Green and Gray Infrastructure

Traditionally, engineers have relied on built structures like stormwater collection systems, levees, dams, and sea walls to protect areas from flooding. These systems are designed to convey water away from important infrastructure and structures (e.g., stormwater pipes, culverts, and drainage ditches), retain and slowly release water to reduce peak flows downstream (e.g., stormwater ponds and dams), or hold back rising water along coasts and rivers (e.g., levees and sea walls). Temporary flood barriers such as sandbags are used as emergency measures when existing flood protections fail. These engineered systems are known as gray infrastructure [29].

While gray infrastructure performs well in many scenarios, it requires large upfront capital investments, is typically single-purpose rather than multifunctional, and is inflexible in terms of placement and management approach.

Green infrastructure uses natural processes to treat, infiltrate, and store water onsite. Well-designed green infrastructure systems can work well to reduce flood risk in highly urbanized environments, especially if they are widely distributed. A 2015 study of flood reduction benefits of green infrastructure indicated that, if implemented consistently at a national level, green infrastructure could generate an annual flood loss reduction benefit of $63 million to $136 million by 2040 [34]. Green infrastructure also provides co-benefits such as air and water quality improvement, reduction of the urban heat island effect, and improved biodiversity.

However, green infrastructure requires proper maintenance and upkeep for optimal performance; for systems that are not well designed and maintained, clogging can be a significant challenge. Typically, smaller green infrastructure systems are not designed to handle extreme weather events. Green infrastructure also has greater performance uncertainty associated with it than gray infrastructure, due to variables such as the type of soil on which the infrastructure is built, the type of vegetation selected, construction parameters, clogging potential and inconsistent maintenance, spatial variability in placement, and ability to handle back-to-back storm events [35] [36] [37].

Despite these drawbacks, a combination of green and gray infrastructure allows cities a balance between flexibility, cost, co-benefits, and stormwater protection for a range of storm events.

Making Systems “Safe to Fail”

Flood management inherently involves some level of risk: no infrastructure can be designed to fully function for all possible flood events. Flood protection infrastructure uses historical rainfall and flood data to design to an acceptable level of risk. Larger structures can protect from larger, less likely events, often with a designed factor of safety built in. When protecting particularly critical infrastructure (like emergency access routes), multiple flood management structures are commonly used to provide systemic redundancy. These systems are designed to be reliable, meaning they have a very low probability of failing to function. In other words, they are designed to be “fail-safe.”

However, large structures are more expensive to construct. In addition, once they are constructed, they cannot easily adapt to changing landscape or climate paradigms. This assumption of climate stationarity may no longer be an appropriate approach to designing flood management systems; studies show that a stationarity assumption can lead to underestimating extreme precipitation by as much as 60 percent [38]. When a large structure fails (as in the case of the levee system in New Orleans during Hurricane Katrina), the damage is catastrophic compared with the failure of a smaller structure [39]. This risk is exacerbated by the so-called “levee effect”: the presence of large flood management structures (like levees) can increase development in the floodplain behind the levee because of the perceived safety it provides [40].

An alternative approach is to design for systemic resilience rather than just reliability. A resilient system is one that can bounce back from a stressor event quickly. One approach to design resilience is looking at ways to make systems “safe to fail” rather than (or in addition to) being “fail-safe.” “Safe to fail” systems prioritize two factors: 1) maintaining critical system-wide services rather than focusing on preventing individual component failure; and 2) minimizing the consequences of an extreme event rather than minimizing the probability of damages. An example of a “safe to fail” system is protecting sections of a river’s floodplain from development to purposefully allow them to flood during storm events, thus protecting other sections of the watershed. Because of their design ethos, “safe to fail” approaches are often dispersed and autonomous rather than centralized and are designed to provide co-benefits besides flood protection [39].

Coastal Management

For communities threatened by coastal erosion, storm surges, and sea level rise, coastal management is integral to protecting assets and infrastructure against future flooding. Generally, coastal management can be subdivided into two types of engineering approaches: hard and soft. Both approaches can be mixed and matched as appropriate for a given project site. 

Hard engineering approaches use structures made of metal, concrete, stone, or other engineered materials to dissipate the energy of breaking waves, protecting inland areas from storm surges and helping prevent coastal erosion. If properly designed, some hard engineering structures can help to replenish beaches by encouraging the deposition of sediments carried by longshore drift (i.e., the transportation of sediments by waves and currents parallel to the shore). Other structures (such as flood gates) have a highly specific purpose, such as protection from storm surges. Examples of hard engineering approaches include groynes, seawalls, revetements, rock armor, breakwaters, gabions, geotextile tubes, and flood gates. While hard engineering approaches are effective, they can be expensive and unsightly. In certain cases, the presence of many engineered structures can actually contribute to beach dissipation further along the coast from the project site [41].

