Risk-Based Management of Hurricane Preparedness and Recovery for a Highway Agency

1
Risk-Based Management of Hurricane Preparedness
and Recovery for a Highway Agency

• Presented to the
• Virginia Department of Transportation Steering
  Committee
• by the
• Center for Risk Management of Engineering Systems
• March 19, 2000

2
VDOT Steering Committee

• Virginia Department of Transportation
• Travis Bridewell
• Mac Clarke
• Perry Cogburn
• Jon DuFresne
• Stephany Hanshaw
• Steve Mondul
• Murali Rao
• Bob Rasmussen
• J.R. Robinson
• Gerald Venable
• Virginia Transportation Research Council
• Wayne Ferguson

3
Stakeholders

• Newport News Police
• Virginia Beach Police
• Norfolk Police
• Franklin Police
• Chester Police
• OBICI Hospital
• Red Cross
• Norfolk Emergency Operations Center
• Virginia Department of Emergency Services

• Virginia State Police
• VDOT Traffic Engineering
• VDOT ITS
• Sentara Hospitals
• Virginia Beach Fire
• Virginia Beach Office of Emergency Management
• Virginia Port Authority
• Virginia Department of Rail and Public
  Transportation

4
University of Virginia

• Graduate Student
• Richard D. Moutoux
• Undergraduate Students
• Ryan M. Finseth
• Linn H. Koo
• Clare E. Patterson
• Timothy J. Zitkevitz
• Faculty
• James H. Lambert, Research Assistant Professor of
  Systems Engineering
• Yacov Y. Haimes, Quarles Professor of Systems and
  Civil Engineering
• Garrick E. Louis, Assistant Professor of Systems
  Engineering
• Project website http//www.virginia.edu/risk/rec
  overy

5
Progress Reviews

• Stakeholders workshop February 28, 2000
• Hampton Roads Planning District Commission
  (HRPDC) April 5, 2000
• Virginia Department of Emergency Management
  (VDEM) October 6, 2000
• Steering committee December 2, 1999 and
  December 12, 2000

6
Overview of Presentation

• Introduction
• Overview of approach
• Risk-based prioritization of recovery
• Decision support for resource allocation
• Agency coordination of recovery schedule
• Trade-off analysis of recovery and preparedness
  alternatives
• Equipment enhancement
• Use of hurricane forecasts

7
Motivation

• Restore mobility as quickly as possible
• Currently no formal highway recovery methodology
  exists
• Justify aid from FEMA and FHWA
• Potential for 30-60 billion in losses for a
  category IV hitting Tidewater, Richmond, or
  Northern Virginia (Source Post Hurricane
  Recovery Workshop, Insurance Institute, 1997)

8
Saffir-Simpson Hurricane Categories
9
U.S. Mainland Hurricane Strikes 1900-1996 (NHC,
1999)
10
Hurricane Floyd

• Hurricane Floyd hit Suffolk District in
  mid-September, causing significant flood damage
• 56 deaths made Floyd the deadliest US hurricane
  since Agnes in 1972
• Roads did not receive much wind damage, but
  flooding closed many roads in Virginia

11
Hurricane Floyd (cont.)
12
Project Goal

• The goal of the effort is to improve
  hurricane preparedness and recovery of the
  Virginia Department of Transportation through the
  identification of planning and management options
  and the assessment and evaluation of the
  associated costs, benefits, and risks

13
Task 1 Review of Literature and Formation of
Advisory Committees

• Review and evaluation of past studies, theory and
  methodology, other agencies' experience, and
  databases
• Two advisory committees (1) Steering Committee,
  consisting primarily of VDOT personnel and (2) a
  Users' Group, made up of localities and other
  government agencies, e.g., emergency services.

14
Task 2. Extension of Prioritization Tool to GIS
Platform

• Integrate the mapping of critical facilities with
  VDOT's capabilities for geographic information
  system (GIS)
• Work closely with the Transportation Information
  Management Steering Committee (TIMSC) and current
  managers of VDOT information systems, including
  GIS, to ensure compatibility.

15
Task 3. Incorporation of Localities and
Additional Critical Facilities

• Local jurisdictions
• Intermodal connections (ports, airports, rail)
• Number of people served
• Logistic points (e.g., food warehouses, power
  generation facilities, water bottling plants,
  natural gas pipeline heads, collection and
  distribution points).

