PACIFIC EARTHQUAKE ENGINEERING

RESEARCH CENTER

Toward Resilient Communities: A Performance-

Based Engineering Framework for Design and
Evaluation of the Built Environment

Michael William Mieler

Department of Civil and Environmental Engineering
University of California, Berkeley

Bozidar Stojadinovic
Department of Civil, Environmental, and Geomatic Engineering
Swiss Federal Institute of Technology (ETH) Zurich

Robert J. Budnitz
Earth Sciences Division
Lawrence Berkeley National Laboratory

Stephen A. Mahin
Department of Civil and Environmental Engineering
University of California, Berkeley

Mary C. Comerio
Department of Architecture
University of California, Berkeley

PEER 2013/19
SEPTEMBER 2013
 Disclaimer

The opinions, findings, and conclusions or recommendations

expressed in this publication are those of the author(s) and
do not necessarily reflect the views of the study sponsor(s)
or the Pacific Earthquake Engineering Research Center.
Toward Resilient Communities: A Performance-
Based Engineering Framework for Design and
Evaluation of the Built Environment

Michael William Mieler

Department of Civil and Environmental Engineering
University of California, Berkeley

Bozidar Stojadinovic
Department of Civil, Environmental, and Geomatic Engineering
Swiss Federal Institute of Technology (ETH) Zurich

Robert J. Budnitz
Earth Sciences Division
Lawrence Berkeley National Laboratory

Stephen A. Mahin
Department of Civil and Environmental Engineering
University of California, Berkeley

Mary C. Comerio
Department of Architecture
University of California, Berkeley

PEER Report 2013/19

Pacific Earthquake Engineering Research Center
Headquarters at the University of California, Berkeley
September 2013
ii
 ABSTRACT

A community is a dynamic system of people, organizations, and patterned relationships and

interactions. Most of these relationships and interactions are physically supported by a
community’s built environment, a complex and interdependent network of engineered
subsystems and components, including buildings, bridges, pipelines, transmission towers, and
other structures. As a result, the built environment plays a crucial role in enabling a community
to function successfully, providing the foundations for much of the economic and social
activities that characterize a modern society. Natural hazards such as earthquakes, hurricanes,
and floods can damage a community’s built environment, which in turn can disrupt the security,
economy, safety, health, and welfare of the public. In response, many communities have
developed and implemented regulatory frameworks to ensure that individual parts of the built
environment attain minimum levels of performance.
This report proposes a performance-based engineering framework for design and
evaluation of the built environment in order to improve the overall resilience of communities to
natural hazards. It begins by examining the regulatory framework currently used in the United
States to design and evaluate a community’s built environment to withstand the effects of
earthquakes and other natural hazards. Specifically, it analyzes building codes and other
engineering standards that establish performance expectations for buildings and lifelines. To this
end, the report first identifies attributes or characteristics of an ideal regulatory framework. Then,
using these attributes as a guide, it discusses both the strengths and shortcomings of the current
regulatory framework. The most significant shortcoming of the current framework is its lack of
an integrated, coordinated, and comprehensive approach to establishing performance
expectations for individual components of the built environment. Consequently, performance
objectives for the individual components are not tied to broader performance targets for the
community, primarily because these community-level performance objectives typically do not
exist.
The growing interest in resilient and sustainable communities necessitates an updated
regulatory framework, one that employs an integrated, coordinated, and comprehensive approach
to account for the built environment’s numerous subsystems, components, and interactions. The
regulatory framework currently used in the United States to design, analyze, and regulate
commercial nuclear power plants to assure their safety offers a promising template for
communities to follow. Despite obvious differences in function and configuration, both
communities and nuclear power plants are multi-faceted, dynamic systems comprising many
interacting subsystems and components that cut across a diverse range of disciplines and
professions. The current nuclear regulatory framework handles these numerous subsystems,
components, and interactions in a consistent and logical manner, informed partly by an explicit
set of system-level performance expectations for the nuclear power plant. Furthermore, the tools
and procedures employed by the current nuclear regulatory framework have been implemented
successfully and refined extensively over the past several decades, resulting in significant
improvements in both the understanding of how these complex, dynamic systems behave and the
efficacy of the regulatory framework itself.
iii
 This report studies the current regulatory framework for nuclear power plants and, using
recent developments from the rapidly evolving fields of community resilience and lifeline
interdependency, adapts it for use in a community setting. To this end, the report proposes and
describes an integrated engineering framework for design and evaluation of a community’s built
environment. This new framework provides a transparent, performance-based, risk-informed
methodology for establishing a consistent set of performance targets for the built environment
and its various subsystems and components to enhance the overall resilience of the community.
This report also presents several conceptual examples that illustrate implementation of the
proposed framework, including a demonstration of how to develop seismic performance targets
for a new residential building from a community-level performance goal.
Ultimately, the work presented herein has the potential to change the way engineers,
planners, and other stakeholders design and evaluate a community’s built environment. The
engineering framework proposed in this report provides a comprehensive, integrated, and
coordinated methodology for planners and policymakers to set community-level performance
targets and, subsequently, for engineers to calibrate the designs of individual components to meet
these community-level performance targets. Though additional work is required, the findings
presented in this report establish the foundations for a much-needed transformation from
engineering individual components of the built environment on a component-by-component
basis to engineering community resilience using an integrated and coordinated approach that
begins at the community level. Future iterations of the framework should aim to expand its scope
beyond disaster resilience to address and incorporate broader sustainability considerations, for
example, carbon footprint concerns, energy efficiency, resource consumption, and the
environmental impact of a community and its built environment.

iv
 ACKNOWLEDGMENTS

The publication of this report by the Pacific Earthquake Engineering Research Center
(PEER) is gratefully acknowledged. Any opinions, findings, and conclusions or
recommendations expressed in this material are those of the authors and do not necessarily
reflect those of PEER or any other sponsors.

v
vi
 CONTENTS

ABSTRACT .................................................................................................................................. iii 

ACKNOWLEDGMENTS .............................................................................................................v 
TABLE OF CONTENTS ........................................................................................................... vii 
LIST OF FIGURES ..................................................................................................................... xi 
LIST OF TABLES ..................................................................................................................... xiii 

1  INTRODUCTION..............................................................................................................1 
1.1  Overview and Background ...................................................................................3 
1.1.1  Current Regulatory Framework for Communities .......................................3 
1.1.2  Community Resilience .................................................................................5 
1.2  Proposed Engineering Framework ......................................................................8 
1.2.1  Scope ............................................................................................................8 
1.2.2  Intellectual Contribution ..............................................................................9 
1.3  Organization of the Report ...................................................................................9 

2  BACKGROUND AND DEFINITIONS .........................................................................13 

2.1  Literature Review ................................................................................................13 
2.1.1  Community Resilience ...............................................................................13 
2.1.2  Lifeline Interdependencies .........................................................................23 
2.2  Definitions .............................................................................................................26 
2.2.1  System ........................................................................................................26 
2.2.1.1  Types of components ...................................................................26 
2.2.1.2  Interactions .................................................................................28 
2.2.1.3  Correlation ..................................................................................28 
2.2.1.4  Phases in the lifecycle of a system ..............................................29 
2.2.1.5  System definition and analysis ....................................................29 
2.2.1.6  Attributes of an ideal system .......................................................30 
2.2.2  Hazard ........................................................................................................31 
2.2.2.1  Multiple effects of hazards ..........................................................32 
2.2.2.2  Hazard analysis ..........................................................................32 
2.2.3  Performance Measures ...............................................................................33 
2.2.4  Regulatory Framework ..............................................................................34

vii
 2.3  Attributes of an “Ideal” Regulatory Framework .............................................34 
2.3.1  Expansionist Framework ...........................................................................35 
2.3.2  Risk-Informed Framework.........................................................................35 
2.3.3  Comprehensive Framework .......................................................................35 
2.3.4  Performance-Based Framework.................................................................36 
2.3.5  Acceptable Provisions and Requirements ..................................................37 

3  COMMUNITIES..............................................................................................................39 
3.1  System Description...............................................................................................39 
3.1.1  Components ...............................................................................................40 
3.1.1.1  Structures ....................................................................................40 
3.1.1.2  Hardware ....................................................................................40 
3.1.1.3  People ......................................................................................40 
3.1.1.4  Organizations..............................................................................41 
3.1.1.5  Procedures ..................................................................................41 
3.1.1.6  Environments ..............................................................................42 
3.1.1.7  Interfaces.....................................................................................42 
3.1.2  Interactions .................................................................................................42 
3.1.3  Correlation .................................................................................................43 
3.2  Vulnerability to Hazards .....................................................................................44 
3.2.1  Types of Effects .........................................................................................44 
3.2.2  Direct Consequences ..................................................................................44 
3.2.3  Cascading Consequences ...........................................................................45 
3.3  Current Regulatory Framework ........................................................................45 
3.3.1  Overview ....................................................................................................46 
3.3.1.1  Buildings .....................................................................................46 
3.3.1.2  Lifelines ......................................................................................51 
3.3.2  Strengths ....................................................................................................53 
3.3.3  Shortcomings .............................................................................................53 

4  NUCLEAR DESIGN PHILOSOPHY ............................................................................57 

4.1  Overview ...............................................................................................................58 
4.2  Definitions .............................................................................................................58 
4.2.1  Undesired Outcomes ..................................................................................58 
4.2.2  Accidents and Accident Sequences ...........................................................59 
4.2.3  Vital Functions ...........................................................................................60 

viii
 4.2.4  Frontline and Support Systems ..................................................................61 
4.3  Design and Regulatory Philosophy.....................................................................61 
4.3.1  The Concept of Defense-in-Depth .............................................................61 
4.3.2  Risk-Informed Regulation .........................................................................62 
4.4  Performance Evaluation Tools ...........................................................................65 
4.4.1  Probabilistic Risk Assessment ...................................................................65 
4.4.2  Event Trees ................................................................................................67 
4.4.3  Fault Trees .................................................................................................69 

5  ADAPTATION OF NUCLEAR DESIGN PHILOSOPHY .........................................73 

5.1  Caveats ..................................................................................................................74 
5.2  Undesired Outcomes ............................................................................................74 
5.3  Vital Functions .....................................................................................................75 
5.3.1  Public Services ...........................................................................................76 
5.3.2  Housing ......................................................................................................76 
5.3.3  Employment ...............................................................................................77 
5.3.4  Education ...................................................................................................78 
5.4  Frontline and Support Systems ..........................................................................78 
5.5  Performance Evaluation Tools ...........................................................................80 
5.5.1  Dependency Matrices.................................................................................80 
5.5.2  Event Trees ................................................................................................82 
5.5.3  Fault Trees .................................................................................................82 

6  COMMUNITY EVENT TREES ....................................................................................83 

6.1  Applicability .........................................................................................................83 
6.2  General Structure ................................................................................................84 
6.3  Event Trees ...........................................................................................................86 
6.3.1  Public Services ...........................................................................................87 
6.3.2  Housing ......................................................................................................88 
6.3.3  Employment ...............................................................................................90 
6.3.4  Education ...................................................................................................92 
6.4  Synthesis................................................................................................................93 

7  EXAMPLE APPLICATIONS ........................................................................................97 

7.1  Example: establishing consistent performance objectives ...............................97 
7.1.1  Performance Targets for a Community......................................................98 
ix
 7.1.2  Performance Targets for the Vital Community Functions .......................100 
7.1.3  Performance Targets for Frontline and Support Systems ........................110 
7.1.4  Performance Targets for Individual Components ....................................111 
7.1.5  Implications..............................................................................................111 
7.2  Example: Computing Disruption to Public Services ......................................113 
7.2.1  Analysis Boundaries ................................................................................114 
7.2.2  System Importance Matrix .......................................................................115 
7.2.3  System Disruption Matrix ........................................................................118 
7.2.4  Service Disruption Matrix........................................................................120 
7.2.5  Service Importance Matrix ......................................................................121 
7.2.6  Total Disruption .......................................................................................122 
7.2.7  Implications..............................................................................................123 

8  CONCLUSIONS ............................................................................................................127 
8.1  Implications ........................................................................................................129 
8.2  Future Work .......................................................................................................130 
8.2.1  Refinement and Expansion of Community Event Trees..........................130 
8.2.2  Refinement and Expansion of the Methodology to Develop
Consistent Performance Targets for the Built Environment ....................131 

REFERENCES...........................................................................................................................135 

x
 LIST OF FIGURES

Figure 1.1  Matrix of performance standards and guidelines for lifelines. ................................7

Figure 2.1  Schematic of disaster recovery. .............................................................................16 

Figure 2.2  Overview of conceptual model of community recovery from earthquakes. ..........16 
Figure 2.3  Important community functions. ............................................................................20 
Figure 2.4  General performance objectives for San Francisco as a function of time. ............21 
Figure 2.5   Specific performance objectives for San Francisco's buildings and
infrastructure as a function of time. .......................................................................22 
Figure 2.6   Relationships among the seven types of components in a system. ........................27

Figure 3.1   Examples of lifeline interdependencies. ................................................................43 

Figure 3.2  Implicit performance objectives corresponding to IBC occupancy
categories. ..............................................................................................................50 
Figure 3.3  Matrix of standards and guidelines for lifelines. ...................................................52 
Figure 4.1  Graphical representation of a simple accident sequence. ......................................60 
Figure 4.2  The three levels of PRA used in the nuclear industry. ...........................................66 
Figure 4.3  Example of a simple event and fault tree. ..............................................................69

Figure 5.1  Dependency matrix corresponding to the support systems portrayed in

Figure 3.1. ..............................................................................................................81

Figure 6.1  Example of an event tree for an individual vital community function. .................86 
Figure 6.2  Housing event tree. ................................................................................................86 
Figure 6.3  Public services event tree. ......................................................................................88 
Figure 6.4  Employment event tree. .........................................................................................90 
Figure 6.5  Education event tree. ..............................................................................................93 
Figure 6.6  Single event tree obtained from combining the public services, housing,
employment, and education event trees. ................................................................95

Figure 7.1  Individual event trees for the vital community functions. ...................................101 
Figure 7.2  Combined event tree (sequences that trigger significant outmigration
identified by red boxes at the end of the branches). ............................................102 
Figure 7.3  Branch probabilities corresponding to 1% probability of significant
outmigration in the design basis earthquake. .......................................................105 

xi
Figure 7.4  Branch probabilities corresponding to 10% probability of significant
outmigration in the maximum considered earthquake. ........................................106 
Figure 7.5  Cumulative distribution functions for each vital community function. ...............108 
Figure 7.6  SPUR analysis of the current building code. .......................................................112 
Figure 7.7  Dependency matrix for public services and frontline and support systems.........117 
Figure 7.8  System importance matrix for public services and frontline and support
systems. ................................................................................................................118 
Figure 7.9  Examples of a service importance matrix. ...........................................................122 
Figure 7.10  Effectiveness of decreasing the vulnerability of individual frontline
systems. ................................................................................................................124 
Figure 7.11  Effectiveness of decreasing the vulnerability of individual support
systems. ................................................................................................................125 

xii
 LIST OF TABLES

Table 2.1  Four properties of resilience. .................................................................................14 

Table 2.2  Four dimensions of resilience. ...............................................................................14 
Table 2.3  Baseline resilience indicators for U.S. communities. ............................................17 
Table 2.4  Components of resilience. ......................................................................................18 
Table 2.5  Characteristics of a disaster-resilient community corresponding to a
specific component of resilience. ...........................................................................19 
Table 2.6  Four principal classes of lifeline interdependencies. .............................................25 
Table 2.7  Three types of lifeline failures. ..............................................................................25 
Table 2.8  Description of the seven types of components in a system. ..................................27

Table 3.1  Occupancy categories for buildings. ......................................................................48 

Table 3.2  Descriptions of the four seismic performance levels implicitly assumed in
the IBC. ..................................................................................................................49

Table 4.1  Levels of defense-in-depth. ....................................................................................62 

Table 4.2  Limitations associated with the use of the PRA methodology. .............................67 
Table 4.3  Explanation of the symbols used in Equation (4.4). ..............................................71

Table 5.1  Frontline systems for each vital function in a community. ....................................79 
Table 5.2  Support systems in a community. ..........................................................................80

Table 6.1  Reasons that businesses can be disrupted after an earthquake. .............................91

Table 7.1  Performance targets for each basic safety function (based on mean values). ......110 
Table 7.2  Methodology to estimate the disruption to public services. ................................113 
Table 7.3  Public services included in the analysis. ..............................................................114 
Table 7.4  Frontline and support systems included in the analysis. ......................................115 

xiii
xiv
1 Introduction

A community is a dynamic system of people, organizations, and patterned relationships and

interactions [Alesch 2005]. Most of these relationships and interactions are physically supported
by a community’s built environment, a complex and interdependent network of engineered
subsystems and components, including buildings, bridges, pipelines, transmission towers, and
other structures. As a result, the built environment plays a crucial role in enabling a community
to successfully function, providing the physical foundations for much of the economic and social
activities that characterize a modern society [O’Rourke 2007]. Natural hazards such as
earthquakes, hurricanes, and floods can damage a community’s built environment, which in turn
may disrupt the security, economy, safety, health, and welfare of the public.
In response, many communities have developed and implemented regulatory frameworks
to ensure minimum levels of performance for individual parts of the built environment. A
regulatory framework provides the legal and technical basis for allowing a system to operate
through all phases of its lifecycle. It comprises three basic elements: regulations, mechanisms for
enforcing the regulations, and guidance for satisfying the regulations. Regulations, which usually
carry the weight of law, include codes, standards, and other documents that specify the rules,
requirements, and provisions for a system and its parts. Enforcement is accomplished by the
agencies and organizations that are charged with promulgating and/or maintaining the
regulations. Guidance, which is typically optional in nature, refers to anything that aids in
satisfying the regulations, providing but one of many possible ways to satisfy the regulations.
This report examines the regulatory framework currently used in the United States to
design and evaluate a community’s built environment to withstand the effects of natural hazards.
In particular, it examines building codes and other engineering standards that establish
performance expectations, either implicit or explicit, for the built environment and its numerous
components and subsystems. This examination reveals several significant shortcomings. Most
crucially, the current regulatory framework approaches the design and evaluation of the built
environment on a component-by-component basis, generally treating each subsystem or
component as if it does not interact with or depend on other parts of the built environment. This
component-by-component approach results in a community in which most individual subsystems
and components of the built environment behave as intended; however, when aggregated, the
performance of and interaction among individual components can result in unacceptable
outcomes for the community.

1
 The growing interest in sustainable and resilient communities necessitates an updated
regulatory framework, one that employs an integrated, coordinated, and comprehensive approach
to account for the built environment’s numerous subsystems, components, and interactions. An
approach like this derives an understanding of the behavior of individual components by first
studying the behavior of the entire system, which stands in contrast to a component-by-
component approach that arrives at an understanding of the behavior of the system by first
studying the behavior of individual components [Bea 2007, 2008]. By focusing attention at the
system level first, performance objectives for individual components within the built
environment can be formulated in a way that is consistent with broad resilience goals for the
community.
The regulatory framework currently used to design, analyze, and regulate commercial
nuclear power plants in the United States offers a promising template for communities to follow.
Similar to a community, a nuclear power plant is a complex, dynamic system comprising many
interacting subsystems and components that cut across a diverse range of disciplines and
professions. The current nuclear regulatory framework handles these numerous subsystems and
components in a consistent and logical manner, informed partly by an explicit set of system-level
performance expectations for the nuclear power plant. To this end, the current nuclear regulatory
framework begins at the system level, identifying key functions that must be available in order
for a nuclear power plant to operate successfully. It then establishes performance targets both for
the overall nuclear power plant as a system and, subsequently, for its numerous subsystems and
components. Finally, in order to ensure the plant satisfies these targets, the framework requires a
detailed analysis of the system and its components in order to verify that the plant design
satisfies the required system-level performance targets. The tools and procedures employed by
the current nuclear regulatory framework have been implemented successfully and refined
extensively over the past several decades, resulting in significant improvements in both the
understanding of how these complex, dynamic systems behave and the efficacy of the regulatory
framework itself.
This report explores the opportunities and challenges that arise when adapting pieces of
the nuclear regulatory framework for use in a community setting. Throughout this process, the
report draws from several major studies from the rapidly evolving fields of community resilience
and lifeline interdependencies, including PCCIP [1997], Rinaldi et al. [2001], Bruneau et al.
[2003], Miles and Chang [2003, 2006, 2007], Poland et al. [2009], Twigg [2009], SERRI and
CARRI [2009], and Cutter et al. [2010]. Using the nuclear framework as a template, it leverages
findings from these studies to create an engineering framework that addresses the built
environment’s numerous components, subsystems, and interactions using an integrated,
coordinated, and comprehensive approach. In particular, the proposed framework uses findings
from Poland et al. [2009], SERRI and CARRI [2009], Twigg [2009] and Cutter et al. [2010], to
identify essential community functions that need to be maintained following a disaster. In
addition, it uses PCCIP [1997], Rinaldi et al. [2001], ALA [2004], and Poland et al. [2009] to
identify components and subsystems within the built environment (including their
interdependencies) that have significant roles in supporting these essential community functions.

2
 Ultimately, the framework described in this report can be used to ensure that individual
components of the built environment perform in a manner that supports broad, community-level
resilience goals. In this new framework, engineers will design individual structures and
components in much the same way they have in the past. However, the performance targets
specified by the framework for these individual components will be compatible with broader
performance targets established for the entire community.

1.1 OVERVIEW AND BACKGROUND

1.1.1 Current Regulatory Framework for Communities

A community is a complex and dynamic system of people, organizations, infrastructure, and

interactions. Equally complex, however, is the regulatory framework that governs how a
community and its numerous components must operate. This report focuses on a small but
important piece of this regulatory framework: the building codes and other engineering standards
that establish performance expectations, either implicit or explicit, for a community’s built
environment. The following paragraphs describe these documents and briefly discuss their
strengths and shortcomings.
At the heart of the current regulatory framework for buildings in the United States is the
International Building Code (IBC), a document that specifies minimum requirements for
buildings and other structures in order to safeguard the health, safety, and general welfare of the
public [ICC 2006]. Historically, the focus of modern building codes like the IBC has been to
“safeguard against major structural failures and loss of life, not to limit damage or maintain
function” [ICBO 1997]. To this end, modern building codes, when properly enforced, have been
effective at reducing casualties, as demonstrated in two recent earthquakes. On 12 January 2010
a magnitude 7.0 earthquake struck near Port-au-Prince, the heavily populated capital of Haiti.
The city lacks both a modern building code and a means to enforce it [DesRoches et al. 2011].
As a result, nearly half the buildings in Port-au-Prince, many of which were constructed using
materials and methods prohibited in the IBC, collapsed during the earthquake. While an exact
estimate may never be possible, the resulting destruction claimed the lives of approximately
300,000 Haitians [DesRoches et al. 2011]. In contrast, the magnitude 6.3 earthquake that struck
Christchurch, New Zealand, on 22 February 2011 claimed the lives of approximately 180 people
[EERI 2011]. Despite both earthquakes having similar intensity, the casualties in New Zealand
were a small fraction of those experienced in Haiti. The stark difference in the performance
stems largely from the New Zealand building code, which closely resembles the IBC and is
strictly enforced.
While the IBC has been effective at reducing the risk posed by earthquakes to life safety,
it has been less effective at addressing other kinds of risk. For example, its provisions allow
significant structural and nonstructural damage to occur in very rare, intense ground shaking as
long as it does not lead to collapse [BSSC 2009]. While the resulting damage may pose minor
risk to life safety, it can significantly impact the functionality of the building, which in turn can

3
impose substantial financial burden on its inhabitants and, ultimately, the community. For
example, the magnitude 6.3 earthquake that struck Christchurch, New Zealand, on 22 February
2011 caused extensive damage to buildings in the city’s Central Business District (CBD).
Approximately 50% of all buildings in the CBD were rendered unusable because they either
sustained significant structural damage or were located near hazardous buildings [EERI 2011].
Consequently, a significant portion of the CBD was closed for over a year [CERA 2012a],
affecting over 50,000 workers (25% of the city’s workforce) and approximately 6000 companies
or institutions [EERI 2011]. The Canterbury Earthquake Recovery Authority [CERA 2012b]
anticipates that it will take twothree years to demolish all heavily damaged buildings in the
CBD, thereby slowing redevelopment prospects. In general, damage that forces a business to
close for weeks or months may strain the finances of the owners, employees, and those who
depend on the goods or services it produces, including other businesses. Closure of many such
businesses, as happened in Christchurch, can result in a precipitous drop in tax revenues for local
governments and even a significant outmigration of residents and businesses. Issues like these
are beyond the consideration of the IBC.
As the 2011 earthquake in Christchurch highlights, the seismic performance levels
specified for individual buildings by the IBC and other modern building codes are inconsistent
and often inadequate when viewed from the perspective of the community. Typically, the
performance levels established by modern building codes reflect choices that balance the desire
to minimize initial construction costs with the need to ensure adequate levels of safety for the
building’s occupants [BSSC 2009]. Absent from this consideration, however, is the impact these
choices have beyond the owners and occupants of the building. For example, if an earthquake
damages a large apartment building and renders it unusable, it can impose significant financial
burden on the local government agencies that provide emergency housing to displaced residents.
It can also impact surrounding businesses, especially if emergency housing is located far from
the damaged apartment building, causing their customer base to disappear overnight.
Consequently, performance objectives that are appropriate for the safety of a building’s
occupants may not be appropriate for the general welfare and overall resilience of the
community.
Buildings, which fall under the purview of the IBC, are but one piece of a community’s
built environment. Lifelines such as electric power, water, and telecommunications also play a
crucial role in a community’s ability to function successfully, both on a daily basis and in the
aftermath of a major disaster. In spite of their importance, “seismic performance standards for
lifelines vary widely and are not tied to generally applicable public policies for reducing risk or
for ensuring community resilience in the face of a major earthquake” [Barkley 2009]. Figure 1.1
lists the various performance standards and guidelines that exist for each of the major lifelines.
Note that for electric power, water, wastewater, and telecommunications, performance standards
that address system reliability do not exist. The American Lifelines Alliance defines system
reliability as “a component of design referring to practices that are specifically developed to
provide reasonable assurance that consequences of a natural hazard on system service will meet
the goals established by stakeholders (owners, operators, regulators, insurers, customers, and
users)” [ALA 2004]. This lack of system-level performance standards for lifelines results in a

4
limited understanding of how these crucial systems are expected to perform in a major
earthquake [Barkley 2009].
Furthermore, because lifelines are distributed systems, they often cross multiple legal and
jurisdictional boundaries [ALA 2005], and can be controlled by either private or public entities.
As a result, individual communities may have little control over how lifelines within their
geographic boundaries are operated and maintained. In addition, lifelines are highly
interdependent, meaning that the serviceability of an individual lifeline following a major
earthquake depends not only its performance but also on the performance of other lifelines.
Therefore coordination among the different lifeline operators and regulators is required in the
development of system-level performance standards.
At the highest level, the current regulatory framework fails to establish an explicit set of
seismic performance objectives for the entire community. This not only makes it difficult for
engineers and planners to communicate the expected seismic performance of the community
with stakeholders and members of the general public, it also makes it impossible to determine
whether the performance levels specified for individual buildings and lifelines by building codes
and other performance standards are appropriate for the surrounding community. This report
addresses this important shortcoming by proposing an integrated, coordinated, and
comprehensive engineering framework—one that establishes broad performance goals for the
community before determining performance targets for individual components and subsystems
within the built environment.

1.1.2 Community Resilience

In the past two decades, the field of resilience has gained traction and received considerable
attention from both researchers and policymakers. Bruneau et al. [2003] define seismic resilience
as “the ability of social units (e.g., organizations, communities) to mitigate hazards, contain the
effects of disasters when they occur, and carry out recovery activities in ways that minimize
social disruption and mitigate the effects of future earthquakes.” To date, most research in this
rapidly evolving field has focused on defining resilience and establishing metrics to measure and
quantify it [Miles and Chang 2003, 2006, 2007; Cutter et al. 2010; Twigg 2009]. These efforts
have been instrumental in shifting the focus of designers and engineers (at least in part) from
how individual components respond to how the entire community performs; however, much of
this work has focused on the evaluation side of resilience. Less effort has gone into the design
side. For example, if communities want to enhance or improve their resilience to disasters,
exactly what level of performance is required from buildings and lifelines? Poland et al. [2009]
establishes a comprehensive set of performance objectives that, if achieved, will make the city of
San Francisco more resilient. Specifically, this set of objectives aims to have the city “back on its
feet” four months after a magnitude 7.2 earthquake on the Peninsula segment of the San Andreas
fault.
This report draws from and builds on these important studies, proposing a
comprehensive, coordinated, and integrated engineering framework that any community can use
to establish a consistent set of performance targets for individual subsystems and components

5
within the built environment in order to enhance overall resilience to natural disasters. In
particular, the proposed framework uses findings from the resilience literature to identify
essential community functions that need to be maintained following a disaster and, subsequently,
to identify components and subsystems within the built environment (including their
interdependencies) that support these essential functions.

6
Figure 1.1 Matrix of performance standards and guidelines for lifelines [ALA 2004].

7
1.2 PROPOSED ENGINEERING FRAMEWORK

The engineering framework proposed in this report is an adaptation of the framework used to
design and evaluate the safety of nuclear power plants in the United States. This report leverages
findings from both the community resilience and lifeline interdependencies fields to create a
transparent, performance-based, and risk-informed engineering framework that can be used to
establish a consistent set of performance targets for the built environment and its many
subsystems and components to enhance the overall resilience of the community. It addresses an
important gap that exists in how the current regulatory framework establishes performance
objectives for individual components within the built environment (e.g., buildings, bridges,
pipelines, electrical grids, etc.). Currently, performance objectives for individual components are
not tied to broader performance goals for the community, resulting in inconsistent and sometimes
inappropriate performance targets for individual components within the built environment.
The proposed engineering framework provides a quantitative methodology for explicitly
linking performance targets for individual components to broader goals for the entire
community. This linkage is especially important in the context of improving community
resilience for two reasons. First, a well-articulated set of performance goals provides
communities tangible targets to strive toward. Second, an explicit set of community performance
goals can serve as the basis for a more consistent set of performance objectives for individual
components within the built environment, thus ensuring that individual components perform in a
manner that is compatible with the best interests of the entire community.

1.2.1 Scope

A community is a dynamic and multi-faceted system of people, organizations, interactions, and

infrastructure. This report, however, focuses primarily on the built environment because of the
important role it plays in enabling a community to function successfully, providing the physical
foundations for much of the economic and social activities that characterize a modern society
[O’Rourke 2007]. At the same time, it is important to recognize that other aspects of a
community (e.g., people, organizations, and political, social, and economic environments) can
significantly influence how a community plans for, responds to, and recovers from a major
disaster. This report acknowledges these other aspects to the extent that it is appropriate.
Much like a community itself, the regulatory framework that dictates how a community
and its numerous components must operate is a complex system of regulations, enforcement
mechanisms, and guidance. Again, this report focuses on a small but important piece of this
regulatory framework: the building codes and other engineering standards that establish
performance expectations, either implicit or explicit, for a community’s built environment. Other
parts of a community’s regulatory framework, including planning, land use, and zoning
regulations, can impact how the built environment develops and performs; however, this report
does not explicitly address these items.
In addition, the proposed engineering framework seeks to improve community resilience
primarily through mitigation, in particular through changes to building codes and other

8
engineering standards that improve how the built environment performs in earthquakes and other
natural hazards. There are other actions that can enhance community resilience, including
development of comprehensive emergency response and recovery plans, but again, these are not
the focus of this report.
Lastly, while the intent of the proposed engineering framework is to remain broadly
applicable to all types of hazards, it is developed with earthquakes in mind. As such, it may
require modification if addressing other types of hazards.

1.2.2 Intellectual Contribution

To date, much of the research in the community resilience field has focused on defining and
measuring resilience from a social sciences’ perspective. This report brings a distinct engineering
perspective to the field of community resilience. Using the nuclear framework as a template, this
report demonstrates how these resilience measures, which are often qualitative in nature, can be
translated into quantitative engineering performance targets for components and subsystems
within the built environment.
The primary intellectual contributions of this report are as follows:
1. Identification and description of a comprehensive list of attributes of an
ideal regulatory framework
2. Adaptation of a methodology originally developed for design and
evaluation of nuclear power plants to be used in the design and evaluation
of the built environment of communities
3. Demonstration of a procedure that enables derivation of consistent
performance objectives for individual components from community-level
performance targets
The work presented in this report has the potential to improve the way engineers,
planners, and other stakeholders design and evaluate the built environment of a community. The
framework and methodology proposed herein provide a transparent, structured blueprint both for
planners and policymakers to set community-level performance targets and, subsequently, for
engineers to calibrate the designs of individual components to meet these community-level
performance targets. Together, the findings presented in this report establish the foundations for
a much-needed transformation from engineering individual components of the built environment
on a component-by-component basis to engineering community resilience using an integrated
and coordinated approach that begins at the community level.

1.3 ORGANIZATION OF THE REPORT

The following report comprises two main parts. The first part, Chapters 2, 3, and 4, provides
background information and demonstrates the need for the proposed engineering framework. The
second part, Chapters 5, 6, and 7, describes the proposed engineering framework and

9
demonstrates several potential applications. The following paragraphs describe each chapter in
more detail.
Chapter 2 reviews existing literature from the community resilience and lifeline
interdependencies fields in order to demonstrate the need for the engineering framework
described in the second part of this report. It also defines important terms and concepts used
throughout the rest of this report, many of which can have different meanings depending on the
context. Last, Chapter 2 introduces a list of attributes that characterize an ideal regulatory
framework. These attributes will be used in Chapter 3 to give structure to a critical analysis of
the current regulatory framework for the built environment.
Chapter 3 describes the current regulatory framework used in the United States to design
and analyze the built environment of a typical community. It examines both the structure of the
framework and the design philosophy it codifies, focusing in particular on the building codes and
other engineering documents that establish seismic performance expectations for the built
environment. Using the list of attributes of an ideal regulatory framework from the previous
chapter, it discusses the strengths and shortcomings of the current regulatory framework,
ultimately providing further justification for the engineering framework proposed in subsequent
chapters of this report
Chapter 4 examines the regulatory framework currently used in the United States to
design and analyze nuclear power plants. It identifies and defines important nuclear terminology
and concepts that are used throughout the rest of the report, including undesired outcomes, vital
functions, and frontline and support systems. The design philosophy codified in the current
regulatory framework is described next. Lastly, it discusses several performance evaluation tools
that nuclear engineers use to analyze the response of nuclear power plants, including
probabilistic risk assessments, dependency matrices, and event and fault trees. These tools are
then adapted for use in the engineering framework proposed in the second part of the report.
Chapter 5 begins the presentation of the proposed engineering framework for design and
evaluation of the built environment. It extends and applies the general nuclear design philosophy
described in Chapter 4 to communities. In particular, it discusses the range of potential undesired
outcomes that can affect a community and the vital community functions that prevent these
undesired outcomes from occurring. It also lists the frontline and support systems within the built
environment that enable the vital community functions. Lastly, it describes how the performance
evaluation tools presented in Chapter 4 are adapted for use in a community setting.
Chapter 6 presents and describes a set of community event trees that forms the backbone
of the proposed engineering framework. Event trees provide a structured methodology for
enumerating and, subsequently, evaluating the numerous combinations of events that can result
in undesired outcomes for a community. Chapter 6 begins by outlining the conditions under
which the event trees should be used before discussing their general structure and organization.
Chapter 6 presents a set of four event trees corresponding to the vital functions in a community
and details the rationale used to develop each one.

10
 Chapter 7 presents two conceptual examples that demonstrate potential applications of
the engineering framework described in Chapters 5 and 6. The first example demonstrates how
the event trees from the previous chapter can be used to establish consistent performance
objectives for individual components from a community-level performance target. More
specifically, the example shows how to develop seismic performance targets for a new
residential building from a community-level performance objective. Ultimately, this example
outlines a procedure that can be used both to modify the implicit performance objectives
contained in building codes and to lay the conceptual foundations of a “community performance
code,” a document that contains explicit performance targets for a community and the numerous
components and subsystems of its built environment. The second example outlines a
methodology that can be used to estimate the disruption to a community’s services caused by an
earthquake or other natural hazard.
Chapter 8 presents conclusions and discusses future work. Specifically, it outlines the
work that remains in refining and expanding the framework presented in previous chapters. In
addition, it discusses implications for the current regulatory framework, including the changes
required before the proposed framework can be implemented in practice.

11
12
2 Background and Definitions

This chapter has three primary objectives. First, it reviews existing literature from the
community resilience and lifeline interdependencies fields to demonstrate the need for the
engineering framework described in later chapters of this report. This literature review reveals
that, to date, most research has focused on defining and measuring community resilience and
lifeline interdependencies (i.e., the evaluation side of resilience and lifeline performance). Less
attention, however, has been given to the design side, including how to establish a consistent set
of specific performance goals for a community and its built environment that, if achieved, will
enhance overall community resilience. Second, to avoid any potential confusion, this chapter
defines important terms and concepts used throughout this report, many of which can have
different meanings depending on the context. Third, this chapter introduces a list of attributes
that characterize an ideal regulatory framework. These attributes will be used in Chapter 3 to
give structure to a critical analysis of the current regulatory framework for the built environment.

2.1 LITERATURE REVIEW

The following subsections discuss several important studies from the fields of community
resilience and lifeline interdependencies. In particular, Section 2.1.1 summarizes the findings of
Bruneau et al. [2003], Miles and Chang [2003, 2006, 2007], Poland et al. [2009], Twigg [2009],
SERRI and CARRI [2009], and Cutter et al. [2010]; Section 2.1.2 summarizes Rinaldi et al.
[2001], Barkley [2009], and others. Ultimately, these two subsections help demonstrate the need
for the engineering framework described in later chapters of this report.

2.1.1 Community Resilience

The field of community resilience developed in response to the observation that, while the
current regulatory framework for the built environment has been successful in reducing
casualties in recent disasters in the United States, it does little to mitigate the damage and
disruption caused by hazards. To date, most research in the field has focused on defining
resilience and establishing metrics to quantify it. The conceptual resilience framework proposed
by Bruneau et al. [2003] has been instrumental in both regards. The authors define resilience as
“the ability of social units (e.g., organizations, communities) to mitigate hazards, contain the
effects of disasters when they occur, and carry out recovery activities in ways that minimize

13
social disruption and mitigate the effects of future earthquakes.” They describe four properties of
resilience (robustness, redundancy, resourcefulness, and rapidity) and four dimensions of
resilience (technical, organizational, social, and economic). Table 2.1 and Table 2.2 explain,
respectively, the four properties and four dimensions in further detail.

Table 2.1 Four properties of resilience (adapted from Bruneau et al. [2003]).

