Resilience in Transportation Planning, Engineering, Management, Policy, and Administration (2018)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

15 This chapter provides a range of resilience definitions from sectors outside as well as within transportation and analyzes the similarities and differences of these definitions. Next, a review of the variety of metrics being applied to assess resilience is discussed. Finally, several frame- works that provide processes for agencies to perform assessments of resilience are examined. This information is provided as a means to begin to understand the relatively immature state of resilience computational methods within the transportation sector. The Department of Homeland Security (DHS) conducted the analysis of the range of defi- nitions of resilience for infrastructure (U.S. DHS, 2011). Within the DHS study, the authors identified several key discriminators (goal of definition, event cycle addressed in definition, and approach to measuring resilience) to analyze and categorize resilience definitions for critical infrastructure, as shown in Table 1. These discriminators have been broadly applied for resilience definitions to help identify simi- larities and differences between the resilience definitions included in this literature review. A review of definitions of resilience from both the highway sector as well as other related industries is provided here. Resilience Definitions from Other Sectors Resilience has many definitions, such as “the ability of a system to resist, absorb, recover from, or successfully adapt to a change in environment or conditions” (Moteff, 2012). Other defini- tions include factors related to a facility being able to continue operating at a diminished level, or where system performance fails gracefully or gradually instead of all at once, again capturing the “bend but not break” concept of resilience. The National Research Council (NRC) defined resilience as “the ability to prepare and plan for, absorb, recover from, and more successfully adapt to adverse events” (NRC, 2012). NIAC published a report in 2009 that defined infrastructure resilience as “the ability to reduce the magnitude and/or duration of disruptive events. The effectiveness of a resilient infrastructure or enterprise depends upon its ability to anticipate, absorb, adapt to, and/or rapidly recover from a potentially disruptive event” (National Infrastructure Advisory Council, 2009). Furthermore, the United Nations (UN) defines resilience as “the ability of a system, community or society exposed to hazards to resist, absorb, accommodate to and recover from the effects of a hazard in a timely and effective manner, including through the preservation and restoration of its essential basic structures and functions” (United Nations, 2009). The National Academies describes the need for local communities to work with both the private sector and all levels of government to ensure there is “the ability to prepare and plan for, absorb, recover from and more successfully adapt to adverse events” [The National Academy of Sciences (NAS), 2012]. C H A P T E R 3 Resilience Definitions, Metrics, and Frameworks

16 Resilience in Transportation Planning, Engineering, Management, Policy, and Administration Resilience is defined by ASME as “the ability to function through an attack or natural event or the speed by which an asset can return to virtually full function” (ASME Innovative Technologies Institute, 2009). Within the American Water Works Association (AWWA), resilience is defined as “the ability of an asset or system to withstand an attack or natural hazard without interruption of performing the asset or system’s function or, if the function is interrupted, to restore the function rapidly” (AWWA Staff, ASCE/AWWA Committee, 2009). AWWA also identifies methods to improve system resilience by finding options that reduce at least one of the three elements that they use to define resilience: service denial, vulnerability, and likelihood. When comparing the various definitions of resilience, one observation that can be made is the focus of the AASHTO Standing Committee on Research (SCOR), ASME, and AWWA defi- nitions on the post-event portion and event cycle, unlike those definitions developed by NRC, NIAC, and UN, which have a broader focus on all three event cycles (before, during, and after an event). The NIAC definition also appears to address the proactive nature of resilience by calling out the need to “reduce the magnitude and/or duration of disruptive events” as well as “rapidly recover.” In addition, the NIAC definition appears to capture the graceful degradation of assets through its note of resilient systems being those that can “anticipate, absorb, adapt to” events. Other research entities and researchers have also developed definitions of resilience as described next. The Multidisciplinary Center for Earthquake Engineering Research sponsored by the National Science Foundation developed a framework (R4 Framework) that examines four main attributes of resilience: • Robustness—the ability of systems, system elements, and other units of analysis to withstand disaster forces without significant degradation or loss of performance; • Redundancy—the extent to which systems, system elements, or other units are substitutable, that is, capable of satisfying functional requirements, if significant degradation or loss of functionality occurs; • Resourcefulness—the ability to diagnose and prioritize problems and to initiate solutions by identifying and mobilizing material, monetary, informational, technological, and human resources; and • Rapidity—the capacity to restore functionality in a timely way, containing losses and avoiding disruptions. These attributes are described in terms of resilience as “the ability of the system to withstand a disaster without significant degradation or loss of performance, the extent to which the system is capable to maintain functional requirements if significant degradation or loss of function occurs, the ability to diagnose and prioritize problems and to initiate solutions by identifying and mobi- lizing material, monetary, informational, technological, and human resources and the capacity to restore functionality in a timely way, containing losses and avoiding disruptions” (Tierney and Bruneau, 2017). This definition also addresses a broader range of the timeline associated with Table 1. Key discriminators used to develop consolidated resilience definitions (U.S. DHS, 2011).

