Evaluation of the 13 Controlling Criteria for Geometric Design (2014)

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.

62 S E C T I O N 5 This section of the report addresses refinement of the defi- nitions for the 13 controlling criteria to reduce overlaps and confusion in how they are applied in a typical design excep- tion process. The discussion also addresses potential modifi- cations to the list of controlling criteria. 5.1 Refinement of Criteria Definitions Specific pairs or sets of design criteria from among the 13 controlling criteria that are closely related to one another and/or need to be carefully distinguished from one another include the following: • Design speed, horizontal alignment, vertical alignment, stopping sight distance, and lane width • Lane width and shoulder width • Shoulder width and horizontal clearance/lateral offset • Bridge width, lane width, and shoulder width • Horizontal alignment, superelevation, and cross slope • Grade, vertical alignment, and stopping sight distance • Horizontal and vertical alignment • Horizontal clearance (lateral offset) and clear-zone width Each of these pairs or sets of design criteria are discussed below. 5.1.1 Design Speed, Horizontal Alignment, Vertical Alignment, Stopping Sight Distance, and Lane Width Design speed differs from the other controlling criteria because it is really a design control rather than a design cri- terion. When a highway agency chooses a design speed for a project, this does not directly affect the design of the project, but it does have an important indirect effect through the role of design speed in determining the criteria for lane width, Refinement of Definitions for the 13 Controlling Criteria horizontal alignment, vertical alignment, and stopping sight distance (see Sections 2.2, 2.6, 2.7, and 2.9 of this report). Given the role of design speed as a design control, rather than a design criterion, it might be desirable to distinguish it from the other controlling criteria in some way. The FHWA policy on design speed (2) states that the pur- pose of this controlling criterion is “to assure that drivers oper- ating at the legal speed limit can do so without unwittingly exceeding the safe design speed.” The use of the words “safe design speed” in this statement appears to presume knowl- edge that the highway engineering profession does not, in fact, have. There is no research that demonstrates that vehicle oper- ations at the design speed are safe or that vehicle operations at speeds above the design speed are unsafe. Indeed, design poli- cies are generally developed with substantial (though usu- ally unquantified) margins of safety. There have been some attempts to quantify such margins of safety. For example, Harwood et al. (33) have demonstrated that current AASHTO policies for horizontal curve design provide substantial mar- gins of safety against skidding and rollover for vehicles trav- eling at, and even above, the design speed. The requirement to use design criteria for a design speed equal to at least the posted or legal speed, or develop a formal design exception, seems a reasonable administrative control given our lack of knowledge in the area, but the use of the term “safe design speed” seems to imply a level of certainty that is not supported by research. It would certainly be desirable for researchers to provide better information to designers about the current lack of knowledge concerning the relationship of design speed to safety and to conduct research to better quantify such effects. A clear weakness of the first edition of the HSM (12) is the absence of factors to quantify explicitly the effect of traffic speed on safety. The HSM does not include such information because such effects are largely undocumented. As discussed in Section 2.1 of this report, Hauer (17) indicates that crash sever- ity increases with vehicle speed, but there is little evidence to support the notion that faster travel speeds necessarily result

63 in greater likelihood of a crash. These issues need to be sorted out in future HSM editions based on valid research. Section 3 of this report indicates that, as a practical mat- ter, few highway agencies seek design exceptions for design speed per se. Rather, design speed is usually left as equal to (or above) the posted or regulatory speed limit and design excep- tions are sought, as needed, for the individual design criteria that are affected by design speed. The FHWA Mitigation Strat- egies guide (7) states that documenting a design exception for design speed should involve analysis of every individual design element that does not meet criteria appropriate for a design speed equal to the posted or legal speed limit. In summary, consideration might be given to the following: • Treating design speed in some different way than the other controlling criteria • Discouraging or prohibiting design exceptions for design speed, and focusing the design exception process on indi- vidual design elements that do not meet established criteria for the design speed. This appears consistent with FHWA guidance and current highway agency practice • Conducting research to better quantify the relationship of traffic safety and speed 5.1.2 Lane Width and Shoulder Width Lane width and shoulder width are closely related design criteria, since they are adjacent elements of the roadway