Property Description

The ability of systems, components, and other units of analysis to

Robustness withstand a given level of stress or demand without suffering
degradation or loss of function
The extent to which systems, components, and other units of
Redundancy
analysis exist that are substitutable
The capacity to identify problems, establish priorities, and mobilize
Resourcefulness resources when faced with conditions that threaten to disrupt some
system, component, or other unit of analysis
The capacity to meet priorities and achieve goals in a timely
Rapidity
manner in order to contain losses and avoid future disruption

Table 2.2 Four dimensions of resilience (adapted from Bruneau et al. [2003]).

Dimension Description

The ability of physical systems (including components, systems,

Technical and their interactions) to perform to acceptable/desired levels when
subject to earthquake forces
The capacity of organizations that manage critical facilities and
have the responsibility for carrying out critical disaster-related
Organizational
functions to make decisions and take actions that contribute to
achieving the four properties of resilience
The capacity to lessen the extent to which earthquake-stricken
communities and governmental jurisdictions suffer negative
Social
consequences due to the loss of critical services as a result of
earthquakes
The capacity to reduce both direct and indirect economic losses
Economic
resulting from earthquakes

14
 Bruneau et al. [2003] proposed measuring resilience using three complementary metrics:
probability of failure, consequences of failure, and recovery time. Figure 2.1 illustrates these
metrics graphically, with the vertical axis measuring the consequences of failure and the
horizontal axis measuring recovery time. It portrays three cases (A, B, and C). Each case has the
same initial loss (i.e., consequences of failure) after the disaster; however, each has a different
time to recovery. For simplicity, a community is assumed to have “fully recovered” when it
restores 100% of its pre-disaster functionality. As shown in Figure 2.1 [Miles and Chang 2006],
the community depicted in Case C never fully recovers, whereas the community in Case B
recovers in the shortest amount of time and actually achieves a higher level of functionality than
existed before the disaster (i.e., Case B is more resilient). In general, resilient communities are
those that have reduced probabilities of failure, reduced consequences of failure, and reduced
time to recovery.
The work done by Bruneau et al. [2003], though conceptual in nature, was an important
step in attempting to quantify community resilience. Other important studies include Miles and
Chang [2003, 2006, 2007], Cutter et al. [2010], Twigg [2009], and SERRI and CARRI [2009].
The following paragraphs summarize the principal contributions of each study to the resilience
literature.
Building upon concepts introduced in Bruneau et al. [2003], Miles and Chang [2003,
2006, 2007] developed ResilUS, a computer program based on a comprehensive conceptual
model of community recovery. This conceptual model “enumerates important relationships
between a community’s households, businesses, lifelines, and neighborhoods” [Miles and Chang
2006]. Figure 2.2 shows schematically the relationships among these four groups. The ultimate
goal of the conceptual model is to “facilitate better understanding of the community recovery
process in hopes that decision makers and citizens can increase their community’s resilience
against disaster” [Miles and Chang 2006]. To this end, ResilUS can be used to track community
recovery at varying levels of detail, ranging from individual households and businesses to entire
lifeline networks and neighborhoods. However, Miles and Chang [2003, 2006, 2007] stop short
of recommending specific recovery targets that will enhance community resilience.
Cutter et al. [2010] established a set of baseline resilience indicators for communities.
These indicators can be used to measure both the resilience of a particular community and the
effectiveness of programs and policies that aim to improve disaster resilience. Cutter et al. [2010]
identify 36 indicators, which are grouped into five main categories: social resilience, economic
resilience, institutional resilience, infrastructural resilience, and community capital. Table 2.3
lists all 36 indicators. The resilience score for a community is an aggregation of each indicator,
and can range between zero and five, with zero being the least resilient and five being the most.
Using this information, specific programs and policies can be developed to target those resilience
indicators that are deficient. However, Cutter et al. [2010] stop short of recommending specific
resilience scores for which communities should aim.

15
 Figure 2.1 Schematic of disaster recovery [Miles and Chang 2006].

Figure 2.2 Overview of conceptual model of community recovery from earthquakes

[Miles and Chang 2006].

16
 Table 2.3 Baseline resilience indicators for U.S. communities (adapted from Cutter
et al. [2010]).

Category Indicator

Educational equity: ratio of percent population with college education to percent

population with no high school diploma
Age: percent non-elderly population
Transportation access: percent population with a vehicle
Social resilience Community capacity: percent population with a telephone
Language competency: percent population not speaking English as second
language
Special needs: percent population without a sensory, physical, or mental disability
Health coverage: percent population with health insurance coverage
Housing capital: percent homeownership
Employment: percent employed
Income and equality: GINI coefficient
Economic Single sector employment dependence: percent population not employed in
resilience farming, fishing, forestry, and extractive industries
Employment: percent female labor force participation
Business size: ratio of large to small business
Health access: number of physicians per 10,000 population
Mitigation: percent population covered by recent hazard mitigation plan
Flood coverage: percent housing units covered by NFIP policies
Municipal services: percent municipal expenditures for fire, police, EMS
Mitigation: percent population participating in Community Rating System for flood
Institutional
resilience Political fragmentation: number of governments and special districts
Previous disaster experience: number of paid disaster declarations
Mitigation and social connectivity: percent population covered by Citizen Corps
programs
Mitigation: percent population in Storm Ready communities
Housing type: percent housing units that are not mobile homes
Shelter capacity: percent vacant rental units
Medical capacity: number of hospital beds per 10,000 population
Infrastructural
Access/evacuation potential: principle arterial miles per square mile
resilience
Housing age: percent housing units not built before 1970 and after 1994
Sheltering needs: number of hotels/motels per square mile
Recovery: number of public schools per square mile
Place attachment: net international migration
Place attachment: percent population born in state that still resides in state
Political engagement: percent voter participation in 2004 election
Social capital (religion): number of religious adherents per 10,000 population
Community
capital Social capital (civic involvement): number of civic organizations per 10,000
population
Social capital (advocacy): number of social advocacy organizations per 10,000
population
Innovation: percent population employed in creative class occupations

17
 Twigg [2009] identified 28 components of resilience, which are organized into five
thematic areas: governance, risk assessment, knowledge and education, risk management and
vulnerability reduction, and preparedness and response. Table 2.4 lists each of these 28
components. For each component of resilience, Twigg [2009] enumerates a more specific and
detailed set of characteristics of disaster-resilient communities that “brings users closer to reality
on the ground.” Table 2.5 lists the characteristics of a disaster-resilient community corresponding
to the “Hazards/risk data and assessment” component of resilience within the “Risk assessment”
thematic area (see Table 2.4). In total, Twigg [2009] lists 167 characteristics of disaster-resilient
communities. The resilience of a particular characteristic or thematic area is evaluated on a scale
from one to five, with one being the least resilient and five being the most. However, Twigg
[2009] stops short of recommending specific targets for which communities to aim.

Table 2.4 Components of resilience [Twigg 2009].

Thematic area Components of resilience

 Policy, planning, priorities and political commitment

 Legal and regulatory systems
 Integration with development policies and planning
 Integration with emergency response and recovery
Governance
 Institutional mechanisms, capacities and structures; allocation of
responsibilities
 Partnerships
 Accountability and community participation
 Hazards/risk data and assessment
Risk assessment  Vulnerability/capacity and impact data and assessment
 Scientific and technical capacities and innovation
 Public awareness, knowledge and skills
 Information management and sharing
Knowledge and
 Education and training
education
 Cultures, attitudes, and motivation
 Learning and research
 Environmental and natural resource management
 Health and well being
Risk management  Sustainable livelihoods
and vulnerability  Social protection
reduction  Financial instruments
 Physical protection; structural and technical measures
 Planning regimes
 Organizational capacities and coordination
 Early warning systems
Disaster
 Preparedness and contingency planning
preparedness and
response  Emergency resources and infrastructure
 Emergency response and recovery
 Participation, voluntarism, accountability

18
 Table 2.5 Characteristics of a disaster-resilient community corresponding to a
specific component of resilience [Twigg 2009].

Thematic Area 2: Risk

Characteristics of a disaster-resilient community
Assessment

 Community hazard/risk assessments carried out which

provide comprehensive picture of all major hazards and risks
facing community (and potential risks)
 Hazard/risk assessment is participatory process including
representatives of all sections of community and sources of
expertise
 Assessment findings shared, discussed, understood and
Component of resilience 1: agreed among all stakeholders, and feed into community
Hazards/risk data and disaster planning
assessment
 Findings made available to all interested parties (within and
outside community, locally and at higher levels) and feed into
their disaster planning
 Ongoing monitoring of hazards and risks and updating of
assessments
 Skills and capacity to carry out community hazard and risk
assessments maintained through support and training

The Southeast Region Research Initiative (SERRI) and Community and Regional Risk
Institute (CARRI) defined three broad groups of community functions that healthy and vibrant
communities provide to their residents [SERRI and CARRI 2009]. The first group includes
infrastructure-based functions like energy, water, and transportation. The second group involves
economic functions like employment opportunities, adequate wages, and affordable housing
options. And the third group includes social functions like community ownership and
participation, education and training opportunities, and a sense of community and place. Figure
2.3 shows these three groups of functions and their interactions. The innermost ring represents
the infrastructure-based functions, which must be restored first following a disaster. The middle
ring represents the economic functions of a community, which cannot be restored until
infrastructure-based functions are recovered. The outermost ring represents the social functions
of a community, which cannot be restored until both infrastructure and economic functions are
recovered.
As mentioned previously, much of the research in the community resilience field has,
thus far, focused on defining and measuring community resilience (i.e., the evaluation side of
resilience). Less effort has gone into the design side: in particular, if communities want to
enhance or improve their resilience to disasters, exactly what level of performance is required
from buildings and lifelines? The San Francisco Planning and Urban Research Association
(SPUR) established a comprehensive set of performance objectives that, if achieved, will
enhance the City of San Francisco’s resilience to earthquakes. Specifically, this set of objectives
aims to have the city “back on its feet” four months after a magnitude 7.2 earthquake on the
Peninsula segment of the San Andreas fault [Poland et al. 2009]. Figure 2.4 displays the set of

19
performance objectives for the entire city as a function of time, while Figure 2.5 displays more
specific performance objectives for important classes of buildings and lifelines. Figure 2.5 also
shows the current level of performance expected from each piece of infrastructure (the “X”
mark) relative to its specified target (the shaded box).

Figure 2.3 Important community functions [SERRI and CARRI 2009].

20
Figure 2.4 General performance objectives for San Francisco as a function of time
[Poland et al. 2009].

21
Figure 2.5 Specific performance objectives for San Francisco's buildings and
infrastructure as a function of time [Poland et al. 2009].

22
 The work of SPUR and Poland et al. [2009] is unique because it establishes explicit
performance objectives for the city to improve its disaster resilience. To date, most resilience
studies, including Miles and Chang [2003, 2006, 2007], Cutter et al. [2010], Twigg [2009], and
SERRI and CARRI [2009], focus on defining and measuring resilience, but stop short of
establishing concrete targets to aim at, thereby leaving the following question unanswered: when
is a community resilient enough? Poland et al. [2009] provides a clear answer for the city of San
Francisco.
This report builds on these efforts, proposing an engineering framework that can be used
by communities to establish a consistent set of performance targets for individual subsystems
and components within the built environment (e.g., buildings and lifelines) in order to enhance
overall community resilience. In the end, this set of performance targets may resemble those
described by Poland et al. [2009] (see Figure 2.4 and Figure 2.5); however, they will be
developed using a more robust, transparent, and technically grounded engineering framework.
As such, the proposed framework provides the technical justification for the performance
objectives described by Poland et al. [2009].

2.1.2 Lifeline Interdependencies

Lifelines are a critical piece of a community’s infrastructure, providing sustenance to both

residents (water networks deliver drinking water to homes; transportation networks deliver food
to grocery stores; energy lifelines deliver the fuel needed to heat residences) and businesses
(roads and highways enable the flow of goods and services; energy networks deliver power to
factories and office buildings). In spite of their importance, both in day-to-day operations and in
recovering after a major disaster, lifelines have received considerably less attention than
buildings. Barkley [2009] described some of the unique challenges associated with lifelines:

In general, a lifeline system incorporates a wide range of elements necessary for system
operation, including linear components; mechanical, electrical, and electronic
equipment; buildings containing system components; operating centers; and other
supporting elements. The circumstances under which individual elements may fail vary
widely, as do applicable design guidelines and standards. The performance of the entire
system is as critical as the performance of individual elements; however, damage to
individual elements may be sufficient to shut down part or all of the system.

Lifeline systems are also distinguished by their interdependency. The continued operation
of a lifeline system, such as the communications network, may be dependent on the
operation of another system, such as the power system. Similarly, the ability for system
owners to restore their respective systems following an earthquake may be dependent on
the condition of highways and other transportation elements.

To date, most research in this field has focused on defining, identifying, monitoring, and
measuring lifelines and their interdependencies. Many studies, including PCCIP [1997], Rinaldi

23
et al. [2001], Barkley [2009], and ALA [2004], identify and enumerate lists of critical lifelines
within a community. This report adapts and combines these lists into the following set of
lifelines: communication, energy (electric power, natural gas, oil, and solid fuels), transportation
(roads and highways, mass transit, ports and waterways, railways, and airports), water, and waste
disposal (waste water and solid waste).
Rinaldi et al. [2001] serves as an excellent primer on lifeline interdependencies. The
authors define four principal classes of interdependencies (physical, cyber, geographic, and
logical) and three types of failures (cascading, escalating, and common cause). Table 2.6 and
Table 2.7 describe, respectively, the four classes of interdependencies and three types of failures
in more detail and provide simple examples to help illustrate the concepts. The authors also
identify the challenges associated with modeling lifelines and their interdependencies; however,
they stop short of developing such a model.
Subsequent studies, including Haimes and Jiang [2001], Zhang et al. [2005], Dueñas-
Osorio et al. [2007], Lee et al. [2007], Svendsen and Wolthusen [2007], Rosato et al. [2008],
Ouyang et al. [2009], and Hernandez-Fajardo and Dueñas-Osorio [2011], have developed and
implemented comprehensive models of lifeline networks and their interdependencies to measure
and monitor their response to earthquakes and other hazards. While this work has been
instrumental in improving the understanding of how these complex systems respond and interact
with each other in the face of natural hazards, much of it has, thus far, focused on the evaluation
side of the equation (i.e., real-time monitoring of network performance). Similar to the field of
community resilience, less attention has been paid to the design side, including establishing
appropriate performance targets for lifelines and their numerous components that support
broader community resilience goals. One notable exception is Poland et al. [2009], which was
described in the previous subsection (see Figure 2.4 and Figure 2.5).
This report leverages the knowledge of lifelines and their interdependencies gained from
recent studies in order to inform the development of an engineering framework that can be used
to establish appropriate performance objectives for lifelines.

24
Table 2.6 Four principal classes of lifeline interdependencies (adapted from Rinaldi
et al. [2001]).

Class Description Example

Arises from a physical linkage

between the inputs and outputs A railway delivers coal to a
of two lifelines: a commodity power generating station; the
Physical produced or modified by one station supplies electricity to the
lifeline (an output) is required by railway’s signals, switches, and
another lifeline for it to operate control centers
(an input)
Arises when the functionality of The electric power lifeline relies
a lifeline depends on information on supervisory control and data
Cyber
transmitted through the acquisition (SCADA) systems to
communications lifeline control the grid
Occurs when elements of An electrical line and a fiber-
Geographic multiple lifelines are in close optic communications cable
spatial proximity slung under a bridge
Arises when the functionality of
each lifeline depends on the Low gas prices motivate more
state of the other via a people to drive, resulting in
Logical
mechanism that is not a increased congestion on roads
physical, cyber, or geographic and highways
connection

Table 2.7 Three types of lifeline failures (adapted from Rinaldi et al. [2001]).

Type Description Example

Disruption of a distribution
Occurs when a disruption in one network within the natural gas
lifeline causes the failure of a lifeline can result in a failure of an
Cascading component in a second lifeline, electricity generating unit located
which subsequently causes a in the service territory of the gas
disruption in the second lifeline system, which can cause power
disruptions
Occurs when an existing A disruption to the
disruption in one lifeline communications network may
exacerbates an independent escalate because of a
disruption of a second lifeline, simultaneous or subsequent
Escalating
generally in the form of increasing disruption in a transportation
the severity or the time for network, which in turn could delay
recovery or restoration of the the arrival of repair crews and/or
second failure replacement equipment
Occurs when two or more lifelines A train derailment that damages
are disrupted at the same time: railroad tracks could also disrupt
Common cause components within each network communications cables and
fail because of some common power lines that are located
cause within the same corridor

25
2.2 DEFINITIONS

The following subsections provide concrete definitions and other background information for
several important terms used extensively in this report, including system, hazard, performance,
and regulatory framework. Each of these terms can have somewhat ambiguous meaning
depending on the context; therefore, the objective of each of the following subsections is to
clarify what these terms mean when used herein.

2.2.1 System

In general, a system is a dynamic entity comprising a collection of interacting, potentially

correlated components assembled to perform an intended function or functions (adapted from
Vesely et al. [1981], Buede [2000], ISO and IEC [2008], and Kossiakoff et al. [2011]). This
report focuses on a particular subset of systems: those that are engineered for a specific purpose
(e.g., buildings and lifelines). Throughout this report, these engineered systems will be referred
to simply as systems. Note that components within a system can themselves be systems [Buede
2000]. For example, a building system comprises, among other things, a structural system,
heating, ventilation, and air conditioning systems, and mechanical, electrical, and plumbing
systems. Each of these systems, in turn, comprises various subsystems and components. For
example, a structural system comprises beams, columns, braces, walls, and floors, to name only a
few.
The following subsections discuss and expand upon the definition given above. This
discussion, which is purposely generic, provides the foundation for a more detailed examination
of communities, which themselves can be considered complex, dynamic, and adaptive systems.

2.2.1.1 Types of components

As defined above, a system is a collection of components (note that the interactions among these
components are discussed in the next subsection). These components can be classified into seven
general types: structures, hardware, people, organizations, procedures, environments, and
interfaces (adapted from Buede [2000], Bea [2007, 2008], and NASA [2007]). Table 2.8
provides brief descriptions of each, listing specific examples from the built environment to
further illustrate each type of component with the caveat that it might not always be possible to
draw clear, unambiguous boundaries for each component. Figure 2.6 graphically portrays the
relationships among the seven types of components. While engineers tend to focus on the
physical elements in a system (i.e., structures and hardware), the other types of components can
have strong influence over how the system behaves and responds.

26
Table 2.8 Description of the seven types of components in a system (adapted from Bea
[2007, 2008]).

Type Description Examples

The physical elements that support or protect Buildings, bridges,

Structures
the system and its functions dams
The physical equipment that enables or Electrical transformers,
Hardware
facilitates the system and its functions pumps, computers
Those who design, analyze, construct,
operate, use, maintain, rehabilitate, and Engineers, architects,
People decommission the system (i.e., anyone who electricians, bus drivers,
is involved in or impacted by the system passengers
during its lifecycle)
Design firms,
The companies, institutions, or agencies
Organizations government agencies,
involved with the system during its lifecycle
professional societies
The rules and guidelines (formal and
Laws, regulations,
Procedures informal) that operators and organizations
codes, ordinances
use to perform their activities
The conditions (external, internal, social,
Weather conditions,
Environments political, economic) in which people and
company culture
organizations perform their activities
The links that connect different components Supply chains, Internet,
Interfaces
in the system together physical proximity

Figure 2.6 Relationships among the seven types of components in a system [Bea
2007 and 2008].

27
2.2.1.2 Interactions
A system is much more than a collection of components or simple sum of parts. It is the
interactions among its components that enable a system to successfully perform its intended
function or functions. These interactions can be very complex, especially if the system comprises
many components, or very simple, as is the case for systems with configurations that are
exclusively parallel or series [Billinton and Allan 1983]. In this report, these interactions will be
referred to as dependencies and interdependencies. Dependency describes a unidirectional
relationship between two components, meaning that one component depends on the other, but
not vice versa [Rinaldi et al. 2001]. For example, a water pump depends on electricity delivered
by the power grid; however, if the pump were to stop functioning it would not affect the
functionality of the power grid. Interdependency, on the other hand, indicates a bidirectional
relationship between two components, meaning both components depend on each other to
function successfully [Rinaldi et al. 2001]. For example, an electrical generating station relies on
natural gas to run its generators and produce electricity; conversely, electricity is required in the
delivery of natural gas to the generating station.
Interactions are dynamic in nature. If the configuration of the system changes, the
interactions among its components can also change. For example, as more and more businesses
migrate from local to web-based computing, the dependence on Internet service providers also
grows. As a result, a service disruption (caused by an earthquake, for example) will have greater
impact on the local economy because of this increased dependency on the Internet.

2.2.1.3 Correlation
In addition to interacting with each other, components in a system may also be correlated.
Correlation measures the relationship between the responses of two distinct objects. It typically
arises from similarities in the design and/or location of the components [Dezfuli and Modarres
1985]. For example, the responses during an earthquake of two identical, adjacent houses will be
highly correlated: if one fails during the earthquake, the other will also likely fail. Correlation
differs from interaction in that there is no functional dependency or interdependency between the
two components. For example, the functionality of one house does not, in general, directly affect
the functionality of another, even if it is identical and located in close proximity. Therefore, the
response of two identical, adjacent houses would be considered highly correlated but not
functionally dependent or interdependent.
Systems with a large number of identical, co-located components are vulnerable to
correlated failures, an event in which a large number of components fail simultaneously during
an earthquake or other hazard, resulting in a potentially significant degradation in performance
[Lin et al. 2012]. One way to address these correlated failures is to build sufficient diversity into
the components of a system. For example, instead of installing two identical diesel generators,
one can procure each from a different vendor to ensure they have varying designs. Alternatively,
instead of placing the generators next to each other, one can locate them far enough apart that
they do not experience similar effects from nearby hazards. Diversity and correlation are
inversely related: as the diversity of components increases, correlation decreases.

28
2.2.1.4 Phases in the lifecycle of a system
A system is a dynamic entity: if the nature of any part of the system changes, the system itself
changes [Vesely et al. 1981]. Its lifecycle comprises many phases, including conception, design,
analysis, construction, operation, maintenance, rehabilitation, and decommission (adapted from
Buede [2000], Wasson [2006], Blanchard and Fabrycky [2006], and Bea [2007, 2008]). This
report focuses primarily on the design phase, examining the design frameworks that establish
performance targets for complex systems, including nuclear power plants and communities. It is
important to note, however, that subsequent phases, especially analysis and operation, need to be
considered when establishing design targets for these systems.

2.2.1.5 System definition and analysis

Perhaps the most crucial characteristic of a system is that it is determinable. Determinable means
that the system is identifiable and, more importantly, can be defined and subsequently analyzed.
System definition begins by establishing appropriate external boundaries for the system
[Vesely et al. 1981]. These boundaries, which need not be purely geographic in nature, depend in
part on the aspects of performance that are of interest. For example, if we want to compute the
probability of a building collapsing in a particular earthquake scenario, the external boundary for
the system will likely coincide with the geographical boundary of the building. If, on the other
hand, we were interested in the probability that a building loses power following a specific
earthquake scenario, the external boundary would need to be expanded to include the electrical
grid that services the building. Consequently, external boundaries determine the
comprehensiveness of the analysis.
It is important to note, however, that a system can be impacted by events and systems
beyond its external boundaries [Buede 2000]. For example, businesses in a community can be
impacted by an earthquake that strikes a distant city or region, especially if it disrupts the
production of goods and services that local business rely on, as happened to carmakers in the
United States following the 2011 Tohoku earthquake and tsunami that struck northern Japan. To
the extent possible, these external events and system should be included in the system analysis.
Another important aspect of system definition involves establishing a limit of resolution
for the system. Limits of resolution serve to define the discrete elements of the system and to
establish the basic interactions within the system [Vesely et al. 1981]. They also limit the detail
of the analysis. For example, if we want to compute the probability that a building collapses in
an earthquake, the system would likely need to include key structural subsystems (e.g., gravity
and lateral force-resisting subsystems) and components (e.g., beams, columns, braces, walls,
floors, etc.), with the resulting fragility curve being quite specific. However, if we want to
evaluate the vulnerability of a large group of buildings (for example, the housing stock of a
community), it will not be practical to resolve the system down to the level of individual
structural members for each building. Instead, it might be satisfactory to specify basic
information about the gravity and lateral force-resisting systems for each building in the system,
or even group the buildings into several broad categories based on key structural properties, and
use generic fragility data.

29
 Note that the external boundaries and/or limits of resolution may need to be updated as a
better understanding of the system or issue under consideration emerges. For example, a
previously unknown interaction with an external system will require the boundaries of the
analysis to expand to include this system. Similarly, the limits of resolution for a system may
need to be refined as its design moves beyond the conceptual stage.
Only after a system has been properly defined can it be analyzed. The type of analysis
performed depends on the aspects of performance that are of interest. In general, two analytical
approaches exist: reductive and expansive [Bea 2007, 2008]. Reductionism involves the
following steps:
 Identify the components in a system
 Study the behavior of individual components
 Derive an understanding of the behavior of the system from the behavior
of individual components
Expansionism is the complement of the reductive process and involves the following
steps [Bea 2007, 2008]:
 Identify the system
 Study the behavior of the entire system
 Derive an understanding of the behavior of individual components from
the behavior of the system
This report proposes and develops an engineering framework that employs an
expansionist (i.e., comprehensive, integrated, and coordinated) approach in order to enhance
community resilience. Specifically, it begins at the system level by establishing explicit
performance goals for the entire community. Then, using these community-level targets, it
studies the system to identify key community functions that need to be available in order for a
community to satisfy its specified targets. Lastly, after identifying these key functions, the
framework identifies components and subsystems within the built environment that support these
functions, and establishes appropriate performance objectives for individual components that are
consistent with previously established community-level goals.

2.2.1.6 Attributes of an ideal system

In the context of this report, the word “ideal” refers to an abstract or hypothetical optimum;
therefore, an ideal system may not be realistic. It does, however, represent a desirable end point
that, to the extent practical, should be aspired to. In general, ideal systems have the following
attributes: they are safe, serviceable, compatible, durable, and analyzable (adapted from Minai et
al. [2006] and Bea [2007, 2008]). The following paragraphs discuss each in further detail.
First, an ideal system is safe, meaning it does not pose undue threat to the health and
safety of the operators, general public, and surrounding environment. This attribute is of
fundamental importance and is often the focus of most regulations for a system. Another aspect

30
of safety is security. Security involves an absence of vulnerability to malevolent events (e.g.,
terrorism, sabotage, etc.).
Second, an ideal system is serviceable, meaning it is highly suited for its intended
purpose. It should not be used for purposes other than those for which it was originally intended.
For example, a building originally designed for lightweight office space should not be used for
heavy manufacturing, at least not without substantial retrofit. In addition, the system should not
operate under conditions that exceed those for which it was initially designed. For example, a
crane with 20-ton capacity should not be used to lift a 25-ton section of a bridge.
Third, an ideal system is compatible, meaning it does not have excessive negative impact
on society and the surrounding environment. Its performance aligns with societal expectations. It
uses resources in an efficient and sustainable manner—one that minimizes operating costs and
consequences while protecting the ability of future generations to operate similar systems
[WCED 1987].
Fourth, an ideal system is durable, meaning it maintains its safety, serviceability, and
compatibility throughout its lifecycle. Consequently, it is reliable, robust, resilient, and
redundant. A reliable system has high likelihood of remaining functional over time. A robust
system can tolerate significant amounts of damage or a large number of defects and errors
without losing functionality and, therefore, is insensitive to small perturbations. A resilient
system can recover functionality quickly following a disruption or disturbance. A redundant
system comprises several independent, diverse paths for ensuring functionality.
Finally, an ideal system is analyzable, meaning it features a design or configuration that
enables reliable and accurate analysis of its response to a wide range of hazard scenarios.
Similarly, the design of a system should also be sufficiently constructible, operable,
maintainable, and repairable.

2.2.2 Hazard

In the most general sense, a hazard is a potential source of danger. Typically, it refers to a threat
that is unrealized but has potential to occur in the future. More specifically, the Federal
Emergency Management Agency (FEMA) defines a hazard as “any event or condition with the
potential to cause fatalities, injuries, property damage, infrastructure damage, agricultural loss,
environmental damage, business interruption, or other loss” [FEMA 1997]. Hazards can be either
natural or human-made. Regarding this distinction, the Organization of American States writes
[OAS 1991]:
A widely accepted definition characterizes natural hazards as "those elements of the
physical environment, harmful to man and caused by forces extraneous to him" [Burton
et al. 1978]. More specifically… the term "natural hazard" refers to all atmospheric,
hydrologic, geologic (especially seismic and volcanic), and wildfire phenomena that,
because of their location, severity, and frequency, have the potential to affect humans,
their structures, or their activities adversely. The qualifier "natural" eliminates such
exclusively manmade phenomena as war, pollution, and chemical contamination.

31
 Hazards to human beings not necessarily related to the physical environment, such as
infectious disease, are also excluded from consideration here.

Notwithstanding the term "natural," a natural hazard has an element of human

involvement. A physical event, such as a volcanic eruption, that does not affect human
beings is a natural phenomenon but not a natural hazard. A natural phenomenon that
occurs in a populated area is a hazardous event… In areas where there are no human
interests, natural phenomena do not constitute hazards… This definition is thus at odds
with the perception of natural hazards as unavoidable havoc wreaked by the unrestrained
forces of nature. It shifts the burden of cause from purely natural processes to the
concurrent presence of human activities and natural events.
The terminology contained in the above excerpt is adopted in this report. Furthermore, the scope
of this report is similar to that outlined above. However, even if a hazardous event does not
immediately or directly affect any human interests, it can still have profound impact. For
example, a volcanic eruption on a remote island may not directly harm any human activities, but
it could destroy important natural habitats and alter global weather patterns, producing crop
failures and food shortages. These effects are important but difficult to address or plan for.
Therefore, this report focuses mainly on the more direct, immediate effects of hazards on a
system. Specifically, this report focuses on the effects arising from natural hazards like
earthquakes.

2.2.2.1 Multiple effects of hazards

In some instances, a natural hazard can produce multiple effects. For example, a hurricane can
produce a combination of violent wind (including tornados), torrential rain, and damaging storm
surge. Similarly, an earthquake can produce ground shaking, surface rupture, lateral spreading,
liquefaction, tsunamis, and landslides. A natural hazard can also induce human-made hazards.
For example, an earthquake can trigger large fires if gas lines rupture throughout a community
[Scawthorn 2003b]. It can also produce extensive flooding if nearby dams or levees fail as a
result of an earthquake. These induced hazards can have as much impact as the primary hazard
and, therefore, should be accounted for when performing a hazard analysis for a system.
Furthermore, the effects of a hazard can vary from location to location. For a spatially distributed
system subject to earthquakes (e.g., the electrical grid in Los Angeles), areas closest to nearby
faults will likely experience stronger shaking than those farther away. Also, portions of the
system founded on soft soil may experience amplified shaking relative to locations founded on
rock.

2.2.2.2 Hazard analysis

Each system faces a unique set of hazards that depends on the surrounding natural and human-
made environment. System designers and operators must carefully analyze the system’s
surroundings in order to properly identify and characterize potential hazards. Only after
thoroughly evaluating these hazards can operators plan and prepare accordingly. A hazard
analysis for a system identifies potential sources of hazard, as well as the range and frequency of
hazard scenarios that each source may produce.

32
 A hazard analysis can be either deterministic or probabilistic. In a deterministic hazard
analysis, one particular hazard scenario is evaluated. This scenario might, for example, postulate
the occurrence of a hazard with a specific size and location (e.g., a magnitude 7.6 earthquake on
a particular fault segment). Such an analysis would be appropriate if attempting to establish a
worst-case scenario for a particular hazard source. A deterministic hazard analysis, however,
neglects to include uncertainties in the hazard such as its size, location, and frequency of
occurrence. A probabilistic hazard analysis, on the other hand, provides a framework that
identifies, quantifies, and combines these uncertainties to obtain a more complete picture of the
hazard [Kramer 1996]. A probabilistic hazard analysis includes all possible hazard scenarios and
combines them using the frequency of occurrence of each scenario. For this reason, a
deterministic hazard analysis corresponds to a particular scenario in a probabilistic hazard
analysis [Thenhaus and Campbell 2003].

2.2.3 Performance Measures

In the context of this report, performance refers to the ability of a system or component to
achieve objectives and targets pertaining to its functionality, safety, or costs. Typical
performance measures for buildings include casualties, lifecycle costs, and time to restore
functionality (i.e., downtime). In contrast, response refers to the physical behavior of a system
when subjected to a stress or stimulus (e.g., earthquake ground shaking or liquefaction).
Traditional response measures for buildings include forces, accelerations, displacements, and
drifts.
As the definition above indicates, system performance is typically evaluated relative to
specified targets or objectives. These performance targets or objectives can take many different
forms, depending on the system or component being considered and the desired outcomes. For
example, if a building owner is only concerned with protecting the safety of occupants during an
earthquake, performance objectives for the building will seek to minimize casualties. These
performance objectives can be achieved, for example, by assuring that the response of the
building during an earthquake remains within certain thresholds (e.g., peak interstory drift ratios
less than 2%). On the other hand, if a building owner is also concerned with maintaining
functionality after an earthquake, performance objectives for the building will seek to minimize
downtime in addition to casualties.
While this report focuses primarily on the process of establishing performance objectives
for a system and its components, an essential corollary to establishing performance objectives is
evaluating whether or not they have been satisfied. Many techniques and methodologies exist for
evaluating system performance; however, this report does not explicitly address them. That said,
it is important to note that performance objectives influence the scope of the evaluation required
for a particular system. For example, if the performance objectives for a building specify that it
minimize casualties during an earthquake, then only its structural system needs to be analyzed, as
most earthquake-related casualties are caused by structural collapse [BSSC 2009]. On the other
hand, if performance objectives for the building specify that it remain functional following an

33
earthquake, then both the building (including its structural and nonstructural systems) and any
supporting lifelines need to be evaluated.

2.2.4 Regulatory Framework

A regulatory framework provides the legal and technical basis for allowing a system to operate
through all phases of its lifecycle. It comprises three basic elements: regulations, mechanisms for
enforcing the regulations, and guidance for satisfying the regulations. The following paragraphs
discuss each element in further detail.
Regulations include codes, standards, and other documents that specify the rules,
requirements, and provisions for a system. Regulations typically exist in the public domain and
carry the weight of law. If a system does not comply with regulations, it can potentially face a
variety of penalties, ranging from fines and lawsuits to temporary or permanent shutdown of the
system. Regulations typically arise in response to societal problems. For example, in the United
States, building codes were developed to protect the public from unsafe living and working
conditions brought about by poorly designed, constructed, and maintained buildings.
Enforcement mechanisms include the agencies and organizations charged with
interpreting and enforcing the regulations. These agencies and organizations are commonly
referred to as regulators. Enforcement is a crucial component in any framework. Without it,
system operators and designers might ignore certain regulations if they impose significant cost or
burden.
Guidance includes anything that aids in satisfying the regulations. It can range from
written documents developed by technical societies (and later adopted by regulators) to
electronic communications with regulators. Guidance is typically optional, providing one of
many possible ways to satisfy the regulations. Often, however, guidance becomes the de facto
means to satisfying the regulations and thus plays a crucial role in a regulatory framework.

2.3 ATTRIBUTES OF AN “IDEAL” REGULATORY FRAMEWORK

A regulatory framework is ideal when it produces systems that are also ideal (see Section 2.2.1
for a discussion of attributes of an ideal system). Again, the word “ideal” refers to an abstract or
hypothetical optimum. Therefore, although an ideal regulatory may not be realistic, it represents
a desirable end point that, to the extent practical, should be aspired to. In general, an ideal
regulatory framework is expansionist, risk-informed, comprehensive, performance-based,
probabilistic, technology-neutral, transparent, acceptable, feasible, consistent, and enforceable
(adapted from the USNRC [1998] and ONRR [2007]). The following subsections describe a
subset of these attributes, focusing on those that are most pertinent in the context of this report.
This subset includes five attributes: expansionist, risk-informed, comprehensive, performance-
based, and acceptable.

34
2.3.1 Expansionist Framework

An ideal regulatory framework employs an expansionist or top-down approach. In other words, it

begins by establishing basic requirements for the entire system first [ONRR 2007]. This stands in
contrast to a reductionist, bottom-up, or component-by-component approach in which
requirements are first established for individual components without consideration of the
performance of the system as a whole. A top-down approach is important because it sets the
stage for the entire regulatory framework. It ensures that provisions for individual components
are consistent with system-level requirements. Furthermore, it facilitates understanding of the
intended performance of the system, unlike a bottom-up approach in which it may be difficult to
determine the performance of the system, especially if the system comprises a large number of
interactive and correlated components.

2.3.2 Risk-Informed Framework

An ideal regulatory framework contains requirements that are risk-informed, where risk is
defined as the product of two quantities: (1) the likelihood or probability that an undesired event
occurs; and (2) the consequences once it occurs. A risk-informed approach considers risk
insights, together with other factors, to establish provisions and requirements for a component
that are commensurate with its importance in protecting the health and safety of the public and
environment [USNRC 1998]. In other words, the regulations for a particular component are
proportional to its overall risk to safety or functionality. For example, if the failure of a particular
component has limited impact on the safety or functionality of the system, a risk-informed
regulatory framework would specify only limited provisions for the component. If, on the other
hand, failure of a particular component has significant impact, a risk-informed regulatory
framework would establish stringent performance requirements for the component. In contrast, a
framework that is not risk-informed would specify the same performance requirements for both
components regardless of their impact on system safety or functionality. In summary, a risk-
informed regulatory framework focuses attention on those components within the system that are
most important to overall safety and functionality.

2.3.3 Comprehensive Framework

An ideal regulatory framework is comprehensive in three respects. First, it establishes provisions

that require consideration of all hazards, both natural and human-made, that can potentially
affect the system. For practical purposes, however, many of these hazards can be “screened out”
because either they do not significantly impact the performance of the system (e.g., earthquakes
of small magnitude) or they have an extremely small chance of occurring (e.g., earthquakes of
extremely large magnitude). Justification for screening out particular hazard scenarios should be
provided. Those that cannot be screened out can be organized into groups with similar attributes
(e.g., earthquakes within a certain magnitude range). An enveloping or worst-case scenario can
then be selected from each group, resulting in a thorough but manageable set of scenarios that
forms the basis for design and evaluation of the system.

