Investing in Transportation Resilience: A Framework for Informed Choices (2021)

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.

19 The increasing threat that natural hazards pose to the nation’s transporta- tion infrastructure and mobility varies by region and by mode. To boost transportation resilience, policy makers and infrastructure decision makers need a solid understanding of the specific hazards that the transportation systems under their purview face. Resilience analysis must therefore begin with evaluations of these hazards, paying special attention to the most acute and severe events—commonly called disasters—but also accounting for the effects of changing environmental conditions such as from climate change. Natural disasters, and their accompanying economic losses, are on the rise. The United States experienced a record-breaking 22 billion-dollar nat- ural disasters in 2020, according to the National Oceanic and Atmospheric Administration (NOAA).1 Hurricane Laura, the California wildfires, and the Midwestern derecho2 were the leading contributors to the $95 billion in losses. During the 1980s, billion-dollar events, even after adjustments 1 NOAA’s calculations of billion-dollar events are adjusted for inflation. See NOAA NCEI (National Centers for Environmental Information). 2021. “U.S. Billion-Dollar Weather and Climate Disasters.” https://doi.org/10.25921/stkw-7w73. 2 The National Weather Service describes a derecho as “a widespread, long-lived wind storm that is associated with a band of rapidly moving showers or thunderstorms. Although a derecho can produce destruction similar to the strength of tornadoes, the damage typically is directed in one direction along a relatively straight swath. As a result, the term ‘straight-line wind damage’ sometimes is used to describe derecho damage. By definition, if the wind damage swath extends more than 240 miles (about 400 kilometers) and includes wind gusts of at least 58 mph (93 km/h) or greater along most of its length, then the event may be classified as a derecho.” See https://www.weather.gov/lmk/derecho. 2 Natural Hazards, Climate Change, and America’s Transportation Infrastructure

20 INVESTING IN TRANSPORTATION RESILIENCE for inflation, averaged only 2.9 per year. By the 2010s, the average reached $11.9 billion in disasters per year. Transportation agencies are on the front lines when natural disasters of all sorts strike. In the wake of the 2020 derecho, the Iowa Department of Transportation (DOT) sent crews from 50 garages to haul tens of thou- sands of loads of debris.3 In January 2021, the nation watched, transfixed, as Caltrans released drone footage of Big Sur’s Highway 1, wiped out in a flood at Rat Creek. Vegetation that resisted erosion had been destroyed in 2020’s Dolan Fire.4 The Rat Creek washout, although the most devastating after the Dolan Fire, was one of 50 landslides on the highway requiring cleanup and repair. The impacts of hurricanes often ripple across wider freight markets and supply chains. In addition to direct damage disrupting service, freight capacity—including trucking and at ports—can be diverted to emergency relief. Hurricane Laura disrupted freight service by damaging the rail net- work around Lake Charles, Louisiana.5 When an unprecedented hurricane struck the coast of California in 2014, the Port of Long Beach saw opera- tions at a standstill for days, and it took several months for the nearby roads and facilities to be fully restored.6 Hurricane Harvey’s tremendous rainfall disrupted most road travel for days, but emergency preparedness efforts among the public sector, industrial sectors, and the Port of Houston prevented, one study concluded, what “could have been some major prob- lems that could have devastated local, regional, and even national supply chains.”7 Modeling conducted for the Fourth National Climate Assessment in- dicates that the increasing danger from natural hazards will be a long- term trend due to increasing emissions and atmospheric concentrations of 3 Iowa DOT (Department of Transportation). 2020. “Iowa DOT Answers the Call for Debris Removal Following Devastating Derecho.” Transportation Matters for Iowa, August 27. https://www.transportationmatters.iowadot.gov/2020/08/iowa-dot-answers-the-call-for- debris-removal-following-devastating-derecho.html. 4 Alexander, K. 2021. “Highway 1 Through Big Sur Will Be Repaired.” San Francisco Chronicle, February 10. https://www.sfchronicle.com/environment/article/In-Big-Sur-rain- came-down-and-so-did-Highway-1-15938072.php. 5 Straight, B. 2020. “Rail Service Still Hampered, But Truck Stops, Roadways Reopened Following Hurricane Laura.” Freight Waves, August 29. https://www.freightwaves.com/news/ rail-service-still-hampered-but-truck-stops-roadways-reopened-following-hurricane-laura. 6 Port of Long Beach. 2016. Climate Adaptation and Coastal Resiliency Plan. https://www. slc.ca.gov/wp-content/uploads/2018/10/POLB.pdf. 7 NASEM (National Academies of Sciences, Engineering, and Medicine) 2020. Strengthen- ing Post-Hurricane Supply Chain Resilience: Observations from Hurricanes Harvey, Irma, and Maria. Washington, DC: The National Academies Press, p. 28. https://doi.org/10.17226/25490.

AMERICA’S TRANSPORTATION INFRASTRUCTURE 21 greenhouse gases.8 For roads, both the Assessment’s high-emission and low- emission scenarios show increased costs, cumulatively up to an additional $230 billion through 2100, just to repair damage attributed to changes in temperature, precipitation, and freeze-thaw cycles. For bridges, the primary danger is scour, where the flow of water undermines the integrity of the bridge piers. Under the high-emission scenario, 4,600 road bridges will be vulnerable in 2050 and 6,000 in 2090. Even in the low-emission scenario, 5,000 bridges will be vulnerable in 2090. For rail, extreme heat threatens to delay freight and passenger trains alike.9 Cumulative costs of increased railroad delays through 2100 are $50 billion in the high-emission scenario and $40 billion in the low-emission scenario.10 Whether a hazard causes harm depends on the characteristics of the infrastructure and a society’s preparation and ability to respond. Resilience analysis, planning, and management require an understanding of natural hazards and climate change effects, including their likelihood and character- istics. Transportation agencies that analyze natural hazards use a range of methods from qualitative descriptions to quantitative probabilistic models. All of these methods must accommodate the reality that while natural haz- ards are a fact of life, there is still a great deal of uncertainty about where, when, and how the next natural hazard will strike. This chapter provides an introduction to natural hazards, a descrip- tion of how some are affected by climate change, and a discussion of the impacts of both on transportation. The chapter begins with an explana- tion of why an understanding of natural hazards—their likelihood and characteristics—is key to building resilience. The chapter provides a brief overview of how meteorological, geological, and climate change–related hazards affect transportation in the United States, including how they vary by region and location. To lay the groundwork for resilience metrics, the chapter then reviews the basics of measuring hazard likelihood, the aspects that go into hazard characterization, and the approaches used to develop hazard scenarios that can be integrated into resilience analysis. The chapter 8 Reidmiller, D.R., C.W. Avery, D.R. Easterling, K.E. Kunkel, K.L.M. Lewis, T.K. Maycock, and B.C. Stewart, eds. 2018. Impacts, Risks, and Adaptation in the United States: Fourth National Climate Assessment, Volume II. Washington, DC: U.S. Global Change Research Program. https://doi.org/10.7930/NCA4.2018. 9 Extreme heat causes the steel in rails to expand and buckle, causing trains to derail. Ex- treme cold causes the steel to contract and crack, similarly causing derailments. 10 Cumulative costs are in addition to a base calculated from 1950 to 2015, in 2015 dol- lars, and discounted 3% annually. See EPA (U.S. Environmental Protection Agency). 2017. Multi-Model Framework for Quantitative Sectoral Impacts Analysis: A Technical Report for the Fourth National Climate Assessment. EPA 430-R-17-001. Washington, DC: U.S. Environ- mental Protection Agency, pp. 74–99.

