A Performance-Based Highway Geometric Design Process (2016)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

49 There are certain guiding principles that should shape and define an effective process into the future. The research team characterizes these guiding principles in three categories—the funda- mental bases for road design, the social and public policy framework within which the process is conducted, and the necessary attributes of an effective process. These guiding principles are defined in the following subsections, and discussed in terms of their implications, opportunities for improvement, and potential barriers or conflicts. 4.1 Fundamental Bases for Roadway Design The following subsections represent what should be the fundamental bases for design of roads, streets, and highways. 4.1.1 Geometric Design Solutions Should Address Objective, Quantitative Measures of Transportation Performance Roadway design projects begin with a stated transportation problem. The purpose of geo- metric design is to provide the necessary three-dimensional framework for a road or highway to address the problem by providing the appropriate level of mobility and/or safety to the road users. Geometric design involves the application of tools, methods, dimensions, and criteria. Dimensional and other design standards and criteria are a means to an end. The end is trans- portation performance and such performance includes mobility, accessibility, safety, and state-of-good repair. Every phase, subprocess, methodology, or model developed and applied to highway design and highway design criteria should be objectively related to one or more measures of transportation performance. The implications of this guiding principle are threefold. First, the mentality of the designer must shift from a dimensional-based approach to a performance-based approach. The tradi- tional philosophical approach to design has been to treat minimum design criteria as adequate to produce an acceptable level of safety (Hauer 1999, Neuman et al. 2002). In traditional design practice, roads are implicitly characterized as being either unsafe or acceptably safe, and the application of minimum design criteria produces (supposedly) a safe highway. This traditional roadway designer’s mindset is that of nominal safety (green line in Figure 15). According to this “nominal safety” mindset, meeting the minimum criteria is all that is necessary or required of the designer. Indeed, designers are discouraged from providing more than the minimum values or dimensions as doing so is presumed to increase the construction cost while producing no added value. The notion of safety as being a fixed attribute (roads are either safe or unsafe) precludes consideration of marginal differences in safety performance associated with marginally different design solutions. To summarize, the nominal safety mindset requires the designer to meet the C h a p t e r 4 Guiding Principles for an Effective 21st Century Highway Design Process

50 a performance-Based highway Geometric Design process minimum dimension, but does not reward or incentivize the designer from using more than the minimum dimension. The application of a green line model (nominal safety) may have been necessary in the past, when safety knowledge was generally lacking, but today, knowledge has advanced such that design for substantive safety has become not only feasible, but desirable. Safety, as understood by the risk of a serious crash, is not an absolute. In fact, the nominal safety model ignores the nuances of marginal differences in performance associated with design values or dimensions. Fambro et al. (1997) postulated that substantive safety is a continuum (the orange line in Figure 15) when discussing the expected differences in crash rates associated with SSD values that differed from the minimum criteria. Those familiar with the science of highway safety understand that the general relationship of the orange line is intuitively adapt- able to other roadway elements and dimensions, including lane and shoulder widths, radius of curve, and grade. The mindset of the green line (the minimum design dimension per criteria) bears a very lim- ited relationship to either operational quality or safety performance for the dimensions associ- ated with a geometric design feature (with perhaps the exception of structural vertical clearance). Moreover, where the minimum design criteria per AASHTO policies lie with respect to actual safety performance was neither known nor applied in the setting of many criteria in the past, with very few exceptions. This is because much of the geometric criteria was, given the state of knowledge at the time, established without knowledge of safety performance, or based on rela- tionships or models unrelated to safety performance. Selecting an arbitrary green line for values associated with a geometric design element does not serve a geometric design process aimed at optimizing performance within a cost-effective design framework. Highway engineers and designers applying design criteria need to understand not just the criteria but the basis behind the criteria. The revised design process envisions that designers will need to approach their work with the performance-based (orange) mindset rather than the dimensional-based mindset of design. Their design objective should be not just to apply a minimum criterion, but rather to find the appropriate combination of criteria that provides Figure 15. Two mindsets for safety in design.

Guiding principles for an effective 21st Century highway Design process 51 optimal performance for an acceptable implementation cost, such combination and cost reflect- ing the specifics of the context. The second of the threefold implications of the guiding principle is that the geometric design process must be viewed as incorporating continuous change. The advances in our knowledge of safety and operational performance should be routinely input to the geometric design process. The design process should readily and seamlessly adapt to the changes in external circumstances (e.g., advances in vehicle technology and changes in the vehicle fleet, social policies, and priori- ties regarding funding of transportation infrastructure). Each project does not exist in a vacuum, but rather contributes to the local, regional, and statewide transportation network. The design process must be adaptable to the evolution of societal investment decisions on highways and indeed all transportation infrastructure. The third implication combines the first two. The design process must transition to one that directs the optimization of a solution (in performance-based terms) within the context of resources allocated to the project. This may be viewed as a cost-effectiveness approach, but an important consideration is that resource allocation should be considered in a life-cycle sense. In other words, the design process should directly consider not only the initial capital cost of the project, but also the long-term O&M costs of the implemented solution. Such costs impose per- manent financial liabilities on the owning agency, and as such produce long-lasting constraints on the agency’s ability to maintain the performance of its system. Limitations in resources pro- vided to the agency must influence the designation or determination of a cost-effective design dimension or solution. This is the essence of recent project development approaches taken by agencies such as the Washington State DOT and the Missouri DOT. With respect to geometric design criteria as put forth by AASHTO, there are significant oppor- tunities to update and revise subprocesses and geometric design models to make them more per- formance based. Table 7 summarizes the results of an analysis of geometric design criteria in the 2011 Green Book (AASHTO 2011a). Although some design criteria have been formulated using transportation performance inputs, many others have not. Moreover, basic models that are theo- retically performance based are either outdated, overly simplistic, or unproven in their relationship to actual performance. A truly performance-based design process cannot be achieved without a comprehensive review and overhaul of many of AASHTO’s geometric design approaches and models that have been unchanged for more than 70 years. Chapter 3 of this report provides an overview of potential new approaches to the development of geometric design criteria. 4.1.2 The Geometric Design Process Should Explicitly Address All Potential, Legal Road Users Roads and road corridors are now understood as serving transportation needs of not only motor vehicle drivers and passengers (including motorcycles), but also cyclists and pedestrians. The geometric design process should direct the evaluation of the need to provide transportation service for each user type, and the spatial and operational design requirements to serve all users. Addressing all legal road users does not merely mean providing space for them within the roadway or right-of-way. It also means designing a corridor in recognition of the unique risks and performance needs of the users. In addressing all legal road users, the design process must provide a means or process for prioritizing what may be conflicting operational needs. Not all roads serve or should serve all road users to the same extent. The context should define what types of users should be expected or explicitly addressed, and the relative importance of each user type. An important aspect of the suggested geometric design process is the recognition of what user types should be included under which contexts. For example, agencies may explicitly choose to exclude consideration of pedestrians and bicycles within certain corridors such as

