A review of emerging strategies for incorporating climate change considerations into infrastructure planning, design, and decision making

1

ABSTRACT

Climate change is affecting infrastructure in complex and uncertain ways. Traditional load factors, safety factors, and design standards appear misaligned with current and anticipated future conditions. Thus, adapting infrastructure for a changing climate will likely necessitate balancing trade-offs between new and old design paradigms. This literature review summarizes advances in the implementation and research of resilient infrastructure within the context of climate change. We identified three categories of adaptation strategies: (1) assessments and frameworks to incorporate climate data and risks into infrastructure design and planning, (2) modelling of decision making under uncertainty and policy analysis, and (3) examples of best practices, case studies, and workarounds to enhance resilience. This work highlights advances in infrastructure decision making under uncertainty and ways to instill resilience into infrastructure systems. It is expected to help form a knowledge basis for transitioning to infrastructure planning, design, and implementation that is congruous with a changing world.

KEYWORDS:

• Infrastructure
• climate change adaptation
• decision making
• resilience assessment

1. Introduction

A changing climate alters the conditions under which our infrastructure systems must be designed and operated. Traditional infrastructure decisions are often based on the most probable conditions and historical data, e.g., the return period of design storms. However, nonstationarity (the concept that past conditions and trends may no longer hold) and uncertainty attributable to climate change are increasingly posing challenges for traditional design and implementation approaches (Markolf et al., 2021). An example of disruption in traditional infrastructure planning can be seen in storm water management. Return periods (e.g., 100-year storm) and load factors commonly used by hydraulic engineers to calculate infrastructure designs appear to be increasingly misaligned with current (and projected) extreme precipitation and storm events (Wright et al., 2019). The decision of engineers to go beyond the ‘by-the-book’ numbers (i.e., 100-year storm event) when re-constructing New Orleans’ flood protection infrastructure after the 2005 Hurricane Katrina further underscores limitations in traditional design approaches (Time, 2021). Partially in response to the availability of additional funding, engineers in New Orleans openly discussed deviating from standards and decided to be overly conservative when anticipating future hazards. This decision has partially helped mitigate disruptions and impacts from subsequent storms. However, using higher margins of safety to account for climate uncertainty can lead to overdesign (Underwood et al., 2020), which may require vast material and financial resources not available for every project. The challenges of accounting for climate change while also avoiding over/under-design calls for alternatives to our standards and conventional design approaches. The issue does not only occur in hydraulic infrastructure, but also structural engineering, power, transportation, and urban infrastructure. Loads like snow, wind, and temperature are important considerations when building highways, power lines, urban parks, and buildings (Al-Rubaye et al., 2022; Brockway & Dunn, 2020; Markolf et al., 2021). These loads manifest in the form of floods, drought, storms, or heat extremes, which are increasingly being influenced and altered by climate change (IPCC, 2021, 2022). Whether using threshold values, return periods, or something else, there appears to be room for advancement with regard to accounting for climate change and its related uncertainties (e.g., timing, location, intensity, and duration of an extreme event).

Given these challenges, the integration of resilience thinking with infrastructure design has emerged as a potential strategy for developing infrastructure systems that are better suited for increasing uncertainty and a changing climate. Resilience has been described in several ways, such as the ‘four Rs’ (Robustness, Redundancy, Rapidity and Resourcefulness) (Bruneau et al., 2003), which primarily reflect capabilities of a system to absorb and recover from disturbances (Minsker et al., 2015). In the context of engineering, the concept of resilience has started to evolve into the four concepts of rebound, robustness, graceful extensibility, and sustained adaptability (Woods, 2015). Under this perspective, resilience is conceptualized as a dynamic and iterative process rather than a static endpoint. In other words, emphasis is placed on a system’s ability to move between robustness and graceful extensibility (as an example), rather than aspiring to reach a desired level of robustness or extensibility. Regardless of one’s conceptualization of resilience, efforts to advance the incorporation of climate data/projections into the infrastructure planning and design process are critical for ensuring that infrastructure systems continue to provide the critical services upon which society has come to rely. However, these efforts can be mired by a murkiness with respect to when, where, and how they can (or should) occur. By synthesizing current approaches, gaps, and opportunities, this review aims to reduce that ambiguity and help delineate some potential pathways forward for researchers and practitioners.