22 INVESTING IN TRANSPORTATION RESILIENCE concludes with a discussion of the role of federal investment in providing data for hazard modeling and projections. TO BUILD RESILIENCE—FIRST, UNDERSTAND THE HAZARD All approaches to evaluating resilience to inform transportation investment decisions require knowledge of the natural hazards. Because natural haz- ards vary across the landscape and in their interaction with transportation modes, transportation agencies are often required to conduct individual- ized analysis of relevant hazards and their likely effects. Climate change compounds the difficulty of analyzing hazards because the analysis can no longer assume that the forces that produce the natural hazards are stable. Climate change also introduces shifts to normal environmental conditions, which must also be taken into account. The importance of understanding hazards is exemplified by Step 1, “Explore Hazards,” of the U.S. Climate Resilience Toolkit’s advisor on Steps to Resilience.11 The word “explore” communicates that best practice is not just to jump into the most detailed analysis possible to build a comprehensive list of all conceivable hazards. Instead, Step 1 includes “investigate regional climate” and “understand exposure.” Hazard analysis typically searches for significant hazards and for infrastructure assets that are most vulnerable to damage and disruption. The toolkit pulls together resources from across the federal government that can aid in identifying the potential natural hazards or climate changes for a given region or community. The analysis of natural hazards focuses on two separate but interrelated questions, both of which wrestle with uncertainty. One inherent feature of natural hazards is that while we know, generally, that they will occur, we do not know specifically where, when, and how severe the effects will be. First, how likely is a specific natural hazard? In the near term? In the long term? For long-lived infrastructure, this question is often stated as follows: How likely is it over the design life of the asset? The second question delves into the interaction between the natural hazard and transportation. If a natural hazard event were to occur, what are the likely effects that will impact transportation assets and functions? The description and analysis of hazard likelihood and effects is called hazard characterization. Both the likelihood and the other characteristics of hazards are necessary inputs to the resilience analysis methods and metrics discussed more fully in Chapters 3, 4, and 5. Hazard characterization also requires comprehensive knowledge of the potentially affected infrastructure assets, including their location, type, function, condition, and maintenance history. Therefore, asset management 11 U.S. Climate Resilience Toolkit. n.d. “Steps to Resilience.” https://toolkit.climate. gov/#steps.

AMERICA’S TRANSPORTATION INFRASTRUCTURE 23 programs, which develop and utilize this knowledge, are vital for integrat- ing resilience into transportation decision making. Asset management is a strategic and systematic process of operating, maintaining, and improving physical assets. Best practices in asset management rely on both engineer- ing and economic analyses, employing high-quality information to identify a structured sequence of maintenance, preservation, repair, rehabilitation, and replacement actions that will achieve and sustain a desired state of good repair over the life cycle of the assets at minimum practicable cost.12 NATURAL HAZARDS AND CLIMATE CHANGE Natural hazards that typically affect transportation in the United States are listed in Table 2-1. To understand the impact of climate change, it is helpful to divide the hazards into meteorological (acute), geological, and climate change–related (chronic) hazards. TABLE 2-1 Types of Natural Hazards13 Meteorological Hazards Geological Hazards Climate Change–Related Hazards Avalanche Debris flow Drought Fire/wildfire Flood/flash flood Hail Heavy rain High wind Ice flow Lightning Mudflow Snow Storm surge Tornado Tree fall Tropical cyclone Water table changes Earthquake Land subsidence Landslide and rockfall Sinkhole Tsunami Volcanic eruption Precipitation: changes in averages, extremes, and seasons Temperature: changes in averages, extremes, and seasons Sea level rise Interaction of precipitation, temperature, and sea level changes with other meteorological hazards 12 MAP-21 (Section 1103(a)(2)). 13 Zaghi, A.E., J.E. Padgett, M. Bruneau, M. Barbato, Y. Li, J. Mitrani-Reiser, and A. McBride. 2016. “Establishing Common Nomenclature, Characterizing the Problem, and Identifying Future Opportunities in Multihazard Design.” Journal of Structural Engineering 142(12). https://doi.org/10.1061/(ASCE)ST.1943-541X.0001586.

24 INVESTING IN TRANSPORTATION RESILIENCE Meteorological Hazards Meteorological hazards are commonly called bad weather: hurricanes and other storms with high winds, heavy rain or snow, and intense lightning; heat waves and severe cold snaps; and drought. High winds can occur during major storms such as hurricanes and nor’easters, as well as more localized windstorms such as tornadoes. Heavy winds can destroy trans- portation structures and facilities, cause injuries and fatalities to people, and topple trees leading to power outages and blocked roads and rail lines. In coastal areas, high winds generate storm surges that can cause flooding. Wind and storm surge can also accelerate coastal erosion, undermining infrastructure. Chlorides from salt water can also intensify corrosion of some infrastructure assets and thus impair their durability and performance in the long term. Rain can cause mudslides and flooding, while snow can block roads and other transportation infrastructure. The combination of high winds and heavy snow can cause a “white-out” condition that reduces visibility and can cause vehicle collisions and other damage. Lightning can damage structures, particularly electric power lines and signaling systems, and can also cause trees to fall and block roads and tracks. Lightning can also ignite wildfires, the severity of which can be worsened by drought. As will be discussed later in this chapter, meteorological hazards can occur simultaneously or in overlapping succession. Geological Hazards Geological hazards include earthquakes, tsunamis, volcanic eruptions, land- slides, and land subsidence. Earthquakes cause injuries, fatalities, and severe damage to transportation facilities not built to withstand them. Tsunamis following earthquakes create damage from both the force of wave action and flooding. Volcanic eruptions cause structural damage from lava flows, gas emissions, and hot cinders that can ignite fires. Landslides, including rockfalls, endanger personal safety and can close transportation routes. They can be triggered by heavy precipitation (e.g., in mudslides) and other meteorological events. Land subsidence typically causes more slow-acting damage. As land sinks, transportation infrastructure may become flooded. Climate Change–Related Hazards Climate change contributes to natural hazards by increasing average tem- peratures and altering historic patterns of extreme temperatures and pre- cipitation. These changes in atmospheric conditions can potentially affect any meteorological hazard. Specific hazards include sea level rise, periods of extreme heat or cold, and changes in freeze-thaw patterns, including

