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.

25 3.1 Introduction This research addresses highway geometric design and the need for a new geometric design process. The first major change should be to acknowledge that “geometric design” as conven- tionally considered has little meaning absent the context in which the design is being completed. Geometric design has meaning and value only as it is applied to the context in which the designer is working—the geography, topography, land use, political, and environmental features—within and adjacent to the roadway in question. Textbooks and agency design manuals typically contain technical content describing how to calculate a feature such as an interchange ramp, or superelevated curve, or high point on a crest vertical curve. The geometry of angles, width and length dimensions, offsets, etc. is described. What is common to virtually all such treatments of geometric design is the absence of any back- ground context information. (Indeed, in some cases concepts may be presented in a simplified format that actually would never occur in a real setting, e.g., development of superelevation runoff on a zero grade roadway.) Geometric designers don’t work in a vacuum. To the contrary, they can accomplish nothing without first understanding completely the full range of conditions, resources, constraints, and opportunities with which they are confronted. So, for example, the development of a geomet- ric alignment for a freeway entrance ramp may require the designer to design the ramp along a horizontal curve, with the diverge portion of the ramp occurring in conjunction with a sag vertical curve, all of which is necessitated by the context. Similarly, guidance to prefer 90 degree intersection crossings may need to be ignored because the intersecting roadways occur at an 82 degree crossing. For the purposes of discussion, geometric design is thus better incorporated within the broader definition of highway project development, a term that includes many considerations beyond the mere geometry of the solution. Properly utilized, the technology available to highway engineers enables the appropriate geometry to be applied to the context. It is the total performance of the solution (and not its geometry per se) that should be of paramount concern. 3.1.1 Highway Design Decision Making Final decisions about the selected alternative and its specific design dimensions will be made by the owning agency, with input as requested from other agencies or external bodies. The deci- sion process will often require the balancing of competing interests and trade-offs. Measurable performance goals will help to balance the trade-off decisions. These may include, for example, raised medians and access management for better traffic operation and greater safety, but at C h a p t e r 3 Highway Geometric Design and Project Development

26 a performance-Based highway Geometric Design process the cost of access to businesses. A footprint that supports better operation through an intersection may require right-of-way acquisition and come at increased cost. The design process should be performed in a manner to provide sufficient data and objective information regarding the unique performance attributes of competing alternatives. The meaning- ful performance attributes of any road design solution can be categorized as follows: • Implementation costs (typically initial cost of construction); • Right-of-way acquisition or impact (total take, partial take, and type of land use acquired); • Traffic operational quality (for motor vehicles, pedestrians, bikes, transit riders); • Environmental effects (noise, air quality, wetlands, threatened and endangered species effects, cultural resources, etc.); and • Safety for both vehicles and vulnerable users (crash frequency and severity). Current practice has evolved such that all of the first four items are or can be characterized in objective ways using the best practices of the design profession. DOTs typically maintain thorough and up-to-date records of construction bids and right-of-way acquisition costs, and have sophisticated cost-estimating models based on quantity and other factors. The TRB HCM has historically been the basis for best practices in traffic operations analysis procedures; current practice has evolved to incorporate sophisticated traffic simulation tools such as VISSIM, CORSIM, and Synchro, which provide extensive quantitative performance data. A robust travel demand model can provide forecast data to get the best estimates of traffic operations for the alternatives under consideration. The traffic forecasts are as good as the assumptions that they are based on. Even in the case of environmental attributes, there are well established, and in some cases set by regulations, applications of predictive methodologies to quantify environmental effects. The fifth area, safety, has historically lacked a consensus method for comparing different alter- natives or evaluating highway geometry. It is common for project decisions on alternatives to be made without an attempt to quantify and compare the expected difference in safety performance. Any analyses that are performed typically fall into one of two categories: • The crash rate for an existing condition (“no-build”) may be compared with that for typical or average conditions in the state, with conclusions drawn regarding whether the existing condition is “safe” or “unsafe.” This analysis may often be used as part of a problem statement that may translate to a NEPA Purpose and Need statement. • A “nominal safety” analysis may be performed, in which an alternative’s adherence to full design criteria is considered sufficient to demonstrate that the alternative will be “safe.” The highway design process should recognize the sensitivity of costs and impacts associated with requirements for marginal dimensions, and should produce outcomes that assure any such requirements will produce measurable benefits and demonstrate cost effectiveness. Finally, most design decisions, while informed by objective measures, are typically both subjec- tive and not transparent and are also typically skewed to favor those project attributes that owners are most sensitive to and for which they have the confidence in the data provided. This is illustrated in Figure 3. Construction costs and right-of-way (both amount and cost) tend to dominate in decision making. Differences in operational and safety performance will most typically not drive the decision, with the latter, safety performance, often not well understood at all. 3.1.1.1 Design Documentation A final choice among multiple alternatives will ideally be fully explained to external stakeholders, be documented for future reference, and in the case of projects in NEPA be codified through a finding of no significant impact (FONSI) or record of decision (ROD). Many agencies also produce Design Study Reports that document the specific design criteria and design decisions.

highway Geometric Design and project Development 27 Design Study Reports are typically prepared in conjunction with environmental documents, and describe what is often referred to as 30% design in which the three-dimensional footprint of the road is set. Bridge studies (referred to as “type, size, and location,” or “TS&L”) may be included. Also included at this stage may be Design Exceptions or Design Deviations reports. Design documentation is an important part of the design process. Complete documentation of the design and reasons behind it serve to defend the owning agency should there be a future tort action alleging negligence. More commonly, final engineering design documents may be prepared by individuals or even entities (consultants) other than those who produced the 30% design. Final engineering should reflect and support decisions made in the process by others. 3.1.1.2 Final Engineering Design Once a preferred solution is selected, the geometric design process becomes more focused on providing details and additional features necessary for bidding and completing construction. This typically includes completion of all geometric and roadway details, traffic control plans, maintenance of traffic during construction, right-of-way plans, lighting and signing plans, utility relocation plans, guardrail and barrier plans, erosion control, pavement design, and construction specifications. During this process revisions to the preliminary geometric plans are common. These typically are associated with increased data obtained, constructability reviews, or final negotiations with affected stakeholders. In a well-executed project such revisions should have at most minor effects on the operational performance of the project that formed the basis for the decision. VE studies are required by FHWA for large projects and recommended for smaller projects. VE as typically integrated into the design process is another source of revisions during the final design phase. From the perspective of effort and cost, the great amount of detail required results in the final engineering stage costing three times or more than the cost of an alternatives study and preliminary design. At the close of this phase, the project will have advanced to a stage where it can be procured by the owner, most typically through a low-bid process involving qualified constructors. 3.1.2 Transportation Values Addressed by the Process Agencies, such as state DOTs, have the fundamental mission to provide transportation. How they do so reflects (or should reflect) those values of their constituents. DOT programs and poli- cies are thus developed to implement projects that address such community values. Figure 3. Costs and right-of-way dominate design decisions.