Groyne

Seawall

Revetment

Gabion

Geotextile Tube

Flood Gate

• Soft engineering approaches leverage the protective properties of dunes and natural coastal buffers to prevent coastal erosion and protect shorelines from storm surges. Coastal dunes form a flexible barrier to high tides and waves into inland areas. However, unstable dunes are highly vulnerable to erosion by wind and water [42].
• Dune stabilization is the practice of using vegetation and structures to prevent dune erosion and migration. Vegetation used in dune stabilization needs to be hardy: the plantings need to cope with low organic matter in the soil, saltwater, sandblasting, drought, flooding, and heat [43].
• Sand fences are wooden slats supported by posts placed perpendicular to the prevailing wind; the structures slow down the wind at ground level, allowing sand to be deposited behind the fence. Other structures such as elevated pedestrian walkways are vital to ensuring that foot traffic does not damage sensitive dune plantings. Dune stabilization may be used in conjunction with beach nourishment, the practice of importing sand to build up a beach that has already been partially eroded by storms [42]. 

Yet another soft engineering practice is the creation of living shorelines (also known as nature-based, green, or soft shorelines) to improve coastal resilience. A living shoreline encompasses a range of techniques but typically uses a combination of native vegetation and harder shoreline structures such as oyster beds or coral reefs to stabilize and protect coastal areas. Living shorelines can include engineered structures that make use of natural processes (like coral reef balls) or coastal buffer zones designed to project natural barriers to wave erosion [42]. Saltwater marshes have been shown to dissipate wave energy by 50 percent within the first 2.5 meters [44]. Mangrove forests reduce wave heights (and energy) by up to 66 percent within the first 100 meters of approach [45]. 

Nonstructural Flood Mitigation

Flooding is not just a meteorological phenomenon: damages from flooding are driven in part by where and how communities choose to develop within an existing floodplain. Land use and zoning practices are thus a crucial aspect of flood management. Local governments may pass laws that prevent construction in areas prone to flooding or can require amendments to building codes (such as building elevation or floodproofing requirements) in these areas. In doing so, governing bodies need to balance the benefits of economic development alongside the long-term risks of flooding [40]. Complicating this analysis is the fact that many communities living in low-lying, vulnerable areas that are at high risk of future flooding tend to be low-income and minority [46]. Stifling economic growth in these areas in the name of flood control can exacerbate existing social justice issues [47]. 

Flood insurance is another tool that reduces flood risk by transferring some of the cost of damages from the homeowner to the entity providing the insurance. In the United States, the National Flood Insurance Program (NFIP) has provided flood protection for nearly 22,000 communities since 1973, when Congress made federal flood insurance mandatory for properties with federally backed mortgages [40]. In October 2021, the Federal Emergency Management Agency (FEMA) initiated Risk Rating 2.0, a strategy designed to shift the basis of insurance premium pricing away from the relative elevation of a building within a designated flood zone and towards a more customized approach. Risk Rating 2.0 premiums will account for flood type, cost to rebuild, distance to water source, property elevation, and discounts for property flood mitigation strategies [48]. 

Flood maps (See "By the Numbers: Flood Mapping") developed by FEMA determine whether a property development is required to acquire flood insurance under the NFIP. These maps are known as Flood Insurance Rate Maps (FIRMs); some communities have digitized maps known as dFIRMs. 

While the NFIP has had some success in buffering property owners from the impacts of devastating flood events, it has encountered significant challenges in meeting its other stated goal: acting as an instrument to limit development in areas prone to high flood risk. Part of the challenge is a lack of enforcement. According to a 2015 FEMA study, less than half of the 1.5 million residential buildings mandated to have flood insurance actually do [49]. Multiple studies have also argued against the “moral hazard” of subsidized federal flood insurance. In general, NFIP premiums are much less expensive than what private insurance would charge (historically, most premiums cost about half the private market rate). Furthermore, nearly 20 percent of NFIP insurance policies receive a subsidy of some kind. Affordable flood insurance sends a distorted market signal that underestimates the true cost of living in a flood-prone area [40]. Over the course of the NFIP’s history, repetitive loss properties (properties for which there have been two or more claims worth over $1000 in any ten-year period) have cost the program $12.5 billion. Partly as a result, since Hurricane Katrina (2005), the NFIP has racked up nearly $25 billion in debt [49]. 

Many of the challenges associated with the NFIP have to do with its underlying flood maps. A FIRM boundary is only as accurate as its underlying topographical map. FIRM maps are typically drawn on top of USGS quadrangle topographical maps, which have a fairly large scale, so the boundaries of the floodplain may not be accurate, particularly in very flat areas [48]. Many FIRMs are also out of date, with nearly 66 percent having not been updated in the last five years and several having not been updated since 1983 [50]. As a result, the maps may not account for development, meteorological events, or other activity that might significantly shift the floodplain boundaries. A 2020 report by the Association of State Floodplain Managers estimated an annual required price tag of between $107 million and $480 million to keep existing FIRMs up to date [51]. 

