flood risk management and the u.s. army corps of engineers actions for change

24
Relevant, Ready, Responsive, Reliable Flood Risk Management and the U.S. Army Corps of Engineers Actions for Change David Moser 1 , Martin Schultz 2 , Todd Bridges 2 and Brian Harper 1 US Army Corps of Engineers 1. Institute for Water Resources, Alexandria, VA 2. Engineer Research and Development Center, Vicksburg, MS 4 th International Symposium on Flood Defence: Managing Flood Risk, Reliability and Vulnerability Toronto, Canada May 6-8, 2008

Upload: vilmos

Post on 10-Jan-2016

33 views

Category:

Documents


0 download

DESCRIPTION

Flood Risk Management and the U.S. Army Corps of Engineers Actions for Change. 4 th International Symposium on Flood Defence: Managing Flood Risk, Reliability and Vulnerability Toronto, Canada May 6-8, 2008. David Moser 1 , Martin Schultz 2 , Todd Bridges 2 and Brian Harper 1 - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Flood Risk Management and the U.S. Army Corps of Engineers Actions for Change

Relevant, Ready, Responsive, Reliable

Flood Risk Management and the U.S. Army Corps of Engineers

Actions for Change

David Moser1, Martin Schultz2, Todd Bridges2 and Brian Harper1

US Army Corps of Engineers1. Institute for Water Resources, Alexandria, VA

2. Engineer Research and Development Center, Vicksburg, MS

4th International Symposium on Flood Defence: Managing Flood Risk, Reliability and Vulnerability

Toronto, CanadaMay 6-8, 2008

Page 2: Flood Risk Management and the U.S. Army Corps of Engineers Actions for Change

Relevant, Ready, Responsive, Reliable

Outline

• Twelve Actions for Change• Risk-informed decision framework (RIDF) for

Louisiana Coastal Protection and Restoration (LaCPR) Project.

• Illustrative example:– How risks, costs, and other tradeoffs are evaluated.– How risk and uncertainty is incorporated into the

decision process.– How stakeholder preferences are considered.

• How RIDF serves the Actions for Change

Page 3: Flood Risk Management and the U.S. Army Corps of Engineers Actions for Change

Relevant, Ready, Responsive, Reliable

Actions for Change (1-8): Systems and Risk-Based Approaches

1. Employ an integrated, systems-based approach

2. Employ risk-based concepts in planning and design

3. Continuously reassess and update policy

4. Employ dynamic independent review

5. Employ adaptive planning and engineering systems

6. Focus on sustainability

7. Review and inspect completed works

8. Assess and modify organizational behavior

Page 4: Flood Risk Management and the U.S. Army Corps of Engineers Actions for Change

Relevant, Ready, Responsive, Reliable

Actions for Change (9-12): Communication and Professionalism

Communication:

9. Effectively communicate risk

10. Involve the public in developing risk reduction strategies

Professionalism:

11.Manage and enhance technical expertise and professionalism

12. Invest in research

Page 5: Flood Risk Management and the U.S. Army Corps of Engineers Actions for Change

Relevant, Ready, Responsive, Reliable

Louisiana Coastal Protection and Restoration Project (LaCPR)

“…develop and present a full range of flood, coastal, and hurricane protection measures exclusive of normal policy considerations for south Louisiana.”

Factors not restricted to of national economic benefits and costs.

Consider full range of risks to people, cultural heritage, environment, property, and economy as well as project costs.

Five Planning Units in So. Louisiana

Charge:

Page 6: Flood Risk Management and the U.S. Army Corps of Engineers Actions for Change

Relevant, Ready, Responsive, Reliable

Risk-Informed Decision Framework

• Multi-Criteria Decision Analysis (MCDA)

• Identify plan that maximizes utility and optimizes the level of risk reduction given preferences

• Provides a means to:– Make tradeoffs between risks, costs, and other

decision outcomes

– Enhance transparency

– Account for aleatory and epistemic uncertainty

– Address stakeholder preferences and values

Page 7: Flood Risk Management and the U.S. Army Corps of Engineers Actions for Change

Relevant, Ready, Responsive, Reliable

Steps of Multi-Criteria Decision Analysis (MCDA)

1. Identify decision alternatives

2. Develop an objectives hierarchy and choose performance metrics

3. Assess preferences over objectives

4. Model performance of decision alternatives

5. Evaluate multi-attribute utility scores

6. Choose alternative that maximizes expected utility

7. Sensitivity analysis

Page 8: Flood Risk Management and the U.S. Army Corps of Engineers Actions for Change