28 a performance-Based highway Geometric Design process 3.1.2.1 Core Values Reflected in Transportation Projects The transportation values that create the priorities and ultimately define and shape the program and each project involve one or more of the following three basic needs: • Maintaining transportation infrastructure in a state-of-good repair. • Providing or enhancing a transportation LOS, which may include multiple modes, such LOS characterized by mobility, accessibility, or both; and such service attending to both persons as well as goods. • Enabling delivery of transportation service in a manner that is safe for all users. Projects involving “state of good repair” may merely reflect the agency’s responsibility to properly manage the infrastructure it owns. They also contribute to a value characterized by “comfort and convenience,” at least with respect to pavement surface. Quite often there may be conflicting values for a project, such as regional mobility, local circu- lation, and property access. The design process should seek to define all values so that trade-offs can be assessed. Note that the above values should be viewed as applying to all potential users of the facility in any mode of travel. The values also should be considered in the context of those who may not directly use the road or corridor, but whose transportation needs may be influenced by the facility. Characterizing all activities of and hence all projects performed by a transportation agency as addressing these fundamental transportation values is an appropriate framework in considering project development. Moreover, note that each basic core value can and should be quantifiable. Only by quantifying in meaningful terms what the project will produce in one or more of the core values can an agency both explain and defend its proposed solution, and be assured that its overall program is producing value. 3.1.2.2 Other Project Investments Some projects may be funded and undertaken that have little apparent direct linkage to the above core transportation values. “Community re-vitalization” projects may involve street scape and landscape features that by themselves do not reflect the above values. Sound wall construc- tion as a standalone project is another example of a project without a specific transportation function. For the purposes of this research effort these projects are acknowledged as being aspects of an agency’s program, but are outside the transportation project framework presented here. 3.1.3 Objective of the Design Process The objective of the design process should be the creation of a financially sustainable road system that delivers transportation values where they are needed, both now and in the future. Full understanding of this concept results in design decisions that do not just reflect the specifics of the individual project, but also consider (1) how the project “fits” and functions within the immediate road network and (2) how the allocation of resources to the project will influence the ability of the agency to complete other worthy projects. 3.1.3.1 Participants in the Design Process Programming and scoping decisions establish type of project. Agency staff including chief engineers set design criteria and design policies. Assigned technical staff conduct the work using

highway Geometric Design and project Development 29 established methods and practices and apply the agency’s design standards and technical guidance. The following are typical technical disciplines involved in road design projects: • Transportation planners, • Traffic engineers, • Environmental planners, • Environmental scientists, • Cultural resource experts, • Pavement and materials specialists, • Road safety engineers, • Highway engineers and geometric designers, • Drainage and hydraulic engineers, • Bridge and retaining wall engineers, • Public involvement and facilitation specialists, and • Construction engineers. External stakeholders may help define the problem, point out constraints or issues for the designers to consider, and review and comment. Some external stakeholders may have regulatory responsibilities and outcomes requiring the owning agency to address under its jurisdiction. Any and all stakeholders may directly influence the design outcome. With respect to the actual solution, as owner, funder, and operator of the project, the owning agency—a state DOT, for example—is ultimately responsible for the final design decisions. 3.1.3.2 Data and Knowledge Requirements Highway design has classically been considered a civil engineering technical discipline. Geometric design, drainage and hydraulics, traffic engineering, geotechnical and materials science, and structural engineering are other core disciplines associated with highway engineering. Each discipline collects and processes data necessary to fulfill its roles on the project. These will include survey and base-mapping, soil borings, environmental resource surveys, traffic counts, travel forecasts or projections, and crash data. In recent years many agencies have acknowledged the role of external non-technical stakeholders in projects. Community surveys, open house meetings, advisory committees, and other efforts have been incorporated into projects to understand values and issues of affected stakeholders. 3.1.4 The Legal Framework in Which the Road and Highway Design Process Exists An important aspect of road design in the U.S. is the legal framework in which DOTs and roadway design professionals work. There are two important aspects to this framework—managing the risk of tort actions against owners and operators of the highway system as well as federal and state laws and regulations governing environmental protections and processes. 3.1.4.1 Legal Liability of Highway Agency Owners and Designers Actions of state governments are not immune from tort lawsuits. Road users involved in crashes may potentially bring tort actions against a state based on allegations of improper or negligent actions by their engineering and maintenance staff. In the context of road design projects, design errors or omissions may be alleged. Although the exact laws, limits, and potential liability vary state by state, in general the following is true: Professionals have what is referred to as a mandatory duty to perform their work according to established standards and practices of the profession, and following the policies of the agency

30 a performance-Based highway Geometric Design process for which they work. Failure to fully perform their duties correctly can leave the agency open to a potential tort action should a crash occur that can be attributed to a problem associated with the road’s design. Decisions involving the professional judgment and discretion of professionals are generally not challengeable, as long as the decisions are fully documented such that they can be explained and defended as being reasonable and not arbitrary. Hauer has noted that a necessary attribute of the highway design process in such a legal setting is the publication of and direct reference to design standards and criteria which are to be followed by design professionals (Hauer 1999). Such standards provide a clear benchmark against which an agency’s discretionary design decisions can be proven to meet the professional standard of care. In simple terms, the design plans and accompanying design reports and documents should be able to be quickly compared to the standards and policies in place when the plans were pro- duced. Again, while state laws and limits on discretionary immunity will vary, a road designer’s first and most basic defense is the ability to prove that the plan as designed met all the relevant published design standards. It is considered a given that the legal framework in which the profession works will remain in place. Any significant design process changes must produce design plans and supporting docu- ments that enable an agency to (1) quality review the plans and documentation to assure it meets or exceeds the standard of professional care and (2) be sufficiently clear and complete to serve to defend the agency against a tort action should that occur in the future. 3.1.4.2 Federal and State Environmental Processes and Regulations Projects of any size and substance are subject to meeting the requirements of NEPA and the many related environmental regulations and laws. Projects that are not federal actions may be subject to comparable state regulations. Highway projects, including design decisions that produce measurable or estimable adverse impacts, may be delayed or halted if opponents can successfully demonstrate that the owning agency and the FHWA did not appropriately identify and/or address such impacts. Outside parties will generally not have standing to sue because they do not support the final decision. Such decisions are discretionary acts of government and immune from suit as long as the overall project process, individual subprocesses and tasks, and technical efforts can be shown to meet the standards of professional practice. Over the past 30 years research on understanding environmental consequences of design actions has produced a large body of knowledge. Sophisticated tools for modeling environ- mental effects such as noise, air quality, economic impacts, and health impacts are now part of professional practice. Most DOTs have experienced at least one instance in which external stake- holders have been able to demonstrate to a court that some technical aspect of a project was not conducted properly. In some cases failure to appropriately consider all reasonable alternatives was sufficient to stop a project. Effective decision-making processes include a stakeholder-inclusive discussion and enumera- tion of what decision criteria, effectiveness measures, and relative importance of the criteria are to be employed to make the decision among competing alternatives. The study and decision processes employed will include meaningful measures of the unique community values that will typically encompass both transportation values (mobility, access, and safety) and other non- transportation values. In virtually every project, there will be issues of environmental sensitivity that will influence the generation and acceptability of alternatives as well as their performance attributes and outcomes.