35
 Second, a comprehensive regulatory framework establishes provisions and requirements
for all subsystems and components important to the system in the selected scenarios. As
discussed in Section 2.3.1, provisions should begin at the level of the system and eventually
work down to the level of individual components. Provisions for a particular subsystem or
component should be based on its risk contribution to the system (see Section 2.3.2). In addition
to including important components and subsystems, provisions should also identify and address
potential interactions between subsystems and components.
Third, a comprehensive regulatory framework accounts for uncertainty, both in
identifying and analyzing potential hazards and in characterizing the system and its numerous
components. Uncertainty stems from two primary sources. The first source, commonly referred
to as aleatory uncertainty, arises from inherent randomness in behavior of the entity under
consideration, while the second, referred to as epistemic uncertainty, arises from limitations in
knowledge [Parry and Winter 1981; Helton 1994; Parry 1996; Ang and Tang 2007]. Over time,
epistemic uncertainty can be reduced as knowledge improves; aleatory uncertainty, however,
always remains [Der Kiureghian and Ditlevsen 2009]. Provisions and requirements for a system
and its components should address both types of uncertainty.

2.3.4 Performance-Based Framework

An ideal regulatory framework is performance-based in character. In general, a performance-

based regulatory approach “specifies the outcome required but leaves the concrete measures to
achieve that outcome up to the discretion of the regulated entity” [Coglianese et al. 2002]. In
contrast, a prescriptive regulatory approach specifies exactly how to achieve compliance. In
other words, performance-based regulations are defined “with respect to desired outcomes rather
than prescribed means or technologies” [May and Koski 2004]. A hallmark of performance-
based regulation is “the explicit statement of goals and objectives that reflect societal
expectations and desires, along with functional statements, operative requirements, and in some
cases performance criteria, which are to be used for demonstrating that goals and objectives have
been met” [Meacham et al. 2005].
May and Koski [2004] summarize the potential advantages and disadvantages of
performance-based regulations. In general, performance-based regulations can increase the
incentive for innovation, increase the flexibility in how regulations can be satisfied, and decrease
the costs of compliance for regulated entities (e.g., designers, architects, utility providers, etc.).
At the same time, however, performance-based regulations can also increase the costs to
government regulators and reduce the predictability in regulatory expectations.
A performance-based framework establishes explicit performance objectives or outcomes
for a system and all necessary subsystems and components. In turn, these objectives can form the
basis for a more detailed set of provisions and requirements. In a truly performance-based
framework, however, only performance objectives would be specified; the designer would be
given complete flexibility in deciding how to satisfy the specified objective. Often, though, a
typical designer will require further, more detailed guidance. In addition, the agency charged

36
with enforcing truly performance-based regulations might struggle to verify compliance without
a more detailed set of performance criteria and requirements.
For these reasons, a truly performance-based regulatory framework is usually not
practical. However, the framework should still use explicit performance objectives as the
foundation for all resulting regulation (i.e., the framework should be performance-informed). For
example, consider the following performance objective: a hospital must remain operational after
an earthquake. This performance objective could be supported by a more detailed set of
performance criteria, including specific limits on peak interstory drift ratios, floor accelerations,
and residual displacements, and provisions requiring onsite backups of all critical utilities.
However, this more detailed set of performance criteria would stop short of prescribing specific
means for satisfying the requirements (e.g., requiring use of a particular structural system or
specifying a particular approach for providing backup electricity).

2.3.5 Acceptable Provisions and Requirements

An ideal regulatory framework contains provisions and requirements that are acceptable. In
particular, regulations represent a level of risk that is consistent with what society expects from
the system and, perhaps more importantly, what costs society is willing to incur. For example,
while it would be ideal for a system to be able to withstand the effects of an extremely large, rare
earthquake without suffering any damage, the costs of doing so might exceed what the public is
willing to pay. It is crucial that these costs be weighed carefully against benefits associated with
preventing the potentially adverse consequences if something goes wrong. For example, in the
case of nuclear power plants, not only could an accident at one facility affect the communities
and environment surrounding the plant, but it could also force other facilities to close or cause a
national or global shift away from nuclear power altogether.
Only after careful consideration of all potential consequences can risk targets for a
system be established. However, no matter what the final target or targets are, residual risk will
always remain. For example, an earthquake exceeding the design basis could occur, even if the
design basis earthquake is extremely rare. An ideal regulatory framework takes necessary steps
to ensure that this risk is acceptable to society. It contains mechanisms and processes to inform
the public about risks and to gather input and feedback. These mechanisms help ensure the
framework achieves an acceptable level of risk that appropriately balances costs and benefits.

37
38
3 Communities

A community is a dynamic system of people, organizations, and patterned relationships and

interactions [Alesch 2005]. Most of these relationships and interactions are physically supported
by a community’s built environment, which is a complex and interdependent network of
engineered subsystems and components, including buildings, bridges, pipelines, transmission
towers, and other structures. Subsequently, the built environment plays a crucial role in enabling
a community to successfully function, providing the physical foundations for much of the
economic and social activities that characterize a modern society. Natural hazards such as
earthquakes, hurricanes, and floods can damage a community’s built environment, which in turn
can disrupt the security, economy, safety, health, and welfare of the public. In response, many
communities have developed and implemented regulatory frameworks to ensure minimum levels
of performance for individual parts of the built environment.
This chapter addresses these issues in greater detail. Specifically, Section 3.1 describes
the configuration of a typical community, including general characterizations of its components,
interactions, and correlations. Section 3.2 summarizes the potential impact earthquakes can have
on a community and its built environment. Finally, Section 3.3 describes the regulatory
framework currently used in the United States to design and evaluate a community’s built
environment to withstand the effects of earthquakes, focusing in particular on building codes and
other engineering standards that establish performance expectations for the built environment.
Using the list of attributes presented in the previous chapter as a guide, Section 3.3 analyzes the
current regulatory framework’s strengths and shortcomings. Ultimately, this analysis will
provide further justification for the engineering framework proposed in subsequent chapters of
this report.

3.1 SYSTEM DESCRIPTION

As detailed in Section 2.2.1, a system is a dynamic entity comprising a collection of interacting,

potentially correlated components assembled to perform an intended function (adapted from
Vesely et al. [1981], Buede [2000], ISO and IEC [2008], and Kossiakoff et al. [2011]). As such,
a community can be considered a system, albeit an incredibly large and multi-faceted one.
Unlike other types of systems (such as nuclear power plants or commercial aircraft), no two
communities are identical. However, many share similar characteristics and configurations. The

39
following subsections describe the basic composition of a typical community, including its key
components, interactions, and potential sources of correlation.

3.1.1 Components

Bea [2007, 2008] defined seven general types of components in a system: structures, hardware,
people, organizations, procedures, environments, and interfaces (see Section 2.2.1 and Table 2.8
for additional discussion). These seven categories will be used to guide the discussion of the
numerous components that comprise a typical community. This discussion, by no means
exhaustive, aims to provide a general sense of the many different components within a
community, focusing in particular on the built environment.

3.1.1.1 Structures
Structures refer to those components that physically support a community and its vital functions.
There are two primary categories of structures: buildings and lifelines. Buildings support a wide
range of functions, including residential, commercial, industrial, and governmental. Lifelines
refer to the systems and facilities that provide services necessary to the function of an
industrialized society and important to emergency response and recovery activities after a
disaster. Lifelines can be grouped into the following five categories (adapted from PCCIP
[1997], Rinaldi et al. [2001], ALA [2004], O’Rourke [2007], and Barkley [2009]): water;
telecommunications; energy (electric power, natural gas, oil, and solid fuels); transportation
(roads and highways, mass transit, ports and waterways, railways, and airports); and waste
disposal (wastewater and solid waste).

3.1.1.2 Hardware
Hardware refers to those components that physically enable the vital functions of a community to
be performed. Typically, hardware works in conjunction with structures to perform these
functions. For example, by themselves, structures like electric transmission towers and lines do
not make a community’s electric power network functional; equipment like generators and
transformers are required in order for the power grid to operate successfully. In general,
hardware has moving parts whereas structures do not. Taken together, structures and hardware
form the built environment.

3.1.1.3 People
In general, people refer to the residents of a community. Residents can serve many different roles
simultaneously, including that of operator, user, and/or member of the general public. Operators
are those residents who actively participate in or enable the vital functions of a community. They
include service and industry workers like truck drivers, firefighters, electricians, custodians,
bankers, city planners, and doctors, to name only a few. Operators typically rely on a specific
subset of structures and hardware to perform their duties successfully. In addition, their behavior
can be strongly influenced by the organizations, procedures, and environments in a community.
For example, firefighters not only require functional communication, transportation, and water

40
infrastructure in order to extinguish fires successfully, but also extensive training and rigorous
command structures.
In addition to operators, residents can also serve as users or customers. In general, users
do not directly participate in the operation of a particular system or service; however, because
they use or consume the service or product, they can be affected if the system or service is
disrupted. For example, an ophthalmologist who relies on public transportation may be unable to
commute to and from work if bus service is disrupted. An especially important group of users
within a community is students, as they have little control over how their community’s education
system is run.
Lastly, residents can serve as members of the general public. Members of the general
public neither operate nor use the service or system under consideration; however, they can still
be affected by its operation. For example, a chemical factory may emit pollutants into the
surrounding environment that affects nearby residents who do not use the chemicals produced by
the factory. In this example, the nearby residents are neither operators nor users, but they are still
affected by operation of the factory.

3.1.1.4 Organizations
Organizations refer to the groups or teams of people that actively participate in the vital
functions of a community. There are two main types of organizations: businesses and
institutions. Businesses provide goods and services to customers for a profit. They include
grocery stores, banks, restaurants, engineering firms, and private utility providers. Institutions
provide vital public services to the residents of a community. They include public and other non-
profit entities like schools, universities, churches, and government agencies (e.g., police and fire
departments, post offices, transit authorities, and public utility providers).
Certain types of organizations specify and enforce the procedures that dictate how people
and other organizations behave. For example, the building department specifies and enforces the
procedures (i.e., building codes) that engineering firms must follow when designing and
constructing buildings. In addition, engineering firms typically specify additional procedures
their engineers must follow; for example, a particular process for analyzing the response of a
building to an earthquake.

3.1.1.5 Procedures
Procedures refer to the formal and informal laws, regulations, guidelines, and customs that
govern a community and its vital functions. They include, for example, legally adopted statutes,
bills, and ordinances, codes and standards, operating manuals, and emergency response plans.
Procedures, which are typically developed, implemented, and enforced by organizations, dictate
the way people and organizations behave. They can also influence how structures and hardware
perform. For example, fuel economy standards affect the types of cars that manufacturers
produce. In light of this discussion, a regulatory framework includes both procedures (i.e.,
regulations and guidance) and the organizations that develop and enforce them.

41
3.1.1.6 Environments
Environments refer to the conditions under which a community and its vital functions are
performed. There are many different types of environments, including natural, economic, social,
and political. Environments can strongly influence the behavior of structures, hardware,
operators, and organizations. For example, the natural environment that surrounds a community
determines the hazards for which its buildings and other structures must be designed. In addition,
the economic environment influences the actions of investment firms, developers, and other
businesses, which in turn can impact the size and condition of a community’s building stock.

3.1.1.7 Interfaces
Interfaces refer to those components in a community that link or connect other components
together. Interfaces enable control in the sense of feedback control, making it possible for
components to interact in a rational manner. An increasingly ubiquitous interface is the Internet,
which can be used, for example, to connect a traffic engineer (i.e., operator) to sensors, cameras,
and other instruments (i.e., hardware) that monitor traffic conditions and loads a bridge (i.e.,
structure). Another example of an interface is the dashboard in a car, which links the driver (i.e.,
operator) to the car’s controls and instrumentation (i.e., hardware).

3.1.2 Interactions

The interactions among its many different components enable a community to perform its vital
functions successfully. These interactions can be extraordinarily complex, especially given the
large number of components in a community. As discussed in Section 2.2.1, these interactions
take the form of dependencies and interdependencies. Figure 3.1 portrays some basic
interdependencies among the lifelines in a community. Note that SCADA stands for supervisory
control and data acquisition. As the figure makes evident, electric power plays a central role in a
community, supplying power to the essential functions of most other lifelines. However, as the
figure also demonstrates, electric power in turn relies on a large number of other lifelines to
operate successfully.
In general, interdependencies increase the vulnerability of lifelines to service disruptions.
For example, telecommunications service can be disrupted on account of internal issues (e.g.,
damage to a switches or cell phone towers); however, because of interdependencies, service can
also be disrupted on account of issues beyond the control of the telecommunications provider
(e.g., power outages). Some of the vulnerabilities arising from these interdependencies can be
mitigated through use of backup or emergency supplies of critical utilities. For example, a
telecommunications center can install onsite diesel generators to supply emergency power to
switches and other vital hardware if the electric power grid goes down. However, as a practical
matter, not all interdependence-related vulnerabilities can be mitigated fully. Furthermore, under
normal operating conditions, interdependencies can serve to increase the operational efficiency
of lifelines.

42
 Figure 3.1 Examples of lifeline interdependencies [Rinaldi et al. 2001].

3.1.3 Correlation

As discussed in Section 2.2.1, correlation measures the relationship between the responses of two
distinct objects. In a complex system like a community, correlation arises when a large number
of its components have similar design or configuration—for example, a neighborhood of
identical apartment buildings. While modular design and construction allows for greater
economies of scale and efficiency, it also increases the vulnerability of the community to the
effects of correlated failures. By incorporating diversity into the design and configuration of its
components, a community can mitigate the impact of correlation.
The procedures developed and utilized by operators and organizations within a
community play an important role in shaping the diversity of its components. For example,
prescriptive provisions in a widely used engineering code or standard may, over time, produce a
large number of structures within a community that have a similar flaw or defect. This
phenomenon was observed in the aftermath of the 1994 Northridge earthquake when a
significant number of welded joints in steel special moment resisting frames failed. These
welded joints were approved for use by the Uniform Building Code (UBC), a national standard
adopted by most communities in California at the time that has since been superseded by the
IBC. In addition, prior to the mid-1970s, provisions in the UBC created a class of buildings,

43
referred to as nonductile concrete buildings, characterized by inadequate seismic detailing that
can result in sudden collapse in an earthquake, therefore posing a serious threat to life and
property [Anagnos et al. 2008; Comartin et al. 2008; Comartin et al. 2011] estimate there are
approximately 17,000 nonductile concrete buildings in the 23 counties with the highest
seismicity and exposure in California. As these examples have illustrated, unintended
correlations that arise as the result of procedures enacted by a community can have significant
impact on the performance of its components (e.g., structures and hardware) in an earthquake.

3.2 VULNERABILITY TO HAZARDS

Communities are vulnerable to a wide range of natural and human-made hazards, including
earthquakes, hurricanes, tornados, floods, economic downturns, pandemics, and terrorist attacks.
This section focuses primarily on earthquakes, which are especially challenging because of their
unpredictability and widespread impact. In particular, the following subsections examine the
types of effects produced by earthquakes, the direct consequences of these effects on
communities, and the cascading consequences that often ensue. Much of the following
discussion can be extrapolated to other types of hazards; however, this extrapolation is beyond
the scope of this report.

3.2.1 Types of Effects

Earthquakes can produce many different effects, though the primary effect is ground shaking
[Scawthorn 2003a]. Depending on the geology of the region, shaking can be felt at great
distances—sometimes hundreds of miles—from the epicenter of an earthquake, though the
intensity of shaking generally decreases as the distance from the epicenter increases. While
ground shaking is typically the most widespread and devastating effect, earthquakes can produce
additional harmful effects, including liquefaction, fault rupture, lateral spreading, landslides, and
tsunamis. Furthermore, when an earthquake occurs, it usually triggers a series of aftershocks.
Sometimes it can even induce additional earthquakes on nearby faults. These aftershocks and
induced earthquakes, which themselves can be sizable, are particularly problematic because they
strike when a community’s built environment is in a weakened state.

3.2.2 Direct Consequences

The most direct consequence of earthquakes involves physical damage to the built environment
of a community. For example, ground shaking can induce significant lateral displacements and
accelerations that damage key structural elements in a building, possibly resulting in partial or
total collapse of the structure. In addition, liquefaction can cause soil instability that ruptures
buried pipelines and damages the foundations of structures. Furthermore, tsunamis can produce
powerful waves that can obliterate entire city blocks. The extent of physical damage caused by
earthquakes depends on many factors, including the location and magnitude of the earthquake
and condition of the community’s built environment. If the physical damage is severe, it can

44
disrupt a large number of a community’s vital functions and result in a significant number of
casualties.

3.2.3 Cascading Consequences

Cascading consequences refer to the sequences of events that result from physical damage to a
community’s built environment. Cascading consequences arise when the direct consequences of
an earthquake cascade through a community, typically following the complex web of component
interactions. For example, damage to gas pipelines can disrupt service to businesses and
residences, and can even trigger large fires that destroy additional infrastructure, including the
water, communication, and transportation systems that firefighters depend on to suppress fires.
In addition, damage to a manufacturing facility can lead to costly downtime that could ultimately
bankrupt the business and force workers to leave town in search of new employment. In turn,
disruption to and potential closure of the plant can impact supply chains throughout the
community, region, and even globe.
Due to an increasingly interconnected global economy, the cascading consequences
caused by an earthquake can extend well beyond areas directly affected by it. The extent of these
consequences depends on several factors, including the extent of the direct consequences and the
importance of the affected community. For example, if an earthquake strikes a community and
causes minor physical damage to its infrastructure, the cascading consequences will also likely
be minor. On the other hand, if an earthquake strikes a city or region and causes extensive
damage, the cascading consequences could be global, as they were following the 2011 Tohoku
earthquake that struck northern Japan.
If the consequences are severe enough, a community may never fully recover after an
earthquake. The combined impact of losses to housing, jobs, schools, and other services may be
too much for a community to handle. Instead of rebuilding, residents may simply choose to leave
and start over elsewhere. Even if the community eventually repairs or rebuilds damaged
infrastructure, the disruption caused by an earthquake may result in irreversible harm to local
businesses as global supply chains shift production to (unaffected) locations, as occurred in
Kobe, Japan, after the 1995 Great Hanshin-Awaji earthquake [Olshansky et al. 2011]. Before the
earthquake, the port of Kobe was the sixth busiest container port in the world. It suffered heavy
damage as a result of the earthquake and, by the time its facilities were fully reconstructed in
1997, the port had dropped to seventeenth busiest [Chang 2000]. Fifteen years after the
earthquake, the volume of containers handled at the port was only 90% of pre-earthquake levels
[City of Kobe 2012].

3.3 CURRENT REGULATORY FRAMEWORK

It may not be possible to fully anticipate all of the cascading consequences, but mitigating the
initial physical damage caused by an earthquake (or other hazard) significantly limits the
consequences that can ensue. To this end, most communities in the United States have developed
and enacted regulatory frameworks to mitigate the direct consequences of earthquakes and other

45
hazards. This section examines these frameworks, giving particular attention to building codes
and other engineering standards that establish performance expectations, either explicit or
implicit, for the built environment in earthquakes. Section 3.3.1 provides an overview of the
current codes and standards for both buildings and lifelines, while Sections 3.3.2 and 3.3.3
describe the strengths and shortcomings, respectively, of these documents. Ultimately, it is these
shortcomings that the engineering framework presented in this report aims to address.

3.3.1 Overview

The regulatory framework that dictates how a community’s built environment (i.e., buildings and
lifelines) should perform comprises a complex web of regulations, enforcement mechanisms, and
guidance. Building codes and other engineering standards play an especially important role
within this framework, specifying provisions and requirements for buildings and lifelines that
ultimately determine how the built environment responds to earthquakes. Therefore, these
documents are the focus of this subsection. It should be noted, however, that other regulations
(e.g., zoning laws, retrofit ordinances, land-use plans) might also affect how a community’s built
environment performs in an earthquake. For example, the Alquist-Priolo Earthquake Fault
Zoning Act, a bill signed into California law in 1972, prohibits construction of buildings on the
surface trace of active earthquake faults with the intent to prevent damage caused by fault
rupture. While important, the following two subsections focus on the codes and standards that
establish performance expectations for buildings and lifelines, respectively, in earthquakes.

3.3.1.1 Buildings
At the heart of the current regulatory framework for buildings in the United States is the building
code, a document that specifies minimum requirements for buildings and other structures in
order to safeguard the health, safety, and general welfare of the public [ICC 2006]. Many
communities in the United States use the IBC. The IBC is a consensus-based document
developed and updated triennially by the International Code Council (ICC), a non-profit, non-
governmental, membership association of engineers, architects, builders, contractors, elected
officials, and others in the construction industry. While the IBC contains many provisions and
requirements of its own, it also references other codes and standards, including ASCE 7
(Minimum Design Loads for Buildings and Other Structures) [ASCE 2006], ACI 318 (Building
Code Requirements for Structural Concrete) [ACI 2011], AISC 360 (Specification for Structural
Steel Buildings) [AISC 2010b], and AISC 341 (Seismic Provisions for Structural Steel
Buildings) [AISC 2010a]. Once adopted by a city, county, or state, the IBC becomes law.
The IBC defines four occupancy categories for buildings: Occupancy Category I, which
includes buildings representing low hazard to human life; Occupancy Category II, which
includes most typical buildings; Occupancy Category III, which comprises high occupancy
structures and buildings containing hazardous materials; and Occupancy Category IV, which
includes facilities essential to emergency response and recovery operations [ICC 2006]. Table
3.1 provides more detailed descriptions of each category. The IBC specifies different design
requirements for each occupancy category. These prescriptive requirements include minimum

46
lateral strength and stiffness for structural systems, as well as guidance for anchoring, bracing,
and accommodation of structural drift for nonstructural systems [BSSC 2009]. In general,
Occupancy Category III and IV buildings are required to have stronger and stiffer structural
systems than Occupancy Category I or II buildings.
Building designs that satisfy the requirements of the IBC are implicitly expected to
achieve certain levels of performance in different earthquake scenarios. These performance
levels, however, are not explicitly stated in the IBC; instead they are discussed in the
commentary to the NEHRP Recommended Provisions [BSSC 2004, 2009], which, via ASCE 7,
serves as the basis for the seismic provisions of the IBC. The NEHRP Recommended Provisions
define four seismic performance levels (operational, immediate occupancy, life safety, and
collapse prevention) and three earthquake hazard scenarios or ground motions (frequent, design
basis, and maximum considered), resulting in twelve possible performance objectives, where a
performance objective comprises a performance level and hazard scenario. Table 3.2 provides a
detailed description of each of the four seismic performance levels.
Figure 3.2 displays the set of performance objectives implicitly assumed for each
occupancy category in the IBC. Again, these performance objectives are not explicitly stated in
the IBC. Figure 3.2 displays three separate lines: one corresponding to Occupancy Category II
buildings (labeled “OC II: Ordinary”); one corresponding to Occupancy Category III buildings
(labeled “OC III: High Occupancy”); and one corresponding to Occupancy Category IV
buildings (labeled “OC IV: Essential”). Each line contains three points, with each point
representing a different performance objective for the particular occupancy category. Therefore,
structures designed in accordance with the provisions of the IBC are expected to satisfy multiple
(three in this case) performance objectives. For example, Occupancy Category IV buildings are
expected to achieve the following three performance objectives: operational performance
following a frequent earthquake; immediate occupancy performance following the design
earthquake; and life safety performance following the maximum considered earthquake (MCE).
As Figure 3.2 demonstrates, the set of performance objectives specified for Occupancy Category
IV buildings are the most stringent, reflecting the essential nature of the functions performed by
these buildings (see Table 3.1).

47
Table 3.1 Occupancy categories for buildings (adapted from ASCE [2006] and ICC [2006]).

Category Nature of occupancy

Buildings and other structures that represent a low hazard to human life in the event of
I
failure
All buildings and other structures except those listed in Occupancy Categories I, III, and
II
IV
Buildings and other structures that represent substantial hazard to human life in the
event of failure, including, but not limited to:
 Buildings where more than 300 people congregate in one area
 Buildings with daycare facilities with a capacity greater than 150
 Buildings with elementary school or secondary school facilities with a capacity
greater than 250
 Buildings with a capacity greater than 500 for colleges or adult education
facilities
 Health care facilities with a capacity of 50 or more resident patients, but not
having surgery or emergency treatment facilities
III
 Jails and detention facilities
Buildings and other structures, not included in Occupancy Category IV, with potential to
cause a substantial economic impact and/or mass disruption of day-to-day civilian life in
the event of failure, including, but not limited to:
 Power generating stations
 Water or sewage treatment facilities
 Telecommunication centers
Buildings and other structures not included in Occupancy Category IV containing
sufficient quantities of toxic or explosive substances to be dangerous to the public if
released
Buildings and other structures designated as essential facilities, including, but not limited
to:
 Hospitals and other health care facilities having surgery or emergency treatment
facilities
 Fire, rescue, ambulance, and police stations and emergency vehicle garages
 Designated earthquake, hurricane, or other emergency shelters
 Designated emergency preparedness, communication, and operation centers
and other facilities required for emergency response
 Power generating stations and other public utility facilities required in an
emergency
IV
 Ancillary structures (communication towers, fuel storage tanks, cooling towers,
fire water storage tanks, etc.) required for operation of Occupancy Category IV
structures during an emergency
 Aviation control towers, air traffic control centers, and emergency aircraft
hangars
 Water storage facilities and pump structures required to maintain water pressure
for fire suppression
 Buildings and other structures having critical national defense functions
Buildings and other structures containing highly toxic substances where the quantity of
the material exceeds a threshold quantity established by the authority having jurisdiction

48
 Table 3.2 Descriptions of the four seismic performance levels implicitly assumed in
the IBC [BSSC 2004].

Performance
Description
level

Represents the least level of damage to the structure. Structures meeting this level when
responding to an earthquake are expected to experience only negligible damage to their
structural systems and minor damage to nonstructural systems. The structure will retain
nearly all of its pre-earthquake strength and stiffness and all mechanical, electrical,
plumbing, and other systems necessary for the normal operation of the structure are
expected to be functional. If repairs are required, these can be conducted at the
Operational
convenience of the occupants. The risk to life safety during an earthquake in a structure
meeting this performance level is negligible. Note, that in order for a structure to meet this
level, all utilities required for normal operation must be available, either through standard
public service or emergency sources maintained for that purpose. Except for very low levels
of ground motion, it is generally not practical to design structures to meet this performance
level.
Similar to the operational level although somewhat more damage to nonstructural systems
is anticipated. Damage to the structural systems is very slight and the structure retains all of
its pre-earthquake strength and nearly all of its stiffness. Nonstructural elements, including
ceilings, cladding, and mechanical and electrical components, remain secured and do not
represent hazards. Exterior nonstructural wall elements and roof elements continue to
provide a weather barrier, and to be otherwise serviceable. The structure remains safe to
occupy; however, some repair and clean up is probably required before the structure can
Immediate be restored to normal service. In particular, it is expected that utilities necessary for normal
occupancy function of all systems will not be available, although those necessary for life safety
systems would be provided. Some equipment and systems used in normal function of the
structure may experience internal damage due to shaking of the structure, but most would
be expected to operate if the necessary utility service was available. Similar to the
operational level, the risk to life safety during an earthquake in a structure meeting this
performance level is negligible. Structural repair may be completed at the occupants’
convenience, however, significant nonstructural repair and cleanup is probably required
before normal function of the structure can be restored.
Significant structural and nonstructural damage has occurred. The structure may have lost
a substantial amount of its original lateral stiffness and strength but still retains a significant
margin against collapse. The structure may have permanent lateral offset and some
elements of the seismic force resisting system may exhibit substantial cracking, spalling,
yielding, and buckling. Nonstructural elements of the structure, while secured and not
Life safety
presenting falling hazards, are severely damaged and cannot function. The structure is not
safe for continued occupancy until repairs are instituted as strong ground motion from
aftershocks could result in life threatening damage. Repair of the structure is expected to
be feasible, however, it may not be economically attractive to do so. The risk to life during
an earthquake in a structure meeting this performance level is very low.
A structure has sustained nearly complete damage. The seismic-force resisting system has
lost most of its original stiffness and strength and little margin remains against collapse.
Substantial degradation of the structural elements has occurred including extensive
cracking and spalling of masonry and concrete elements and buckling and fracture of steel
Collapse
elements. The structure may have significant permanent lateral offset. Nonstructural
prevention
elements of the structure have experienced substantial damage and may have become
dislodged creating falling hazards. The structure is unsafe for occupancy as even relatively
moderate ground motion from aftershocks could induce collapse. Repair of the structure
and restoration to service is probably not practically achievable.

49
 Figure 3.2 Implicit performance objectives corresponding to IBC occupancy
categories [BSSC 2009; ASCE 2000].

In summary, the primary intent of the IBC is to “prevent, for typical buildings and
structures, serious injury and life loss caused by damage from earthquake ground shaking”
[BSSC 2009]. Because most earthquake-related injuries and deaths are caused by structural
collapse of buildings, the focus of code provisions centers on preventing collapse during ground
shaking associated with the maximum considered earthquake (MCE). Specifically, building
designs that satisfy the provisions and requirements of the IBC are expected to have a 1%
probability of collapse in 50 years, which is roughly equivalent to a 10% probability of collapse
in MCE ground shaking [BSSC 2009].
Another important piece of a community’s regulatory framework is the local department
or agency that enforces the building code and its provisions. While the exact structure of these
departments varies from community to community, their duties typically include reviewing plans
and drawings, issuing permits, and inspecting buildings and other structures during and after
construction. Without proper enforcement, building designers and constructers may choose to
ignore certain code requirements if they consider them too onerous or costly. The resulting
buildings will likely have lower quality and less reliable performance in an earthquake than
code-compliant structures. While enforcement is an important component, this report focuses
primarily on the regulations in a framework (i.e., building codes and other documents).
Often a community’s regulatory framework also includes a local hazard mitigation plan.
Federal law requires a community to develop such a plan as a condition for receiving certain
types of non-emergency disaster aid. In general, a hazard mitigation plan provides a long-term
strategy for a community to reduce the risks arising from natural hazards. It must contain, among

50
other things, a risk assessment and a mitigation strategy [USNARA 2009]. A risk assessment
identifies the natural hazards that can affect a community and describes their impact. A
mitigation strategy provides a blueprint for reducing the potential losses identified in the risk
assessment, including a description of mitigation goals to reduce or avoid long-term
vulnerabilities to hazards, an analysis of specific mitigation projects being considered, and an
action plan for prioritizing and implementing the identified projects. Once it is approved,
however, a hazard mitigation plan is not legally binding, unlike the adopted building code.

3.3.1.2 Lifelines
The current regulatory framework that establishes performance expectations for lifelines in
earthquakes is not as easy to characterize as the one for buildings. Figure 3.3 summarizes the
patchwork of codes and standards that currently exist for the different types of lifelines. Barkley
[2009] provides an excellent summary of the current regulatory framework for lifelines:
Guidelines, standards, and code requirements for the seismic performance of lifelines
vary widely. The range of functions and designs of these systems, as well as the range of
potentially damaging hazards, necessitates sector and hazard specific approaches to
reducing damage, ensuring safety, and facilitating system restoration. Consequently,
development of these standards occurs among numerous code development entities, other
professional organizations and private sector entities, and Federal, state, and local
government agencies. These entities have made great strides in developing standards to
reduce risk to lifeline systems in all sectors.

Most sectors have progressed to system-based approaches in order to assess risk and
reduce disruptions the performance of systems and delivery of services to customers.
Nevertheless, achieving a consistent level of resilience is complicated by the many
different regulating bodies to which system operators must answer. The general tendency
toward sector and hazard specific development of standards results in the following
problems:
 A lack of commonly understood definitions for acceptable seismic
performance
 Different standards for performance among different sectors
 A lack of inter-sector coordination for the development of standards,
setting of priorities, and implementation of mitigation
 Limited understanding by political leadership and the general public of
the potential performance of lifelines during an earthquake – and whether
the performance of lifelines will meet expectations
The sector specific natural hazards provisions are generally based on varying levels of
risk (for example, in terms of the design earthquake or probability of occurrence).
Additionally, most sectors do not have standards for reliability – that is, practices that
have been developed to ensure system restoration in accordance with goals set by
stakeholders. According to the American Lifelines Alliance, such standards have been
developed only for highways/roads, ports, and railroads [ALA 2004].

51
Figure 3.3 Matrix of standards and guidelines for lifelines [ALA 2004].

52
3.3.2 Strengths

One of the strengths of the current regulatory framework involves the efficacy in which it
reduces the risk to life. As detailed in Table 3.2 and Figure 3.2, buildings designed in accordance
with the provisions of the IBC are expected to respond to earthquakes in a manner that poses
very low risk to life. Occupancy categories II, III, and IV all achieve life safety performance or
better for earthquakes as severe as the design basis scenario, which has an expected return period
of 475 years.
Recent earthquakes highlight the dramatic improvement in life safety provided by the
current regulatory framework and, in particular, the building code. On 12 January 2010, a
magnitude 7.0 earthquake struck near Port-au-Prince, the heavily populated capital of Haiti. The
city lacks both a modern building code and a means to enforce it [DesRoches et al. 2011]. As a
result, nearly half the buildings in Port-au-Prince, many of which were constructed using
materials and methods prohibited in the IBC, collapsed during the earthquake. While an exact
estimate may never be possible, the resulting destruction claimed the lives of approximately
300,000 Haitians [DesRoches et al. 2011]. In contrast, the magnitude 6.3 earthquake that struck
Christchurch, New Zealand on 22 February 2011 claimed the lives of approximately 180 people
[EERI 2011]. Despite both earthquakes having similar intensity, the casualties in New Zealand
were a small fraction of those experienced in Haiti. The stark difference in the performance
stems largely from the New Zealand building code, which closely resembles the IBC and is
strictly enforced.
Another strength of the current regulatory framework involves the risk-based approach it
uses to establish design requirements for buildings. Instead of specifying a universal set of
provisions that apply to all structures, the building code assigns requirements based on a
structure’s risk to the health, safety, and general welfare of the public. The IBC requires all
structures normally occupied by people (i.e., Occupancy Categories II, III, and IV) to remain
safe during an earthquake, where safe is defined as collapse prevention performance or better.
However, for buildings that pose greater risk or provide vital community services (e.g.,
Occupancy Category III or IV buildings), the IBC requires enhanced performance (see Figure
3.2). Table 3.1 lists examples of buildings in Occupancy Categories III and IV.

3.3.3 Shortcomings

In spite of these strengths, the current regulatory framework has several significant
shortcomings. First, it does not employ an integrated, coordinated, and comprehensive approach
(i.e., expansionist or top-down). Instead, it approaches the design and evaluation of a community
on a component-by-component basis, often treating each component as if it does not interact
with or depend on other components in the community. This approach produces communities in
which most individual components behave as intended during an earthquake; however, when
aggregated, the performance of components can result in unacceptable outcomes for the
community.

53
 As discussed in Section 3.3.1, the provisions contained in the IBC produce individual
buildings that achieve implicit levels of performance (see Figure 3.2). However, these
performance levels are not tied to broader performance objectives for the community, making it
difficult to understand exactly how the community will perform in an earthquake. For example,
if all of the residential buildings in a community achieve life safety performance after a major
earthquake (see Table 3.2), the number of casualties should be small. However, it is less clear
how many residents will be displaced and how long it will take to repair damaged homes.
The local hazard mitigation plan, despite taking a more expansive view, also fails to
establish more detailed performance objectives for a community beyond simply reducing
potential earthquake losses. Consequently, neither it nor the IBC looks at a community
systematically to establish a framework that details exactly how it should perform in an
earthquake. An integrated, coordinated, and comprehensive framework would first specify
performance objectives for the community and then, using these objectives, would establish
targets for individual subsystems and components.
Second, the current regulatory framework is not declaratively performance-based.
Instead, it is highly prescriptive, specifying a large number of requirements and provisions for a
building without first establishing explicit performance objectives for it. The performance
objectives in Figure 3.2 are implicit in nature, meaning that if the prescriptive provisions in the
IBC are satisfied, the building is expected to achieve its corresponding performance objectives.
The IBC, however, does not require an explicit performance evaluation to verify whether these
objectives have been satisfied. Furthermore, the prescriptive requirements of the IBC only ensure
that a typical building (i.e., Occupancy Category II) remains safe after a major earthquake. They
do nothing to address the functionality or reparability of the structure after an earthquake
[Karlinsky 2009]. This makes it difficult to communicate with the public regarding exactly how
individual buildings are expected to perform in an earthquake. Most people believe that a
building designed in accordance with the current building code is “earthquake proof.” In general,
this is not true.
Third, the current regulatory framework fails to account for certain types of risk (i.e., it is
not fully risk-informed). As discussed in Section 3.3.1, the current building code focuses on
safeguarding the health, safety, and general welfare of the public. Consequently, its provisions
allow significant structural and nonstructural damage to occur, provided it does not threaten the
safety of the building’s occupants. While the resulting damage may pose minor risk to life safety,
it can significantly impact the functionality of the building and, ultimately, the entire community.
As detailed in Table 3.2, a building that achieves life safety performance in an earthquake is not
safe for continued occupancy until repairs are made. While repair of the structure is expected to
be feasible, it may not be economically attractive to do so [BSSC 2004]. As a result, residents
may be forced out of their homes for extended periods of time, imposing significant financial
burden on both the families who must relocate and the government agencies that must shelter
them. Similarly, businesses may be forced to close for lengthy periods of time, straining the
finances of business owners, employees, and those who depend on their goods and services (e.g.,
other businesses). A significant outmigration of businesses and people may prevent a community
from fully recovering after an earthquake.

54
 Finally, the current framework fails to account for important components and interactions
within a community. Buildings, which fall under the purview of the building code, are but one
piece of a community’s physical infrastructure. Lifelines also play a crucial role in a
community’s ability to function, both on a daily basis and in the aftermath of an earthquake. In
spite of this, performance standards for lifelines vary widely and are not tied to generally
applicable public policies for reducing risk in the face of a major earthquake [Barkley 2009;
Poland et al. 2009]. Furthermore, the standards that do exist are not performance-based, making
it difficult to determine exactly how these crucial systems will respond and interact with other
components in the community [Barkley 2009]. This, in turn, makes it difficult to fully
understand how a community will perform.
The engineering framework presented in subsequent chapters of this report addresses
these shortcomings. It employs a transparent, performance-based, risk-informed methodology in
order to establish performance targets for the numerous subsystems and components within the
built environment that are consistent with broad resilience goals for the community. The
proposed engineering framework is an adaptation of the regulatory framework used to design
and evaluate the safety of nuclear power plants in the United States. This nuclear framework is
described in the next chapter.

