U.S. Department of Transportation
Federal Highway Administration
1200 New Jersey Avenue, SE
Washington, DC 20590

Skip to content
Facebook iconYouTube iconTwitter iconFlickr iconLinkedInInstagram

Federal Highway Administration Research and Technology
Coordinating, Developing, and Delivering Highway Transportation Innovations

Public Roads
This magazine is an archived publication and may contain dated technical, contact, and link information.
Public Roads Home | Current Issue | Past Issues | Subscriptions | Article Reprints | Guidelines for Authors: Public Roads Magazine | Sign Up for E-Version of Public Roads | Search Public Roads
Back to Publication List        
Publication Number:  FHWA-HRT-18-001    Date:  Autumn 2017
Publication Number: FHWA-HRT-18-001
Issue No: Vol. 81 No. 3
Date: Autumn 2017


Rivers, Rainfall, and Resilient Roads

by Brian Beucler

An updated FHWA manual on highways in floodplains provides methods for assessing their vulnerability to extreme events. Check out this timely publication.

Dave Claman, Iowa DOT
Photo showing a flooded section of a multilane highway.
In June 2008, the Cedar River at Cedar Rapids in Iowa had catastrophic flooding that exceeded the 500-year flood by 1.4 times. Downstream, I–80 was closed for 4 days and required a 120-mile (193-kilometer) detour.

Almost every day, it seems, news outlets show scenes of homes and communities flooded somewhere in the United States. Given the formidable power of rushing water, the floods could have substantial impacts on the Nation’s communities, highways, and bridges.

What are the impacts of extreme floods, are dangerous floods more frequent, and what can State departments of transportation do to lessen future damage to their road networks? To find out, read on.

The phrase used by scientists and transportation planners–the “riverine environment”–refers to waterways, rivers, streams, lakes, wetlands, and other natural resources that convey water. The riverine environment also includes floodplains, which are land areas that are susceptible to inundation resulting from the overflow of adjacent rivers and streams. Many roads, bridges, and other components of transportation systems are located adjacent to or near the riverine environment.

Floodplains are subject to extreme floods, such as the 1,000-year flood of June 2016 in West Virginia, which killed 23 people, and the June 2008 flooding in the Iowa and Cedar River basins in Iowa, which necessitated $1.3 billion in Federal disaster assistance and an additional $583 million obligated for improvements to infrastructure. Extreme floods are driving the need to develop more resilient approaches to the planning, design, and operation of transportation systems located within floodplains.

To address this need, State DOTs are scrutinizing their budgets to ensure that investments in highway infrastructure consider the risks to roads, bridges, and culverts in the face of extreme floods. With risks increasing because of development in floodplains, DOTs are looking to improve the resilience of transportation infrastructure to dangerous floods.

Consequently, transportation agencies need updated and expanded guidance on methods of assessing the risks, techniques for estimating extreme floods (including changing precipitation patterns), and strategies for reducing the vulnerability of transportation assets to extreme events.

To fulfill these needs for guidance and strategies, the Federal Highway Administration published the second edition of Hydraulic Engineering Circular No. 17 (HEC-17) Highways in the River Environment–Floodplains, Extreme Events, Risk, and Resilience (FHWA-HIF-16-018) in June 2016. This updated publication contains more than 150 pages and numerous illustrations.

Why a New HEC-17 Manual?

The manual was in need of an update. For one thing, the first edition of HEC-17 was published in 1981– more than 35 years ago.

Cover of the Hydraulic Engineering Circular 17, second edition: Highways in the River Environment—Floodplains, Extreme Events, Risk, and Resilience.
HEC-17 is one in a series of technical reference manuals published by FHWA. The updated second edition addresses risk assessment methods and strategies for reducing the vulnerability of transportation assets to extreme events.

That initial edition focused on the risk analysis of encroachments, such as highway embankments, in the floodplain. In particular, the publication described a decision-making process referred to as the “least total expected cost” method, which could be applied to a specific asset with its own unique conditions.

