Climate change has been identified as a likely existential threat to human civilization and ecological systems over a time period of decades [1
]. Climate change pollutants, dominated by CO2
from combustion of fossil fuels, is often emitted alongside criteria air pollutants that have direct and immediate effects on human health including reduced life expectancy and childhood respiratory problems, to name but a few [3
]. Global, national, and regional governments have developed targets for mitigating greenhouse gases (GHGs) to reduce or avert climate change impacts [9
]. GHG emissions and mitigation targets are often distinguished by the broad economic sector to which they belong, including industry, transportation, and residential/commercial buildings. Transportation, the large majority of which is combustion of fossil fuels by on-road vehicles, produces more than 40 percent of GHG emissions in California [16
], though transportation’s contribution varies by state and country, and is a lower percentage in most other states and countries [17
The United States has over 4 million miles of public roads ([18
] and 2.65 million miles of paved roadways [19
], supporting nearly 3 trillion vehicle miles traveled annually [20
]. The nation’s roadway system is one part of a transportation network that provides mobility and access to a range of users (e.g., access to schools, services, and work; leisure travel; and general mobility) [21
]. The roadway system is also vital to the economy because it enables the movement of freight and commodities and is a major source of employment. Roads carry about 65% of all freight in the nation, in terms of both tons and dollar value [22
]. More than 300,000 people were employed in the road and bridge industry in 2014, and even more were employed before recent cuts in national transportation infrastructure funding. Most of these jobs do not require a college degree and typically offer higher wages than jobs requiring similar educational backgrounds [23
However, operation of the nation’s pavement network, which includes both its construction and its maintenance, is costly. The total annual construction and maintenance expenditure for U.S. highways (pavements and bridges) in 2014 was $
165 billion [24
]). Highway construction and maintenance also requires large inputs of energy and natural resources, and causes significant emissions of GHGs, criteria air pollutants, and water pollutants. Vehicle operation on the nation’s roadways consumes more than 181 billion gallons of fuel [25
] and the amount of energy consumed by vehicles is affected by the pavements they roll on.
Taken together, these numbers demonstrate the magnitude of the investment in public roadways, and the system’s importance in supporting movement, access, and mobility. At the same time, there is increased recognition of the harm caused by pollutants from roadway construction and demolition of worn-out materials; of the influence of pavement on the fuel use of vehicles and on the surrounding environment. The transportation sector of the US economy produces about 28.5 percent of national GHG emissions [26
California has established a series of mandated targets for reducing GHG emissions contributing to global warming. Governor’s Executive Order S-3-05 (2005) required a reduction of GHG emissions to 1990 levels by 2020, and a reduction to 80 percent below 1990 levels by 2050. California’s 2006 Climate Change Solutions Act (Assembly Bill 32) made the 2020 reductions law and tasked many government entities, including local governments and government agencies, with helping to meet those goals. Governor’s Executive Order B-30-15 (2015) requires a reduction of 40 percent below 1990 levels by 2030, which was codified into law with Senate Bill 32 in 2016. Executive Order B-55-18 (2018) requires carbon neutrality for the state by 2045.
The most recent California Climate Inventory published by the California Air Resources Board (CARB) found that in 2016 the state emitted 429.4 million metric tons (MMT) of carbon dioxide equivalent (CO2-eq
; calculated by CARB using the global warming potential (GWP) factors published in 2007 by the Intergovernmental Panel on Climate Change Fourth Assessment Report over a 100-year time horizon), achieving a 30 percent reduction from 2005 levels and meeting the 2020 goal of a reduction to 1990 levels four years ahead of time [27
]. The 2016 inventory showed that the transportation, industrial, and electricity generation sectors of the economy were the three largest sources of emissions, with 41%, 23%, and 16 percent, respectively. The large majority of the emissions in the transportation sector is from combustion of gasoline and diesel, and in the electricity sector are from combustion of natural gas in in-state power plants, and importation of energy from combustion of coal in out-of-state plants. In the industrial sector, a large portion of the emissions are from oil and natural gas production and oil refining, of which a portion can be attributed to an asphalt binder used in transportation infrastructure, and there are contributions from the production of cement and steel used in transportation infrastructure.