52 a performance-Based highway Geometric Design process controlled access facilities. Another example may be the design of a parkway for which trucks are prohibited. Finally, pedestrian-only corridors or transit-only corridors are solutions applicable to many urban contexts. There are many implications of this guiding principle. Perhaps the most important is the clear difference in providing for transportation performance—both safety and mobility—for vul- nerable users versus those traveling in motor vehicles. Historically, AASHTO geometric design approaches and criteria have evolved based on meeting the user needs of those traveling in motor vehicles, with the primary transportation value being vehicular mobility. Providing for perfor- mance in motor vehicle mobility historically translated to designing roads for as high an operat- ing speed as was reasonable given the context. Indeed, highway engineers have learned to equate design speed with design quality (the higher the better). In a multimodal corridor serving not only motor vehicles but also pedestrians and bicyclists, higher motor vehicle speeds may be incom- patible with the mobility and safety needs of such vulnerable users. A future geometric design process must provide a direct means for design engineers to produce a high-quality design based on whatever is an appropriate speed given the context and composition of the traveling public. In some cases, this may mean taking explicit actions to produce lower, not higher, vehicle speeds. Another implication (linked to the above guiding principle) is the difference in operating per- formance associated with the wide range of legal vehicles on the road network. Differences in operations and safety performance associated with passenger cars versus larger, heavy vehicles (single unit trucks, tractor-semi-trailers, buses) are well documented in the literature (Harwood et al. 2003, Fitzpatrick and Wooldridge 2001). One important challenge of a revised process is the determination of how to best serve the needs of the full range of vehicle types that differ so much in their individual performance characteristics. This may mean, for example, using a vehicle other than a passenger car as the basis for horizontal alignment for some contexts. Such an approach is really not new. For example, the operational bases for roadside barrier design are routinely changed as the vehicle fleet changes, with varying combinations of vehicle type, size, and weight applied first in the NCHRP Report 350 design guidelines (Ross et al. 1993), and now with the Manual for Assessing Safety Hardware design basis for barriers (AASHTO 2011b). 4.1.3 The Geometric Design Process Should Integrate Operational Solutions with Geometric Elements Solutions to performance problems include features or concepts other than roadway geome- try. Operational solutions include traffic control, signing and warning devices, traveler informa- tion using intelligent transportation systems (ITS), and real-time demand-responsive systems to help regulate speed and reallocate signal priorities. Operational solutions are now evolving that incorporate in-vehicle technology. The maturation of ITS and the ability to operate a corridor or network in real time offers benefits to highway agency customers. It also offers opportunities to adjust approaches to geometric design. (See Figures 16 and 17.) Consider, for example, the traditional rationale and requirements for full-width shoulders on urban freeways: shoulders provide space for emergencies, enforcement, and maintenance, all of which drive the width dimensions in AASHTO Policies. But with the ability to monitor traffic continuously via ITS and apply automated enforcement where enabling legislation exists, the need for full shoulders on some corridors and in some contexts may be greatly lessened. More- over, an optimal design solution may be the reallocation of the space previously allocated for full shoulders to travel lanes. The qualitative trade-offs associated with an urban freeway cross section with full shoulders versus one without full shoulders that incorporates ITS technologies are further illustrated in Figure 18. These qualitative trade-offs will vary in magnitude and relative importance; they will

Figure 16. Value proposition: reliable travel time. Source: http://www.wsdot.gov/smarterhighways/vsl.htm Figure 17. Smarter highways variable speed limits. 3 Lanes plus Full Outside Shoulder 4 Lanes with Full Outside Shoulder 4 Lanes with No or Minimal Shoulder 4 Lanes with No or Minimal Shoulder and ITS Mobility (Capacity) Safety Performance Maintenance of Roadway Cost Maintenance of Roadside Cost Operaon (Incident Management) Snow Removal Cost Law Enforcement Cost to Construct *Colors represent qualitative rang, with green being best, yellow next, and red worst for a given attribute Freeway Cross Section Alternatives* Figure 18. Qualitative demonstration of trade-offs for cross-section alternatives incorporating varying lane and shoulder width strategies and ITS.