The alternative is analyzing risk in terms of a specified design event, such as a 100-year storm (one that has a 1 percent probability of occurring in any given year). Although the first edition of HEC-17 includes important risk concepts, the second edition significantly expands on the subject of risk analysis, while also focusing on extreme events, resilience, climate science, and designing in the face of uncertainty.

In the manual’s introduction, FHWA states that it intends HEC-17 to represent the “best available and actionable engineering and scientific data and approaches” that can be applied to the design of riverine highway infrastructure, so that it will be resilient to extreme events in the future. Data and methods that do not meet FHWA’s definition of “best available” or “actionable” were purposely excluded from the manual. The approaches outlined in the manual are grounded in sound science and can be clearly implemented without great complexity or ambiguity.

“As climate science and floodplain regulations continue to evolve, FHWA will be updating its technical publications, including HEC-17, to keep pace,” says Michael Culp, team leader of the Sustainable Transportation and Resilience Team in FHWA’s Office of Planning, Environment, and Realty.

HEC-17 also provides a consistent, nationwide approach to incorporating the potential effects of extreme events and changes in climate, with respect to the hydrologic design of roads, bridges, and culverts. Rather than completely abandon traditional hydrologic design methods, HEC-17 recommends that information from climate models incrementally augment or inform the traditional design, based on the criticality and remaining service life of the asset being considered.

For example, the longer a culvert remains in service, the likelihood increases that it will see a damaging flood. Also, the more expensive the culvert is, and the stronger the increase in precipitation projected by climate models, the more significance the design needs to place on outputs from those models. The manual includes tables and specific numerical recommendations to help the designer determine when, and to what extent, those outputs should be incorporated into designs.

HEC-17 limits its focus to riverine hydrology, floodplains, extreme events, risk, and resilience. A related publication, HEC-25, Highways in the Coastal Environment: Assessing Extreme Events (FHWA-NHI-14-006), addresses the vulnerability assessment of highways exposed to sea level rise and coastal extreme storm events. State DOTs and the American Association of State Highway and Transportation Officials frequently cite and use the concepts contained in HEC manuals to inform their hydraulic manuals and design procedures.

Contents of HEC-17

The manual begins with an introductory chapter on the publication’s scope, purpose, and target audience, plus a list of related FHWA guidance materials and Web sites. The next chapter describes the evolution of Federal floodplain regulations and executive orders. The third chapter defines terms and reviews hydrologic methods to compute floods from rainfall/runoff models and statistical methods using streamgages. This chapter also discusses uncertainties associated with traditional hydrologic methods, irrespective of climate impacts.

The fourth chapter defines “nonstationarity” (meaning the past is not necessarily a predictor of the future) and lists sources of nonstationarity, including projected future changes in temperatures and precipitation patterns. Also outlined are statistical techniques to adjust datasets for nonstationary effects.

Chapter 5 presents a brief overview of climate science, climate models, and the tools available to extract temperature and precipitation data relevant to the hydrologic design of highway drainage infrastructure. Chapter 6 defines the concepts of risk and resilience and the consequences of failure to modify the design process. Also listed are potential adaptation strategies to increase resilience and reduce vulnerability to extreme events.

Chapter 7 presents a five-level hydrologic design/analysis framework that incrementally increases the consideration of the effects of climate change based on remaining service life, asset criticality, and strength of projected trends (increases) in precipitation amounts. This chapter is the heart of the manual, providing the reader with a practical method to incorporate extreme events and nonstationarity into the hydrologic design of highway assets.

The final chapter outlines several case studies conducted by State DOTs and Federal agencies illustrating the various levels of analyses presented in the previous chapter.

Climate science is a complex and rapidly changing field. Computer models are becoming more sophisticated, and more measurements are being taken of the Earth’s reaction to increasing temperatures and changing precipitation patterns. Most aspects of climate science and nonstationarity and even some of the risk and resilience concepts might be new to some DOT designers, so digesting the chapters that precede the five-level analysis framework is critical.

