A review of climate change-induced flood impacts and adaptation of coastal infrastructure systems in the United States

Climate change-induced sea level rise, storm surge and extreme precipitation in coastal regions of the United States (US) are affecting coastal infrastructure systems, including transportation, defense, energy, buildings, water supply, wastewater, stormwater and shoreline infrastructure. The interdependencies among these systems further worsen the climate change risks affecting infrastructure reliability and resiliency. Evaluating the current state of scientific research focus on climate change-induced coastal flood risk and the adaptation of US coastal infrastructure systems helps in understanding the current progress in critical coastal infrastructure adaptation and guides future research in the necessary direction. In this review, we synthesize the scientific literature through a metadata analysis within the scope of US coastal infrastructure system risk due to climate change-induced recurrent flooding in seven key coastal infrastructure systems across different coastal regions, namely, New England, the Mid-Atlantic, the Southeast and Gulf, and the West Coast. Our review found that coastal stormwater and shoreline protection systems and transportation systems are the most studied, while water supply and defense systems are the least studied topics. Over the last decade of scientific contributions, there has been a distinct shift in focus from understanding and quantifying coastal flood risks towards adapting coastal infrastructure systems. The majority of the studies are based in the Mid-Atlantic, Southeast and Gulf, while national scale studies are very limited. Although critical to resilient coastal infrastructure systems, the consideration of interdependencies or studies expanding across multiple infrastructure systems are limited. Out of the forward-looking studies that consider future climate scenarios, 39% considered only long-term (year 2100) scenarios, while 27% considered all short-, medium- and long-term scenarios. Considering finite resources and finite infrastructure life span, the ultimate focus on the end of the century climate scenarios extending beyond most of the existing infrastructure’s design life is a challenge to adaptation planning.


Introduction
Coastal civil infrastructure systems provide important economic, health, safety, and social benefits.Climate change reveals the vulnerabilities of these infrastructure systems by altering the bio-physical and environmental variables influencing infrastructure design, construction and operation.Globally, around 40% of the population live within 100 km (62 miles) of the coast (UN 2017), 250 million people live below annual coastal flood levels, which could rise to 340 million by 2050 (Kulp and Strauss 2019), most of the global megacities are in the coastal zones (Neumann et al 2015a), and, in the United States (US), alone, coastal shoreline counties are home to 133.2 million people (42% of the US population) (Kildow et al 2016, Fleming et al 2018).Settlements and built infrastructure in the coastal zone are exposed to increased flood damages due to the combined effect of sea level rise (SLR), storm surge, which is often a devastating tail water from outlets with increasing saline water, groundwater infiltration, and salinization, yielding major corrosion (Allen et al 2019).Built structures like tide gates, seawalls and bridge ramps may suffer functional failure as SLR and storm-driven flooding becomes more frequent (Johnston et al 2014).Additionally, climate change and its impacts are a national security threat with strategic implications for the US Navy, Air Force, Coast Guard and other services that maintain a coastal presence.Climate adaptation is a key priority of the Department of Defense (DOD), and climate-related implications affect many of the DOD's activities and decisions related to future operating environments, military readiness, stationing, environmental compliance and stewardship, and infrastructure planning and maintenance (Garfin et al 2021).By 2008, more than 30 US military installations were already facing risks from rising sea levels (Ayyub and Kearney 2012).Coastal infrastructure systems are at the forefront of climate change-induced recurrent flooding, thus the effective adaptation and management of such systems is critical.
Infrastructure systems provide interconnected networks that deliver resources, remove waste, move people, information and goods (Jacob et al 2000).Complex interrelationships, dependencies and interdependencies exist among critical infrastructure systems or within their sub-components, often crossing infrastructure boundaries, either through direct connectivity, policies and procedures, or geospatial proximity (Pederson et al 2006).However, such interdependencies and their dynamics are not fully understood, and this shortfall could lead to incomplete understanding of infrastructure vulnerability.Critical infrastructure operation, serviceability and integrity are at risk, not only due to the increasing frequency of extreme events, but also due to the strong reliance of infrastructure systems upon each other, which propagate local disturbances in one system to cascading and catastrophic failures across systems and society as a whole (Bigger et al 2009).These linkages should be evaluated as dependencies in a causal chain relationship that could cause a cascading failure of first, second, or nth order dependency (Bigger et al 2009).The robustness of infrastructure systems depends on their design, state of maintenance, and the man-made, environmental and natural stresses to which they are exposed.Future management and governance of coastal infrastructure systems must rely upon the adequate understanding of climate implications (for example, SLR), system interdependencies and cascading impacts.
This review catalogs previous literature studies on US coastal infrastructure vulnerability to climate-induced coastal flooding and coastal infrastructure adaptation.We investigate how vulnerabilities in coastal infrastructure systems and adaptation strategies are evaluated in scientific findings, and we identify key knowledge gaps affecting infrastructure adaptation planning.As adaptation planning is required to maintain coastal infrastructure services in the future, we identify literature solutions to four key research questions surrounding successful coastal infrastructure adaptation: (1) How are infrastructure systems along coastal zones in the US affected by climate change-driven SLR and recurrent flooding?(2) What is the current status of coastal infrastructure system adaptation?(3) How are system interdependencies considered for risk identification and adaptation planning?(4) What knowledge gaps should be addressed in coastal infrastructure adaptation research?