54 a performance-Based highway Geometric Design process or should be quantifiable in any geometric design context. The point of this comparison and qualitative analysis is to demonstrate that incorporation of operational solutions or features can and should be part of the geometric design process and decision-making process, rather than a separate or after-the-fact process. Under the current geometric design process, the conversion of shoulders to travel lanes, or their use during certain time periods, such as peak periods (e.g., on Interstate I-66 in Northern Virginia), or for transit vehicles as is shown in Figure 19, require design exceptions. In the context of resource limitations, highway agencies are employing ITS solutions to achieve mobility and safety perfor- mance goals. The future geometric design process should enable and indeed encourage an agency to take full advantage of operational solutions to adjust its approach to geometric design decisions; and the design process should support such approaches rather than treat them as exceptions. Similarly, a run-off-road safety problem at a horizontal curve may be addressed by curve flat- tening, shoulder widening, or a combination of the two; but it also may be addressed by advanced speed warning or special high-friction pavement, particularly if the crash history suggests wet or icy pavement as a contributing factor. If such less costly solutions reasonably address the stated problem, the design process should lead to their acceptance. Such solutions should be considered part of the geometric design, not as an alternative to geometric design. The geometric design process should be more broadly defined as a design and operational solutions process. The preceding discussion is intended to be illustrative. Fundamental to a successful process is the ability to quantify the benefits and costs of incorporating (or not incorporating) ITS solu- tions as part of the alternatives evaluation process. 4.1.4 The Design Process Is Forward Looking Road infrastructure is designed and constructed to accommodate the future transportation needs. Right-of-way purchased for transportation is removed from other productive uses. Well- constructed and maintained roads and bridges should last 75 years or more. Right-of-way acquired may be the maximum available into the future, depending on how the adjacent land develops and matures over time. The value and utility of right-of-way will last for 100 years or more. This guiding principle is not new, but its implications are often overlooked. A forward-looking process contains built-in uncertainty. Actual traffic will differ from travel demand forecasts, Figure 19. Transit service operation on freeway shoulders in Minneapolis.

Guiding principles for an effective 21st Century highway Design process 55 unforeseen changes in the context will occur, and roads designed using models will not always perform as expected or intended. Such uncertainty cannot be overcome with additional data, greater investment in technology, or more detail in the engineering. The amount of uncertainty and the implications of different futures emerging will vary widely based on the local and regional context. Major portions of the U.S. have experienced little or no growth for 30 years or more. The land use in many urban areas is fully built out and stable. Also, certain project types are inherently more uncertain in their outcomes than others. • New roads on new alignments are clearly the most difficult to forecast the future. New roads will generally affect land development patterns in very significant ways. They will change com- muting, freight movement, and other travel patterns. Major new roads may significantly spur regional development in ways that influence the road network beyond the project itself. Safety prediction and operational models based on research elsewhere may be applicable to new roads, but the extent to which actual outcomes mirror modeled ones is inherently uncertain. • Reconfigured and reconstructed roads may produce some level of uncertainty in future traffic patterns if they include substantial increased capacity. Such projects may result in intensifi- cation of abutting land use, and regional travel pattern shifts. However, reconfigured and reconstructed roads will generally produce more predictable and less severe changes in traffic patterns than new roads on new alignment. • Reconstructed roads in their current configuration and 3R projects will generally produce little measurable change in patterns or volume of traffic. Post-construction outcomes will be more predictable from a safety and operational perspective. A significant design process challenge regarding future uncertainty is the conflict between the typical nominal design year for infrastructure and the useful physical life of such infrastructure. Highway projects are sized based on forecast travel demands, which reflect forecasts of land use, local and regional socioeconomic changes, evolution in social trends related to transportation, and national and local policies. Travel demand forecasts typically reflect a nominal 30-year timeframe from initial studies, which represents a consensus maximum ability to forecast the cumulative effects of the above factors on travel demand. Complex and/or substantial projects can take up to 10 years or longer from the original planning phases to construction, thus frequently resulting in a project sized to meet as little as a 20-year future post-construction time. This contrasts with the useful life of newly constructed pavement, bridges, and other infrastructure, which is at a minimum of 50 years and, depending on many factors, may be as great as 100 years. Indeed, much strategic research is being conducted by AASHTO and the FHWA on long-lasting pavements and bridges (FHWA SHRP 2 Solutions). This strong emphasis in future infrastructure policy would further widen the gap between a nominal design year under current processes and the actual physical useful life of road infrastructure. Based on the complexity of the project, a risk assessment of the forecast traffic should be conducted. The geometric design process must address and reconcile differences in the timeframes asso- ciated with the ability to forecast future demands and needs and the intent of infrastructure investments to be more long-lasting. This is especially true for new construction and recon- struction projects. 4.1.5 The Design Process Must Be Context Sensitive to the Extent Possible The experience of designers working with stakeholders in the CSS environment was captured in NCHRP Report 480 (Neuman et al. 2002), Context as broadly defined varies widely. Every project is unique. The current AASHTO context framework is composed of area type as defined