Evolving Federal Floodplain Policy

FHWA developed the manual in 2016 partly as a response to the requirements set forth in sections of Moving Ahead for Progress in the 21st Century (MAP-21) related to floodplain development; Federal Emergency Management Agency (FEMA) regulations governing the National Flood Insurance Program; FHWA orders, such as Order 5520 “Transportation System Preparedness and Resilience to Climate Change and Extreme Weather Events”; and 23 CFR 650 Subpart A, the section in the Code of Federal Regulations that deals with the “Location and Hydraulic Design of Encroachments on Flood Plains.” For reference, these are all contained in one of the manual’s appendices.

The second chapter of the manual helps the reader navigate the history of floodplain regulation. It answers the question, “How did we get here?” It is important to know the lines of authority that FHWA has in floodplain regulation and the mechanisms it can use to reduce the impacts of highway encroachments on floodplains.

National Floodplains Policy

Graphic. The figure provides a timeline from 1930 through 2020 divided into three periods: (1) “Recognizing the Need” (1930 through 1970), (2) “Maturation of Policy” (1970 through 2000), and (3) “21st Century Policy” (2000 through 2020). Specific key milestones are shown: 1936 - Flood Control Act; 1954 - Watershed Protection and Flood Prevention Act; 1966 - EO 11296/House Document 465; 1977 - EO 11988; 1978 – Water Resources Council Guidance; 1979 - 23 CFR 650; and 2015 - EO 13690.

* HD 465 is House Document 465. WRC is the Water Resources Council.
Source: FHWA.

Joe Krolak, principal hydraulic engineer in FHWA’s Office of Bridges and Structures, says, “In 2012, MAP-21 added ‘extreme events’ as a new category of considerations, such as scour countermeasures or seismic retrofits, that transportation folks could consider during project delivery. HEC-17 provides FHWA’s understanding of how to properly account for extreme events in the riverine environment.”

Traditional Hydrologic Design

The manual’s third chapter defines key terms and describes traditional methods that engineers use to estimate flood discharges. Hydrologists and hydraulic engineers employ many terms to describe rainfall, runoff, flow in rivers, and flow large enough to be considered a design event or an extreme flood event. In addition, rainfall is more precisely defined as precipitation, which can be in the form of rain, hail, sleet, or snow. Rainfall over a watershed can vary temporally and spatially, especially in larger watersheds.

In addition to considering the size of an event, a second component of hydrologic design is the watershed’s reaction to precipitation. How much becomes runoff? How much is intercepted by leaves and other vegetation? How much infiltrates directly into the ground based on soil types in the watershed or impervious coverings such as pavements and buildings? How wet and saturated (from recent rainstorms) is the watershed before the precipitation falls? Has the watershed been burned by a wildfire? All these interactions contribute to the accumulation of runoff that eventually produces a peak flow at a specific point in a river, perhaps at a bridge or a culvert.

Russell Pankratz
Photo. Shown is a raging flood. Not seen: any roads, but a dramatic photo of the river.
Shown is flooding on the Pedernales River in Texas in August 2007. Dangerous flooding like this puts bridges and culverts at risk.

Floods occur when accumulated flows exceed a river’s carrying capacity and land that is normally dry becomes inundated. When scientists describe a flood as a 100-year flood, they are describing the return period of the flood–the average length of time between occurrences in which the magnitude of that flood is equaled or exceeded. The time between two 100-year floods can be shorter or longer than 100 years.

A more useful way of characterizing a 100-year flood is to define the annual exceedance probability (AEP) of the flood. The AEP is the probability that the magnitude of a flood will be equaled or exceeded in any given year. The 100-year flood has a 1-percent chance of being met or exceeded in any given year.