Methodological overview
To answer these four key research questions, we review the existing work from multiple repositories by selecting and filtering articles based on the scope of this study.Figure 1 presents a methodological overview of this study.First, we define the coastal infrastructure systems, and create a database for the literature review.We conceptualize seven infrastructure systems in the coastal zone as coastal infrastructure systems, i.e. transportation, defense, energy, buildings, water supply, wastewater, and stormwater and shoreline systems, which are critical to economy, health, security, production and the distribution of goods and services.The literature search was done using a combination of keywords-climate change, SLR, storm surge, coastal flooding, inundation, coastal infrastructure, coastal infrastructure adaptation, defense and military, searched within Scopus, Google Scholar, Science Direct, and the American Society of Civil Engineers' online library.Keywords and phrases were combined using Boolean logic to find and download a large number of published works based on the title, keywords and abstract.These articles were then sorted based on relevancy to the scope of our review and subset based on three criteria: (1) peer-reviewed articles, (2) focus on coastal flood risk, resilience, and adaptation surrounding coastal infrastructure systems, and (3) US-based studies published from 2000-2022.A summary illustration of the literature search process is presented in a supplemental file (figure S1, section S2).Over 459 potentially relevant journal articles were initially downloaded, and 94 articles were then reviewed that met the scope of our study through a title and abstract screening.A standardized form was created and used to evaluate each literature study uniformly, creating metadata to synthesize these articles.The process, showing a summary of the targeted literature review, the list of reviewed articles, and a summary of the metadata, is presented in a supplemental file (sections S1-S4).(3) the types of infrastructure system being studied; (4) goals, the underlying primary focus of the collected literature, identified here based on common themes such as damage reduction, infrastructure adaptation and resilience, transportation disruption, economic impact; (5) a summary of the methods being applied; (6) climate scenarios being studied based on the representative concentration pathways (RCPs) or low, intermediate, and high designations; and (7) whether the study is solution-oriented with recommendation for adaptation planning.Using the collected articles in the metadata, we reviewed the climate change-induced flood impact on different coastal infrastructure systems, the existing focus of the literature on infrastructure adaptation strategies, the cascading impacts and interdependencies considered across infrastructure systems, and the key gaps in adaptation decision making.
Based on a systemic review and deductive reasoning, we synthesized the metadata to identify common themes and gaps, either methodological, information, intellectual or geographical, with the aim of understanding the current progress and identifying future research needs.

Impacts on coastal infrastructure systems
Since the primary climate change impact on coastal infrastructure is coastal flooding, due to SLR, storm surge, and extreme precipitation, the majority of the studies are centered around shoreline or flood protection and stormwater infrastructure systems, followed by a focus on transportation and buildings.The remaining studies evaluated wastewater, energy, and water supply systems as past hurricanes showed the vulnerabilities of such systems, while defense systems were given less focus in published journal articles (as shown in figure 2). Figure 2 also shows the common pairing or interaction between infrastructure systems in the literature.The nature of such interactions can be characterized into a few categories, including common failure modes-such as inundation of facilities, for example, inundation of critical infrastructure such as hospital buildings (here: grouped within buildings) and wastewater treatment plants (WWTPs); loss of service-such as the physical damage to stormwater and shoreline protection structures and energy supply/distribution systems; and reduced system efficiency-such as partial loss of function.Coastal flooding does not only have singular impacts on systems, but the dependencies and interdependencies across infrastructure systems can result in cascading impacts.The number of studies that investigate multiple infrastructure systems are in the minority, as only 15 papers out of 94 studied three or more infrastructure systems, and only 35 studied at least two systems.While this is not an insignificant sample size that evaluates multiple systems, the joint evaluation of systems is still in the minority.The limited number of studies that evaluate multiple infrastructure systems is due to the complexity of the interdependent infrastructure systems itself in relation to the level of detailed data.However, the evaluation of these interdependent systems merits further understanding and quantification due to cascading failures.
Recurrent flooding and climate change impacts coastal infrastructure systems based on their exposure and vulnerability.Recognizing that the exposure, vulnerability and risk vary across infrastructure systems, we summarize the climate impacts on each type of infrastructure system.Within each subsection, the cited articles are given as selective examples that are relevant to the discussion.Additionally, the discussion is aided by some articles outside the scope of the metadata analysis for further clarification.For an extensive list of selected articles within the metadata and summary of each article, readers are referred to the supplementary file (sections S3 and S4).

Stormwater and shoreline protection
The majority of the evaluated studies (40 articles) included stormwater and shoreline protection infrastructures (figure 2).SLR and storm surge have significant impacts on shoreline flood control and protection infrastructure such as dams and levees.Furthermore, stormwater systems, designed to safeguard people and other infrastructures in coastal urban watersheds are often strained in their ability to transport and store excess runoff due to the combined effects of extreme rainfall and high tides prolonging the flooding events (Johnston et al 2014, Sweet et al 2014, Gold et al 2022).Frequent exposure of such systems to extreme events physically damages the infrastructure, requiring costly and frequent maintenance over its design life.For example, storm surge flooding during Hurricane Katrina in Biloxi, Mississippi, clogged stormwater systems with debris, including sand and building materials, along Highway 90, and all of the 105 lift stations were inoperable immediately after the storm.Only 71% of the stations were able to be restored within a few days.In addition, during the same extreme event in Pascagoula, Mississippi, flooding of up to 4.88 m (16 ft) affected stormwater drainage joints, causing leaks and exfiltration of groundwater into the stormwater drains (Chisolm and Matthews 2012).Such risks are particularly common for buried infrastructure, such as storm and waste water collection conduits, as they degrade faster due to storm surge and flooding (Hao et al 2012, Matthews 2016), requiring higher investment in asset management.Stormwater systems with outfalls discharging into tidal waters are prone to tidal wave effects (such as during high or king tides), affecting designed hydraulic grade lines and causing a backwater effect, propagating effects of reduced storage and transport capacity of drainage systems further upstream.This propagation upstream can further deteriorate the system due to frequent inundation and exposure to saltwater, potential road closures, and disruption to critical services.
The analyzed articles that focus on stormwater and shoreline protection can be categorized into four general types of studies: the effect of local to multi-scale variation of SLR and tidal waves on coastal inundation analysis (for example, Johnston et al 2014, Wang et al 2017, Xie et al 2019); physical-based and coupled storm drainage and coastal inundation modeling (for example, Goudreau 2016, Joyce et al 2017, Shen et al 2019); an economic evaluation tool to support adaptation decision making (for example, Reguero et al 2018, Wong et al 2022); and an evaluation of infrastructure resilience (for example, Small et al 2016).While stormwater and shoreline protection is the most commonly studied system, model-based studies combining several coastal infrastructure systems and considering cascading climate change impacts are quite limited (for example, Frazier et al 2010), while studies that do consider several infrastructure systems are mostly conceptual framework papers or reviews without technical models or analyses (for example, Azevedo de Almeida and Mostafavi 2016, Singh et al 2021).