56 a performance-Based highway Geometric Design process by urban and rural; three levels of functional classification; and three levels of terrain is insuf- ficient to fully capture all relevant context factors. The geometric design process needs a more robust set of context definers, and the design process should support or direct design decisions, dimensions, and choices consistent with the defined context. In particular, context definers are needed that differentiate how the full range of legal users should be prioritized. Research supports the intuitive notion that the type and intensity of land use strongly influences the propensity for pedestrian activity. Explicit land use policies and ordi- nances employed by cities are now common in promoting walkable or pedestrian-friendly envi- ronments (Association of Metropolitan Planning Organizations, nd; National Conference of State Legislators, nd). The road design process must encompass such land use policies and actions. Potential new approaches to a more robust context framework are presented in the following subsection. 4.1.6 The Design Process Must Be Financially Sustainable at Both the Program and Project Level The design process should recognize both the initial and long-term financial implications of the constructed solution. Initial capital costs represent one aspect; agencies must commit resources for ongoing maintenance and operational costs of the project as well. The design process should be sufficiently flexible such that the agency can execute it over time under a range of available resources. The typical challenges as related to the sustainability of the highway trust fund is demonstrated in Figure 20, highlighting a short fall in highway trust fund revenues and evaluation of policy level solutions to address the issues. A practical design approach is one that delivers value within the budgetary limitations of the agency. Each project is not independent of other projects completed by the agency. Design deci- sions that impose immediate and long-term costs will influence the ability for other projects to be performed either at all or in an optimal manner. As an example, if 90% of the identified problem could be solved for $1 million and 100% will require $10 million, the less than perfect solution should be considered. 4.2 The Design Process Must Be Conducted Within the Prevailing Social and Public Policy Framework Highway design and construction were considered a core function of government. Construc- tion funds were generated through taxes, fees, and appropriations at the federal and state levels. The legal system provided the principal of eminent domain, enabling the acquisition of neces- sary right-of-way for construction. Highway engineering professionals operated within the governmental framework. When the highway design profession came into being, and road design and construction emerged in the early 20th century, the road design process was purely technical. Highway engineers developed technical tools, methods, and models independent of external inputs. The process was wholly a technical, internal process. The technical disciplines included geometric design, pavements and materials, bridge design, geotechnical engineering, hydrology and drainage, and construction. Accountability for finished projects was through budget and schedule completion. The work of professional engineers was considered beyond questioning, and their recommendations or requirements were generally unchallenged. As roads were built, the political landscape changed in the U.S., and important societal issues and external pressures emerged. The influence and interaction with nonprofessional stakeholders

Guiding principles for an effective 21st Century highway Design process 57 began to influence project development. The progression of questioning and input by these stake- holders lead to the passing of the NEPA and other subsequent legislation. The road design process is no longer an independent, wholly technical engineering process. Projects that involve public allocation of resources, acquisition and use of public rights-of-way, and provide a core value or service to the general public are considered to be political projects. A guiding principle behind the 21st century geometric design process is the explicit recogni- tion of a social or societal public policy framework that influences and directs the process and its execution. 4.2.1 Accountability and Responsibility Every project is the responsibility of an owning agency, which is typically a unit of state or local government. All decisions regarding every project must be made or endorsed by the entity that is given the responsibility, including the funding and other resources for design, construc- tion, operations, and long-term maintenance. The accountability for the outcome of a proj- ect rests with the policymakers who direct or manage the transportation agency. Furthermore, accountability must reflect objective measures of performance (safety, mobility and accessibility, Source: Joint Committee on Taxation, Long-Term Financing of the Highway Trust Fund, (JCX-92-15), June 15, 2015. Figure 20. Sustainability of highway trust fund.

58 a performance-Based highway Geometric Design process and lasting functional value), resource allocation (funds allocated for construction and mainte- nance), and meeting expectations regarding delivery (timing and schedule, direct and indirect effects of construction on others). Such accountability measures apply at the programmatic and individual project levels. At the individual project level, the notion of accountability rests with the responsibility of knowledgeable, properly trained, and professionally licensed design professionals to undertake or oversee the project. The process can be made more rigorous; and it can employ more technical models, methods, or analytical procedures. But the geometric design process will always involve some measure of uncertainty, require assumptions where data are incomplete or unavailable, and require judgments that must reside with those fully trained in all aspects of the road design process. External stakeholder involvement in highway design project development emerged with the CSS/design initiatives in the 1990s. The value and necessity of input from external stakeholders is unquestioned; it is considered essential. However, input and involvement should not be con- fused with ultimate responsibility and accountability. Highway engineering and design decision making are and should continue to be the responsibilities of professional engineers and other licensed or similarly qualified professionals who are engaged on behalf of the owning agency. In many countries, and now in certain states within the U.S., there are roads or networks for which the design, construction, and O&M may be outsourced to private entities for consider- able lengths of time. Privatization does not eliminate the public nature of the infrastructure; it merely changes the manner in which the transportation service is delivered. The ultimate public owning agency communicates its values and establishes accountability through the contracting terms of the privatized facility and the extent to which the owning agency monitors and enforces its contracts. 4.2.2 Legal Framework There are five distinct aspects of the legal framework in the U.S. that affect the road design process. These are tort laws and civil actions involving professionals and agencies, laws and regu- lations associated with licensing and enforcement of traffic laws, the environmental and social regulatory process at the federal and state levels, the regulation of motor vehicle manufacturing and sales, and governmental policies that direct resources and their use toward transportation facilities. 4.2.2.1 Tort Laws and Professional Liability The legal framework includes the establishment of laws and procedures for holding pro- fessional engineers accountable for their design, construction, and maintenance efforts. The acceptance of professional liability for design errors or omissions is a centerpiece of the design process. All states require that design plans be sealed by professional engineers licensed in the state in which the project is to be constructed. An important element of the prevailing tort law framework is the liability associated with types of actions that engineers and agencies undertake. Generally, actions involving design decisions are considered discretionary in nature, and as such are typically immune from tort actions, as long as the professional engineer and agency apply current practices, implement agency policies, avoid errors or omissions in their work, demonstrate appropriate professional judgment, and fully document their work (Glennon 1996; AASHTO 2004). Professional engineers must employ a process that reflects best practices of the profession. Such best practices are science-based and transparently published in peer reviewed research and

