Download PDF 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. Resiliently Engineered Flood and Hurricane Infrastructure: Principles to Guide the Next Generation of Engineers Monday, July 1, 2019 Author: Gregory Baecher, Michelle Bensi, Allison Reilly, Brian Phillips, Lewis (Ed) Link, Sandra Knight, and Gerald Galloway The hurricane and flood disasters of recent decades have created a paradigm shift in how engineers approach natural hazards. The federal and some local governments have moved from standards-based approaches to developing risk-informed project plans, recognizing the importance of uncertainty in planning, the inevitability of disaster events, and the inability to provide absolute protection. Coupled with this is the concept of resilience—the -ability of a system to function when exposed to disruptions—a concept long considered in ecology, sociology, and other disciplines. As the 21st century presents both challenges from climate change and population growth and opportunities through technology acceleration, the task is to determine how engineers will leverage risk-informed approaches to mitigate flood and hurricane impacts. This article discusses the transition from the 20th to a 21st century paradigm, and the impact of that transition on the education of engineers. Figure 1 Background The number and cost of disaster events in the United States has risen significantly over the past few decades (NOAA/NCEI 2019), and those related to floods and hurricanes are especially high (figure 1). For example, just since 2012 the damage from Superstorm Sandy in 2012 was record-breaking in physical impacts and caused $47.5 billion in inflation-adjusted losses in New York and New Jersey. a 2014 pluvial flooding event in Michigan, caused by exceptional rainfall and inadequate drainage, was estimated to cost over $1 billion (NOAA/NCEI 2019). Hurricane Harvey in 2017 brought unprecedented rainfall, basinwide flooding, and windstorms that devastated infrastructure and flooded more than 150,000 homes, 46 percent of which were outside the FEMA 500-year floodplain (Galloway et al. 2018). Hurricane Michael, which devastated northwest Florida in 2018, was the strongest storm on record for that area and led to an estimated $25 billion in costs (NOAA/NCEI 2019). These weather and climate events are diverse in their phenomenological origin as well as their impacts on communities. But it is expected that their frequency, severity, and cost will continue to increase because of climate change and factors such as land use and land cover changes, short-sighted planning and development, and federal disaster policy that emphasizes recovery rather than loss reduction. The Transformation of Disaster Mitigation During the 20th century engineering design -primarily used resistance to minimize damage from natural -hazards. Such designs provide adequate solutions when systems are exposed to hazards within the envelope of the scenarios and loads for which the systems were designed. However, when systems are exposed to scenarios that differ from those envisioned, resistance-based solutions are often inadequate. Moreover, such solutions are often associated with “cliff-edge” effects in which the adverse consequences of events may be exacerbated by changes in hazard conditions. The escalation of disaster losses has resulted in recognition of the need for a 21st century paradigm that supports resilience for a broad range of evolving natural threats. A resilient community or organization is one that is prepared for hazard events, is able to respond and recover when events occur, and can adapt to changes in hazard exposure (NRC 2012). Resilience engineering is integral to this new paradigm: It approaches disaster loss reduction and avoidance using multifaceted approaches that cross traditional disciplinary boundaries. How can engineering accommodate the transition to a new paradigm, and what will be the impact of that transition on education? What steps is the profession taking to recognize the realities of hazards, understand how people and infrastructure respond to hazards, and work within resource limitations? Engineering approaches to dealing with floods and hurricanes are constantly evolving, albeit in a reactionary way. Nations have moved from structural flood control to multiapproach flood damage reduction to flood risk management. After devastating losses in the -Netherlands in 1953 from the North Sea flood and in the United States in 2005 from Hurricane Katrina, design approaches and engineering decision making began to change with new techniques. Safety factors have given way to reliability-based codes and are beginning to yield to performance-based analysis. The realization has come that there is no such thing as certain protection. What Is Resilience Engineering? The terms resilience and resilience engineering are used in a number of disciplines with definitions and approaches that differ by field and research focus. In general, resilience engineering focuses on the complexities of real-world systems (e.g., inter-dependencies, human-system interactions, and regulatory constraints), the evolving nature of hazards and systems, the strengths and limitations of conventionally engineered solutions, and the importance of working across disciplines to ensure that communities, organizations, and individuals are able to