AMERICA’S TRANSPORTATION INFRASTRUCTURE 25 melting permafrost.14 These chronic changes in the natural environment, which are happening today and are expected to be exacerbated by cli- mate change, can alter the context under which transportation operates. Such hazards may affect transportation directly or they may interact with meteorological and geological events, affecting their frequency and severity. To the extent that transportation networks have been designed using norms derived from historical weather data, they are likely to be unprepared to withstand these climate change impacts. Sea level rise leads to repeated nuisance flooding, increases the height of high tides, and may also raise the water table beneath coastal land and possibly destabilize landforms. In many coastal communities, roads are located at a lower elevation than the surrounding lands to allow water to drain into the streets and away from homes and businesses. Rail lines follow waterways to reduce grades. As sea level rises, local drainage sys- tems become less effective, causing increased flooding on low-lying roads and costly delays to the transportation system. More than 7,500 miles of roadway on the Eastern seaboard are located in high tide flooding zones.15 Sea level rise may also be a hazard to airports, which are commonly built along tidal waters. Railroads in coastal regions often cut across marsh areas and run along the coastline as well. Sea level rise may also reduce the clearance under bridges, affecting or blocking navigation. A rise in the water table can flood tunnels, including their entrances and vents, and other infrastructure that is below grade. Tunnels, especially rail transit tunnels, may be vital links for communities and travelers with few other travel op- tions. All of the effects of sea level rise can be compounded by increased rainfall intensity triggered by climate change.16 As the case of Devils Lake in North Dakota demonstrates, increasing water levels do not only affect coastal communities and their transportation infrastructure. The increase in precipitation over the past 80 years has had a dramatic effect on the water level in the lake, because it has no natural 14 The base layer of roads can expand, contract, and shift during freeze-thaw temperature cycles, causing the surface to crack. Increases in the number of freeze-thaw cycles dur- ing the winter season because of climate change may more quickly degrade the quality of road surfaces. See EPA. 2017. Multi-Model Framework for Quantitative Sectoral Impacts Analysis: A Technical Report for the Fourth National Climate Assessment. EPA 430-R-17-001. Washington, DC: U.S. Environmental Protection Agency, pp. 79–81. 15 Reidmiller, D.R., C.W. Avery, D.R. Easterling, K.E. Kunkel, K.L.M. Lewis, T.K. Maycock, and B.C. Stewart, eds. 2018. Impacts, Risks, and Adaptation in the United States: Fourth National Climate Assessment, Volume II. Washington, DC: U.S. Global Change Research Program. https://doi.org/10.7930/NCA4.2018. 16 The content in this paragraph draws from Titus, J. 2002. “Does Sea Level Rise Matter to Transportation Along the Atlantic Coast?” https://research.fit.edu/media/site-specific/ researchfitedu/coast-climate-adaptation-library/united-states/east-coast/regional---us-east-coast/ Titus.-2002.-US-Transportation--SLR-on-the-Atlantic-Coast.pdf.

26 INVESTING IN TRANSPORTATION RESILIENCE outlet. Since 1964, the water level of the lake has risen by 13 meters, the area of the lake has expanded by 10 times, and the volume of water in the lake has expanded by 32 times. As a result, local farms have been flooded, the local towns have been protected by levees, and highways and key rail lines for freight and passenger train service have been washed out.17 In the United States, high temperature records over the past two de- cades far exceed the number of low temperature records.18 Recent data from NOAA indicate that a warming pattern occurred in all of the contigu- ous United States with the exception of portions of the Upper Midwest and Northern Plains (see Figure 2-1).19 Nonetheless, changes in temperature patterns—extreme hot as well as extreme cold—can affect infrastructure assets and the experience of employees and customers. The 2021 polar vortex that affected the south-central United States had a severe effect on energy infrastructure, caused an estimated $200 billion in economic losses, and, as SwissRe reports, “is on track to rival and perhaps even surpass the likes of intense climate disasters more well acquainted to the state such as Hurricane Harvey (2017) and Ike (2008).”20 Extreme temperatures can cause health emergencies for employees and customers exposed to the ele- ments. Extreme heat can melt asphalt on roads and airport tarmacs. For rail infrastructure, extreme heat can lead to track buckling and extreme cold can cause brittle fracture of track. Changes in freeze-thaw patterns can af- fect the life span and maintenance needs of roads and runways. Changes in temperature patterns are a particular concern for all types of transportation infrastructure and facilities in Alaska. When permafrost thaws, land in the melted area subsides.21 17 Larson, D. 2012. “Runaway Devils Lake.” American Scientist 100(1):46. https://www. americanscientist.org/article/runaway-devils-lake. 18 Reidmiller, D.R., C.W. Avery, D.R. Easterling, K.E. Kunkel, K.L.M. Lewis, T.K. Maycock, and B.C. Stewart, eds. 2018. Impacts, Risks, and Adaptation in the United States: Fourth National Climate Assessment, Volume II. Washington, DC: U.S. Global Change Research Program. https://doi.org/10.7930/NCA4.2018. 19 NOAA. 2021. “NOAA Delivers New U.S. Climate Normals.” https://www.ncei.noaa.gov/ news/noaa-delivers-new-us-climate-normals. 20 Pui, A., and S. Horie. 2021. “Polar Vortex: A Counter Intuitive Threat of Climate Change?” SwissRe Corporate Solutions, April 13. https://corporatesolutions.swissre.com/ insights/knowledge/polar-vortex-a-counter-intuitive-threat-of-climate-change.html. 21 EPA. 2017. Multi-Model Framework for Quantitative Sectoral Impacts Analysis: A Tech- nical Report for the Fourth National Climate Assessment. EPA 430-R-17-001. Washington, DC: U.S. Environmental Protection Agency, pp. 100–107.

AMERICA’S TRANSPORTATION INFRASTRUCTURE 27 FIGURE 2-1 Average annual temperature change for the contiguous United States from the 1981–2010 climate normals to the newest data in the 1991–2020 normals.22 Hazards Vary by Region and Location Meteorological and geological hazards and the effects of climate change vary by region and location. Resilience analysis, planning, and management processes need to account for this variation. In the United States, it is generally well understood that the country’s diverse regions experience different mixes of natural hazards. Transpor- tation agencies adopt practices adapted to these regional circumstances. Hurricanes are tropical cyclone storms that form in the North Atlantic and North Pacific, affecting shipping and bordering coastal regions, and they commonly traverse far inland to cause damage far from coasts. Storm surge from high winds is confined to areas bordering large bodies of water. Wildfires are typically the most dangerous on the West Coast and in the Rocky Mountain region but also occur in the south-central and southeast- ern states. Tornadoes occur frequently in the central plains, Florida, and the Gulf Coast states. Severe thunderstorms capable of producing tornadoes 22 NOAA. 2021. “NOAA Delivers New U.S. Climate Normals.” https://www.ncei.noaa.gov/ news/noaa-delivers-new-us-climate-normals.