highway Geometric Design and project Development 31 Roadway design decisions inherently involve trade-offs and judgments, with many of these involving environmental issues. The level of detail associated with quantifying environmental effects is increasingly greater. Future design processes should strive to match detail with better and more complete information on performance attributes of a design. 3.2 Key Findings on the Highway Design Development Process and Need for Design Process Changes Chapter 2 outlined the evolution and progression of highway design in the U.S. in response to societal changes, public policy initiatives, and technological advances. The framework outlined above captures values, objectives, and public policy within which highway and road projects must be developed. With this introductory background, the research team, with confirmation of the research panel, posits 16 basic findings that form the basis for suggested changes to the road design process in the U.S. These are summarized below. 3.2.1 Finding 1: Interdisciplinary Project Development Is Here to Stay (Institutionalization of CSS) The stakeholder-driven process referred by some as “CSS” is here to stay, whether by that name or some other names. Direct involvement of stakeholders, many of them external to the owning agency and many with varying roles, is now accepted as how highway projects are to be developed. The AASHTO policy document A Guide to Achieving Flexibility in Highway Design endorses the concept of stakeholder involvement in the design process (AASHTO 2004). Many state DOT project development processes now directly incorporate reference to stakeholders and their input, but the experience in full implementation is uneven. Stakeholder interests include both transportation and non-transportation values and concerns. Project development—including especially the highway design process—must address such concerns through workshop and public involvement tasks, targeted analyses and design studies, and formal evaluations, many of which must meet regulatory requirements. Listening and gaining a true understanding of concerns is an important part of this process. Projects with even modest budgets and scopes now routinely include technical input from multiple disciplines. The stakeholder process potentially increases the complexity and often the time needed to complete the project. It demands that owners fully investigate a range of design alternatives including cross section, alignment, and traffic operational options. Multiple interests translate to competing priorities, especially for projects with important limitations of budget, space, or right-of-way as well as with multiple problems or needs to address. In most projects in urban settings the highway or road is no longer designed for motor vehicles alone, but must address other modes including vulnerable users such as pedestrians and bicyclists, often within a pre-defined limited right-of-way envelope. All of the above reinforce the notion that road and highway design is in many ways the exer- cise of choices by the designer (the owner), such choices reflecting judgments about meeting the project’s objectives while addressing the full range of often conflicting stakeholder interests. This process is not separate from design—it is central to it. Highway designers and engineers need to understand that this is part of their job—to seek, assemble, review, and deal with stakeholder input. Such understanding also translates to a need for designers to have good communication skills in addition to their technical skills. The AASHTO policy on geometric design ideally will explicitly describe both the substance of and process by which highway designers seek and incorporate stakeholder input as an integral part of the highway design process.

32 a performance-Based highway Geometric Design process 3.2.2 Finding 2: Context Matters—And It Varies A project’s context as broadly defined matters greatly with respect to what is physically pos- sible to construct, what is reasonable to expect in terms of both operational and safety perfor- mance, what performance in fact will occur, what direct implementation costs are incurred, and what socioeconomic and environmental effects may result. The AASHTO highway context framework is limited in that it only reflects general considerations of the cost or difficulty of construction, and the traffic operational trade-off associated with mobility vs. access. It also lacks sufficient depth. Context varies greatly, and in many more ways and dimensions than the current AASHTO design process defines. The AASHTO context framework as currently defined—by location (urban, rural), functional classification (arterial, collector, local road), and terrain (flat, rolling, mountainous)—does not adequately capture the full range of context variance as it influences roadway or highway design development. Two important context features should be included—the presence of vulnerable highway users and the type of project. 3.2.2.1 Context and Vulnerable Users (Pedestrians and Bicyclists) Where service for motor vehicles is of paramount interest the operational performance of the road is measured by travel time, which in turn is influenced by speed. Higher design speed facilities are thus favored and promoted by policy. However, there are clearly situations in which the right overall solution should promote and enforce operating speeds that reflect the presence of road users particularly vulnerable to conflicts with vehicles at moderate to higher speeds. It is current national policy to promote walking and bicycle riding. Trends for both types of travel are strong and expected to continue. As of 2012, one in six of the total number of highway fatalities nationally is pedestrian and bicycle related. The AASHTO design process must explicitly acknowledge the importance and relevance of these road users. In the urban environment vulnerable user activity (pedestrians and bicyclists) must become increasingly more important as a fundamental control and input to road design. For some projects the transportation problems or needs will be focused on such users. For many other projects, the mere presence of vulnerable users in sufficient numbers clearly shapes the setting of performance measures and development of reasonable solutions. What is needed is a process, based on readily attainable data, that defines contexts in which meeting the needs of vulnerable users is of primary importance. Initial thoughts on this subject are offered here, to be more fully developed. Land use as a context definer offers a simple and intuitively appealing approach to defining the propensity for significant pedestrian activity. The extent of pedestrian activity varies widely based on both the type of land use adjoining a roadway and the density of the land use. Streets designed in downtown urban cores fronted by high-rise commercial properties experience high volumes of pedestrians walking along and crossing the roadway on a regular and frequent basis. Expectations of both motor vehicle operators and pedestrians themselves reflect the land use context. Both spatial demands for pedestrians as well as operational considerations where pedes- trians and motor vehicles compete for mobility will influence road design solutions. The current road design process lacks a means of sufficiently differentiating and defining land use contexts in which pedestrian and vulnerable user considerations are of such importance that they should “drive” the design process and solutions. AASHTO’s reference to “urban” is insufficient. The Institute of Transportation Engineers’ (ITE) recently published Recommended Practice for Walkable Thoroughfares outlines a Context Zone framework that is useful and could be directly applicable as an expanded context identifier, as illustrated in Figure 4.

highway Geometric Design and project Development 33 Seven context zones are described by both type and intensity of land use. Guidance for appropriate design controls that reflect sensitivity to vulnerable users, including specifically speed, is presented. This framework offers a starting point for discussion about how to define the full range of contexts. A formal process is needed for identifying contexts in which vulnerable users should be expected, and hence different approaches to speed and operational needs adopted. 3.2.2.2 Context As Defined by Project Type—The Differences Between Reconstruction and New Construction Projects The AASHTO policy considers new construction and reconstruction projects to be the same with respect to applicability of design policy. The historic basis for this assertion is unclear, but it presumably relates to the concept that, geometrically, a road undergoing complete reconstruction should be designed to current, updated design criteria. Such a judgment, though, ignores important aspects of the differing contexts around both project types. By definition a reconstruction project involves a roadway already in place and functioning. Similarly, a new construction project (one on new alignment) introduces a new transportation corridor where none exists. The following differences are evident: • With an existing road the abutting land use has developed around the mobility and access provided by the road; with a road on new alignment the road itself always produces significant change (both positive and potentially negative). • With an existing road there is a clear and well-understood operational and safety performance history; none exists for a road on new alignment. • The costs of construction and constructability of each project type are based on significantly different factors. 3.2.2.2.1 Land Use Context Differences. Abutting land uses evolve and develop around the traffic service provided by an existing road. In more developed locations land use will be fully occupied by buildings, parking, plazas, or setbacks established by local land use ordinances. Devel- opments form the economic and social fabric of the community, and are the source of employment and tax revenues. Reconstruction of an existing road should bear a burden of minimizing damage Source: Institute of Transportation Engineers, Designing Walkable Urban Thoroughfares: A Context Sensitive Approach—ITE Recommended Practice. Figure courtesy Duany Plater-Zyberk and Company. Figure 4. Roadway context zones.