Resilience Definitions, Metrics, and Frameworks 17 events (before, during, post) and addresses more of the goal areas as outlined in Table 1, includ- ing maintaining functionality, recovery of functionality, and inhibiting change of state. Resilience definitions have also been developed by researchers for the freight industry. When defining resilience for freight transportation systems, Ta et al. (2009) describe it as “The ability for the system to absorb effects from a disruption and continue moving traffic in an uninhibited manner . . .” and also say that this definition involves the flexibility, elas- ticity, and ability to recover from a disturbance. In this definition, resilience is defined by the measurement of reliability, travel time, travel speed, and vehicle count, which are used to measure overall system performance. The resilience of the freight system depends on the ability of the system to recover from/reduce the effects of disruptions in order to maintain freight mobility. Resilience Definitions from Transportation Resilience is a term that is relatively new in the transportation sector, therefore, it still has a range of definitions in literature. AASHTO’s SCOR has defined resilience as “the ability of the transportation system to recover and regain functionality after a major disruption or disaster” (U.S. DOT, 2014b). In 2009, the AASHTO–TRB Transportation and Security Sum- mit proposed that resilience in the transportation sector should be defined as “The ability of a system to provide and maintain an acceptable level of service or functionality in the face of major shocks or disruptions to normal operations” (AASHTO, 2016b). Another definition from FHWA Order 5520 explains resilience as “the ability to anticipate, prepare for and adapt to changing conditions and withstand, respond to and recover rapidly from disruptions” (U.S. DOT, 2014b). While the AASHTO-TRB definition focuses primarily on the ability to function during and post event, the FHWA definition also includes the period before the event by stressing the need to “anticipate, prepare for, . . . changing conditions” as well as “recover rapidly from disruptions” (U.S. DOT, 2014b). Other researchers within the transportation sector have identified various components or properties of resilience, for example, Murray-Tuite (2006) developed a resilience approach that describes the following 10 properties of a resilient transportation system: 1. Redundancy, 2. Diversity, 3. Efficiency, 4. Autonomous components, 5. Strength, 6. Adaptability, 7. Collaboration, 8. Mobility, 9. Safety, and 10. The ability to recover quickly. Murray-Tuite (2006) went on to state that because resilience is tied to these 10 components, it is difficult to obtain a comprehensive measure due to the complexities and interactions between the components. This researcher also defined resilience as “a characteristic that enables the system to compensate for losses and allows the system to function even when infrastructure is damaged or destroyed.” This definition appears to be most focused on the performance of a system during an event and also on graceful degradation of a system to continue to function despite an event occurring.