Reviews on the definition of infrastructure resilience (Meerow et al., 2016), extreme events (McPhillips et al., 2018) and advances in certain infrastructure topics (Monteiro et al., 2020; Oikonomou et al., 2021; Twumasi-Boakye & Sobanjo, 2019) have been conducted. However, to the best of our knowledge, there are not any summarizing publications on the current status of and advances in planning and design for a changing climate. This review seeks to identify gaps and opportunities for the use of climate data/projections and alternative infrastructure performance concepts in the planning, design, and implementation of infrastructure systems. Additionally, this review explores strategies employed by practitioners and researchers to implement resilience within infrastructure systems as a response to climate change. Examples are found in multiple sectors and stages of infrastructure planning, design, and decision making. Overall, we aim to provide a holistic overview of the state of research and practice in order to advance the transition toward climate-adapted infrastructure systems.

2. Methods

We conducted a literature review using journal articles indexed in the Web of Science (WOS) Clarivate Database. The following query was used to retrieve relevant articles: TS = (Infrastructure AND Climate Change) AND ALL = (Decision AND Resilience). This query means articles had to be within the topic of Infrastructure and Climate Change to be within the scope of this review and include the words decision and resilience in the article text or abstract.

The keyword ‘decision’ is used to select research related to the planning process instead of ‘design’, which returns only a few results. The keyword ‘resilience’ represents characteristics of infrastructure that can cope with a changing climate (Chester et al., 2020). To reduce the potential for an overly narrow search, this keyword was substituted with the alternatives ‘vulnerability’ and ‘adaptation’ and search results compared. Adaptive capacity, vulnerability and resilience are linked concepts (Cutter et al., 2008). A summary of this analysis can be found in Figure 1. Results indicate that the selected keyword for infrastructure resilience is representative of this cluster of keywords (69% of articles include either one or both other keywords). This ensures that a query for resilience yields relevant articles that overlap with, and do not exclude, similarly used vocabulary.

Figure 1. Results of the web of science (WOS) search of keywords often used in combination to resilience showing the overlap of resulting articles.

Additionally, we conducted a similar search on the American Society of Civil Engineers (ASCE) library. The ASCE library articles and conference proceedings are also indexed in the WOS. However, resources aimed more at practitioners and educators, such as ‘Manuals of Practice’ (MOP) and textbook collections, are not part of the WOS results. Therefore, a dedicated search of the ASCE library appeared warranted. The ASCE search is a flat keyword search, where results can be selected by topic and document type category. The same keywords ‘climate change’, ‘infrastructure’, ‘resilience’, and ‘decision’ were used.

Results of these publication database searches were then screened according to several criteria such as relevance, quality, appropriate referencing, and use of plausible methods (see, Figure 2). The articles were selected based on their mention of strategies and policies concerning a direct response of the infrastructure sector to climate change. Publications focusing on financial modelling, community-based participatory research, developmental work, or the peripheries of infrastructure (such as food and agricultural systems or institutional infrastructure) were not selected. From the ASCE library, 1328 results were returned. The titles of the ‘state of the art reviews’ (12), technical papers (574) and case studies (95) were read. Relevant publications that complemented topics only briefly touched by the WOS search results (e.g., Envision, review of design approaches used in practice) were added to the analysis. The selected articles included reviews (12), conference proceedings (2), journal articles (1) and books (5), bringing the total number of reviewed publications in this study to 69. Two of the books (Ayyub, 2018; Ayyub et al., 2021) are from the Manuals of Practice (MOPs) series, which are compilations of publications that give recommendations on how to adapt infrastructure in practice.

Figure 2. Workflow of selecting the publications included in this study.

To identify infrastructure areas with high research activity, publications were grouped by the type of infrastructure mentioned in the article. This is not an exhaustive list of critical infrastructure sectors, but merely a reflection of the sectors represented in the reviewed publications. Similarly, the outcomes of the research were grouped by comparable articles, e.g., policy analysis and recommendations versus assessment frameworks.