Transportation
Transportation systems are the second most studied (32 studies) coastal infrastructure system (figure 2).Transportation systems are mostly affected by storm surge, which inundate and wash out roads, and flood subways and airports.The analyzed articles combine coastal inundation analysis with road damage or traffic disruption using either hydrodynamic coastal flood simulation models (for example,  (Bellis et al 2019).With SLR, several sections of the road network are at risk of inundation.For example, along the coast of the Gulf of Mexico, with SLR of 1.2 m (3.94 ft), 27% of major roads in the region, 9% of rail lines, and 72% of ports are currently at risk (Chinowsky et al 2013).A study by Sadler et al (2017) combining SLR projections and with tidal data to determine the most critical major roadways in Norfolk and Virginia Beach, Virginia, found that under the intermediate SLR scenario, around 10% of major roads in Virginia Beach and Norfolk might regularly flood by 2100 due at tides reaching mean higher high water, which might increase to over 15% of major roads with a 99% tide and to over 65% of major roads with the addition of a 100-year storm surge.These flooding impacts primarily affect the transportation of freight and commuter vehicles, but they can have significant economic effects.
SLR will increase groundwater levels in coastal zones, which will interfere with the conditions of coastal road infrastructure, weakening the pavement structure and reducing its service life (Knott et al 2017).Hurricane-induced wind and storm surge can also cause temporary closure and even structural damage to bridges as hydrodynamic loads from waves and tidal currents cause wave scour.Approximately 60,000 road miles in the US are exposed to coastal waves and storm surge (Douglass and Webb 2020), and the Federal Highway Administration estimates that 36,000 bridges are within 24.14 km (15 nautical miles) of the US coastline (Snaiki et al 2020).For example, Hurricane Irene, in 2011, caused damage to over 200 roads and bridges in New England, while Hurricane Katrina, in 2005, caused damage to over 40 bridges in Louisiana and neighboring states (Snaiki et al 2020).Climate change is poised to directly and indirectly impact the transportation networks of the US with recurrent flooding and rising sea levels.

Buildings
Buildings are the third most common infrastructure system studied (29 studies) (figure 2).Extreme precipitation, SLR, and storm surge can wreak havoc on buildings in coastal cities.Evidence from Hurricane Sandy showed at least 650,000 homes were damaged or destroyed (Blake et al 2013).Hurricane Katrina showed significant structural damage to several building types due to storm surge, hydrodynamic uplift, and floating debris from wave action, which might require considering different loading conditions to avoid bending and shear effects for building design (Robertson et al 2007).Currently, a significant number of buildings are at substantial flood risk (defined as inundation >1 cm in 100-year flood), particularly in the coastal states of Louisiana, Florida, Georgia, Maine, Massachusetts, New York, Virginia, North Carolina and Texas (First Street Foundation 2020).In comparison to 2020, by 2050, the number of properties at substantial flood risk could increase in key coastal cities: New Orleans, LA (+207.6%),Jacksonville, FL (+65.6%),Tampa, FL (+22.4%),Cape Coral, FL (+20.5%),New York, NY (+20.2%),Houston, TX (+17.1%), and Los Angeles, CA (+3%) (First Street Foundation 2020).This increase in vulnerability particularly affects people living in low-lying affordable housing, which is often structurally vulnerable.The number of affordable units exposed to coastal flooding across the US could increase by more than three times by 2050 compared to 2000 (Buchanan et al 2020).Coastal flooding further compromises housing affordability as the loss of affordable housing increases housing cost in the adjacent communities to meet housing demands (Buchanan et al 2020).
The application of typical depth-damage relations or hydrodynamic models that consider the effects of wave action and flood velocity with infrastructure (parcel to building level) data to study coastal inundation is a common approach in the evaluated studies (for example, Strauss et al 2012, Hatzikyriakou and Lin 2017, Davlasheridze et al 2019, Li et al 2020).Such inundation-based impact analysis is often combined with other infrastructure systems that have geographic interdependency, such as those existing in close spatial proximity (for example, Small et al 2016, Allen et al 2019).Other evaluated studies include the inundation of other facilities such as schools, health care, transport and energy, highlighting the diversity of structures at risk from recurrent flooding and SLR in the US.Additionally, the studies estimate damage relative to inundation using the Federal Emergency Management Agency's building stock inventory in the nationally standardized risk modeling HAZUS tool (for example, Heberger et al 2011).