Because of the work involved in developing flood maps, FEMA focuses its mapping efforts on the highest-density developed areas (i.e., those with the highest risk of flood damages). As a result, mapping follows development rather than precedes it, meaning that communities allow development in unmapped areas without taking flood risk into consideration. FIRMs also are based on historical data rather than future risk scenarios. In other words, FIRMs make the inaccurate assumption that “what happened yesterday will happen tomorrow,” rather than accounting for changes in future flood risk due to factors such as development, land subsidence, increased precipitation, or sea level rise [52]. 

[29] L. Mays, Stormwater Collection Systems Design Handbook, 1st ed., McGraw Hill, 2001.

[34] Atkins, “Flood Loss Avoidance Benefits of Green Infrastructure for Stormwater Management,” USEPA, Calverton, Md., 2015.

[35] R. M. Elliott, et al., “Green roof seasonal variation: comparison of the hydrologic behavior of a thick and a thin extensive system in New York City,” Environmental Research Letters, vol. 11, no. 7, p. 074020, 2016.

[36] S. Le Coustumer, et al., “The influence of design parameters on clogging of stormwater biofilters: A large-scale column study,” Water Research, vol. 46, no. 20, pp. 6743–6752, 2012.

[37] R. William, P. Gardoni and A. Stillwell, “Predicting rain garden performance under back-to-back rainfall conditions using stochastic life-cycle analysis,” Sustainable and Resilient Infrastructure, vol. 6, no. 3-4, pp. 143–155, 2021.

[38] L. Cheng and A. A. , "Nonstationary Precipitation Intensity-Duration-Frequency Curves for Infrastructure Design in a Changing Climate," Scientific Reports, vol. 4, p. 7093, 2014.

[39] Y. Kim, et al., “Fail-safe and safe-to-fail adaptation: decision-making for urban flooding under climate change,” Climate Change, vol. 145, p. 397–412, 2017.

[40] National Research Council, “Chapter 6: Implementing Flood Risk Management Strategies,” in Levees and the National Flood Insurance Program: Improving Policies and Practices., Washington, D.C., National Academies Press, 2013.

[41] D. Reeve, A. Chadwick and C. Fleming, Coastal Engineering: Processes, Theory and Design Practice, 3rd ed., CRC Press, 2018.

[42] National Oceanic and Atmospheric Administration, “Guidance for Considering the Use of Living Shorelines,” NOAA, Washington, D.C., 2015.

[43] USDA Soil Conservation Service, “Measures for Stabilizing Coastal Dunes,” USDA, Americus, Ga., 1992.

[44] C. Currin, et al., “Shoreline change in the New River Estuary, NC: rates and consequences,” Journal of Coastal Research, vol. 31, no. 5, pp. 1069–1077, 215.

[45] M. Spalding, et al., “Mangroves for coastal defence: Guidelines for coastal managers & policy makers,” Wetlands International and The Nature Conservancy, 2014.

[46] O. Wing, et al., “Inequitable patterns of US flood risk in the Anthropocene,” Nature Climate Change, vol. 12, p. 156–162, 2022.

[47] L. Bakkensen and L. Ma, “Sorting over flood risk and implications for policy reform,” Journal of Environmental Economics and Management, vol. 104, p. 102362, 2020.

[48] Federal Emergency Management Agency, “National Flood Insurance Program (NFIP) Floodplain Management Requirements,” FEMA, Washington, D.C., 2005.

[49] V. Lee and D. Wessel, “The Hutchins Center Explains: National Flood Insurance Program,” Oct. 10, 2017. [Online]. Available: https://www.brookings.edu/blog/up-front/2017/10/10/the-hutchins-center-explains-national-flood-insurance-program/. [Accessed June 6, 2022].

[50] Scata, J, “FEMA’s Outdated and Backward-Looking Flood Maps,” Oct. 12, 2017. [Online]. Available: https://www.nrdc.org/experts/joel-scata/femas-outdated-and-backward-looking-flood-maps. [Accessed June 6, 2022].

[51] American Society of Floodplain Managers, “Flood Mapping for the Nation: A Cost Analysis for Completing and Maintaining the Nation’s NFIP Flood Map Inventory,” ASFPM, Madison, Wis., 2020.

[52] C. Sarmiento and T. Miller, “Costs and Consequences of Flooding and Impacts of the National Flood Insurance Program,” Pacific Institute for Research and Evaluation, Beltsville, Md., 2006.