18 Resilience in Transportation Planning, Engineering, Management, Policy, and Administration According to Heaslip et al. (2012), transportation resilience can be defined in ways that include the following: • A system’s ability to maintain its demonstrated level of service or return to that level of service in a specified time frame; • A system’s ability to compensate for losses to allow functionality, even when that system is damaged or destroyed; • A system’s ability to cope with unexpected conditions without complete failure; and • A system’s ability to absorb consequences of disruptions to reduce effects and maintain freight mobility. This definition again is broader in terms of the event cycle and is focused on returning to functionality despite damage and without complete failure referencing the desire for graceful degradation. The ability to cope or absorb unexpected events or disruptions is also called out in the definition that reflects the broader range of goals as described in Table 1. Internationally, the New Zealand Transport Agency has taken steps to measure resilience of the transport infrastructure and has also developed a definition of resilience that is more estab- lished than most with regard to asset management, though qualitative in nature. Their publica- tion states that “the concept of resilience is wider than natural disasters and covers the capacity of public, private and civic sectors to withstand disruption, absorb disturbance, act effectively in a crisis, adapt to changing conditions, including climate change, and grow over time” (Hughes and Healy, 2014). This definition appears to reflect all three event cycles and goal areas related to maintaining continuity of functionality, and inhibiting state of change by absorbing distur- bances. It also begins to address graceful degradation through the note to adapt to changing conditions but puts less emphasis on recovery time. As part of this synthesis, a survey of state DOTs was conducted and is fully discussed in chap- ter 4. Of note, however, is the response provided by eight state DOTs that provided a definition of resilience as shown in Table 2. A review of the definitions provided in Table 2 from state DOTs reveals that most fall short of capturing the before, during, and post time periods of emergency events, with most being focused on the post-event time period. Also missing in most is the graceful degradation of trans- portation assets to reduce or eliminate catastrophic failures. All of the definitions provided for resilience within the transportation sector describe a sys- tem’s ability to recover from a disaster, whether natural or attack based, with minimal loss of that system’s functionality. As noted in the introduction of this synthesis, there are competing definitions of resilience that can be categorized in multiple ways. The work by DHS to develop a consolidated definition of resilience may be of use as noted in the introduction to this chapter. In general, the key components for a definition of resilience for highways appear to capture preparing for, absorbing, and recovering quickly from an event that disrupts service. Following the DHS key discriminators of definitions of resilience, most definitions appear to miss the mark on graceful degradation and inhibiting a basic state of change, meaning a system that does not fail or is damaged during an event. The key discriminators recognized by DHS for critical infra- structure include the full range of time periods in which resilience needs to be addressed by asset managers and owners. The key discriminators also reflect four goal areas to address that are not only focused on quick recovery but also address the need to restrict the state of change of some critical assets (hardening these assets to withstand events). Furthermore, recognizing that funds are limited, it is recommended that assets be built to “gracefully degrade” to avoid catastrophic failures that may increase injuries and loss of life. Finally, maintaining continu- ity of function through pre-event planning is addressed through these key discriminators. This type of analysis may be of use to state DOTs as they move forward with their resilience

Resilience Definitions, Metrics, and Frameworks 19 programs as a way to help identify resilient practices and policies that address a broader range of activities that can be implemented to provide more than quick recovery post event to trans- portation assets. An outstanding issue as related to resilience analysis in highways is defining the relationship between resilience and risk. In order to understand this relationship, a short summary of risk definitions is provided next. Risk Definitions In order to understand resilience of the highway system, one must have a solid understanding of what risk assessment and management are and their relationship to resilience. Risk entails four elements: hazards, exposure, vulnerability, and consequence. A comprehensive definition of risk states that risk can be defined as “the potential for adverse effects from the occurrence of a particular hazardous event, which is derived from the combination of physical hazards, the exposure, and vulnerabilities” (The National Academy of Sciences, 2012). FEMA’s definition of risk focuses on the likelihood that a threat will harm an asset with some severity of consequences (FEMA, 2014). NIPP defines risk as the “potential for an unwanted State Agency Resilience Definition Minnesota Department of Transportation Reducing vulnerability and ensuring redundancy and reliability to meet essential travel needs. Oregon Department of Transportation Uses a definition from the Oregon DOT (ODOT) seismic report: “To achieve rapid recovery, require government continuity, resilient physical infrastructure, and business continuity.” Arizona Department of Transportation ADOT developed the Resilience Program to support its mission to provide a safe, efficient, cost-effective transportation system that can be compromised from the effects of heat extremes, dust storms, wildfires, flooding, landslides, rockfall incidents, and slope failures, and cope with the ever-growing cost of these threats. Delaware Department of Transportation Uses the concept in the Delaware Executive Order 41, which outlines resiliency practices to help mitigate climate impacts and reduce emissions. Colorado Department of Transportation Uses a definition provided by the Governor’s Resiliency Framework: “The ability of communities to rebound, positively adapt to or thrive amidst changing conditions or challenges – including disasters and climate change – and maintain quality of life, healthy growth, durable systems, and conservation of resources for present and future generations.” New York State Department of Transportation Uses the recommended guidance provided by the NYS 2100 Commission, which outlines how the state plans to identify areas where further resilience practices are needed. New Hampshire Department of Transportation Uses the NAS definition by default, which states “resilience is the ability to plan, absorb, recover and adapt.” Illinois Department of Transportation Plan and invest in the state’s transportation system to ensure that infrastructure is prepared for extreme weather events. Table 2. State DOT definitions of resilience.