3. Results

The number of publications associated with the topic of this study has increased exponentially over the past years (Figure 3). For the 69 selected articles, emphasis appeared to be placed on stormwater, urban systems, transportation, and green infrastructure. These are often interconnected, e.g., connections in water-energy-food systems or water-transportation systems. The outputs of the selected articles include policy analysis and recommendations, case studies, and resilience assessment frameworks for decision making (see, Table 1). As a result of this review, we identified three categories of outcomes: (1) assessments and frameworks to incorporate climate data and risks into infrastructure design and planning, (2) modelling of decision making under uncertainty and policy analysis, and (3) examples of best practices, case studies, and workarounds to enhance resilience. Following is a summary of key insights for each of these three categories. The results span multiple infrastructure sectors and underscore the interconnectedness of modern infrastructure (Markolf et al., 2018; Rinaldi et al., 2001).

Table 1. Infrastructure sector and deliverables of selected studies.

	Grouped Articles
Area of Infrastructure Research	Stormwater (15), Urban Systems (10), Transportation (7), Green Infrastructure (7), Coastal Infrastructure (6), Power Infrastructure (2), Other (2)
Deliverables/Outcome	Assessment and Frameworks (12), Policy Analysis (11), Case Study Analysis (9), Review (4), Scenario Modelling (4), Other (9)

Figure 3. Annual publication trend and geographical distribution of 399 articles under the keyword climate change, infrastructure, resilience, and decision up to date of WOS search (March 2022).

3.1. Assessment of resilience

Resilience assessment frameworks typically entail qualitatively and/or quantitatively measuring specific aspects like the ability to rebound from disasters. One version of resilience assessment is focused on qualitative description of resilience. This includes matrix structures for criteria to explore indicators and relationships between resilience indicators and multiple criteria, which can be social, economic, and material (Sharifi & Yamagata, 2016). There are also many qualitative assessment tools for community resilience (Lavelle et al., 2015; Summers et al., 2017). These tools often describe general community resilience for physical and social systems, but do not necessarily place an emphasis on climate change and/or climate uncertainty. In recent years, this approach to resilience assessment has been supplemented by assessment tools that use quantifiable resilience indexes for single infrastructure projects, cities, watersheds, or communities. For the remainder of this review, we focus on these quantitative resilience assessment approaches.

Multiple quantitative resilience assessments have been developed within the last few years (Gerges et al., 2022; Heinzlef et al., 2019; Hsieh & Feng, 2020; Javadpoor et al., 2021; Kaaviya & Devadas, 2021; Poo et al., 2021). They often become frameworks for enhancing knowledge and guiding infrastructure adaptation. These assessment approaches typically include comparisons between spatial scales, temporal scales, and different adaptation strategies, with the goal of engaging stakeholders (Cardoso et al., 2020). As discussed in more detail in the next section, these assessment frameworks can also be used as decision support tools to aid the identification and prioritization of resilience efforts in critical infrastructure.

Main takeaways from publications in this category are a trend towards the use of technology, e.g., spatial modelling software for flood risks and heat exposure, as well as use of ‘smart sensors’ (Berglund et al., 2020; Reda Taha et al., 2021). These sensors can be used to monitor traffic, water flow, heat, and wind hazards. Fast communication and bundling of information on infrastructure surroundings and performance helps decision makers and engineers react to disturbances and model future scenarios for infrastructure planning. Another trend is towards economic cost analysis over the lifetime of the infrastructure. This includes a comparison between the cost of (future) adaptation and the cost of (future) failure (Fischbach et al., 2019).

3.2. Dimensions of resilience assessments

The presented assessment frameworks often include multiple dimensions in their analysis. Some assessments utilize spatial information to compare infrastructure sites, while others include temporal considerations such as the life cycle of the infrastructure or costs and system performance under different climate change scenarios. Furthermore, approaches can be based on qualitative criteria and/or quantitative metrics. For example, a qualitative resilience factor has been combined with cost-based calculations (Espinet et al., 2016), and resilience has been ‘measured’ as the ability to absorb change with a value between 0 and 1 (Song et al., 2018). Compared to earlier resilience assessment reviews (Lavelle et al., 2015; Summers et al., 2017) there is a trend toward quantifiable resilience, which creates opportunity for spatial and temporal comparison.