Wastewater
SLR and storm surge-induced coastal flooding also impact WWTPs and utilities, which were evaluated within 17 studies.Indeed, WWTPs are mostly located at the lowest elevation of the wastewater collection network (Tram et al 2014).Evidence from Hurricane Katrina in Covington, Louisiana showed around 10%-15% of WWTPs were rendered inoperable, requiring maintenance, replacement of sewer pipelines, or large cleanup efforts (Chisolm and Matthews 2012).Furthermore, in New Orleans, due to the lack of redundancy in wastewater collection and storage systems, the infrastructure was overloaded due to the high volume of water and untreated wastewater, which eventually flooded into the street.Similarly, during Hurricane Rita in Lake Charles, Louisiana, 100% of the WWTPs were without power (Matthews 2016).In total, 200 WWTPs in Louisiana, Mississippi, and Alabama, including three large WWTPs serving more than 150,000 customers, were affected during Hurricane Katrina and Rita in 2005 (Esworthy et al 2005).Vulnerability assessments of wastewater systems in Oahu, Hawaii showed sewer mains and onsite disposal systems are the most vulnerable components of the system, and, in the scenario 0.3 m (1 ft) of SLR, over 16.1 km (10 miles) of sewer mains, 475 onsite disposal systems are at risk (Spirandelli et al 2018).
A study by Hummel et al (2018) showed that, at a SLR of 0.3 m (1 ft), 60 WWTPs across the US will be affected, impacting 4.1 million residents, and, at a SLR of 1.83 m (6 ft), 394 WWTPs will be affected, impacting 31.6 million residents.Extreme precipitation, higher runoff, flooding conditions, inflow and infiltration further pose risks to sanitary sewer overflows, system degradation and water quality impairment (Spirandelli et al 2018).Based on our review, coastal climate change links with wastewater systems have been mostly studied together with other water systems (figure 2) (for example, Matthews 2016, Ridha et al 2022).

Energy
Only 13 peer-reviewed articles in the analysis focused on energy systems (figure 2).Energy infrastructure systems are at high risk since low lying coastal zones are home to oil refineries, coal and natural gas processing facilities, and a large portion of US energy production (nearly 4,000 oil and gas platforms) is located on the Gulf Coast (USGAO 2014).These systems are in addition to electric infrastructure such as power plants, substations, and transmission infrastructure in coastal communities.Such infrastructure is at risk from SLR, increasing the intensity of storms and higher storm surge and flooding, potentially disrupting oil and gas production, refining, and distribution by pipelines, rail or barge, as well as electricity generation and distribution (US DOE 2013).Past examples have shown that Hurricane Katrina caused damage to 100 platforms, 558 pipelines, and shut down several refineries, oil and gas production for several weeks (USGAO 2014).Bradbury et al (2015) studied the effects of storm surge and sea level rise on energy infrastructure by examining historical hurricane data along the Texas and Louisiana coast, and combined storm surge with SLR projections for 2100.Their results suggest that, between 1992 and 2060, the number of energy facilities exposed to a weak (category 1) storm might increase by 15%-67% under the NOAA's highest SLR scenario.Bradbury et al (2015) also suggested that during the same period, there could be a significant increase in flood exposure under category 1 and 3 storms to all energy infrastructure types, namely, oil refineries, strategic petroleum reserves, electric power substations, electric power plants, and the natural gas sector.
Frequent exposure to storm surge can directly impact electrical infrastructure and power systems, causing widespread outages.For example, Hurricane Sandy in 2012 caused significant flooding in New York City, which led to the failure of several electric power systems, eventually resulting in the evacuation of New York University's Langone Medical Center, the destruction and damage to houses in fire incidents due to the lack of electricity, shortage of fuel supply, business interruptions, and the breakdown of heating networks, security systems, telecommunication services, and emergency power generators.The impact of these outages showcase the cascading failures and interconnected nature of coastal infrastructure systems, which are not always captured in the modeling literature.Moreover, the risk of physical damage from more intense and frequent storm events, lightning strikes, wildfire damage, altered vegetation growth, wood pole decay and water damage could increase (US DOE 2013, Fant et al 2020).This additional exposure will significantly increase (as much as 25%) the total electricity infrastructure cost throughout the contiguous US due to the reduction in the lifespan of substation transformers and an increase in vegetation management expenditures (Fant et al 2020).Only a few studies acknowledge the significance of electricity systems, including sub-stations, distribution, transmission, and other electrical components and their failure risk (figure 3

Water supply
Climate change affects coastal water supply infrastructure due to the inundation of the water supply infrastructure, such as treatment plants and pumping stations, and saltwater intrusion due to SLR and associated chemical pollution (Patterson et al 2007, Kolb et al 2017).Hurricane Katrina destroyed local water supplies in Mississippi and Louisiana due to the loss of power, loss of pressure, and damage to water treatment and distribution systems, affecting millions of people (Patterson et al 2007).Additionally, land subsidence, soil swelling, and uprooted trees incrementally damaged underground water distribution pipelines.The storm also damaged a few fire hydrants, requiring several days to re-pressurize the water network and make the water potable, while saltwater corroded mechanical systems such as pumps and hydrants (Chisolm and Matthews 2012).The US Environmental Protection Agency estimated over 1,220 small drinking water infrastructure systems were affected during Hurricane Katrina, disrupting services to millions of people, and a restoration cost of $2.25 billion (Esworthy et al 2005).Such impacts are likely in the Gulf Coast region, requiring a large restoration and cleanup process.
Saltwater intrusion into ground water aquifers, which is a key source of drinking water in several coastal cities, is also exacerbated by SLR and storm surge.Over 15% of the West Coast (central and Southern California) groundwater well elevation, nearly 22.6% of Gulf Coast well levels (Texas, Louisiana, Mississippi, and Florida), and 34.7% of the East Coast (Miami-Florida, Georgia, South Carolina, Virginia) water wells are below sea level (Jasechko et al 2020).This vulnerability could elevate the concentration of chemicals such as bromide, leading to the formation of trihalomethane, requiring the treatment process to be adapted to meet regulatory threshold levels (Kolb et al 2017).As a result, a more energy intensive water treatment process is required.Coastal water supply systems are mostly studied in combination with wastewater and stormwater systems, as mentioned above in section 3.4.Unlike WWTPs, water supply systems are not necessarily located at the lowest elevation.Additionally, challenges in water security due to climate-induced and anthropogenic drought often originate at the headwater or upstream watershed, unless coastal groundwater or desalination is the major water supply source, partially explaining the limited number of studies in the review.