20 Resilience in Transportation Planning, Engineering, Management, Policy, and Administration outcome resulting from an event or occurrence, as determined by its likelihood and the associated consequence” (U.S. DHS, 2009). The Office of Management and Budget defines risk as “a hazard, a probability, a consequence, or a combination of probability and severity of consequences.” According to AWWA, risk is defined as “a function of consequences, hazard frequency, or likelihood and vulnerability, which with point estimates is the product of the terms. It is the expected value of the consequences of an initiating event weighted by the likelihood of the event’s occurrence and the likelihood that the event will result in the consequences, given that it occurs. Risk is based on identified events or event scenarios” (AWWA Staff, ASCE/AWWA Committee, 2009). Risk is estimated by the consequences that are expected to happen if a particular event occurs as well as the likelihood of the event occurrence. A 2010 publication by the NRC reviewed the DHS’s approach to risk analysis and noted that DHS relies heavily on quantitative models for risk assessment, which the committee felt this approach may underemphasize unquantifiable characteristics of risk. The committee rec- ommended that both quantitative and non-quantitative attributes of risk be considered in the decision-making process (National Research Council, 2010b). National and international organizations have also developed definitions for risk. The International Organization for Standardization’s (ISO) ISO 31000 standard defines risk as “the effects of uncertainty on objectives” (ISO, 2009). Similarly, ASME developed a seven- step approach for Risk and Resilience Analysis and Management for Critical Asset Protection (RAMCAP PlusSM) in which risk is defined as “the potential for loss or harm due to an untow- ard event and its adverse consequences” (ASME, 2009). The New Zealand Transport Agency developed a risk assessment tool based on AS/NZS 31000 which is equivalent to the ISO 31000 standard. Within this framework, risk is defined as “the chance of something happening that will have an impact on objectives. It is measured in terms of a combi- nation of the likelihood of an event and its consequence” (New Zealand Transport Agency, 2014). Within the context of highway asset management, FHWA has published a series of reports about risk-based transportation asset management (FHWA, 2013b). Within these reports, FHWA compares definitions of risk from different sources at the national and international levels. FHWA emphasizes that risk can be considered not only as threats or dangers but also as possible opportunities. FHWA formally defines risk as “the positive or negative effects of uncertainty or variability upon agency objectives” (FHWA, 2013b). Similarly, the Transporta- tion Research Board (TRB) report Guidebook on Risk Analysis Tools and Management Practices to Control Transportation Project Costs defines risk as “an uncertain event or condition that, if it occurs, has a negative or positive effect on a project’s objectives” (Molenaar et al., 2010). The primary factors captured in this range of definitions are the acknowledgement and management of potential hazards and the probability of occurrence of applicable hazards that may result in losses dependent on present vulnerabilities. Next, understanding the relationship between risk and resilience is important to fully understand what is required to develop and manage a resilient highway transportation system. Relationship Between Risk and Resilience Evolving policy has led to some confusion as to the relationship and responsibility of infra- structure managers to reduce risk reduction and increase system resilience. The Congressional Research Service documented evolving policy that has moved away from protection and more toward resilience post 9/11 (Moteff, 2012). As noted, “improving resiliency reduces risk pri- marily by reducing vulnerability to and potential consequences of an attack or natural event.”

Resilience Definitions, Metrics, and Frameworks 21 This evolution of the relationship between risk and resilience has been seen through various federal reports and directives since 2006. As noted in the 2010 National Security Strategy, encouraging incentives to design structures and systems that can withstand threats, incorpo- rating redundancy and decentralizing critical operations to reduce vulnerability, developing and practicing emergency response plans, and investing in infrastructure to maintain condition and performance are critical to improve resilience and reduce potential losses from applicable threats. Subsequently, Presidential Decision Directive (PPD-8), National Preparedness, issued by the Obama Administration, sought to improve “the security and resilience of the United States through systematic preparation for threats that pose the greatest risk . . . (U.S. Department of Homeland Security, Online, c).” Efforts have been made to conceptualize the quantitative relationships between risk and resilience in a graphic that depicts the degradation of performance after the event occurs, the bottom out time, and the time required to recover to initial performance. The metrics in the conceptual quantitative model included time and percent performance, referring to the level of performance the asset or system is able to perform at after an event occurs and before full performance is achieved. Figure 2 illustrates a quantitative model. As noted previously, as risk or the expected financial loss decreases, resilience increases. By reducing vulnerability to a threat (moving toward a value of 0), resilience is highest and the expected risk tends toward zero. By understanding the susceptibility of an asset to loss or failure and taking steps to reduce that vulnerability, owners can move toward a more resilient, less risk- adverse state. As anticipated financial losses or risk increase, system resilience decreases. Also, as an asset or system becomes more vulnerable to threats, the risk increases and resilience decreases. Fig- ure 3 provides an example of the conceptual relationship between risk, resilience, and vulnerability. It is important to understand the relationship between risk and resilience to allow the reader to reflect on the materials presented in this synthesis as related to highway resilience within the context of more well-established risk management and assessment methods. Identifying risks is an important step in determining the resilience of a transportation system. By using risk analysis and performance-based measurement approaches, transportation agencies are able to identify weaknesses within their respective systems to determine the potential system level of resilience. According to Herning et al. (2016), “Risk analysis and performance-based approaches can be used to quantify reliability of infrastructure, inform prioritization of inter- ventions, assess life-cycle outcomes related to infrastructure management decisions, and achieve performance objectives. . . .” When transportation agencies analyze risk to inform resilience, they are better able to identify areas that need the most improvement to become more resilient. Figure 2. Resilience profile (U.S. Department of Homeland Security, 2011).