This observation can be summarized into three main dimensions of assessment tools: spatial considerations, temporal considerations, and the depth of analysis. The spatial dimension includes the use of GIS data, which enables comparison between infrastructure locations as well as scales of assessment. While the local scale (e.g., neighbourhood and street level) is often the main focus of resilience assessments, a comparison with the whole city or region can aid the consideration of trade-offs and the prioritization of strategies. For flood risk management, the mapping of risk areas and flood plains plays a significant role for community development (Adedeji et al., 2019). A detailed summary of the dimensions of resilience assessment can be found in Table 2.

Table 2. Dimensions of resilience assessment with description and respective literature examples.

Dimension of Assessment	Description	Examples
Spatial	Use of geographic information systems (GIS) to analyze resilience assessment metrics on various scales Enables comparison within organizational level (neighborhood, local, state, national) to identify priorities for infrastructure placement and adaptation Visualization of infrastructure locations (e.g., bridge), network (transportation or water) or community (neighborhood, city)	(Bertilsson et al., 2019; Gerges et al., 2022; Heinzlef et al., 2019; Kaaviya & Devadas, 2021; Lamaury et al., 2021)
Temporal	Using projected hazards or climate scenarios to assess infrastructure Some include different timelines of decisions, potentially measures regret and path dependency Mostly medium and long term (20 years or more), often same terms as climate predictions	Includes future climate scenarios (Espinet et al., 2017; Underwood et al., 2020)
Depth of Analysis	Utilization of different approaches to measure resilience: Comparison of hazards and exposure, inclusion of risk analysis and economic values (investment needed to adapt vs. costs of damages), comparison of different decision/hazard scenarios Communication of results through visualization, qualitative or quantitative resilience assessment matrices Utilization of community and stakeholder knowledge, web-applications, and software development Trend towards calculating a resilience value, quantifying resilience as potential to rebound	Evaluation of economics, comparing strategies, assessing regret and risk (Espinet et al., 2017), visualization, cascading effects, modelling what-if scenarios (Vamvakeridou-Lyroudia et al., 2020) Resilience rating factor based on multiple hazards and infrastructure parameters used to compare and prioritize infrastructure (Hsieh & Feng, 2020) Resilience rating factor between 0 and 1 for a larger system (Poo et al., 2021; Song et al., 2018)
(Model) Complexity, Interconnectedness	Model includes multiple infrastructure sectors, multiple hazards, or compound hazards/interconnectedness of infrastructure E.g., precipitation hazards impact multiple sectors, while one infrastructure sector gets impacted by multiple hazards	Multi-sector model (Gerges et al., 2022) Multi-Criteria model (Kaaviya & Devadas, 2021; Poo et al., 2021)
Data and Inputs for Assessment	Does the assessment framework include compound or multiple hazards? Does the assessment include high quality data, compare different data? Use of downscaled global or regional climate models, hazard models or empirical data	Includes different return periods for rainfall events to assess resilience of stormwater management system (Valizadeh et al., 2019) Multiple hazards and parameters (Poo et al., 2021)

Many resilience assessment approaches include at least two of these dimensions (often multi-sector/criteria and spatial dimension), but do not include any future changes or evaluation of economic risk or alternative designs (Kaaviya & Devadas, 2021). On the other hand, when economic risk assessment is included, the spatial dimensions are usually not (Espinet et al., 2017). The importance of the temporal dimension and planning for a future climate is acknowledged by most publications, but rarely incorporated into the assessments, which usually take a more static approach.

An earlier review article of transportation network resilience concluded that insufficient practical models for resilience quantification during recovery and the development of a resilience index are key problems (Twumasi-Boakye & Sobanjo, 2019). In combination with this literature review, results indicate that this key problem is a valid concern for other infrastructure sectors as well. Finally, given varying depth and dimensions of these assessment, further research aimed at convergence appears warranted.