Hydraulic designers may specify a given AEP as the design flood for the infrastructure being designed, depending on the infrastructure’s criticality and the risks associated with physical damages and other losses such as traffic interruption. Extreme flood events are not normally used as design floods because designing for them might not be economically justifiable.

Source: NOAA.

NOAA's Precipitation Frequency Data Server

Screenshot. An example of data available from NOAA's Precipitation Frequency Data Server.

This screenshot shows precipitation intensities for different durations and annual exceedance probabilities. Ninety-percent confidence intervals are shown in parentheses.

The manual’s third chapter also classifies two main methods to calculate flood discharges: rainfall/runoff and statistical. Rainfall/runoff models use precipitation in the form of rain as the primary input. Rainfall data are obtained from local or national weather service sources such as the National Oceanographic and Atmospheric Administration’s (NOAA’s) Precipitation Frequency Data Server.

The other method for calculating flood discharges is statistical models, such as regression equations and the Log-Pearson TypeIII method, which are based on statistical analysis of records taken directly from streamgages. Statistical methods are considered superior to rainfall/runoff methods because they are based directly on observed streamflow measurements. Various AEPs can be derived and graphed from these measurements, as represented by a flood frequency curve.

Source: FHWA.

Flood Frequency Curve

Graph. The figure shows an example flood frequency curve with confidence limits. The vertical axis is labeled “Annual Peak Discharge (cubic feet per second)” and ranges from 100 at the bottom to 100,000 at the top (logarithmic scale). The horizontal axis is labeled “Annual Exceedance Probability, Percent Station 03606500 Sandy River at Bruceton, TN.” The curve on the plot ranges from the lower left to the upper right, showing an increase in discharge with a decrease in annual exceedance probability. The upper and lower confidence limits are plotted above and below the curve, respectively.

The figure shows an example of a flood frequency curve with confidence limits. The red line curve represents the best fit of a single line to the datapoints showing annual peak flow (the highest daily flow of any year) recorded at a streamgage through its operational history. The measure of uncertainty associated with this line is called the “confidence interval.” This interval is defined by an “upper confidence limit” and a “lower confidence limit” represented by the curved blue lines. As annual peak datapoints become more scarce, toward the extremely low (dry) or extremely high (wet) values, the confidence interval becomes wider (more uncertain). HEC-17 encourages designers to consider the full range of values within the blue confidence limits rather than just the computed value (red line).

HEC-17 calls for the designer to consider a range of possible discharge values, especially for assets that are to remain in service for longer than 30 years. Later under step 3 in chapter 7, HEC-17 uses the upper 90 percent confidence limit from historical precipitation records to determine whether future precipitation projections from climate models should be weighted more heavily in the design. If the future precipitation projections are not deemed significantly different from the historical observations, design decisions can be made based on the historical observation dataset. HEC-17 still advises the designer to consider the uncertainty associated with the historical observations, especially with important assets or assets that are to remain in service for a long time.

Nonstationarity: The Past Offers No Guarantees

Traditional hydrologic design operates on the assumption of a climate that exhibits stationarity, which holds that data collected from the past is representative of the future. By contrast, nonstationarity means that past patterns or trends may not be valid in the future. As discussed in the fourth chapter of the manual, nonstationarity essentially introduces uncertainty into the use of historical observations for estimating future flooding.

Causes of nonstationarity include changes in watershed land cover from pests, the migration of plant species, wildfires, and urban development (increased impervious surfaces); changes in precipitation patterns; introduction or removal of water storage ponds or dams within a watershed; and stream diversions for municipal (drinking water) or agricultural (irrigation) purposes. Nonstationarity caused by changes in precipitation patterns often can be minor compared to some of these other causes.

Source: USGS.

Trends in Annual Peak Streamflow

Map. This is a map of the contiguous United States showing 1,491 streamgage locations with triangles pointing either upward to represent significant upward trends or downward to represent significant downward trends, or dots representing locations exhibiting no trends. Approximately 36 triangles are pointing upward, 54 triangles are pointing downward, and the remaining approximately 1,401 stations are dots. There are no regions of the country showing consistent trends.