22 Resilience in Transportation Planning, Engineering, Management, Policy, and Administration However, as noted by Jackson (2008), addressing inherent risks can result in the layering of measures for each identified, relevant threat, which often is not sustainable, whereas focusing on resilience and reducing the consequences should a threat come to fruition, may offer more methods to reduce risk. In addition, as noted by the Congressional Research Service, “Reduc- ing risks by building higher fences and deploying more guards around a particular facility or asset (protection) is very different than reducing risks by building a second facility or asset somewhere else or strategically stockpiling replacement equipment that can get the facility or asset up and running again quickly (resilience).” In some cases, it may also be impossible to protect from all applicable threats, and thus incorporation of resilient practices into a system may be the only plausible solution. After reviewing the range of resilience definitions and the relationship between risk and resil- ience, the next section of this chapter focuses on how to measure resilience and those studies that have been published reflecting a range of approaches to measuring resilience. Resilience Metrics and Indices The concept of resilience is still being formalized within many sectors, and in many cases is dependent on the preference of measurement of resilience (e.g., time, number of vehicles). Cur- rently, there is no standard measurement for resilience within highway analysis, which is echoed in the state DOT survey results in the survey conducted for this research, in which 92% of the states responding noted that they do not currently have specific metrics for resilience in place. The method by which resilience is measured often depends on what is most important to policy makers. For example, “if monetary losses are important, it may be more appropriate to measure the total (or net) loss of revenue associated with the disruption” (Moteff, 2012). On the other hand, if policy makers are more concerned about the public need to use the damaged facility or asset, it may be appropriate to measure resilience in terms of how quickly the damaged asset or facility was able to resume normal operations. The following review has been categorized into empirical field measurements of resilience metrics and measurements of resilience gathered through modeling efforts. Very few studies were identified that measured resilience during an event and were able to compare the performance before and during an event. Most studies appear to be seeking to use existing data sources to reflect resilience through operational mea- sures such as travel time. One study noted that when agencies are defining resilience metrics, it is important that the metrics chosen are broad enough to be used to measure resilience across the Figure 3. Relationship between risk, resilience, and vulnerability (U.S. Department of Homeland Security, 2011).