16
Task 4. Use of Hurricane Forecasts for VDOT
Operations

• National Hurricane Center and others
• The effort will demonstrate the efficacy of
  probabilistic hurricane forecasts in support of
  various VDOT planning and management functions.
• Capture the impacts of current decisions to
  future options

17
Task 5. Modeling for Agency-Wide Preparedness and
Recovery

• Hierarchical holographic modeling (HHM) will be
  used to classify overlapping and connected
  functions, divisions, and performance metrics
• Similar studies for the DoD, FBI, and PCCIP
• Foundation for resource allocation and
  coordination within and outside agency

18
Task 6. Resources, Databases, and Software
Task 7. Reports, Presentations, and Workshop
19
Overview of Approach
20
HHM for Preparedness
21
Abbreviated HHM
22
Organizational HHM
23
Sample Trade-Off Analysis
24
Trade-Off Analysis (cont.)
Option3
Option2
Option1
Status Quo
25
Risk-Based Prioritization of Recovery
26
Features of Priority Setting

• Importance of roads and intersections based on
  critical facilities and condition of road network
• Restoring or replacing damaged equipment
• Critical facilities are those necessary for a
  communitys well-being

27
Classification of Critical Facilities
28
Critical Facilities
29
Geographic Information Systems

• Build on an existing GIS database of VDOT roads
• Network modeling used to create prioritization
  tool
• Using Microsoft Excel for optimization and user
  functionality

30
Electronic Road Map

• Two sources for electronic road maps
• VDOTs Network Level Basemap
• Used to establish road system from which network
  model will be created
• Census maps obtained from UVA Library
• Used for geocoding addresses

31
Example of Arcview Map
32
Facility Data Acquisition

• Worked with the Hampton Roads Planning District
  Commission to get data on locations of critical
  facilities
• Facility data also collected by geocoding
  addresses in Arcview
• Critical facilities stored by category and
  sub-category

33
Norfolk Roads and Facilities
34
Data Summary

• Data gathered includes critical facilities,
  population, highway mileage, and connections
• Data is imported into Excel using a grid system
• Data is used to create a prioritization model in
  Excel

35
Excel Grid Method

• Microsoft Excel is used for the prioritization
  model
• A map is in the background of a spreadsheet for
  spatial reference
• In addition to being much easier for VDOT to use,
  the spreadsheet format makes modeling
  straightforward

36
Sample Excel Grid Method
37
Arcview Grid
38
Prioritization in Excel

• Spreadsheet layers contain facilities, road
  mileage, population, etc.
• Priority metrics are calculated across the grid
• Priorities are reflected by color-coding grid
  elements

39
Sample Categorical Prioritization
40
Decision Criteria

• The following types of data can be used for
  spreadsheet prioritization
• Population
• Roadway mileage by road type (Interstate, US
  highway, or primary route) three different
  measurements
• Critical facilities by category (eight
  categories)
• All combinations are possible

41
Operating the Prioritization Tool

• All that is required is an entry of one or zero
  in each box of this table
• A one includes that category in the calculation
• This example does not break down the road types
  or the many categories of facilities

42
Prioritization Tool Example
43
Prioritization Example (cont.)
44
Calibration of Tool
45
Local and Region-wide Models
46
Sensitivity Analysis
47
Temporal Priorities

• The model will prioritize recovery differently
  based on short, medium, and long-term goals
• Certain facilities are most important immediately
  after a disaster, and others are needed further
  down the road
• Since the Excel model includes data for all
  categories of facilities, roads, and population,
  these changes will be simple to implement

48
Phases of Preparedness
.
.
.
49
Decision Support for Resource Allocation
50
Motivation

• Hurricane Floyd struck Virginia in September 1999
  and resulted in severe damage from flooding
• Over 100,000 phone calls to Emergency Operations
  Center as a result
• Large scale of disaster makes it difficult to
  appreciate where the resources are going

51
Objective

• To develop a methodology to support in
  decision-making for resource allocation and in
  communicating and substantiating its rationale
  for its decisions regarding resource allocation
  for highway recovery from a natural disaster.

52
Factors Impacting Resource Allocation

• Tool tests recovery projects against multiple
  objectives in order to offer an idea of priority
  of the projects
• Effective performance criteria needed
• Risk reduction
• Performance gain
• Resources

53
Data Collection

• Case study performed using data from Hurricane
  Floyd recovery supplied by VDOT
• Data includes statistics from 39 counties/cities
  and 15 types of recovery/preparedness activity
• Data provided by Perry Cogburn, 2000.