prepare for, respond to, recover from, and adapt to disruptive events. These dimensions require new first principles to guide resilience engineers, as explained in a later section. From the perspective of resilience engineering, a key element of resilient societies is functioning and reliable infrastructure to provide the essential services on which individuals, communities, and organizations depend. From transportation to health care, infrastructure comprises the facilities, systems, and networks necessary for society to function (DHS 2013). Yet many infrastructure systems are facing increasing service demands while deteriorating as a result of aging and inadequate maintenance (ASCE 2017). Based on a new paradigm of risk-informed design and recognition of the interconnectivity of infrastructure and societies, infrastructure systems of the future will be designed differently, in tandem with land-use planning, environmental consideration, and social and equity factors. Risk will inform planning and decision making, and there will be greater integration of objectives with multiple attributes and more focus on lifecycle planning, flexibility and future options, and resilience. How Does Resilient Infrastructure Differ from Today’s Infrastructure? In the past, infrastructure was designed to meet basic public needs using codes and guidelines. It now faces increasing natural hazards and other uncertainties, and must remain viable both in the face of such events and within a systems context. Going forward, design and analysis decisions need to include social, economic, and environmental factors in addition to code requirements. Resilient infrastructure design must accommodate lifecycle relationships, consider destructive forces of natural hazards, incorporate end-of-use disposal, and be based on not only the engineering features but also the interrelationship among these features and other systems, including environmental systems. Furthermore, community, values, equity, and social responsibilities play increasing roles in engineering planning and design. Because resilient development for a community or organization requires integration of the multiple sectors that support those entities, no aspect of infrastructure development can be accomplished in isolation (NRC 2012). Failure to incorporate social, environmental, economic, and engineering elements and to adapt to new technologies and approaches will limit, and perhaps even negate, the effectiveness of resilience efforts. To be successful, the engineering community must join other disciplines in eliminating silos in education, engineering, government, and other sectors. Dealing with Complexities of Interconnected Systems A particular challenge is that infrastructure systems may be distributed over large geographies, so their hazard exposure is greater than that of single-site facilities and they are exposed to a wider variety of hazards. Moreover, hazards can affect multiple components of an infrastructure system simultaneously. Neglecting to account for correlated loads and responses and for network effects and physical dependencies may result in inefficient allocation of resources and little improvement in system reliability. While sector-specific interdependencies are often well known, cross-sector interdependencies are not. A failure in one sector typically leads to failures in others -(Rinaldi et al. 2001), so infrastructure must be designed and regulated with consideration of both hazards and dis-tributed network effects created by tightly coupled systems. -Lifeline networks such as transportation, water distribution, and communications are particularly susceptible to natural hazards and may exhibit cascading failures. Infrastructure is also broadly owned or managed by distinct owners and operators, and thus subject to many, at times conflicting, regulations. Planning objectives may focus on bettering one owner’s system with little consideration of how it interrelates with other infrastructure and societal function. This leads to resource inefficiencies and stymies collaborative approaches to resilience. Current regulatory frameworks reinforce these tendencies by focusing on sector-specific performance (Reilly et al. 2015). Moving Beyond Codes The adoption and enforcement of building codes is left to individual states. In 2018 the Insurance Institute for Business and Home Safety ranked 18 Atlantic and Gulf Coast states on a scale of 0 (lowest) to 100 (highest) on the effectiveness of their residential building code adoption and enforcement programs (IBHS 2018). Scores ranged from 17 in Delaware to 95 in Florida, and the study reported that Alabama, Delaware, Georgia, Maine, Mississippi, New Hampshire, New York, and Texas have no statewide enforced residential building code. Mitigation of damage from hurricanes will require more consistent adoption and enforcement of building codes. It is clear from recent hurricanes that buildings designed to modern codes demonstrate less frequent and less severe damage (IBHS 2004). Often simple solutions such as better window protection, stronger roof-to-wall connections, and stricter roof sheathing schedules significantly reduce wind-induced damage. But current building codes do not address the uncertain hazard landscape of the future. Engineers need to better understand the impacts of climate change and how to