There is no single change that will achieve the ambitious goals set by the state for itself, instead multiple changes must be made in the state’s economy by many actors. Many changes are being proposed by a multitude of sources, with the proposals based to varying degrees on science, economics, the potential to grow markets or shrink the markets of competitors, regulatory strategies, and attractiveness often based on the simplicity of the idea and the ability to easily communicate it to policy-makers and the general public. Identifying, quantifying, and then selecting among the many possible strategies to achieve GHG reductions is difficult, especially without a standardized approach for comparison.
Over the last five years, state and local policy leaders and transportation system operators have identified the difficulty of prioritizing the tens of strategies and tactics that are being proposed for changes in how they should design and operate systems to reduce GHG emissions to meet the state’s goals. One method of doing so is by considering policy packaging, which capitalizes on the synergistic relationship between various policies and maximizes effectiveness through appropriate bundling. Researchers have proposed a framework to group policies according to this synergistic relationship [29
]. Others have developed a ranking method that considers more than just synergistic relationships, such as one-way facilitative relationships and contradictory relationships [30
]. One shortfall of these methods is that they are inherently qualitative. For example, the ranking framework developed by Taeihagh et al. [30
] relies on, among other things, expected public approval, and the perceived extent to which policies are interlinked. Additionally, they only look at the direct, “use stage” impacts of the projects, which is sufficient for some policies but certainly not for all.
In parallel with integrated packages of policies to address environmental issues, it is the experience of the authors that at the state and local levels there is a stream of essentially ad hoc proposed regulations, laws, specifications, and policies being proposed that are intended to reduce environmental impacts that appear to be beneficial. These proposed changes are often unquantified, and have been considered within narrow system boundaries and without consideration of life cycle implications. The authors have worked extensively supporting state and local government in California with “life cycle thinking” data and tools for implementation of life cycle assessment (LCA) for environmental impacts (including social impacts) and life cycle cost analysis (LCCA) for financial impacts. Therefore, a process is proposed that considers the full system and complete life cycles, as this provides the most beneficial and sustainable solutions while minimizing the likelihood of unintended negative consequences. It is recognized that the impacts and implementation of considered strategies can change based on the concurrent policies. However, the scope of this project is not to propose policy decisions, but to inform them. The final prioritization of strategies will differ across jurisdictions depending on their goals, funds, and timeline.
The focus of this study is to examine two strategic options that the California Department of Transportation (Caltrans) could adopt to lower its GHG emissions in operating the California state highway network so it can help meet the state’s climate change goals as examples of the use of the approach used to evaluate them. Although many GHG reduction strategies appear to be attractive, simple, and effective, the following limitations are also true for many of them:
The net GHG reductions that result from implementing the strategies have often not been quantified;
Few of the cases where GHG reductions have been quantified have used a system-wide perspective for their estimates;
The time it will take to implement a strategy and begin achieving GHG reductions has not been considered;
The process and difficulty of making the change have not been estimated, and;
Most importantly, the costs, or in some cases savings, of implementing both initial and life cycle strategies have rarely been estimated in a way that prioritizes the most cost-effective strategies that would allow maximal emissions reductions with minimal costs.
A life cycle perspective is required for GHG accounting because benefits achieved during one stage of a system’s life cycle may be reduced or reversed by unforeseen increases in GHG emissions simultaneously caused in upstream or downstream stages. Similarly, if an incomplete system view is taken, benefits in one part of the system may be reduced or reversed (i.e., more carbon is emitted than business as usual) in another part of the system that was not considered. In some cases, two or more potential changes in operations are incompatible with each other in ways that will negate the benefits, and a full system view can help identify these conflicts.