Guiding principles for an effective 21st Century highway Design process 59 other reports. Best practices change over time as the knowledge base and technology improve- ments come into being. Agencies must commit in an ongoing manner, at a minimum, to moni- tor advances in knowledge and ideally sponsor and conduct research to continuously improve the knowledge base and resulting design process. The road design process is sufficiently complex, specifically with respect to the context of the road, such that a simple, formulaic, or rote design process is not possible. Professional engineers cannot rely solely on criteria or dimensions, but rather must exercise judgment and be accountable for those judgments. 4.2.2.1.1 Design Exceptions. The concept of design exceptions, which evolved in response to the combination of loss of sovereign immunity in the 1950s and difficulties in applying pub- lished design criteria to every project, is a special aspect of designer discretion and agency risk. Design exceptions have become integral to the geometric design process. In many project types they are routinely applied. Design exceptions involve the use of a dimension that is outside the applicable dimension for the context as understood by functional classification, area type, and design speed (Neuman and Stein 2007). The need for a design exception is intended for special circumstances such as environ- mental conflicts, terrain, or other features that may preclude application of applicable geometric design criteria. The purpose of a design exception process is to document where a designer could not or need not apply the relevant dimension. By documenting the exception the designer avoids the possibility of the design being interpreted as an error or misapplication of design standards, a situation that could arise in the future should a crash occur and that could be linked to the design feature in question. Such documentation would demonstrate the professional engineer’s unique dilemma, the design choices considered, and effects of such choices, and the engineer’s judgment as to the appropriate solution for the context. (Absent such specific documentation, the only evidence in an adversarial tort setting of an agency’s design decisions are the contents of the road plans. Should the plans include a design dimension outside that of the applicable design policy, a judge or jury could reasonably interpret the plans as containing an error, in which case a potential finding of negligence may result.) A basic problem with highway designer decision making regarding design exceptions is the presumptive linkage between the criteria and safety performance (see Figure 1). When a design exception is considered or applied, the designer bears the burden of proving that the exception will not adversely affect the safety performance of the road or that whatever compromise in safety occurs is necessitated by an unavoidable impact. Designers also are expected to mitigate the potential adverse consequences as well (Neuman and Stein 2007). Risk management poli- cies employed by most agencies stress that design exceptions should not be justified based solely on construction cost savings, but rather on other quantifiable effects such as right-of-way or environmental consequences. Design exceptions are generally recognized as increasing tort liability risk to agencies, increas- ing the time and cost to reach a decision, and producing less than optimal design outcomes. Some designers believe that design exceptions are indicative of failure, or that they will produce an inherently inferior design, or that the agency or even individual will be open to a tort lawsuit should a crash occur, even if documentation is produced. Risk management processes that are employed to review, approve (or not approve), and document design exceptions can increase project development time and cost. Design records maintenance for future reference also is vital for an agency, creating an additional administrative cost. The design process did not always require design exceptions. They evolved in response to the voluntary ceding of sovereign immunity by states, with the resulting need to provide defense against tort suits. Research has demonstrated both the frequency and routine application of

60 a performance-Based highway Geometric Design process certain design exceptions (Mason and Mahoney 2003), and the lack of widespread adverse safety performance consequences when these are appropriately applied (Stamatiadis et al. 2005). A geometric design process that routinely incorporates exceptions to what are supposed to be appropriate best practices, dimensions, or standards is suboptimal. Such a process over- emphasizes the physical dimensions of the roadway, which are not the end but the means to the end, which is performance. Where design exceptions are routinely considered and granted, this demonstrates either inappropriate geometric standards, lack of creativity by the designer, or insufficient flexibility in the application of the standards. For example, shoulder widths on freeways, SSD (or length of vertical curve) on two-lane roads, and lane widths are often designed as exceptions on reconstruction projects (Mason and Mahoney 2003). Despite the firm, stated minimum 12-foot lane widths for freeway lanes per current AASHTO policy and FHWA Inter- state standards, less than 12-foot lanes are so prevalent that there was sufficient data in recent research on safety performance of freeways and interchanges to enable establishment of the relative crash risk of lane widths less than 12 feet (Bonneson et al. 2012). Agency designers and approvers involved with such projects demonstrate an implicit under- standing of the lack of risk (or lack of cost effectiveness in applying the full criteria), yet the design criteria remain unchanged, and the design process requires such treatments as exceptions. A design process that treats published dimensions as firmly associated with some level of safety deemed appropriate (and that invites the inference that an exception will not provide that appropriate level of safety) ignores the reality of safety performance related to design elements and context. It needlessly increases project development time and cost, and unnecessarily injects a measure of tort liability risk (associated with incomplete documentation, or documentation lost or not retained when needed in future years). The geometric design process should produce an outcome that is optimal given the con- text, which is the desired expectation, and as such should not need to be labeled or considered an exception. The process by which the geometric dimensions and elements are obtained and included in the design should suffice to establish the basis for the design. If done properly and documented completely, an ideal geometric design process should not require an exception pro- cess, but rather a complete and thorough optimization analysis and documentation. 4.2.2.1.2 Ministerial and Mandatory Duties of Highway Agencies. Design decisions are discretionary in nature and generally immune from tort actions when performed and docu- mented properly. Maintenance functions, however, are typically considered ministerial or mandatory, and present a different risk profile to transportation agencies, which are at risk of successful tort actions if they fail to maintain and operate their road systems in a reasonable manner, consistent with their agency’s policies. The importance of understanding this aspect of the legal framework is twofold. First, it must be assumed that the tort law framework will continue, as it is imbedded in U.S. laws and legal traditions. Second, as many agencies transition to greater maintenance functions and fewer reconstruction or new construction activities, the importance of direct consideration of the potential impacts of design decisions on maintenance, and vice versa, becomes that much greater. Geometric design decisions can have a meaningful impact (positive or negative) on the difficulties and costs of maintenance. The design process and designers should understand and account for these impacts and not consider them merely the business of others within the agency who will be engaged only after design and construction is completed. There is no comparable exceptions process with respect to maintenance activities. Agencies may adjust certain mainte- nance policies to reflect budget constraints, but many activities must be undertaken regardless of costs or consequences.

