In This Issue
Summer Bridge Issue on Engineering for Disaster Resilience
July 1, 2019 Volume 49 Issue 2
The articles in this issue present examples of engineering innovation to develop resilient infrastructure.

A Fully Integrated Model of Interdependent Physical Infrastructure and Social Systems

Wednesday, July 3, 2019

Author: Bruce R. Ellingwood, John W. van de Lindt, and Therese P. McAllister

Common to the many definitions of resilience in the literature and in policy statements is the notion that resilience is the ability to prepare for and adapt to changing conditions and withstand and recover rapidly from disruptions. The performance of the built environment and the support of social, economic, and public institutions are essential for a community’s immediate response and long-term recovery after a disruptive natural hazard event.

This article describes research in the Center for Risk-Based Community Resilience Planning, which is sponsored by the National Institute of Standards and Technology (NIST). The Center’s objective is to advance measurement science to (a) understand the factors that make a community resilient, (b) assess the -likely impacts of hazards on communities, and (c) develop risk-informed decision strategies that optimize planning for and recovery from natural hazard events. Research has already laid the groundwork for advances in community resilience science and implementation.

Introduction and Background

The resilience of a community is determined by its ability to return to a level of normalcy within a reasonable time after a disruptive hazard event; it reflects the community’s preparedness and capacity to respond to and recover from damage to physical, social, and economic systems.

Civil infrastructure, on which the economic and social well-being of any community depends, is susceptible to damage due to natural hazards such as hurricanes, floods, tornadoes, earthquakes, and wildland-urban-interface fires. Significant damage may occur even from hazard events with less than “design-level” forces, and may produce disproportionate economic and social -losses, especially for lower-income households, the elderly, and other vulnerable segments of society. The potential exists for even larger US losses due to shifts of population and economic development to hazard-prone coastal regions, increasing population density in urban areas, and the ever-increasing interdependence of physical, social, and economic networks.

Current Practices for Resilience of the Built Environment

The resilience of communities has garnered significant attention from practitioners, researchers, and policy-makers over the last decade. Earthquakes in Haiti, Chile, New Zealand, and Japan, Superstorm Sandy in 2012, and more recent hurricanes and wildfires in the United States have revealed the importance of mitigation, response, and recovery policies that focus on the resilience of the community as a whole, rather than on the safety and functionality of individual civil infrastructure facilities. Resilience assessment has become a national imperative (NRC 2012; White House 2013) not only in the United States but in Europe and countries in the Asia-Pacific Rim.

A community’s resilience needs and objectives, particularly those related to postdisaster functionality and recovery, are not generally reflected in the codes, standards, and other regulatory documents that engineers typically apply in designing individual facilities for natural hazards (e.g., ASCE 2016; ICC 2018). While the performance of individual facilities (e.g., buildings, bridges, buried piping, electrical substations) and infrastructure systems (e.g., electrical, gas, and water distribution systems) during specific natural hazards is reasonably well understood, there has been less attention to the fact that such facilities and systems are interconnected and interdependent.

Natural hazards have varying temporal and spatial scales and are highly uncertain in occurrence and intensity. Similarly, while each facility and infrastructure system has its own characteristic response to a natural hazard, the performance of these systems during and after disruptive hazard events is positively correlated because of the interconnected nature of their functions in the community and the extended spatial scale of the event. These spatiotemporal correlations are not reflected in current risk and loss estimation platforms.

A New Approach to Community Resilience

In light of the numerous uncertainties associated with the performance of community infrastructure, a new risk-informed decision-making approach to community resilience assessment and enhancement is essential (Lounis and McAllister 2016). The new approach should reflect the interdisciplinary nature of the problem and the complex interdependencies among the physical, social, and economic systems in a community.

The integrated effects of physical and social infrastructure performance on the resumption of community normalcy after a hazard event and the uncertainties in recovery are depicted by figure 1 (see also Bruneau et al. 2003). Given the many dimensions of community resilience, this curve may be difficult to understand.

 Figure 1

Furthermore, most research on community resilience in the past two decades has focused on the impacts of severe earthquakes on physical infrastructure (e.g., -Bruneau et al. 2003; Koliou et al. 2018; Mieler et al. 2015). In recent years, attention has shifted to other natural hazards, including those that might be impacted by climate change (ASCE 2015).