The last point in the bullet list above is considered equally important with the calculation of emissions, because state government and the state’s overall economy have a finite capacity and political will to pay for change. The approach used in the studies described in this paper is that the greatest and fastest GHG reduction will occur if there is a prioritization in terms of GHG reduction benefit based on the cost of the proposed system change. Prioritization based on the benefit to cost will result in the most efficient use of existing funds to achieve the maximum reduction possible, in other words the most “bang for the buck”. The capacity of the public and the state’s economy to implement the needed GHG reductions may be exceeded before the goals are reached without a prioritization of this type. It is also considered important to be able to demonstrate to the public that efforts are being made to achieve GHG reduction goals in the most cost-effective ways possible in order to help maintain public support for those goals.
The ability to quantify the full-system and life cycle effects of decisions and changes in systems is advancing and improving using the LCA approach and related analysis processes. LCA is a structured evaluation methodology that quantifies environmental impacts over the full life cycle of a product or system, including impacts that occur throughout the supply chain. LCA provides a comprehensive approach to evaluating the total environmental burden of a product, examining all the inputs and outputs over its life cycle, from raw material production to the end of the product’s life [31
]. The limitations and problems with LCA are also being identified so that more robust and trustworthy results can be produced. With regard to LCCA, the methodology is already mature and used within Caltrans for support of decision-making regarding infrastructure choices [32
The timeframe for change is also important because emission reductions that occur sooner will have greater beneficial impact than emission reductions that occur later or are spread out over a longer period of time. This is not accounted for in current GWP calculations. Time-adjusted warming potential (TAWP, [33
]) as well as GWP are used to account for the timing of emission reductions in the approach proposed in this paper. Use of TAWP will help identify strategies providing the “fastest bang for the buck”.
This study discusses and shows early examples of use of a GHG reduction “supply curve” framework to support decision-making by Caltrans. The supply curve, as used in these studies, provides a method for selecting the most cost-effective strategies for GHG reductions by undertaking the following process for each strategy: (1) quantification of the net effects on GHG emissions over the strategy’s lifecycle, (2) consideration of the time required to make the change happen, (3) exploration of the change process and difficulty of making the change happen, and (4) calculation of the initial and lifecycle costs of the strategy. This approach is currently being used for two studies:
To evaluate possible changes that Caltrans can make in its operations to reduce GHG emissions
To evaluate proposed actions for transportation in climate action plans that have been developed by cities and counties in California to reduce GHG emissions.
2. The Approach
The approach used was to support strategic prioritization of alternative changes for reducing GHG emissions using what are called “marginal abatement curves”, “supply curves”, or “McKinsey curves” after the company that has made extensive use of them [34
]. Supply curves illustrate the economics associated with changes and policies made for climate change reduction. In particular, the work done by Lutsey and Sperling [35
] demonstrated how alternatives within the transportation sector could be quantified and compared using the available information, and also compared with alternatives in other sectors of the economy. Transportation is particularly important in California because it is responsible for approximately 41 percent of annual GHG emissions in the state. This percentage has increased and actual transportation emissions have increased as other sectors of the economy, particularly generation and use of electrical energy, have decreased [28
A generic example of a supply curve, recreated from Lutsey [36
], is shown in Figure 1
. To implement the development of supply curves, a set of questions are answered, and calculations are completed using the best available information about the proposed changes, with each proposed change shown as a box on the plot (the complete set of questions and calculations are described later in the paper). The supply curve uses the best estimate of the benefit on the x-axis, with each box representing a proposed change and the width of the box indicating the size of the benefit. Reduction of GHG emissions is shown in the example, however this could be a performance metric for other environmental goals, such as air pollution, which is a major concern in California, as well.
The y-axis of the supply curve shows the cost of the change per unit of benefit. Two values are calculated for each proposed change using the best available information: the initial cost of implementation and the long-term or life cycle cost (LCC). As with the LCA information, the economic analysis of the proposed changes for the supply curve is developed with the best available information and documentation is required of the assumptions, calculations and quality of the information used.