Guiding principles for an effective 21st Century highway Design process 61 4.2.2.2 Driver Licensing and Traffic Law Enforcement Other aspects of the legal framework include laws and regulations associated with licensing of drivers, passage and enforcement of traffic laws, and definitions and restrictions on legally operable vehicles. The engineering profession must assume that driving will continue to be a highly regulated privilege in which drivers are repeatedly tested and licenses are issued, suspended, and revoked based on driver behavior and physical capabilities. Roadway designs should consider the needs of traffic enforcement. Similarly, regulations and laws at the federal level governing the characteristics of motor vehicles—their performance in crashes and dimensions—is a key element of concern. The legal framework associated with licensing and operation of vehicles allows the design profes- sion to make reasonable assumptions for the purposes of designing and operating infrastructure. The onset of what is referred to as driverless vehicles is acknowledged. How this new technol- ogy will influence geometric design remains to be seen. It seems clear that the road system will continue to be driven with human input for many years to come. Some corridors or special roads may, over time, be fully automated but what is more likely is the evolution of a vehicle fleet that is highly interactive with roadway infrastructure to limit or minimize the adverse effects of driver errors. To the extent that future driverless technology will include some component of roadway infrastructure (sensing), the vast size of the U.S. highway system suggests complete implementa- tion is decades away at best. The increase in bicycle usage and promotion of walking as a transportation mode for which design is necessary creates some challenges. There are typically no legal restrictions on bicycle riding in public roadways, by age or capability. There also are no limitations or legal require- ments on walking speed or capability. Indeed, the Americans with Disabilities Act (ADA) legally sets requirements for accommodation of pedestrians and in particular disabled and blind pedes- trians. In both instances the geometric design process may require tailoring to the unique demo- graphics of the area in which a project exists. 4.2.2.3 Environmental Regulations and Laws A third aspect of the legal framework is in the environmental regulatory process that exists at the federal and state levels. Laws and regulations governing the processes by which environmental issues are disclosed, studied, and influence the outcome are now integral to the geometric design process. Early versions of the AASHO Policies on Geometric Design were published prior to the passage of such laws and regulations. Multiple governmental agencies have regulatory approval over various details of projects. Their input will shape the design solutions that emerge during the design process. States also may pass regulations or policies requiring accommodation of pedestri- ans and cyclists for certain roads or contexts; the ADA clearly influences the road design process. This aspect of the legal framework emerged after the early years of road building in the United States, and most notably after initial design and construction of much of the Interstate System. The environmental regulatory framework that evolved following passage of NEPA in 1969 is now widely recognized as central to project development, and its permanence as part of the geometric design process must be acknowledged. NEPA and subsequent legislation place specific requirements on the highway design process, and the environmental and social performance outcomes of a design. The geometric design process now includes any number of analysis tools, data, and, in many cases, infrastructure to address regulatory issues. AASHTO has long acknowledged the social, economic, and environmental issues in design in the foreword to its policy documents. However, the road design criteria and process has yet to fully incorporate what are not merely values but, in many cases, legal or regulatory requirements. An array of technical agency stakeholders has the responsibility of assuring that the laws and regulations referenced in Table 8 are adhered to properly. These technical agency stakeholders

62 a performance-Based highway Geometric Design process play a direct role in the geometric design process that is no less important than that of the high- way design engineers. Again, what often results is a design exception to address a regulatory constraint or issue. An optimal geometric process must directly include both the identification and specific influence on the three-dimensional roadway footprint of applicable environmental regulations and laws. And, it should not label design outcomes driven by legal or other formal policy issues, which are important elements of the project context, as being exceptions. 4.2.2.4 Laws and Regulations on Legal Motor Vehicle Manufacturing and Sales A fourth aspect of the legal framework is the progression of requirements on motor vehicle manufacturing and sales in the U.S. There are strict standards for vehicle performance, for incor- poration of features such as seat belts, airbags, and side-impact protection, for performance in a collision (i.e., protection of the passenger compartment), for both passenger cars and trucks. Such regulations have had measurable impact on the survivability of crashes that formerly pro- duced fatalities. Other aspects of vehicle regulation include height and width limitations, which influence vertical clearance and roadway or lane-width criteria. Vehicle attributes and their contribution to occupant safety and traffic operational efficiency have evolved significantly over the years. Features such as automatic braking systems, improved tires, and better suspensions mean the operation of the fleet is significantly different than it was 50 years ago. With the exception of a few variables, such as driver eye height, AASHTO geo- metric design policies have remained unchanged in the face of advances in vehicle technology. The practical effect of motor vehicle regulation on geometric design is to influence (reduce) the severity of crashes that occur. A more performance-based geometric design process account- ing for such effects may result in less costly infrastructure design, as the vehicle technology itself Table 8. NEPA legal and regulatory “umbrella.”