cross section. In new construction, or in the absence of design con- straints (such as existing development), lane and shoulder width can be determined independently. However, recon- struction projects often have substantial design constraints, which limit the total cross-section width. In such a constrained situation, any increase in lane width may decrease shoulder with, and vice versa. And, there is often a need to consider reducing both lane and shoulder widths to provide space for other features that clearly enhance safety such as median treat- ments, left- and right-turn lanes, bicycle lanes, parking lanes, and shorter pedestrian crossing distances. Currently Green Book policies present lane- and shoulder- width criteria independently. The lane- and shoulder-width effects on safety in the HSM are presented as independent effects, although most researchers presume that lane and shoul- der widths have effects that interact in ways that we do not yet fully understand. The HCM addresses the operational effects of lane and shoulder widths with a combined table (see Table 5) for two-lane highways, but addresses their effects separately for other facility types. There has been some research to look at combined effects of lane- and shoulder-width, including recent research on combined lane- and shoulder-width effects on safety for two-lane highways for FHWA by Gross et al. (44). However, there does not yet appear to be any work sufficiently complete and comprehensive to serve as a basis for policy on combined lane- and shoulder-width criteria. The time may come when lane and shoulder width may be dealt with by a combined set of controlling criteria for design or a combined set of operational and safety analysis procedures for use in the design process. For example, at some future time, design criteria might be established for total road- way width, with guidelines for optimally allocating space to lane width, shoulder width, and other cross-section features. However, there is not yet sufficient knowledge to achieve this goal. For the present, there appears to be a need to work toward the development of such combined operational and safety effectiveness measures for lane and shoulder width, while improving our design policies to emphasize even fur- ther the need for flexibility in allocating space to lane width, shoulder width, and other features. In summary, consideration might be given to the following: • More strongly encouraging flexibility in allocating avail- able cross-section width to lane width, shoulder width, and other features that enhance safety, particularly in recon- struction projects, by either: – Simplifying the design exception approval process for lane and shoulder width where space is provided in the cross-section width for other features that enhance safety such as median treatments, left- and right-turn lanes, bicycle lanes, parking lanes, and shorter pedes- trian crossing distances, or – Revising the design criteria for roadway cross sections to address lane width, shoulder width, and these other features in a combined, flexible design criterion • Encouraging individual agencies to develop design manuals and internal design and design exception policies to take advantage of the flexibility that is already available in the Green Book (4) and Guidelines for Geometric Design of Very Low-Volume Local Roads (ADT ≤ 400) (11) • Encouraging research to better understand the interactions among lane width, shoulder width, and other cross-section design features 5.1.3 Shoulder Width and Horizontal Clearance/Lateral Offset The controlling criteria for design include separate criteria for shoulder width and horizontal clearance/lateral offset (see Sections 2.3 and 2.13). However, these two criteria are closely related. Usable shoulder width is defined as the width of the paved or unpaved area from the edge of the traveled way to the point of intersection of the shoulder slope and mild roadside slope (for example, 1V:4H or flatter) or to the beginning of rounding

64 for slopes steeper than 1V:4H. The horizontal clearance, being renamed lateral offset in the 2011 edition of the Green Book (5), is defined as the distance immediately outside the trav- eled way, face of curb, shoulder, or other designated point to a vertical roadside element or obstruction. Given these defini- tions, the minimum shoulder widths, where provided, should ensure that the required lateral offset is available. Thus, lateral offset really becomes directly applicable to traffic operations and safety only where no shoulder is present, such as where the roadway is designed with a curb-and-gutter section. Since shoulder width and lateral offset are so closely related, their definitions as controlling criteria might be better coor- dinated, as follows: • For roadways with open cross sections, minimum shoulder widths are needed in accordance with Green Book criteria. Where the minimum shoulder widths are not provided or retained, a design exception that addresses traffic opera- tional and safety effects is needed. Where the minimum shoulder width is not provided, it is still desirable that the minimum lateral offset be provided, and a design excep- tion for lateral offset would also be needed. • For roadways with curb-and-gutter sections, minimum lat- eral offsets from the face of curb should be provided, and a design