The proposed changes are put in rank order of cost effectiveness, with color coding to identify the level of uncertainty of the information used for the analysis (not shown in the example in Figure 1
). All changes have an implementation cost, indicated by the unpatterned portion of the box, but some changes will potentially result in an LCC savings. Those changes that are to the left on the curve should be considered for implementation first, because they provide the most improvement for the least cost. Those that have negative LCCs are what Lutsey [36
] refers to as “no regrets” choices because they reduce costs over the life cycle. Moving to the right along the x-axis of the curve identifies the cumulative effect of changes towards the overall GHG reduction goal, and the increasing cost of achieving that goal. As with all economic analyses regarding public policy, the economic analysis should consider not only the overall costs, but who pays the costs or receives the savings, and whether those costs or savings are equitable.
The purpose of developing supply curves to review alternatives is to bring a full system analysis, life cycle thinking, and above all, quantification, to their development in a decision-making environment where they are often absent, and to support decision-making for prioritization that includes consideration of economics.
In the approach being used, LCA was used to estimate the benefit by comparing GHG emissions from the proposed change over the life cycle analysis period versus current practice. The LCA was performed using the best available information, which could range from very poor to very good based on ISO 14044 [37
] data quality parameters as discussed related to pavements in the Federal Highway Administration Pavement LCA Framework [31
]: time-related coverage, geographical coverage, technology coverage, precision, completeness, representativeness, consistency, and reproducibility. The documentation of the LCA for the supply curve includes a data quality assessment, which must be taken into consideration when comparing alternative proposed changes on the supply curve.
The details of the process for developing the supply curves included in this paper consisted of:
Definition of the functional unit and system boundaries for the technology;
Identification of available information;
Technology of the strategy;
Life cycle, including maintenance, rehabilitation, replacement or end-of-life.
Costs of the strategy;
Constraints on implementation relevant to implementation by Caltrans.
Creation of information;
By analogous estimating from existing sources about similar technologies, different scales of research, development or implementation, or implementation in different contexts;
By bottom-up estimation from existing sources about components of the technology.
Life cycle inventory and impacts;
Life cycle costs.
Assessment of data quality;
Inclusion of the strategy on the supply curve.
The best available information that could be obtained was used for both the LCA and the LCCA. Data quality for both was assessed as part of the data analysis. There was a wide range of data quality across the different strategies, primarily dependent on the maturity of the research and development for the strategy, and the amount of implementation that has occurred to provide real world costs.
A questionnaire was developed to help the analyst develop the LCA and LCCA analyses for the supply curve and to also develop information regarding the potential for implementation, including the definition of the technology and the system it would change, the state of readiness of the strategy, the responsible stakeholders, and the factors that would drive the change. The questions to support the LCA and LCCA are standard practice in these methods. The questions used to identify, document, and analyze the implementation of the proposed change come from the experience of the authors working with state and local government on the implementation of a wide range of technologies, primarily for transportation infrastructure, and from critical evaluation by the authors of climate action plans produced by local governments. They were not based on the literature of large-scale policy development. The questions in the questionnaire were:
Define the change/technology; this question requires the proposer to specifically define the proposed change;
Define the state of readiness of the change of technology using approach adapted from the NASA Technology Readiness Level system [38
]; decision-makers are often not aware of the readiness of the proposed technology;
TRL 1: basic principles observed;
TRL 2: technology concept formulated;
TRL 3 and 4: experimental proof of concept/ technology validated in lab;
TRL 5 and 6: technology validated or demonstrated in relevant environment at less than full scale (industrially relevant environment in the case of key enabling technologies);
TRL 7: system prototype demonstration in the operational environment (full scale);
TRL 8: actual system completed and “flight qualified” through test and demonstration;
TRL 9: actual system proven in operational environment elsewhere or less than full market penetration.