28 INVESTING IN TRANSPORTATION RESILIENCE FIGURE 2-2 Expected frequency of earthquake occurrences in the United States.23 and hail appear in every state. Mountainous regions create the conditions for landslides and rockfalls. For earthquakes, the highest hazard areas are in Alaska, Hawaii, Puerto Rico, the West Coast, and a small region in the central United States (see Figure 2-2).24 Similarly, the significance of flooding will vary by both the region and the specific locations of infrastructure assets. The significance of flooding will also vary by the type of flood, such as flash floods with little warn- ing, storm surges from cyclones and tsunamis, hurricane driven rain, or snow melt. Figures 2-3 and 2-4 present the historical flood risk in New York and California, respectively. The maps were generated using NOAA’s 23 USGS. n.d. “Introduction to the National Seismic Hazard Maps.” https://www.usgs.gov/ natural-hazards/earthquake-hazards/science/introduction-national-seismic-hazard-maps?qt- science_center_objects=0#qt-science_center_objects. 24 NOAA. n.d. “National Centers for Environmental Information.” https://www.ncei.noaa. gov; USGS (U.S. Geological Survey). n.d. “Earthquake Hazards.” https://www.usgs.gov/natural- hazards/earthquake-hazards.

AMERICA’S TRANSPORTATION INFRASTRUCTURE 29 FIGURE 2-3 Number of flood events reported for a county or zone in New York 1996–2019.25 FIGURE 2-4 Number of flood events reported for a county or zone in California 1996–2019.26 25 FEMA. 2021. “Historical Flood Risk and Costs.” https://www.fema.gov/data-visualization/ historical-flood-risk-and-costs. 26 FEMA. 2021. “Historical Flood Risk and Costs.” https://www.fema.gov/data-visualization/ historical-flood-risk-and-costs.

30 INVESTING IN TRANSPORTATION RESILIENCE interactive data tool, which presents historical flood risk using data from 1996 to 2019.27 Increases in the Frequency of Extreme Weather Events Climate change can lead to shifts in extreme weather, and trends indicate that large areas of the United States are being subject to such extremes. Because conventional design, material, and operational standards in trans- portation are built around historic weather data, increases in the likelihood of a hazard can turn a distant threat into an imminent disaster. NOAA’s climate extreme index tracks changes in extremes for the contiguous United States and its regions. The index consolidates extremes in temperature and precipitation (i.e., days when temperature or precipitation are in the top or bottom 10% of the historical average) and is reported as a percentage of the total number of days for the region. Going back more than a century, on average, 21% of the United States experiences extremes in any given year. Over the past 20 years, however, this average has risen to 28%. Regionally, increases in the average area affected by extremes over the past 20 years range from a low of 1.5 percentage points in the Northwest to a high of 13.7 percentage points in the Northeast. In terms of years above the long- run average, the Ohio Valley ranks first, with 15 out of the past 20 years above the long-run average. The trend in extremes may be accelerating. Over the past 5 years, only four of the nine regions experienced any single year below the long-run average for extremes.28 EXPOSURE AND EVENT LIKELIHOOD Evaluating resilience to natural hazards starts with exposure. Because the likelihood of natural hazards varies by region and location, the first pass at a comprehensive analysis of exposure can be a simple question of whether a particular hazard ever occurs in a particular location. The next level of analyzing exposure is to categorize the hazards on a scale from low to high likelihood of occurrence. However, quantitative analyses of resilience typically require describing the likelihood of a natural hazard as a speci- fied event with a defined probability. For example, the Federal Emergency 27 FEMA (Federal Emergency Management Agency). 2021. “Historical Flood Risk and Costs.” https://www.fema.gov/data-visualization/historical-flood-risk-and-costs. 28 The four regions with at least 1 year below the long-run average from 2016 to 2020 are Rocky Mountains and Northern Plains (2 years), Southwest (1 year), West (1 year), and Northwest (3 years). Study committee analysis of Climate Extremes Index data; contiguous United States and regional Climate Extremes Index averages from 1910–2020 and 2001–2020 were compared. See NOAA. n.d. “U.S. Climate Extremes Index (CEI).” https://www.ncdc. noaa.gov/extremes/cei.

AMERICA’S TRANSPORTATION INFRASTRUCTURE 31 Management Agency (FEMA) defines a “base flood” as “a flood having a one percent chance of being equaled or exceeded in any given year.” FEMA’s base flood also has a metric for flood elevation.29 Measuring Likelihood Measuring likelihood is an integral step to producing the scenarios required for most approaches to resilience analysis. Metrics that capture hazard likelihood usually require data on past frequency and projections of future frequency. The first step is to turn a natural hazard into something that can be counted, usually defined as an “event.” Measures of occurrence, and thus frequency of events, differ for different hazards. For example, the likelihood of floods is usually measured in annual probabilities, but the frequency of earthquakes is reported in events over 10,000 years.30 In addition, measuring frequency typically requires threshold values of severity to indicate when the magnitude of an event is sufficiently great to make it count as a hazard event. For example, for Atlantic hurricanes there are thresholds for named storms and for five categories indicating in- creasing severity. Earthquakes and tornadoes also have measurement scales that categorize events by severity. As knowledge about natural hazards improves, the categories and scales used to define thresholds for events are periodically revised.31 Comprehensive approaches to resilience analysis and planning require a way to put the likelihood of all hazards on the same frequency scale. The National Institute of Standards and Technology (NIST), in its guide for community resilience, advises using three categories: routine, design, and extreme. NIST uses a 50-year analysis period. The routine level is for hazards that have a 50% or greater probability of occurring over the next 50 years. The design level specifies the event with a 10% chance of happen- ing over 50 years, and the extreme level events have a probability of 2–3% over 50 years. For earthquakes, NIST’s extreme level is typically called the “maximum considered event.” (For comparison, FEMA’s base flood of 1% annual probability would have a roughly 40% chance of occurring over 29 FEMA. n.d. “National Flood Insurance Program Terminology Index.” https://www.fema. gov/flood-insurance/terminology-index. 30 USGS. n.d. “Introduction to the National Seismic Hazard Maps.” https://www.usgs.gov/ natural-hazards/earthquake-hazards/science/introduction-national-seismic-hazard-maps?qt- science_center_objects=0#qt-science_center_objects. 31 The Weather Channel. 2020. “The Enhanced Fujita Scale: How Tornadoes Are Measured.” https://weather.com/storms/tornado/news/enhanced-fujita-scale-20130206; USGS. n.d. “Moment Magnitude, Richter Scale.” https://www.usgs.gov/faqs/moment-magnitude-richter-scale-what-are- different-magnitude-scales-and-why-are-there-so-many?qt-news_science_products=0#qt-news_ science_products.