exception is needed where the minimum lateral offset is not provided. • Design exceptions for lateral offset need to address opera- tional or safety considerations only where the available lat- eral offset is less than 1.5 ft from the edge of the traveled way. Where the design exception is related to lack of a 1.5-ft lateral offset from the face of a curb or from the edge of a shoulder, but there is a 1.5-ft offset available from the edge of the traveled way, the design exception should address mitigation strategies, but need not address traffic opera- tions and safety. Otherwise, the controlling criterion would be interpreted as a clear-zone criterion, which it is not. The change in the 2011 RDG (40) to presenting a lateral offset of 4 to 8 ft rather than the previous 1.5 ft needs to be carefully considered in relation to the controlling criterion for lateral offset. This issue is addressed more fully in Section 5.1.8. 5.1.4 Bridge Width, Lane Width, and Shoulder Width It is clearly desirable at bridges to carry the full lane and shoulder width available on bridge approaches across the bridge. However, this is not always practical, especially in reconstruction projects involving existing bridges. Section 2.4 of this report discuss the provisions in AASHTO policy that allow bridges with less than the full approach lane and shoul- der widths to be constructed or retained without the need for a design exception. The research reported in Section 4.3 found no statistically significant effect of bridge-width difference on crash frequency and severity for bridges on rural two-lane highways. Thus, there is no indication of any safety concern at rural two-lane highway locations where the bridge width is narrower than the approach roadway width. This implies that existing bridges in good structural condition on rural two-lane highways can remain in place in reconstruction projects. No similar results are available for bridge widths on other roadway types. 5.1.5 Horizontal Alignment, Superelevation, and Cross Slope FHWA policy includes controlling criteria for horizontal alignment, superelevation, and cross slope. These controlling criteria actually incorporate four separate design elements: • Horizontal curve radius • Superelevation • Cross slope • Cross-slope breaks between the pavement and shoulder at the outside of superelevated horizontal curves The relationships of these four design elements to the con- trolling criteria are not always understood because (1) three of these four design elements (all but normal cross slope) are related to horizontal alignment design; (2) three of these four design elements (all but horizontal curve radius) are cross- section elements rather than alignment elements; and (3) three of these four design elements (all but horizontal curve radius) address cross-slope issues. These criteria overlap and relate to one another in such a complex manner that not everyone reading the list of the 13 controlling criteria imme- diately understands that all four of the design elements are included in the controlling criteria. The definition of the controlling criterion for horizontal alignment is not self-evident because it does not include all aspects of horizontal alignment design. It essentially includes only horizontal curve radius, because horizontal curve length and transition design details are not part of the controlling cri- teria, and superelevation and pavement/shoulder slope breaks are addressed by separate controlling criteria. Still another controlling criterion, stopping sight distance, is closely related to horizontal alignment, because roadside obstruc- tions on the inside of horizontal curves can limit stopping sight distance. It would be helpful if the definitions of the controlling criteria were realigned so that • There are separate controlling criteria for horizontal curve radius, superelevation, and cross slope (including normal cross slope and pavement/shoulder cross-slope breaks), or

65 • There is one controlling criterion for horizontal alignment that includes horizontal curve radius, superelevation, and pavement/shoulder slope breaks and a separate controlling criterion for normal cross slope Of these alternatives, the former approach appears preferable because the latter preserves the term “horizontal alignment” as a controlling criterion, even though not all elements of horizontal alignment are included in that criterion. 5.1.6 Grade, Vertical Alignment, and Stopping Sight Distance FHWA policy includes separate controlling criteria for grade, vertical alignment, and stopping sight distance. These controlling criteria are closely related and all, in one way or another, are included within the term “vertical alignment.” Grade is clearly an element of vertical alignment, and crest vertical curves, another vertical alignment element, are features that commonly limit stopping sight distance. The control- ling criterion for vertical alignment is potentially confus- ing because grade and stopping sight distance are part of, or closely related to, vertical alignment but are treated as separate controlling criteria. In fact, the only aspect of vertical align- ment design considered part of the controlling criteria, but not included explicitly within grade or stopping sight distance, is sag vertical curve length. The controlling criteria might be clearer if the term “vertical alignment” were eliminated and separate criteria were estab- lished for the following: • Grade • Stopping sight distance (which includes consideration of crest vertical curve length, sight obstructions on the inside of horizontal curves, and overpass structures) • Sag vertical curve length 5.1.7 Horizontal and Vertical Alignment Discussions in Sections 5.1.5 and 5.1.6 of this report have suggested eliminating the controlling criteria named “horizon- tal alignment” and “vertical alignment” and focusing instead on separate controlling criteria for selected critical elements of horizontal and vertical alignment design. 