55
56
4 Nuclear Design Philosophy

The regulatory framework currently used to design, analyze, and regulate commercial nuclear
power plants in the United States offers a promising template for communities to follow. Despite
obvious differences in function and configuration, both communities and nuclear power plants
are multi-faceted, dynamic systems comprising many interacting subsystems and components
that cut across a diverse range of disciplines and professions. The current nuclear regulatory
framework handles these numerous subsystems and components in a consistent and logical
manner, informed partly by an explicit set of system-level performance expectations for the
nuclear power plant. Furthermore, the tools and procedures employed by the current nuclear
regulatory framework have been implemented successfully and refined extensively over the past
several decades, resulting in significant improvements in the understanding of how these
complex, dynamic systems behave.
This chapter examines the regulatory framework used to design and analyze nuclear
power plants and their numerous subsystems and components. This examination is by no means
exhaustive; instead it focuses on the parts of the framework with the most potential applicability
to communities. Section 4.1 provides a brief overview of the current regulatory framework for
nuclear power plants. Section 4.2 defines important nuclear terms and concepts that will be used
throughout this chapter. Section 4.3 describes the design philosophy codified in current
regulations. Section 4.4 discusses several important performance evaluation tools used to analyze
the response of nuclear power plants, including probabilistic risk assessments (PRAs), event
trees, and fault trees.
The engineering framework presented in Chapters 5, 6, and 7 of this report adapts the
nuclear framework for use in a community setting. In particular, Chapter 5 describes how the
nuclear concepts defined in Section 4.2 can be reinterpreted, subsequently providing a more
transparent, integrated, and consistent basis for the design and evaluation of communities.
Chapter 6 presents a set of event trees that can be used to link community-level resilience goals
to specific performance objectives for individual components and subsystems within the built
environment. This set of event trees, together with the definitions presented in Chapter 5, forms
the foundations of the engineering framework proposed in this report. Finally, Chapter 7
demonstrates two potential applications of the proposed engineering framework using conceptual
examples.

57
4.1 OVERVIEW

A nuclear power plant is a complex, multi-faceted, dynamic system. The focus of this chapter,
however, is not the configuration or inner workings of the plant itself but rather the regulatory
framework that dictates how it is designed, analyzed, constructed, and operated. More
specifically, it is the general design and analysis philosophy codified in the current regulatory
framework that is of particular interest. This philosophy, which is described in more detail in the
following sections, provides a structured methodology for evaluating and mitigating the impact
of natural hazards and other adverse events on the safety of a nuclear power plant. It also handles
these numerous subsystems, components, and interactions in a consistent and logical manner,
informed partly by an explicit set of system-level performance expectations for the nuclear
power plant. It is this general methodology that will be adapted and applied to communities in
subsequent chapters of this report.
In summary, the nuclear design philosophy begins at the system level, defining undesired
outcomes whose occurrence should be avoided to the extent possible. It then identifies both the
vital plant functions that must be maintained in order to prevent these undesired outcomes from
occurring and the components and subsystems within the plant that support these vital functions.
Next, it establishes performance targets for both the overall nuclear power plant system and,
subsequently, its numerous subsystems and components. Finally, in order to ensure the plant
satisfies these targets, the philosophy requires a detailed analysis of the system and its
components in order to verify that the plant design satisfies the required system-level
performance targets.

4.2 DEFINITIONS

The following subsections explain important terms and concepts from the nuclear regulatory
framework, including undesired outcomes, accidents and accident sequences, vital functions, and
frontline and support systems. In Chapter 5 these concepts will be extrapolated to communities.

4.2.1 Undesired Outcomes

In general, an undesired outcome represents a situation that inhibits the ability of a system to
maintain functionality, the consequences of which can adversely impact the safety and welfare of
the general public and surrounding environment. As such, a primary focus of the regulatory
framework that governs the design and operation of such systems should involve minimizing the
occurrence of these undesired outcomes to the extent possible. A wide range of events, including
natural and human-made hazards, can trigger these undesired outcomes.
The nuclear regulatory framework defines two such undesired outcomes: core damage
and large release of radioactivity. Core damage refers to damage to the nuclear fuel assemblies in
the reactor core. It occurs when the reactor core losses sufficient cooling, resulting in heating of
the core to the point that it damages the nuclear fuel [ANSI and ANS 2003]. Core damage ranges
in severity depending on the length of time the core goes without cooling. If enough of the core

58
is damaged, large amounts of radioactive material can be released from the reactor core and,
possibly, into the surrounding environment. Subsequently, core damage is a necessary precursor
to a large release of radioactivity. Large release represents the worst-case outcome for a plant,
not only dooming future operation of the plant, but also resulting in offsite contamination,
casualties, and other adverse health effects. This hierarchy of undesired outcomes (i.e., core
damage followed by large release) serves as the basis for many of the provisions and
requirements that regulate the design and operation of nuclear power plants.
It is important to note that core damage and large release events may still take place (e.g.,
the nuclear accident at Fukushima Daiichi nuclear power plant after the 2011 Tohoku
earthquake). However, these events should occur infrequently, depending in part on the risk
targets codified in the regulatory framework.

4.2.2 Accidents and Accident Sequences

The International Atomic Energy Agency (IAEA) broadly defines an accident as “any
unintended event, including operating errors, equipment failures and other mishaps, the
consequences or potential consequences of which are not negligible from the point of view of
protection or safety” [IAEA 2007]. An accident sequence is the representation of an accident as a
series of events that may or may not result in a specified undesired outcome or end state, which
for a nuclear power plant is either core damage or large release. The first event in an accident
sequence, referred to as the initiating event, is any event, either internal or external, that perturbs
normal operations of the plant, whether operating or not [ANSI and ANS 2003]. Initiating events
include random subsystem and component failures, earthquakes, floods, fires, tornadoes, and
aircraft impact, to name a few. The initiating event triggers a sequence of events that, depending
on the combination of component, function, and operator failures or successes, may or may not
result in the specified undesired outcome. If the undesired outcome does not result at the end of
an accident sequence, the plant has avoided core damage or large release and is in a safe state.
For the purposes of illustration, Figure 4.1 displays a graphical representation of a simple
accident sequence resulting in core damage. The sequence, which begins with an initiating event,
comprises three additional events: Component A fails, Component B fails, and Component C
fails. Core damage results only if all three events occur. Therefore, in this particular example, the
accident sequence describes a simple parallel configuration with three components, A, B, and C.
In a parallel configuration, all components must fail for the system to fail (i.e., core damage
occurs). In contrast, in a series configuration, the system will fail if any one of its components
fails. Note that Figure 4.1 represents only one of a potentially large number of accident
sequences for the system being analyzed. These other sequences can involve additional
components within the system, and can also feature both parallel and series configurations of
components.

59
 Figure 4.1 Graphical representation of a simple accident sequence.

Accident sequences can also be represented using Boolean expressions. A Boolean

expression is mathematical construct that uses logical operators (intersection, union, etc.) to
combine and give order to the events in an accident sequence. The symbol  represents the
intersection logical operator and denotes the occurrence of both events (i.e., both event X and
event Y occur). The symbol  represents the union logical operator and denotes the occurrence
of either event (i.e., event X or event Y occurs).
Equation (4.1) shows the Boolean expression for the accident sequence in Figure 4.1. In
words, Equation (4.1) says that if Component A fails and Component B fails and Component C
fails, then core damage results. Again, note that Equation (4.1) represents only one of many
potential accident sequences for the system. Accident sequences will be discussed further in
Section 4.4.2, which describes event trees.

Component A fails  Component B fails  Component C fails (4.1)

4.2.3 Vital Functions

When an initiating event occurs at a nuclear power plant, three vital safety functions need to be
maintained in order to prevent core damage and subsequent large release [IAEA 2009]. The first
vital function involves reactivity control: the self-sustaining chain reaction in the reactor core
must be shut down. For most reactors, this involves insertion of control rods that absorb neutrons
and prevent further fission. The second vital function involves cooling the fuel. Once the chain
reaction is shut down, heat generated by the fuel needs to be removed from the core to ensure
that the core does not incur damage such as melting, which would release radioactivity. In most
reactors, this requires circulating a steady supply of cool water through the reactor core for
several months, possibly longer. The third and final vital function involves confinement:
radioactive material must not be allowed to escape into the environment. This typically involves
relieving pressure in the core to ensure the steel reactor vessel does not rupture or the concrete
containment structure does not crack. Failure to perform one of these functions does not
necessarily result in core damage and/or large release; however, maintaining all of them ensures
that these undesired events do not occur.

60
4.2.4 Frontline and Support Systems

Most nuclear power plant designs feature multiple redundant systems to perform each vital plant
function. Depending on their role in an accident, these systems can be classified as either
frontline or support systems. In a nuclear power plant, frontline systems are the engineered safety
systems that deal directly with preventing an accident [Fullwood 2000]. In other words, they
directly enable the vital functions in a nuclear power plant. Support systems, on the other hand,
refer to those systems that support frontline systems and even other support systems. For
example, the containment spray system, a frontline system that prevents over-pressurization of
the containment structure, depends on several support systems, including electrical and water
systems. Most nuclear power plant designs feature multiple redundant frontline systems for each
vital safety function, as well as multiple redundant support systems for each frontline system. For
example, for frontline systems requiring power, there are typically several support systems
available to provide electricity, including offsite AC power, onsite AC power (often via diesel
generators), and onsite DC power.

4.3 DESIGN AND REGULATORY PHILOSOPHY

In the context of this report, the structure of the regulatory framework currently used to design
and analyze nuclear power plants is not as important as the philosophy that developed and
shaped it because it is this philosophy that will be referred to and used in subsequent chapters of
this report. Therefore, this section focuses primarily on the design and regulatory philosophy
used in the nuclear industry in the United States. This philosophy has evolved substantially since
the first nuclear power plants were built in the 1960s; the traditional, prescriptive, deterministic
approach has given way (partly) to a more risk-informed, performance-based, probabilistic
approach. The following subsections discuss two key concepts: defense-in-depth and risk-
informed regulation.

4.3.1 The Concept of Defense-in-Depth

The concept of defense-in-depth is central to the current nuclear design and regulatory
philosophy. Despite its fundamental role, there exists no official or preferred definition of the
term [Sorenson et al. 1999]. The concept has evolved and expanded significantly since its initial
development in the 1950s. One common interpretation defines defense-in-depth as the multiple
physical barriers that prevent escape of radioactive material. These barriers typically include
cladding on the fuel assemblies, the steel reactor vessel, and the concrete containment structure.
Another common interpretation defines defense-in-depth as the high-level lines of defense in a
nuclear power plant. These lines of defense are typically threefold: preventing the initiation of
accident sequences; rapidly terminating those sequences that do occur; and mitigating the
consequences of sequences that cannot be terminated successfully. A common thread in both
interpretations is the deployment of successive levels of defense to ensure that safety of the plant
is not dependent on only a single function or system [Sorenson et al. 1999]. This idea of multiple
levels of protection is the central feature of defense-in-depth [INSAG 1999].

61
 Table 4.1 Levels of defense-in-depth [INSAG 1999].

Level Objective Essential means

1 Prevention of abnormal operation and Conservative design and high quality

failures in construction and operation
2 Control, limiting, and protection
Control of abnormal operation and
systems and other surveillance
detection of failures
features
3 Control of accidents within the design Engineered safety features and
basis accident procedures
4 Control of severe plant conditions, Complementary measures and
including prevention of accident accident management
progression and mitigation of the
consequences of severe accidents
5 Mitigation of radiological Offsite emergency response
consequences of significant releases
of radioactive material

Over the years the concept has expanded into an overall safety strategy for the nuclear
industry. When applied properly, defense-in-depth ensures that no single human or equipment
failure will lead to harm to the public, and even most combinations of failures will result in little
or no harm [INSAG 1999]. Defense-in-depth is structured in five levels. Table 4.1 describes the
objective of each level and the essential means for satisfying each objective. The five levels of
defense are successive in that failure of one level calls into action the subsequent level. Events
like earthquakes and fires, however, can impair multiple levels of defense simultaneously. For
this reason, these hazards typically receive special consideration in order to limit their impact.
The current regulatory framework embodies the defense-in-depth strategy [Sorenson et
al. 1999]. Its regulations are derived by repeated application of the question: what if a particular
barrier or safety feature fails? The resulting set of provisions, which is prescriptive and
deterministic in nature, ensures that nuclear power plants have multiple lines of defense for each
barrier or safety feature, regardless of the probability that it may be required. In general, this
conservative approach has served the nuclear industry in the United States well in terms of
safety; however, in certain instances, it has resulted in excessive regulatory burden [Sorenson et
al. 1999].

4.3.2 Risk-Informed Regulation

While most of its regulations are prescriptive and deterministic in nature, the Nuclear Regulatory
Commission (NRC), the government agency that oversees commercial nuclear power plants in
the United States, has been working over the past few decades to make the current nuclear
regulatory framework more risk-informed. This change stems from the need to better understand
and quantify the risks posed by current and future nuclear power plants to the health and safety
of the public, where risk is defined as the product of two quantities: (1) the probability or

62
likelihood of an event occurring (i.e., equipment failure or human error) and (2) the
consequences associated with its occurrence. As mentioned in the previous subsection, most
current regulations were developed through consideration of questions that focus on only the
second half of the risk equation: namely, what can go wrong and what are the consequences?
Risk-informed regulation, on the other hand, involves asking a third question: how likely is it
that something goes wrong? This third question helps ensure that the various burdens imposed by
regulations are appropriate to their importance in protecting the health and safety of the public
and the environment.
As an initial step towards a more risk-informed framework, the NRC issued a policy
statement in 1986 that established an acceptable level of radiological risk to the public from
nuclear power plant operation. In support of this risk target, the statement specifies two
qualitative safety goals, which, in turn, are supported by two quantitative health objectives
[USNRC 1986]. The two qualitative safety goals are:
 Individual members of the public should be provided a level of protection
from the consequences of nuclear power plant operation such that
individuals bear no significant additional risk to life and health.
 Societal risks to life and health from nuclear power plant operation should
be comparable to or less than the risks of generating electricity by viable
competing technologies and should not be a significant addition to other
societal risks.
The two quantitative health objectives are:
 The risk to an average individual in the vicinity (within one mile) of a
nuclear power plant of prompt fatalities that might result from reactor
accidents should not exceed one-tenth of 1% (0.1%) of the sum of prompt
fatality risks resulting from other accidents to which members of the U.S.
population are generally exposed.
 The risk to the population in the area near (within ten miles) a nuclear
power plant of cancer fatalities the might result from nuclear power plant
operation should not exceed one-tenth of 1% (0.1%) of the sum of cancer
fatality risks resulting from all other causes.
These qualitative safety goals and quantitative health objectives can be thought of as
system-level performance targets for a nuclear power plant. The qualitative safety goals provide
the basis for the quantitative health objectives, which in turn provide the basis for more specific
numerical performance targets that focus specifically on the avoidance of core damage and large
release. Core damage frequency (CDF) measures the number of core damage occurrences
expected per year of operation for an individual reactor. The CDF for a plant is calculated using
the probabilistic risk assessment (PRA) methodology, which is discussed more detail in Section
4.4.1. Similarly, large release frequency (LRF) measures the number of large release occurrences
expected per year of operation. The LRF for a plant is also calculated using a PRA. The NRC has
set as a target the expectation that all operating nuclear power plants have mean CDFs less than

63
1×10-4 per reactor-year and mean LRFs less than 1×10-5 per reactor-year [USNRC 2002]. This
translates into less than one core damage event for every 10,000 reactor-years and less than one
large release event for every 100,000 reactor-years, where a reactor-year refers to a year of plant
operation. In philosophy, these targets can be considered equivalent to implicit performance
expectation that buildings designed according to the provisions of the IBC have a 1% probability
of collapse in 50 years [BSSC 2009].
Development in the mid-1980s of these system-level targets for nuclear power plants was
driven primarily by the capabilities of the technology that existed at the time, not by a public
policy decision as to what might be adequately safe, although this consideration did play a role.
The NRC developed the targets using results from PRAs of several nuclear power plants built
before 1980. As a result, most nuclear reactors designed and constructed before the NRC
established these targets were expected to satisfy them and, when evaluated, they all did. New
plants, whose designs include a variety of advanced technology features, will likely perform
better than the NRC targets; however, like previously constructed plants, they are not required to
meet a specific risk target.
While they do not replace the deterministic regulations based on defense-in-depth
principles, the system-level goals and targets help provide a more rational and transparent
foundation for the current regulatory framework for nuclear power plants. In other words, they
represent the expected level of performance achieved by satisfying the prescriptive requirements
of the current framework. All nuclear power plants currently operating in the U.S. meet these
goals and objectives with considerable margin. Crucially, if it were to be found that a reactor did
not meet one of these safety goals or health objectives, the NRC and the reactor owner/operator
would perform a detailed investigation to understand the reason in order to achieve enough
improvements to bring the reactor back in compliance with the specified requirements.
Since the 1986 policy statement, the NRC has gradually been updating its regulations to
make them more risk-informed, though currently most regulations are still deterministic in
nature. This is due primarily to reluctance to rely fully on the results of PRAs. In spite of this, the
PRA methodology has been an essential tool in facilitating the gradual shift from a deterministic
to a more risk-informed regulatory structure. In 1995, the NRC issued a policy statement
addressing the use of PRA in nuclear regulatory activities. It states: “The use of PRA technology
should be increased in all regulatory matters to the extent supported by the state-of-the-art in
PRA methods and data and in a manner that complements the NRC's deterministic approach and
supports the NRC's traditional defense-in-depth philosophy” [USNRC 1995].
While the 1995 policy statement clearly embraces the increased use of PRA, it does so
with caution. Limitations of the PRA methodology, coupled with uncertainty and incompleteness
in the understanding of how nuclear power plants behave during severe accidents, make the NRC
reluctant to trust fully the results obtained from PRAs. Consequently, the NRC uses defense-in-
depth to compensate for these shortcomings in understanding. This approach has resulted in a
regulatory framework that is risk-informed rather than risk-based. A risk-informed framework
uses risk information to develop regulations for those items most important to safety; however, it

64
reserves the right to impose additional regulations (i.e., extra lines of defense) in order to
compensate for any potential uncertainty and/or incompleteness of knowledge.

4.4 PERFORMANCE EVALUATION TOOLS

In order for a regulatory framework to establish and successfully implement quantitative design
targets, adequate performance evaluation tools must be available. Since its conception in the
mid-1970s, the PRA methodology has gained increasing prominence in the nuclear industry as
one such analysis tool. The following subsections describe the methodology in detail, focusing
on key elements that will be used later in this report to evaluate communities.

4.4.1 Probabilistic Risk Assessment

In general, a PRA tries to answer the three following questions: (1) what can go wrong; (2) how
likely it is; and (3) what are the consequences if it occurs [Kaplan and Garrick 1981]. For a
complex system like a nuclear power plant, the answers to these questions can be highly
uncertain, stemming primarily from limitations in our knowledge of how the systems behave
during severe accidents. The PRA methodology provides a rational, consistent framework by
which to evaluate these uncertainties and, ultimately, produce an estimate of risk.
The scope of a PRA depends on the nature of the risk being evaluated. The nuclear
industry uses the PRA methodology to estimate three different levels of risk: a Level 1 PRA
evaluates the risk of core damage at a plant; a Level 2 PRA estimates the risk of radioactive
release at a plant; and a Level 3 PRA, or consequence analysis, quantifies the risk of radiation
exposure to the public and the environment arising from plant operation [ANS and IEEE 1983].
Figure 4.2 shows the relationship between the three levels of PRA. As can be seen in the figure,
a Level 3 PRA uses results from a Level 2 PRA, which in turn uses results from a Level 1 PRA.
In principle, the PRA methodology is relatively straightforward. After establishing the
scope of the analysis (i.e., Level 1, 2, or 3 PRA), the first step involves identifying initiating
events that have potential to disrupt steady-state operation of the reactor. Initiating events can be
either internal or external to the plant. Internal initiating events typically involve operator errors
and random failures of important equipment and components. External initiating events include
earthquakes, hurricanes, tornados, fires, floods, and aircraft impact. Incomplete or ill-defined
initiating events can lead to inaccuracies in the PRA results; therefore a great deal of emphasis
must be placed on selecting a comprehensive and appropriate set of initiating events.
The second step in the PRA methodology involves identifying all frontline and support
systems (including component failure rates, fragilities, and dependencies) that can be called upon
during an accident sequence. As discussed in Section 4.2.4, frontline systems directly enable the
basic safety functions in a plant, while support systems enable the frontline systems. Once all
relevant systems have been identified, dependency matrices can be developed. A dependency
matrix portrays any direct dependencies that exist between frontline and support systems or

65
among different support systems. These matrices will be used in the next step to determine how
the failure of certain systems impacts the availability of others.
The third step in the PRA methodology involves performing a systems analysis of the
plant to enumerate all possible accident sequences that can result from the initiating events
identified in the first step. Each accident sequence involves a different series of events that,
depending on the combination of component, function, and operator failures or successes, may or
may not result in the specified undesired outcome (i.e., core damage or large release). There can
be thousands or even millions of sequences for a plant. Accident sequences are portrayed
graphically using event trees, which are described in more detail in Section 4.4.2.

Figure 4.2 The three levels of PRA used in the nuclear industry [USNRC 2012a].

66
 The final step in the PRA methodology involves compiling the information obtained in
previous steps to produce an estimate of risk. For a Level 1 PRA, this means computing the CDF
for the plant, which is obtained by summing the frequencies of all core damage accident
sequences identified in the third step. For a Level 2 PRA, this involves computing the LRF for
the plant. And for a Level 3 PRA, this means estimating the consequences (e.g., physiological,
environmental, and economical issues) of core damage and large release on the surrounding
environment and the population in neighboring communities. When these consequences are
considered together with the CDF and LRF, an estimate of risk can be obtained.
While in principle the methodology is relatively straightforward, in practice, performing
a PRA can be quite difficult. Much of this difficulty stems from limitations in current knowledge
(i.e., epistemic uncertainty). Table 4.2 lists several of these limitations. It is important to
emphasize, however, that these are not limitations in the PRA methodology itself, but rather
limitations associated with its use and application. In other words, PRAs have made these
shortcomings more transparent. Over time, with increased knowledge and better data, many of
these limitations can be overcome. In the meantime, by exposing these limitations, performing a
PRA serves as an effective way in figuring out areas where current knowledge needs most
improvement.

Table 4.2 Limitations associated with the use of the PRA methodology.

Limitation

1 Inability to anticipate fully all possible initiating events and their subsequent effects
2 Insufficient data to quantify accurately the frequency of occurrence of initiating events
Insufficient understanding of the failure mechanics and modes for systems and
3
components
4 Insufficient data to quantify accurately the failure rates or fragilities of components
5 Inability to anticipate fully all possible dependencies among systems and components
6 Inability to enumerate fully all potential accident sequences
7 Inability to understand fully the consequences of severe accidents

4.4.2 Event Trees

An event tree is a graphical representation of the various accident sequences that can occur as a
result of an initiating event [USNRC 2012b]. It is an essential tool in analyzing whether a
nuclear power plant satisfies its system level design targets (i.e., CDF and LRF targets). It
provides a rational framework for enumerating and, subsequently, evaluating the myriad events
and sequences that can affect a nuclear power plant.
The top half of Figure 4.3 shows a simple example of an event tree and will be used to
explain its basic structure and logic. While the event tree in Figure 4.3 is much simpler than one

67
for an actual nuclear power plant, the principles remain the same. All event trees begin with an
initiating event—in this example, jumping from an airplane (see the red box in Figure 4.3). In
general, for a nuclear power plant, an initiating event is anything that perturbs steady-state
operation (e.g., an earthquake, fire, flood, etc.). After the initiating event, a series of top events
follows (see the yellow box in Figure 4.3). Each top event corresponds to a subsystem or
component required to prevent the undesired outcome from occurring. In this example, the
undesired outcome involves injury or death of the person jumping from the airplane. For a plant,
it is typically core damage or large release.
The event tree in Figure 4.3 comprises two top events. The first involves the main
parachute while the second involves the reserve chute. After the initiating event occurs, the main
parachute is called upon. This first top event can either fail or succeed. A downward branch in an
event tree indicates that the corresponding top event has failed to occur, while an upward branch
indicates the event has occurred successfully. In this example, a downward branch means the
main parachute fails. An upward branch, on the other hand, means the main chute succeeds. Note
that the upward branch results in the jumper landing safely, which is the first of three possible
outcomes shown in Figure 4.3.
After the first top event, the second top event is called upon. Note, however, that for this
particular example, if the first event is successful then the second top event is not called upon. In
other words, the jumper does not need the reserve chute if the main one succeeds; only if the
main parachute fails will the reserve chute be called upon. This second top event can either fail
or succeed. Again, a downward branch represents failure and an upward branch success. If the
reserve chute succeeds, the jumper lands safely (the second outcome in Figure 4.3). If, on the
other hand, the reserve chute fails, the undesired outcome results, which is the third outcome
shown in Figure 4.3.
In the end, the event tree in Figure 4.3 contains three accident sequences, with one
resulting in the undesired outcome (i.e., injury or death of the jumper). This failure sequence is
represented by the Boolean expression in Equation (4.2).

 Main chute fails   Reserve chute fails (4.2)

In order to compute the probability of the undesired outcome, Pf, we need to compute the
probability of the accident sequence in Equation (4.2). If the two events are independent, Pf is
simply:

Pf  Pf ,mc  Pf ,rc (4.3)

Where Pf,mc is the probability the main chute fails and Pf,rc is the probability the reserve
chute fails. In order to compute these two quantities, we need to perform analyses of both the
main and reserve chutes. Fault trees, which are discussed in the next subsection, provide one
such methodology for doing so.

68
 Figure 4.3 Example of a simple event and fault tree [USNRC 2012a].

4.4.3 Fault Trees

A fault tree is an analytical model that graphically depicts the logical combinations of faults (i.e.,
hardware failures and/or human errors) that can lead to an undesired state (i.e., failure mode) for
a particular subsystem or component [Vesely et al. 1981]. This undesired state serves as the
topmost event in the fault tree and usually corresponds to a top event in an event tree. Thus, a
fault tree provides a rational framework for identifying the combinations of hardware failures
and/or human errors that can result in a particular failure mode of a subsystem or component.

69
Once fully developed, a fault tree can be used to evaluate the subsystem or component
quantitatively.
The bottom half of Figure 4.3 shows a simple example of a fault tree and will be used to
explain its basic structure and logic. While the tree in Figure 4.3 is much simpler than one for an
actual subsystem in a nuclear power plant, the principles remain the same. As mentioned in the
previous paragraph, the topmost event in a fault tree corresponds to a top event in an event tree.
The green box in Figure 4.3 explicitly highlights this connection. The top event in this example
involves failure of the reserve chute. From the topmost event, the reader works downward
through the fault tree. Directly beneath the top event is an OR-gate. To pass through an OR-gate,
one or more of the events directly connected to the gate must occur. In this example, there are
two connected events: “Chute Not Deployed” and “Chute Tangled.” In other words, the reserve
chute will fail if it does not deploy or if it gets tangled.
Directly beneath the “Chute Not Deployed” event in Figure 4.3 is an AND-gate. To pass
through an AND-gate, all events directly connected to the gate must occur. In this example, there
are two connected events: “Rip Cord Breaks” and “Auto Activation Device Fails.” Both of these
events must occur for the reserve chute to not deploy. Lastly, beneath the “Auto Activation
Device Fails” event is another OR-gate, which is connected to two events: “Altimeter
Malfunctions” and “Battery Is Dead.” If either of these events occurs, the auto activation device
will fail.
The fault tree in Figure 4.3 can be represented using a Boolean expression. See Equation
(4.4). Table 4.3 explains the symbols used in Equation (4.4).

T  A E1  A (B  E2 )  A  B  (C  D)  A (B C)  (B  D) (4.4)

According to Equation (4.4), the top event (failure of the reserve chute) will occur if any
one of the following occurs: the chute tangles; the ripcord breaks and the altimeter malfunctions;
or the rip cord breaks and the battery is dead. In Section 4.4.2, we were interested in computing
Pf,rc, the probability that the reserve chute fails. Using the Boolean expression in Equation (4.4),
we can now do so. See Equation (4.5).

    
Pf ,rc  P A  B  (C  D)  P( A)  P B  (C  D)  P A  B  (C  D)  (4.5)

After some manipulation, Equation (4.5) expands to the following:

Pf ,rc  P( A)  P(B C)  P(B  D)  P(B C  D)  P( A B C )  (4.6)

P( A B  D)  P( A B C  D)

If we assume A, B, C, and D are all independent, Equation (4.6) simplifies to the following:

Pf ,rc  PA  PB  PC  PB  PD  PB  PC  PD  PA  PB  PC  PA  PB  PD  PA  PB  PC  PD (4.7)

70
where PA = P(A) is the probability the chute tangles; PB = P(B) is the probability the ripcord
breaks; PC = P(C) is the probability the altimeter functions; and PD = P(D) is the probability the
battery is dead. These quantities can be estimated from historical data, laboratory testing, and/or
analytical modeling.
In summary, the event and fault trees presented Figure 4.3 provide a structured, rational
methodology for identifying and quantifying the risk associated with a particular activity (in this
case, jumping out of an airplane). These performance evaluation tools have been implemented
successfully in the design and analysis of nuclear power plants for several decades. In the next
two chapters, these tools will be extended and applied to communities in order to create an
engineering framework that explicitly links community-level resilience goals to specific
performance targets for individual components and subsystems within the built environment.

Table 4.3 Explanation of the symbols used in Equation (4.4).

Symbol Event

T Reserve chute fails (i.e., top event)

 AND-gate (i.e., intersection)
 OR-gate (i.e., union)
E1 Chute not deployed
E2 Auto activation device fails
A Chute tangled
B Ripcord breaks
C Altimeter malfunctions
D Battery is dead

71
72
5 Adaptation of Nuclear Design Philosophy

This chapter marks the beginning of the presentation of the proposed engineering framework.
The need for this framework arises from the observation that, under the current regulatory
framework for communities, performance objectives for components and subsystems within the
built environment are not tied to broader performance goals for the community. To remedy this
shortcoming, the engineering framework presented in this chapter and the next seeks to develop
an integrated, comprehensive, and consistent methodology for establishing performance targets
for individual components. This methodology properly accounts for both the numerous
interactions among components and subsystems and also broader community resilience goals.
The proposed engineering framework adapts parts of the regulatory framework used in
the United States to design, analyze, and regulate commercial nuclear power plants. In particular,
it extends and applies the general design philosophy described in the previous chapter to
communities. In summary, this philosophy, as adapted to communities, comprises three main
steps. The first step defines undesired outcomes for a community whose occurrence should be
avoided to the extent practical. The second step identifies both the vital community functions
that must be maintained in order to prevent these undesired outcomes from occurring and the
frontline and support systems within the built environment that support these vital functions.
Lastly, the third step establishes performance targets for both the overall community and,
subsequently, its numerous frontline and support systems.
This chapter focuses on the first two steps in the above philosophy: Section 5.2 discusses
the range of potential undesired outcomes that can affect a community; Section 5.3 identifies and
describes the vital community functions that prevent the undesired outcomes from occurring; and
Section 5.4 lists the frontline and support systems within the built environment that enable the
vital community functions. Chapter 6 focuses on the last step of the philosophy, outlining a
methodology that can be used to link community-level performance goals to specific
performance targets for individual components and subsystems within the built environment.
Section 5.5 helps set the stage for Chapter 6, describing how the performance evaluation tools
presented in Chapter 4 are adapted for use in a community setting.

73
5.1 CAVEATS

Before describing the conceptual foundations of the proposed engineering framework, it is

important to recognize that not all of the concepts presented in Chapter 4 lend themselves
perfectly to extension to communities. Important differences exist between nuclear power plants
and communities. One distinction involves physical scale. Communities occupy much larger
geographic areas than nuclear power plants, meaning certain subsystems and components in a
community, especially lifelines, can be spatially distributed over a potentially large area. As a
result, it becomes necessary to account for partial failures of these subsystems and components.
For example, an earthquake may cause damage to portions of a community’s electric power grid,
resulting in service disruptions to particular neighborhoods or city blocks. The evaluation of
nuclear power plants does not account for partial failures: in the safety analysis of these plants,
the assumption is generally that a particular subsystem or component is either functional or has
failed completely.
Another distinction involves external boundaries of the system. Most components and
subsystems in a nuclear power plant reside within the well-defined physical boundaries of the
plant. A community, on the other hand, can rely on components and subsystems that fall outside
its jurisdictional boundaries. For example, a community’s electric power grid may draw
electricity from a generating station hundreds of miles away. An event that disrupts the
functionality of the station may cause service disruptions in the community, even though its
electric power grid is not directly affected by the event. In general, these types of interactions do
not exist in nuclear power plants.
The final distinction involves time scale. A community’s built environment is constructed
over time, expanding and evolving over the course of decades or even centuries as the
community’s population grows and/or its needs change. Consequently, individual components
within the built environment have likely been designed and constructed using substantially
different specifications and standards, meaning that the expected performance of similar
components (e.g., residential buildings or highway bridges) within a community can vary
drastically. In comparison, nuclear power plants are built over a relatively short period of time,
with most of their components and subsystems being designed and constructed using a common
set of specifications and standards.
In spite of these differences, the nuclear design philosophy, with appropriate
modification, is still suitable for use in a community setting. The next four sections, which detail
how several key nuclear concepts and tools are adapted, also discuss how the above differences
can be addressed.

5.2 UNDESIRED OUTCOMES

As described in Section 4.2.1, an undesired outcome is one that inhibits the ability of a system
(e.g., a nuclear power plant or a community) to maintain functionality, the consequences of
which can adversely impact the safety and welfare of the general public and surrounding

74
environment. As such, a primary focus of the regulatory framework that governs the design and
operation of such systems should involve minimizing the occurrence of these undesired
outcomes to the extent possible. The nuclear regulatory framework defines two undesired
outcomes, core damage and large release, and specifies performance targets that establish
numerical limits regarding the likelihood of their occurrence (i.e., less than one core damage
event in 10,000 years).
The first step in adapting the nuclear design philosophy to communities involves
identifying possible undesired outcomes for a community. Because of their diverse nature,
undesired outcomes will likely vary from community to community. This report focuses on one
particular undesired outcome: a significant and rapid outmigration of residents. This outcome is
particularly problematic because residents serve as both a community’s workforce and customer
base. If a large number of residents leave suddenly, the effects can ripple through the community
and its economy. Businesses lose both workers and customers. In response, some might close
permanently or decide to relocate, taking additional workers with them. As businesses and
residents disappear, tax revenue for local government shrinks, forcing layoffs and cuts to
essential community programs. This, in turn, might induce even more residents to leave.
It is important to reiterate that a community may choose whatever undesired outcome (or
outcomes) it feels is appropriate given its particular circumstances. This report focuses on a
significant and rapid outmigration of residents because this phenomenon has been observed (to
varying degrees) following several major natural disasters. After the 1995 Great Hanshin
earthquake, the population of Kobe, Japan, shrank by 2.5% and took 10 years to return to pre-
earthquake levels [Chang 1996, 2010; Horwich 2000]. A year following Hurricanes Katrina and
Rita, the population of New Orleans, Louisiana, was approximately 9% to 21% lower than pre-
hurricane levels, though in certain neighborhoods it was significantly lower [Hori et al. 2009;
Olshansky and Johnson 2010]. This report makes no attempt to define numerical boundaries for
what constitutes a significant outmigration of residents, as these boundaries will likely vary from
community to community. Consequently, individual communities need to determine thresholds
that are appropriate for them.

5.3 VITAL FUNCTIONS

As detailed in Section 4.2.3, three vital safety functions stand in the way of core damage and
large release during an accident at a nuclear power plant. In a similar fashion, this section
identifies four vital community functions that prevent a significant and rapid outmigration of
residents caused by an earthquake or other natural disaster. These four vital functions include
public services, housing, employment, and education (adapted from Poland et al. [2009], SERRI
and CARRI [2009], Twigg [2009], and Cutter et al. [2010]). It is important to note that these four
items refer to functions, not physical infrastructure; the frontline and support systems that
physically enable these vital functions are described in Section 5.4. If a community chooses an
undesired event other than an outmigration of residents, then the corresponding vital community
functions will likely need to be modified. The following four subsections describe each vital
function in more detail.

75
5.3.1 Public Services

The first vital community function involves providing essential public services to the residents of
a community, where public services refer to those services considered so important to a modern
society that they are typically provided, subsidized, or regulated directly by the government.
They commonly include police, fire and rescue, emergency medical care, non-emergency health
care, food, water, energy, transportation, communication, banking, sanitation, and other essential
community services (including building permit and inspection, planning, government finance
and taxation, social services, defense, mail delivery, and recordkeeping). This list focuses on
those services most crucial to disaster response and recovery because they play a fundamental
role in protecting the physical health, safety, and security of the general public both in day-to-
day operations and in the aftermath of a disaster.
In the immediate aftermath of a major accident or event, certain public services are
critical to emergency response operations. For example, emergency services like police, fire, and
medical are especially important in the first few hours and days, as they are responsible for
rescuing and treating injured residents, extinguishing fires, evacuating unsafe areas, and
maintaining general law and order. These services need to be at or near full capacity immediately
following an initiating event in order to respond successfully to the potentially significant
increase in demand caused by the accident or event.
Furthermore, certain public services play important roles in recovering from a major
accident or event. For example, inspection services provided by the local building department
will likely be required for a substantial number of a community’s buildings before they can be
reoccupied. In addition, mass transit systems need to be operational so residents can commute to
and from work. Other public services are less critical to the disaster response and recovery
process. For example, most residents can go without access to public libraries or museums for
several months. However, this should not downplay the importance of these services beyond the
response and recovery phases. Often, libraries, museums, and similar institutions play a vital role
in preserving a community’s history and culture. In some cases, they can also serve as major
tourist attractions and thus play an important role in the local economy.
In the long term, failure to provide any public service after an event is unacceptable,
representing a significant breakdown in one of the most fundamental functions of society.
However, in the immediate aftermath of a major disaster, the public services most important to
the response and recovery process must be given highest priority. Chapters 6 and 7 will discuss
this prioritization in more detail.

5.3.2 Housing

The second vital community function involves providing housing to the residents of a
community. Housing is particularly important because it helps keep in place both the workforce
and customer base of a community’s local economy. In a modern society, housing also includes
basic utilities typically available at a residence, including water, sanitation, electricity, natural
gas, and communications (e.g., telephone, Internet, and/or television). The building code

76
stipulates various provisions and requirements that establish minimum habitability requirements
for residential buildings [ICC 2006; BSSC 2004; BSSC 2009], many of which pertain to the
availability of essential utilities like water, power, and sanitation.
In the aftermath of a major accident or event, it is ideal for residents to shelter-in-place in
their own homes [Poland et al. 2009; SPUR 2012]. Shelter-in-place is a new performance level
for buildings developed by SPUR that proposes to relax certain habitability requirements during
emergency periods. Whereas the building code would prohibit residents from occupying
buildings without electricity or water, the SPUR shelter in place performance level would allow
residents to shelter in such buildings provided they are structurally safe and important utilities
are restored within a specified time period. Neighborhood support centers would provide shelter-
in-place residents with interim access to important utilities.
However, for many different reasons, not all residential buildings will be safe to occupy
after a major accident. If residents cannot shelter in place, the next best option is to move them to
emergency shelters in their original neighborhoods. If this is not possible, then the next best
option is to locate them somewhere else in the community, preferably in adjacent neighborhoods.
And if this is not possible and residents are forced to leave, then it is important for the
community to develop a plan for their return [Johnson and Eckroad 2001].