Relevant, Ready, Responsive, Reliable

Illustrative Example

• Nine (9) flood risk-reduction alternatives and a no-action alternative

• Five (5) decision objectives and eight (8) performance metrics

• Three (3) uncertain planning assumptions– Stage frequency curve (ADCIRC Model)– Development pattern (employment growth rate and

population distribution)– Rate of relative sea-level rise

Page 9: Flood Risk Management and the U.S. Army Corps of Engineers Actions for Change

Relevant, Ready, Responsive, Reliable

Objectives Hierarchy and MetricsSuper-objectives Sub-objectives Performance Metric

Maximize human health and safety

Minimize life loss & health impacts among residents

1Risk to population within the inundation footprint (n)

Minimize economic losses from storm surge

Minimize residential and commercial property losses

2Risk to residential and commercial property ($)

Minimize disruptions to the local economy

3Risk to gross regional sales output ($)

Minimize costMinimize the cost of risk reduction projects

4 Life-cycle project costs ($)

Minimize environmental impacts

Maximize wetland acreage 5Net change in wetland acreage (acres)

Promote a sustainable ecosystem

6 Spatial integrity index

Minimize indirect impacts 7 Indirect impacts index

Minimize other social effects

Sustain cultural heritage 8Number of unique cultural centers protected (n)

Page 10: Flood Risk Management and the U.S. Army Corps of Engineers Actions for Change

Relevant, Ready, Responsive, Reliable

Preferences and ConsequencesC

DF

Stage

Stage frequencycurve (t)

Dam

age

Stage

Stage damage function (t)

Risk metric

mF 125.0

CD

F

mF 175.0

CDF for riskmetric (t)

(a) (b) (c)

Forecast ofE[damages]

Time (t)

E[d

amag

es]

(d)

CD

F

Stage

CD

F

Stage

Stage frequencycurve (t)

Dam

age

Stage

Dam

age

Stage

Stage damage function (t)

Risk metric

mF 125.0

CD

F

mF 175.0

Risk metric

mF 125.0

CD

F

mF 175.0

CDF for riskmetric (t)

(a) (b) (c)

Forecast ofE[damages]

Time (t)

E[d

amag

es]

(d)

Forecast ofE[damages]

Time (t)

E[d

amag

es]

(d)

• Assess stakeholder and decision maker’s preferences over objectives– Conduct interviews using an elicitation instrument to assess relative

importance. – Obtain relative preference weight on the sub-objective

• Model decision outcomes for each alternative in terms of performance metrics

Page 11: Flood Risk Management and the U.S. Army Corps of Engineers Actions for Change

Relevant, Ready, Responsive, Reliable

Multi-Attribute Utility

• Utility: An aggregate measure of relative satisfaction with modeled decision outcomes given objectives– Transform metrics to a 0 (worst) to 1 (best)– Weight transformed metrics by relative importance– Compensatory

• Six Planning Scenarios

iN

ii mVwU

1

Rate of Sea-Level Rise

(Relative to Observed)

Development Pattern None Moderate High

High employment & compact population

k = 1 k = 2 k = 3

BAU employment& dispersed population

k = 4 k = 5 k = 6

Page 12: Flood Risk Management and the U.S. Army Corps of Engineers Actions for Change

Relevant, Ready, Responsive, Reliable

Multi-Criteria Decision Analysis

Kkkjkj

Jj|UpUEMAXMaximize Expected Utility:

DevelopmentPattern

RelativeSea-levelRise

DecisionAlternatives

Performance Outcome (Utility score)

UncertaintyIn stage frequency

n

ii mFVwU

1

125.0

n

ii mFVwU

1

175.0

High

None

High & compact

BAU & dispersed

Alternative 1

Alt 2

Alt 3

Alt n

Alt N

Moderate

Page 13: Flood Risk Management and the U.S. Army Corps of Engineers Actions for Change

Relevant, Ready, Responsive, Reliable

Sensitivity of the Decision

• Choose alternative that maximizes expected utility

– Optimal level of risk reduction is the level associated with the chosen alternative given the decision maker’s preferences.

• How sensitive is the decision?– to stakeholder and decision maker’s preferences?– to the distribution of probability over sea-level rise

scenarios?

– to the pattern of development?