3.3. Decision making

Modelling decision making under uncertainty and analysing different policy options under different scenarios are important considerations for adapting to a changing climate. Decision making under uncertainty is a complex and nonlinear process that requires examination of multiple (often conflicting) criteria, as well as consideration of the decision-making process itself as part of the analysis (Fischbach et al., 2019; Groves et al., 2018; Hall et al., 2012; Helmrich & Chester, 2022; Lempert & Groves, 2010; Sriver et al., 2018; Webber & Samaras, 2022). As discussed below, publications in this category include advances on decision making and modelling as well as policy analysis.

3.3.1. Decision modeling

Topics in decision modelling for infrastructure under climate change include multi-criteria decision analysis, decision making under (deep) uncertainty, and complex systems analysis (e.g., infrastructure interdependencies, cascading failure, path dependency, regret; Chester & Allenby, 2019; Espinet et al., 2017; Gilrein et al., 2021; Markolf et al., 2018). Planning seeks to answer questions related to what and who, while decision modelling seeks to answer questions related to if, when, where, (sometimes) how, and (sometimes) why to adapt existing and/or new infrastructure (Meerow & Newell, 2019). Scenario modelling, which aligns with climate scenarios and often incorporates future climate conditions, is another common approach (Espinet et al., 2017; Underwood et al., 2020). Multicriteria decision analysis is relevant and predominant in urban flood risk management (Da Silva et al., 2020) and present in many resilience assessments (see, Table 2). Decision support tools based on multiple criteria, interdependencies, and modelling the impact of decisions across a timeline of changing conditions are emerging (Moallemi et al., 2020; Mostafavi, 2018; Oikonomou et al., 2021). Increasingly, economic considerations are a critical component of the design phase and can be represented as constraints in decision modelling (Singh et al., 2020). More extreme hazards and higher load factors can lead to larger constructions and use of more resources, which can in turn be weighed against the risks of failure. An example is modelling the costs and regrets associated with infrastructure adaptation strategies for various future conditions, using robustness-based prioritization (Espinet et al., 2017). These approaches can also be helpful for avoiding path dependency, maladaptation, counter-effects, and regret that can emerge from rapid decision-making or ill-conceived incentives (Pittock & Finlayson, 2013; Sadiq et al., 2020).

Flexibility in the decision process and during infrastructure planning, construction, and reconstruction after disasters plays a key role in resilient decision making under uncertainty. This flexibility is incorporated in dynamic adaptive planning methods based on decision trees and scenarios (Singh et al., 2020), as well as in the modified observational method, which includes continuous monitoring and modifications (Ayyub, 2018). Flexibility can be instilled through proactive planning (Abunnasr et al., 2015; Berke et al., 2021), as well as the use of diagnostic frameworks to identify the strategic actions needed to facilitate infrastructure transitions (Ferguson et al., 2013a, 2013b).

3.3.2. Policy analysis and stakeholder perspectives

Many articles analyzed policies of European and North American states and communities in reaction to climate change and aimed at developing resilient infrastructure. Common contents were policy impact analyses (especially for nature-based solutions; Anderson & Gough, 2022), reviews of flood risk management (Adedeji et al., 2019), and evaluation of stakeholder responsibilities (Birchall & Bonnett, 2021). Interviews of practitioners for their perception of issues (Mclean & Becker, 2020; Schulz et al., 2017) and research needs (Molino et al., 2020) were also included. These articles highlight several shortcomings in current practices and policies, with emphasis on increasing the flexibility of local regulations and institutions, more decentralized decision making, better integration of state-of-the-art data/analysis with local information, and increased attention to processes in the design-phase (Bender et al., 2020; Cradock-Henry & Frame, 2021; Tullos et al., 2021). Local standards and policies neither encourage nor require alternatives to traditional planning nor the consideration of climate change. Thus, systematically revising codes and operations will be a crucial step towards infrastructure resilience.

Publications also expressed the need for assessment tools for resilience to be usable during the planning process of new infrastructure. Recommendations for reform include flexibility of decision making and planning options, research on innovative technology and methods, decentralizing infrastructure, and educating and training stakeholders in addition to revising standards (Adedeji et al., 2019; Davies & Lafortezza, 2019). It has also been recognized that the window of opportunity for reforming these rules is greater if no infrastructure yet exists, and new construction invites the most opportunity to pursue resilience (Goytia et al., 2016).