Streamgages with positive trends are shown by upward-pointing triangles and those with negative trends are shown by downward-pointing triangles. Dots represent gages with no significant trends. There are no regions of the country showing consistent trends.

Moreover, nonstationarity could occur as an abrupt change, a periodic variability, or a long-term trend. An abrupt change might be due to a wildfire scorching a watershed. A periodic variability might be a multiyear weather oscillation such as the El Ninjo–Southern Oscillation or ENSO. A long-term trend could be a change in precipitation patterns, which climate models project to increase in the northern latitudes of the United States. Sometimes a long-term trend can be masked by periodic variabilities.

The fourth chapter provides information on how to statistically detect significant gradual and abrupt trends and how to adjust a project’s design for these trends. Chapter 4 also warns readers to be careful when calling something a trend. The length of record being considered could determine whether a trend is significant or just an intermediate fluctuation.

Climate Modeling

Chapter 5 covers global climate models (GCMs), emissions scenarios that drive the GCMs, and downscaling of the GCM outputs to make them more useful in transportation design. Global climate models are sophisticated computer models used to represent atmospheric physics and the connections between the Earth’s atmosphere, land surface, sea surface, and polar ice.

By design, GCMs operate on large spatial and temporal scales and are useful at modeling long-term trends of average global temperature and, to a lesser extent, precipitation changes. Many GCMs have been developed worldwide and run on different assumptions and formulations to simulate atmospheric processes. GCM outputs are called “projections” and are commonly expressed as future estimates of daily precipitation and minimum and maximum daily temperatures. Climate scientists often examine outputs from multiple GCMs or an “ensemble” of models, in order to account for the range of possible results.

Because global climate models are designed to model atmospheric conditions for the entire planet, their computational grid cells are large, usually between 120 to 190 miles (193 to 306 kilometers) per side. In order to be useful at a transportation design scale, one of two types of downscaling is used to break the cells into smaller units. Dynamic downscaling uses a smaller regional climate model that is nested within a larger climate model. Statistical downscaling uses statistics to relate the GCM outputs to historical measurements of temperature and precipitation.

Source: FHWA.

Emissions Scenarios

Graphs. Two graphs compare global climate models (GCM) output for temperature change for selected Special Report on Emissions Scenarios (SRES) and RCP scenarios. The SRES A2 and B1 scenarios are plotted on the left, along with historical data. On both graphs, the horizontal axis is labeled “Year” and ranges from 1900 through 2100, and the vertical axis is temperature change in degrees Fahrenheit and is divided into 2 degree Fahrenheit increments from -2 to 12. The projected increase for the A2 scenario is about 7 degrees by 2100 and for the B1 scenario is about 4 degrees. Both are shown with mean and range of values. The RCP 2.6 and 8.5 scenarios are plotted on the right with the same labeling. The projected increase for the 8.5 scenario is about 8.5 degrees by 2100 and for the 2.6 scenario is about 2.6 degrees. Both increases are shown with mean and range of values.

These two graphs compare GCM output for temperature change for selected (1) special report on emissions scenarios (SRES) and (2) representative concentration pathway (RCP) scenarios. Some of the scenarios are more severe than others. HEC-17 recommends looking at scenarios of RCP 6.0 and greater, especially for long-lived, critical assets.

FHWA recommends retrieving statistically downscaled climate projections at the “Downscaled Climate and Hydrology Projections” (DCHP) Web site, http://gdo-dcp.ucllnl.org/downscaled_cmip_projections. Instructions for retrieving these data are contained in FHWA’s CMIP Tool (see www.fhwa.dot.gov/environment/sustainability/resilience/tools). CMIP refers to the World Climate Research Programme’s Coupled Model Intercomparison Project, which shares, compares, and analyzes the latest outcomes of GCMs. Grid cells from the DCHP site are 7.5 miles (12kilometers) per side, which will better account for local variations in temperature and precipitation, such as between a valley and a mountainous ridge.