54
Data Collection (contd)
55
(No Transcript)
56
(No Transcript)
57
Data Collection (contd)

• Performance indices must be generated from the
  data
• Lack of risk and performance data lead to the
  development of new indices
• Average Daily Traffic(ADT)
• For each road type(interstate, primary,
  secondary)
• Population Density
• For all 39 counties/cities

58
Hurricane Floyd Projects
59
(No Transcript)
60
(No Transcript)
61
POPULATION PER SQUARE MILE
AVERAGE DAILY TRAFFIC
62
Comparative Analysis of Delays in Hurricane
Recovery

63
Motivation

• Individual delays can be modeled in PERT or CPM
  as in prior UVa/VTRC efforts
• Need for systematic reporting of delay conditions
• Need for coordination to reduce the
• overall time to recovery

64
Data Collection

• Interviews of Virginia agencies involved in the
  recovery process
• Specifically the situations in which they are
  depending or dependent on VDOT
• Currently conducting interviews of Florida
  agencies

65
Data Collection (cont.)

• Meeting with Susan Maddox Toth of VDOTs EOC in
  Fall 2000
• Liaison with VDEM in Fall 2000 who are making
  similar efforts
• Past report by UVa/VTRC regarding a specific
  delay

66
Outline of Interviews with DOTs and Agencies

1. What are the cases in which your agency waited on
   VDOT to be able to start a recovery activity?
   What are the cases in which VDOT was waiting on
   your agency? (For example, authorization,
   materials, personnel, etc.)
2. If not in the past, can you see your agency
   waiting on VDOT or VDOT waiting on your agency
   for a recovery activity in the future? Under what
   circumstances?

67
Outline of Interviews with DOTs and Agencies
(cont.)

• Is there anything VDOT could have done to better
  minimize the delay of recovery? Does your agency
  have any suggestions for improvement?
• Is there a particular system or geographic area
  with which your agency is concerned for the
  recovery?

68
Virginia Agencies Contacted
Attended Stakeholders Meeting 2/00
69
Virginia Agencies Contacted (cont.)
Attended Stakeholders Meeting 2/00
70
Virginia Agencies Contacted (cont.)
Attended Stakeholders Meeting 2/00
71
Sample of Delay Scenarios
72
Sample of Delay Scenarios (cont.)
73
Delay Scenarios

• The following are the delay scenario descriptions
  that are indicated in the table above
• 2A A potential delay could take place between
  the time when the Henrico County Division of Fire
  requests materials over the phone from VDOT such
  as sand bags and the time when they actually
  receive it.
• (R. C. Dawson, Jr., Deputy Fire Chief, Henrico
  County Division of Fire, Oct. 9, 2000)  
• 2B A potential delay could take place between
  the time when the Henrico County Division of Fire
  requests equipment over the phone from VDOT such
  as traffic barricades and the time when they
  actually receive it.
• (R. C. Dawson, Jr., Deputy Fire Chief, Henrico
  County Division of Fire, Oct. 9, 2000)

74
Delay Scenarios (cont.)

• 2C A potential delay could take place between
  Henrico County Division of Fire if there is an
  inadequate number of on-call personnel at VDOT
  during emergency response.
• (R. C. Dawson, Jr., Deputy Fire Chief, Henrico
  County Division of Fire, Oct. 9, 2000)  
• 11A Obici Hospital had to wait on VDOT for the
  availability to current, updated road status and
  closure information during the recovery of
  Hurricane Floyd.
• (Randy Vick, Obici Hospital, Oct. 17, 2000) 

75
Delay Scenarios (cont.)

• 11B Obici Hospital experienced delays during
  recovery from Hurricane Floyd because the road
  status information that was provided by VDOT was
  inaccurate.
• (Randy Vick, Obici Hospital, Oct. 17, 2000) 
• 12A The response of the Office of Emergency
  Medical Services in the future to an isolated
  area could be delayed if there is a road or
  bridge failure that is waiting to be repaired by
  VDOT.
• (C. Everette Vaughan, Jr., Director of Emergency
  Operations at the Office of Emergency Medical
  Services, Oct. 9, 2000)

76
Delay Scenarios (cont.)

• 12B The Office of Emergency Medical Services had
  to wait on VDOT for the availability to current,
  updated road status information during the
  recovery of Hurricane Floyd.
• (C. Everette Vaughan, Jr., Director of Emergency
  Operations at the Office of Emergency Medical
  Services, Oct. 9, 2000)
• 12C The Office of Emergency Medical Services
  response to Franklin, VA during the recovery from
  Hurricane Floyd was delayed because the road
  status and closure information provided by VDOT
  was inaccurate.
• (C. Everette Vaughan, Jr., Director of Emergency
  Operations at the Office of Emergency Medical
  Services, Oct. 9, 2000)

77
Delay Scenarios (cont.)

• 29A The Department of Conservation and
  Recreation could potentially be waiting on VDOT
  in the future to make a bridge or road passable.
•   (Corey Garyotis, Senior Floodplain Engineer,
  Dept. of Conservation and Recreation, Oct. 10,
  2000)
• 36A VDOT could potentially be waiting on the
  Dept. of Mines, Minerals, and Energy in the
  future to provide geological information or
  information on where road building materials can
  be found.
• (Cheryl Cashman, Dept. of Mines, Minerals, and
  Energy, Oct. 20, 2000)
• 36B see 36A