Define the system in which the change occurs this information defines the specific context in which the change is proposed to be implemented;
Will the market change or is it just changes in market share; this information addresses the approach to be used for the life cycle assessment, either attributional (changes within the existing market) or consequential (new markets appear and/or old markets disappear or are fundamentally changed);
Who is responsible for the change; implementation is often stopped because the responses of critical stakeholders, particularly those who must make changes, are not explicitly identified and resistance or buy-in planned for;
Who is responsible for implementing the change; most change requires a champion to push it, unless it is completely market driven;
Who pays for the change (this information is needed to identify the financial capacity and willingness of those responsible to pay for change):
What will drive the change (this identifies the implementation approach):
Market incentives (example, tax break);
Public programs incentivizing change;
What will the change do to these other environmental indicators (this identifies unintended consequences in other impact areas, particularly when the proposed change has one specific goal such as GHG reduction):
What are the performance metrics; this information is needed to assess progress and success during implementation, and make required changes in the implementation strategy if needed;
Supply curve calculation questions (the results needed to build the supply curve):
Expected change in GHG output per unit of change in system;
Expected maximum units of change in system;
Time to reach maximum units of change;
Expected rate of implementation;
Total estimated initial cost (to be used with total change in GHG to calculate initial cost per unit of change);
Estimated LCC per unit of change (to be used with total change in GHG to calculate the initial cost per unit of change).
Methodology for developing information to answer questions; this information is needed so that critical reviewers and stakeholders can review, understand, and critique the supply curve and implementation plans;
Any available documentation for answers to all questions; this is documentation needed for the transparency of scope, goals, methodology, data, and data quality;
Data quality assessment; this information is needed for decision-makers and stakeholders to understand the limitation of the quantitative information used for the supply curve, and can also be used to identify where additional effort should be made to develop better data for promising proposed changes that have high uncertainty;
Critical review of results; this is documentation of the critical review.
The information used to develop the answers to all questions needs to be fully documented, including:
Development of optimistic, best, and pessimistic estimates to the extent possible to permit sensitivity analysis; to help assess the robustness of the supply curve information;
Identification of the level of disagreement between different sources of information; needed to help assess the robustness of the supply curve information;
A ranking of the data and estimation quality such as excellent, good, fair, poor, and completely unknown; needed to help assess the robustness of the supply curve information.
The recommendation is to submit supply curves and their documentation to outside critical review by interested stakeholders before using them for decision-making and documentation of the critiques and responses by the supply curve developers, following ISO LCA principles.
The information collected in the framework includes both quantitative and qualitative data, which decision-makers are expected to use in a process of multi-criteria alternative prioritization. The supply curve provides the primary data, providing a rank ordering of strategy cost-effectiveness in terms of cost per unit change of impact, and quantification of contribution towards achieving impact reduction goals. However, to be implemented, strategies must be ready for implementation, and those that are not yet developed to a point at which they can be implemented have more uncertainty in the quantification of their costs and impact changes and will need further investment in development. Information to assess technology readiness comes from the NASA Technology Readiness Level. Decision-makers should consider TRL when selecting from the most promising strategies on the supply curve, both in terms of risk of achieving the expected outcomes, and time to achieve those outcomes.
For a strategy to be implemented, the system in which the implementation will occur, what implementation will change in the system, the process of implementation (who is responsible, the mechanism (regulation, market, etc.), and funding) need to be understood. The framework calls for this information to be included in the decision making so that planning, including risk assessment, of the expected approaches or alternative approaches for implementation can be done. The performance metrics for the evaluation of the success of implementation also need to be identified in the framework. If any of this information cannot be identified, then the strategy is not viable.