5.1.8 Horizontal Clearance (Lateral Offset) and Clear-Zone Width The 2004 Green Book (4) and earlier editions have speci- fied in Chapter 4 (Cross Section Elements) a horizontal clear- ance of 1.5 ft from the face of the curb to roadside objects for curbed sections on urban arterials, collectors, and local streets. The purpose of this horizontal clearance is to pro- vide an “operational offset” with sufficient space so that pas- sengers would have space to open doors of parked vehicles, even where a roadside object was present, and so that mir- rors of turning vehicles would not strike roadside objects. This horizontal clearance is one of FHWA’s 13 controlling criteria. The 2004 Green Book states that, in addition to the “opera- tional offset,” an additional clear-zone distance commensu- rate with prevailing traffic volumes and vehicle speeds should be provided where practical. FHWA guidance also states that horizontal clearance is an operational offset and is not intended as a clear-zone criterion. FHWA policy (2) states explicitly that clear-zone width or other roadside design features intended to reduce the sever- ity of run-off-road collisions are not part of the controlling criteria for design. Roadside features are designed based on guidance in the RDG (39), and roadside design, other than hardware testing addressed by the Manual for Assessing Safety Hardware (MASH) (45), is normally based on benefit-cost analysis using the Roadside Safety Analysis Program (RSAP) (46,47,48) rather than on standards. The 2011 edition of the Green Book (5), in Section 4.6.2, has renamed horizontal clearance as lateral offset. The 2011 edition no longer presents the previous 1.5-ft criterion for horizontal clearance/lateral offset, but rather refers to the 2011 RDG (40) for lateral offset criteria. Revised language in Chapter 10 of the 2011 RDG states the following criteria for lateral offset: • Lateral offset of 1.5 ft from the face of the curb should be provided where curb is used, with 3 ft lateral offset at intersections. • In high-risk urban roadside corridors, a recommended goal is to achieve 6-ft lateral offset to roadside objects on the outside of horizontal curves and at other locations and 4-ft lateral offset to roadside objects elsewhere. The accom- panying text states that lateral offsets of 12 ft on the outside of horizontal curves and 8 ft at tangent locations are rea- sonable goals where the clear-zone widths in RDG Chapter 3 cannot be achieved. • Lateral offset of 12 ft is suggested where feasible at merge points on curbed roadways in high-risk urban roadside corridors. • Lateral offset of 10 to 15 ft is suggested along the roadway within 10 to 15 ft immediately downstream of driveways. • Lateral offset should be 6 ft at intersection curb returns in high-risk urban roadside corridors with a minimum value of 3 ft. The rationale presented in the RDG for the 4- and 6-ft lat- eral offset criteria for high-risk urban roadside corridors is

66 based on the likelihood of vehicles running off the road and striking roadside objects. Thus, the 4- and 6-ft lateral offset criteria are not intended as operational offsets, but rather appear to represent “mini-clear-zone” criteria. The 4-ft lat- eral offsets along the inside of horizontal curves and down- stream of driveways are specifically shown in the RDG as providing appropriate sight distance for through or turning vehicles. These new RDG guidelines are based on research by Dixon et al. (49), who found that approximately 80 percent of road- side crashes in an urban environment involved an offset from the curb face less than or equal to 4 ft, and more than 90 percent of urban roadside crashes involved lateral offsets less than or equal to 6 ft. The research by Dixon et al. (49) specifically discusses minimizing roadside objects in the border area of urban arterials between the traveled way and the shoulder to reduce run-off-road crashes. The RDG does not define “high-risk urban roadside corridors” explicitly, but related language defines critical urban roadside locations as being best determined by identifying locations with a history of over-representation of roadside crashes. Other relevant fac- tors mentioned are operating speed, functional purpose, and other specific road features. Several issues arise with these changes: • FHWA policy identifies horizontal clearance as one of the 13 controlling criteria, but there are no longer any correspond- ing AASHTO criteria for horizontal clearance. Instead, the Green Book now uses the term lateral offset instead of hori- zontal clearance. • The current Green Book has no numerical criteria for lat- eral offset. The 1.5-ft criterion that appeared in previous editions has been replaced by a