34 a performance-Based highway Geometric Design process or disruption to property owners (businesses, residents, government agencies) who invested in their land based on the right-of-way and footprint of the existing road. Reconstruction projects that require taking of new right-of-way will inevitably produce adverse impacts beyond the direct cost of the right-of-way itself. Design philosophies and solutions for reconstruction projects should thus properly focus on what will “fit” within the existing right-of-way, an exercise that may influence not only cross section but also alignment. Finally, reconstruction projects will frequently produce changes in access. In most cases these will be con- sidered adverse by the abutting property owners (e.g., closing of driveways, or their consolidation or relocation). By contrast roads on a new alignment involve right-of-way acquisition for the entire corridor. Such projects are typified by studies of alternative independent alignments, with the availability, adverse effects such as severances, and cost of the right-of-way key factors in selection of a preferred alignment. Designers will generally have more latitude in identifying feasible corri- dors. Finally, the addition of the road itself to the local context almost always represents a net improvement in transportation service (mobility or accessibility) given that no such service preceded the road. 3.2.2.2.2 Operational and Substantive Safety Performance History. Perhaps the most significant difference between reconstruction and new construction concerns the transportation performance history of the facility. In the latter case it simply does not exist. Indeed, the design process must rely on travel demand forecasts for volume and pattern of traffic that are inherently uncertain. There is no way to directly observe how drivers will respond to the finished solution in terms of their speed behavior, navigation, or decision making. Reliance on simulation and other models of expected driver behavior must suffice. Finally, the expected safety performance can only be estimated based on the knowledge base and models from the HSM or other sources. With existing roads to be reconstructed the designer will have a complete understanding of the traffic operational and safety performance of the existing road over time. This can include direct observations of traffic volume by time of day, delay, speed and speed behavior, travel time, gap acceptance, and conflicts. It should also include a complete set of data and observations on the frequency, types, severity, and other characteristics of crashes over whatever time period is necessary to reach full understanding of the safety performance. Most roads undergoing reconstruction will be 30 years old or older. In most cases their geo- metric features will reflect AASHTO or equivalent design policies from 1984 or earlier. Given the periodic changes in design policy since 1984, it will be commonplace for one or more geometric features not to meet current criteria. The existing road thus may or may not include geometric features that are “nominally unsafe,” i.e., do not meet current criteria. Roadway design standards— nominal safety—are a means to an end, that end being performance. The presence of a nominally unsafe condition or feature is not a transportation problem—it is a condition of the context. In the absence of an associated measurable transportation problem (safety or operational) there will be no performance value obtained in upgrading the road to current standards. Moreover, as noted above, given the context of fixed right-of-way any such upgrading, whether for cross section or alignment, will invariably create significant adverse impacts and costs to adjoining properties and stakeholders. Any and all reliable data describing actual transportation performance that can inform the design process should be directly incorporated into that process. Highway design for reconstructed roads thus offers a level of knowledge and understanding not available for new alignment proj- ects. Incorporation of such knowledge into the reconstruction design process is a significant differentiator that should be explicitly acknowledged in design policy and process.

highway Geometric Design and project Development 35 3.2.2.2.3 Construction Costs and Constructability. Design criteria per AASHTO are based in part on the notion of cost effectiveness, which is based on cost elements of the quantities of new pavement, structures, and earthwork balance. Such considerations will drive the design process of a roadway on new alignment as they have historically. The difficulty and resultant costs of reconstructing an existing road are much more heavily influenced by factors other than material costs or pavement or earthwork. Reconstruction in most cases must accommodate existing traffic service on the corridor while construction occurs. Costs of multiple project phases with traffic control switches, detours, temporary pavement, and closures can amount to 30% or more of a reconstruction project, and in many cases be the deciding factor in a final design solution. Vertical alignment controls reflect the need to maintain access points and intersections during construction; earthwork balance is at best a secondary concern. The implied cost-effectiveness relationship of criteria based on construction of new alignment does not translate to reconstruction. 3.2.2.2.4 Potential Design Process Improvements for Reconstructed Highways and Roads. The ability and indeed responsibility to fully study and incorporate knowledge of the existing performance of a road to be reconstructed is a substantial benefit to the design process for such projects. The design process for reconstructed roads should address this information, with solutions reflecting specific performance. Such a process may produce the following: • Existing geometry regardless of whether it meets criteria or not could presumptively be iden- tified as adequate based on a thorough review of the safety and operational performance. • A road with a “nominally unsafe” geometric element (i.e., one not meeting current criteria) could retain such element if a thorough review of the safety and operational-performance history deems the geometry to be satisfactory, or if a demonstrated performance problem was determined to be unrelated to the geometry. • Retention of existing geometry could by policy be approved with such analyses and without the labeling and need for a design exception. Design criteria could be expressed in a way that directly allows retention of a design feature based on performance analyses. Note that there are hints of differentiating for reconstruction already in the AASHTO Green Book. Table 6—showing the headings for lane and shoulder Metric U.S. Customary Design Speed (km/h) Minimum Width of Traveled Way (m)a for Specified Design Volume (veh/day) Design Speed (mph) Minimum Width of Traveled Way (ft)a for Specified Design Volume (veh/day) under 400 400 to 1500 1500 to 2000 over 2000 under 400 400 to 1500 1500 to 2000 over 2000 aOn roadways to be reconstructed, an existing 6.6-m [22-ft] traveled way may be retained where the alignment is satisfactory and there is no crash pattern suggesting the need for widening. b Preferably, usable shoulders on arterials should be paved; however, where volumes are low or a narrow section is needed to reduce construction impacts, the paved shoulder width may be a minimum of 0.6 m [2 ft] provided that bicycle use is not intended to be accommodated on the shoulder. Source: Table 7-3 AASHTO Green Book Table 6. Example of performance-based criteria from current AASHTO Green Book.