Page 14: Flood Risk Management and the U.S. Army Corps of Engineers Actions for Change

Relevant, Ready, Responsive, Reliable

Sensitivity to Preference Patterns

0.0

0.2

0.4

0.6

0.8

1.0

NoAction

A1 A2 A3 A4 A5 A6 A7 A8 A9

Decision Alternative

Exp

ecte

d U

tility

Metric 1 Metric 2 Metric 3 Metric 4

Metric 5 Metric 6 Metric 7 Metric 8

(c)

0.0

0.2

0.4

0.6

0.8

1.0

NoAction

A1 A2 A3 A4 A5 A6 A7 A8 A9

Decision Alternative

Exp

ecte

d U

tility

Metric 1 Metric 2 Metric 3 Metric 4

Metric 5 Metric 6 Metric 7 Metric 8

0.0

0.2

0.4

0.6

0.8

1.0

NoAction

A1 A2 A3 A4 A5 A6 A7 A8 A9

Decision Alternative

Exp

ecte

d U

tility

Metric 1 Metric 2 Metric 3 Metric 4

Metric 5 Metric 6 Metric 7 Metric 8 Metric Weight

1 0.202 0.353 0.204 0.055 0.056 0.057 0.058 0.05

Preference Pattern A

0.0

0.2

0.4

0.6

0.8

1.0

NoAction

A1 A2 A3 A4 A5 A6 A7 A8 A9

Decision Alternative

Exp

ecte

d U

tility

Metric 1 Metric 2 Metric 3 Metric 4

Metric 5 Metric 6 Metric 7 Metric 8 Metric Weight

1 0.302 0.303 0.004 0.305 0.006 0.007 0.008 0.10

Preference Pattern B

(d)

(a) (b)

0.0

0.2

0.4

0.6

0.8

1.0

NoAction

A1 A2 A3 A4 A5 A6 A7 A8 A9

Decision Alternative

Exp

ecte

d U

tility

Metric 1 Metric 2 Metric 3 Metric 4

Metric 5 Metric 6 Metric 7 Metric 8

(c)

0.0

0.2

0.4

0.6

0.8

1.0

NoAction

A1 A2 A3 A4 A5 A6 A7 A8 A9

Decision Alternative

Exp

ecte

d U

tility

Metric 1 Metric 2 Metric 3 Metric 4

Metric 5 Metric 6 Metric 7 Metric 8

0.0

0.2

0.4

0.6

0.8

1.0

NoAction

A1 A2 A3 A4 A5 A6 A7 A8 A9

Decision Alternative

Exp

ecte

d U

tility

Metric 1 Metric 2 Metric 3 Metric 4

Metric 5 Metric 6 Metric 7 Metric 8 Metric Weight

1 0.202 0.353 0.204 0.055 0.056 0.057 0.058 0.05

Preference Pattern A

0.0

0.2

0.4

0.6

0.8

1.0

NoAction

A1 A2 A3 A4 A5 A6 A7 A8 A9

Decision Alternative

Exp

ecte

d U

tility

Metric 1 Metric 2 Metric 3 Metric 4

Metric 5 Metric 6 Metric 7 Metric 8 Metric Weight

1 0.302 0.303 0.004 0.305 0.006 0.007 0.008 0.10

Preference Pattern B

(d)

(a) (b)

BAU employmentDispersed population

High employmentCompact population

Page 15: Flood Risk Management and the U.S. Army Corps of Engineers Actions for Change

Relevant, Ready, Responsive, Reliable

Prob(RSLR = High)

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Prob(RSLR = Moderate)

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Prob(RSLR = Low)

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Alternative A7Alternative A8Alternative A1Not feasible

How sensitive is the decision to sea-level rise assumptions?

BAU Employment Growth Rate and Dispersed Population

Page 16: Flood Risk Management and the U.S. Army Corps of Engineers Actions for Change

Relevant, Ready, Responsive, Reliable

Prob(RSLR = High)

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Prob(RSLR = Moderate)

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Prob(RSLR = Low)

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Alternative A7Alternative A8Alternative A1Not feasible

Changing the Development Scenario alters the Landscape

High Employment Growth Rate and Compact Population

Page 17: Flood Risk Management and the U.S. Army Corps of Engineers Actions for Change

Relevant, Ready, Responsive, Reliable

Risk-Informed Planning and the USACE Actions for Change

1. An integrated, systems-based approach to problem solving

2. Uses risk-based concepts in planning and decision making

3. Suitable for adaptive planning and engineering systems

4. Helps planners effectively communicate risks

5. Involves stakeholders in developing risk reduction strategies and decision making

Page 18: Flood Risk Management and the U.S. Army Corps of Engineers Actions for Change