32 INVESTING IN TRANSPORTATION RESILIENCE 50 years.) Furthermore, each level is tied to a performance goal. Routine hazards should lead to minimal disruptions. The design hazard should be built into building and construction standards. Planning for the extreme hazard event should protect life but may require rescue and a significant recovery period.32 Defining the relevant event for measuring frequency may also depend on the technology operated by a transportation agency. For example, the Washington Metropolitan Area Transit Authority (WMATA) operates both heavy rail and bus service. For metrorail, even the rail lines above ground can support close to normal service in up to 6 inches of snow. Only in- creases in the frequency of snow events above 6 inches would worsen the resilience of WMATA’s rail service. WMATA’s bus service, however, is dependent on roads maintained by others. Bus routes may begin to be detoured or cut back with as little as 2 inches of snow.33 Likelihood with Climate Change Measuring likelihood should also incorporate the effects of climate change. However, measuring changes in likelihood is also not a straightforward exercise. For example, the Atlantic hurricane season in 2020, breaking the record set in 2005, produced 30 named storms, and 2020’s 13 hurricanes and 6 major hurricanes exceeded the average. 2020 was also the fifth consecutive year with an above average number of named storms.34 Still, the era of good data on tropical cyclone storms begins only in the 1980s. Climate change could be affecting the frequency of all named storms or the intensity of major hurricanes or both. In addition, climate change may be affecting where major storms intensify, changing the frequency for some locations but not others. Similarly, for severe thunderstorms producing tornadoes, trends since the 1970s indicate a reduction in the number of days with at least one tornado but increases in the number of days with outbreaks of a large number of tornadoes. Climate change models predict continued increases in the number of severe thunderstorms in the Midwest and Great Plains states, especially in March, April, and May.35 32 NIST. 2016. Community Resilience Planning Guide for Buildings and Infrastructure Sys- tems, Volume 1. NIST Special Publication 1190. http://dx.doi.org/10.6028/NIST.SP.1190v1. 33 WMATA. n.d. “Rail Snow Service.” https://www.wmata.com/rider-guide/weather/rail. cfm; WMATA. n.d. “Bus Snow Service.” https://www.wmata.com/rider-guide/weather/bus/ index.cfm. 34 NOAA. 2020. “Record-Breaking Atlantic Hurricane Season Draws to an End.” https:// www.noaa.gov/media-release/record-breaking-atlantic-hurricane-season-draws-to-end. 35 USGCRP (U.S. Global Change Research Program). 2017. “Chapter 9: Extreme Storms” in Climate Science Special Report: Fourth National Climate Assessment. https://science2017. globalchange.gov/chapter/9.

AMERICA’S TRANSPORTATION INFRASTRUCTURE 33 Examples of Measuring Exposure and Likelihood The U.S. Department of Transportation’s (U.S. DOT’s) Vulnerability Assess- ment Scoring Tool (VAST), which includes tools to analyze exposure to natural hazards, acknowledges that location-specific modeling incorporat- ing climate change is the best way to produce projections for the likelihood of a natural hazard event. Tools are available that make the output of the climate change models useful at a local scale for transportation planning. For example, U.S. DOT’s Climate Model Intercomparison Project (CMIP) Climate Data Processing Tool uses statistical methods to produce projec- tions of changes in temperature and precipitation, including extreme heat and rainfall. The tool produces projections for changes in environmental conditions that then need to be integrated into models projecting the likeli- hood of meteorological and geological hazards.36 If models such as CMIP are not available, VAST offers indicators that transportation agencies can use to score an asset’s exposure to a natural hazard. For storm surge, for example, the tool’s indicator library provides a scale for scoring exposure based on miles from the coastline and elevation. The scores, from 1 to 4, do not represent probabilities but rather indicators that allow for comparing the relative exposure of different assets.37 The National Cooperative Highway Research Program report Main- streaming System Resilience Concepts into Transportation Agencies: A Guide also provides step-by-step guidance on how to conduct an assess- ment of a transportation agency’s exposure to natural hazards.38 Minnesota DOT, recognizing that its current infrastructure and prac- tices already take into account past patterns of hazards, frames its evalu- ation of hazard likelihood in terms of the change expected over the next 20 years. Heavy precipitation leading to flooding and warmer winters received “very high” ratings for likelihood of worsening over the next 20 years. Vegeta tion patterns received a “high” rating for likelihood of change, leading to concerns about vegetation loss and invasive species caus- ing soil erosion and wetland failure. On the other hand, wildfires and severe wind received “low” likelihood of change ratings.39 36 FHWA (Federal Highway Administration). n.d. “Climate Change Adaptation Tools: CMIP Climate Data Processing Tool.” https://www.fhwa.dot.gov/environment/sustainability/ resilience/tools. 37 FHWA. n.d. “Climate Change Adaptation Tools: Vulnerability Assessment Scoring Tool.” https://www.fhwa.dot.gov/environment/sustainability/resilience/tools. A further discussion of VAST will be found in Chapter 3 of this report. 38 NASEM. 2021. Mainstreaming System Resilience Concepts into Transportation Agencies: A Guide. Washington, DC: The National Academies Press. https://doi.org/10.17226/26125. 39 Meek, J. 2020. “MnDOT Transportation Resilience.” Presentation to the Committee on Transportation Resilience Metrics, June 26.

34 INVESTING IN TRANSPORTATION RESILIENCE HAZARD CHARACTERIZATION The significance of a natural hazard depends not only on how likely it is to occur but also on how serious and widespread its effects are likely to be. Resilience analysis, therefore, must also incorporate knowledge about how specific natural hazards interact with specific transportation assets (includ- ing nodes, networks, and systems). Damage to infrastructure and facilities may not be the only important effect of a natural hazard. Essential person- nel unable to report to work may also disrupt service. Failures in power supplies, water services, or communication technologies can affect entire systems. For intermodal nodes, damage to one mode can force closures of services on other modes, such as maritime and surface freight operations at a port or connecting transportation modes at a station, port, or airport. To be able to assess the potential for damage or disruption, resilience assessment must first develop qualitative and quantitative descriptions of the loading (or stress) that the natural hazard puts on infrastructure assets. Physical forces such as the speed of the wind, the height of the flood, the type of debris flow, the amount of rain, or the number of days of extreme heat or cold are analyzed. Some of the categories and scales used to define hazard events, such as for hurricanes and tornadoes, already integrate knowledge about likely damage. Hazard characterization describes the geographic distribution of load- ing intensity (or stress) generated by one or more natural hazard events. In addition to likelihood, hazard characterization includes the effects that di- rectly cause damage and disruption and also must account for differences in the duration and scale of natural hazard events. Methods for hazard char- acterization vary from general descriptions of common hazards to detailed quantitative models of the specific effects on assets. Even general descrip- tions can still be useful for formulating mitigation strategies and plans. Case studies and historical patterns can also help characterize specific hazards. Affected Area or Region Hazard characterization also requires an analysis of the geographic area affected by the hazard event, which should be as spatially explicit as pos- sible. Spatial analysis includes identifying specific locations for damage and the larger areas or regions affected by the damage and disruption. Again, different hazards have different conventions for measuring the area affected. For earthquakes, the load effect is often described in terms of the joint occurrence of shaking intensity over the region of interest.40 40 Jayaram, N., and J.W. Baker. 2010. “Efficient Sampling and Data Reduction Techniques for Probabilistic Seismic Lifeline Risk Assessment.” Earthquake Engineering & Structural Dynamics 39(10):1109–1131.