Risk and Resilience

Although many definitions of risk can be found, they all involve the consequences associated with hazards (including climatic) and the probabilities of those hazards occurring. Design criteria or standards are based on a community’s tolerance for risk. Critical roads such as interstate highways will have lower tolerance for risk than collector system roads with less traffic. Consequences can be in the form of loss of life, physical damage to the infrastructure, or interruption of service requiring detours.

The manual’s sixth chapter discusses the expected performance of an asset over its design life. An example would be that the probability of a 50-year storm striking within a 75-year design life is 78 percent. The longer the design life (say 100 years), the higher the probability of that 50-year storm occurring (now 87 percent).

Risk may change over the lifetime of an asset, especially if the asset is located in a watershed that is experiencing nonstationarity caused by changes in precipitation patterns or land use changes in the watershed itself. Consider the case of a watershed where the calculated magnitude of the 2 percent AEP flood is 1,000 cubic feet (28 cubic meters) per second today, but in the future the 2 percent AEP may be 1,200 cubic feet (34 cubic meters) per second. This case demonstrates the need to consider nonstationary changes seriously if a critical asset is expected to withstand higher flows long into the future.

Source: FHWA.

Probability of Storm Occurrence

Graph. Plot with the horizontal axis labeled “n equals Years of Service,” ranging from 1 to 100 on a logarithmic scale, and the vertical axis labeled “Probability of Occurrence,” ranging from 0.01 to 1.00 on a logarithmic scale. A family of five curves is shown moving from the lower left to the upper right of the plot. From top to bottom, the five curves are T equals 10 years, T equals 25 years, T equals 50 years, T equals 100 years, and T equals 500 years.

This chart shows the probability of occurrence as a function of return period, T, and years of service. There is an 87-percent chance that a 50-year return period storm (green dashed line) will occur within a 100-year service life.

Chapter 6 also recommends that the designer look at a range of possible storm events based on the statistical uncertainties described in chapter 4 and based on a range of emissions scenarios and ensembles of multiple datasets of downscaled climate model projections. The designer then needs to examine the consequences of particular flow events within that range. Those consequences, or that risk, may change over time. The designer also needs to judge how resilient that asset is to the range of possible flows. A resilient asset, such as a culvert or bridge, will be able to pass the design event without serious consequences.

“Resilience” is defined as the ability to anticipate, prepare for, and adapt to changing conditions and withstand, respond to, and recover rapidly from disruptions. A culvert may be more resilient if it is under a deep roadway fill, where it might take more flow to overtop and damage the roadway. This culvert might still suffer consequences, such as damage to the roadway embankment, flooding of upstream properties, and erosion at the culvert outfall, but it might perform better than a culvert under a shallow fill where an equivalent flow may wash out the roadway section entirely.

© NikonShutterman, Getty Images
Photo. Striped center lines on the road are visible underwater. The flooded road appears to be in a low floodplain.
Shown is a flooded road in North Carolina. This road could be raised to possibly increase its resilience.

Moving Stepwise Toward A More Resilient Design

As mentioned earlier, chapter 7 is the heart of the manual. This chapter provides a straightforward framework or procedure to look at an existing hydrologic design and systematically incorporate downscaled climate projections, based on the criticality and design life of the asset, and the significance of the nonstationary trend in the climate.

The framework consists of five levels of analysis, ranging from traditional hydrologic design to sophisticated, specialized analyses, including downscaled climate projections. At the conclusion of each level, the designers evaluate whether they are comfortable with the level of risk determined for the asset, or whether they need to move ahead and further examine the significance of climate impacts on the asset.

The chapter also includes information on the types of hydrologic tools and methods that may be appropriate for different levels of analysis, recommendations on confidence intervals to examine for different lengths of design life, and even advice on the composition and skill sets of the members of the design team.