Defense
Climate change also affects defense infrastructure systems, including the Air Force and other military bases and their critical facilities, due to the proximity along coastal areas, mostly through SLR and storm surge.Studies evaluating defense infrastructure systems considered coastal inundation impacts at the installation scale using hydrodynamic models (Li et al 2013); surveys to identify gaps in action plans against SLR for the Coast Guard (Lassiter and Shealy 2017); and multi-criteria risk assessment framework for the Lower Chesapeake Bay (Ratcliff and Smith 2011).Two recent hurricanes have impacted military installations, Hurricane Florence and Hurricane Michael, both in 2018.Hurricane Florence caused 2.74-3.96m (9-13 ft) of surge and precipitation of 508-762 mm (20-30 in) over several days, flooding three Marine Corps installations with an estimated $2.6 billion in damage.Hurricane Michael destroyed Tyndall Air Force Base in Florida with over $4.7 billion in damage only a few weeks later.Several military facilities and the world's largest naval facility are located in the Hampton Roads area in southern Virginia, which is under consistent threat of coastal flooding as the average sea level in the last 100 years has risen by 0.46 m (1.5 ft), and it is expected to rise by 0.3-0.9m (1-3 ft) by 2050 (U.S.Department of Defense 2021).In terms of regional frequency at Sewells Point, VA, there were 14 d with high tides (0.53 m or 1.74 ft higher than daily average high tide) in 2019, which per NOAA is expected to increase to Based on a Department of Defense report, 20 Air Force installations and several Navy and Army installations are currently being affected by recurrent flooding, and this number will only grow in coming decades (U.S.Department of Defense 2019).A recent study showcased how Eglin Air Force Base in Florida has an increasing susceptibility to damage from storm surge with intensifying hurricanes  (Baldwin et al 2023).This exposure particularly challenges the US DOD's activities, operating environments, military readiness, stationing, environmental compliance, stewardship and infrastructure planning and maintenance (Garfin et al 2021).Changing environmental conditions through flooding, erosion by waves and high tides and frequent extreme high temperatures also affect personnel lost training time, increased cost of facility maintenance and military readiness (Garfin et al 2021).

Research focus Definition of research focus
Infrastructure adaptation and resilience Coastal flooding impacts on infrastructure systems' function, vulnerability, resilience and cascading impacts, including qualitative and quantitative frameworks, with the aim of addressing adaptation needs and strategies.

Damage reduction
Managing infrastructure and mitigating risks through assessments, frameworks and applications supporting the way forward to coastal adaptation.

Flood analysis
Flood mechanism, propagation and quantification framework and approach including methodological advances, tools and applications.

Economic impact
Quantification through the economic lens of coastal flood damage from past to future climate change effects, cost of adaptation, including conceptual, analytical and numerical approaches, tools and applications.

Transportation disruption
Quantifying the disruption of transport systems from nuisance flooding to damage of physical infrastructure systems from coastal and groundwater flooding, including tools and applications.This special focus is identified as one of the most recurrent themes in the literature and one of the critical infrastructure systems affected due to coastal flooding.

Focus and scales of research
Beyond understanding which types of infrastructure are studied, it is important to evaluate the context of the evaluation.This section evaluates three facets of the existing research on coastal infrastructure threats to recurrent flooding.(1) First, we broadly categorize the literature's focus by summarizing the underlying primary focus or motivations of the studies, which are identified as: (i) infrastructure adaptation and resilience, (ii) damage reduction, (iii) flood analysis, (iv) economic impact, and (v) transportation disruption.
(2) Second, we discuss the spatial scale of the study, which is important in understanding the scope of the analyses and the relevance in local, regional, or national policy.
(3) Finally, we evaluate the temporal scale of the study, which provides context on the time horizon the study evaluates (i.e.short-, medium-or long-term timeframe).For an extensive list of articles within each focus category, readers are referred to the supplementary file (section S4).

Focus of research
We identify five key focus areas of research, presented in figure 3 and defined in table 1.The evolving pattern of the literature in coastal flooding (figure 3(a)) showed that during the earlier period of the decades (prior to 2013), the majority of studies focused on understanding climate change-induced coastal flood risk through quantifying flooding; sensing and communicating flood risks; modeling flooding scenarios under different SLR projections; and evaluating the economic cost of flood damage.These studies demonstrated a sense of urgency and set the direction for future work on infrastructure resilience.More recent studies have considered infrastructure adaptation and resilience, while advancing flood and impact analysis methods, tools and models (figure 3(a)).Transportation disruption has been an emerging key focus area and a significant issue, as an inundation of roads, physical damage to roads and bridges, and traffic delays directly affect a significant number of people (figure 3(a)).There is also noticeable combined focus on infrastructure adaptation and resilience with economic impact and flood analysis to evaluate the infrastructure's fragility, risk, and resilience from coastal flooding (figure 3(b)).