AMERICA’S TRANSPORTATION INFRASTRUCTURE 35 For flooding or storm surge, spatial analysis usually focuses on the area inundated and the depth of the water.41 Spatial analysis of hazard effects also must include the geographic extent of the transportation system under study and capture disruption to nodes and network links, as well as infra- structure assets. Duration of the Hazard Event Hazard characterization includes the entire arc of time from first warn- ing to when the event is no longer actively producing effects. Analysis of duration focuses on the evolution of the loading intensity or effects of the hazard over time. For earthquakes, although the hazard may be active for just seconds to a minute, the duration and intensity of shaking (as well as potential aftershocks) are still critical to understanding the extent of damage. Post-event recovery from a damaging earthquake also requires a considerable amount of time; for example, the transportation system of San Francisco was impacted for several years after damage from the Loma Prieta earthquake.42 For floods, wildfires, hurricanes, heat waves, and cold waves, the hazard event may be active for days to weeks. For flooding, the time required for water levels to subside, for example, significantly affects post-event recovery and thus needs to be part of characterizing duration. For chronic natural hazards associated with climate change, such as sea level rise or changing temperature and precipitation patterns, the duration is likely to be indefinite. Forecasting Hazard characterization includes the ability to forecast an event in a way that provides information on specific time and place and thus allows for taking temporary actions to reduce damage and save lives. Hurricane fore- casting, for example, has advanced to the point that warnings go out 3–4 days in advance, advising that specific locations are likely to experience certain levels of intensity. Disaster preparations start ahead of hurricane landfall: windows are boarded up, sandbags positioned, and popula- tions evacuated, all of which reduce the damage resulting from the storm. Improved forecasting of major winter storms allows road maintenance crews to pre-treat to reduce the disruption from snow and ice. Earthquake 41 Apel, H., G.T. Aronica, H. Kreibich, and A.H. Thieken. 2009. “Flood Risk Analyses— How Detailed Do We Need to Be?” Natural Hazards 49:79–98. https://doi.org/10.1007/ s11069-008-9277-8. 42 SPUR (San Francisco Planning and Urban Research Association). 2010. “Transportation and Rebuilding.” The Urbanist 494. https://www.spur.org/publications/urbanist- article/2010-07-06/ transportation-and-rebuilding.

36 INVESTING IN TRANSPORTATION RESILIENCE forecasting, by contrast, is much more limited. Earthquake “shaking” alert systems can only provide seconds of warning. The lack of advance warning for a specific place and time is a norm that feeds into resilience analysis and planning for earthquakes.43 Seasonality Seasonality occurs when the frequency of natural hazard events varies throughout the year in a regular and predictable pattern. Atlantic hurri- canes (June–November), Arizona monsoons (June–September), and severe winter weather all exhibit seasonality. Seasonality can be important for re- silience planning and should be included in hazard characterization. How- ever, especially with climate change, seasonality may produce a false sense of security. Climate change can produce what is known a bit irreverently as “weather weirding.” A summer-like day in February may be fun, but a heavy rain—when normally the precipitation falling on frozen ground is snow—may lead to disaster. For severe thunderstorms producing tornadoes, trends indicate that their occurrence is becoming more volatile, and the “high season” is shifting to earlier in the year. The Atlantic hurricane season is not absolute either. Named storms regularly occur outside of the season. Since 2015, there has been at least one out-of-season named storm every year, occurring in the months of January (one), April (one), and May (six).44 Understanding the Psychology of Uncertainty To the extent that uncertainty varies among hazards, and depending on the degree of risk aversion, this variation may affect how people perceive the significance of a particular type of hazard. If two types of disasters have the same mean risk, but one has a larger variance in the extent of damage, people are likely to assess the disaster with the larger variance to be more significant. MULTIPLE AND CASCADING EVENTS Most regions in the United States are prone to multiple natural hazards. Comprehensive approaches to resilience analysis and planning must not 43 USGS. 2021. “ShakeAlert Earthquake Early Warning Delivery for the Pacific Northwest.” https://www.usgs.gov/news/shakealert-pacific-northwest-rollout?. 44 USGCRP. 2017. “Chapter 9: Extreme Storms” in Climate Science Special Report: Fourth National Climate Assessment. https://science2017.globalchange.gov/chapter/9; Wikipedia. n.d. “List of Off Season Atlantic Hurricanes.” https://en.wikipedia.org/wiki/List_of_off-season_ Atlantic_hurricanes.

AMERICA’S TRANSPORTATION INFRASTRUCTURE 37 only characterize all relevant hazard events but also analyze the potential for events to occur simultaneously or in quick succession. Multiple-hazard analysis evaluates the effects of two or more separate hazard events, as opposed to looking at the multiple effects of a single event. (A tsunami that generates loading on a bridge from both moving water and debris is an example of multiple effects from a single event.) Mul- tiple hazards may be concurrent or in overlapping sequence, such that the asset has not recovered from the first event before the second event occurs. The multiple hazards may be the same type of event, such as the main shock and the aftershocks of earthquakes or two successive hurricanes.45 Multiple-hazard analysis also includes the same effects from different types of events. For example, storm surge from high winds combined with heavy rain farther up the river valley can increase the size of the area inundated with flood water. Potentially more dangerous are multiple hazard events where the haz- ards interact. One hazard may compound the effect of another. A heat wave is likely to be more intense during a drought. Sea level rise may mean that port facilities designed for short-term flooding may no longer be adequate for both periodic flooding and the loading associated with long-term rise. Cascading events occur when one hazard event triggers another, like a series of toppling dominoes. Wildfire destabilizes vegetation, so even mod- erate rainfall after wildfire can lead to landslides, heightened floods, and debris flow. Similarly, major hurricanes that damage vegetation can also lead to landslides and increased flooding after subsequent storms. Climate change is increasing the likelihood of cascading events.46 MEASURING AND MODELING HAZARD SCENARIOS A commonly used technique for hazard characterization is to designate and describe a hazard scenario or a range of hazard scenarios. A magni- tude 7 earthquake with a specified epicenter location is an example of a hazard scenario. Several different methods are available to integrate the hazard scenarios into resilience analysis. 45 ASCE (American Society of Civil Engineers). 2019. Resilience-Based Performance: Next Generation Guidelines for Buildings and Lifeline Standards. https://ascelibrary.org/doi/ book/10.1061/9780784415276. 46 Vahedifard, F., and A. AghaKouchak. 2018. “The Risk of ‘Cascading’ Natural D isasters Is on the Rise.” The Conversation, October 22. https://theconversation.com/the-risk-of- cascading-natural-disasters-is-on-the-rise-104192.