adapt as it unfolds. Current Tools and Principles Design practice for water resources infrastructure has been based on either limits specified in codes or the best practice application of analytical tools using experience as a guide. While informed by statistical treatment of data, application of probabilistic methods has been limited to hazard characterization. Factors of safety have been the primary approach to dealing with uncertainty, which was seldom fully estimated or understood. Hurri-cane Katrina and other major hazard events demonstrated the limitations of this approach and stimulated interest in adopting reliability-based approaches to improve planning and design. Federal flood insurance rate maps (FIRMs) are the most common means of communicating flood hazard information. FIRMs delineate flooding hazards from -fluvial and coastal flooding, but do not generally capture hazards from pluvial flooding or capacities of urban storm water systems to manage heavy onslaughts of water. About 25 percent of all US flood insurance claims come from areas that FIRMs indicate have low to moderate flood risks (Galloway et al. 2018). In fact, most homes flooded during Hurricane Harvey were outside the FEMA-designated 500-year floodplain. Use of the FIRMs without recognition of their limitations may lead individuals, businesses, or communities to make suboptimal or even harmful decisions about mitigation and avoidance of pluvial flood hazards. Hazard Characterization Hurricanes, storms, and other meteorological events present complex threats. Floodwater-induced damage can be caused by surge, waves, inundation, and other effects (e.g., debris, sediment deposition, and erosion). Wind-induced damage results from wind pressure, debris impact, and wind-driven rain. Hazard characterizations often focus on the meteorological storm (e.g., the Saffir-Simpson category of a hurricane[1]) or the total rainfall volume and intensities of storms. But the hurricane category or rainfall volume are incomplete characterizations of the forces that impact infrastructure and people—hurricane intensity does not accurately convey storm surge potential nor is rainfall fully representative of potential flood elevations for an area. Datasets are available for tropical cyclone events, rainfall, and wind, but challenges arise when filtering these datasets for applicability to individual locations. Probabilistic hazard approaches are evolving to (a) integrate statistical analysis of historical data with physical process knowledge and models and (b) leverage new tools such as surrogate modeling based on machine learning -methods (e.g., Jia et al. 2016) and ensemble modeling. Moreover, while single hazard assessment approaches often dominate design, events may involve combined hazards, as in the case of Superstorm Sandy and Hurricane Harvey. There have been significant advances in characterizing hazard forces such as storm surge, particularly in terms of coastal geometry, storm path, and storm metrics beyond central pressure deficit. High-resolution computational grids and depth-limited two-dimensional circulation models coupled with sophisticated wave models have raised the bar considerably. These model packages have been applied to ensembles of hypothetical hurricanes to develop distributed hurricane storm surge datasets for application with joint probability methods to advance both probabilistic hazard definition and risk assessments (Irish et al. 2009). This combined approach was a fundamental tool for the postevent analysis of Hurricane Katrina and the development of distributed risk estimates to inform the design of new risk reduction structures that were completed in 2011 (IPET 2009, p. 61). Climate change is expected by many researchers to alter the frequency, duration, and intensity of North Atlantic hurricanes (e.g., Cui and Caracoglia 2016; Mallakpour and Villarini 2015; Webster et al. 2005). Warmer sea surface temperatures will extend the Atlantic hurricane season and broaden the area over which hurricanes form and strengthen. Climate change is already bringing wetter hurricanes (figure 2), with greater storm surge and freshwater flooding. Figure 2 Increased sophistication in the application of probabilistic risk analysis to flood risk reduction provides a more comprehensive picture of near-term situations and significantly aides decision making. However, nonstationarity is of growing concern with hurricanes and floods. Nonstationarity can be briefly defined as -changes in long-term means and other trends. Although risk analysis techniques can, in principle, be used to grapple with it, they have not been. Furthermore, while non-stationarity is often associated with climate change, other forms—such as changing land use, population, and property values—are also relevant. One option to deal with limited data, uncertainty, and nonstationarity is scenario analysis. An example of this approach is the Netherlands’ Delta Program, which uses adaptive scenarios to deal with the implications of climate change for water management.