Spatial scale of research
While localized studies present concise conclusions that can provide actionable risk or adaptation portfolios, studies at a larger spatial scale, regionally or nationally, offer insight into patterns for policy and funding.Over half (53%) of the studies were at the city scale, which is understandable as cities are key consumers of resources, contributors to climate change, and have a concentration of infrastructure.Moreover, human and infrastructure interaction and the impact of flooding occurs at a localized scale.For example, a study by Singh et al (2021) compares flood resilience institutions, strategies and outcomes in New York City, Tokyo and Rotterdam, evaluating best management practices.Around 14% of the studies are at the national scale.Such studies lack fine resolution data, but they contribute to laying the framework for evaluating coastal flood risk and developing actionable flood policies to support stakeholders.Decision-makers benefit from both or a mix of small and large spatial-scale studies, as large spatial-scale studies provide more diverse knowledge and spatial trends, while small-scale studies provide localized examples and evidence for adaptation.Regional-scale studies, combining multiple cities or coastal zones, such as the Gulf of Mexico, are more precise than national-scale studies, provide a higher resolution of data and outcomes to stakeholders, and cover linkages between multiple coastal communities in the region in the face of potential large-scale impacts of SLR and storm surge.For example, a study by Dismukes and Narra (2018) showed that particular areas along the Gulf Coast, coastal areas in Louisiana and the Mississippi River delta plain, are at significant risk from SLR; and natural gas processing is particularly at a higher risk than other sub-components.
Of the coastal infrastructures being studied, city-scale or localized studies focus predominantly on buildings (23), transportation (20), and stormwater and shoreline protection (19) systems.Regional studies focus more on stormwater and shoreline protection (21) and transportation ( 12) systems.Given the nature of the distribution of infrastructure systems, buildings are isolated individual elements relevant for local-scale studies; water and transport systems are moderately connected across multiple cities and service areas; while energy systems have a very wide network and interconnection across and beyond coastal zones.Thus, larger spatial scale studies tend to focus on geographically widespread and connected systems like energy and shoreline protection infrastructures.
Figures 4 and 5 summarize the distribution of the literature in metadata focused on geographic regions and exclude studies that focus on coastal infrastructure, without being specific to any geographic region.Figure 4(a) shows the spatial distribution of the total number of studies across four coastal regions (New England, the Mid-Atlantic, the Southeast and Gulf, and the West Coast) and on a national scale.The Mid-Atlantic and the Southeast and Gulf regions have the largest prevalence of studies with 24 and 20, respectively.The number of studies covering the national scale is lower, and there are regional variations in the type of study accomplished (figures 4(b) and 5(a)).In New England, a higher number of studies focused on stormwater and shoreline infrastructure followed by transportation systems, buildings, and energy.In the Mid-Atlantic region, a greater number of studies focused on buildings, transportation systems, and stormwater and shoreline infrastructure.Water supply and wastewater system evaluations are mainly focused along the Southeast and Gulf and Mid-Atlantic regions.It was also found that a larger number of studies occur in cities with a higher population density where the exposure and risks are also higher, for example, New York City in the Mid-Atlantic (10 in total) and San Francisco on the West Coast (9 in total), as could be expected.However, this does mean that lower density communities are underrepresented in the literature.

Temporal scale of research
One of the most important considerations in climate change research is the temporal scale of the analysis, or how far out the study projects (figure 5(b)).We define these temporal scales into short-term (2021-2040), medium-term (2041-2060), and long-term (2081-2100) studies.Overall, out of 94 studies in the metadata, only ∼58% of the studies are forward-looking, with the remaining focusing on assessment frameworks, historic events utilizing past data, or studies considering hypothetical SLR numbers, not attributed to a timescale.Of the forward-looking studies, 77% considered long-term or end of the century (year 2100) scenarios, only 27% considered all short-, medium-and long-term scenarios, while 39% considered only long-term scenarios, without considering short-or medium-term climate scenarios.Even in those studies that consider all temporal climate scales, their ultimate focus is on 2100.Figure 5(b) shows the distribution of how the studies within five research foci considered short-, medium-and long-term climate scenarios.The majority of the studies tend to consider long-term climate scenarios.While not shown in this figure, the majority of the studies (26 out of 42) focusing on infrastructure adaptation and resilience did not consider the future time horizon as those include conceptual papers or frameworks with limited evidence-based applications.
However, this raises a key question regarding the rationality of the majority of the studies focusing on the long-term future time frame, as many infrastructure systems are built for a specific lifespan.For example, roads, railways, and energy transmission are built for at least a 50 year life-cycle.However, many bridges, water supply systems, wastewater systems and treatment plants have life-cycles of 100 years or more (Gibson 2017).Given that the current age of many infrastructure systems is approaching end-of-life, adaptation decisions on upgrades and replacements need to match the life-cycle of the system (ASCE 2021, Doyle 2021).
This means that much of our existing infrastructure will reach its end-of-life by the middle of the century, while new systems should be constructed with a time horizon of the end of the century and, potentially, beyond.Coastal infrastructure systems require more maintenance and upkeep, and adaptation costs for such systems are costly; therefore, appropriate adaptation pathways should be developed to meet both short-term goals while acknowledging the larger uncertainty in long-term projections.Allowing for dynamic adaptation pathways to support continued adaptation in an economical and low-risk portfolio can support these incremental adaptations as certainties become clearer (Haasnoot et al 2013, Lawrence and Haasnoot 2017, Lawrence et al 2018, Ramm et al 2018, Lawrence et al 2019).The analysis revealed that a significant number of studies were not forward-looking or did not have specific time horizons in their projections, highlighting the need to match timescales and projections to policy timelines and infrastructure lifespans.