Relevant, Ready, Responsive, Reliable

Back-up Slides

Page 19: Flood Risk Management and the U.S. Army Corps of Engineers Actions for Change

Relevant, Ready, Responsive, Reliable

High Employment Growth Rate andCompact Population Distribution

Prob(RSLR = High)

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Prob(RSLR = Moderate)

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Prob(RSLR = Low)

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Alternative A7Alternative A8Alternative A1Not feasible

Low Employment Growth Rate andDispersed Population Distribution

Prob(RSLR = High)

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Prob(RSLR = Moderate)

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Prob(RSLR = Low)

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Alternative A7Alternative A8Alternative A1Not feasible

(a) (b) High Employment Growth Rate andCompact Population Distribution

Prob(RSLR = High)

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Prob(RSLR = Moderate)

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Prob(RSLR = Low)

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Alternative A7Alternative A8Alternative A1Not feasible

Low Employment Growth Rate andDispersed Population Distribution

Prob(RSLR = High)

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Prob(RSLR = Moderate)

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Prob(RSLR = Low)

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Alternative A7Alternative A8Alternative A1Not feasible

(a) (b)

Page 20: Flood Risk Management and the U.S. Army Corps of Engineers Actions for Change

Relevant, Ready, Responsive, Reliable

Aleatory and Epistemic Uncertainty

• Aleatory uncertainty: – Natural variability that cannot be reduced– Example is the variability in storm surge

• Epistemic uncertainty:– Lack of knowledge, can in principle be reduced– Examples used in modeling performance outcomes

for LaCPR:• Characterization of the stage frequency curve for storm

surge• Projections of the future rate of sea-level rise• Population and employment growth rate projections

Page 21: Flood Risk Management and the U.S. Army Corps of Engineers Actions for Change

Relevant, Ready, Responsive, Reliable

Model Performance of Alternatives in Terms of Risk Metrics

CD

F

Stage

Stage frequencycurve (t)

Dam

age

Stage

Stage damage function (t)

Risk metric

mF 125.0

CD

F

mF 175.0

CDF for riskmetric (t)

(a) (b) (c)

Forecast ofE[damages]

Time (t)

E[d

amag

es]

(d)

CD

F

Stage

CD

F

Stage

Stage frequencycurve (t)

Dam

age

Stage

Dam

age

Stage

Stage damage function (t)

Risk metric

mF 125.0

CD

F

mF 175.0

Risk metric

mF 125.0

CD

F

mF 175.0

CDF for riskmetric (t)

(a) (b) (c)

Forecast ofE[damages]

Time (t)

E[d

amag

es]

(d)

Forecast ofE[damages]

Time (t)

E[d

amag

es]

(d)

• Calculate risk metrics under each plan– Integrate stage frequency curve w/ the damage function

• Characterize parameter uncertainty in the risk metrics– Integrate realizations of the stage frequency curve w/ damage

function

Page 22: Flood Risk Management and the U.S. Army Corps of Engineers Actions for Change

Relevant, Ready, Responsive, Reliable

Multi-attribute Utility (U )

worse)is (more

)()(

)(1

better) is (more )()(

)(

)(

mMINmMAX

mMINm

mMINmMAX

mMINm

mVi

i

i

U Multi-attribute utility scorew Preference weight on objectiveV(mi) Risk-neutral value scoreM Metric for the objectivei Index on objectives

iN

ii mVwU

1

Val

ue V

(mi)

Metric (m)

Averse

Seeking

Neutral

0

1

Value Function

BestWorst

Page 23: Flood Risk Management and the U.S. Army Corps of Engineers Actions for Change

Relevant, Ready, Responsive, Reliable

Six Planning Scenarios

Rate of Sea-Level Rise

(Relative to Observed)

Development Pattern None Moderate High

High employment & compact population

k = 1 k = 2 k = 3

BAU employment& dispersed population

k = 4 k = 5 k = 6

Page 24: Flood Risk Management and the U.S. Army Corps of Engineers Actions for Change

Relevant, Ready, Responsive, Reliable

Assess Preferences over Objectives

1. RANK OBJECTIVE 2. RATE

1 Minimize risks to residents 20

2 Minimize property losses 20

3 Minimize regional economic losses 15

4 Minimize project cost 15

5 Maximize wetland acreage 10

6 Promote sustainable ecosystem 10

7 Minimize indirect environmental impacts 5

8 Historic Properties Protected 5

100

Stakeholder and Decision Maker’s