Resilience Definitions, Metrics, and Frameworks 23 board. Hughes and Healy (2014) note that studies tend to fall into one of two groups: qualita- tive frameworks that reflect resilience and quantitative studies that seek to measure metrics of resilience through more elaborate modeling approaches. In order for an agency to measure resilience, it must be understood what components com- pose resilience. Parkany and Ogunye (2016) used previous work to identify multiple potential metrics for each of the four components of resilience (robustness, redundancy, resourcefulness, and rapidity), including hours of congestion, travel time index, pavement condition, distance to alternative routes, and availability of safety (courtesy) service patrols (Tierney and Bruneau, 2017). Next, information was gathered from readily available data sources, and the metrics were applied to 10 corridors in the Commonwealth of Virginia. The authors concluded that con- tinuous data were difficult to work with, however, they found some additional value in the metrics when these data were converted to categorical data, but they noted that additional work is needed to finalize the metrics and data sources. Of note is the use of existing performance metrics cast in the light of resilience without the need to create additional metrics from scratch, which should reduce barriers to measuring resilience (Murray-Tuite, 2006). The Multidisciplinary Center for Earthquake Engineering Research also recognized the four Rs of resilience but expanded the metric categories to include four components of resilience: technical, organizational, social, and economic (TOSE) to develop the TOSE method. These components are used to develop resilience metrics that can be used by any sector to mitigate the effects caused by natural hazards (Tierney and Bruneau, 2017). Transportation agencies have a need to analyze threats that cause failures in their systems, but they also need a way to be able to measure their systems’ resilience to those threats. Resilience metrics provide a way for an agency to take what they have learned from their risk analyses to measure the resilience of their transportation system. Tierney and Bruneau (2017) state that “The R4 approach highlights the multiple paths to resilience. Investments can improve all four resilience components—robustness, redundancy, resourcefulness, and rapidity. The TOSE framework emphasizes a holistic approach to community and social resil- ience, looking beyond physical and organizational systems to the effect of disruptions on the social and economic systems.” By using both of these frameworks, transportation agencies will be able to create resilience policies that not only will take into account the functionality of the asset but also will look at the social and economic implications that come along with natural and manmade threats. The Mountain-Plains Consortium developed a framework and metric index to evaluate the resiliency of the light rail system in the Denver Metro area (Marshall, 2015). The Transporta- tion Economic Resilience (TER) rating system was developed and reflects the effect of spikes in fuel prices on individuals’ income across the region for home-to-work trips. Using 2,832 trans- portation analysis zones (TAZ), the authors assigned a TER score to each TAZ and were able to assess how resilient each zone is to sudden spikes in fuel costs that reflected the ability to mode shift, the increased costs associated with each mode, and the distance and time associated with travel between home and work in each TAZ. Not surprisingly, three primary contributors to the TER scores include household proximity to downtown, median household income, and avail- ability of multimodal transportation options. A similar study was conducted for Mississippi DOT after Hurricane Katrina that sought to develop metrics to reflect the resilience of regions post event (Zhang et al., 2010). Several resil- ience indicators, including travel time, average trip length, percentage of freight vehicles travel- ing under the 85th percentile speed, percentage of total length of highway open to freight traffic in the network, and proportion of freight traffic that is able to maintain schedule were considered in the study. The authors used a measure of resilience (MOR) to capture the reduction in any

24 Resilience in Transportation Planning, Engineering, Management, Policy, and Administration of the resilience indicators before and after an event. The MOR can be calculated for each of the resilience indicators using the following equation: ( )( )= − + α MOR 1 % RI RI t RI Before After Before where RIBefore = resilience indicator before an event; RIAfter = resilience indicator after an event; t = total time required to restore the capacity (year); and α = system parameter related to network size, socioeconomic status, government policy, etc.; used α = 0.5 in case study. Ip and Wang (2011) proposed a measure of resilience that reflects the weighted average number of reliable passageways with all other nodes in a network. The method presented takes into account the population of the various nodes being serviced by the rail system analyzed. The authors introduced a concept known as “friability,” which represents the reduction in network resilience caused by a disruption to the network such as a disaster or an attack. The authors applied their approach to the Chinese rail network and assessed the reliability of the network. Jenelius et al. (2006) introduced a metric that reflects the change in travel costs before and after an event that may limit travel on particular links. Using a sparse network in the northern part of Sweden, the authors used traditional models, such as user equilibrium, to assign traf- fic to the network based on the assumption that travelers will choose the least costly travel route between their origin and destination. They also assumed that travel time is independent of the traffic demand on each link and justified their assumption by noting the very low traffic demands in the northern Sweden area studied. The authors tested their approach by randomly closing a link on the network, and then closing the most important link on the network, which is identified as the link that results in the highest increase in travel time when removed from the network. The authors also captured unmet demand, those trips that could not be completed, through their analysis. Finally, by using visualization in geographic information systems (GIS), the authors differentiated between the importance of each link that serves origin–destination pairs in northern Sweden based on each link’s effect on increased travel costs and the unmet demand that remained when each link was removed from the network. A similar metric was developed by Chen and Miller-Hooks (2012), who studied the ability of intermodal freight networks to withstand and recover from disruptions. Their metric takes into account predisaster planning and recovery post disaster and reflects the demand that can be serviced before and after a disaster. Adams et al. (2012) also studied freight resilience metrics by studying truck speeds and counts along the I-90/94 corridor in Wisconsin during two weather events in 2008. The authors focused their efforts on measuring two periods referred to as reduction and recovery; they represented the time passage between the start of the event and when the system reaches its lowest perfor- mance level, and the time passage between the lowest performance level and the time to recovery (pre-event conditions). By sampling Global Positioning System–equipped trucks on the cor- ridor during two major weather events (a blizzard and a flooding event), the authors were able to develop two metrics, referred to as alpha and beta, which represent the angle of deterioration and recovery triangles to represent freight resilience on the I-90/94 corridor during significant weather events. Resilience metrics for transit systems have also been developed by researchers at the Mineta National Transit Research Consortium (Nassif et al., 2017). The study addressed transit system