78
Inter-Agency Delays
79
Measuring Delays
80
Measuring Severity
The severity of the length of time of delay is
relative to the time horizon - Short Term
(hours, days), - Medium Term (days, weeks) -
Long Term (months, years)
81
Example Delay Scenario

• A power line is in a tree blocking a primary road
• VDOT waits on the electric utility to check
  status of line before clearing tree from road
• The road is impassable until the tree is removed

82
Example Delay Scenario
83
Example Delay Scenario
HIGH
MODERATE
LOW
84
Comparing Multiple Delay Scenarios
85
Trade-Off Risk-Based Analysis
86
Tradeoff Analysis
87
Examples of Vulnerable Systems

• Equipment
• Signs, signals, and lights
• Road systems
• Roads, bridges, and tunnels
• Smart highway systems
• Motion detectors, cameras, and traffic alert signs

88
Types of Preparedness

• Redundancy Keeping additional spares on hand or
  adding extras to the system
• Resilience A system that can bounce back after
  a hurricane.
• Robustness The physical strength of a system

89
Planning Horizons

• Pre-event planning horizons
• Short term Preparing the system once a
  hurricane watch has been announced (hours to days
  before)
• Medium term Preparing the system at the start
  of the upcoming hurricane season (weeks to months
  before)
• Long term Preparing the system for an unknown
  hurricane in the future (years before)

90
Planning Horizons (cont.)

• Post-event planning horizons
• Short term Repairing the system immediately
  after or during the hurricane (hours to days
  after)
• Medium term Repairing the system after all
  short term problems are fixed (weeks to months
  after)
• Long term Repairing the system after all other
  non-long term problems are fixed (months to
  years after)

91
Assessments of Benefits

• Time savings How much time VDOT saves in man
  hours from the alternative
• Cost savings How much money is saved from each
  alternative
• Lives saved Number of human lives saved from
  the alternative
• Economic impact Impact on the society and the
  commerce in the area
• Environmental impact Impact on not only the
  environment but also the animals in the area too
• Private equipment saved Amount of personal
  property saved from the alternative

92
Mechanisms of Hurricane Impacts and Preparedness

• Maximum wind velocity Top maximum stainable
  wind that the road system can withstand
• Storm surge height Maximum amount of water in
  flooding and storm surge height that the road
  system can withstand
• Sustainable traffic flow Maximum amount of
  traffic that the road system can handle without
  failing

93
Saffir-Simpson Hurricane Categories
94
Example 1 Smart Highways
95
Example 1 Planning Horizons
96
Example 1 Assessment of Benefits
97
Example 1 Mechanisms of Hurricane Impact and
Preparedness
98
Equipment Enhancement
99
Example of Damage Assessment
100
Multiple Enhancement Levels

• Three enhancement levels for each grade or
  performance measure are shown in this example

101
Mechanism of Hurricane Impact to Road Systems
Assessments of (High, Med, Low) are tentative
102
Example of Data Requirements
103
Example of an Enhancement Trade-Off Analysis
104
Example Trade-Off Analysis
105
Sample User Input Worksheet
This is only a screen shot. Values within model
are not realistic
106
Sample Impact Values
This is only a screen shot. Values within model
are not realistic
107
Use of Hurricane Forecasts
108
Use of Hurricane Forecasts
109
Use of Forecasts (cont.)
110

• Forecasts
• Policy
• Options

111
Forecasts Tradeoff Analysis
112
Forecasts Tradeoff Analysis
113
Summary Recommendations
114
Priority Setting

• Consider a systematic approach to
  priority-setting for recovery
• Adopt the grids for priority-setting with various
  grid-size resolutions (District, Residency,
  smaller)
• Adopt the demonstrated metrics (populations,
  mileages,stakeholder facilities, etc.)
• Consider adding a metric to represent the degree
  of recovery
• Use the developed software downloads

115
Resource Allocation

• Consider a systematic approach to resource
  allocation for recovery
• Represent the variety of recovery projects across
  regions
• Discover the balance among all project impacts
  and costs
• Use the approach to improve the allocation of
  resources to diverse projects
• Project from past storms to the needs arising
  from future storms

116
Management of Delays

• Consider a systematic approach to delays in
  preparedness and recovery
• Characterize delays in the short, medium, and
  long terms, pre- and post-event
• Identify intra- and inter-agency dependencies
• Use multiple metrics to compare the diverse
  sources of delay
• Focus resources on the delays that are most
  critical to the overall recovery

117
Risk-Cost-Benefit Analysis

• Consider a systematic approach to cost-benefit
  analysis of recovery/preparedness
• Identify the variety of options for redundancy,
  robustness, and resilience
• Use the developed software to prepare
  cost-benefit analyses for diverse systems
• Consider the use of seasonal and monthly
  forecasts to aid in preparedness
• Continue to update the hierarchical holographic
  model of hurricane preparedness and recovery