36 a performance-Based highway Geometric Design process width design values—has a footnote that addresses the safety and operational performance of existing roads within the context of establishing a need for widening. Such process improvements would represent an appropriate emphasis on performance rather than roadway geometry as a surrogate for performance. They would eliminate costly, no value solutions (“upgrade to standards”) and reduce the bureaucracy of needless design exceptions. The process could be suitably flexible in its construct to allow agencies to select decision-performance thresholds based on availability of funds, priorities for certain road types, or any other factors deemed important. Finally, such a process would directly incorporate a strong incentive for a designer to conduct proper evaluation of the performance of the existing road in order to garner the value in cost savings associated with retaining the existing geometry. 3.2.3 Finding 3: Providing Multimodal Solutions Is Now the Rule and Not the Exception The geometric design process, historically focused solely on motor vehicles, must evolve to more directly and routinely address the needs of all potential users of a facility or corridor. Both process and cultural change within the road design community are needed. There are inherent conflicts and choices to be made in prioritizing the amount and manner of transportation service afforded general purpose traffic, transit, truck and freight traffic, bicyclists, and pedestrians. The design process should direct the resolution of such conflicts and the establishment of choices among all needs. With respect to process, more refined context definitions are needed to identify corridors and conditions in which, for example, pedestrian needs should take precedence over motor vehicles. The law requires pedestrian facilities to comply with ADA Requirements. Similarly, the presumptive need to design cross sections, intersections, and vertical alignment recognizing the presence of bicycles is also desirable. As noted above, land use variables, including type and density of use are possible context definers that may be included. 3.2.4 Finding 4: AASHTO Dimensional Criteria Should Ideally Be Based on Known and Proven Measurable Performance Effects The AASHTO policy is over 1,000 pages long and continues to grow with every edition. Much of the growth in contents stems from advances in research knowledge, but also expansion of considerations that were not important or even existed in previous years. Designers are con- fronted with increasing demands and urged to be “flexible” in their approaches. DOT manuals based on the AASHTO policy are generally written in a manner that removes flexibility. Too many designers don’t understand the relative importance of a given criteria, or are not allowed to exercise judgment in ignoring or violating a criterion. There remain within the AASHTO policy and state DOT manuals examples of design criteria that are based on outdated rational models or assumptions that in practical terms offer no mean- ingful value. Criteria for minimum length of curve and for ratio of compounded horizontal curves are two examples. Dimensional criteria published by AASHTO will ideally reflect known, proven, and mean- ingful operational or safety performance effects. The guidance in the AASHTO Green Book that is not based on research but is based on past practices and professional judgment should have research conducted to either verify that the guidance provides for the desired operations and safety results or a revision to that guidance made accordingly.

highway Geometric Design and project Development 37 3.2.5 Finding 5: Speed Is an Essential Input to Determination of Design Values and Dimensions The design process is significantly reliant on speed as a central input or control. Current AASHTO processes incorporate the concept of “design speed” and others have suggested the use of “target speed.” Regardless of the term used, development of design dimensions and details will to a great extent require the setting of speed control or variable. Speed clearly influences distances vehicles travel while maneuvering. Speed directly influences the severity of conflicts and crashes. The design process requires a framework and starting point (i.e., “design controls”). Speed, or the “speed regime” in which the road is to operate, is arguably the most important control to establish. Moreover, the need to specify a speed as the basis for a design includes documentation of the design within the legal framework. Besides being a critical component to design criteria, speed must be acknowledged as having conflicting contributions to transportation performance. Historically, speed has been a surrogate measure of quality in that the prevailing transportation value was travel time (i.e., its minimization). Travel time and hence speed continues to be important in this regard. However, the adverse effects of speed on safety performance must also be considered. Nontechnical stakeholders recognize what is documented in the research (see Figure 5). The survivability of vulnerable users in crashes decreases dramatically as the speed of the collision increases. Providing or encouraging a high-speed environment under every possible context does not appear to be an appropriate approach. In some situations, reducing the speed could be considered an effective solution to reduce crashes. The design process and AASHTO guidance will by necessity continue to require the setting and application of selected or assumed speeds as a fundamental design control. Figure 5. Adverse effects of speed on safety performance.

38 a performance-Based highway Geometric Design process The design process and AASHTO guidance will also by necessity require the identification of those contexts and circumstances in which high-speed operations should not be encouraged because of the increased risk to vulnerable road users. 3.2.6 Finding 6: AASHTO Design Criteria Produce Uneven Outcomes Re: Performance AASHTO design criteria have evolved over the years. In some cases models and assumptions have remained unchanged over time (e.g., SSD, horizontal curvature); in others research has resulted in wholesale changes to the design approach (e.g., intersection sight distance, passing sight distance); and in still others significant changes in design values have resulted (lane and shoulder widths on two-lane rural roads). As far as the research team knows, the entire set of basic geometric design criteria have never undergone a thorough, holistic review. Under current policy the models and approaches vary widely. The input assumptions and basis for design criteria reflect differing value judgments, which in turn produce varying performance. As summarized in Table 3: • Horizontal curve design is based on providing comfort, • Lane and shoulder width design is based on safety performance for two-lane rural highways, • SSD is based on a theoretical model lacking evidence of its relevance to actual events and operations, • Maximum grade is based on heavy vehicle operations, and • Roadside design is based on explicit analysis of safety performance. Many basic AASHTO models apply across the full range of road types and contexts. This approach to design has yet to be questioned. It may be that more context-specific models and input parameters would yield more cost-effective and reasonable solutions within application of the policy. As a follow up of this finding, a detailed evaluation of the design criteria was conducted as part of this research. Refer to Section 6.5 for additional discussions on this topic. Given the above it is not possible to assert that the application of all design criteria as currently published in AASHTO will produce uniformly or programmatically cost-effective outcomes. 3.2.7 Finding 7: Many AASHTO Criteria Are Not Sensitive to Key Context Attributes That Are Proven Influencers of Performance and Cost Effectiveness AASHTO criteria are presumed by users to be directly related to safety performance. The underlying principle of cost effectiveness is also presumed to apply to the use of AASHTO criteria. As noted previously, two fundamental variables central to the concept of safety performance and overall cost effectiveness—traffic volume and type of road—are not present in the formulation of many basic AASHTO design models. Per Table 7, horizontal curvature, SSD, and lane width for urban and multilane facilities are derived independently of design year traffic and road type. Horizontal curves are designed per AASHTO using the same model for roads with 500, 5,000, 50,000, or more vehicles per day, and the same also for two-lane highways, multilane highways, and the full array of access-controlled facilities alike. Yet the expected safety performance of any curve, which translates directly into cost effectiveness, will clearly differ in meaningful ways for the volume and road-type conditions noted above.