Resilience Definitions, Metrics, and Frameworks 25 resilience by focusing in on three key areas: transit bridge infrastructure resiliency, public transit system resiliency, and efficiency of transit systems with a focus on disability paratransit service. The study used structural health monitoring, finite element analysis, and remote sensing to estab- lish resilience metrics for transit bridge infrastructure. Vulnerability, resilience, and efficiency in recovering from a major natural disaster were studied for the public transit system as a whole by adapting readily available metrics to transit. Specific metrics included change in travel time, recovery time, change in travel speed, and number of reliable routes. And travel time and trip delay were noted as the more significant metrics for modeling the efficiency of transit systems. Indices Some research has been conducted on the development of indices to measure resilience. Foster (Online) developed a Resilience Capacity Index that captures an area’s ability to respond to future stressors into a single metric that includes 12 indicators, including economic, socio- demographic, and community connectivity. The researchers went on to apply the index to the United States as a whole and rated various metro areas on a scale from very low to very high by categorizing scores. The University of South Carolina has developed a similar index: the Social Vulnerability Index for the United States (SoVI). This index captures 29 socioeconomic vari- ables at the county level that are believed to reflect a county’s ability to prepare for, respond to, and recover from hazardous events ranging from natural disasters or disease outbreaks (Uni- versity of South Carolina, Online). Figure 4 includes a recent version of the SoVI metric for the United States. Of note from the review of the resilience metrics and indices is the lack of consensus on resil- ience metrics. Many studies have pulled metrics from operational areas (such as reliability) and related metrics (such as travel time, delay, and travel speed) to reflect the resilience of transporta- tion systems; most of these metrics relate to the speed at which the system is brought back into full service or the amount of time operating at partial service capacities. Metrics appear to be lacking for the bulk of the Rs of resilience, including robustness, redundancy, resourcefulness, and rapidity, and exclusively relate to the ability of the agency to rebound. The last section of this chapter focuses on methods by which agencies can begin to develop a repeatable process to assess resilience of their systems often referred to as frameworks. Frameworks for Resilience Some of the challenges faced by transportation agencies when beginning to address system or asset resilience is how to integrate resilient solutions into their existing programs, includ- ing asset management, operations, design, and maintenance. The broad range of publications from federal agencies that tackle the topic of resilience reflects the potential confusion and lack of full integration of resilient practices among transportation agencies. For example, FHWA has published multiple guidance documents on the topic of resilience, ranging from risk-based asset management, climate change and extreme weather guidance, the FHWA Emergency Relief Program Manual, and planning and design guidance (FHWA, 2013a, 2013b, and 2016). In this section, models for structuring the assessment of resilience are described along with documented efforts by transportation agencies to begin the assessment of system resilience. In order to understand resilience, it is important to know the different range of dimensions, principles, and measures of resilience. To develop an implementable approach for resilience measurement and improvement, it is important to define its dimensions. Different approaches have been developed incorporating dimensions of resilience at different levels such as individual, community, design, economic, and strategic planning.

26 Resilience in Transportation Planning, Engineering, Management, Policy, and Administration NIAC differentiates between approaches related to people and processes, and the structure of infrastructure and assets (National Infrastructure Advisory Council, 2010). Other studies have identified different principles for resilience such as robustness, redundancy, resourceful- ness, rapidity, and adaptability (Bruneau et al., 2003). Figure 5 summarizes some potential attri- butes of resilience. The incorporation of resilience management into an agency’s decision-making process requires a broad approach that considers long-term infrastructure resilience and economic optimization rather than the short-term return on investment perspective. In addition, risk and resilience management approaches should be considered complementary and are not mutually exclusive (New Zealand Transport Agency, 2014). Without the inclusion of resil- ience management, an owner may inadvertently invest in risk reduction across the system without addressing the needs of the system in terms of service delivery. A resilience manage- ment plan allows for the identification of critical infrastructure corridors or networks that provide for the required service levels of the overall system. One requirement of any resilience management plan is the prioritization of assets in terms of their ability to withstand specific Figure 4. Social Vulnerability Index for the United States (University of South Carolina, Online).

