Climate-resilient Infrastructure Adaptive Design And Risk Management

[Endike Baur]

Infrastructuredevelopment requires a broad range of actors, including policymakers, planners, funders, engineers, researchers, and communities. Together they shape the physical structures and services intended to provide critical support to the public for decades. Some of these actors have attempted to integrate consideration of future climate risk into infrastructure decisions. For example, the American Society of Civil Engineers (ASCE) developed a top-down approach for adaptive risk management that starts with the climate scenario and projections shown in figure 1. But despite these promising efforts, innovative ideas and strategies for addressing climate risks are not yet common practice.

How can resilience be designed, funded, and incorporated in public infrastructure investments? This question puts engineers and other professionals in a race against a changing climate to enhance planning and design practices.

Climate and weather are exhibiting intensifying extremes, but infrastructure systems traditionally have been designed, constructed, operated, and maintained according to assumed stationary climate and weather conditions in stochastic terms, without taking account of future climate change and associated uncertainties that increase and intensify hazards. Adaptation technologies, adaptive designs, and appropriate new policies (e.g., addressing risk management) are necessary to reduce climate impacts and risks.

In 2014 the Intergovernmental Panel on Climate Change (IPCC) concluded that warming of the climate system is unequivocal and that it is extremely likely that the dominant cause of the observed warming since the mid-20th century is human influence. Graphs from the congressionally mandated Fourth National Climate Assessment (NCA4; USGCRP 2018) depict these trends in figures 2 and 3. Since the 1950s, anthropo-genic greenhouse gas (GHG) emissions have driven many of the observed changes, which are unprecedented over decades and even millennia (IPCC 2014). Future GHG emissions are uncertain and depend on many factors, including policy, population growth, globalization, economic activity, and human behavior, all of which require the use of projections to define emissions pathways (European Commission 2015; Hausfather 2018).

Figure 2

Figure 3

According to the NCA4, the global annually averaged surface air temperature increased by about 1.8F (1.0C) from 1901 to 2016. This 115-year period is now the warmest in the history of modern civilization, and the last few years have also seen record-breaking climate-related weather extremes. Researchers have documented changes in surface, atmospheric, and oceanic temperatures; melting glaciers; diminishing snow cover; shrinking sea ice; rising sea levels; ocean acidification; and increasing atmospheric water vapor.

Without major emission reductions from policy, behavioral, or technological change, annual average global temperature relative to preindustrial times could increase by 9F (5C) or more by 2100. With significant reductions, it could be limited to 3.6F (2C) or less, which would still be associated with serious climate impacts.

In light of the uncertainties surrounding how much climate will change, climate scientists and engineers struggle to accurately characterize future climate and weather extremes even in probabilistic terms, requiring projections based on numerous assumptions about GHG pathways.

While the evidence that climate is changing is very strong, the engineering community has found it difficult to incorporate consideration of climate change at temporal and spatial scales relevant to engineering practice. Development of adaptation and mitigation technologies and strategies has proved elusive based on current practices.

As an example, relative sea level rise (SLR) is a locally observable phenomenon that reflects changes in the eustatic sea level, the subsidence or uplift of the sea floor, and the accumulation, erosion, or compaction of sediment along the coast. Sediment accumulation and erosion are greatly affected by subsidence and uplift, so these processes are often mutually reinforcing (ASCE 2018). The tectonic setting of continental margins plays a primary role in determining whether a section of coastline experiences uplift or subsidence due to large landmasses breaking up with geological interactions. In addition, glacial loading contributes through (a) glaciation depression, with associated isostatic adjustments during periods of widespread continental glacial coverage, or (b) rebounding due to glacial retreat, creating uplift at rates slower than that of the retreat. Besides long-term changes on geologic time scales, relative SLR is affected by fluid withdrawal, diversion or elimination of sediment sources, and other human activities.

High-resolution SLR projections are important for the development of durable engineering designs. They are useful tools to support coastal hazard mitigation design criteria and communicate projected changes to stakeholders. However, engineering works must also consider the significant unknowns in future SLR, including those resulting from possible GHG pathways and those based on the behavior of ice sheets, which remain uncertain.