3.4. Advances in practice and education

In structural engineering, power infrastructure, stormwater and flood protection, heuristic techniques for updating loads from climate parameters, such as precipitation, temperature, and wind have emerged in practice. These techniques can either be advances on statistical methods like traditional return periods, or performance-based design and risk-based evaluation methods (Olsen, 2015; Shadabfar et al., 2022). Advances on statistical methods are diverse. Loads can be adjusted with a simple percentage increase, additional safety factor(s), adaptive percentage increase, a percentage increase based on relationships of the load/hazard to climate change, or loads based on climate models (Martel et al., 2021). There is also the idea of obtaining load values from modelling of single extreme events outside of continuous hazards, instead of using probabilities and risk-based approaches (Al-Rubaye et al., 2022). If annual maxima of nonstationary hazards are calculated with time varying parameters, seasonality of hazards can also be included. Many load adjustment methods assume that climate change impacts are continuous (Mishra et al., 2022).

There is a trend toward flexible design alternatives like safe-to-fail infrastructure (Kim et al., 2019, 2017), modular systems, phased implementation, and platform designs (Gilrein et al., 2021; Spiller et al., 2015), as well as new planning principles (Monteiro et al., 2020). There are propositions for approaches to devise practical resilience metrics. These include resilience, recovery, disruption time and failure (Ayyub, 2015), which can then be used in conjunction with resilience assessment frameworks.

There is also a growing body of work developing methods for downscaling global and regional climate models to predict future loads (Martel et al., 2021; Miro, DeGaetano, Lopez-Cantu et al., 2021; Westra et al., 2014; Wright et al., 2019). The availability of data downscaled from climate models for immediate use in infrastructure planning expands upon traditional methods for establishing infrastructure design criteria. One example are projected precipitation data compared to traditional data sources like the official Atlas 14 precipitation frequency estimates (Miro, DeGaetano, Samaras et al., 2021). Utilizing this data for infrastructure planning and design will be a crucial step in creating resilient infrastructure.

A key challenge for the use of downscaled climate data is potential misalignment of the spatial-temporal resolution of the data and the infrastructure design decisions they inform. Oftentimes, relevant regional climate conditions are influenced by local geography and atmospheric conditions. The output scale of <4 km x 4 km is necessary and available, but have high computational requirements (Martel et al., 2021). Model choices of climate modelling influence the calculated hazards (Cook et al., 2020). Furthermore, hazards used for load calculations need to be estimated at a sub-daily timescale (Westra et al., 2014). Daily averages of precipitation events, storms and temperature spikes cannot be used for planning and design since infrastructure must be able to function under the extremes.

Additionally, these strategies to utilize climate model data still lack supporting policy frameworks. Nonetheless, growing awareness has led to the emergence of holistic sustainability rating systems, which include climate resilience planning and encouragement of new methods (Goytia et al., 2016; Minsker et al., 2015). Multiple frameworks for sustainability ratings like Envision (Dunford & Gillis, 2019), LEED and SITIES (Minsker et al., 2015) are mentioned as supportive tools for a paradigm shift in planning and design of infrastructure. The Envision framework is described as flexible with the benefits of documentation of design decisions and monitoring over the project lifecycle, which facilitates knowledge transfer for future projects (Sheppard et al., 2019). It is deemed important that sustainability ratings include the decision-making process and not only the final product.

Furthermore, practitioner-oriented publications summarize design examples and best practice options like nature-based solutions, including their co-benefits. Increasing emphasis is also being placed on the interconnectedness of infrastructure designs and their performance (or respective benefits) under multiple hazards (Kiel et al., 2016; Van de Ven et al., 2016). Another notable publication compiles a list of agile and flexible infrastructure cases (Gilrein et al., 2021). Public collection of case studies like these, which include descriptions and characteristics of infrastructure examples, can serve as a valuable reference for practitioners and researchers.