In addition, the chapter talks about programmatic information that results from formal large-scale climate studies performed within a State or vetted by its DOT. This programmatic information can help the agency make broad policy decisions as to how to assess risk and incorporate resilience into projects within the State and when projects warrant a higher level of climate analysis. This information also would also provide designers with standardized, easy-to-use methods to include local climate projections into their projects.

The five levels of analysis are discussed in detail in chapter 7:

Source: FHWA.

Five Levels of Analysis

Graphic. Five color blocks from the top to the bottom are labeled: “1. Historical Discharges”; “2. Historical Discharges and Confidence Limits”; “3. Precipitation Projection Trend Test”; “4. Projected Discharges Using CMIP Tool”; and “5. Customized Projected Discharges with Climate Scientist”

The five levels of analysis range from traditional hydrologic design to sophisticated specialized analyses that include downscaled climate projections.

Level 1–Historical discharges. At level 1, the design team applies standard hydrologic design techniques based on historical data to estimate the design discharge. In addition, the team qualitatively considers changes in the estimated design discharge based on possible future changes in land use (runoff coefficient) and climate. Note that no climate projection data are required for this level. This step could be as simple as making sure that the latest hydrologic observations are used in the design.

Level 2–Historical discharges/confidence limits. Here, the design team estimates the design discharge based on historical data and qualitatively considers future changes in land use and climate, as in level 1. In addition, the design team quantitatively estimates a range of discharges (confidence limits) based on historical data to evaluate the performance of a plan or project. Again, no climate projection data are required, but a designer might consider a higher magnitude event (above the design event) within the range of the confidence limits.

Level 3–Historical discharges/confidence limits with precipitation projections. At level 3, the design team performs all level 2 analyses and quantitatively estimates projected changes in precipitation for the project location. The team evaluates the projected precipitation changes to determine whether a higher level of analysis is appropriate. This level is really just a test to determine whether the change in projected precipitation because of a changing climate is large, as compared to the uncertainty in the current observed data (as represented by the upper 90 percent confidence limit). If the change is large, move on to level 4. This level requires obtaining downscaled climate projections, preferably from the DCHP Web site.

Level 4–Projected discharges/confidence limits. At level 4, the team completes all level 3 analyses and develops projected land use and climate data, where feasible. The team performs hydrologic modeling using the projected land use and climate data to estimate projected design discharges and confidence limits. Downscaled climate projections are required and are used to evaluate the design. Consultation from experts in climate science is recommended but not required.

Level 5–Projected discharges/confidence limits with expanded evaluation. Here, the design team performs the equivalent of the level 4 analyses based on customized projections of land use and climate. The team also expands to include appropriate expertise in climate science and/or land use planning to secure site-specific custom projections.

Examples of the levels summarized above can be found in chapter 8 and were also provided in the HEC-17 webinars.

HEC-17 Webinars

In early 2017, FHWA produced a series of three webinars outlining the HEC-17 manual in detail, including stepping through an example involving the design of headwater elevation for a culvert, using three of the five levels of analysis. A culvert example from the webinar shows results progressing from the single flow calculated in level 1, to the range bounded by confidence limits in level 2, to a further expanded range considering climate projections in level 4.

Where Do We Go From Here?

Chapter 8 contains examples of analyses performed by various State DOTs and categorized in HEC-17 as examples of various levels of analysis. The chapter on these projects also contains commentary highlighting good practices and lessons learned.

The projects range from a level 2 bridge analysis in Connecticut to a sophisticated level 5 analysis of six bridges in two watersheds in Iowa. The Iowa examples used a continuous daily simulation of rainfall from customized, downscaled climate projections developed at Iowa State University and a complex watershed model developed at the University of Iowa. This project was featured as one of the FHWA resiliency pilots, which can be found at www.fhwa.dot.gov/environment/sustainability/resilience/pilots.