38 INVESTING IN TRANSPORTATION RESILIENCE Deterministic and Probabilistic Methods The deterministic approach chooses a set of hazard scenarios that could affect the transportation asset or system. Using methods specific to the type of hazard, the approach generates the loadings from each scenario and chooses the one that presents the worst case. The loadings in the worst-case event scenario then become the controlling event for the next steps of the resilience analysis. The probabilistic approach was developed because, in practice, the worst-case loading can be difficult to identify. The probabilistic approach uses all possible events, assigning the loadings associated with each event a weight based on its frequency of occurrence.47 The result is a range of loading values with a probability assigned to each value. Techniques also allow the inclusion of uncertainties related to randomness and lack of informa tion. The insurance industry uses the probabilistic approach. Scenario-Based, Event-Based, or Time-Based Approaches Scenario-based, event-based, and time-based approaches are methods to integrate assessments of damage, disruption, and recovery into resilience analysis. Scenario-based hazard characterization uses one event or a small set of historical or hypothetical events to model the spatial distribution of loading intensity, such as where and how the natural hazard will interact with the natural landscape and built environment of the region of interest. Scenario-based approaches are often used to develop disaster mitigation and recovery plans.48 The event-based strategy, which is used in building codes and stan- dards, including for infrastructure,49 uses maps to designate areas that experience the same likelihood of a hazard event (e.g., events that have a 10% or greater probability over 50 years). FEMA’s flood hazard maps are event-based. This strategy can identify areas that are exposed to the specified hazard event, but it is less useful for analyzing the effects of a specific hazard event. An actual hazard event will not have a spatial dis- tribution of loadings that affects all areas equally. Techniques to address 47 Bommer, J.J. 2002. “Deterministic vs. Probabilistic Seismic Hazard Assessment: An Exag- gerated and Obstructive Dichotomy.” Journal of Earthquake Engineering 6(Spec 01):43–73. 48 Jones, L.M., R.L. Bernknopf, D.A. Cox, J. Goltz, K.W. Hudnut, D.S. Mileti, S. Perry, et al. 2008. “The ShakeOut Scenario: Effects of a Potential M7.8 Earthquake on the San Andreas Fault in Southern California.” U.S. Geological Survey. https://pubs.usgs.gov/of/2008/1150. 49 See, for example, ASCE. 2017. “Minimum Design Loads and Associated Criteria for Buildings and Other Structures.” ASCE/SEI 7-16.

AMERICA’S TRANSPORTATION INFRASTRUCTURE 39 this shortcoming involve generating a set of realistic loading scenarios that correspond to approximately the same hazard level.50,51 The time-based approach requires generating the spatial distribution of loading intensity for all events that could impact the region of interest over a specific time horizon. The time-based approach, which is also used by the insurance industry, is the most complex, and its application in transporta- tion is currently limited to research. However, because the approach can produce life-cycle impact assessments over a pre-defined time horizon (e.g., over 50 years), the approach could be used to evaluate the probability of exceeding a specified level of functional loss over the design life of an infra- structure asset. The annualized impact is then computed by weighting the damage and disruption from each event based on its rate of occurrence.52 DATA FOR MODELS AND PROJECTIONS Complete and accurate data and up-to-date models of natural hazards that integrate climate change are critical to characterizing hazards for resil- ience analysis. Transportation agencies depend on federal sources of data for meteorological, geological, and climate change–related hazards. They supplement federal data with specialized information tailored to their own unique circumstances. For example, the Colorado DOT collected data on the location and extent of burn scars and combined it with FEMA flood hazard maps to create models of likelihood and character of debris flow.53 The committee reviewed government sources of the data required for hazard modeling and resilience analysis; while only two important data sources are presented here for explanation, other select resources are listed in Appendix C. Drawing on the interviews and their own experiences, com- mittee members raised specific concerns about the need to update federal information on precipitation and flood hazards. 50 Jayaram, N., and J. Baker. 2010. “Considering Spatial Correlation in Mixed-Effects Re- gression and the Impact on Ground-Motion Models.” Bulletin of The Seismological Society of America 100:3295–3303. https://doi.org/10.1785/0120090366. 51 Bocchini, P., V. Christou, and M.J. Miranda. 2016. “Correlated Maps for Regional Multi-Hazard Analysis: Ideas for a Novel Approach” in Multi-Hazard Approaches to Civil Infrastructure Engineering (P. Gardoni and J. LaFave, eds.). Cham, Switzerland: Springer. https://doi.org/10.1007/978-3-319-29713-2_2. 52 Tomar, A., and H.V. Burton. 2021. “Risk-Based Assessment of the Post-Earthquake Functional Disruption and Restoration of Distributed Infrastructure Systems.” International Journal of Disaster Risk Reduction 52:102002. 53 Kemp, L. 2020. “Transportation Resilience Metrics.” Presentation to the Committee on Transportation Resilience Metrics, September 14.

40 INVESTING IN TRANSPORTATION RESILIENCE FEMA Flood Maps FEMA produces maps of flood hazards to serve the National Flood In- surance Program (NFIP) and the associated regulatory requirements for county-level flood zone management. Historically, FEMA has targeted its investments in mapping flood hazards at populated areas or nearby areas likely to be developed. As a result, only one-third of the nation’s miles of rivers and streams are mapped. An estimated 2.3 million miles of rivers and streams and 50,000 miles of coastal land remain unmapped (see Fig- ure 2-5). As of 2019, more than 6,500 counties and communities have no FEMA flood maps and for 3,300 communities, the FEMA flood maps are more than 15 years old. Most of the unmapped areas are rural, meaning that transportation networks that cross these areas may suffer from a lack of information about flood hazards. The Biggert-Waters Flood Insurance Reform Act of 2012 set modern conditions for flood hazard mapping, including incorporating climate change. However, FEMA has struggled to keep pace with mapping needs. The Association of State Floodplain Manag- ers has estimated that an additional infusion of $3.2–$11.8 billion is needed to complete the flood hazard mapping program.54 In addition, the flood hazard mapping methods that FEMA pioneered in the early decades of its flood programs are now more widely avail- able. The private sector is capable of creating its own flood hazard maps, adapted to specific needs. However, there can be substantial differences in the outcomes of different hazard modeling processes. Figure 2-6 illustrates the differences between FEMA’s official maps of flood hazard and flood hazard analysis produced by the models of the First Street Foundation, a nonprofit dedicated to “accurate, property-level, publicly available flood risk information.”55 One significant difference between the processes used by the private sector and those used by FEMA is that the private sector typically uses proprietary models. FEMA maps are limited in that they are probabilistic, using historical data only, and therefore do not incorporate the effects of climate change. FEMA is also required to map hazards using a public process, and its hazard determinations can be appealed.56 54 Association of State Floodplain Managers. 2020. Flood Mapping for the Nation: A Cost Analysis for Completing and Maintaining the Nation’s NFIP Flood Map Inventory. https://asfpm- library.s3-us-west-2.amazonaws.com/FSC/MapNation/ASFPM_MaptheNation_Report_2020.pdf. 55 First Street Foundations. n.d. “First Street Foundations Mission.” https://firststreet.org/ mission. 56 Eby, M., and C. Ensor. 2019. “Understanding FEMA Flood Maps and Limitations.” https:// firststreet.org/flood-lab/published-research/understanding-fema-flood-maps-and-limitations.