Guiding principles for an effective 21st Century highway Design process 63 produces the intended driver/vehicle/roadway system performance. This also may be an out- come produced by self-driving vehicles. The evolution of self-driving vehicles is rapidly occurring. The viability of the technology is now proven, but how such technology is transitioned within the vehicle fleet and the roadway infrastructure is uncertain. Claims of substantial improvement to safety performance are made with respect to many of the in-vehicle and infrastructure technologies. Such claims are associ- ated with eliminating driver errors, which are a contributing factor in over 90 percent of crashes (AASHTO 2010). Vehicles that are truly driverless and error-free presumably would negate geometric require- ments based on human needs for sight distance, extensive roadside infrastructure (guardrails, barriers, attenuation), and lane or roadway widths above a bare minimum. Such outcomes are years away, and the overall viability of truly driverless vehicles throughout the entire public roadway system is uncertain. A geometric design process based on performance outcomes, and that is continuously changing to reflect current conditions, will automatically adjust to real safety performance benefits that emerge. Crash modification factors or functions for geometric improvements would over time gravitate toward 1.0, meaning many roadway countermeasures would lose their effectiveness or have it significantly reduced. 4.2.2.5 Public Policy Resource Allocation The fifth and final aspect of the legal and policy framework is the allocation of resources to transportation agencies by their governing citizens, at the federal, state, and local levels. From the 1950s to the late 20th century, geometric design and road or highway design policies and standards at the national level have historically evolved independently of budgetary or resource limitations, or special transportation programs. The development of 3R criteria in the 1980s was the first attempt to recognize and adopt some measure of cost effectiveness in design decision making. The need for 3R criteria evolved over a time when many agencies began to transition from new road building to reconstruction of existing roads, with apparent difficulties in applying the published criteria. The evolution of the criteria involved some controversy, which mirrored the aforementioned conflicts in under- standing the difference between nominal and substantive safety. A major research effort was required, which produced TRB Special Report 214 (TRB 1987). This study focused on the need to relate road safety performance to physical dimensions or standards as a basic premise behind cost effectiveness in design. Approximately 15 years later, another similar controversy emerged in the design profession that had its roots in dissatisfaction over the cost effectiveness of geometric design criteria. County engineers who had long relied on the AASHTO policies for design criteria of their road systems began to express discontent with the costs of the resultant designs and lack of apparent value. County engineers did not have access to federal funds for their systems and, in many cases, the overall funding available to them was far short of that available to their state DOT peers. County road systems are lower volume in nature, with substantial mileage of lower classification facilities. County engineers did not have the resources or organization to develop their own design cri- teria, and so approached the American Society of Civil Engineers to conduct a study and make recommendations for revised design criteria for VLVLR. The criteria were developed specifically to recognized principles of cost effectiveness and risk associated with lower-volume, lower-speed roads. They were intended to promote the concept of requiring less construction than higher- volume, higher-class facilities. AASHTO intervened and agreed to perform such a study, which was completed in 1998 (Neuman 1998). That resulted in the eventual completion and adoption by AASHTO of criteria for VLVLR (AASHTO 2001), which is referenced by many county engineers.

64 a performance-Based highway Geometric Design process As a final example, consider the concept of the design domain, which is evident in design practices in many Canadian provinces. This principle acknowledges the practicalities of selecting design dimensions for a reconstruction road project to be in concert with the prevailing dimen- sions of the road and network to which the reconstructed road will be attached. Widening or upgrading a roadway to a higher standard when it will transition to the older standard may not make sense from both an economic or performance basis. In the cases of 3R criteria development and VLVLR criteria development, the geometric design process was disrupted or altered only following a widespread acknowledgment of a process flaw (i.e., the judgment that the process and geometric criteria were not producing systematically cost-effective solutions and were placing undue resource burdens on agencies using the criteria). The lesson of these initiatives is that the roadway geometric design process should expressly acknowledge the inherent limitations in resources provided the agency by whatever funding programs or mechanisms are established. Such limitations include the amount of funding, designation of specific programs to be completed, allocations by geography, and other policy restrictions. A road design process that is conducted at the project level independently of overall program funding limitations cannot be assured of being sustainable over the long term. The process should continuously be tested and adapt to resources and resource limitations. Many of the management initiatives begun in recent years at the state level (most notably, the concepts of practical design and the Washington State DOT’s design matrices) reflect an under- standing that the historic design processes were insufficiently sensitive to what were perceived as permanent or long-term resource limitations. Agencies can no longer afford a road design process that leads to solutions that are clearly unaffordable, or that requires the labeling of cost- effective solutions as special cases or design exceptions. 4.2.3 The Design Process Should Support the Financial Sustainability of the Agency’s Program The road design process should be aimed at providing the best possible solution for a road or project that is part of an overall jurisdictional network of transportation to be maintained and operated by the owning agency over an indefinite future. This will require a fundamental change in both programming and project-level design decision-making processes and approaches. Road design decision making in the U.S. has traditionally been driven by a least initial capi- tal cost decision model. Highway engineers design a road to meet a typically static, qualitative performance criterion using established design dimensions and criteria. The geometric criteria to a great extent define the right-of-way, cost, and other resources necessary for project con- struction. The criteria are typically unchanged or minimally changed from year to year, and are to be used in a consistent manner for all projects. Finally, the geometric criteria in most cases are derived independent of maintenance or operating cost considerations in a meaningful or quantitative way. A sustainable road design process is one that is aimed at minimizing the total life-cycle costs (construction and M&O over the total project life) within the context of achieving the purpose and need for the project (i.e., addressing the problems). In achieving the purpose and need, the project is expected to produce measurable societal transportation benefits over the total project life, such benefits to include lives saved, travel time reduced, and vehicle operating costs minimized. The agency’s investment in the project should be commensu- rate with these forecast or quantifiable societal benefits. Moreover, a sustainable geometric design process is one in which the performance criteria that drive the geometric design reflect