highway Geometric Design and project Development 39 Table 7. AASHTO criteria evaluation. Design Element Functional Model Basis Model or Design Assumptions Incorporates Design Speed? Incorporates Empirical Substantive Safety or Human Factors Research? Sensitive to Traffic Volume? Sensitive to Functional Classification or Road Type? Sensitive to Variances in Vehicle Type? Incorporates Interactive Effects of Other Roadway Elements? Comments Research Basis for AASHTO Policy Horizontal Alignment Radius of Curve Passenger car produces acceptable threshold of comfort operating at design speed Vehicle operates at constant design speed tracking the center of the designed curve throughout its length Yes Yes*—dated studies of driver discomfort operating on curves No No No No Studies of actual operations on curves conflict with AASHTO assumptions Length of Curve Length, radius, and central angle between tangents are interrelated geometrically Length and/or central angle are not currently included in design policy as controls NA NA NA NA NA Trade-off involving curve radius and length produces varying quantitative safety performance for the same central angle Not explicitly covered in design policy Superelevation Superelevation counteracts side friction on “one-to-one” basis Passenger car exactly tracks road as designed at constant design speed Yes No No No No No Studies of actual operations on curves conflict with AASHTO assumptions Vertical Alignment Maximum Grade Truck operations on upgrades Ability to reach sustained speed for assumed weight to horsepopwer (WT/HP) for heavy vehicles Yes (initial speed) Yes—research relating speed differentials to rear-end conflicts No Yes Yes Grade and length of grade combine to produce the speed of the vehicle Criteria do not reflect ability of bicyclists to operate on steep and/or long grades Minimum Grade Pavement drainage No No No No NA Minimum grade, superelevation and cross slope all produce cross section that drains Criteria do not reflect frequency and/or intensity of precipitation Crest Vertical Curve Provide SSD See sight distance above; operation at design speed with assumed eye height and object height No No No No Vertical curve length is directly related to intersecting tangent grades Sight-distance profiles produced by combinations of grade and vertical curve length vary NCHRP Report 400 Sag Vertical Curve Visibility of pavement at night; headlight beams; also comfort in the extreme if headlight beam criteria cannot be met Assumed headlight height and beam spread Yes No No No No Vertical curve length is directly related to intersecting tangent grades Cross Section Lane Width Widths for two-lane rural highways based on cost-effectiveness analysis including both substantive safety and traffic operations NA Yes Rural two-lane highways only Rural two-lane highways only Yes No (exception is guidance for lane widening on horizontal curves where trucks are present) Rural two-lane highways research basis reflects combined effects of lane and shoulder width NCHRP Report 362 Shoulder Width Widths for two-lane rural highways based on cost-effectiveness analysis including both substantive safety and traffic operations “Full-width” shoulders are considered 10 feet or more Yes Rural two-lane highways only Rural two-lane highways only Yes No Rural two-lane highways research basis reflects combined effects of lane and shoulder width NCHRP Report 362 Cross Slope Pavement drainage No NA No Yes (greater slopes on lower class roads; and on wider pavements) NA Median Width Separate opposing traffic Varying widths associated with intended function(s); e.g., separation, incorporation of left-turn lane(s), access control, enable provision for physical barriers No No Yes—applies to multilane roads only; freeways require minimum dimension for shoulders and barriers No Median type (raised vs flush) Medians not required for nonfreeway facilities Sight Distance Stopping Collision avoidance with object in road Passenger car operating at design speed brakes to full stop; single values for object height, eye height, reaction time, and deceleration rate/Horizontal sight line assumed to be center of roadway/lane with sight line to object at center of roadway/lane Yes Changes in model parameters based in part on review of crash records of objects struck by size No No No Yes—effect of grade on stopping length Original derivation of object height based on cost-effectiveness calculations of construction costs; NCHRP Report 400 resulted in change to a meaningful object height NCHRP Report 400 Intersection Gap acceptance for vehicles on minor approach Driver requirements for gaps based on range of cases based on type of maneuver (turning, crossing) with sight lines based on design driver eye height and vehicle height for range of vehicle types Yes Yes No Yes (cases categorized by urban, rural) Yes Yes—presence of intersections; effect of approach grade NCHRP Report 383 Passing Distance required for vehicle to complete or abort a passing maneuver Passenger car undertaking passing maneuver assuming speed differentials, acceleration capabilities, and available sight distance with design driver eye height and design vehicle height Yes Yes—observations of passing maneuvers No Applies to two-lane rural highways only No No Decision Distance for driver to detect unexpected or difficult to perceive information source or condition; recognize it, select speed and path, and initiate complex maneuvers. Human factors requirements for complex actions Human factors—based on time requirements for five cases involving range of contexts and operating conditions. Yes Yes No Yes (cases categorized by urban, rural) No Yes—e.g., presence of intersections Decision sight distance (DSD) is not required but considered advisory or good practice Roadside Design Lateral Offset Operational offset for car doors, side mirrors, etc.; also potential for impact with roadside objects Yes Yes No Yes No Yes—4 feet lateral offset on moderate to higher speed roads without vertical face curb Side Slope Ability to recover given encroachment Vehicle leaving roadway at design speed Yes Yes No No Yes Yes—clear zone Clear Zone Ability to recover without overturn or striking an object given encroachment beyond edge of pavement Yes Yes Yes No No Yes—side slope and shoulder width Vertical Clearance Vertical Clearance Provide at least 1-foot vertical clearance to maximum legal-height vehicle Legal height of vehicle is 13 feet NA Yes* No Yes, greater dimensions for freeways and Interstates No No Tunnel clearance is special case

40 a performance-Based highway Geometric Design process Empirical models of safety performance based on HSM research confirm that the risk of a crash varies by road type. For both segments and intersections, the following is known: • The frequency of crashes by type varies widely for two-lane versus multilane roads and roads in urban vs rural areas. • The effect of specific geometric variables and dimensions on safety performance also varies by roadway context. Design criteria properly applied should generally produce cost-effective solutions in any combination of context and traffic volume. The formulation of current design criteria does not do this. It is possible that in some contexts design criteria are overly restrictive, producing no meaningful performance benefits for the increment of additional costs required. In other cases, application of current design policy may inadvertently miss the garnering of performance benefits with no or minimal increase in cost. In the view of the research team, reaching the goal of having cost-effective design criteria must mean the incorporation of traffic volume and road type in the formulation of such criteria. Any geometric criterion or design model lacking sensitivity in its formulation to traffic volume and road type cannot be assumed to produce cost-effective solutions when applied across the full range of design conditions. 3.2.8 Finding 8: Some AASHTO Criteria Are Unnecessarily Simplistic in Their Formulation, or Are Based on Models That Are Lacking a Proven Science Basis The AASHTO policy emerged from a recognized need in the 1930s and 1940s for formal design direction. At that time there was little if any scientific or empirical knowledge about human factors, traffic operations, or safety performance. Originators of the AASHTO policies relied on simplifying assumptions and simple rational models based on fundamental concepts of physics. Examples include: • AASHTO horizontal curve model, which assumes constant operation of a passenger car at design speed independent of upstream alignment effects, with the vehicle tracking the center of the roadway. • SSD model, which assumes passenger car operation at design speed and hard braking to a full stop to avoid collision with a fixed object. • Horizontal sight-distance model, which defines the offset assuming a driver eye location at the center of the lane, and a point fixed object at the center of the lane. • Cross-slope rollover criteria based on the assumption that the rollover itself is the critical operating condition. • Design guidance for interchange ramp design speed expressed as a percentage of the freeway’s design speed. In each of the above examples either the models themselves or the assumptions used in them are overly simplistic, have been proved to be incorrect in actual operation, or in some cases not based on any science. Some simplifying assumptions reflect the pre-computer era in which graphical techniques were used or calculations were made by hand. (For example, there is no reason why horizontal sight lines cannot be defined based on more appropriate placement of the driver’s eye in the lane, and also on placement of the object at different spots across the width of the lane. These assumptions would provide significantly different horizontal offsets based on the direction of the curve itself.) Glennon (1987c) confirmed 25 years ago that drivers do not track horizontal curves as assumed, but rather “overdrive” them. Krammes and Otteson (2000) showed that the assumed speed behaviors produced through curves did not replicate AASHTO assumptions (see Figure 6).