Civil engineers have dealt with uncertainty in geotechnical engineering practice by employing the observational method (OM), originally proposed by Karl Terzaghi (Nicholson et al. 1999; Peck 1969; Terzaghi and Peck 1948). The specific steps in a climate change OM (modified after Terzaghi and Peck 1948) are as -follows (also see Ayyub and Wright 2016):

The uncertainty associated with future climate is not completely quantifiable, and therefore accounting for it in engineering practice requires understanding and treatment of uncertainty as well as engineering judgment. Uncertainty can be broadly classified as follows for convenience (Ayyub and Klir 2006):

Climate projections entail similar types of -uncertainty (Stainforth et al. 2007), although they are expressed differently for the purpose of communication to the public; however, these five classes offer a basis for engineering and planning guides and standards.

It is common practice in engineering to classify uncertainties as (a) aleatory and (b) epistemic. The former is considered inherent to the situation and irreducible with data collection, although its characterization can be enhanced. The latter is reducible with data collection or investigation, although the economic cost of investigating the data might not justify the reduction.

Planning for a changing climate entails planning not only for climatic uncertainty but also for uncertainty about regulatory, environmental, economic, social, and other conditions affecting an engineering project.

Risk methods provide practical means for managing uncertainty (Ayyub 2014). Risk is commonly measured in simple terms as the probability of occurrence of an event or scenario and the outcomes or consequences associated with the occurrence. Risk assessment is primarily concerned with answering three questions (Kaplan and Garrick 1981):

Engineers should develop a new paradigm for engineering practice as climate change, population growth, and development patterns alter the risk profiles of projects, communities, and even nations. The effects of climate change may be difficult to estimate with a high degree of certainty in many instances, but clear indications are available of some effects, such as SLR, more frequent extreme heat events, and an increase in the number of extreme events in areas where they have rarely been encountered. These suggest a changing footprint of risk, regardless of the magnitudes of the events on regional or national scales.

Engineers design infrastructure by accounting for uncertainties to achieve acceptable safety levels and appropriate physical and economic efficiencies. Uncertainty is thus foundational in developing a design philosophy. Engineering design evolves based on an enhanced understanding of uncertainty.

The five types of uncertainty noted above drive engineers toward practice enhancements. When -uncertainty is recognized and well characterized, engineers use tradi-tional factors of safety, followed by -reliability-based design in building codes. For cases 2 and 3, where uncertainty is recognized but moderately or poorly characterized, engineers use reliability-based or risk-informed designs.

When dealing with climate change, it is important to recognize that it is not possible to define a hazard with probability distributions a priori. Scenario modeling, which can be used to perform sensitivity analysis and address variability, incorporates conditional probabilistic information such as uncertainties owing to spatial variability of seismic demand, random phasing of ground motion, local soil conditions, and performance levels of civil infrastructure (which can be gauged from fragility curves as conditional probabilities for varied hazard levels). Scenario modeling is helpful to design for a changing climate, but may be insufficient.

Robust decision making (RDM) can address cases where deep uncertainty is recognized but either poorly characterized or incapable of being characterized. It provides an analytic framework to identify strategies that can perform over a wide range of poorly characterized uncertainties. In turn, RDM strategies may provide a basis for a number of scenarios to be analyzed while incorporating probabilistic information.

The RDM framework could identify strategies that would be insensitive to vulnerabilities associated with deep uncertainty in future climate projection models. However, it might not produce cost-effective solutions.

When uncertainty is unrecognizable or it is not possible to fully define and estimate the risks and potential costs for a project to reduce the uncertainty in the timeframe in which action should be taken, engineers should use adaptive design or risk management. These are most effective in cases 4 and 5, although they can be used in others. Wilby and Dessai (2010) present a robust framework called adaptive management of climate risks, which involves monitoring of the environment and systematic assessment of the performance of measures installed. The approach calls for inventorying preferred adaptation strategies that are then synthesized into a subset of measures that reduce vulnerability under the current climate regime. The resulting strategies should be able to perform well over a variety of scenarios, regardless of climate change and conditions.

3a8082e126