Introducing the Center for Risk-Based Community Resilience Planning

The Center for Risk-Based Community Resilience Planning (the Center; http://resilience.colostate.edu/), headquartered at Colorado State University in Fort Collins, was established as a Center of Excellence by NIST in 2015. The Center involves a dozen universities and has benefited from the participation of more than 100 investigators (faculty, graduate students, and postdoctoral fellows). Consistent with NIST research priorities, the Center’s goals are to

  • establish the measurement science for identifying and understanding the factors that make a community resilient,
  • develop a computational environment with fully integrated supporting databases to assess the likely impacts of natural hazards on communities, and
  • develop risk-informed decision support strategies with discrete sets of optimal solutions.

The Center is engaged in three major research thrusts aimed at accomplishing these goals: (1) an Inter-dependent Networked Community Resilience Modeling Environment (IN-CORE) to assess alternative strategies for improving community resilience; (2) a standard data ontology, robust architecture, and management tools that support IN-CORE; and (3) testbeds, hindcasts, and field studies to validate the advanced modeling environment.

Community Resilience Measurement Science: Overview of the Center’s Programs

NIST and the Center define communities as regional entities with a common governance structure that allows coordinated decision making and policy implementation. The Center’s research takes a broadly integrated and interdisciplinary approach toward modeling natural hazards, interdependent physical infrastructure systems, and social and economic institutions that are supported by the built environment. The natural hazard models include hurricanes/storm surge/coastal inundation, earthquakes, and riverine flooding, as well as hazards that may be more localized and less well defined for engineering analysis, such as tornadoes, tsunamis, and wildland-urban-interface fires.

To lay the groundwork for advances in community resilience science, research at the Center is addressing

  • single, multiple, and cascading hazards, modeled as scenarios at the community scale;
  • the performance of physical components and -systems—buildings, transportation networks, water and wastewater systems, energy networks, tele-communication networks, and their geospatial and logical interdependencies;
  • economic modeling using computable general equilibrium (CGE) models;
  • social systems and cascading effects, including event impacts and recovery;
  • optimization strategies for enhancing the selection of community resilience strategies that consider uncertainties in natural hazard occurrence, intensity, and physical-socioeconomic response;
  • exploratory testbeds, full validation of the modeling environment using hindcasts, and field studies of communities of differing sizes;
  • a resilience data management structure and a community resilience glossary and taxonomy; and
  • standard tools and protocols for longitudinal post-event community surveys.

A measurement science–based approach to community resilience assessment and decision making also requires community goals, or aspirational statements, quantified by resilience metrics to determine whether the goals have been achieved. The Center, in collaboration with NIST research staff, has developed tentative community resilience goals and metrics (table 1) informed by the NIST Community Resilience Planning Guide (NIST 2016).

The Interdependent Networked Community Resilience Modeling Environment (IN-CORE)

IN-CORE provides the structure and direction to the Center’s research programs and is one of the Center’s most notable accomplishments to date (-Gardoni et al. 2018; van de Lindt et al. 2018a).

Built on the recognition that the resilience of a community depends on interconnected physical, social, and economic systems, IN-CORE provides novel and comprehensive modeling to enhance understanding of the decision points, actions, and resources that communities use to improve their resilience. It is being developed as an open-source research tool that integrates estimates of the impacts of natural hazards on the built environment to support collaborative research by experts in community resilience modeling and assessment. It is also intended to inform users about the (a) potential impacts of natural hazard events on communities, (b) effectiveness of policy and project decisions aimed at enhancing community resilience, and (c) development of best practices for achieving community resilience.

 Figure 2

The structure of IN-CORE is illustrated in figure 2. Some modules are core modules while others may be user supplied, facilitating use as a research tool. -Modules must be called in a specific order, a requirement necessitated by the input and output for information flow and integration. Users can enter and exit the workflow at any point, depending on their analysis needs.

 Figure 3

IN-CORE can simulate the effects of various strategies for improving community resilience (e.g., reduced times to restore functionality; Zhang et al. 2017). -Feasible or optimal strategies for providing risk-informed decision support can also be identified using algorithms based on one or more hazard scenarios, as indicated in figure 3, which shows the results of a multiobjective optimization analysis aimed at minimizing cost and household dislocation (Zhang and Nicholson 2016).

The decision framework is not intended to endorse specific resilience improvement strategies; these involve local social, economic, political, and cultural values that require community consensus. Rather, it is intended to provide choices that improve community resilience related to physical and socioeconomic -systems and the availability of services (e.g., health care).