AMERICA’S TRANSPORTATION INFRASTRUCTURE 41 FIGURE 2-5 Unmapped stream miles by county, as of fiscal year 2019.57 FIGURE 2-6 Difference in number of properties at substantial flood risk compared to FEMA’s data.58 57Association of State Floodplain Managers. 2020. Flood Mapping for the Nation: A Cost Analysis for Completing and Maintaining the Nation’s NFIP Flood Map Inventory. https://asfpm- library.s3-us-west-2.amazonaws.com/FSC/MapNation/ASFPM_MaptheNation_Report_2020.pdf. 58 Eby, M., and C. Ensor. 2019. “Understanding FEMA Flood Maps and Limitations.” https:// firststreet.org/flood-lab/published-research/understanding-fema-flood-maps-and-limitations.

42 INVESTING IN TRANSPORTATION RESILIENCE Atlas 14 Atlas 14 is the most recent edition of a database produced by NOAA’s Hydrometeorological Design Studies Center that provides detailed precipita- tion frequency data for most regions of the United States. Although Atlas 14 is used at the federal, state, and local levels for planning activities, engineering design, modeling of flood risks, and managing floodplain devel opment for NFIP, its data are out of date. Importantly, the methodology used for Atlas 14 does not incorporate climate change projections. The first regional volume of Atlas 14, for the semiarid southwest, was released in 2004. Although volumes for the northeastern states and Texas were first released in 2015 and 2018, respectively, the volumes for most of the regions are more than 8–10 years old and in need of updating. The northwestern states are not covered by Atlas 14, and no funding is available to complete their volume.59 The weaknesses of Atlas 14 have consequences for transportation agen- cies. Recent studies conducted for the Virginia Transportation Research Council indicate that Virginia’s rainfall index has increased, represent- ing a significant and ongoing change from the precipitation frequencies documented by Atlas 14 for the Ohio Valley and surrounding states, last completed in 2004 and revised in 2006. More recent trends and simulated future conditions show the inadequacy of Atlas 14. The City of Virginia Beach, following analysis that extreme rainfall events will be occurring more frequently in the coming decades, recently revised its design standards manual to increase all of the volumes of design storms by 20%.60 Because precipitation frequency data are critical to Virginia’s ability to adapt and protect coastal and riverine regions of the state from the impacts of climate change, the Commonwealth is currently collaborating with Dela- ware, Maryland, North Carolina, and NOAA’s National Weather Center on 59 HDSC (Hydrometeorological Design Studies Center). n.d. “Precipitation Frequency Data Server.” https://hdsc.nws.noaa.gov/hdsc/pfds/index.html; HDSC. 2019. “Progress Re- port for Period October 2018 to March 2019.” https://www.weather.gov/media/owp/oh/hdsc/ docs/201904_HDSC_PR.pdf. 60 Morsy, M.M., Y. Shen, J.M. Sadler, A.B. Chen, F.T. Zahura, and J.L. Goodall. 2019. “Incorporating Potential Climate Change Impacts in Bridge and Culvert Design.” FHWA/ VTRC 20-R13. http://www.virginiadot.org/vtrc/main/online_reports/pdf/20-r13.pdf; City of Virginia Beach, Virginia. 2017. “Joint Occurrence and Probabilities of Tides and Rainfall.” CIP 7-030, PWCN-15-0014, Work Orders 2 and 5A, Final Report. https://www.vbgov.com/ government/ departments/public-works/comp-sea-level-rise/Documents/joint-occ-prob-of-tides- rainfall-4-24-18.pdf; Smirnov, D., J. Giovannettone, S. Lawler, M. Sreetharan, J. Plummer, and B. Workman. 2018. “Analysis of Historical and Future Heavy Precipitation: City of Virginia Beach, Virginia.” CIP 7-030, PWCN-15-0014, Work Order 9A. https://www.hrpdcva.gov/uploads/ docs/5A_ Attachment_AnalysisofHistoricalandFutureHeavyPrecipitation_ Finalrev_20180326. pdf; City of Virginia Beach, Department of Public Works. 2020. Design Standards Manual. https://www.vbgov.com/government/departments/public-works/standards-specs/pages/default. aspx.

AMERICA’S TRANSPORTATION INFRASTRUCTURE 43 a four-state effort to update Atlas 14’s precipitation estimates. Updated data are essential to support accurate estimates for what communities can expect from storm events; for Commonwealth agencies to have accurate forecast- ing projections to prepare for future rain, storm, and other climatic events; and to ensure accurate regulatory processes. Although paid for through a Federal Highway Administration pooled funding process, the updated data will be available for download from NOAA’s Precipitation Frequency Data Server. A similar project is under way to update precipitation frequency data for the State of Louisiana.61 CHAPTER SUMMARY Ensuring the resilience of transportation systems requires preventing natural hazards from creating the damage and disruption that leads to disastrous outcomes. Resilience analysis starts with methods to determine hazard like- lihood and characteristics and must address the uncertainty about where, when, and how a hazard event is likely to occur. Comprehensive approaches to resilience analysis cover multiple hazards, including their interactions. When transportation agencies conduct resilience analysis, they need to ac- count for regional and location-specific variations in exposure to different types of natural hazards. The changes associated with climate change make resilience invest- ments more pressing while also increasing the importance of integrating uncertainty into resilience analysis. In the face of this uncertainty, scenario- based approaches can be effective strategies for analyzing changing natural hazards because they consider a range of possible threats, rather than relying on point estimates. However, scenario-based analysis approaches are still affected by data and modeling quality. Transportation agencies de- pend on the federal government and others for up-to-date data on hazards and their effects and for modeling climate change. Currently, outdated and deficient precipitation data are a major risk for accurate damage estimates, and, similarly, incomplete and out-of-date FEMA flood maps hinder the ability to assess and prepare for major impacts from floods. The extent to which these data and modeling capabilities are lacking will affect the abil- ity of transportation agencies to engage resilience analysis with confidence. 61 Commonwealth of Virginia, Office of the Governor. 2020. “Virginia Coastal Resilience Master Planning Framework, Principles and Strategies for Coastal Flood Protection and Adapta- tion.” https://www.naturalresources.virginia.gov/initiatives/resilience--coastal- adaptation; Trans- portation Pooled Fund Program. 2020. “Update Precipitation Frequency Estimates for Delaware, Maryland, North Carolina and Virginia (NOAA Atlas 14 Volume 13).” Solicita tion 1534. https://www.pooledfund.org/Details/Solicitation/1534; Transportation Pooled Fund Program. 2020. “Update Precipitation Frequency Estimates for Louisiana (NOAA Atlas 14 Volume 14).” Solicitation 1543. https://www.pooledfund.org/Details/ Solicitation/1543.