5.3.3 Employment

The third vital community function involves providing adequate employment opportunities to the
residents of a community. Private employment is important because it is the primary driving
force behind a local economy, servings two crucial functions (note that public employment is
captured in the public services basic safety function). First, employment is responsible for
producing, distributing, and selling many of the goods and services required by residents, local
governments, businesses, and organizations in the community. Second, employment provides
residents with a source of income to purchase the goods and services provided by the local
economy.
In the immediate aftermath of a major accident or event, certain types of employment are
more important than others. For example, engineers, contractors, and materials suppliers play
critical roles in repairing and replacing damaged infrastructure after a disaster. Banks, insurance
agencies, and other financial institutions finance these reconstruction projects. Therefore, these
(and other) types of employment need to be available quickly following an initiating event.
If employment is disrupted for an extended period of time following an initiating event,
some residents may struggle to afford even basic necessities like food, water, energy, and health
care. In response, some will leave the community in search of employment elsewhere. As their
customer base shrinks, businesses that sell goods and services to local residents may start to fail,
initiating a potentially adverse cycle of further outmigration and additional business failures.
Globalization complicates this process. The growing interconnectedness of local
economies means that certain types of businesses are less dependent on the residents of local
communities to serve as their workforce and/or customer base. For example, if an earthquake

77
disrupts operations at a local factory, the owner might decide to shift production, either
temporarily or permanently, to a location unaffected by the earthquake. Similarly, a local
business that exports its goods and services is less vulnerable to a collapse in local demand
caused by a major disaster than a business that sells only to local residents. However, these
export businesses are now vulnerable to disasters outside their local community. In spite of these
complications, employment remains a crucial function that communities must maintain in order
to prevent disruption to the local economy and significant outmigration of residents.

5.3.4 Education

The fourth vital community function involves providing residents with adequate access to
schools and education. A community’s education system is an important factor in attracting
potential residents. It also plays a crucial role in preventing residents from leaving the
community after a major incident. Without functional schools or day care facilities for their
children, some residents will be unable to return to work. If the disruption lasts long enough,
some will leave the community to enroll their children elsewhere. Education can also play an
important role in a community’s local economy. Colleges, universities, and trade schools train
and educate a community’s workforce, which attracts businesses and other employers eager to
leverage this highly skilled workforce.

5.4 FRONTLINE AND SUPPORT SYSTEMS

As discussed in Section 4.2.4, frontline systems in a nuclear power plant refer to those systems
that directly enable its vital safety functions. Sometimes, frontline systems can support multiple
functions. Frontline systems, in turn, are supported by support systems. Support systems are
especially important because they often support multiple frontline systems, as well as other
support systems, meaning that their failure can have widespread impact on the vital safety
functions in a nuclear power plant.
In general, in a community, frontline systems refer to buildings while support systems
refer to lifelines. Table 5.1 lists some of the frontline systems that support each of the four vital
functions in a typical community. Table 5.2 lists the support systems in a typical community
(adapted from PCCIP [1997], Rinaldi et al. [2001], ALA [2004], and Barkley [2009]). These lists
are by no means exhaustive, but rather give a general indication of the types of components in
each system.

78
Table 5.1 Frontline systems for each vital function in a community (adapted from
ASCE [2006] and Poland et al. [2009]).

Vital community
Frontline systems
function

 Hospitals, clinics, medical provider offices, and other health care

facilities
 Fire, police, rescue, and ambulance stations
Public services  Dispatch and emergency operations centers
 City hall and other administrative offices
 Military bases and other defense facilities
 Grocery stores and pharmacies
 Permanent residences
o Single-family housing (including mobile homes)
o Multi-family housing (apartments, condominiums, dormitories,
public housing)
o Institutional housing (nursing homes, assisted living facilities,
Housing correctional facilities, prisons, rehabilitation facilities)
 Short-term residences
o Transient housing (hotels, motels, boarding houses)
o Emergency housing (community centers, schools, convention
centers, arenas, other designated emergency shelters)
o Interim housing (FEMA trailers, tents)
 Commercial buildings (offices, retail shops, restaurants, banks,
Employment warehouses)
 Industrial buildings (factories, hazardous facilities)
 Preschools and day care facilities
 Primary and secondary schools (elementary, middle, and high
Education schools)
 Post-secondary schools (universities, colleges, trade schools,
institutes)

79
 Table 5.2 Support systems in a community (adapted from Rinaldi et al. [2001],
Barkley [2009], PCCIP [1997], and ALA [2004]).

Support system Components

Generation stations; transmission substations, towers, lines, and conduits;

Electric power
distribution substations, towers, lines, and conduits; control centers
Well facilities; processing plants; compressor stations; storage facilities;
Natural gas
pipelines; control centers
Well facilities; pumping stations; refineries; storage facilities; pipelines;
Oil
control centers
Solid fuels Mines; processing/preparation plants; storage facilities
Bridges; tunnels; roadways; traffic signs and signals; embankments;
Roads and
culverts; retaining walls; operation and control centers; maintenance
highways
facilities
Buses: stations; operation and control centers; fuel, dispatch, and
maintenance facilities
Mass transit
Light rail: tracks; bridges; tunnels; DC power substations; dispatch and
maintenance facilities
Tracks; bridges; tunnels; stations; signs and signals; fuel, dispatch, and
Railways
maintenance facilities
Runways; control towers; terminal buildings; hangars; fuel and maintenance
Airports
facilities
Waterfront structures (docks, piers, wharves, sea walls, breakwaters,
Ports and
jetties); cranes and cargo handling equipment; warehouses; fuel facilities;
waterways
locks and other engineered waterways
Well facilities; desalination plants; dams; reservoirs; canals; pipelines;
Water
pumping stations; treatment facilities; storage tanks
Waste water Pipelines; pumping/lift stations; treatment facilities
Transfer stations; materials recovery facilities; waste combustion facilities;
Solid waste
disposal sites
Central offices; data centers; network operations centers; transmitter
Telecommunications
stations; towers and poles; cables, lines, and conduits; satellite dishes

5.5 PERFORMANCE EVALUATION TOOLS

The following three subsections explain how the performance evaluation tools presented in
Section 4.4 are adapted for use in a community setting. Specifically, Section 5.5.1 describes
dependency matrices, Section 5.5.2 describes event trees, and Section 5.5.3 describes fault trees.

5.5.1 Dependency Matrices

As briefly mentioned in Section 4.4.1, dependency matrices provide a simple, convenient way to
capture the interactions among the various systems in a nuclear power plant. Dependency

80
matrices can be used in a similar fashion in a community. Figure 5.1 displays a simple
dependency matrix developed for the lifelines (i.e., support systems) depicted in Figure 3.1. The
matrix is constructed one row at a time, with an “x” mark indicating a dependency between the
two systems under consideration. For example, as shown in Figure 3.1, the transportation lifeline
depends on oil (for fuels and lubricants), electric power (for power to signals and switches), and
telecommunications [for supervisory control and data acquisition (SCADA) and
communication]. Therefore, the transportation row of the dependency matrix in Figure 5.1
contains three “x” marks corresponding to these three dependencies.

Figure 5.1 Dependency matrix corresponding to the support systems portrayed in

Figure 3.1.

A dependency matrix not only captures the direct dependencies between each of the
systems included in the analysis, but also provides a relative indication of which systems are
most important and which are most vulnerable. By summing the number of “x” marks in each
column, a general measure of the importance of each system can be ascertained. In Figure 5.1,
both electric power and telecommunications systems support five lifelines, meaning a service
disruption to either of these systems can have widespread impact on other lifelines. In contrast, a
disruption to the natural gas lifeline would only affect two other systems (electric power and
telecom). Similarly, by summing the number of “x” marks in each row, a general indication of
the vulnerability of each system can be obtained. In Figure 5.1, electric power, natural gas, and
telecommunications systems each depend on five other lifelines, meaning they are more
vulnerable to disruptions caused by disruptions to other systems. In contrast, the water support

81
system only relies on two other systems (electric power and telecommunications), meaning it is
less susceptible to disruptions caused by external systems.
As Figure 5.1 demonstrates, dependency matrices make for useful planning tools. They
can be adapted to fit other situations: for example, analyzing the interactions between frontline
systems and support systems or between individual components within a particular system. In
spite of these many potential applications, dependency matrices have a few important limitations.
First, they do not capture the type of dependence (e.g., physical, cyber, geographic, or logical)
between the systems (see Table 2.6). Second, they do not capture the strength of dependence
between the systems. For these reasons, when performing in-depth analyses of community
systems and their interactions, dependency matrices should only be used during the preliminary
stages to identify general interactions.

5.5.2 Event Trees

The framework proposed in Chapter 6 makes extensive use of event trees. Traditionally, event
trees have been used to analyze the response of a system or component (i.e., the analysis side of
the equation); however, in this report they will be used to link broad performance goals for a
community to specific performance targets for an individual component or subsystem (i.e., the
design side of the equation). Regardless, the event trees described in Chapter 6 function in the
same fashion as described in Section 4.4.2, with one key difference in how they are constructed.
In a nuclear power plant, the response of frontline and support systems is binary: each system
either fails or succeeds. In a community, the response of frontline and support systems (i.e.,
buildings and lifelines) is not as binary. Due to their spatial distribution, some systems can fail
partially. For example, damage to a community’s electric power system caused by an earthquake
may disrupt service to only a small number of neighborhoods. Event trees can be easily modified
to account for these partial failures, as will be demonstrated in Chapter 6.

5.5.3 Fault Trees

As detailed in Section 4.4.3, fault trees can be used to evaluate the failure modes of specific
subsystems and components within a nuclear power plant. They can be used in a similar fashion
to analyze particular subsystems and components within a community. However, this report does
not demonstrate this extension for the following reason: unlike event trees, fault trees cannot be
adapted for use in a generic fashion. For example, a fault tree cannot be developed for a generic
electric power system because the structure of the fault tree requires detailed information about
the system’s components and configuration, which can be obtained only after selecting a
particular system to study (e.g., the electric power grid in San Francisco). At this stage of
development, the engineering framework presented in this chapter and the next is kept
purposefully generic in order to maximize its applicability and also demonstrate its benefits more
readily.

82
6 Community Event Trees

This chapter presents and describes a set of event trees that forms the backbone of the proposed
engineering framework. To this end, Section 6.1 outlines the conditions under which the event
trees should be used, while Section 6.2 discusses their general structure and organization. Section
6.3 introduces the event trees and details the rationale used to develop each one. Section 6.4
synthesizes and combines the event trees from the previous section into a single tree.
The event trees presented in this chapter can be used for many different purposes. For
example, they can be used as an analysis tool to quantify the vulnerability of a community and its
built environment to natural hazards. In addition, they can be used as a design tool to link
community-level resilience goals to specific performance objectives for individual components
and subsystems (e.g., buildings and lifelines) within the built environment. This application is
given particular attention in Chapter 7, which presents a conceptual example demonstrating how
to develop seismic performance targets for a new residential building from a community-level
performance objective.

6.1 APPLICABILITY

A primary objective in developing the event trees presented in this chapter is to maximize their
applicability. In general, the event trees can be used to evaluate a wide range of communities and
hazards; however, several important limitations must first be acknowledged.
First, the event trees have been developed with earthquakes in mind. Earthquakes are
unique for several reasons. Unlike most other natural hazards, they offer little or no advance
warning. Hurricanes, on the other hand, can be forecast several days before landfall, allowing
residents sufficient time to perform last-minute mitigation activities and evacuate to safer land.
Even tornadoes give a few minutes of warning in most cases. In addition to their
unpredictability, earthquakes can have extremely widespread impact. Tornadoes, in contrast,
affect a much narrower geographic area. Furthermore, large-magnitude earthquakes are typically
followed by sizable aftershocks, some of which can be as devastating as the initial shock,
especially since the built environment is already in a weakened state. As a result of these unique
factors, the event trees developed in this chapter may require some modification in order to apply
to other natural hazards.

83
 Second, the event trees have been developed for hazards that have the potential to affect
an entire community. As a result, the event trees should not be used to evaluate the impact of
small-scale hazards, like an isolated windstorm that damages an office building or a landslide
that destroys two or three houses. Instead, the event trees should be used to evaluate hazards with
more widespread impact, like hurricanes, earthquakes, and large tornadoes, wildfires, and floods.
Third, the event trees have been developed for application at the level of the community,
which can range in size from large towns to major metropolises. For larger communities, the
event trees presented in this chapter can also be used at the neighborhood level. However, at
further levels of refinement (e.g., city block or parcel level), the event trees presented in this
chapter begin to lose applicability. Additional refinement is possible, but it will likely require
that a new set of event trees be developed, which is beyond the scope of this report. In the
opposite direction, the event trees can be used to evaluate clusters of nearby communities or
geographic regions (e.g., the San Francisco Bay Area). However, the event trees require
modification in order to be applied at the state or national level. Again, this extrapolation is
beyond the scope of this report.
And fourth, the event trees have been developed for mitigation and planning purposes
only. More specifically, they have been developed to quantify the vulnerability of communities
to hazards and, consequently, to help communities make better-informed policy decisions to
address this vulnerability before a disaster strikes, through, for example, modifications to locally-
adopted building codes, development of retrofit programs, and/or improved land use regulations.
The event trees should not be used to track or assess the real-time response of a community
during an actual disaster or accident.

6.2 GENERAL STRUCTURE

Event trees provide a structured framework for enumerating and, subsequently, evaluating the
numerous combinations of events that can result in undesired outcomes for a system. In a nuclear
power plant, they help identify specific combinations of events, or accident sequences, that can
produce core damage and/or large release. The event trees presented in this chapter are
developed for a similar purpose: to identify the combinations of events that can result in a
significant and rapid outmigration of residents from a community. As described in Section 5.3,
four vital functions stand in the way of this undesired outcome: public services, housing,
employment, and education. Therefore, this chapter develops an individual event tree for each of
the four vital community functions.
Figure 6.1 illustrates the general structure of an individual event tree. In essence, the tree
describes the range of possible outcomes for a single vital function following an initiating event,
which in the context of communities can range from natural phenomena like earthquakes,
hurricanes, and floods to human-made events like terrorist attacks, economic downturns, and
random system or component failures. This report, however, focuses primarily on earthquakes.
The event tree in Figure 6.1 comprises three top events, each defined in terms of a different limit

84
on a tracking variable. As a result of this construction, these three top events enable the tree to
capture partial failure of the vital community function.
The tracking variable in Figure 6.1 refers to a quantifiable parameter that describes the
status of the vital function after the initiating event. This variable must be chosen carefully so
that it adequately captures and summarizes the overall status of the vital community function.
Using a single variable to track a complex, multi-faceted process carries with it inherent
limitations; however, because the primary intent of the event tree is to serve as a summary of the
impact of damage on the chosen vital function, a single variable, when chosen carefully, is
appropriate.
The three top events in Figure 6.1 produce four distinct outcomes. Outcome #1 results if
the first top event fails to occur (i.e., the parameter fails to satisfy Limit #1). Outcome #2 results
if the first top event occurs but the second one does not (i.e., the parameter satisfies Limit #1 but
fails to satisfy Limit #2), and so on. In general, Outcome #1 represents a worst-case scenario, one
that has catastrophic impact on the vital community function, hence the red color. Outcome #4,
on the other hand, represents a best-case scenario, one that has minor impact on the availability
of the vital function, hence the green color. Outcomes #2 and #3 fall between these two
extremes. Additional branches can be added to the event tree if more than four outcomes are
desired. However, if chosen correctly, three top events should adequately encompass the range of
possible outcomes for a vital community function, while at the same time limiting the
complexity of the tree.
Figure 6.2, which portrays the event tree for the housing vital community function, helps
make these concepts and ideas more concrete by providing a specific tracking variable and top
events. The tracking variable for the event tree is the percentage of residents displaced from their
homes. The first top event in the tree asks whether less than 20% of residents have been
displaced from their homes. Figure 6.2, which portrays the event tree for the housing vital
community function, helps make these concepts and ideas more concrete by providing a specific
tracking variable and top events. The tracking variable for the event tree is the percentage of
residents displaced from their homes. The first top event in the tree asks whether less than 20%
of residents have been displaced from their homes. If the answer to this question is no (i.e., more
than 20% of residents are displaced), the resulting outcome has catastrophic impact on the
housing vital community function, potentially resulting in a significant and rapid outmigration of
residents from the community, especially if other vital community functions suffer similar levels
of disruption. If, on the other hand, the answer to this question is no (i.e., less than 20% of
residents have been displaced), the second top event is called upon, which asks whether less than
10% of residents have been displaced. If yes, the resulting outcome has significant impact on the
housing vital function; if no, the third top event is called upon, which asks whether less than 2%
of residents have been displaced. Section 6.3.2 discusses this event tree in additional detail.

85
 Figure 6.1 Example of an event tree for an individual vital community function.

Figure 6.2 Housing event tree.

6.3 EVENT TREES

In general, when an initiating event such as an earthquake occurs, it causes damage to a

community’s built environment. More specifically, the initiating event causes physical damage
to components (e.g., structures and hardware) within a community’s frontline and support
systems. The extent and scale of damage depends on many factors, including the characteristics
of the initiating event (e.g., magnitude and location of the earthquake) and the vulnerability of
the built environment. This damage has several immediate consequences. First, it can cause
serious physical and psychological harm to residents in the community. For example, debris may
fall on people, or buildings may collapse on their inhabitants. And second, damage to structures
and equipment can cause frontline and support systems in the community to partially or
completely fail. Through a complex web of interdependencies and interactions, these direct
failures can cause additional systems to fail. Ultimately, if the initiating event causes enough
direct and indirect system failures, it can disrupt one or more of the vital functions of a
community.
The event trees presented and described in the following subsections aim to quantify and
summarize the extent to which an initiating event impacts the vital functions of a community.
Four event trees are described, one for each of the four vital community functions (public
services, housing, employment, and education). The following subsections describe each tree in

86
more detail, including discussion of the rationale for selecting both the tracking variable and top
events for each tree.

6.3.1 Public Services

The public services event tree, depicted in Figure 6.3, captures the effect of damage to frontline
and support systems on the availability of the public services vital function. Specifically, the
event tree tracks the percentage of capacity disrupted by the initiating event, where capacity is
benchmarked to pre-event service levels. This measure is an aggregation of the disruption to
individual public services. Only the most essential public services for response and recovery
should be included in this aggregation. Recall that public services refer to those services
considered so important to a modern society that they are typically provided, subsidized, or
regulated directly by the government. Section 5.3.1 enumerates a baseline set of public services,
including police, fire and rescue, emergency medical care, non-emergency health care, food,
water, energy, sanitation, transportation, communication, banking, and other essential
community services (including building permit and inspection, planning, government finance
and taxation, social services, defense, mail delivery, and recordkeeping). Some of these services
are provided by the private sector but are included in this formulation due to their extremely
important nature.
Because of the diversity of services included in the public services vital community
function, the tracking variable needs to be an aggregation. Of course, this aggregation could be
avoided by developing a separate event tree for each public service, but the complexity involved
with doing so is not appropriate at this stage of development and would likely overwhelm the
analyst or decision maker. Furthermore, since the framework in this chapter has been developed
for application at the community level, there is benefit to describing the public services vital
function with a single measure.
In addition, the selected tracking variable is appropriate because it can capture
disruptions to the capacity of public services arising from several different sources. First,
frontline systems (i.e., buildings) may suffer structural damage to the extent that they are not safe
to occupy. For example, a hospital that sustains significant permanent lateral displacement after
an earthquake will be rendered unusable due to the collapse risk it poses. Second, frontline
systems may lose access to important utilities due either to nonstructural damage to the buildings
themselves or to damage to external support systems (i.e., lifelines). For example, a hospital’s
water supply can be disrupted if pipes and conduits throughout the hospital break, or if damage
to a community’s water infrastructure causes a service disruption to the hospital. And third,
support systems may suffer physical damage. For example, damage to a community’s water
infrastructure can disrupt water service to certain neighborhoods, which can inhibit the ability of
firefighters to battle blazes. The chosen tracking variable is robust enough to capture all of these
effects.
At the same time, it is important to recognize a limitation associated with the choice of
tracking variable presented in Figure 6.3. Because it aggregates the performance of many
different services, the tracking variable does not capture the source or nature of the disruption.

87
For example, consider a situation in which 10% of capacity is disrupted. This disruption could be
the result of a 10% disruption to each individual service, or it could be the result of a 100%
disruption to one particular service (assuming there are 10 individual services and each is
weighted equally in the aggregation). As this simple example demonstrates, the tracking variable
obscures which services have been disrupted and the extent to which each has been impacted.
Again, this limitation can be overcome by simply creating separate event trees to track each
public service.
The three top events displayed in Figure 6.3 establish four possible outcomes for the
public services vital function. The top outcome, less than 5% capacity disrupted, represents a
situation that has limited impact on the community’s ability to provide public services to its
residents. The bottom outcome, more than 50% of capacity disrupted, on the other hand,
represents a scenario that can have catastrophic impact on public services. It is important to note,
however, that the numerical limits corresponding to each top event in Figure 6.3 may require
refinement in order to properly distinguish the range of possible outcomes. Currently,
insufficient data exists to verify the appropriateness of these numerical targets (i.e., whether a
disruption of more than 50% of capacity actually represents a catastrophe). In addition, these
targets may need to be adjusted on a community-by-community basis. However, at this stage of
development, the structure of the public services event tree is more important than the numerical
targets associated with its top events, as the numbers can be refined further in the future.

Figure 6.3 Public services event tree.

6.3.2 Housing

The housing event tree, depicted in Figure 6.2, captures the effect of damage to frontline and
support systems on the availability of the housing vital community function. Specifically, the
event tree tracks the percentage of residents displaced by the initiating event. This tracking
variable captures the two main reasons residents can be displaced. First, frontline systems (i.e.,
residential buildings) may suffer structural damage to the extent that they are not safe to occupy
following an initiating event. However, even if a residence is safe to occupy, some residents
might still choose to leave for personal or psychological reasons, while others might be forced
out by landlords who want to make repairs before allowing residents to reoccupy. Second,
frontline systems may lose access to important utilities, due either to nonstructural damage to the
buildings themselves or to damage to external support systems (i.e., lifelines). This second

88
reason is especially applicable to multi-family housing (e.g., apartments and condominiums) and
institutional housing. After an initiating event, landlords may not want tenants occupying
apartments that lack power or water for liability reasons, while certain types of institutional
housing (e.g., nursing homes) may require utilities in order to remain operational. The chosen
tracking variable is robust enough to capture these effects.
In addition, the percentage of residents displaced is an appropriate choice of tracking
variable because in the days, weeks, and months following a major disaster, a primary concern of
community leaders and decision makers involves sheltering displaced residents. Unlike tracking
variables that focus on the physical damage to a community’s housing stock, the percentage of
displaced residents provides a direct measure of the affected population, which is important for
emergency response and recovery planning (e.g., determining the number of public shelters
required, preparing emergency food supplies, etc.).
Despite its appropriateness, it is important to recognize the limitations of the chosen
tracking variable. First, it does not capture which residents are displaced. Vulnerable populations
within a community are more likely to seek shelter at public facilities than affluent, well-
connected populations [Yelvington 1997; Wisner et al. 2003]. Therefore, it is helpful for
communities to know which types of residents are displaced so that they can plan accordingly.
Second, the chosen tracking variable does not capture the reason each resident is displaced. As
detailed in the first paragraph of this subsection, residents can be displaced for several reasons,
the implications of which have varying impact on housing recovery. For example, it typically
takes more time to repair buildings with significant structural damage than those with minor
nonstructural issues [Comerio 1998; SPUR 2012; ATC 2010]. Knowing the nature of the
physical damage to the housing stock helps a community determine (approximately) how long
residents will be displaced. Third, the chosen tracking variable does not capture which types of
housing are damaged. Again, the implications of this affect recovery time. For example, multi-
family housing typically takes longer to restore than single-family housing [Comerio 1998; ATC
2010; SPUR 2012]. However, because the primary intention of event tree is to summarize the
overall impact of damage on the housing vital function, these limitations are secondary in nature.
The three top events displayed in Figure 6.2 establish four possible outcomes for the
public services vital function. The top outcome, less than 2% of residents displaced, represents a
situation that has limited impact on the community’s ability to provide housing to its residents.
The bottom outcome, where more than 20% of residents are displaced, represents a scenario that
has catastrophic impact on housing. These numerical targets are based on data from SPUR
[2012]. However, they may need to be adjusted on a community-by-community basis. In general,
the first top event (20%) corresponds to a percentage of the population that substantially exceeds
the community’s emergency shelter capacity; the second top event (10%) corresponds to the
community’s emergency shelter capacity; and the third top event (2%) corresponds to the
community’s vacancy rate, which refers to the percentage of rental units that are unoccupied. In
the aftermath of an earthquake, these vacant units can be made available to displaced residents,
provided they are not rendered uninhabitable from damage caused by the earthquake.

89
6.3.3 Employment

The employment event tree, depicted in Figure 6.4, captures the effect of damage to frontline and
support systems on the availability of the employment vital function. Specifically, the event tree
tracks the percentage of businesses disrupted by the initiating event. The tracking variable is
appropriate because it can capture the many different reasons that businesses can be disrupted
following an earthquake or other natural hazard. Table 6.1 summarizes several of these reasons.
As the table demonstrates, there exist a wide variety of reasons a business can be disrupted,
reflecting the high degree of interconnectedness within a community’s local economy. Again, the
chosen tracking variable is robust enough to capture these numerous sources of disruption.

Figure 6.4 Employment event tree.

In addition, the percentage of businesses disrupted is an appropriate choice of tracking

variable because businesses are drivers of the local economy. In the aftermath of a major
disaster, it is usually business owners, not workers, who make decisions about closing or
relocating, either temporarily or permanently. Therefore, the tracking variable should focus on
the percentage of businesses disrupted, not the percentage of workers unemployed. Furthermore,
a tracking variable like the unemployment rate would be unable to distinguish between those
unemployed before the disaster and those unemployed because of the disaster.
Despite its appropriateness, it is important to note several limitations associated with the
chosen tracking variable. First, it does not capture which businesses are disrupted. For example,
consider a scenario in which 10% of businesses are disrupted after an earthquake. If the disrupted
businesses include several large employers, it will impact more workers and likely have graver
consequences for the community than if the disrupted businesses are small employers. In
addition, if the disrupted businesses are concentrated in a single employment sector, the
consequences will likely impact a community more profoundly than if the disrupted businesses
are spread across all sectors. Furthermore, if the business disruption affects a critical industry or
sector, the consequences will likely be more profound than if a less critical sector is disrupted.
Second, the chosen tracking variable does not capture the nature of the business disruption. Table
6.1 lists several causes of business disruption, with each one having varying impact on the speed
of recovery (if recovery is even possible). While important, these limitations are secondary in

90
nature, as the primary purpose of the employment event tree is to broadly summarize the impact
of damage on businesses.
The three top events displayed in Figure 6.4 establish four possible outcomes for the
employment vital community function. The top outcome, less than 5% of businesses disrupted,
represents a situation that has limited impact on the community’s ability to provide employment
opportunities to its residents. The bottom outcome, where more than 50% of businesses are
disrupted, represents a scenario that can have catastrophic impact on employment. It is important
to note, however, that the numerical limits corresponding to each top event in Figure 6.4 may
require refinement in order to properly distinguish the range of possible outcomes. Currently,
insufficient data exists to verify the appropriateness of these targets (i.e., whether a disruption of
more than 50% of businesses actually represents a catastrophe). In addition, these targets may
need to be adjusted on a community-by-community basis. Again, at this stage of development,
the structure of the employment event tree is more important than the numerical targets
associated with its top events, as the numbers can be refined in future iterations.

Table 6.1 Reasons that businesses can be disrupted after an earthquake.

Description Example

Frontline systems (i.e.,

buildings) may suffer structural A restaurant suspends operations because its building
damage to the extent that they suffers extensive damage (i.e., receives a red tag)
are not safe to occupy
Frontline systems may lose
access to important utilities, due
either to nonstructural damage Loss of electricity from the power grid suspends
to the buildings themselves or to operations at an otherwise undamaged factory
damage to external support
systems (i.e., lifelines)
Damage to roads and highways disrupts the
Support systems may suffer transportation of goods and services, causing supply
physical damage chains to shut down and preventing workers from
commuting to work
A factory suspends activity because the building of one
Other businesses may be
of its suppliers has suffered damage and, as a result,
disrupted
cannot supply a key input
A coffee shop near a large apartment complex closes
Customers may relocate because the apartment complex suffers extensive
damage, forcing residents to seek shelter elsewhere
A factory shuts down because its workers, displaced
Workers may relocate from their homes, relocate at great distances from the
factory

91
6.3.4 Education

The education event tree, depicted in Figure 6.5, captures the effect of damage to frontline and
support systems on the availability of the education vital function. Specifically, the event tree
tracks the percentage of students displaced by the initiating event. This tracking variable is
appropriate because it can capture the many different reasons students can be displaced after an
initiating event such as an earthquake. First, frontline systems (i.e., schools) may suffer structural
damage to the extent that they are not safe for students to occupy. Second, frontline systems may
lose access to important utilities, due either to nonstructural damage to the schools themselves or
to damage to external support systems (i.e., lifelines). Additionally, support systems may suffer
damage. For example, roads and highways may suffer damage that prevents students from
getting to school. Furthermore, students may be displaced because their schools also serve as
public shelters during a disaster or emergency. Lastly, students may be displaced if their homes
are damaged and they are forced to move to locations that lack adequate schools. The chosen
tracking variable is sufficiently robust to capture these effects. In contrast, variables that focus on
physical damage to schools are unable to capture student displacement caused by external factors
(e.g., damage to roads and highways, schools doubling as public shelters, etc.). Furthermore,
these alternate tracking variables obscure important factors like the size of the schools damaged.
Despite its appropriateness, it is important to recognize two limitations associated with
the chosen tracking variable. First, it does not capture which students have been displaced. For
example, consider a scenario in which 15% of students have been displaced after an earthquake.
If the majority of displaced students come from vulnerable neighborhoods, it will have different
consequences than if most displaced students come from affluent neighborhoods. Similarly, if
the majority of displaced students are in grade school, it will have different consequences than if
displaced students are in high school or college, in part because younger children need constant
care if not in school, meaning their parents will likely be unable to return to work until their
children are back in school. Second, the chosen tracking variable does not capture why students
have been displaced. As described above, students can be displaced for many reasons, each one
having varying impact on how quickly students can get back in the classroom. However, because
the primary intention of event tree is to summarize the overall impact of damage on the
education vital function, these limitations are secondary in nature.
The three top events displayed in Figure 6.5 establish four possible outcomes for the
education vital function. The top outcome (less than 5% of students displaced) represents a
situation that has limited impact on the community’s ability to provide educational opportunities
to its residents. The bottom outcome (more than 20% of students displaced), on the other hand,
represents a scenario that can have catastrophic impact on education. It is important to note,
however, that the numerical limits corresponding to each top event in Figure 6.5 may require
refinement in order to properly distinguish the range of possible outcomes. Currently,
insufficient data exists to verify the appropriateness of these targets (i.e., whether displacing
more than 20% of students actually represents a catastrophe). In addition, these targets may need
to be adjusted on a community-by-community basis. Again, at this stage of development, the

92
structure of the education event tree is more important than the numerical targets associated with
its top events, as the numbers can be refined in future iterations.

Figure 6.5 Education event tree.

6.4 SYNTHESIS

In order to obtain a complete picture of the impact of damage to frontline and support systems on
the community, the four event trees need to be combined into a single tree. This combined event
tree, which is depicted in Figure 6.6, summarizes the numerous outcomes possible in a
community following an initiating event such as an earthquake. There are 44 = 256 possible
outcomes. The topmost outcome in Figure 6.6 results from the following sequence of events: less
than 5% of public services capacity disrupted, less than 2% of residents displaced, less than 5%
of businesses disrupted, and less than 5% of students displaced. Because each vital community
function suffers only minor disruption, this outcome does not trigger a significant and rapid
outmigration of residents from the community. In contrast, the bottommost outcome in Figure
6.6 will likely result in a significant and rapid outmigration of residents because each vital
community function suffers catastrophic disruption: more than 50% of public services capacity
disrupted, more than 20% of residents displaced, more than 50% of businesses disrupted, and
more than 20% of students displaced.
The combined event tree in Figure 6.6 provides a structured methodology for identifying
combinations of events that can result in a significant and rapid outmigration of residents (or any
other undesired outcome of interest). Identification of these sequences is a crucial step in the
proposed engineering framework, as it links damage to frontline and support systems to broader
outcomes in a community. Because these sequences will likely vary from community to
community, this report does not attempt to definitively identify them here. In general, sequences
that trigger a significant and rapid outmigration of residents can be identified using data from
previous disasters or, in its absence, the expert judgment of those with extensive knowledge of
communities and the built environment (e.g., engineers, planners, policymakers, and
economists).
The combined event tree in Figure 6.6 can be used for both design and analysis
applications. On the analysis side, the combined event tree can be used to synthesize the results
of separate analyses of frontline and support systems in order to determine the overall impact on

93
the community and its vital functions. For example, an analysis of the housing stock using
HAZUS, FEMA’s methodology for estimating potential losses from disasters, can be used in
conjunction with the housing event tree (see Figure 6.2) to understand the contribution of
potential housing losses to the likelihood of outmigration following an earthquake scenario. This
report, however, does not demonstrate an application of this nature. Instead it focuses on design
applications. Chapter 7 presents a conceptual example that uses the combined event tree to
develop seismic design targets for a new residential building from a community-level
performance goal. Ultimately, the combined event tree can serve as a mechanism for linking
community-level resilience goals to specific design targets contained in building codes and other
engineering standards for buildings and lifeline systems.

94
Figure 6.6 Single event tree obtained from combining the public services, housing,
employment, and education event trees.

95
96
7 Example Applications

This chapter represents the culmination of the work in this report, presenting two conceptual
examples that highlight potential applications of the framework described in Chapters 5 and 6.
The first example, presented in Section 7.1, describes a methodology for establishing consistent
performance targets for individual components from a community-level target using the event
trees developed in the previous chapter. The second example, presented in Section 7.2, outlines a
methodology for estimating the capacity of public services disrupted using dependency matrices.

7.1 EXAMPLE: ESTABLISHING CONSISTENT PERFORMANCE OBJECTIVES

The first example demonstrates how the event trees presented in the previous chapter can be used
to develop consistent performance targets for individual components from a community-level
target. In particular, the example shows how to develop seismic performance targets for a new
residential building from a community-level performance objective using the combined event
tree in Figure 6.6. Ultimately, this example outlines a procedure that can be used to modify the
implicit performance objectives contained in building codes, or even to lay the conceptual
foundations of a “community performance code,” a proposed document that specifies explicit
performance targets for a community and the numerous components and subsystems of its built
environment (see Chapter 8 for further discussion).
The following subsections outline a basic methodology for creating a consistent hierarchy
of performance objectives for a community. The first step, described in Section 7.1.1, establishes
performance targets for the entire community. The second step, described in Section 7.1.2, uses
these community-level targets together with the event trees presented in the previous chapter to
determine performance objectives for each of the four vital community functions. The third step,
outlined in Section 7.1.3, uses the objectives for each vital community functions to establish
performance targets for frontline and support systems in the built environment. And the fourth
step, discussed in Section 7.1.4, establishes targets for individual components within each
frontline and support system. Several important simplifications and assumptions need to be made
during each step; each one will be identified and detailed in the following subsections when
appropriate.

97
7.1.1 Performance Targets for a Community

The first step in the methodology involves establishing a performance target (or targets) for the
entire community. These targets can take many different forms. For a nuclear power plant,
system-level performance targets take the form of a mean annual frequency of core damage and
large release: less than 1x10-4 and 1x10-5 per reactor year, respectively (see Section 4.3.2 for
additional detail). System-level performance targets for a community can mirror those for a
nuclear power plant. In other words, community-level performance targets could take the form of
a mean annual frequency of occurrence of an undesired outcome, which in the context of this
report is a significant and rapid outmigration of residents.
The target chosen for the mean annual frequency of significant outmigration directly
impacts the resilience of a community. In general, a community that selects a more stringent
performance target (e.g., 1×10-4 instead of 1×10-3) is less vulnerable to events that can cause a
significant and rapid outmigration of residents, including earthquakes, hurricanes, and floods.
Consequently, the mean annual frequency of significant outmigration can be considered a proxy
for community resilience, with smaller targets equating to improved levels of resilience.
For the purposes of this example, we will assign a mean annual frequency of significant
outmigration equal to 1×10-4 or less per year, which translates to one undesired outcome
occurring every 10,000 years, on average. Note that this value represents a design target. As
such, it does not reflect the actual level of performance achieved by the community as it
currently exists. Instead, it represents the level of performance the community ultimately desires.
If an evaluation reveals that the community and its existing built environment do not satisfy their
specified performance targets, the community needs to make investments to retrofit or replace its
infrastructure in order to improve performance.
It is essential that community stakeholders (e.g., politicians, engineers, and planners)
establish these community-level performance targets in a public process so that they accurately
reflect the level of risk acceptable to society. There are many considerations to weigh during this
process, including the expected lifetime of individual subsystems and components within the
built environment (e.g., buildings, bridges, power grid, etc.), the level of risk aversion of
stakeholders, and sustainability/environmental issues. The mean annual frequency target chosen
for this example, 1×10-4, likely resides at the conservative end of the spectrum of possible
targets, as it is the same as the target for averting core damage in a nuclear power plant.
In order to verify that a community satisfies the target chosen for this example, its
performance needs to be evaluated across the entire range of possible hazard types and intensities
(e.g., small-magnitude and large-magnitude earthquakes, frequent and rare floods, etc.). An
explicit and comprehensive analysis of this scope and scale is not practical. Furthermore,
engineers and planners traditionally use specific hazard scenarios when designing individual
buildings or when modeling damage and loss at the community level. Therefore, to be consistent,
the annualized performance target needs to be converted to targets corresponding to specific
scenarios for each type of hazard that can impact a community. For each type of hazard, an
appropriate number of scenarios should be selected. This number should enable an accurate

98
picture of response to emerge without burdening designers and analysts with unnecessary work.
Furthermore, hazard scenarios should be well separated and effectively encompass the range of
intensities with most potential to impact a community. For example, for most hazards, small
magnitude events do not need to be considered.
In this example, we will make the following simplification: earthquakes are the only type
of hazard that can affect the hypothetical community. Furthermore, to be consistent with the
IBC, we will select only two earthquake scenarios. The first scenario, referred to as the design
basis earthquake (DBE), has a 475-year mean recurrence interval, corresponding to an annual
frequency of exceedence of 2×10-3, or 10% probability of being exceeded in 50 years [BSSC
2004]. This scenario represents an earthquake that can reasonably be expected to occur during
the lifetime of a building, which is typically assumed to be 50 years. The second scenario,
referred to as the maximum considered earthquake (MCE), has a 2,450-year mean recurrence
interval, corresponding to an annual frequency of exceedence of 4×10-4, or 2% probability of
being exceeded in 50 years [BSSC 2004]. This scenario represents the “worst-case” event
expected during the lifetime of a structure. Note, however, that the MCE is not truly a worst-case
event as larger-magnitude earthquakes are still possible, though they are very unlikely.
Once specific hazard scenarios have been selected, conditional probabilities of significant
outmigration need to be chosen for each one. These conditional probabilities should be assigned
in such a way that the original mean annual frequency of significant outmigration (1×10-4) is
satisfied. In this example, we will select the following conditional probabilities of significant
outmigration: 1% for the 475-year earthquake and 10% for the 2450-year earthquake, which are
similar in structure to the performance objectives selected for nuclear power plants in ASCE 43
(Seismic Design Criteria for Structures, Systems, and Components in Nuclear Facilities) [ASCE
2005]. If we assume these two hazard scenarios (DBE and MCE) are independent, the mean
annual frequency associated with these choices of conditional probabilities is:

Pf  0.01 2 103  0.10  4 104  6 105  Pf,target  1104 (7.1)

As Equation (7.1) demonstrates, the resulting mean annual frequency of significant

outmigration for the community (6×10-5) is less than the original target (1×10-4) by a factor of
1.67. Because only two scenarios have been selected, this level of conservatism is warranted.
Note that the ratio of Pf,target to Pf can be interpreted as the confidence with which the original
community-level performance target is satisfied. Consequently, values greater than one indicate
higher confidence that the original community-level performance target is satisfied. Again, it is
important to note that other combinations of numbers can be selected if desired, provided that the
resulting Pf is less than Pf,target. Community stakeholders and policymakers can adjust this ratio as
they see fit.
In summary, in this example, we have established two community-level performance
objectives:
 1% probability of significant outmigration in the 475-year earthquake
 10% probability of significant outmigration in the 2450-year earthquake

99
 These two community-level performance targets provide the foundation for the hierarchy
of performance objectives developed in the next three subsections.