The ASCE library includes Manuals of Practice, which promote adaptation methods, coping strategies, and workarounds for practical engineering. As compilations of state-of the art and practice for infrastructure analysis in the context of uncertainty and risk, they are well-suited as references for consulting engineers and textbooks for educators. Two recent examples concerning infrastructure resilience towards hazards and climate change have been published in the last 4 years (Ayyub, 2018; Ayyub et al., 2021). These manuals of practices are complemented with books (Kelly et al., 2017; Ojha et al., 2017) as well as a white paper (Olsen, 2015), which reference the ASCE Policy Statement 360 on Climate Change. This statement declares support for a revision of engineering standards, research, development, and government policies to prepare for climate change impacts (ASCE, 2021).

The MOP by the ASCE Committee on Adaptation to a Changing Climate defines the problem of climate uncertainty, characterizes extremes, reviews climate-changed informed loads, summarizes adaptive designs and risk management (including life cycle costing and resilience metrics), and gives guidance on the use of climate models and data produced by entities, such as National Oceanic and Atmospheric Administration (NOAA), US Geological Survey (USGS), US Bureau of Reclamation (USBR), Department of Energy (DOE) and other (Ayyub, 2018). The more recent MOP by the Infrastructure Resilient Division lists and explains methodologies and attributes for resilience assessment approaches, as well as mechanisms and design principles of resilience. Among them are flexibility, decentralization, nature-based solutions, community culture, transcending spatial and temporal scales, diversity, and redundancy (Ayyub et al., 2021). The 2015 white paper on infrastructure adaptation to climate change includes thoughts on incorporating climate science in infrastructure sectors (Olsen, 2015). Recommendations include extreme events research, understanding of future projections, adoption of risk and alternative strategies as new design paradigms, as well as resilience assessment research. It is recognized that approaches like safety factors, adding freeboards against wind and flood, and employing statistical methods can help account for and address uncertainty. However, model uncertainty still inhibits assessment and selection of appropriate monitoring techniques and subsequent infrastructure modifications (Ayyub, 2018; Olsen, 2015).

4. Discussion

Analysed publications in this literature review were limited to the WOS and ASCE search results. Using publication databases with a keyword search is a limitation for this study. Publications in languages other than English or German were not considered, and keywords can unintentionally exclude knowledge that is relevant but associated with different keywords. There is potential that practitioners are developing solutions and workarounds to cope with climate uncertainty that are not published in peer reviewed articles or case study literature. Stakeholder interviews like (Mclean & Becker, 2020) can include these perspectives to some extent. There is also a dearth of studies and analysis related to the adaptation approaches and perspectives of the consulting industry, which often plays a critical role in infrastructure planning and design.

For the category of resilience assessments, multiple approaches and dimensions have emerged. Research to converge assessment methods into a set of frameworks operating under comparable principles seems necessary. For this research, applicability of assessment frameworks should be tested with case studies, and increased emphasis should be placed on translating knowledge to stakeholders. Another opportunity for advancement is clearly delineating the circumstances and context for which different resilience assessment metrics are applicable. If resilience metrics require certain data or methods that are not available for other hazards or sectors, interconnected systems resilience might not be measurable. Relatedly, efforts to better contextualize and improve the interpretability of resilience metrics appear warranted. Current resilience indices are often calculated as a percentage or absolute numbers with no defined scale (Hsieh & Feng, 2020). In either case, it is often not clear which insights and actions should result from resilience quantification. Hypothetically, how should stakeholders respond when a system is found to have a resilience index of 77%, or if one component has a resilience index of 20, while another has a resilience index of 10? Different resilience definitions can also cause opposite interpretation of values. In some cases, a high index value can be interpreted as low resilience (Poo et al., 2021), while in others, it can correspond to high resilience (Argyroudis et al., 2020; Valizadeh et al., 2019). It is also unknown how many assessment frameworks are making their way to practitioners as opposed to being exclusively used for research.