Prioritizing adaptation research to advance implementation
The adaptation of coastal infrastructure systems is critical to improve the resilience of these systems and reduce vulnerability to catastrophic damage from climate change.The global annual cost of SLR in 2100, assuming optimal adaptation, is expected to be $180-520 billion between moderate emission scenario or RCP 4.5 to worst case scenario in RCP 8.5 (Depsky et al 2022).Without any adaptation, the expected global annual damage could increase by a factor of 150 between 2010 and 2080 (Tiggeloven et al 2020).Adaptation is expected to save over $2.5 trillion dollars per year in coastal property damage across the contiguous US by the end of the century (U.S. EPA 2017, Fleming et al 2018).Coastal infrastructure adaptation strategies, often categorized as protect (for example, low lying areas from inundation), accommodate (for example, allow rising water but minimizing damage), and retreat (for example, moving assets out of harm's way) (Diaz 2016), are often expected to result in a strong return on investment by minimizing the cost of climate change induced damage and avoiding catastrophic effects to society.
There are different forms of adaptation: anticipatory versus reactive, autonomous versus planned, and incremental versus transformational; with generally five key stages: awareness, assessment, planning, implementation, monitoring and evaluation (Po ¨rtner et al 2022).In practice, most of the coastal adaptation efforts in the US are in the planning phase and common adaptation efforts and strategies across coastal infrastructure systems include: the fortification of existing infrastructure or flood protection barriers to maintain their functions using hard structures or nature-based solutions (CDEEP 2011, Gregg et al 2011, Hill 2015); identifying and upgrading alternate options such as alternate routes and sources of transportation (CDEEP 2011, Johnson 2012); increasing redundancy in design to maintain service delivery in network infrastructures, such as power and water (DEP 2009, CDEEP 2011, USGAO 2014, Gregg and Braddock 2021); and the adoption of advanced metering, smart grids and technology to increase energy efficiency (Ayyub and Kearney 2012).Despite the benefits of adaptation, studies focused on adaptation measures with practical relevance are limited to the conceptual level and only 25% of the studies evaluated in this meta-analysis included adaptation strategies in their model.While most of the articles do mention adaptation strategies, many do so without modeling a specific solution or making recommendations for further research.Studies that focus on adaptation research are based on small-(city) scale studies.Adaptation against climate change impacts is critical for all coastal infrastructure systems; however, our review presents an unequal distribution of the current literature focus (figure 2), and an unequal distribution across geographic regions (figure 4).The current literature has also evolved from understanding climate risk to quantifying risk and infrastructure adaptation and resilience (figure 3).More evidence-based and use-inspired research is necessary to support practical implementation efforts.Although expanding the scope of adaptation is daunting, it is an essential step to advancing coastal infrastructure resilience.
Implementing adaptation strategies could involve significant investments, as infrastructure adaptation costs are expected to be 60%-80% of total global climate change adaptation, which could average $150-$450 billion per year in 2050 (Woetzel et al 2020).It is worth noting that there are inherent challenges to economic analyses associated with climate change adaptation due to the complex interaction with human systems and stakeholder usage.A complete review of the economics of climate change adaptation is outside the scope of the current effort.Advances in computing, the availability of high-resolution data, and access to these resources significantly contributed to increasing the pool of knowledge on coastal climate change impacts and flood assessment.For example, Olympia, Washington has invested in high resolution LiDAR elevation data and used LiDAR maps to run flooding simulations during high-tide SLR scenarios to assess the impact SLR could have on downtown Olympia (Gregg et al 2011).Such assessments are critical to better understanding infrastructure fragility and criticality.Future research should leverage improved datasets and tools to evaluate adaptation options and benefit-cost analysis of adaptation strategies (Tiggeloven et al 2020) at the scale appropriate for decision making and to support evidence-based practices.

Frameworks to support infrastructure adaptation
Implementing adaptation strategies requires an accurate understanding of infrastructure systems and its components' vulnerability, fragility, influence of external environmental factors and effectiveness of adaptation measures, which helps guide the adaptation strategies and process.A key challenge is the inability to evaluate uncertainties in the future climate and incorporate them into the infrastructure planning process.Uncertainty in the estimation of future SLR could lead to uncertain distribution of coastal risks, and inaccuracies in adaptation planning and cost (Wong et al 2022).This has global significance as flood damage cost, cost of no adaptation, and adaptation investment cost differs significantly at the warming of 1.5 • C-2 • C or higher, and under such uncertainty, adaptation decisions need to be made (Jevrejeva et al 2018).The key requirement for effective adaptation planning is the uncertainty quantification of future risks in adapting the decision, design and operation of infrastructure for future conditions within infrastructure lifetime.A safety margin strategy in infrastructure design, which is preparing systems to cope with worst-case climate change scenarios, might be inexpensive during the construction phase but it is expensive and difficult to modify existing systems (Hallegatte 2009).For example, to increase the resilience of the nuclear power plant in Turkey Point, Florida, the Florida Power and Light, in 2009, while planning new nuclear reactor units for the next 100-year SLR, considered 20% more conservative estimates of SLR (compared to units built in 1960s) by extrapolating historical data and adding safety margins to account for the limited accuracy, quantity and period upon which the historical data were based (USGAO 2014).Adapting such critical infrastructure incurs significant cost and resources (Kopytko and Perkins 2011).
Although the safety margin approach is one of the basic engineering design principles to mitigate infrastructure failure risks under uncertainties, it is not an optimal design approach as it is difficult to predict appropriate safety factors due to the deep uncertainty and non-stationarity of climate change.Additionally, it is not always appropriate for every coastal infrastructure system as the cost of over and under design can be extremely high.Coastal infrastructure adaptation involves increasing the capacity of a system, increasing redundancy and interconnections among networked infrastructure, hardening structures and flood proofing facilities through flood barriers, or implementing risk averse design standards.Due to the significant presence of infrastructure assets in the coastal zone, both the cost of over-design and the cost incurred due to under-adaptation will be high.Instead, to avoid the risk of over-design and under-adaptation, a different coastal adaptation and planning approach is necessary.Flexible and adaptive infrastructure planning, which incorporates uncertainty into the planning process, balances design cost and risk of failure, as infrastructure systems are incrementally upgraded as needs arise (De Neufville andScholtes 2011, Spiller et al 2015).Such a planning approach relies upon more accurate new information over time as climate projections reduce in uncertainty over time.In adaptation studies, it is critical to include all time scales from short-to long-term, not just the long-term climate projection scenarios as discussed in section 4.3, to avoid a mismatch between existing infrastructure design lifespans and planning horizons.