7.1.2 Performance Targets for the Vital Community Functions

The second step in the methodology involves establishing performance targets for each of the
four vital community functions (public services, housing, employment, and education). To this
end, the community-level performance objectives from the previous step (see Section 7.1.1) need
to be translated into targets for each vital community function. The event trees presented in
Chapter 6 and reproduced in Figure 7.1 provide the means for this translation. Specifically, the
event trees can be used to identify the sequences of events that produce a significant and rapid
outmigration of residents from a community. Once these sequences have been identified,
probabilities can be assigned to each branch of the tree in such a way that the original
community-level performance objectives are satisfied. These branch probabilities can then be
used to establish performance objectives for each of the four vital community functions.
Continuing the example from the previous subsection, the two community-level
performance objectives (1% probability of significant outmigration in the 475-year earthquake
and 10% probability of significant outmigration in the 2450-year earthquake) need to be
translated into targets for each vital function. In order to do so, the event trees in Figure 7.1 need
to be combined into a single tree so that sequences that result in a significant and rapid
outmigration of residents can be identified. Figure 7.2 shows this combined event tree. It
comprises 44 = 256 possible outcomes or accident sequences. Sequences that trigger a significant
outmigration of residents can be identified using several different techniques, including expert
judgment or data from previous disasters. In this example, we will identify these sequences using
the following simple rule: a significant outmigration results either when at least two of the
individual event trees in Figure 7.1 are “in the red” or when at least three are “in the orange.” An
individual event tree is “in the red” when its bottom outcome has occurred and “in the orange”
when its second-from-the-bottom outcome has occurred. As a result of this rule, there are 104
sequences that trigger significant outmigration. In Figure 7.2, a red box at the end of a branch
indicates such a sequence.

100
Figure 7.1 Individual event trees for the vital community functions.

101
Figure 7.2 Combined event tree (sequences that trigger significant outmigration
identified by red boxes at the end of the branches).

102
 In order to compute the conditional probability of significant outmigration for a
community in a specific hazard scenario (e.g., a DBE or MCE), the probability of each of the
104 significant outmigration sequences needs to be computed and then summed. Consider, for
example, the bottom outcome in the combined event tree in Figure 7.2. This outcome, which is a
sequence that triggers significant outmigration, can be represented with the following Boolean
expression:

C  0.50   R  0.20   B  0.50   S  0.20 (7.2)

Where C is the fraction of capacity disrupted [i.e., the tracking variable in Figure 7.1(a)],
R is the fraction of residents displaced, B is the fraction of businesses disrupted, S is the fraction
of students displaced, and the Boolean symbol  indicates the intersection of events (i.e., the
occurrence of two or more events). Note that each of these tracking variables can be considered a
random variable bounded between zero and one. In words, Equation (7.2) represents a sequence
in which more than 50% of the capacity of public services has been disrupted, more than 20% of
residents have been displaced, more than 50% of businesses have been disrupted, and more than
20% of students have been displaced. Equation (7.3) can be used to calculate P256, the probability
that this sequence—the 256th sequence—occurs.

       
P256  P  C  0.50  R  0.20  B  0.50  S  0.20  (7.3)

Similarly, Equation (7.4) can be used to compute P255, the probability that the second-
from-the-bottom outcome in Figure 7.2 occurs. This outcome is another sequence that results in a
significant and rapid outmigration of residents.

       
P255  P  C  0.50  R  0.20  B  0.50  0.10  S  0.20  (7.4)

Similar equations can be developed for the other 102 significant outmigration sequences.
Because each sequence is mutually exclusive, the conditional probability of significant
outmigration for the community is simply the sum of these 104 equations. In this example, the
sum needs to be less than 1×10-2 (1%) for the DBE and 1×10-1 (10%) for the MCE. In order to
ensure these requirements are satisfied, an appropriate multivariate probability distribution for
the tracking variables (C, R, B, and S) needs to be selected. A multivariate distribution is required
because, in general, the tracking variables are not pair-wise independent. In other words, the
tracking variables can be correlated. This correlated behavior arises from the fact that each of the
four vital community functions commonly relies on a shared network of frontline and support
systems in order to operate successfully. A multivariate distribution can capture this correlation.
In this example, however, we will assume that each of the tracking variables is
independent of the others. As a result, a multivariate probability distribution is no longer
required; instead, each tracking variable can be described individually with a separate univariate
distribution. This assumption greatly simplifies the computations of Equation (7.3) and Equation
(7.4); see Equation (7.5) and Equation (7.6), respectively, for the simplified expressions that

103
result. This assumption is justified because, while it impacts the final numbers, accuracy of
results is not the focus of this example. Instead, the focus is on the process, which is not affected
by the independence assumption. Future iterations of this example should address this
assumption by using a multivariate probability distribution to capture correlation among the four
tracking variables.

      
P256  P C  0.50  P R  0.20  P B  0.50  P S  0.20  (7.5)

      
P255  P C  0.50  P R  0.20  P B  0.50  P 0.10  S  0.20  (7.6)

Employing these simplifications and assumptions, it is relatively straightforward to

assign branch probabilities for each vital community function in a way that satisfies the
conditional probability targets for significant outmigration (1% in the DBE or 10% in the MCE).
There are many possible ways to assign branch probabilities. Again, community stakeholders
and policymakers have the freedom to choose branch probabilities that most appropriately reflect
the priorities and preferences of the community. Figure 7.3 and Figure 7.4 demonstrate one
possibility.

104
Figure 7.3 Branch probabilities corresponding to 1% probability of significant
outmigration in the design basis earthquake.

105
Figure 7.4 Branch probabilities corresponding to 10% probability of significant
outmigration in the maximum considered earthquake.

106
 Figure 7.3 shows an example of branch probabilities that satisfy the specified target of
1% probability of significant outmigration in the DBE. Note that the sum of branch probabilities
for each event tree must equal one. Also note that each vital community function is given equal
importance. In other words, the corresponding branches for each vital function are assigned the
same probability (e.g., the top outcome in each of the event trees in Figure 7.3 is assigned a
probability of 0.69). Using these quantities, it is straightforward to compute the probability
associated with each sequence. For example, Equation (7.5), which corresponds to the bottom
sequence in Figure 7.2, computes to:

P256  0.03 0.03 0.03 0.03  8.1107 (7.7)

If this calculation is repeated for the other 103 sequences that trigger significant
outmigration, the conditional probability of outmigration in the DBE can be computed by
summing each probability. For the branch probabilities in Figure 7.3, the conditional probability
of outmigration is 0.92%, which is less than the specified target of 1%. Recall, however, that
Figure 7.3 depicts only one of many possible ways to assign branch probabilities. Community
stakeholders may, for example, determine that housing is the most important vital community
function and, subsequently, assign branch probabilities so that the combined likelihood of the
two bottom outcomes in the housing event tree is smaller than the two bottom outcomes in the
three other event trees.
Similarly, Figure 7.4 shows an example of branch probabilities that satisfy the specified
target of 10% probability of significant outmigration in the MCE. As before, each vital
community function is given equal importance (e.g., the top outcome in each of the event trees in
Figure 7.4 is assigned a probability of 0.32). A similar procedure as outlined in the previous
paragraph can be used to calculate the probability of each sequence and, subsequently, the
conditional probability of significant outmigration in the MCE. For the branch probabilities in
Figure 7.4, the conditional probability of outmigration is 10.05%, which is very close to the
specified target of 10%. Again, Figure 7.4 depicts only one of many possible ways to assign
branch probabilities.
Before proceeding, it is important to note that the branch probabilities in Figure 7.3 and
Figure 7.4 represent design targets for each of the vital community functions and, subsequently,
the built environment that supports them. As such, branch probabilities do not reflect the actual
level of performance achieved by the built environment as it currently exists. Instead, they
represent the level of performance the community ultimately desires. If an evaluation reveals that
the existing built environment does not satisfy its specified performance targets, the community
needs to make investments to retrofit or replace its infrastructure in order to improve
performance.
The branch probabilities for each vital community function can be used to plot the
cumulative distribution function for each tracking variable. Figure 7.5 displays the cumulative
distribution function for each vital community function in both earthquake scenarios. It is
important to note upfront that the curves in Figure 7.5 are not the same as seismic fragilities, as
their interpretations are different. Seismic fragilities plot the conditional probability of failure of

107
a particular component or structure given a seismic demand parameter (e.g., peak ground
acceleration). The next paragraph describes how the curves in Figure 7.5 are to be interpreted.
Figure 7.5(b) shows the cumulative distribution function for the housing vital community
function in both the DBE and MCE. On the horizontal axis is r, the fraction of residents
displaced, which ranges between zero and one. Note the difference between R and r. R refers to
the random variable that describes the tracking variable, while r refers to a specific value that R
can take. On the vertical axis is FR(r), the cumulative distribution function for R, which is the
probability that R is less than or equal to r. To make these concepts more concrete, consider r =
0.20. From Figure 7.5(b), for the DBE, FR(0.20) = 0.97, which means that the community
requires a 97% probability that less than 20% of residents be displaced in the DBE, or a 3%
probability that more than 20% of residents be displaced. Similarly, for the MCE, FR(0.20) =
0.90, which means that the community requires a 90% probability that less than 20% of residents
be displaced in the MCE, or a 10% probability that more than 20% of residents be displaced.
These targets make sense: we would expect the probability of more than 20% of residents being
displaced to be higher in the MCE (10%) than the DBE (3%).

Figure 7.5 Cumulative distribution functions for each vital community function.

108
 Again, the cumulative distribution functions in Figure 7.5 represent design targets for a
community, and can serve as benchmarks either for measuring the performance of the vital
community functions as they currently exist in a community or for evaluating the effect of
engineering actions to improve their performance. Specifically, each cumulative distribution
function can be used to compute the mean, median, standard deviation, or any other statistic for a
particular tracking variable and hazard scenario. These statistics can, in turn, be used as the basis
for establishing performance targets for the corresponding vital community functions—the
ultimate goal of this subsection. In this example, we will use the mean to establish these design
targets. In general, the mean, or expected value, of a continuous random variable X can be
computed using the following equation:

E  X    x  f (x) dx (7.8)


Where f(x) is the probability density function of X and E[X] is the mean of X. If X is
strictly positive (P(X ≥ 0) = 1), the mean can be computed using the equivalent formula:

E  X    1 FX
(x) dx (7.9)
0

Where FX(x) is the cumulative distribution function of X. In this example, because each
tracking variable is bounded between zero and one, Equation (7.9) simplifies to:
1 1
E  X    1 FX
(x) dx  1 F X
(x) dx (7.10)
0 0

Where the integral of the cumulative distribution function, the second term on the right-
hand side of Equation (7.10), is simply the area under each of the curves in Figure 7.5. This area
is a function of not only the assigned branch probabilities in Figure 7.3 and Figure 7.4, but also
the specified branch limits (e.g., 2%, 10%, 20%, etc.). See Section 6.3 for a discussion of how
the branch limits for each of the event trees in Figure 7.1 were selected.
Table 7.1 summarizes the performance targets for each vital community function in both
hazard scenarios. These targets are based on the mean values of each tracking variable and were
calculated using Equation (7.10). Alternatively, the median or other statistic may also be used to
establish these design targets. From Table 7.1, we obtain the following two performance
objectives for the housing vital community function: (1) less than 4.9% of residents be displaced
in the DBE; and (2) less than 11.4% of residents be displaced in the MCE. Similar objectives can
be formulated for the other vital community functions.
In review, the performance targets in Table 7.1 result from four primary choices:
 The original community-level performance target (1×10-4 mean annual
frequency of significant outmigration)

109
  The hazard scenarios and corresponding conditional probabilities (1%
probability of significant outmigration in the DBE; 10% probability of
outmigration in the MCE)
 The technique for identifying sequences that trigger significant
outmigration
 The event tree branch probabilities (Figure 7.3 and Figure 7.4)
Again, each of these choices is at the discretion of community stakeholders and
policymakers. Changing any of the decisions made previously in this example will change the
results in Table 7.1.

Table 7.1 Performance targets for each basic safety function (based on mean values).

Vital community
Public services Housing Employment Education
function
(% of capacity (% of residents (% of businesses (% of students
(tracking
disrupted) displaced) disrupted) displaced)
variable)

475-year
9.3 4.9 9.3 6.2
earthquake
2450-year
19.6 11.4 19.6 12.5
earthquake

7.1.3 Performance Targets for Frontline and Support Systems

The third step in the methodology involves establishing performance targets for each frontline
and support system in the community. To this end, the performance objectives from the previous
step (see Section 7.1.2) need to be translated into targets for each frontline and support system.
Because the goal of this example involves developing performance targets for a new residential
building, we only need to focus on the frontline and support systems related to the housing vital
function.
Section 5.4 identified important frontline and support systems that support the housing
vital function, including single-family housing, multi-family housing, and water, electric power,
and natural gas lifelines (see Table 5.1 and Table 5.2). Physical damage to any of these systems
can displace residents from their homes, as described in Section 6.3.2. For simplicity, in this
example, we will assume that residents will be displaced from their homes only if the buildings
are not structurally safe enough to occupy immediately following an earthquake. In other words,
residents can stay in their homes even if they lack important utilities like water and power,
provided they are structurally safe to occupy. This would be the case for a community that has
adopted shelter-in-place performance requirements for its residential buildings [Poland et al.
2009; SPUR 2012]. As a result of this simplification, we only need to establish performance

110
objectives for frontline systems (i.e., the housing stock), while recognizing that an actual real-
world example would be more complicated.
To this end, if we assume that the percentage of residents displaced is roughly equivalent
to the percentage of the housing stock not safe to occupy, it is straightforward to establish
performance targets for the housing stock. The equivalence of these two measures is true for
communities with a large percentage of single-family homes, but starts to break down for
communities with large concentrations of multi-family residences. For the purposes of this
example, however, it is a sufficient approximation. As a result of this approximation, we now
have two performance targets for the housing stock: (1) less than 4.9% of residences not
structurally safe to occupy in the DBE; and (2) less than 11.4% of residences not structurally safe
to occupy in the MCE.

7.1.4 Performance Targets for Individual Components

The final step in the methodology involves establishing performance targets for individual
components within each frontline and support system. To this end, the performance objectives
from the previous step (see Section 7.1.3) need to be translated into targets for individual
components within each frontline and support system. In this example, the performance
objectives for the housing stock need to be translated into targets for an individual residential
building. On account of the assumptions and simplifications made in previous subsections, this is
a straightforward task. If the performance target is that no more than 4.9% of a community’s
housing stock will be structurally unsafe to occupy after the DBE, then an individual residential
building needs to have a 4.9% probability (or less) of being structurally unsafe. Similarly, if no
more than 11.4% of a community’s housing stock can be structurally unsafe to occupy after the
MCE, then an individual residential building needs to have an 11.4% probability (or less) of
being structurally unsafe.
In summary, we have established the following two performance targets for an individual
residential building: (1) less than 4.9% probability of being structurally unsafe to occupy in the
DBE; and (2) less than 11.4 % probability of being structurally unsafe to occupy in the MCE.
These targets are consistent with the community-level performance objectives developed in
Section 7.1.1.

7.1.5 Implications

The example described in prior subsections outlined a methodology for establishing a consistent
hierarchy of performance objectives for a community and its built environment. Specifically, it
developed a set of performance objectives for an individual residential building that, if achieved,
will satisfy community-level performance targets. These performance objectives can be used to
check whether the implicit performance levels achieved by the current building code (e.g., the
IBC) are satisfactory from the perspective of the community. Figure 7.6 displays results from an
analysis of the current building code performed by SPUR. Of the five performance categories
defined in the figure, three (C, D, and E) represent outcomes in which a building is not usable

111
following an earthquake. SPUR estimates that the provisions of the current building code result
in a new building having approximately 65% probability of being unsafe to occupy after the
DBE, where 65% is the sum of the black bars corresponding to Categories C, D, and E in Figure
7.6.

Figure 7.6 SPUR analysis of the current building code [Poland et al. 2009].

Recall that in the previous subsection we established the following performance target for
a new residential building in the DBE: 4.9% probability of being structurally unsafe to occupy
(or a 95.1% probability of being structurally safe to occupy). In light of this requirement, the
provisions of the current building code are inadequate, as they provide only a 35% probability
that a new building will be safe to occupy following the DBE (compared to the 95.1%
requirement). This discrepancy is significant; however, many simplifications and assumptions
were made in establishing the 95.1% requirement. Furthermore, the original community-level
performance target (1×10-4 mean annual frequency of significant outmigration) is probably too
conservative. A less stringent target for the community will likely lower the 95.1% requirement

112
for individual buildings in the DBE. Regardless, the example presented in this section illustrates
the potential mismatch in performance objectives that can arise when broader performance goals
for the community are ignored, as is the case in the IBC.
In summary, the methodology presented in this section can be used to create a set of
performance objectives for a wide range of components and subsystems within the built
environment, including not only residential buildings but also hospitals, factories, and electric
power grids. Ultimately, this set of performance objectives can serve as the basis for revisions to
the building code and can even provide the foundations for a “community performance code”
(see Chapter 8).

7.2 EXAMPLE: COMPUTING DISRUPTION TO PUBLIC SERVICES

The second example outlines a methodology for estimating the capacity of public services
disrupted following an initiating event. The methodology can be used for several purposes. First,
a community can use it to estimate the vulnerability of its public services to different hazard
scenarios, the results of which can be used to develop and implement targeted retrofit programs
or other mitigation activities that address its most critical vulnerabilities. Second, the
methodology can be used to calibrate the branch limits of the public services event tree [see
Figure 6.3 or Figure 7.1(a)]. As mentioned in Section 6.3.1, the branch limits for the public
services event tree may require refinement in order to properly distinguish the range of possible
outcomes (i.e., minor impact, moderate impact, significant impact, and catastrophic impact).
Using data from previous earthquakes and other hazards, the methodology outlined in this
example can be used to compute the initial disruption to public services that occurred during
these events. Once enough data have been compiled, the branch limits can be adjusted
appropriately. As the example in Section 7.1 demonstrated, branch limits are important because
they influence the performance objectives for each vital community function (see Section 7.1.2).
The methodology comprises six steps, as outlined in Table 7.2. The following
subsections describe each step in more detail.

Table 7.2 Methodology to estimate the disruption to public services.

Step Action

1 Define boundaries for the analysis

2 Determine the system importance matrix
3 Compute the system disruption matrix
4 Calculate the service disruption matrix
5 Determine the service importance matrix
6 Calculate the total disruption

113
7.2.1 Analysis Boundaries

The first step in the methodology involves defining appropriate boundaries for the analysis. Most
importantly, the individual public services to be included in the analysis need to be identified.
Table 7.3 lists the individual public services included in the analysis. It is adapted from Section
5.3.1 and reflects a baseline set of services essential both in normal, day-to-day operations and in
the aftermath of a major disaster. If desired, the list in Table 7.3 can be expanded to include
additional services. Note that essential community services, the last row in Table 7.3, include
building permit and inspection, planning, government finance and taxation, social services, mail
delivery, and public recordkeeping, to name only a few.
Once individual public services have been identified, the frontline and support systems
(i.e., buildings and lifelines) that enable these services need to be identified. Table 7.4 lists the
frontline and support systems included in the analysis (note that italics are used to differentiate
support systems from frontline systems). The list, which draws from Table 5.1 and Table 5.2 in
Section 5.4, can be expanded (or trimmed) if desired. Note that if a community has an auxiliary
water system for firefighting purposes (separate from the system that delivers potable water), it
needs to be added to the list. Also note that non-emergency medical facilities include medical
provider offices, clinics, and other outpatient facilities, while government facilities include
prisons, post offices, and administrative offices.

Table 7.3 Public services included in the analysis.

Public service

1 Police
2 Fire and rescue
3 Emergency medical care
4 Non-emergency health care
5 Food
6 Water
7 Energy
8 Sanitation
9 Transportation
10 Communication
11 Banking
12 Essential community services

114
 Table 7.4 Frontline and support systems included in the analysis.

Frontline/support system

1 Police stations
2 Fire stations
3 Hospitals
4 Dispatch centers
5 Emergency operations centers
6 Non-emergency medical facilities
7 Grocery stores
8 Banks
9 City hall
10 Government facilities
11 Water (potable)
12 Electric power
13 Natural gas
14 Oil
15 Solid fuels
16 Roads and highways
17 Mass transit
18 Railways
19 Airports
20 Ports and waterways
21 Telecommunications
22 Wastewater
23 Solid waste

7.2.2 System Importance Matrix

The second step of the methodology involves developing the system importance matrix. The
system importance matrix is an adaptation of the dependency matrix described in Section 5.5.1.
Whereas a dependency matrix identifies the basic interactions among the systems in a
community (see Figure 5.1), the system importance matrix also captures the strength of the
interaction or dependence. Specifically, it captures the extent to which the public services
identified in Table 7.3 depend on the frontline and support systems identified in Table 7.4.

115
 The system importance matrix, Isystem, is an m × n matrix, where m corresponds to the
number of public services included in the analysis and n corresponds to the number of frontline
and support systems. From Table 7.3 and Table 7.4, m = 12 and n = 23. Entries in Isystem range
between zero and one. Isystem(i,j) = 0 indicates that the ith public service does not depend on the jth
frontline or support system. On the other hand, Isystem(i,j) ≠ 0 indicates dependence to some
degree, with values close to one representing strong dependence. The ith row of Isystem describes
the dependence of the ith public service on the frontline and support systems included in the
analysis. The nonzero entries in the ith row must all sum to one.
The system importance matrix can be developed to capture basic dependencies either
during normal, steadystate conditions or in the aftermath of a major disaster. In an emergency
situation, certain public services may be able to operate in a limited capacity even if some of the
frontline and support systems they normally depend on are unavailable. For example, under
normal circumstances, banks serve as the headquarters for most of the banking services provided
to residents; however, after a major earthquake, it may be possible to provide these services even
if banking buildings are damaged and unsafe to occupy. For example, portable ATMs can be
brought in to provide cash to residents. As a result, the banking services row in the system
importance matrix might assign a smaller value for banks in the aftermath of a disaster than
during normal, steadystate conditions.
Before developing the system importance matrix, it is helpful to construct a dependency
matrix to first identify basic interactions. Figure 7.7 shows an example of a dependency matrix
that captures the interactions among the public services in Table 7.3 and the frontline and support
systems in Table 7.4. The matrix is constructed one row at a time, with an “x” mark indicating
that a particular public service depends on the corresponding frontline or support system. The
dependency matrix in Figure 7.7 is developed for application during a disaster scenario.
Furthermore, it is developed for general applicability to a wide range of communities; therefore,
the matrix will likely require adjustment in order to capture the interactions within a particular
community. These adjustments should be made by community stakeholders who have intimate
knowledge of particular public services (e.g., fire and police chiefs, utility operators, etc.).
The first row of the dependency matrix in Figure 7.7 shows that the police service
depends on eight frontline and support systems: police stations, dispatch centers, emergency
operations centers, government facilities (which include prisons), electric power, oil, roads and
highways, and telecommunications. Police stations serve as the primary frontline systems in law
enforcement operations in a community. To remain fully operational, they require access to basic
utilities like electricity and telecommunications. However, a significant percentage of police
services involve responding to emergencies throughout the community. Therefore, dispatch
centers and emergency operations centers, which direct officers to where they are needed, also
impact the availability of police services. Furthermore, the condition of transportation (in
particular, roads and highways), energy (in particular, oil/gasoline to fuel patrol vehicles), and
telecommunications infrastructure is vitally important in determining the disruption to the police
service.

116
 Figure 7.7 Dependency matrix for public services and frontline and support
systems.

Figure 7.8 shows an example of a system importance matrix developed using the
dependency matrix in Figure 7.7. The specific values assigned in the matrix will vary from
community to community and, again, should be developed by stakeholders who have intimate
knowledge of particular public services. If the values in each column of the matrix are summed,
a measure of the importance of each frontline and support system can be obtained. As Figure 7.8
demonstrates, support systems like telecommunications, electric power, roads and highways, and
water feature some of the highest totals, reinforcing the crucial role these lifelines play in
supporting multiple public services. For illustrative purposes, the rows of the matrix in Figure
7.8 sum to 100 instead of one.

117
 Figure 7.8 System importance matrix for public services and frontline and support
systems.

7.2.3 System Disruption Matrix

The third step in the methodology involves computing the system disruption matrix, Dsystem,
which measures the disruption to frontline and support systems caused by an initiating event. It is
an n × 1 matrix, where n is the number of frontline and support systems included in the analysis.
In this example, n = 23. The disruption to each system is represented by a single variable, as
shown in Equation (7.11). These variables can range between zero and one, where zero indicates
no disruption to the system and one represents complete disruption.

118
 D 
 police stations 
 D fire stations 
 
 Dhospitals 
 
 Ddispatch centers 
D 
 emergency operations centers 
D 
 nonemergency medical facilities 
 Dgrocery stores 
 
 Dbanks 
D 
 city hall 
D 
 government facilities  (7.11)
 Dwater ( potable ) 
 
Dsystem   Delectric power 
 
 Dnatural gas 
D 
 oil 
 Dsolid fuels 
 
 Droads and highways 
 
 Dmass transit 
D 
 railways 
 Dairports 
 
 D ports and waterways 
 
 Dtelecommunications 
D 
 wastewater 
 Dsolid waste 
 

An analysis of each frontline and support system must be performed in order to fully
determine the system disruption matrix. The nature of each analysis will depend on the overall
purpose of the study and the particular frontline or support system under consideration. For
example, if the purpose of the study is to calibrate the branch limits in the public services event
tree, then the analysis of each frontline and support system will require corresponding data sets
from previous disasters. On the other hand, if the purpose of the study involves identifying a
community’s vulnerabilities before disaster strikes, then analysis tools like HAZUS or the
lifeline interdependency models listed in Section 2.1.2 can be used to estimate the disruption
expected for each system. For example, the disruption to the hospitals frontline system can be
measured by the fraction of hospitals in the community that are not operational following an
initiating event, or, alternatively, by the fraction of hospital beds unavailable. See Equation
(7.12).

119
 number of beds unavailable (7.12)
Dhospitals 
total number of beds

7.2.4 Service Disruption Matrix

The fourth step in the methodology involves computing the service disruption matrix, Dservice,
which measures the disruption to individual public services caused by disruptions to frontline
and support systems. It is the product of the system importance matrix, Isystem, and the system
disruption matrix, Dsystem. See Equation (7.13).

Dservice  I system  Dsystem (7.13)

Dservice is an m × 1 matrix, where m is the number of public services included in the

analysis (m = 12 in this example). Equation (7.14) shows the individual entries of Dservice.

D 
 police 
 D fire and rescue 
 
 Demergency medical care 
 
 Dnonemergency health care 
D 
 food  (7.14)
D 
Dservice   water 
 Denergy 
 
 Dsanitation 
D 
 transportation 
 Dcommunication 
 
 Dbanking 
 
 Dessential community services 

Each entry in Dservice can be considered a weighted average or aggregation of the

disruptions to frontline and support systems, with the weights being specified in the system
importance matrix. For example, disruption to the police service, Dpolice, is a weighted average of
the disruption to police stations, dispatch centers, emergency operations centers, government
facilities (i.e., prisons), electric power, oil, roads and highways, and telecommunications. See
Equation (7.15).

D police  0.25 D police stations  0.05 Ddispatch centers 

0.05 Demergency operations centers  0.05 Dgovernment facilities  (7.15)
0.05 Delectric power  0.05 Doil  0.25 Droads and highways 
0.25 Dtelecommunications

120
 As can be seen in Equation (7.15), even a significant disruption to dispatch centers,
emergency operations centers, government facilities, electric power, or oil infrastructure has only
minor impact on Dpolice. In contrast, an extensive disruption to police stations, roads and
highways, or telecommunications infrastructure can have a substantial effect on Dpolice. Again,
the specific weights assigned to each frontline and support system (via the system importance
matrix) will likely require refinement in order to better reflect the interactions within a particular
community.

7.2.5 Service Importance Matrix

The fifth step in the methodology involves determining the service importance matrix, Iservice,
which measures the relative importance of each public service to the community. It is a 1 × m
matrix, where m is the number of public services included in the analysis (m = 12 in this
example). The sum of all entries in Iservice must equal one. Similar to the system importance
matrix, the service importance matrix can be developed to reflect the importance of individual
public services either during normal, steady state conditions or during the emergency response
phase after a major disaster. Figure 7.9 displays an example of each. During normal, steady state
conditions [Figure 7.9(a)], each service is weighted approximately the same. In contrast, during
the emergency response phase [Figure 7.9(b)], services like police, fire and rescue, and
emergency medical care are assigned higher weights because of their importance in responding
to an emergency. As elsewhere, the numbers presented in Figure 7.9 can be adjusted to reflect
the priorities and preferences of stakeholders in a particular community.

121
 Figure 7.9 Examples of a service importance matrix.

7.2.6 Total Disruption

The final step in the methodology involves computing the total disruption to public services,
Dtotal, which is an aggregation of the disruption to individual public services. Note that Dtotal is
equivalent to C, the fraction of capacity disrupted (as defined in Section 7.1.2), which, in turn, is
equivalent to the tracking variable in the public services event tree for the initial damage phase
[see Figure 6.3 or Figure 7.1(a)]. Dtotal is the product of the service importance matrix and the
service disruption matrix. See Equation (7.16). It can be considered a weighted average of the

122
disruptions to each individual public service, with the weights coming from the service
importance matrix.

Dtotal  I service  Dservice (7.16)

If Equation (7.13) and Equation (7.16) are combined, Equation (7.17) results.

Dtotal  I service  I system  Dsystem (7.17)

In summary, Dtotal, as computed using Equation (7.16) or Equation (7.17), can be used to
determine the branch of the public services event tree on which a community resides ([see Figure
6.3 or Figure 7.1(a)]. For example, if Dtotal = 0.36 = 36%, the community resides on the third
branch of the public services event tree, which indicates the initial disruption to public services
will have significant impact on the community.

7.2.7 Implications

Previous subsections have outlined a methodology that aggregates the disruption to public
services using a single measure, Dtotal. In spite of its simplicity, the implications of the
methodology are significant. As discussed at the beginning of Section 7.2, it can be used to
calibrate the branch limits of the public services event tree for the initial damage phase [see
Figure 6.3 or Figure 7.1(a)]. This calibration process requires development of an extensive
database of observations gathered from a wide range of previous disasters. The development of
this database, while beyond the scope of this report, is one of the future tasks identified in
Section 8.2. Proper calibration of the branch limits in the public services event tree will enable
communities to more accurately gauge the expected performance of their public services. In
other words, it will allow them to get a better sense of whether the total disruption to public
services (computed using the methodology outlined in previous subsections) has catastrophic,
significant, moderate, or minor impact.
Furthermore, the methodology can be used to measure the effect of different mitigation
strategies once a community has estimated the total disruption expected to its public services.
For example, consider a community that performs an analysis that estimates each of its frontline
and support systems will have 25% of its capacity disrupted in a particular hazard scenario (i.e.,
Dsystem(i,1) = 0.25 for i = 1, 2, … , n). Using the system importance matrix in Figure 7.8, the
service importance matrix in Figure 7.9(b), and Equation (7.17), an estimate of Dtotal can be
computed. In this example, because Dsystem (i,1) = 0.25 for i = 1, 2, … , n, Dtotal is simply 0.25. In
order to decrease the overall vulnerability of its public services (i.e., Dtotal), the community can
perform a wide range of mitigation activities that reduce the vulnerability of various frontline
and support systems (i.e., the entries in Dsystem). The methodology can be used to measure and
compare the effectiveness of each of these mitigation activities. Figure 7.10 shows the
effectiveness of reducing the disruption to each frontline system from 0.25 to zero on Dtotal.
Similarly, Figure 7.11 displays the effectiveness of reducing the disruption to each support
system from 0.25 to zero.

123
 0.25

0.245

0.24

0.235

0.23
Dtotal

0.225

0.22

0.215
police stations
fire stations
0.21 hospitals
dispatch centers
emergency operations centers
non−emergency medical facilities
0.205 grocery stores
banks
city hall
government facilities
0.2
0.25 0.2 0.15 0.1 0.05 0
disruption to frontline system

Figure 7.10 Effectiveness of decreasing the vulnerability of individual frontline

systems.

124
 0.25

0.245

0.24

0.235

0.23
Dtotal

0.225

0.22

water (potable)
electric power
0.215
natural gas
oil
solid fuels
0.21 roads and highways
mass transit
railways
airports
0.205 ports and waterways
telecommunications
wastewater
solid waste
0.2
0.25 0.2 0.15 0.1 0.05 0
disruption to support system

Figure 7.11 Effectiveness of decreasing the vulnerability of individual support

systems.

In this example, the effect of decreasing the vulnerability of frontline systems is not as
significant as decreasing the vulnerability of support systems (see Figure 7.10 and Figure 7.11).
Even so, improving the performance of hospitals, police stations, and grocery stores has the most

125
effect. However, the most effective way to reduce Dtotal is to decrease the vulnerability of
telecommunications, electric power, roads and highways, and water support systems. Therefore,
if faced with limited resources for mitigation activities, a community should focus efforts on
retrofitting or replacing these particular systems, at least in this example.
However, it is important to note that additional factors, including cost, environmental
impact, and political or legal constraints, need to be considered when choosing among potential
mitigation activities. For example, while it might produce significant reductions to Dtotal,
upgrading a community’s water infrastructure is typically expensive and disruptive to nearby
residents and businesses, as it may require replacement of lengthy sections of pipeline buried
under city streets. Furthermore, sometimes the water infrastructure within a community can be
part of a larger network that serves other communities, meaning an individual community may
have limited control over retrofit decisions.
In summary, the example presented in this section outlines a straightforward
methodology for capturing the interdependencies that exist between a community’s public
services and its frontline and support systems to estimate the total disruption to a community’s
public services. This methodology, when combined with analysis tools like HAZUS-MH, may
be used to make better decisions regarding how to address the seismic vulnerabilities of a
community’s built environment. In addition, it may be used to calibrate the public services event
tree in order to improve the rigor of the performance objectives established for a community and
the numerous components and subsystems within its built environment.