For decision making, actions and needs arise out of the policy analysis. Consideration of climate data/projections and the establishment of risk and performance-based approaches can play a vital role in updating policies and regulations for changing and uncertain conditions. Well informed selection of adaptation strategies can be supported by the development of guidelines and pathways for the implementation of alternative design paradigms (e.g., safe-to-fail, risk-based design, modular systems). There is also opportunity for advancing approaches for evaluating the timeliness of decisions, i.e., consideration of differences in costs and performance if infrastructure adaptation strategies are implemented before, during, or after disruptive events.

In the practitioner space, use of the Envision framework is reported as currently being the only holistic sustainability framework for large-scale projects. Limits, measured success, and co-benefits of uses with other sustainability ratings should be examined. The authors attribute particular importance to the ASCE Manuals of Practice publications as means for reaching practitioners and educators. They seem suitable to be used in the classroom for training future engineers as well as references for consultants. Although this ASCE material exists, it is unknown to what extent climate change adaptation is taught in the classroom. Do civil and environmental engineering degree programs teach standardized and traditional design calculations based on safety factor for water, power, transport and other infrastructure, or the skills needed to plan for uncertainty and complexity? It would be illuminating to explore if and how education and trainings can cope for missing policies and guidelines reported by stakeholders.

5. Conclusion

This paper has outlined gaps and advances in research and implementation of infrastructure planning, design, and implementation that is congruous with a changing climate. The literature revealed three main categories of research publications: resilience assessment, decision-making and policy, and examples of best practices to enhance resilience. The main takeaway from this study is that the different approaches and definitions in the field of infrastructure adaptation need to be phased into cooperative methods. This is especially true for resilience assessments including quantitative resilience indices, which is a key research priority. The field of infrastructure research under climate uncertainty is very diverse. Overarching concepts like risk approaches, performance-based design, robust decision-making, multicriteria decision analysis and decentralization exist, but the development of practical solutions for respective hazards and sectors diverges. Also, the translation of research and practical knowledge into policy is urgent. Due to old standards, load assumptions based on climate change knowledge are rarely utilized, although new methods exist. A priority in practice must be the implementation of standards that describe and encourage usage of climate data or alternatives like risk-based design. Further gaps in research include the development of models for infrastructure resilience, engineering education, and considerations within the engineering consulting industry.

Disclosure statement

The Coalition for Disaster Resilient Infrastructure (CDRI) reviewed the anonymised abstract of the article, but had no role in the peer review process nor the final editorial decision.

Additional information

Funding

The Article Publishing Charge (APC) for this article is funded by the Coalition for Disaster Resilient Infrastructure (CDRI).

Notes on contributors

Marie Buhl

Marie Buhl is a PhD student in the Environmental Systems program at UC Merced. She is a German Diplom-Ingenieurin in Civil Engineering specialized in hydraulic engineering and has worked as a consulting engineer before coming to UC Merced. Her research interests include how climate change is impacting the planning and design process for infrastructure, especially water infrastructure. Having witnessed the struggle of engineers to adapt the infrastructure process to new conditions within the framework of existing standards, she wants to help develop strategies and toolkits that help local stakeholders respond to disasters and strengthen their infrastructure.

S. Markolf

Dr. Samuel (Sam) Markolf is an Assistant Professor within the Department of Civil and Environmental Engineering at UC Merced. Prior to joining UC-Merced, Sam was a Research Fellow on the NSF-sponsored Urban Resilience to Extremes Sustainability Research Network (UREx SRN) at Arizona State University. Broadly, his research applies systems thinking to sustainability and resilience challenges facing cities and infrastructure systems. Example projects include exploring the incorporation of climate projections into infrastructure design processes, examining impacts and responses to extreme events (e.g., flood, extreme heat) within transportation systems, and analyzing the extent to which interconnected social-ecological-technological systems (SETS) can enhance (or hinder) the resilience of cities and infrastructure systems. Ultimately, his aim is to help decision-makers become more adept at identifying, anticipating, alleviating, and responding to accelerating climatic, technological, and social change.

Sam holds a B.S. in Chemical Engineering from the University of Texas at Austin, M.S. in Civil & Environmental Engineering from Carnegie Mellon University, and a joint-Ph.D. in Civil & Environmental Engineering and Engineering & Public Policy from Carnegie Mellon University.