highway Geometric Design and project Development 41 In particular, the research team has concern over the simple models for horizontal curve design and SSD, which have not been challenged nor substantially changed over the years. As both of these are core criteria that heavily influence geometric design, they warrant substantial further study. A highway design process applicable to the wide range of contexts, and intended to produce cost-effective results, requires design models and criteria sufficiently sophisticated and robust, and directly linked to research knowledge on operations and/or substantive safety. 3.2.9 Finding 9: The Legal Framework Requires the Provision of Threshold Limits for Design Criteria and Design Values; the Question Is—How Should Such Lower Limits Be Set? Even assuming perfect knowledge of safety and operational performance, it is questionable whether a design process that relied totally on performance models would suffice. Risk managers confirm the importance of published criteria. The design profession has relied on such criteria throughout its history. The question of what should constitute minimum dimensional criteria is to an extent a philo- sophical one. The authors of NCHRP Report 400 (Fambro et al. 1997) introduced the theoretical risk model of SSD, shown in Figure 7. Even if it is agreed that a safety risk-based approach is appropriate for SSD, and the actual shape of the curve could be established for a range of contexts through research, one still must decide where to “draw the line,” i.e., how much risk should be codified in the minimum design dimensions? A risk-free highway cannot and should not be promised. Hauer’s (1999) observations provide some measure of direction. It may be that minimum dimensions should be based on fundamentally human factors considerations. Finally, where and how one chooses to establish minimum design dimensions also affects decisions about exceptions and design exception processes. Again, the answers to these questions need careful discussion. A design process that derived and codified dimensions that were as low/ minimal as can be justified may satisfy some who argue that designers should always stay within published criteria. But others may not be comfortable with a process that has no “out” or means Source: Fambro et al. 1997 Figure 6. Comparison of actual vs assumed driver speed behavior by AASHTO curve design model.

42 a performance-Based highway Geometric Design process of allowing an exception. Moreover, setting a very low bar for minimum criteria carries with it a responsibility for agencies and their design staff to be more knowledgeable and exercise judgment in ways that they currently do not typically do. Additional research is needed to set criteria based on the risk for the different contexts encountered on various projects. 3.2.10 Finding 10: Nominal and Substantive Safety Differ in Meaningful Ways For reasons noted by Hauer, a “nominal safety” threshold approach is a necessary element of design policy (Hauer 1999). The legal framework in which the profession operates demands this. Owners and designers need to have firm, quantifiable, and documentable thresholds against which their chosen design can be compared and defended as meeting the standard of care. Unfortunately, the necessary concept of “minimum” criteria or dimensions has taken on unintended meaning. In applying the AASHTO policy, designers have evolved to a “mental model” as was shown previously in Figure 3. A design that meets the minimum addresses any and all potential risks. Designs above the minimum will cost more with no quantifiable benefits. As Hauer (1999) has observed, what is characterized as minimums actually becomes maximums. This mindset precludes a process that would seek an optimization of performance vs. imple- mentation cost. Such optimization would reflect what is well known and documented about traffic operations and substantive safety: that they will tend to vary over the range of reasonable design dimensions and that they will vary under different contexts. Given that implementation costs are clearly site-specific, strict adherence to minimum dimensions in every case (the nominal safety mindset) will not produce programmatically cost-effective designs. A highway design process that provides optimal results requires design criteria to be applied based on research knowledge for substantive safety (Figure 8). As demonstrated earlier, nominal safety models, equations, and approaches are not related to substantive safety performance. As such a one-to-one relationship between a condition in which design criteria were substandard and a history of crashes should not be expected. For an exist- ing road in which either reconstruction or 3R is being contemplated, a simple condition matrix describes what a designer may encounter (Figure 9). Depending on which quadrant the project falls, one should expect a fundamentally different approach to both defining the problem and proposing a solution. Source: Fambro et al. 1997 Figure 7. What Is acceptable risk as the basis for design criteria?

highway Geometric Design and project Development 43 The design process, while requiring nominal safety thresholds, should be focused not on pro- ducing minimum designs but rather on the optimization of substantive safety (and substantive performance) within an overall framework of implementation cost effectiveness. 3.2.11 Finding 11: AASHTO Criteria Should More Completely Reflect Known Interactive Safety and Operational Effects of Geometry Research has established significant interactive effects of, for example, grade and alignment on speed, roadside design and alignment on safety performance, and speed change effects on safety performance. There is sufficient anecdotal evidence to suggest interactive effects worthy of investigation (e.g., the effect of grade and direction of grade on loop ramp operations for trucks). In the urban environment the combined effects of medians, access control, and lane or roadway Figure 8. The two dimensions of safety. Is the designated project “Nominally Safe”? (Do its design characteriscs meet current criteria?) YES NO Is th e de sig na te d pr oj ec t “S ub st an v el y Sa fe ”? (Is th e hi st or y of cr as he s a lo ng th e pr oj ec t w ith in a d es ig na te d th re sh ol d of a cc ep ta bi lit y? ) YES NO Figure 9. A decision matrix based on the two dimensions of safety.

44 a performance-Based highway Geometric Design process width are evident. Indeed, there is a striking disparity between the known importance of access control and management on both traffic operations as well safety performance, and the extent to which it is referenced in the development of other geometric elements, most notably urban cross section design. In the two-lane highway rural environment the primacy of roadside design in establishing safety performance is well understood, yet roadside design and both alignment and cross section are considered independent of each other. The formulation and application of design models, dimensions, and approaches should to the extent possible directly incorporate meaningful interactive performance effects (speed, operations, substantive safety). 3.2.12 Finding 12: Replace Dimensional Guidance with Direct Performance Guidance Where Possible Within the AASHTO Policy Technology advances in traffic simulation models are now well established in practice. Their use should be explicitly acknowledged and referred to in the AASHTO policy. Moreover, consid- eration should be given to their use with appropriate traffic data to replace current dimensional values in the policy. Consider, for example, Figure 10, which shows design guidance for ramp spacing and weaving as expressed in feet. These dimensions actually are based on both human factors knowledge and traffic operational analyses of freeway operations. This table could be replaced by perfor- mance outcome guidance (e.g., lane density in lane one, speed change between segments, speed differentials across lanes) with reference to appropriate simulation models for obtaining such performance measures, rather than citing physical dimensions. This concept is illustrated in the AASHTO Green Book in Figure 10-2, which provides qualitative performance criteria rather than dimensions for access control near an interchange. Other research (design-consistency module of the IHSDM) provides designers the ability to produce speed profiles along a 3D alignment for a range of vehicle types. Other examples of Source: Figure 10-68 Recommended Minimum Ramp Terminal Spacing, AASHTO Green Book Figure 10. Dimensioned minimum ramp terminal spacing.