126
8 Conclusions

A community is a dynamic system of people, organizations, and patterned relationships and

interactions [Alesch 2005]. Structures and hardware, referred to as the built environment in this
report, play a particularly important role in enabling a community to successfully function,
providing the physical foundations for much of the economic and social activities that
characterize a modern society [O’Rourke 2007]. The built environment is a complex, dynamic,
interdependent network of engineered subsystems and components, including buildings, bridges,
pipelines, and other structures. Natural hazards such as earthquakes, hurricanes, and floods can
damage a community’s built environment, which in turn may disrupt the security, economy,
safety, health, and welfare of the public. In response, many communities have developed and
implemented regulatory frameworks to ensure minimum levels of performance for individual
parts of the built environment. For buildings subject to earthquakes, these minimum levels of
performance center on preventing collapse during very rare, intense seismic ground motion
[BSSC 2009].
This report has examined the regulatory framework currently used in the United States to
design and evaluate a community’s built environment to withstand the effects of earthquakes and
other natural hazards. Using the attributes of an ideal regulatory framework as a guide, it has
identified and described several important shortcomings. The most significant shortcoming of the
current regulatory framework involves its lack of an integrated, coordinated, and comprehensive
approach to establishing performance expectations for individual components of the built
environment. Consequently, performance objectives for individual components within the built
environment are not tied to broader performance targets for the community. This divergence
results in a community in which most individual components behave as intended; however, when
aggregated, the performance of and interaction among components can result in unacceptable
outcomes for the community (i.e., insufficient levels of resilience).
To address the shortcomings of the current regulatory framework, this report has studied
the philosophy used in the United States to design and analyze nuclear power plants and has
adapted it for use in a community setting. Most crucially, the nuclear design philosophy features
an integrated, coordinated, and comprehensive approach. It begins at the system level, specifying
performance objectives that result in a small probability of unacceptably large radiation release
affecting the nearby population. The framework then identifies the vital safety functions that
must be available during a postulated accident in order for a nuclear power plant to avert core

127
damage and/or large release of radioactivity. It then identifies the frontline and support systems
that enable each vital safety function to operate successfully. This systematic, top-down
approach ensures that performance requirements for individual components and subsystems are
consistent with system-level performance targets. In order to verify that a nuclear power plant
satisfies these performance objectives, nuclear engineers use tools like probabilistic risk
assessments, event trees, and fault trees to analyze the response of the plant and its various
components and subsystems.
In adapting the nuclear design philosophy to communities, this report has drawn
extensively from the rapidly evolving fields of community resilience and lifeline
interdependency. The adaptation begins with defining undesired outcomes for a community
whose occurrence, because of their adverse consequences, should be minimized to the extent
possible. This report has selected a significant and rapid outmigration of residents as the
undesired outcome of interest because it has been observed in the aftermath of several major
disasters, including the Great Hanshin earthquake in 1995 and Hurricane Katrina in 2005. Using
the work of Poland et al. [2009], SERRI and CARRI [2009], Twigg [2009], and Cutter et al.
[2010], the report has identified four vital functions, public services, housing, employment, and
education, that a community must maintain in the aftermath of a major earthquake or other
natural hazard in order to prevent a significant and rapid outmigration of residents. It then
identifies the many frontline and support systems within the built environment that enable each
vital function to operate successfully. This list was adapted from PCCIP [1997], Rinaldi et al.
[2001], ALA [2004], ASCE [2006], Barkley [2009], and Poland et al. [2009]. In summary, these
vital community functions and their corresponding frontline and support systems prevent a
significant and rapid outmigration of residents from occurring in a community after an initiating
event (e.g., earthquakes, hurricanes, and floods).
Furthermore, this report has developed a set of event trees for a community that can be
used to identify combinations or sequences of events that result in a significant and rapid
outmigration of residents. The event trees track the status of each vital function after an initiating
event. Together, these concepts and tools form the foundations of a performance-based
engineering framework that can be used for many different purposes, ranging from the revision
of building code provisions to the evaluation of competing retrofit strategies.
Lastly, this report has presented several conceptual examples that illustrate application of
the proposed engineering framework. The first example outlined a methodology for creating a
consistent hierarchy of performance objectives for a community. Specifically, it illustrated how
the event trees developed for a community can be used to establish performance objectives for a
residential building from a community-level performance target. Subsequently, these
performance objectives can be used to update building code provisions so that new residential
buildings will perform in a manner consistent with community-level resilience goals. The second
example outlined a methodology for estimating the capacity of public services disrupted by an
earthquake or other natural hazard. This methodology can be used to estimate the vulnerability of
a community’s public services and, subsequently, to evaluate the effectiveness of different
mitigation activities or strategies. It can also be used to refine the branch limits of the public

128
services event tree so that each outcome better delineates the scale of potential consequences
(i.e., minor, moderate, significant, or catastrophic).

8.1 IMPLICATIONS

The work presented herein has the potential to change the way engineers, planners, and other
stakeholders design and evaluate the built environment of a community. The growing interest in
sustainable and resilient communities necessitates an updated regulatory framework, one that
employs an integrated, coordinated, and comprehensive approach to account for the numerous
subsystems, components, and their interactions. The framework presented in this report provides
a transparent, performance-based, risk-informed methodology for planners and policymakers to
set community-level performance targets and, subsequently, for engineers to calibrate their
designs to meet these community-level performance targets. It provides the missing link between
community-level resilience goals and component-level performance objectives. Together, the
findings presented herein establish the foundations for a much-needed transformation from
engineering individual components of the built environment on a component-by-component
basis to engineering community resilience using an integrated and coordinated approach that
begins at the community level.
For example, the framework proposed herein may provide the basis for a document that
plays a role in the design and evaluation of a community’s built environment similar to the role
of building codes in the design and evaluation of buildings. This “community performance code”
would contain provisions that explicitly spell out the performance expectations for a particular
community. These explicit performance statements might take many different forms; Section 7.1
presented an example in which the community performance target took the form of an
annualized probability of significant outmigration. Performance targets could also take the form
of specific timetables for recovery. Regardless of their particular form, these community-level
performance targets need to be established by community stakeholders in a public process so that
they most effectively represent the level of risk acceptable to society.
In turn, this “community performance code” may provide the foundations for revisions to
buildings codes and/or the development of new design standards for lifelines and other
infrastructure. Using the framework and methodology proposed in this report, community-level
performance targets specified in the “community performance code” may be used to develop a
consistent set of performance objectives for each of the various components and subsystems
within the built environment. In the end, these performance targets may resemble those described
by Poland et al. [2009] (see Figure 2.4 and Figure 2.5); however, they will be developed using a
more robust, transparent, and technically grounded engineering framework. The example in
Section 7.1 demonstrated how to develop performance objectives for a residential building from
a community-level resilience goal. Performance targets for other components and subsystems
within the built environment (e.g., hospitals, electric power grids, etc.) can be developed in a
similar fashion. These performance objectives, in turn, can be used to develop appropriate
provisions for inclusion in the corresponding design standards or codes.

129
 The “community performance code” can also be used as the basis for a community-wide
resilience rating system. Communities that meet or exceed the performance objectives specified
in the “community performance code” would receive higher resilience ratings than those that do
not. A key element in this rating system involves having the capability to evaluate the response
of an entire community relative to its specified performance targets. At the moment, an analysis
like this is beyond the capabilities of most communities because they lack both the necessary
inventory data about the built environment and sufficient performance evaluation tools. The
framework presented herein, once refined and tested further, can function as one such tool.

8.2 FUTURE WORK

Much future work remains to be done. There are two general areas that require further attention.
First, the community event trees presented in Chapter 6 require further refinement and possible
expansion. Second, the methodology outlined in Section 7.1 needs to be refined and expanded.
The following subsections discuss each of these tasks in more detail.

8.2.1 Refinement and Expansion of Community Event Trees

One of the most important next steps involves additional refinement of the community event
trees presented in Chapter 6. In particular, the branch limits for several event trees require
calibration in order to more effectively delineate the impact of each of the four outcomes in an
event tree. As defined in Section 6.2 and Figure 6.1, the four outcomes of an individual event
tree have varying impact, ranging from minor to catastrophic. The branch limits in an event tree
establish boundaries that distinguish each outcome. Therefore, branch limits must be chosen
carefully so that they effectively demarcate each of the four possible outcomes.
The branch limits for several of the event trees presented in Chapter 6 require refinement
in order to more effectively differentiate the range of possible outcomes. In particular, the event
trees for public services, employment, and education need most attention (see Figure 6.3, Figure
6.4, and Figure 6.5, respectively). For each of these trees, the chosen branch limits may not
properly delineate the range of possible outcomes. For example, for the employment event tree
(see Figure 6.3), it is unclear whether a disruption of more than 50% of businesses actually has
catastrophic impact. With additional research, including analyses of previous disasters, it may be
discovered that a disruption of 35% of businesses has catastrophic impact.
In order to establish more appropriate branch limits, a comprehensive analysis of
previous disasters must be performed. This analysis requires development of two items. First,
methodologies for estimating the initial disruption to public services, employment, and education
need to be developed, verified, and validated. Section 7.2 described a methodology for
estimating the disruption to public services; similar methodologies need to be developed for
calculating the disruption to employment and education. Second, an extensive database of
observations from a wide range of previous disasters needs to be developed. This database would
contain information like the percentage of businesses without immediate access to electricity or
natural gas and the percentage of schools damaged to an extent that they are not safe to occupy.

130
Once these two analyses have been completed, they can be used to compute the disruption to
public services, employment, and education that occurred during previous disasters. Ideally,
trends in the results of this analysis will emerge and, subsequently, provide the basis for more
appropriate branch limits for each event tree. For example, an analysis of previous disasters may
reveal that if more than 35% of businesses are disrupted, the impact on the community is likely
to be catastrophic. As a result, the first branch limit in the employment event tree would be set to
35%. As a result of these efforts, a generic set of community event trees can be obtained.
Individual communities can modify them if they have reason to believe that different branch
limits are more appropriate.
The event trees presented in Chapter 6 capture the impact of damage on the availability
of each of the four vital community functions. As such, they can be used to develop performance
targets that focus on limiting the immediate damage and disruption caused by an initiating event.
However, limiting initial damage is but one aspect of resilience; another important element
involves containing the effects of disasters when they occur [Bruneau et al. 2003]. Therefore, in
order to more adequately address the multi-faceted nature of resilience, additional sets of event
trees can be developed to track recovery following an initiating event. Ultimately, these event
trees can be used to develop an additional set of performance objectives for a community that
focus on the restoration of vital functions.
Consider, for example, the set of performance objectives established for a residential
building in Section 7.1: (1) less than 4.9% probability of being structurally unsafe to occupy in
the DBE; and (2) less than 11.4% probability of being structurally unsafe to occupy in the MCE.
These two objectives aim to minimize the initial damage and disruption caused by an earthquake;
however, they do little to ensure that full functionality is restored to the building in a timely
manner. Therefore, an additional set of performance objectives that address restoration of
functionality needs to be developed using both the methodology outlined in Section 7.1 and yet-
to-be-developed event trees that track community recovery. The resulting set of performance
objectives for residential buildings would address not only initial damage caused by an
earthquake but also restoration of functionality, thereby enhancing community resilience in a
more complete fashion.
In addition to expanding the framework to include additional facets of resilience, future
iterations of the framework should also aim to expand its scope beyond disaster resilience to
address and incorporate broader sustainability considerations like carbon footprint, energy
efficiency, resource consumption, and environmental impact of a community and its built
environment.

8.2.2 Refinement and Expansion of the Methodology to Develop Consistent

Performance Targets for the Built Environment

The second major task involves refining and expanding the methodology used to develop
consistent performance objectives for individual components within the built environment.
Section 7.1 outlined the foundations of this methodology and presented an example in which a
set of performance objectives for a residential building was derived from a community-level

131
performance target. Many simplifications and assumptions were made throughout the course of
the example. The following paragraphs discuss the future work required to address the most
critical of these simplifications and assumptions.
The first step in the methodology involves establishing explicit performance targets for
the community under consideration (see Section 7.1.1). For illustrative purposes, a community
performance target of 1x10-4 mean annual frequency of significant outmigration was selected,
though it was noted that this target could take other values (e.g., 1×10-3 or 1×10-2). Additional
research is required to determine how often a community is willing to tolerate a disaster with
catastrophic local and regional consequences (e.g., Hurricanes Katrina and Sandy). This
conversation may need to take place at the national level, as the implications of this decision can
have profound impact on the economic security of the United States, mainly because the federal
government typically bears a significant share of the costs associated with major disasters.
The following simple example illustrates some of the factors that need to be considered
when selecting these community-level performance targets. If there are 100 major metropolitan
areas in the United States, then a community-level performance target of 1×10-4 mean annual
frequency of significant outmigration equates to approximately one catastrophic event
somewhere in the United States every 100 years. In contrast, a performance target of 1×10-3
translates into roughly one event every 10 years. So while an individual community may be
comfortable with a target of 1×10-3 (i.e., one catastrophic event every 1000 years), from a
national perspective, this target may be insufficient. Detailed cost-benefit analyses can help
illuminate which targets are most appropriate, both at local and national levels.
The second step in the methodology involves establishing performance objectives for
each of the four vital community functions (see Section 7.1.2). In order to simplify this process,
it was assumed that each of the vital functions was pair-wise independent, though it was noted
that this assumption was not realistic because each function commonly relies on a shared
network of frontline and support systems in order to operate successfully. Additional research is
required to identify a multivariate distribution that can adequately capture the correlations among
the four vital community functions. Once an appropriate distribution is selected, parametric
studies that investigate the impact of correlation on the performance objectives required for each
vital function can be performed.
The third and fourth steps in the methodology involve establishing performance
objectives for frontline and support systems and individual components, respectively (see
Section 7.1.3 and Section 7.1.4). The example focused on developing performance objectives for
the housing stock and, subsequently, individual residential buildings. This process was relatively
straightforward, largely due to simplifications and assumptions that allowed important
interactions to be ignored. Future work is required to establish performance objectives for
additional frontline and support systems (e.g., electric power grids, communication networks,
etc.) and individual components (e.g., electrical substations, hospitals, bridges, etc.) within the
built environment. For many of these systems and components, interactions cannot be ignored:
for example, a hospital cannot function without power and water. At the moment, however, it is
unclear how to account for these interactions in a generic fashion in order to develop a set of

132
design targets for these components. This is probably the most important task moving forward,
as it is a major impediment to the development of the aforementioned “community performance
code.”

133
134
 REFERENCES

ALA (2004). American Lifelines Alliance Matrix of Standards and Guidelines for Natural Hazards, American
Lifelines Alliance, retrieved from
http://www.americanlifelinesalliance.com/pdf/CombinedMatrices061404AsPosted.pdf.
ALA (2005). Protecting Our Critical Infrastructure: Findings and Recommendations from the American Lifelines
Alliance Roundtable, American Lifeline Alliance, retrieved from
http://www.americanlifelinesalliance.com/pdf/RoundtableReportWebPosted081705.pdf.
Alesch D.J. (2005) Complex urban systems and extreme events: Toward a theory of disaster recovery, Proceedings,
1st International Conference on Urban Disaster Reduction, Kobe, Japan.
Anagnos T., Comerio M.C., Goulet C., Na H., Steele J., Stewart J.P. (2008). Los Angeles inventory of nonductile
concrete buildings for analysis of seismic collapse risk hazards, Proceedings, 14th World Conference on
Earthquake Engineering, Beijing, China.
Ang A.H., Tang W.H. (2007). Probability Concepts in Engineering: Emphasis on Applications in Civil and
Environmental Engineering, 2nd, ed., John Wiley & Sons, Hoboken, NJ.
ACI (2011). Building Code Requirements for Structural Concrete and Commentary (ACI 318-11), American
Concrete Institute, Farmington Hills, MI.
AISC (2010a). Seismic Provisions for Structural Steel Buildings (ANSI/AISC 341-10), American Institute of Steel
Construction, Chicago, IL.
AISC (2010a). Specification for Structural Steel Buildings (ANSI/AISC 360-10), American Institute of Steel
Construction, Chicago, IL.
ANSI, ANS (2003). External-Events PRA Methodology (ANSI/ANS-58.21-2003), American National Standards
Institute, American Nuclear Society, La Grange Park, IL.
ANS, IEEE (1983). PRA Procedures Guide: A Guide to the Performance of Probabilistic Risk Assessments for
Nuclear Power Plants (NUREG/CR-2300), American Nuclear Society, Institute of Electrical and Electronics
Engineers, United States Nuclear Regulatory Commission, Washington, D.C.
ASCE (2000). Prestandard and Commentary for The Seismic Rehabilitation of Buildings (FEMA 356), American
Society of Civil Engineers, Federal Emergency Management Agency, Washington, DC.
ASCE (2005). Seismic Design Criteria for Structures, Systems, and Components in Nuclear Facilities (ASCE 43-
05), American Society of Civil Engineers, Reston, VA.
ASCE (2006). Minimum Design Loads for Buildings and Other Structures (ASCE 7-05), American Society of Civil
Engineers, Reston, VA.
ATC (2010). Here Today – Here Tomorrow: The Road to Earthquake Resilience in San Francisco: Potential
Earthquake Impacts (ATC-52-1), Applied Technology Council, Redwood City, CA.
Barkley C. (2009). Lifelines: Upgrading Infrastructure to Enhance San Francisco’s Earthquake Resilience, San
Francisco Planning and Urban Research Association (SPUR), San Francisco, CA.
Bea R.G. (2007). Margins of Quality in Engineered Systems: Reliability-Based Engineering for Design,
Requalification, Construction, Operation, Maintenance, and Decommissioning, CE205B course manual.
Department of Civil & Environmental Engineering, University of California, Berkeley, CA.
Bea R.G. (2008). Human and Organizational Factors: Quality and Reliability of Engineered Systems, CE290A
course manual, Department of Civil & Environmental Engineering, University of California, Berkeley, CA.
Billinton R., Allan R.N. (1983).Reliability Evaluation of Engineering Systems. Plenum Press, London.
Blanchard B.S., Fabrycky W.J. (2006). System Engineering and Analysis, 4th ed., Pearson Prentice Hall, Upper
Saddle River, NJ.

135
Bruneau M., Chang S.E., Eguchi R.T., Lee G.C., O’Rourke T.D., Reinhorn A.M., Shinozuka M., Tierney K.,
Wallace W.A., von Winterfeldt D. (2003). A framework to quantitatively assess and enhance the seismic
resilience of communities, Earthq. Spectra, 19(4): 733–752.
Buede D.M. (2000). The Engineering design of Systems: Model and Methods, John Wiley & Sons, New York, NY.
BSSC (2004). NEHRP Recommended Provisions for Seismic Regulations for New Buildings and Other Structures
(FEMA 450), Building Seismic Safety Council, National Institute of Building Sciences, Washington, D.C.
BSSC (2009). NEHRP Recommended Provisions for Seismic Regulations for New Buildings and Other Structures
(FEMA P-750), National Institute of Building Sciences, Washington, D.C.
Burton I., Kates R.W., White G.F. (1978). The Environment as Hazard, Oxford University Press, New York, NY.
CERA (2012a). CBD Cordon Reduction, Canterbury Earthquake Recovery Authority, retrieved from
http://cera.govt.nz/maps/cordon-reduction
CERA (2012b). Recovery Strategy for greater Christchurch, Canterbury Earthquake Recovery Authority, retrieved
from http://cdn.cera.govt.nz/sites/cera.govt.nz/files/common/recovery-strategy-for-greater-christchurch.pdf.
Chang S.E. (1996). Regional economic impact of the January 17, 1995 Kobe, Japan, earthquake, Proceedings, 43rd
North American Meeting of the Regional Science Association International, Washington, D.C.
Chang S.E. (2000). Disasters and transport systems: Loss, recovery and competition at the Port of Kobe after the
1995 earthquake, J. Transport Geography, 8(1): 53–65.
Chang S.E. (2010). Urban disaster recovery: a measurement framework and its application to the 1995 Kobe
earthquake, Disasters, 34(2): 303–327.
Chang S.E., Shinozuka, M. (2004). Measuring improvements in the disaster resilience of communities, Earthq.
Spectra, 20(3): 739–755.
City of Kobe (2012). The Great Hanshin-Awaji Earthquake Statistics and Restoration Progress, retrieved from
http://www.city.kobe.lg.jp/safety/hanshinawaji/revival/promote/january.2012.pdf
Coglianese C., Nash J., Olmstead T. (2002). Performance-Based Regulation: Prospects and Limitations in Health,
Safety, and Environmental Protection (Regulatory Policy Program Report No. RPP-03), Harvard University,
Cambridge, MA.
Comartin C.D., Anagnos T., Faison H., Greene M., Moehle J.P. (2008). The Concrete Coalition: Building a network
to address nonductile concrete buildings, Proceedings, 14th World Conference on Earthquake Engineering,
Beijing, China.
Comartin C.D., Bonowitz D., Greene M., McCormick D., May P., Seymour E. (2011). The Concrete Coalition and
the California Inventory Project: An Estimate of the Number of Pre-1980 Concrete Buildings in the State,
retrieved from http://www.concretecoalition.org/wp-content/uploads/2012/05/concrete_coalition_final-revise-
0512.pdf.
Comerio M.C. (1998). Disaster Hits Home: New Policy for Urban Housing Recovery. University of California
Press, Berkeley, CA.
Cutter S.L., Burton C.G., Emrich C.T. (2010). Disaster resilience indicators for benchmarking baseline conditions.,
J. Homeland Security and Emergency Management, 7(1): Article 51.
Definitions (2000). Title 44. C.F.R. pt. 206.401.
Der Kiureghian A., Ditlevsen, O. (2009). Aleatory or epistemic? Does it matter? Struct. Safety, 31(2): 105–112.
DesRoches R., Comerio M., Eberhard M., Mooney W., Rix G.J. (2011). Overview of the 2010 Haiti earthquake,
Earthq. Spectra, 27(S1): S1–S21.
Dezfuli, H. and Modarres, M. (1985). “Uncertainty analysis of reactor safety systems with statistically correlated
failure data.” Reliability Engineering, 11(1), 47–64.
Dueñas-Osorio L., Craig J.I., Goodno B.J. (2007). Seismic response of critical interdependent networks, Earthq.
Eng. Struct. Dyn., 36(2): 285–306.

136
EERI (2011). Learning from earthquakes: The M 6.3 Christchurch, New Zealand, Earthquake of February 22,
2011, Earthquake Engineering Research Institute, Oakland, CA.
FEMA (1997). Multi-Hazard Identification and Risk Assessment, Federal Emergency Management Agency,
Washington, DC.
Fullwood R.R. (2000). Probabilistic Safety Assessment in The Chemical And Nuclear Industries, Butterworth–
Heinemann, Woburn, MA.
Haimes Y.Y., Jiang P. (2001). Leontief-based model of risk in complex interconnected infrastructures, J.
Infrastructure Systems, 7(1): 1–12.
Helton J.C. (1994). Treatment of uncertainty in performance assessments for complex systems, Risk Analysis, 14(4):
483–511.
Hernandez-Fajardo I., Dueñas-Osorio, L. (2011). Sequential propagation of seismic fragility across interdependent
lifeline systems, Earthq. Spectra, 27(1): 23–43.
Hori M., Schafer M.J., Bowman, D.J. (2009). Displacement dynamics in Southern Louisiana after Hurricanes
Katrina and Rita, Population Res. Policy Review, 28(1): 45–65
Horwich G. (2000). Economic lessons of the Kobe earthquake, Economic Development and Cultural Change, 48(3):
521–542.
IAEA (2007). IAEA Safety Glossary: Terminology Used in Nuclear Safety and Radiation Protection, International
Atomic Energy Agency, Vienna, Austria.
IAEA (2009). INES: The International Nuclear and Radiological Event Scale User’s Manual, International Atomic
Energy Agency, Vienna, Austria.
ICBO (1997). Uniform Building Code. International Code Council, International Conference of Building Officials
Country Club Hills, IL.
ICC (2006). International Building Code, International Code Council, Country Club Hills, IL.
INSAG (1999). Basic Safety Principles for Nuclear Power plants 75-INSAG-12 Rev. 1 (INSAG-12), International
Nuclear Safety Advisory Group, International Atomic Energy Agency, Vienna, Austria.
ISO, IEC (2008). Systems and Software Engineering — System Life Cycle Processes (ISO/IEC 15288). International
Organization for Standardization and International Electrotechnical Commission, Geneva, Switzerland.
Johnson L., Eckroad L. (2001). Summary Report on the City and County of San Francisco’s Post-Disaster Interim
Housing Policy Planning Workshop. San Francisco Department of Emergency Management, San Francisco,
CA.
Kaplan S., Garrick B.J. (1981). On the quantitative definition of risk, Risk Analysis, 1(1): 11–27.
Karlinsky S. (2009). The ground beneath our feet, in: Urbanist, 479, 2.
Kossiakoff A., Sweet W.N., Seymour S.J., Biemer S.M. (2011). Systems Engineering Principles and Practice, 2nd
ed., John Wiley & Sons, Hoboken, NJ.
Kramer S.L. (1996). Seismic hazard analysis, in: Geotechnical Earthquake Engineering. Prentice Hall, Upper
Saddle River, NJ, 106–142.
Lee E., Mitchell J., Wallace W. (2007). Restoration of services in interdependent infrastructure systems: A network
flows approach, IEEE Transactions on Systems, Man, and Cybernetics Part C: Applications and Reviews,
37(6): 1303–1317.
Lin Y., Chang, P. Fiondella L. (2012). A study of correlated failures on the network reliability of power
transmission systems, Inter. J. Electric Power & Energy Systems, 43(1): 954–960.
May P.J., Koski, C. (2004). Performance-based regulation and regulatory regimes, Report No. PEER 2004/06,
Pacific Earthquake Engineering Research Center, Berkeley, CA.
Meacham B., Bowen R., Traw J., Moore, A. (2005). Performance-based building regulation: Current situation and
future needs, Building Research & Information, 33(2): 91–106.

137
Miles S.B., Chang S.E. (2003). Urban disaster recovery: A framework and simulation model, Report No. MCEER-
03-0005, Multidisciplinary Center for Earthquake Engineering Research, Buffalo, NY.
Miles S.B., Chang S.E. (2006). Modeling community recovery from earthquakes, Earthq. Spectra, 22(2): 439–458.
Miles S.B., Chang S.E. (2007). A simulation model of urban disaster recovery and resilience: Implementation for
the 1994 Northridge earthquake, Report No. MCEER-07-0014, Multidisciplinary Center for Earthquake
Engineering Research, Buffalo, NY.
Minai A.A., Braha D, Bar-Yam Y. (2006). Complex engineered systems: A new paradigm, in: Complex Engineered
Systems: Science Meets Technology. D. Braha, A.A. Minai, and Y. Bar-Yam, eds. Springer, Berlin.
NASA (2007). Systems Engineering Handbook (NASA/SP-2007-6105 Rev1), National Aeronautics and Space
Administration, Washington, D.C.
OAS (1991). Primer on Natural Hazard Management in Integrated Regional Development Planning, Organization
of American States, Washington, D.C.
ONRR (2007). Feasibility Study for a Risk-Informed and Performance-Based Regulatory Structure for Future Plant
Licensing (NUREG-1860), Office of Nuclear Regulatory Research, United States Nuclear Regulatory
Commission, Washington, D.C.
Olshansky R.B., Johnson L.A. (2010). Clear as Mud: Planning for the Rebuilding of New Orleans, American
Planning Association, Chicago, IL.
Olshansky R.B., Johnson L.A., Topping K.C. (2011). Opportunity in Chaos: Rebuilding after the 1994 Northridge
and 1995 Kobe Earthquakes. Retrieved from http://www.urban.uiuc.edu/faculty/olshansky/chaos/Opportunity-
in-Chaos-March2011.pdf.
O’Rourke T.D. (2007). Critical infrastructure, interdependencies, and resilience, The Bridge, 37(1): 22–29.
Ouyang M., Hong L., Zijun M., Ming-Hui Y., Fei Q. (2009). A methodological approach to analyze vulnerability of
interdependent infrastructures, in: Simulation Modeling Practice and Theory, 17(5): 817–828.
Parry G.W. (1996). The characterization of uncertainty in probabilistic risk assessments of complex systems,
Reliability Engineering and System Safety, 54(2-3), 119–126.
Parry G.W., Winter P.W. (1981). Characterization and evaluation of uncertainty in probabilistic risk analysis,
Nuclear Safety, 22(1), 28–42.
PCCIP (1997). Critical Foundations: Protecting America’s Infrastructures, President’s Commission on Critical
Infrastructure Protection Retrieved from http://www.cyber.st.dhs.gov/docs/PCCIP%20Report%201997.pdf.
Poland C., Bonowitz D., Maffei J., Barkley,C. (2009). The resilient city: Part 1: Before the disaster, Urbanist, 479,
4–21.
Rinaldi S.M., Peerenboom J.P., Kelly T.K. (2001). Identifying, understanding, and analyzing critical infrastructure
interdependencies, IEEE Control Systems, 21(6): 11–25.
Rosato V., Tiriticco F., Issacharof L., De Porcellinis S., Setola R. (2008). Modeling interdependent infrastructures
using interacting dynamical models,. Int. J. Critical Infrastructures, 4(1/2): 63–79.
Scawthorn C. (2003a). Earthquake risk management: An overview, in: Earthquake Engineering Handbook. W.
Chen and C. Scawthorn, eds., CRC Press, Boca Raton, FL.
Scawthorn C. (2003b). Fire following earthquakes, in: Earthquake Engineering Handbook. W. Chen and C.
Scawthorn, eds., CRC Press, Boca Raton, FL.
SERRI, CARRI (2009). Creating Resilient Communities: The Work of SERRI and CARRI, Southeast Region
Research Initiative and Community and Regional Resilience Institute, Oak Ridge National Laboratory, Oak
Ridge, TN.
Sorenson J.N., Apostolakis G.E., Kress T.S., Powers D.A. (1999). On the role of defense in depth in risk-informed
regulation, Proceedings, PSA ’99: International Topical Meeting on Probabilistic Safety Assessment,
Washington, D.C.
SPUR (2012). Safe Enough to Stay. San Francisco Planning and Urban Research Association, San Francisco, CA.

138
Svendsen N., Wolthusen, S. (2007). Connectivity models of interdependency in mixed-type critical infrastructure
networks, Information Security Technical Report, 12(1): 44–55.
Thenhaus P.C., Campbell, K.W. (2003). Seismic hazard analysis, in: Earthquake Engineering Handbook. W. Chen
and C. Scawthorn, eds. CRC Press, Boca Raton, FL.
Twigg J. (2009). Characteristics of a Disaster-Resilient Community: A Guidance Note, 2nd ed., Disaster Risk
Reduction Interagency Coordination Group, London.
USNARA (2009). Local mitigation plans, Code of Federal Regulations, 44 CFR 201.6, U.S. National Archives and
Records Administration, Washington, DC.
USNRC (1986). Safety goals for the operations of nuclear power plants; Policy statement, United States Nuclear
Regulatory Commission, Federal Register, 51 FR 30028, Washington, D.C.
USNRC (1995). Use of probabilistic risk assessment methods in nuclear regulatory activities; Final policy
statement, United States Nuclear Regulatory Commission, Federal Register, 60 FR 42622, Washington, D.C.
USNRC (1998). White Paper on Risk-Informed, Performance-Based Regulation (SECY-98-144), United States
Nuclear Regulatory Commission, Washington, D.C.
USNRC (2002). An Approach for Using Probabilistic Risk Assessment in Risk-Informed Decisions on Plant-Specific
Changes to the Licensing Basis (Regulatory Guide 1.174), United States Nuclear Regulatory Commission,
Washington, D.C.
USNRC (2012a). Probabilistic Risk Assessment (PRA),” United States Nuclear Regulatory Commission, retrieved
from http://www.nrc.gov/about-nrc/regulatory/risk-informed/pra.html.
USNRC (2012b). Event Tree, United States Nuclear Regulatory Commission, Washington, D.C., retrieved from
http://www.nrc.gov/reading-rm/basic-ref/glossary/event-tree.html.
Vesely W.E., Goldberg F.F., Roberts N.H., Haasl D.F. (1981). Fault Tree Handbook (NUREG-0492). United States
Nuclear Regulatory Commission, Washington, D.C.
Wasson C.S. (2006). System Analysis, Design, and Development: Concepts, Principles, and Practices, John Wiley
& Sons, Hoboken, NJ.
WCED (1987). Our Common Future, World Commission on Environment and Development Oxford University
Press, New York, NY.
Wisner B., Davis I., Blaikie P., Cannon T. (2003). At Risk: Natural Hazards, People’s Vulnerability and Disasters,
Routledge, London.
Yelvington K.A. (1997). Coping in a temporary way: The tent cities, in: Hurricane Andrew: Ethnicity, Gender, and
the Sociology of Disasters. W.G. Peacock, B.H. Morrow, and H. Gladwin, eds. Routledge, New York, NY.
Zhang P., Peeta S., Friesz T. (2005). Dynamic game theoretic model of multi-layer infrastructure networks,
Networks and Spatial Economics, 5(2): 147–178.

139
140
 PEER REPORTS

PEER reports are available as a free PDF download from http://peer.berkeley.edu/publications/peer_reports_complete.html. Printed
hard copies of PEER reports can be ordered directly from our printer by following the instructions at
http://peer.berkeley.edu/publications/peer_reports.html. For other related questions about the PEER Report Series, contact the
Pacific Earthquake Engineering Research Center, 325 Davis Hall mail code 1792, Berkeley, CA 94720. Tel.: (510) 642-3437; Fax:
(510) 665-1655; Email: peer_editor@berkeley.edu

PEER 2013/19 Toward Resilient Communities: A Performance-Based Engineering Framework for Design and Evaluation of the
Built Environment. Michael William Mieler, Bozidar Stojadinovic, Robert J. Budnitz, Stephen A. Mahin and Mary C.
Comerio. September 2013.

PEER 2013/18 Identification of Site Parameters that Improve Predictions of Site Amplification. Ellen M. Rathje and Sara Navidi.
July 2013.

PEER 2013/17 Response Spectrum Analysis of Concrete Gravity Dams Including Dam-Water-Foundation Interaction. Arnkjell
Løkke and Anil K. Chopra. July 2013.

PEER 2013/16 Effect of hoop reinforcement spacing on the cyclic response of large reinforced concrete special moment frame
beams. Marios Panagiotou, Tea Visnjic, Grigorios Antonellis, Panagiotis Galanis, and Jack P. Moehle. June 2013.

PEER 2013/15 publication pending

PEER 2013/14 publication pending

PEER 2013/13 publication pending

PEER 2013/12 Nonlinear Horizontal Site Response for the NGA-West2 Project. Ronnie Kamai, Norman A. Abramson, Walter J.
Silva. May 2013.

PEER 2013/11 Epistemic Uncertainty for NGA-West2 Models. Linda Al Atik and Robert R. Youngs. May 2013.

PEER 2013/10 NGA-West 2 Models for Ground-Motion Directionality. Shrey K. Shahi and Jack W. Baker. May 2013.

PEER 2013/09 Final Report of the NGA-West2 Directivity Working Group. Paul Spudich, Jeffrey R. Bayless, Jack W. Baker, Brian
S.J. Chiou, Badie Rowshandel, Shrey Shahi, and Paul Somerville. May 2013.

PEER 2013/08 NGA-West2 Model for Estimating Average Horizontal Values of Pseudo-Absolute Spectral Accelerations
Generated by Crustal Earthquakes. I. M. Idriss. May 2013.

PEER 2013/07 Update of the Chiou and Youngs NGA Ground Motion Model for Average Horizontal Component of Peak Ground
Motion and Response Spectra. Brian Chiou and Robert Youngs. May 2013.

PEER 2013/06 NGA-West2 Campbell-Bozorgnia Ground Motion Model for the Horizontal Components of PGA, PGV, and 5%-
Damped Elastic Pseudo-Acceleration Response Spectra for Periods Ranging from 0.01 to 10 sec. Kenneth W.
Campbell and Yousef Bozorgnia. May 2013.

PEER 2013/05 NGA-West 2 Equations for Predicting Response Spectral Accelerations for Shallow Crustal Earthquakes. David
M. Boore, Jonathan P. Stewart, Emel Seyhan, Gail M. Atkinson. May 2013.

PEER 2013/04 Update of the AS08 Ground-Motion Prediction Equations Based on the NGA-West2 Data Set. Norman
Abrahamson, Walter Silva, and Ronnie Kamai. May 2013.

PEER 2013/03 PEER NGA-West2 Database. Timothy D. Ancheta, Robert B. Darragh, Jonathan P. Stewart, Emel Seyhan, Walter
J. Silva, Brian S.J. Chiou, Katie E. Wooddell, Robert W. Graves, Albert R. Kottke, David M. Boore, Tadahiro
Kishida, and Jennifer L. Donahue. May 2013.

PEER 2013/02 Hybrid Simulation of the Seismic Response of Squat Reinforced Concrete Shear Walls. Catherine A. Whyte and
Bozidar Stojadinovic. May 2013.

PEER 2013/01 Housing Recovery in Chile: A Qualitative Mid-program Review. Mary C. Comerio. February 2013.

PEER 2012/08 Guidelines for Estimation of Shear Wave Velocity. Bernard R. Wair, Jason T. DeJong, and Thomas Shantz.
December 2012.

PEER 2012/07 Earthquake Engineering for Resilient Communities: 2012 PEER Internship Program Research Report Collection.
Heidi Tremayne (Editor), Stephen A. Mahin (Editor), Collin Anderson, Dustin Cook, Michael Erceg, Carlos
Esparza, Jose Jimenez, Dorian Krausz, Andrew Lo, Stephanie Lopez, Nicole McCurdy, Paul Shipman, Alexander
Strum, Eduardo Vega. December 2012.
PEER 2012/06 Fragilities for Precarious Rocks at Yucca Mountain. Matthew D. Purvance, Rasool Anooshehpoor, and James N.
Brune. December 2012.

PEER 2012/05 Development of Simplified Analysis Procedure for Piles in Laterally Spreading Layered Soils. Christopher R.
McGann, Pedro Arduino, and Peter Mackenzie–Helnwein. December 2012.

PEER 2012/04 Unbonded Pre-Tensioned Columns for Bridges in Seismic Regions. Phillip M. Davis, Todd M. Janes, Marc O.
Eberhard, and John F. Stanton. December 2012.

PEER 2012/03 Experimental and Analytical Studies on Reinforced Concrete Buildings with Seismically Vulnerable Beam-Column
Joints. Sangjoon Park and Khalid M. Mosalam. October 2012.

PEER 2012/02 Seismic Performance of Reinforced Concrete Bridges Allowed to Uplift during Multi-Directional Excitation. Andres
Oscar Espinoza and Stephen A. Mahin. July 2012.

PEER 2012/01 Spectral Damping Scaling Factors for Shallow Crustal Earthquakes in Active Tectonic Regions. Sanaz Rezaeian,
Yousef Bozorgnia, I. M. Idriss, Kenneth Campbell, Norman Abrahamson, and Walter Silva. July 2012.

PEER 2011/10 Earthquake Engineering for Resilient Communities: 2011 PEER Internship Program Research Report Collection.
Eds. Heidi Faison and Stephen A. Mahin. December 2011.

PEER 2011/09 Calibration of Semi-Stochastic Procedure for Simulating High-Frequency Ground Motions. Jonathan P. Stewart,
Emel Seyhan, and Robert W. Graves. December 2011.

PEER 2011/08 Water Supply in regard to Fire Following Earthquake. Charles Scawthorn. November 2011.

PEER 2011/07 Seismic Risk Management in Urban Areas. Proceedings of a U.S.-Iran-Turkey Seismic Workshop. September
2011.

PEER 2011/06 The Use of Base Isolation Systems to Achieve Complex Seismic Performance Objectives. Troy A. Morgan and
Stephen A. Mahin. July 2011.

PEER 2011/05 Case Studies of the Seismic Performance of Tall Buildings Designed by Alternative Means. Task 12 Report for
the Tall Buildings Initiative. Jack Moehle, Yousef Bozorgnia, Nirmal Jayaram, Pierson Jones, Mohsen Rahnama,
Nilesh Shome, Zeynep Tuna, John Wallace, Tony Yang, and Farzin Zareian. July 2011.

PEER 2011/04 Recommended Design Practice for Pile Foundations in Laterally Spreading Ground. Scott A. Ashford, Ross W.
Boulanger, and Scott J. Brandenberg. June 2011.

PEER 2011/03 New Ground Motion Selection Procedures and Selected Motions for the PEER Transportation Research Program.
Jack W. Baker, Ting Lin, Shrey K. Shahi, and Nirmal Jayaram. March 2011.

PEER 2011/02 A Bayesian Network Methodology for Infrastructure Seismic Risk Assessment and Decision Support. Michelle T.
Bensi, Armen Der Kiureghian, and Daniel Straub. March 2011.

PEER 2011/01 Demand Fragility Surfaces for Bridges in Liquefied and Laterally Spreading Ground. Scott J. Brandenberg, Jian
Zhang, Pirooz Kashighandi, Yili Huo, and Minxing Zhao. March 2011.