[2] The -scenarios explore four types of future situations to determine where and when alternative measures will be required in the future. New First Principles of Resilient Design Resilient infrastructure design requires a focus on first principles of a new paradigm. Whereas earlier generations of engineers were trained in balancing loads versus capacities and in optimizing benefits versus costs, resilience engineers will need to use the first principles of resilient design: integrate physical and social design considerations, quantify and incorporate uncertainty, use systems-level thinking and planning for diversity and redundancy, explicitly include options and adaptability in design decisions, and leverage nonstructural and nature-based alternatives. Integration of physical and social design considerations requires that behavioral responses and vulnerabilities be incorporated explicitly in planning and design. Engineers must recognize and embrace the knowledge that effective resilience means that not only physical systems but also individuals, communities, and organizations can cope with and respond to disasters. Creating resilient communities entails enhancing and building adaptive capacities through investments in preparedness and understanding the social factors that mediate community response. While engineers are not typically trained to have in-depth knowledge of these social dimensions and the need for community engagement, it is important that they understand the need to consider these factors and seek support and guidance from experts in the social sciences, economics, citizen outreach, and related fields as part of design and planning processes. Recognizing and accommodating uncertainty is central to the principles above but remains challenging because of a lack of both guidance on standard tools and understanding of nonstationarity. Resilience engineers must consider systems-level interactions in design and plan for diversity and redundancy. Understanding such interactions means understanding system inter-dependencies and how humans and physical systems interact. Redundancy means that parallel and independent means are adopted. The inclusion of options and adaptability in planning allows for flexibility to cope with uncertain futures; scenarios of how the future might unfold provide a platform for including future options. Nonstructural interventions and nature-based design increase resilience by reducing the consequences of flooding rather than its probability (the latter is commonly the goal of structural measures such as levees and seawalls). Nonstructural measures might include modifications to public, regulatory, or pricing policy in the National Flood Insurance Program; buyouts and relocations; changes to land-use regulations and permitting; flood proofing of structures; warning systems; and preparedness planning. Nature-based measures might include preservation of wetlands, development of vegetation dams on dunes, introduction of aqua-culture, reforestation, and park creation along waterways (Bridges et al. 2015, p. 479). Educating Resilience Engineers The need to train the next generation of engineers for resilience faces an old tension in engineering education: the balance between breadth and specialization. Undergraduate engineering programs must maintain broad traditional disciplinary coverage at appropriate depth while both incorporating emerging areas such as data science and computing and adhering to the broad and demanding requirements of professional associations like ASCE, accrediting authorities such as ABET, and professional licensure. To address the needs of engineering for resilience, programs must add exposure to the social sciences, uncertainty and risk analysis, economics and finance, ecological engineering, and decision making under uncertainty. This is a substantial challenge. Moreover, strategies for engineering for resilience continue to evolve, requiring curricula that can be adapted to this changing knowledge base. Finally, administrative and financial barriers often exist when students are encouraged or required to take courses across conventional academic divides. New administrative strategies and a reimagining of curricula (e.g., through development of new courses that transcend conventional engineering silos) are needed to ensure that academic programs offer a more comprehensive, up-to-date undergraduate curriculum while remaining sensitive to the rising costs of education. Finally, educators must carefully consider whether master’s-level engineering specialization is needed to prepare competent resilience engineers. Conclusions Infrastructure needs have traditionally been addressed with assumptions and practices that oversimplify uncertainties underlying complex interactions among social, environmental, and physical domains. Engineers have assumed stationarity, relied on short datasets, applied one-size-fits-all standards and factors of safety, and generally designed and built for “human” increments of time. Knowledge has grown and demands on infrastructure have changed, and current approaches and tools do not meet the challenges of the future. The concept of major hurricane and flood risk reduction infrastructure projects with an n-year life, designed for a very specific set of loads and demands, is contrary to the increased pace of change and inherent uncertainties in the three domains. What’s needed is a paradigm shift to engineering that embraces uncertainty and exploits multifaceted scenario planning and analysis, leading to incremental adaptive designs and infrastructure. The shift toward this new paradigm is underway. The new risk reduction