Climate science continues to progress as global climate models become faster and more comprehensive. Downscaling techniques and datasets included on the DCHP Web site continue to improve. Also, floodplain regulations continue to evolve as local, State, and Federal governments struggle with how to build resilient infrastructure and deal with extreme events and their impacts on U.S. society.

Source: FHWA.

Level 1: Traditional Hydrologic Design

Photo. Shows two men standing in front of a culvert, with a green line superimposed on the photo about at head height, approximately 4 feet below the top of the culvert. The green line indicates the flow elevation at 371 cubic feet per second computed from traditional hydrologic design without the effects of climate.

Level 2: Confidence Limits

Photo. Shows two men standing in front of a culvert, with a dotted green line superimposed on the photo approximately 4 feet below the top of the culvert. Also superimposed on the photo is a block of yellow shading from the top of the culvert to about waist height on the men. There is a solid yellow line at the top of the yellow shaded block, indicating the flow elevation at the 90-percent upper confidence limit flow of 659 cubic feet per second. The shaded yellow block represents the range of elevations that could be produced from the 90-percent upper confidence limit (the solid yellow line at the top of the block) to the lower 90-percent confidence limit indicated by the bottom of the block.

Level 4: Expanded Confidence Limits

Photo. Shows two men standing in front of a culvert, with a dotted green line superimposed on the photo approximately 4 feet below the top of the culvert. Also shown is a yellow dotted line at the height of the top of the culvert, and there is a red shaded block from about two feet above the top of the culvert (at about road level) to about the height of the men's waists. At the top of the red block is a solid red line, indicating the 90-percent upper confidence limit of flow elevations from climate-adjusted flow of 755 cubic feet per second. The bottom of the red shaded block indicates the elevation of the lower 90-percent confidence limit associated with the climate-adjusted flow of 755 cubic feet per second.

During the webinars, this example of designing headwater elevation for a culvert used three of the five levels of analysis from the HEC-17 manual: Level 1 (traditional hydrologic design), Level 2 (consideration of confidence limits), and Level 4 (expanded confidence limits after incorporation of climate projections).

FHWA is participating in several research efforts with science agencies, including the U.S. Geological Survey and NOAA, to better refine the agency’s guidance in this field. In addition, the National Cooperative Highway Research Program’s Project 15-61, “Applying Climate Change Information to Hydrologic and Hydraulic Design of Transportation Infrastructure,” is a major research effort involving a multiagency panel of State DOT hydraulic engineers, planners, climate scientists, and hydrologists.

In 2018, FHWA will turn its attention back to coastal highways, updating and expanding the companion coastal hydrology circular HEC-25. When this is completed, most of the research listed above will be ready to be incorporated into a third edition of HEC-17. Recent extreme events such as Hurricane Harvey in the Houston, TX, area highlight the need for more resilient future highway infrastructure investments to be made considering possible changes in precipitation patterns as well as land use changes occurring in surrounding watersheds. HEC-17 provides a scientifically informed, comprehensive, and balanced approach to designing for future uncertainties.

Brian Beucler is a senior hydraulics engineer with the Hydraulics and Geotechnical Engineering Team in the FHWA Office of Bridges and Structures, where he focuses on technical guidance in the areas of inland and coastal hydrology and resilience. He holds an M.S. degree in civil and environmental engineering from The George Washington University and a B.S. degree in civil engineering from the University of Virginia.

For more information, see www.fhwa.dot.gov/engineering/hydraulics/pubs/hif16018.pdf or contact Brian Beucler at 202–366–4598 or brian.beucler@dot.gov. The recorded webinar series on HEC-17 can be found at www.fhwa.dot.gov/engineering/hydraulics/media.cfm.



Federal Highway Administration | 1200 New Jersey Avenue, SE | Washington, DC 20590 | 202-366-4000
Turner-Fairbank Highway Research Center | 6300 Georgetown Pike | McLean, VA | 22101