As noted at the introduction to this section of the paper, this methodology was based on the experience of the authors supporting local and state government decision-making, and was intended to be used by state and local government to quantify changes in environmental impact from proposed changes, and help them identify, document, and analyze a proposed path for implementation. The example case studies presented in this paper were analyzed at the proposed implementation scale for the state of California. Smaller scale analysis should be done for specific local governments, as the quantitative information needed for the approach is often context-specific. The proposed methodology was not intended to be used for simple scaling up for integrated national or international policy development. However, it is likely that this approach of quantitative evaluation of proposed changes at the scale of implementation could provide useful information for a bottom-up building of integrated policy plans, provided appropriate inter-system effects were considered.
6. Summary and Critiques of Supply Curves
Governments and road agencies have goals for reducing GHG emissions and other environmental impacts and also face cost constraints. In democracies, there is a need to maintain public support for policies and practices to achieve these critical environmental goals by choosing the most cost-effective alternatives, and honestly, transparently, and effectively communicate the approach used in decision making, the expected benefits and costs, and the metrics for measuring the performance of the decision makers in delivering the results. Many potential changes in the policies and practices of road agencies are being proposed, both internally and externally. However, there is often a lack of quantitative information regarding the benefits and costs of these proposals, and a lack of definition regarding how the changes will interact in a larger system in which they will occur and their long-term effects, and who they will affect, which has equity implications.
A promising approach, called supply curves, that has been used at a national level for developing abatement strategies for GHG reduction with mixed results was proposed for use in this paper to compare strategies at the state and local levels. Some of the critiques of the past use of supply curves were being addressed through the use of the principles of consequential LCA and LCCA. Pilot studies currently underway for a large state road agency and local governments will provide initial feedback on the ability to use this approach at a conceptual level for initial prioritization of alternatives. Initial results indicate that sufficient data could be gathered in a reasonable amount of time to compare alternatives and that the results could be compared on a much more consistent basis than had occurred previously. It is apparent from work to date that a number of important assumptions need to be made, that need to be fully documented, and assessed for quality, for consideration in decision making.
Regardless of whether the reduction in GHG emissions is small or large, every government and industrial sector should find ways to reduce GHG emissions and at the lowest cost possible because cumulative contributions can bring a major change towards a cleaner environment. This paper gave a detailed methodology for developing supply curves for different GHG reduction strategies and presented two examples of strategy assessment for inclusion in a supply curve. A questionnaire was developed that could help in understanding the strategies and became the basis for the supply curve development. The results indicate that keeping the California road network smoother could reduce much more GHG emissions than installing PV panels and wind turbines on Caltrans right of ways. The results also indicate that there was a potential for a net reduction in Caltrans’ life cycle operating cost, dependent on the value of the electricity generated.
However, supply curves must be used with caution, and are only one of the tools available to support decision-making regarding GHG and other pollutant reduction, not the only one. A number of limitations of supply curves were identified, including omission of ancillary benefits of GHG emission abatement, poor consideration of uncertainty in the data, a lack of consideration of dynamic interactions over time, and a lack of transparency concerning their assumptions. Supply curves based on the individual assessment of abatement measures suffer from additional shortcomings such as not considering interactions, non-economic costs, and behavioral changes, as well as incorrect counting of benefits, and inconsistent baselines [66
]. It has been suggested that supply curves be used more for comparisons of alternatives than for quantifying cumulative progress to abatement [67
]. The ability of supply curves to predict future abatement has been critiqued because of the lack of considerations of longer-term changes in markets driven by consumer changes, the timing of policy actions, actions taken by other actors in the market, and changes in future technologies [68
]. Most of these critiques have focused on national-level supply curves, rather than more granular and often less complex curves for agency- and local-level curves, but they must be kept in mind when using supply curves to support decision-making.
These critiques were intended to be addressed somewhat by the use of LCA and LCCA approaches analyzed at the state and local levels, by the additional information that was intended to be gathered as part of the development of the supply curves, and in particular by the use of consequential LCA, which assumed that decisions would result in changes in the market rather than attributional LCA, which assumed that the market would not change.