Need for integrated system analysis
Infrastructure system interdependencies, the dynamic evolution of the built environment, and its interaction with the natural environment add further complexities to the adaptation process.Our review has shown that there are a limited number of studies that consider multiple infrastructure systems (figure 2), and current scientific works do not fully evaluate an integrated coastal adaptation approach.Often, research in adaptation planning does not consider how future coastal development could evolve and influence future risk as well as the design of adaptation strategies.Indeed, adaptation scenarios should consider the dynamic changes in the system, local characteristics, spatial variations, and changes in policies (Reguero et al 2018) as well as human responses.In interdependent complex systems, impacts on one coastal system can cascade to other systems, and adaptation to one system could be of benefit or detriment to other systems or sub-systems.For example, tidal hydrodynamics could come into play while planning the adaptation.As shown by Wang et al (2018), the measures to protect a shoreline may increase inundation elsewhere, and these interactions and feedbacks should be carefully evaluated, especially in larger coastal cities.Therefore, coordinated efforts from different entities are necessary to plan well-informed adaptation strategies.
Finally, while considering an integrated system approach for coastal infrastructure adaptation and resilience, the nature of interdependencies and the level of complexities that decision makers should consider depend on several factors.For example, adapting water distribution systems should consider some level of redundancy to ensure water supply to other critical facilities or buildings within the coverage area.Similarly, adapting wastewater utilities should consider an emergency power supply to ensure uninterrupted functioning of pumping stations or increased capacity of treatment plants in case of sewer overflows coupled with surface flooding.However, the key limitations of integrated studies are complex when fully quantifying such system-wide feedback, challenges with data requirements, availability and accessibility, or the computational resources and efforts required for analytical and numerical methods.Therefore, this warrants future investigation on overcoming challenges, frameworks and adaptation practices within an integrated coastal infrastructure system.

Conclusions
This study synthesizes the peer-reviewed literature on coastal flood risk and its impact on coastal infrastructure systems in the US.These impacts were categorized to identify the current status of coastal infrastructure adaptation, any considerations of system interdependencies, and knowledge gaps that exist.SLR and storm surge-induced coastal flooding are increasingly impacting critical coastal infrastructure systems across all coastal regions in the US.A higher number of studies covering stormwater and shoreline protection, and transportation systems imply a high degree of concern surrounding those systems, while energy systems are less studied, despite their centrality in network dependencies.Understanding the pattern of focus in existing literature suggests that key focus areas have emerged in the literature, ranging from coastal flood risk assessment, risk communication and economic evaluation of flood damage cost, to adaptation-focused studies addressing coastal infrastructure adaptation needs and strategies to build infrastructure resilience under climate change.We believe this is partly due to the advancement in computational power and access to resources for hydrological, hydrodynamical, and integrated modeling studies, which have enabled better ways to analyze and communicate flood risk, infrastructure resilience and the urgency of adaptation actions.
Systemic understanding of interdependencies within coastal infrastructure systems is a critical element of successful coastal adaptation as climate change impacts propagate across systems and sub-systems.As is evident by this review, there are limited studies covering multi-infrastructure systems, interdependencies and cascading impacts, while every adaptation decision warrants coordinated and integrated action plans with interdependencies under consideration.It was also found that studies that aim to formulate assessment frameworks or decision-making frameworks often do not explicitly consider climate projections in their studies.A significant number (∼42%) of the total studies did not tie their projections to a specific time horizon, yielding unclear pathways on adapting current infrastructure systems to oncoming future climate uncertainties.Adapting existing and new infrastructure to future climate change requires reasonably reliable forecasts and scenarios relevant for infrastructure design across the planning horizon.

Figure 1 .
Figure 1.Review methodological overview.Note: Dotted lines represent core collective components of this review process which forms the metadata.

Figure 2 .
Figure 2. Common themes surrounding infrastructure impact and adaptation focus across the coastal infrastructure systems in the existing literature.The figure was made using the 'UpSetR' R package (Conway et al 2017), which is an alternative to a Venn diagram to visualize the intersection between multiple sets.The left (horizontal) bar plot represents the total size of each set, the bottom plot represents every possible interaction (multiple joint or individual) and the top (vertical) bar plot represents their occurrence.
Frazier et al 2010, Papakonstantinou et al 2019, Shen et al 2022), GIS-based inundation analysis (for example, Johnston et al 2014), or reviews and conceptual papers toward adaptation frameworks and strategies (for example, Rosenzweig et al 2011, Azevedo de Almeida and Mostafavi 2016, Singh et al 2021).The road network is an important national asset, which required an annual maintenance of $134 billion in 2013 (Chinowsky et al 2013) and $169 billion in 2019 ) (for example, Rosenzweig et al 2011, Lane et al 2013, Johnston et al 2014, Azevedo de Almeida and Mostafavi 2016, Matthews 2016), neglecting an important component in cascading failures, since electrical systems are broadly connected to other infrastructures.

Figure 3 .
Figure 3. (a) Histograms showing the evolving focus of peer-reviewed and scientific articles and (b) UpSet plot showing studies focus.

Figure 4 .
Figure 4.The current literature focuses on different infrastructure systems showing (a) the total number of studies across four coastal regions of the contiguous US, (b) the distribution of infrastructure focus by coastal regions.Note: this includes only the studies specific to geographic regions.

Figure 5 .
Figure 5. Current literature focus (a) distributed by coastal regions and (b) on temporal scale showing the number of studies considering short (2021-2040), medium (2041-2060) and long (2081-2100) terms.Note that some studies cover more than one future time horizon, which are then counted as duplicates in the bar plots.

Table 1 .
Definition of research focus.