infrastructure in New Orleans incorporated uncertainty in the fundamental design; probabilistic approaches were used to reduce risks associated with overtopping and catastrophic breaching. But far from being a model for future design, it only addressed contemporary challenges. And as with all new paradigms, it creates conflict and confusion with respect to existing standards and practice as well as educational content. The Netherlands’ Delta Program and Delta Model initiatives incorporate risk analysis in a multifaceted scenario analysis that includes social, environmental, and physical factors to develop one of the first examples of a risk-informed, large-scale, adaptive strategy for changes in the water regime due to social/cultural and climate changes. The program depends, however, on avoiding unreasonable restrictions on the magnitude of the changes examined in the scenarios, which can limit the infrastructure approaches (structural and non-structural) considered for adaptation. This could be a significant challenge to achieving the necessary paradigm, given the propensity of current practice to simplify and optimize based on limited resources, time constraints, stove-piped organizational responsibilities, and convoluted governance systems. Ultimately, any approach that is adopted must recognize that a cornerstone of resilient communities is robust infrastructure and infrastructural policies that consider the interconnections among people, hazards, and the natural and built environment. References ASCE [American Society of Civil Engineers]. 2017. ASCE’s 2017 Infrastructure Report Card | GPA: D+. Online at www.infrastructurereportcard.org. Bagstad KJ, Stapleton K, D’Agostino JR. 2007. Taxes, subsidies, and insurance as drivers of United States coastal development. Ecological Economics 63:2–3. Bridges TS, Wagner PW, Burks-Copes KA, Bates ME, -Collier ZA, Fischenich CJ, Gailani JZ, Leuck LD, Piercy CD, -Rosati JD, and 5 others. 2015. Use of Natural and Nature-Based Features (NNBF) for Coastal Resilience (ERDC SR-15-1). Vicksburg MS: US Army Engineer Research and Development Center. Cui W, Caracoglia L. 2016. Exploring hurricane wind speed along US Atlantic coast in warming climate and effects on predictions of structural damage and intervention costs. Engineering Structures 122:209–225. DHS [US Department of Homeland Security]. 2013. -National Infrastructure Protection Plan (NIPP) 2013: Partnering for Critical Infrastructure Security and Resilience. -Washington. Galloway GE, Reilly A, Ryoo S, Brody S, Highfield W, Gunn J, Rainey J, Parker S. 2018. The Growing Threat of Urban Flooding: A National Challenge. College Park and -Galveston: University of Maryland and Texas A&M University. IBHS [Insurance Institute for Business & Home Safety]. 2004. Hurricane Charley: Nature’s Force vs. Structural Strength. Tampa. IBHS. 2018. Rating the States 2018: An Assessment of Residential Building Code and Enforcement Systems for Life Safety and Property Protection in Hurricane-Prone Regions—Atlantic and Gulf Coast States. Tampa. IPET [Interagency Performance Evaluation Task Force]. 2009. A General Description of Vulnerability to Flooding and Risk for New Orleans and Vicinity: Past, Present, and Future. Washington: US Army Corps of Engineers. Irish JL, Resio DT, Cialone MA. 2009. A surge response function approach to coastal hazard assessment. Part 2: Quantification of spatial attributes of response functions. Natural Hazards 51(1):183–205. Jia G, Taflanidis AA, Nadal-Caraballo NC, Melby JA, -Kennedy AB, Smith JM. 2016. Surrogate modeling for peak or time-dependent storm surge prediction over an extended coastal region using an existing database of synthetic storms. Natural Hazards 81(2):909–938. Mallakpour I, Villarini G. 2015. The changing nature of flooding across the central United States. Nature Climate Change 5(3):250–254. Melillo JM, Richmond TC, Yohe GW, eds. 2014. Climate Change Impacts in the United States: The Third National Climate Assessment. US Global Change Research Program. Washington: US Government Printing Office. NOAA/NCEI [National Oceanic and Atmospheric Administration/National Centers for Environmental Information]. 2019. US Billion-Dollar Weather and Climate Disasters. Asheville NC. NRC [National Research Council]. 2012. Disaster Resilience: A National Imperative. Washington: National Academies Press. Reilly A, Samuel A, Guikema S. 2015. Gaming the “system”: Decision making by interdependent critical infrastructure. Decision Analysis 12(4):155–172. Rinaldi SM, Peerenboom JP, Kelly TK. 2001. Identifying, understanding, and analyzing critical infrastructure interdependencies. IEEE Control Systems 21(6):11–25. Webster PJ, Holland GJ, Curry JA, Chang H-R. 2005. -Changes in tropical cyclone number, duration, and intensity in a warming environment. Science 309(5742):1844–1846. [1] This is the wind scale ranking behind the designation of, for example, a Category 4 hurricane. [2] https://www.government.nl/topics/delta-programme About the Author:Gregory Baecher (NAE) is a professor, Michelle Bensi and Allison Reilly are assistant professors, Brian Phillips is an associate professor, Lewis (Ed) Link and Sandra Knight are senior research engineers, and Gerald Galloway (NAE) is a professor, all in the Department of Civil and Environmental Engineering at the University of Maryland Clark School of Engineering.