Exploratory Testbeds, Hindcasts, and Longitudinal Field Studies for Architecture Validation

The fundamental measurement science (e.g., methods and metrics to quantify resilience) and the IN-CORE computational environment have been developed through testbeds, hindcasts, and field studies involving interdisciplinary teams of engineers, social and -economic scientists, and information technologists. The Center took this approach for three primary -reasons: to require interdisciplinary collaborations, which seldom happened in the past; to advance integrated temporal (longitudinal) and spatial (regional) modeling of community recovery; and to ensure that the supporting research and development are relevant to real communities. Each study was created for a specific purpose.

The Centerville Virtual Community Testbed is an idealized community of 50,000 residents with an economy, infrastructure, and demographics intended to be representative of similarly sized communities exposed to seismic hazards in the central United States. It provides a platform suitable for interdisciplinary team training (Ellingwood et al. 2016).

The Seaside, OR Testbed involves a small coastal resort community that is being used as a model to develop multihazard damage and loss assessments for earthquake and tsunami scenarios originating from the Cascadia Subduction Zone (Attary et al. 2017). Impacts considered include population dislocation as a result of structural damage to buildings as well as loss of building functionality due to lack of potable water, and -temporal reduction in water demand due to population dis-location (Guidotti et al. 2017).

The Galveston and Bolivar Peninsula, TX Testbed integrates models that include the effects of hurricane wind and storm surge and performance of interdependent physical infrastructure systems on housing and business disruption and population dislocation. Postevent housing recovery modeling is being advanced using empirical findings from Hurricane Ike in 2008. Factors that affect housing recovery include resource availability, social capacity, business continuity, infrastructure system performance, and access (Hamideh et al. 2018).

The Metropolitan -Memphis Statistical Area Testbed involves a nine-county region of approximately 1.4 million people exposed to earthquake and flood hazards and is the largest testbed in the Center. The inclusion of surrounding counties allows examination of the impact of support from adjacent communities on resilience and the degree to which algorithms developed on smaller testbeds can be scaled to an urban area. A CGE model of eight employment and residential areas subject to varying effects was developed as part of this testbed.

The Joplin, MO Hindcast was designed to validate the accuracy of IN-CORE modeling of both individual physical sectors and coupled physical, social, and economic sectors in responding to the twin-vortex EF-5 tornado of May 22, 2011. The Center’s tornado and -fragility models were applied to a series of analyses, from spatial building damage analysis (compared to building inspection data after the event) to population dislocation data (compared to data collected both immediately after the event and during Joplin’s path to recovery) (Attary et al. 2018).

Finally, the Lumberton, NC Longitudinal Field Study is an example of synergies achieved by collaboration between NIST and the Center. After Hurricane -Matthew caused severe flood damage in Lumberton in September 2016, a Center-NIST team visited to assess building damage, survey a representative sample of households to document their dislocation and early recovery efforts, and meet with community leaders, infrastructure stakeholders, and public officials to discuss overall impacts and recovery decisions and issues. Figure 4 shows estimated days of dislocation for sampled Lumberton housing/households.

 Figure 4

The team returned to Lumberton in early 2018 to follow up on recovery and to initiate a survey of business recovery, and has plans to return periodically to measure recovery progress. Such longitudinal studies, where observations are collected over time for the same cases (e.g., buildings, households, organizations), enhance understanding of interdependency challenges in hazard recovery and community resilience. This in turn supports efforts to quantify correlations between population dislocation and race/ethnicity, income, and education level, and thus inform recovery modeling in the IN-CORE computational environment (van de Lindt et al. 2018b).

Community Resilience Challenges

Challenges to community resilience assessment are numerous and vary among members of the resilience community. The following are generally recognized as common challenges.

Verification and Validation

The IN-CORE computational environment integrates engineering, social science, and economic models, and their coupling raises a number of practical difficulties. Among them, engineering and social science models are typically of different scales (e.g., region, parcel, neighborhood, buildings); and while engineering models are based on laws of physical science or mechanics, social science models are highly data driven. These differing models must communicate with each other within and across IN-CORE modules.

Because investments in community resilience and risk mitigation may be significant and must be amortized over decades, confidence in IN-CORE’s ability to capture the likely performance of physical and social infrastructure with reasonable accuracy is essential. Verification and validation are a multifaceted process that involves testing the individual modules in figure 2 and their linkages, and comparing predictions with data from hindcasts and field studies.

Financing Community Recovery and Resilience

How will communities pay for resilience improvements and recovery? This question is difficult to address.