Resilience Definitions, Metrics, and Frameworks 27 threats. For example, to ensure continuity of service, agencies must study their system as a whole and identify high priority origin–destination pairs to ensure service on critical segments of the system to allow for flow of essential traffic, freight movement, and defense movement, if applicable. The level of resilience could be further refined by segment or corridor with design standards that reflect a required passage of flood events, debris flow, or embankment scour. The identified level of resilience could then be incorporated into a risk and resilience asset management plan. DHS developed the National Mitigation Framework as a reference for agencies to learn to lessen the effects that are brought on by threats in hazards through mitigation strategies in order to promote resilience for their systems. The framework states that “Demonstrating clear and measurable returns on investment through mitigation is essential to that dialogue and necessary to build a resilient, risk-conscious culture” (U.S. Department of Homeland Security, 2016). The authors also note that becoming a resilient society is dependent on managing the entire spectrum of possible risks, which includes natural threats as well as manmade threats. The New Zealand Transport Agency developed and implemented a framework to measure resilience of transport infrastructure (road and rail) based on the New Zealand Treasury’s National Infrastructure Plan. The framework can be applicable at various scales, including asset, network/route, and regional and produces a range of resilience scores reflecting the following for each agency: • A robustness score, • A redundancy score, • A safe to fail score, • A change readiness score, • A networks score, and • A leadership and culture score. The framework provides an approach to combine these scores to produce an overall agency resilience score ranging from 1 (low resilience) to 4 (very high resilience) (New Zealand Transport Agency, 2014). Figure 5. Resilience attributes (New Zealand Transport Agency, 2014).

28 Resilience in Transportation Planning, Engineering, Management, Policy, and Administration ASME (2009) published a framework for the assessment of critical infrastructure, RAMCAP, which is applicable to 16 critical infrastructures, including transportation. The framework pro- vides a seven-step process to analyze and mitigate risks from potential terrorist attacks on critical infrastructure assets. The seven steps are listed below: 1. Asset characterization and screening, 2. Threat characterization, 3. Consequence analysis, 4. Vulnerability analysis, 5. Threat assessment, 6. Risk assessment, and 7. Risk management. Using this seven-step framework, agencies from all sectors are able to identify, analyze, and potentially mitigate risks associated with terrorist attacks to provide a much more resilient system. AWWA has taken the RAMCAP approach to assess resilience in the water/waste water sector and developed an industry standard referred to as the AWWA J-100-10 (AWWA Staff, ASCE/ AWWA Committee, 2009). The approach has been used to meet three major objectives in the water sector: • To define a common framework that can be used by the water sector to assess human-caused and natural hazards risk to their systems; • To develop risk-based vulnerability analyses and value-based prioritized actions to reduce risk and enhance resilience; and • To provide an efficient and consistent mechanism that can be applied to both private and governmental (federal, state, and local) sectors to report essential risk and benefit information to operators of utilities, local and state governments, DHS, U.S. EPA, and others with a need to know (AWWA Staff, ASCE/AWWA Committee, 2009). When used for the water sector, RAMCAP has provided a “consistent, efficient, and techni- cally sound methodology to identify, analyze, quantify, and communicate the level of risk and resilience” (AWWA Staff, ASCE/AWWA Committee, 2009). FHWA (2012) produced a framework to assess vulnerabilities due to climate change and extreme weather. The framework is made up of three steps: defining study objectives and scope, assessing vulnerability, and incorporating results into decision making. By employing these three steps, transportation agencies are better able to prepare for and recover from extreme weather events and effects of climate change; this will lead to a more resilient system. Figure 6 details the process. Summary This chapter provided an overview of resilience metrics, indices, and frameworks for assess- ment. It should be noted that resilience metrics within the transportation sector appear to be lacking and are often drawn from existing operational measures related to system reliability, which fails to provide methods to address all four components of resilience: robustness, redun- dancy, resourcefulness, and rapidity. It may be possible to draw from other sectors to address the components that currently lack metrics, such as the SoVI metric to address resourcefulness, though this metric is currently measured at the county level and may not fully address a highway agency’s ability to respond to emergency events.

Resilience Definitions, Metrics, and Frameworks 29 Figure 6. FHWA Framework for Vulnerability Assessment (FHWA, 2012a). Frameworks were also introduced to demonstrate potential approaches to measuring system resilience. In some cases, the frameworks address risk assessment as well as system resilience, although others may put an agency on the right path to resilience assessment but require addi- tional assessment beyond vulnerability assessment. In the next chapter, a review of the current use of resilient practices in state DOTs is provided based on information gathered through a national survey.