highway Geometric Design and project Development 45 integrating simulation into design include turning lanes at intersections (based on, for example, 95th percentile queues from simulation rather than fixed dimensions). An interesting parallel to this suggested approach already exists, in the process associated with design decisions involving environmental criteria. Warrants for sound walls are based on land use types (residences, parks, schools, hospitals) and a noise-level performance criteria—decibel levels at the receptor. The required design dimensions for the walls (locations, height) are obtained from noise modeling, not through specific look-up tables or published dimensions. This suggestion has the benefit of requiring designers or those assisting them to compute a design dimension from a model or algorithm that expresses the performance intent. Designers thus must understand and apply the knowledge base behind the performance in order to dimen- sion their plans. Physical dimensions are the means to the end, the end being some intended level of perfor- mance. The design process should (1) require use of best technology practices in confirming a design and (2) replace reference to physical dimensions, with specifications for technical dynamic, performance-based approaches that will produce the appropriate dimensions for the specific project and context. 3.2.13 Finding 13: Advances in Technology Should Be Incorporated into the Geometric Design Process Designers will always need to follow a basic “linear” order of completing their work. They start with the cross section, then develop horizontal alignment, and next vertical alignment. This practical approach is unchanged by technology. However, the criteria designers are given and the manner in which they apply them have not yet been formulated or revised to take full advantages of the advances in technology. Both the formulation and application of AASHTO policy criteria reflect the pre-computer age design process. For example, design policy provides look-up tables as opposed to functions or formulas. More importantly, in the pre-computer age the time and labor cost to complete all the technical work of roadway engineering was the controlling factor in the design process. Many technical approaches were established to minimize the time and chances for error in calculating coordinate geometry (e.g., the use of parabolic vertical curves). An alignment, once calculated and plotted, required substantial time and effort to revise—both the engineering and drafting. Computer-aided engineering and design tools and methods have radically changed the profession of highway engineering in positive ways. Survey and mapping is much less time consuming than previously. The time to produce an alignment is fractions of what was required previously. Changes, new alternatives, and fundamentally different concepts can be produced in a matter of hours rather than weeks or months. See Figures 11–13. Geometric design is three-dimensional. The ability to produce a 3D alignment and “drive” it using simulation software is now easily done. The ability to produce sight-distance profiles is similarly readily done. Human factors research has produced driver workload models. Finally, advances in engineering approaches to traffic operations analysis and now sub- stantive safety analysis have significantly improved our ability to predict and understand the performance of a design, under as many different scenarios of traffic and other factors as may be needed. The current project development process includes traffic operations analysis, but mostly at the “front end.” Most importantly, complete integration of dynamic operations analysis in an iterative manner is not routinely done. With respect to highway safety performance analysis,

46 a performance-Based highway Geometric Design process Figure 11. Pre-automated drafting environment. Figure 12. CAD work environment, 2015. Figure 13. LiDAR monitoring equipment for data collection in the field.

highway Geometric Design and project Development 47 DOTs are just now beginning to learn the methods and approaches. Typical applications are in design exception evaluations, and to a limited extent alternatives analysis. All the technology growth advantages offer the means for a dynamic, iterative design process as illustrated in Figure 14 and summarized here: 1. Project-specific, CSD, and performance criteria are established, 2. A design solution is developed in sequence (typical section, horizontal alignment, vertical alignment), 3. The dynamic performance of the solution is modeled as is the life-cycle cost, with a cost- effectiveness index or optimization exercise completed, 4. Based on model results the geometry is revised, and 5. Performance and cost are re-modeled and optimization re-computed until a consensus or satisfactory outcome is evident. The effort, time, and cost to redesign a project in such an iterative manner is no longer a limit- ing factor or constraint, given the advances in design technology. The more costly and complex the project (e.g., a multi-level system interchange) the more significant the potential benefits of such iteration. The above process is ideally suited to a project development process in which stakeholder involvement occurs throughout alternatives and preliminary design. Note that this process was that envisioned (in more simple terms for two-lane rural highways) by FHWA in the original concept behind its IHSDM. Advances in design technology and performance-based analysis of geometric designs should be integrated within the design process in an iterative manner. 3.2.14 Finding 14: The Notion of “Conservatism” in Policy and Leadership in the Highway Engineering Field Needs a 180 Degree Shift Much of design policy, including not only design models and dimensions, but attitudes about design and professional responsibilities, represents a legacy of the times up until the 1970s. In the very early days of road construction and through the Interstate construction era, road designers Figure 14. A performance-based, iterative design process.

48 a performance-Based highway Geometric Design process made do with somewhat limited technical knowledge of operations and safety. The primary value to be met was speed. Funds for road construction were sufficient. The concept of “more is better” in width, right-of-way, and footprint prevailed. Designer mindsets need to change. Being conservative should mean not reconstructing or adding cost to a project unless there is clear evidence of a performance improvement, using state-of-the-art methods for traffic operational or substantive safety analysis. “Upgrade to standards” should be characterized for what it is—an administrative decision that often results in a wasteful expenditure of limited funds. 3.2.15 Finding 15: Geometric Design Should Be Understood, Taught, Executed, and Communicated as Iterative in Nature with Performance at the Center of All Iterations and Optimizing Following from all of the above, the geometric design process should clearly be understood as a dynamic and iterative one, as illustrated in Figure 14. The design itself should be viewed as the means to an end, the end being the expected or desired performance, as determined from the problem definition and scoping process. At the university level as well as within DOTs, the teaching of highway engineering should not be limited to use of the design software and the contents of the Green Book. The engineering of a highway is fundamentally about the providing of transportation. Highway engineers should know traffic operational performance and how to predict it as well as substantive safety performance. 3.2.16 Finding 16: More Explicitly Incorporate Maintenance and Operation Costs The design and decision process in the U.S. has historically been driven primarily by initial implementation costs (right-of-way and construction costs). The financial sustainability of any agency’s program is highly dependent on the maintenance and operating (M&O) costs of their infrastructure. Implementation of agency asset management is the dominant emerging trend among forward thinking agencies. There is a need to assemble and integrate the knowledge associated with M&O costs, particu- larly as pertaining to design decisions and trade-offs that are common in projects. For example, costs of roadside maintenance (mowing, ditch maintenance) vs. guardrail; signing, lighting, and delineation; snow removal; retaining walls and structures vs. embankment; shoulder widths for maintenance and enforcement operations; and raised vs. flush medians with plantings are all routine decisions that can have measurable maintenance costs. An iterative optimization process should include the ability to model or estimate annual M&O costs as a function of key design decisions—geometric and other—that are proven to influence such costs.