Disaster response and recovery are financed mainly by the federal government (e.g., the Federal Emergency Management Agency, Housing and Urban Development, Small Business Administration), nonprofits, and insurance. These programs may be reactive, in that they generally are not mobilized until after a disaster has occurred, or they may support proactive measures before a natural hazard event. Community resilience planning to reduce risk can be improved significantly with proactive measures. New opportunities for proactive resilience planning and implementation were addressed by the Disaster Recovery Reform Act of 2018 (Ellard et al. 2018).

Financing to enhance community resilience typically requires a variety of funding sources in the public and private sectors. Little consideration has been given to drawing on the capital markets, although the funds potentially available overshadow resources from government programs or the insurance industry (Goda 2015).

Risk Analysis and Risk-Informed Decision Support

The risk-informed decision process is poorly understood for community resilience planning with integrated physical, social, and economic models. IN-CORE will facilitate risk-informed decisions that take into account likelihoods and consequences of natural hazard events. Risk attitudes of decision makers also vary by context; the risk tolerance of communities whose inhabitants have been exposed to a prior disruptive event may be different from that of inhabitants who have not.

In addition, the manner in which risk is communicated from the results of the IN-CORE analysis is at least as important as the manner in which it is assessed technically (NRC 1989); risk goals must be adaptable to individual communities. A typical community resilience goal and metric might read: “The population stability goal will be met if 95% of the population can resume residency four weeks after event H with 90% confidence.”

Finally, existing decision methods to support community resilience goals require modifications to achieve sustainable long-term decisions for enhancing community resilience, when the investment horizon extends to 50–100 years or more.

Best Practices for Community Resilience

Community resilience assessment and enhancement may require a variety of measures, such as

  • improved codes and standards that meet resilience goals for community buildings, bridges, and other infrastructure and are consistent with current engineering practice (Ellingwood et al. 2019; Lin et al. 2016);
  • land use policies;
  • incentives such as tax credits and discounts on insurance for individual homeowners and businesses to enhance the resilience of their properties; and
  • educational programs for property owners and future resilience practitioners on methods of resilience enhancement.

Meeting community resilience goals often involves an economic analysis component to understand the cost of policy decisions. Questions of whether up-front costs are justified by future risks and who pays the costs and who receives the benefits drive the debate in most communities.

Conclusion

Modern communities consist of closely integrated civil infrastructure systems and social and economic institutions that are essential for the health and welfare of their inhabitants. The Center’s overarching goals are to establish the measurement science for identifying and understanding the factors that make a community resilient, to develop a computational environment with fully integrated supporting databases to assess the likely impact of natural hazards on communities, and to develop risk-informed decision support strategies for optimal solutions. These goals are supported by collaborative field studies and research to improve IN-CORE and its supporting databases.

Readers interested in learning more are encouraged to visit the Center’s website,[1] which includes a list of Center publications and yearly webinars hosted for the resilience community.

Acknowledgments and Disclaimer

Funding for this study was provided as part of Cooperative Agreement 70NANB15H044 between NIST and Colorado State University. The paper represents the views of the authors, not necessarily those of NIST or the US Department of Commerce. The authors acknowledge the numerous researchers and students working on behalf of the Center of Excellence.[2] Cameron Fletcher’s comments and edits were helpful in making our article more accessible to the readers of The Bridge.

References

ASCE [American Society of Civil Engineers]. 2015. Adapting Infrastructure and Civil Engineering Practice to a Changing Climate. Reston VA.

ASCE. 2016. Minimum Design Loads and Associated Criteria for Buildings and Other Structures (Standard 7-16). Reston VA.

Attary N, Unnikrishnan V, van de Lindt JW, Cox DT, Barbosa A. 2017. Performance-based tsunami engineering methodology for risk assessment of structures. Engineering Structures 141:676–686.

Attary N, van de Lindt JW, Mahmoud H, Smith S. 2018. Hindcasting community-level damage to the inter-dependent buildings and electric power network after the 2011 Joplin, Missouri, tornado. Natural Hazards Review 20(1):04018027.

Bruneau MS, Chang S, Eguchi R, Lee G, O’Rourke T, -Reinhorn AM, Shinozuka M, Tierney K, Wallace WA, Wintervelt DV. 2003. A framework to quantitatively assess and enhance the seismic resilience of communities. Earthquake Spectra 19(4):733–752.

Ellard WH, Zaltsberg MF, Abbott EB. 2018. The Disaster Recovery Reform Act of 2018: Key Provisions Affecting FEMA Funding, October 12. Jackson MS: BakerDonelson.

Ellingwood BR, Cutler H, Gardoni P, Peacock WG, van de Lindt JW, Wang N. 2016. The Centerville virtual community: A fully integrated decision model of interacting physical and social infrastructure systems. Sustainable and Resilient Infrastructure 1(3-4):95–107.

