimum dam release.
• AMPL model produces 32 water years of flow data
  for these control points
• Flow data and other models used to calculate
  performance measures.
• Performance measures summarize consequences of
  alternatives for objectives.

17
Cheakamus WUP Consequences
18
Tradeoffs (or not)
Win-Lose
Win-Win
19
Cheakamus WUP Filtering

• Use PMs to Eliminate clearly inferior
  alternatives.
• Drop insensitive PMs (e.g., rafting).
• Drop Objectives that dont help the decision
  (e.g., flooding).
• Tradeoff analysis Even Swaps
• Elicit values behind decisions (e.g., rating
  exercises)
• Develop new alternatives to address concerns
  (e.g., chum spawning vs. rainbow trout rearing).

20
Keenleyside Problem Increased egg mortality
from dam operation
21
Problem II Uncertainty True whitefish
recruitment dynamics?
Given typical egg mortality, LARGE differences
in abundance associated with these curves
No reliable baseline information
22
Stage 1 - Decision Analysis w current uncertainty
23
Stage 1 Results Current Uncertainty
Objective Maintain least cost whitefish
population nearest to or greater than 45,000
adults
24
Stage 2 - Simulated learning from flow
experiments and monitoring
Uses same model and uncertain components but...
Actions are now alternative experimental flow
regimes monitoring programs
Assume a true relationship for population
dynamics with process error
25
What would you change if you knew the truth?If
population insensitive, then maximize power
revenues (85 kcfs)If population sensitive, then
minimize biological risk (60 kcfs)
26
Example Stage 2 Results Good monitoring is
critical for differentiating hypotheses flow
manipulation had less effect than expected.
27
AM can pay for itself
28
Is AM and monitoring worth it?
Yes If New information leads to choice of a
different management action that better satisfies
a particular objective, or rigorously confirms
that current management action is appropriate.
29
No definitive yes/no
Under AM practitioners control
Can evaluate implications using decision analysis?
Factor
Management objective (fish vs. power ) Ability
to do well designed experiments Initial level of
uncertainty in alternative hypotheses Magnitude
of natural variability in the system What
truth really is Inherent sensitivity of best
action to uncertainty
Yes Yes Maybe No No (cant know without
doing the experiment) No
Yes Yes Yes Yes Yes Yes
30
General Conclusions

• Value of AM potentially large
• Whether to proceed depends on the kind of
  system you are in (i.e. previous factors)
• Decision Analysis is very helpful for evaluating
  these benefits
• Determine which uncertainties have strongest
  effect on choice of best management decision
• Decisions more robust to uncertainties (reduces
  risk - integrates broader range of possible
  outcomes included)
• Include new information as revised probabilities
  on hypotheses

31
Decision Analysis - Summary