Guiding principles for an effective 21st Century highway Design process 65 the practical limitations of both the project’s context and the available resources set by the agency’s policymakers and sponsors. Transportation agencies do not just design and construct roads. They incur other substantial costs that relate to the functioning of their road system. Agencies also are responsible for the maintenance and upkeep of each project and their overall transportation system. For many agencies a substantial part of their annual budgets are now allocated to 3R, O&M functions, with actual new road construction and reconstruction expenditures much less. For example, a recent Government Accountability Office Report documented that 90 percent of FHWA’s 2013 fiscal year obligations were spent on reconstruction or 3R type projects, and only 10 percent on new road construction (U.S. Government Accountability Office 2014). Figure 21 shows the 2013 budget for the Missouri DOT, which is a typical midwestern state primarily rural in character but with two major metropolitan areas (Kansas City and St. Louis). Maintenance costs are almost 20 percent of the total DOT budget, and more than one-third of the budget for construction (Missouri Department of Transportation nd). Geometric highway design decision making results in incurring costs to produce the measur- able benefits. Different design solutions may create different or unique challenges or require- ments associated with M&O. Once a project is constructed, the M&O requirements, whatever they may be, become necessary outlays forever by the owning agency. A geometric design decision process that ignores impacts on the M&O costs, particularly when these may be highly influenced by the geometric design itself, robs the agency of the ability to manage the long-term sustainability of the project and its system. The evaluation of M&O functions and costs and their incorporation into project development and design decision making is thus an important recommendation for a financially sustainable road design process. 4.3 Attributes of an Effective Geometric Design Process The following are core attributes of an effective geometric design process: • The process must be efficient, • The process should be scalable, Construcon 53% Maintenance 19% Fleet, Facilies & Info Systems 3% Debt Service 13% Administraon 2% Other State Agencies 10% Figure 21. Allocation of 2013 annual budget for the Missouri DOT.

66 a performance-Based highway Geometric Design process • The process must be executable by properly trained professionals in a consistent manner, • The process should be transparent, and • The process must be defensible. 4.3.1 Efficiency Efficiency refers primarily to the data required and tools or methods that are core to process execution. Advances in technology now enable the efficient use of methods, tools, and approaches that were uneconomic 10 to 20 years ago. It should be expected that continued advances in both knowledge and data gathering and maintenance will produce greater efficiencies over time. The advances in computer-aided design and data gathering technologies offer the greatest ben- efits to a new design process. From the early days of design and construction to the early 1980s, the vast majority of time, labor, and expense associated with road design development was associated with the technical delivery of engineering construction drawings and associated documents such as specifications and construction quantity estimates. Calculations by hand were tedious and sub- ject to error, requiring additional time for independent checking. Actual design plans including individual cross sections had to be hand drawn. Even minor revisions required costly significant re-engineering effort and redrafting. In this environment, resistance to considering design alter- natives or continuously adjusting a design based on stakeholder comments were understandable, given the time and cost of completing the plans and specifications for construction. Some of the time and effort to develop highway engineering design drawings is now devoted to application of other technical processes, models, and tools that are either new to the pro- fession or that represent significant advances in technology. Table 9 summarizes some of the important design tools used to support or characterize the safety, operational, or environmen- tal performance impacts associated with the roadway design process. Many of these tools were developed in response to the need to respond to requirements of the aforementioned laws and regulations. Others represent advances in knowledge that allow more quantitative and empirical knowledge to replace assumptions or mere qualitative judgments. Highway engineers now have the ability to apply a full array of design tools and models to both rapidly and completely address not only the geometric features of the road, but many of its performance characteristics. The actual engineering design process is so much more efficient today that the production of engineering drawings is no longer on the critical path with respect to a typical project schedule. The efficiency of any project is now measured by the extent to which all stakeholder inter- ests are properly accommodated, design alternatives appropriately vetted for their perfor- mance objectives, and the risks of unforeseen construction or other problems are eliminated or Safety Operations Environmental HSM Spreadsheets Highway Capacity Software TNM IHSDM Synchro MOVES CMF Clearinghouse Sidra EMFAC RSAP CORSIM ArcGIS RSAR VISSIM PBCAT Paramics ISATe Table 9. Summary of safety, operations, and environmental tools used in the design process.

Guiding principles for an effective 21st Century highway Design process 67 minimized. This is done through the appropriate, timely application of the full range of tools listed in Table 3, using readily available agency and project-specific databases. 4.3.2 Scalability The term scalable refers to a project’s size and scope, the complexity of the project and, to an extent, the capabilities and resources of the owning agency. Fundamental guiding principles apply to all projects, but the extent to which they create complexity and the need for multiple sub- processes or analyses will vary. For many projects, the cost or time for collection and evaluation of much data may not be justified based on the nature of the project, thus requiring assumptions or default values. What is important is that the geometric design process takes full advantage of the knowledge base regarding all aspects of the roadway infrastructure (design, construction, performance, and maintenance) and employs suitably sophisticated models and relationships for those projects or programs that require significant investments. Many agencies, particularly county and local governments, will be challenged by the data and technical needs associated with the design tools. Such challenges should not constrain the goal of a robust and data-centric performance approach. Rather, the needs of resource-constrained agencies can be met by careful development of shortcut procedures, programmatic data assump- tions, and approaches that reflect the agency’s context circumstances. 4.3.3 Executable The term executable refers to the need for the roadway design process to be successfully com- pleted by knowledgeable engineering professionals in a manner that is consistent within the agency. Properly trained and equipped design professionals using the same process and subpro- cesses should produce reasonably similar results, assuming the availability and quality of data are comparable. Similar results do not imply that the actual dimensions for road projects are the same, but rather that the manner in which solutions are studied and decisions made should be the same. Indeed, given that every location has unique context features, some measure of vari- ance in the physical dimensions should be expected. 4.3.4 Transparency and Defensibility The importance of transparency requires roadway design projects to involve both the alloca- tion of limited resources that could be allocated to other projects or programs and imposition of impacts to multiple stakeholders. The roadway design process is conducted in a public setting. Key subprocesses, data, and methods should be readily accessible to stakeholders and explain- able in terms that can be understood. The importance of transparency translates to a process that is defensible. A defensible design process may produce an outcome that some may find objectionable, but the process by which the design was completed and final decisions made should be defensible. Central to this attribute is the importance of all core technical models and subprocesses to have a firm, science-based, proven background. Defensibility includes the appropriate documentation of all data inputs, subprocesses, and value judgments made in reaching the design decision. An important aspect of defensibility is the ongoing protection of the agency against future tort actions based on allegations of error or negligence.