Ellingwood BR, Wang N, Harris JR, McAllister TP. 2019. -Performance-based engineering to achieve community resilience. In: Routledge Handbook of Sustainable and Resilient Infrastructure, ed Gardoni P. New York: -Routledge.

Gardoni P, van de Lindt JW, Ellingwood BR, McAllister TP, Lee JS, Cutler H, Peacock W, Cox DT. 2018. The Interdependent Networked Community Resilience Modeling Environment (IN-CORE). Proceedings of 16th European Conference on Earthquake Engineering, Jun 18–21, -Thessaloniki, Greece.

Goda K. 2015. Seismic risk management of insurance portfolio using catastrophe bonds. Computer-Aided Civil and Infrastructure Engineering 30:570–582.

Guidotti R, Gardoni P, Chen Y. 2017. Multi-layer heterogeneous network model for interdependent infrastructure systems. Proceedings, 12th International Conference on Structural Safety and Reliability (ICOSSAR 2017), Aug 6–10, Vienna.

Hamideh S, Peacock WG, Van Zandt S. 2018. Housing recovery after disasters: Primary versus seasonal and vacation housing markets in coastal communities. Natural Hazards Review 19(2).

ICC [International Code Council]. 2018. International Building Code. Washington.

Koliou M, van de Lindt JW, McAllister TP, Ellingwood BR, Dillard M, Cutler H. 2018. State of the research in community resilience: Progress and challenges. Sustainable and Resilient Infrastructure, January.

Lin P, Wang N, Ellingwood BR. 2016. A risk de-aggregation framework that relates community resilience goals to building performance objectives. Sustainable and Resilient Infrastructure 1(1):1–13.

Lounis Z, McAllister TP. 2016. Risk-based decision making for sustainable and resilient infrastructure systems. ASCE Journal of Structural Engineering 142(9):F4016005.

Mieler M, Stojadinovic B, Budnitz R, Comerio M, Mahin S. 2015. A framework for linking community-resilience goals to specific performance targets for the built environment. Earthquake Spectra 31:1267–1283.

NIST [National Institute of Standards and Technology]. 2016. Community Resilience Planning Guide for Buildings and Infrastructure Systems, Vols 1 and 2. Special Publication 1190. Gaithersburg MD.

NRC [National Research Council]. 1989. Improving Risk Communication. Washington: National Academy Press.

NRC. 2012. Disaster Resilience: A National Imperative. Washington: National Academies Press.

van de Lindt JW, Ellingwood B, McAllister T, Gardoni P, Cox D, Peacock W, Cutler H, Dillard M, Lee J, Peek L, and 2 others. 2018a. Modeling community resilience: Update on the Center for Risk-Based Community Resilience -Planning and the computational environment IN-CORE. Proceedings of the 17th US-Japan-New Zealand Workshop on the Improvement of Structural Engineering and Resilience, Nov 12–14, Queenstown, New Zealand.

van de Lindt JW, Peacock WG, Mitrani-Reiser J, eds. 2018b. Community resilience-focused technical investigation of the 2016 Lumberton, North Carolina Flood: Multi-disciplinary approach. Special Publication 1230. -Gaithersburg MD: National Institute of Standards and Technology.

White House. 2013. Presidential Policy Directive/PPD-21: Critical Infrastructure Security and Resilience. -Washington.

Zhang W, Nicholson C. 2016. A multi-objective optimization model for retrofit strategies to mitigate direct economic loss and population dislocation. Sustainable and Resilient Infrastructure 1(3-4):123–136.

Zhang W, Lin P, Xue X, Wang N, Nicholson CD. 2017. Probabilistic prediction of postdisaster functionality loss of community building inventories considering utility disruptions. ASCE Journal of Structural Engineering 144(4).


[1]  www.resilience.edu/publications_list.php

[2]  A full list is available at http://resilience.colostate.edu/-researchers.shtml.

About the Author:Bruce Ellingwood (NAE) is College of Engineering Eminent Scholar and codirector of the Center for Risk-Based Community Resilience Planning in the Department of Civil and Environmental Engineering at Colorado State University. John van de Lindt is Harold H. Short Endowed Chair in the Walter Scott Jr. College of Engineering at Colorado State University and codirector of the Center for Risk-Based Community Resilience Planning. Therese McAllister is Community Resilience Group leader and program manager in the Materials and Structural Systems Research Division of the Engineering Laboratory at the National Institute of Standards and Technology.