PEER 2010/05 Guidelines for Performance-Based Seismic Design of Tall Buildings. Developed by the Tall Buildings Initiative.
November 2010.
PEER 2010/04 Application Guide for the Design of Flexible and Rigid Bus Connections between Substation Equipment Subjected
to Earthquakes. Jean-Bernard Dastous and Armen Der Kiureghian. September 2010.
PEER 2010/03 Shear Wave Velocity as a Statistical Function of Standard Penetration Test Resistance and Vertical Effective
Stress at Caltrans Bridge Sites. Scott J. Brandenberg, Naresh Bellana, and Thomas Shantz. June 2010.
PEER 2010/02 Stochastic Modeling and Simulation of Ground Motions for Performance-Based Earthquake Engineering. Sanaz
Rezaeian and Armen Der Kiureghian. June 2010.
PEER 2010/01 Structural Response and Cost Characterization of Bridge Construction Using Seismic Performance Enhancement
Strategies. Ady Aviram, Božidar Stojadinović, Gustavo J. Parra-Montesinos, and Kevin R. Mackie. March 2010.
PEER 2009/03 The Integration of Experimental and Simulation Data in the Study of Reinforced Concrete Bridge Systems
Including Soil-Foundation-Structure Interaction. Matthew Dryden and Gregory L. Fenves. November 2009.

PEER 2009/02 Improving Earthquake Mitigation through Innovations and Applications in Seismic Science, Engineering,
Communication, and Response. Proceedings of a U.S.-Iran Seismic Workshop. October 2009.

PEER 2009/01 Evaluation of Ground Motion Selection and Modification Methods: Predicting Median Interstory Drift Response of
Buildings. Curt B. Haselton, Ed. June 2009.

PEER 2008/10 Technical Manual for Strata. Albert R. Kottke and Ellen M. Rathje. February 2009.

PEER 2008/09 NGA Model for Average Horizontal Component of Peak Ground Motion and Response Spectra. Brian S.-J. Chiou
and Robert R. Youngs. November 2008.
PEER 2008/08 Toward Earthquake-Resistant Design of Concentrically Braced Steel Structures. Patxi Uriz and Stephen A. Mahin.
November 2008.

PEER 2008/07 Using OpenSees for Performance-Based Evaluation of Bridges on Liquefiable Soils. Stephen L. Kramer, Pedro
Arduino, and HyungSuk Shin. November 2008.

PEER 2008/06 Shaking Table Tests and Numerical Investigation of Self-Centering Reinforced Concrete Bridge Columns. Hyung
IL Jeong, Junichi Sakai, and Stephen A. Mahin. September 2008.

PEER 2008/05 Performance-Based Earthquake Engineering Design Evaluation Procedure for Bridge Foundations Undergoing
Liquefaction-Induced Lateral Ground Displacement. Christian A. Ledezma and Jonathan D. Bray. August 2008.

PEER 2008/04 Benchmarking of Nonlinear Geotechnical Ground Response Analysis Procedures. Jonathan P. Stewart, Annie
On-Lei Kwok, Yousseff M. A. Hashash, Neven Matasovic, Robert Pyke, Zhiliang Wang, and Zhaohui Yang.
August 2008.

PEER 2008/03 Guidelines for Nonlinear Analysis of Bridge Structures in California. Ady Aviram, Kevin R. Mackie, and Božidar
Stojadinović. August 2008.

PEER 2008/02 Treatment of Uncertainties in Seismic-Risk Analysis of Transportation Systems. Evangelos Stergiou and Anne S.
Kiremidjian. July 2008.

PEER 2008/01 Seismic Performance Objectives for Tall Buildings. William T. Holmes, Charles Kircher, William Petak, and Nabih
Youssef. August 2008.

PEER 2007/12 An Assessment to Benchmark the Seismic Performance of a Code-Conforming Reinforced Concrete Moment-
Frame Building. Curt Haselton, Christine A. Goulet, Judith Mitrani-Reiser, James L. Beck, Gregory G. Deierlein,
Keith A. Porter, Jonathan P. Stewart, and Ertugrul Taciroglu. August 2008.

PEER 2007/11 Bar Buckling in Reinforced Concrete Bridge Columns. Wayne A. Brown, Dawn E. Lehman, and John F. Stanton.
February 2008.

PEER 2007/10 Computational Modeling of Progressive Collapse in Reinforced Concrete Frame Structures. Mohamed M. Talaat
and Khalid M. Mosalam. May 2008.

PEER 2007/09 Integrated Probabilistic Performance-Based Evaluation of Benchmark Reinforced Concrete Bridges. Kevin R.
Mackie, John-Michael Wong, and Božidar Stojadinović. January 2008.

PEER 2007/08 Assessing Seismic Collapse Safety of Modern Reinforced Concrete Moment-Frame Buildings. Curt B. Haselton
and Gregory G. Deierlein. February 2008.

PEER 2007/07 Performance Modeling Strategies for Modern Reinforced Concrete Bridge Columns. Michael P. Berry and Marc
O. Eberhard. April 2008.

PEER 2007/06 Development of Improved Procedures for Seismic Design of Buried and Partially Buried Structures. Linda Al Atik
and Nicholas Sitar. June 2007.

PEER 2007/05 Uncertainty and Correlation in Seismic Risk Assessment of Transportation Systems. Renee G. Lee and Anne S.
Kiremidjian. July 2007.

PEER 2007/04 Numerical Models for Analysis and Performance-Based Design of Shallow Foundations Subjected to Seismic
Loading. Sivapalan Gajan, Tara C. Hutchinson, Bruce L. Kutter, Prishati Raychowdhury, José A. Ugalde, and
Jonathan P. Stewart. May 2008.

PEER 2007/03 Beam-Column Element Model Calibrated for Predicting Flexural Response Leading to Global Collapse of RC
Frame Buildings. Curt B. Haselton, Abbie B. Liel, Sarah Taylor Lange, and Gregory G. Deierlein. May 2008.

PEER 2007/02 Campbell-Bozorgnia NGA Ground Motion Relations for the Geometric Mean Horizontal Component of Peak and
Spectral Ground Motion Parameters. Kenneth W. Campbell and Yousef Bozorgnia. May 2007.

PEER 2007/01 Boore-Atkinson NGA Ground Motion Relations for the Geometric Mean Horizontal Component of Peak and
Spectral Ground Motion Parameters. David M. Boore and Gail M. Atkinson. May. May 2007.

PEER 2006/12 Societal Implications of Performance-Based Earthquake Engineering. Peter J. May. May 2007.

PEER 2006/11 Probabilistic Seismic Demand Analysis Using Advanced Ground Motion Intensity Measures, Attenuation
Relationships, and Near-Fault Effects. Polsak Tothong and C. Allin Cornell. March 2007.

PEER 2006/10 Application of the PEER PBEE Methodology to the I-880 Viaduct. Sashi Kunnath. February 2007.

PEER 2006/09 Quantifying Economic Losses from Travel Forgone Following a Large Metropolitan Earthquake. James Moore,
Sungbin Cho, Yue Yue Fan, and Stuart Werner. November 2006.

PEER 2006/08 Vector-Valued Ground Motion Intensity Measures for Probabilistic Seismic Demand Analysis. Jack W. Baker and
C. Allin Cornell. October 2006.
PEER 2006/07 Analytical Modeling of Reinforced Concrete Walls for Predicting Flexural and Coupled–Shear-
Flexural Responses. Kutay Orakcal, Leonardo M. Massone, and John W. Wallace. October 2006.

PEER 2006/06 Nonlinear Analysis of a Soil-Drilled Pier System under Static and Dynamic Axial Loading. Gang Wang and
Nicholas Sitar. November 2006.

PEER 2006/05 Advanced Seismic Assessment Guidelines. Paolo Bazzurro, C. Allin Cornell, Charles Menun, Maziar Motahari,
and Nicolas Luco. September 2006.

PEER 2006/04 Probabilistic Seismic Evaluation of Reinforced Concrete Structural Components and Systems. Tae Hyung Lee
and Khalid M. Mosalam. August 2006.

PEER 2006/03 Performance of Lifelines Subjected to Lateral Spreading. Scott A. Ashford and Teerawut Juirnarongrit. July 2006.

PEER 2006/02 Pacific Earthquake Engineering Research Center Highway Demonstration Project. Anne Kiremidjian, James
Moore, Yue Yue Fan, Nesrin Basoz, Ozgur Yazali, and Meredith Williams. April 2006.

PEER 2006/01 Bracing Berkeley. A Guide to Seismic Safety on the UC Berkeley Campus. Mary C. Comerio, Stephen Tobriner,
and Ariane Fehrenkamp. January 2006.

PEER 2005/16 Seismic Response and Reliability of Electrical Substation Equipment and Systems. Junho Song, Armen Der
Kiureghian, and Jerome L. Sackman. April 2006.

PEER 2005/15 CPT-Based Probabilistic Assessment of Seismic Soil Liquefaction Initiation. R. E. S. Moss, R. B. Seed, R. E.
Kayen, J. P. Stewart, and A. Der Kiureghian. April 2006.

PEER 2005/14 Workshop on Modeling of Nonlinear Cyclic Load-Deformation Behavior of Shallow Foundations. Bruce L. Kutter,
Geoffrey Martin, Tara Hutchinson, Chad Harden, Sivapalan Gajan, and Justin Phalen. March 2006.

PEER 2005/13 Stochastic Characterization and Decision Bases under Time-Dependent Aftershock Risk in Performance-Based
Earthquake Engineering. Gee Liek Yeo and C. Allin Cornell. July 2005.

PEER 2005/12 PEER Testbed Study on a Laboratory Building: Exercising Seismic Performance Assessment. Mary C. Comerio,
editor. November 2005.

PEER 2005/11 Van Nuys Hotel Building Testbed Report: Exercising Seismic Performance Assessment. Helmut Krawinkler,
editor. October 2005.

PEER 2005/10 First NEES/E-Defense Workshop on Collapse Simulation of Reinforced Concrete Building Structures. September
2005.

PEER 2005/09 Test Applications of Advanced Seismic Assessment Guidelines. Joe Maffei, Karl Telleen, Danya Mohr, William
Holmes, and Yuki Nakayama. August 2006.

PEER 2005/08 Damage Accumulation in Lightly Confined Reinforced Concrete Bridge Columns. R. Tyler Ranf, Jared M. Nelson,
Zach Price, Marc O. Eberhard, and John F. Stanton. April 2006.

PEER 2005/07 Experimental and Analytical Studies on the Seismic Response of Freestanding and Anchored Laboratory
Equipment. Dimitrios Konstantinidis and Nicos Makris. January 2005.

PEER 2005/06 Global Collapse of Frame Structures under Seismic Excitations. Luis F. Ibarra and Helmut Krawinkler. September
2005.

PEER 2005//05 Performance Characterization of Bench- and Shelf-Mounted Equipment. Samit Ray Chaudhuri and Tara C.
Hutchinson. May 2006.

PEER 2005/04 Numerical Modeling of the Nonlinear Cyclic Response of Shallow Foundations. Chad Harden, Tara Hutchinson,
Geoffrey R. Martin, and Bruce L. Kutter. August 2005.

PEER 2005/03 A Taxonomy of Building Components for Performance-Based Earthquake Engineering. Keith A. Porter.
September 2005.

PEER 2005/02 Fragility Basis for California Highway Overpass Bridge Seismic Decision Making. Kevin R. Mackie and Božidar
Stojadinović. June 2005.

PEER 2005/01 Empirical Characterization of Site Conditions on Strong Ground Motion. Jonathan P. Stewart, Yoojoong Choi,
and Robert W. Graves. June 2005.

PEER 2004/09 Electrical Substation Equipment Interaction: Experimental Rigid Conductor Studies. Christopher Stearns and
André Filiatrault. February 2005.

PEER 2004/08 Seismic Qualification and Fragility Testing of Line Break 550-kV Disconnect Switches. Shakhzod M. Takhirov,
Gregory L. Fenves, and Eric Fujisaki. January 2005.

PEER 2004/07 Ground Motions for Earthquake Simulator Qualification of Electrical Substation Equipment. Shakhzod M.
Takhirov, Gregory L. Fenves, Eric Fujisaki, and Don Clyde. January 2005.
PEER 2004/06 Performance-Based Regulation and Regulatory Regimes. Peter J. May and Chris Koski. September 2004.

PEER 2004/05 Performance-Based Seismic Design Concepts and Implementation: Proceedings of an International Workshop.
Peter Fajfar and Helmut Krawinkler, editors. September 2004.

PEER 2004/04 Seismic Performance of an Instrumented Tilt-up Wall Building. James C. Anderson and Vitelmo V. Bertero. July
2004.

PEER 2004/03 Evaluation and Application of Concrete Tilt-up Assessment Methodologies. Timothy Graf and James O. Malley.
October 2004.

PEER 2004/02 Analytical Investigations of New Methods for Reducing Residual Displacements of Reinforced Concrete Bridge
Columns. Junichi Sakai and Stephen A. Mahin. August 2004.

PEER 2004/01 Seismic Performance of Masonry Buildings and Design Implications. Kerri Anne Taeko Tokoro, James C.
Anderson, and Vitelmo V. Bertero. February 2004.

PEER 2003/18 Performance Models for Flexural Damage in Reinforced Concrete Columns. Michael Berry and Marc Eberhard.
August 2003.

PEER 2003/17 Predicting Earthquake Damage in Older Reinforced Concrete Beam-Column Joints. Catherine Pagni and Laura
Lowes. October 2004.

PEER 2003/16 Seismic Demands for Performance-Based Design of Bridges. Kevin Mackie and Božidar Stojadinović. August
2003.

PEER 2003/15 Seismic Demands for Nondeteriorating Frame Structures and Their Dependence on Ground Motions. Ricardo
Antonio Medina and Helmut Krawinkler. May 2004.

PEER 2003/14 Finite Element Reliability and Sensitivity Methods for Performance-Based Earthquake Engineering. Terje
Haukaas and Armen Der Kiureghian. April 2004.

PEER 2003/13 Effects of Connection Hysteretic Degradation on the Seismic Behavior of Steel Moment-Resisting Frames. Janise
E. Rodgers and Stephen A. Mahin. March 2004.

PEER 2003/12 Implementation Manual for the Seismic Protection of Laboratory Contents: Format and Case Studies. William T.
Holmes and Mary C. Comerio. October 2003.

PEER 2003/11 Fifth U.S.-Japan Workshop on Performance-Based Earthquake Engineering Methodology for Reinforced
Concrete Building Structures. February 2004.

PEER 2003/10 A Beam-Column Joint Model for Simulating the Earthquake Response of Reinforced Concrete Frames. Laura N.
Lowes, Nilanjan Mitra, and Arash Altoontash. February 2004.

PEER 2003/09 Sequencing Repairs after an Earthquake: An Economic Approach. Marco Casari and Simon J. Wilkie. April 2004.

PEER 2003/08 A Technical Framework for Probability-Based Demand and Capacity Factor Design (DCFD) Seismic Formats.
Fatemeh Jalayer and C. Allin Cornell. November 2003.

PEER 2003/07 Uncertainty Specification and Propagation for Loss Estimation Using FOSM Methods. Jack W. Baker and C. Allin
Cornell. September 2003.

PEER 2003/06 Performance of Circular Reinforced Concrete Bridge Columns under Bidirectional Earthquake Loading. Mahmoud
M. Hachem, Stephen A. Mahin, and Jack P. Moehle. February 2003.

PEER 2003/05 Response Assessment for Building-Specific Loss Estimation. Eduardo Miranda and Shahram Taghavi.
September 2003.

PEER 2003/04 Experimental Assessment of Columns with Short Lap Splices Subjected to Cyclic Loads. Murat Melek, John W.
Wallace, and Joel Conte. April 2003.

PEER 2003/03 Probabilistic Response Assessment for Building-Specific Loss Estimation. Eduardo Miranda and Hesameddin
Aslani. September 2003.

PEER 2003/02 Software Framework for Collaborative Development of Nonlinear Dynamic Analysis Program. Jun Peng and
Kincho H. Law. September 2003.

PEER 2003/01 Shake Table Tests and Analytical Studies on the Gravity Load Collapse of Reinforced Concrete Frames. Kenneth
John Elwood and Jack P. Moehle. November 2003.

PEER 2002/24 Performance of Beam to Column Bridge Joints Subjected to a Large Velocity Pulse. Natalie Gibson, André
Filiatrault, and Scott A. Ashford. April 2002.

PEER 2002/23 Effects of Large Velocity Pulses on Reinforced Concrete Bridge Columns. Greg L. Orozco and Scott A. Ashford.
April 2002.
PEER 2002/22 Characterization of Large Velocity Pulses for Laboratory Testing. Kenneth E. Cox and Scott A. Ashford. April
2002.

PEER 2002/21 Fourth U.S.-Japan Workshop on Performance-Based Earthquake Engineering Methodology for Reinforced
Concrete Building Structures. December 2002.

PEER 2002/20 Barriers to Adoption and Implementation of PBEE Innovations. Peter J. May. August 2002.

PEER 2002/19 Economic-Engineered Integrated Models for Earthquakes: Socioeconomic Impacts. Peter Gordon, James E.
Moore II, and Harry W. Richardson. July 2002.

PEER 2002/18 Assessment of Reinforced Concrete Building Exterior Joints with Substandard Details. Chris P. Pantelides, Jon
Hansen, Justin Nadauld, and Lawrence D. Reaveley. May 2002.

PEER 2002/17 Structural Characterization and Seismic Response Analysis of a Highway Overcrossing Equipped with
Elastomeric Bearings and Fluid Dampers: A Case Study. Nicos Makris and Jian Zhang. November 2002.

PEER 2002/16 Estimation of Uncertainty in Geotechnical Properties for Performance-Based Earthquake Engineering. Allen L.
Jones, Steven L. Kramer, and Pedro Arduino. December 2002.

PEER 2002/15 Seismic Behavior of Bridge Columns Subjected to Various Loading Patterns. Asadollah Esmaeily-Gh. and Yan
Xiao. December 2002.

PEER 2002/14 Inelastic Seismic Response of Extended Pile Shaft Supported Bridge Structures. T.C. Hutchinson, R.W.
Boulanger, Y.H. Chai, and I.M. Idriss. December 2002.

PEER 2002/13 Probabilistic Models and Fragility Estimates for Bridge Components and Systems. Paolo Gardoni, Armen Der
Kiureghian, and Khalid M. Mosalam. June 2002.

PEER 2002/12 Effects of Fault Dip and Slip Rake on Near-Source Ground Motions: Why Chi-Chi Was a Relatively Mild M7.6
Earthquake. Brad T. Aagaard, John F. Hall, and Thomas H. Heaton. December 2002.

PEER 2002/11 Analytical and Experimental Study of Fiber-Reinforced Strip Isolators. James M. Kelly and Shakhzod M. Takhirov.
September 2002.

PEER 2002/10 Centrifuge Modeling of Settlement and Lateral Spreading with Comparisons to Numerical Analyses. Sivapalan
Gajan and Bruce L. Kutter. January 2003.

PEER 2002/09 Documentation and Analysis of Field Case Histories of Seismic Compression during the 1994 Northridge,
California, Earthquake. Jonathan P. Stewart, Patrick M. Smith, Daniel H. Whang, and Jonathan D. Bray. October
2002.
TM
PEER 2002/08 Component Testing, Stability Analysis and Characterization of Buckling-Restrained Unbonded Braces .
Cameron Black, Nicos Makris, and Ian Aiken. September 2002.

PEER 2002/07 Seismic Performance of Pile-Wharf Connections. Charles W. Roeder, Robert Graff, Jennifer Soderstrom, and Jun
Han Yoo. December 2001.

PEER 2002/06 The Use of Benefit-Cost Analysis for Evaluation of Performance-Based Earthquake Engineering Decisions.
Richard O. Zerbe and Anthony Falit-Baiamonte. September 2001.

PEER 2002/05 Guidelines, Specifications, and Seismic Performance Characterization of Nonstructural Building Components and
Equipment. André Filiatrault, Constantin Christopoulos, and Christopher Stearns. September 2001.

PEER 2002/04 Consortium of Organizations for Strong-Motion Observation Systems and the Pacific Earthquake Engineering
Research Center Lifelines Program: Invited Workshop on Archiving and Web Dissemination of Geotechnical
Data, 4–5 October 2001. September 2002.

PEER 2002/03 Investigation of Sensitivity of Building Loss Estimates to Major Uncertain Variables for the Van Nuys Testbed.
Keith A. Porter, James L. Beck, and Rustem V. Shaikhutdinov. August 2002.

PEER 2002/02 The Third U.S.-Japan Workshop on Performance-Based Earthquake Engineering Methodology for Reinforced
Concrete Building Structures. July 2002.

PEER 2002/01 Nonstructural Loss Estimation: The UC Berkeley Case Study. Mary C. Comerio and John C. Stallmeyer.
December 2001.

PEER 2001/16 Statistics of SDF-System Estimate of Roof Displacement for Pushover Analysis of Buildings. Anil K. Chopra,
Rakesh K. Goel, and Chatpan Chintanapakdee. December 2001.

PEER 2001/15 Damage to Bridges during the 2001 Nisqually Earthquake. R. Tyler Ranf, Marc O. Eberhard, and Michael P.
Berry. November 2001.

PEER 2001/14 Rocking Response of Equipment Anchored to a Base Foundation. Nicos Makris and Cameron J. Black.
September 2001.
PEER 2001/13 Modeling Soil Liquefaction Hazards for Performance-Based Earthquake Engineering. Steven L. Kramer and
Ahmed-W. Elgamal. February 2001.

PEER 2001/12 Development of Geotechnical Capabilities in OpenSees. Boris Jeremić. September 2001.

PEER 2001/11 Analytical and Experimental Study of Fiber-Reinforced Elastomeric Isolators. James M. Kelly and Shakhzod M.
Takhirov. September 2001.

PEER 2001/10 Amplification Factors for Spectral Acceleration in Active Regions. Jonathan P. Stewart, Andrew H. Liu, Yoojoong
Choi, and Mehmet B. Baturay. December 2001.

PEER 2001/09 Ground Motion Evaluation Procedures for Performance-Based Design. Jonathan P. Stewart, Shyh-Jeng Chiou,
Jonathan D. Bray, Robert W. Graves, Paul G. Somerville, and Norman A. Abrahamson. September 2001.

PEER 2001/08 Experimental and Computational Evaluation of Reinforced Concrete Bridge Beam-Column Connections for
Seismic Performance. Clay J. Naito, Jack P. Moehle, and Khalid M. Mosalam. November 2001.

PEER 2001/07 The Rocking Spectrum and the Shortcomings of Design Guidelines. Nicos Makris and Dimitrios Konstantinidis.
August 2001.

PEER 2001/06 Development of an Electrical Substation Equipment Performance Database for Evaluation of Equipment
Fragilities. Thalia Agnanos. April 1999.

PEER 2001/05 Stiffness Analysis of Fiber-Reinforced Elastomeric Isolators. Hsiang-Chuan Tsai and James M. Kelly. May 2001.

PEER 2001/04 Organizational and Societal Considerations for Performance-Based Earthquake Engineering. Peter J. May. April
2001.

PEER 2001/03 A Modal Pushover Analysis Procedure to Estimate Seismic Demands for Buildings: Theory and Preliminary
Evaluation. Anil K. Chopra and Rakesh K. Goel. January 2001.

PEER 2001/02 Seismic Response Analysis of Highway Overcrossings Including Soil-Structure Interaction. Jian Zhang and Nicos
Makris. March 2001.

PEER 2001/01 Experimental Study of Large Seismic Steel Beam-to-Column Connections. Egor P. Popov and Shakhzod M.
Takhirov. November 2000.

PEER 2000/10 The Second U.S.-Japan Workshop on Performance-Based Earthquake Engineering Methodology for Reinforced
Concrete Building Structures. March 2000.

PEER 2000/09 Structural Engineering Reconnaissance of the August 17, 1999 Earthquake: Kocaeli (Izmit), Turkey. Halil Sezen,
Kenneth J. Elwood, Andrew S. Whittaker, Khalid Mosalam, John J. Wallace, and John F. Stanton. December
2000.

PEER 2000/08 Behavior of Reinforced Concrete Bridge Columns Having Varying Aspect Ratios and Varying Lengths of
Confinement. Anthony J. Calderone, Dawn E. Lehman, and Jack P. Moehle. January 2001.

PEER 2000/07 Cover-Plate and Flange-Plate Reinforced Steel Moment-Resisting Connections. Taejin Kim, Andrew S. Whittaker,
Amir S. Gilani, Vitelmo V. Bertero, and Shakhzod M. Takhirov. September 2000.

PEER 2000/06 Seismic Evaluation and Analysis of 230-kV Disconnect Switches. Amir S. J. Gilani, Andrew S. Whittaker, Gregory
L. Fenves, Chun-Hao Chen, Henry Ho, and Eric Fujisaki. July 2000.

PEER 2000/05 Performance-Based Evaluation of Exterior Reinforced Concrete Building Joints for Seismic Excitation. Chandra
Clyde, Chris P. Pantelides, and Lawrence D. Reaveley. July 2000.

PEER 2000/04 An Evaluation of Seismic Energy Demand: An Attenuation Approach. Chung-Che Chou and Chia-Ming Uang. July
1999.

PEER 2000/03 Framing Earthquake Retrofitting Decisions: The Case of Hillside Homes in Los Angeles. Detlof von Winterfeldt,
Nels Roselund, and Alicia Kitsuse. March 2000.

PEER 2000/02 U.S.-Japan Workshop on the Effects of Near-Field Earthquake Shaking. Andrew Whittaker, ed. July 2000.

PEER 2000/01 Further Studies on Seismic Interaction in Interconnected Electrical Substation Equipment. Armen Der Kiureghian,
Kee-Jeung Hong, and Jerome L. Sackman. November 1999.

PEER 1999/14 Seismic Evaluation and Retrofit of 230-kV Porcelain Transformer Bushings. Amir S. Gilani, Andrew S. Whittaker,
Gregory L. Fenves, and Eric Fujisaki. December 1999.

PEER 1999/13 Building Vulnerability Studies: Modeling and Evaluation of Tilt-up and Steel Reinforced Concrete Buildings. John
W. Wallace, Jonathan P. Stewart, and Andrew S. Whittaker, editors. December 1999.

PEER 1999/12 Rehabilitation of Nonductile RC Frame Building Using Encasement Plates and Energy-Dissipating Devices.
Mehrdad Sasani, Vitelmo V. Bertero, James C. Anderson. December 1999.
PEER 1999/11 Performance Evaluation Database for Concrete Bridge Components and Systems under Simulated Seismic
Loads. Yael D. Hose and Frieder Seible. November 1999.

PEER 1999/10 U.S.-Japan Workshop on Performance-Based Earthquake Engineering Methodology for Reinforced Concrete
Building Structures. December 1999.

PEER 1999/09 Performance Improvement of Long Period Building Structures Subjected to Severe Pulse-Type Ground Motions.
James C. Anderson, Vitelmo V. Bertero, and Raul Bertero. October 1999.

PEER 1999/08 Envelopes for Seismic Response Vectors. Charles Menun and Armen Der Kiureghian. July 1999.

PEER 1999/07 Documentation of Strengths and Weaknesses of Current Computer Analysis Methods for Seismic Performance of
Reinforced Concrete Members. William F. Cofer. November 1999.

PEER 1999/06 Rocking Response and Overturning of Anchored Equipment under Seismic Excitations. Nicos Makris and Jian
Zhang. November 1999.

PEER 1999/05 Seismic Evaluation of 550 kV Porcelain Transformer Bushings. Amir S. Gilani, Andrew S. Whittaker, Gregory L.
Fenves, and Eric Fujisaki. October 1999.

PEER 1999/04 Adoption and Enforcement of Earthquake Risk-Reduction Measures. Peter J. May, Raymond J. Burby, T. Jens
Feeley, and Robert Wood.

PEER 1999/03 Task 3 Characterization of Site Response General Site Categories. Adrian Rodriguez-Marek, Jonathan D. Bray,
and Norman Abrahamson. February 1999.

PEER 1999/02 Capacity-Demand-Diagram Methods for Estimating Seismic Deformation of Inelastic Structures: SDF Systems.
Anil K. Chopra and Rakesh Goel. April 1999.

PEER 1999/01 Interaction in Interconnected Electrical Substation Equipment Subjected to Earthquake Ground Motions. Armen
Der Kiureghian, Jerome L. Sackman, and Kee-Jeung Hong. February 1999.

PEER 1998/08 Behavior and Failure Analysis of a Multiple-Frame Highway Bridge in the 1994 Northridge Earthquake. Gregory L.
Fenves and Michael Ellery. December 1998.

PEER 1998/07 Empirical Evaluation of Inertial Soil-Structure Interaction Effects. Jonathan P. Stewart, Raymond B. Seed, and
Gregory L. Fenves. November 1998.

PEER 1998/06 Effect of Damping Mechanisms on the Response of Seismic Isolated Structures. Nicos Makris and Shih-Po
Chang. November 1998.

PEER 1998/05 Rocking Response and Overturning of Equipment under Horizontal Pulse-Type Motions. Nicos Makris and
Yiannis Roussos. October 1998.

PEER 1998/04 Pacific Earthquake Engineering Research Invitational Workshop Proceedings, May 14–15, 1998: Defining the
Links between Planning, Policy Analysis, Economics and Earthquake Engineering. Mary Comerio and Peter
Gordon. September 1998.

PEER 1998/03 Repair/Upgrade Procedures for Welded Beam to Column Connections. James C. Anderson and Xiaojing Duan.
May 1998.

PEER 1998/02 Seismic Evaluation of 196 kV Porcelain Transformer Bushings. Amir S. Gilani, Juan W. Chavez, Gregory L.
Fenves, and Andrew S. Whittaker. May 1998.

PEER 1998/01 Seismic Performance of Well-Confined Concrete Bridge Columns. Dawn E. Lehman and Jack P. Moehle.
December 2000.
 ONLINE PEER REPORTS

The following PEER reports are available by Internet only at http://peer.berkeley.edu/publications/peer_reports_complete.html.

PEER 2012/103 Performance-Based Seismic Demand Assessment of Concentrically Braced Steel Frame Buildings. Chui-Hsin
Chen and Stephen A. Mahin. December 2012.
PEER 2012/102 Procedure to Restart an Interrupted Hybrid Simulation: Addendum to PEER Report 2010/103. Vesna Terzic and
Bozidar Stojadinovic. October 2012.
PEER 2012/101 Mechanics of Fiber Reinforced Bearings. James M. Kelly and Andrea Calabrese. February 2012.
PEER 2011/107 Nonlinear Site Response and Seismic Compression at Vertical Array Strongly Shaken by 2007 Niigata-ken
Chuetsu-oki Earthquake. Eric Yee, Jonathan P. Stewart, and Kohji Tokimatsu. December 2011.
PEER 2011/106 Self Compacting Hybrid Fiber Reinforced Concrete Composites for Bridge Columns. Pardeep Kumar, Gabriel Jen,
William Trono, Marios Panagiotou, and Claudia Ostertag. September 2011.
PEER 2011/105 Stochastic Dynamic Analysis of Bridges Subjected to Spacially Varying Ground Motions. Katerina Konakli and
Armen Der Kiureghian. August 2011.
PEER 2011/104 Design and Instrumentation of the 2010 E-Defense Four-Story Reinforced Concrete and Post-Tensioned
Concrete Buildings. Takuya Nagae, Kenichi Tahara, Taizo Matsumori, Hitoshi Shiohara, Toshimi Kabeyasawa,
Susumu Kono, Minehiro Nishiyama (Japanese Research Team) and John Wallace, Wassim Ghannoum, Jack
Moehle, Richard Sause, Wesley Keller, Zeynep Tuna (U.S. Research Team). June 2011.
PEER 2011/103 In-Situ Monitoring of the Force Output of Fluid Dampers: Experimental Investigation. Dimitrios Konstantinidis,
James M. Kelly, and Nicos Makris. April 2011.
PEER 2011/102 Ground-motion prediction equations 1964 - 2010. John Douglas. April 2011.
PEER 2011/101 Report of the Eighth Planning Meeting of NEES/E-Defense Collaborative Research on Earthquake Engineering.
Convened by the Hyogo Earthquake Engineering Research Center (NIED), NEES Consortium, Inc. February
2011.
PEER 2010/111 Modeling and Acceptance Criteria for Seismic Design and Analysis of Tall Buildings. Task 7 Report for the Tall
Buildings Initiative - Published jointly by the Applied Technology Council. October 2010.
PEER 2010/110 Seismic Performance Assessment and Probabilistic Repair Cost Analysis of Precast Concrete Cladding Systems
for Multistory Buildlings. Jeffrey P. Hunt and Božidar Stojadinovic. November 2010.
PEER 2010/109 Report of the Seventh Joint Planning Meeting of NEES/E-Defense Collaboration on Earthquake Engineering.
Held at the E-Defense, Miki, and Shin-Kobe, Japan, September 18–19, 2009. August 2010.
PEER 2010/108 Probabilistic Tsunami Hazard in California. Hong Kie Thio, Paul Somerville, and Jascha Polet, preparers. October
2010.
PEER 2010/107 Performance and Reliability of Exposed Column Base Plate Connections for Steel Moment-Resisting Frames.
Ady Aviram, Božidar Stojadinovic, and Armen Der Kiureghian. August 2010.
PEER 2010/106 Verification of Probabilistic Seismic Hazard Analysis Computer Programs. Patricia Thomas, Ivan Wong, and
Norman Abrahamson. May 2010.
PEER 2010/105 Structural Engineering Reconnaissance of the April 6, 2009, Abruzzo, Italy, Earthquake, and Lessons Learned. M.
Selim Günay and Khalid M. Mosalam. April 2010.
PEER 2010/104 Simulating the Inelastic Seismic Behavior of Steel Braced Frames, Including the Effects of Low-Cycle Fatigue.
Yuli Huang and Stephen A. Mahin. April 2010.

PEER 2010/103 Post-Earthquake Traffic Capacity of Modern Bridges in California. Vesna Terzic and Božidar Stojadinović. March
2010.
PEER 2010/102 Analysis of Cumulative Absolute Velocity (CAV) and JMA Instrumental Seismic Intensity (IJMA) Using the PEER–
NGA Strong Motion Database. Kenneth W. Campbell and Yousef Bozorgnia. February 2010.

PEER 2010/101 Rocking Response of Bridges on Shallow Foundations. Jose A. Ugalde, Bruce L. Kutter, and Boris Jeremic. April
2010.
PEER 2009/109 Simulation and Performance-Based Earthquake Engineering Assessment of Self-Centering Post-Tensioned
Concrete Bridge Systems. Won K. Lee and Sarah L. Billington. December 2009.
PEER 2009/108 PEER Lifelines Geotechnical Virtual Data Center. J. Carl Stepp, Daniel J. Ponti, Loren L. Turner, Jennifer N. Swift,
Sean Devlin, Yang Zhu, Jean Benoit, and John Bobbitt. September 2009.
PEER 2009/107 Experimental and Computational Evaluation of Current and Innovative In-Span Hinge Details in Reinforced
Concrete Box-Girder Bridges: Part 2: Post-Test Analysis and Design Recommendations. Matias A. Hube and
Khalid M. Mosalam. December 2009.
PEER 2009/106 Shear Strength Models of Exterior Beam-Column Joints without Transverse Reinforcement. Sangjoon Park and
Khalid M. Mosalam. November 2009.

PEER 2009/105 Reduced Uncertainty of Ground Motion Prediction Equations through Bayesian Variance Analysis. Robb Eric S.
Moss. November 2009.

PEER 2009/104 Advanced Implementation of Hybrid Simulation. Andreas H. Schellenberg, Stephen A. Mahin, Gregory L. Fenves.
November 2009.

PEER 2009/103 Performance Evaluation of Innovative Steel Braced Frames. T. Y. Yang, Jack P. Moehle, and Božidar
Stojadinovic. August 2009.

PEER 2009/102 Reinvestigation of Liquefaction and Nonliquefaction Case Histories from the 1976 Tangshan Earthquake. Robb
Eric Moss, Robert E. Kayen, Liyuan Tong, Songyu Liu, Guojun Cai, and Jiaer Wu. August 2009.

PEER 2009/101 Report of the First Joint Planning Meeting for the Second Phase of NEES/E-Defense Collaborative Research on
Earthquake Engineering. Stephen A. Mahin et al. July 2009.

PEER 2008/104 Experimental and Analytical Study of the Seismic Performance of Retaining Structures. Linda Al Atik and Nicholas
Sitar. January 2009.

PEER 2008/103 Experimental and Computational Evaluation of Current and Innovative In-Span Hinge Details in Reinforced
Concrete Box-Girder Bridges. Part 1: Experimental Findings and Pre-Test Analysis. Matias A. Hube and Khalid M.
Mosalam. January 2009.

PEER 2008/102 Modeling of Unreinforced Masonry Infill Walls Considering In-Plane and Out-of-Plane Interaction. Stephen
Kadysiewski and Khalid M. Mosalam. January 2009.

PEER 2008/101 Seismic Performance Objectives for Tall Buildings. William T. Holmes, Charles Kircher, William Petak, and Nabih
Youssef. August 2008.

PEER 2007/101 Generalized Hybrid Simulation Framework for Structural Systems Subjected to Seismic Loading. Tarek Elkhoraibi
and Khalid M. Mosalam. July 2007.

PEER 2007/100 Seismic Evaluation of Reinforced Concrete Buildings Including Effects of Masonry Infill Walls. Alidad Hashemi
and Khalid M. Mosalam. July 2007.
The Pacific Earthquake Engineering Research Center (PEER) is a multi-institutional research and
education center with headquarters at the University of California, Berkeley. Investigators from over 20
universities, several consulting companies, and researchers at various state and federal government
agencies contribute to research programs focused on performance-based earthquake engineering.

These research programs aim to identify and reduce the risks from major earthquakes to life safety and
to the economy by including research in a wide variety of disciplines including structural and geotechnical
engineering, geology/seismology, lifelines, transportation, architecture, economics, risk management, and
public policy.

PEER is supported by federal, state, local, and regional agencies, together with industry partners.

PEER Core Institutions:

University of California, Berkeley (Lead Institution)
California Institute of Technology
Oregon State University
Stanford University
University of California, Davis
University of California, Irvine
University of California, Los Angeles
University of California, San Diego
University of Southern California
University of Washington

PEER reports can be ordered at http://peer.berkeley.edu/publications/peer_reports.html or by contacting

Pacific Earthquake Engineering Research Center

University of California, Berkeley
325 Davis Hall, mail code 1792
Berkeley, CA 94720-1792
Tel: 510-642-3437
Fax: 510-642-1655
Email: peer_editor@berkeley.edu

ISSN 1547-0587X