reference to the RDG. • The RDG still includes the 1.5-ft lateral offset criterion that has historically been used as FHWA’s controlling criterion for horizontal clearance, but the RDG also uses the same term, lateral offset, to refer to “mini-clear-zone criteria” for high-risk urban roadside corridors. The dual use of the term lateral offset needs clarification because FHWA docu- mentation states explicitly that the controlling criterion for horizontal clearance is not intended to include clear-zone considerations. Clarification of the language in the RDG is needed so that the RDG criteria for lateral offsets greater than 1.5 ft are not mistakenly considered as part of the controlling criterion for horizontal clearance. Potential courses of action toward resolving this issue that may be considered include the following: • Horizontal clearance could be dropped from the list of controlling criteria, as recommended in Sections 7 and 8 of this report. This would resolve the potential uncertainty about whether the RDG language on lateral offsets larger than 1.5 ft is part of the controlling criterion for horizontal clearance. • If horizontal clearance is retained as a controlling criterion, its name could be charged to lateral offset for consistency with the Green Book and the RDG, with the accompanying clarification that this controlling criterion applies only to the 1.5-ft operational offset and not to wider lateral offsets presented in the RDG intended to reduce fixed-object collisions for vehicles that run off the road. • The RDG could be changed to use different terms for the 1.5-ft offset intended as an operational offset and the wider offsets intended to reduce collisions with vehicles that run off the road. 5.2 Suggested Renaming of the 13 Controlling Criteria If all of the current controlling criteria are retained, it is suggested that they be renamed to minimize any poten- tial confusion over which design elements are, or are not, included as part of the controlling criteria. The rationale for this renaming of the controlling criteria has been presented in Section 5.1. The suggested names are the following: • Design speed • Lane width • Shoulder width • Bridge width • Structural capacity • Horizontal curve radius • Superelevation • Grade • Stopping sight distance • Sag vertical curve length • Cross slope • Vertical clearance • Lateral offset If these suggested names are used, it would be useful if the accompanying documentation made it clear that the stop- ping sight distance criterion includes stopping sight distance as limited by any roadway or roadside feature—including crest vertical curves, sight obstructions on the inside of hori- zontal curves, and overpass structures. Thus, the controlling criterion for stopping sight distance directly influences the minimum crest vertical curve length for any given algebraic difference in grade and the offset to roadside sight obstruc- tions for any curve radius on horizontal curves. The potential need to add other controlling criteria is addressed in Section 5.3. The potential need to drop specific controlling criteria is addressed in Sections 6 and 7.

67 5.3 Other Potential Controlling Criteria for Design Consideration was given in the research to whether addi- tional design elements might be considered in the future as con- trolling criteria for design. It is suggested that factors for adding a design element as a controlling criterion would include the following: • Known traffic operational and safety effects • Traffic operational and safety effects large enough to warrant inclusion as a controlling criterion for design • Within the scope of this research and the existing con- trolling criteria (which exclude design elements related to intersection design, roadside design, and access control) The research team identified three design elements with potential traffic operational or safety effects that appear to be important enough to merit consideration for inclusion as controlling criteria for design. These were • Intersection sight distance • Decision sight distance • Spacing between crossroad ramp terminals at interchanges and the nearest access point Ultimately, however, the research team’s assessment con- cluded that none of these three design criteria currently meet the tests set out by the factors above. Intersection sight distance (particularly the provision of clear sight triangles with dimensions sufficient to provide clear sight lines for drivers) is presumed by most researchers and practi- tioners to be important to safety, but no CMFs or safety effec- tiveness measures are available for intersection sight distance. In fact, research is currently being conducted under NCHRP Project 17-59 to develop such CMFs. Therefore, it cannot be stated at this time that the safety effects of intersection sight distance are known. Intersection sight distance is determined by both roadway alignment design characteristics and inter- section design characteristics, so it might also be ruled out as being an intersection design element. Decision sight distance, while having obvious implications for safety, does not have a documented CMF or a defined rela- tionship to safety. The current design criteria for decision sight distance involve a great deal of judgment in their appli- cation; two experienced designers applying the current crite- ria might reach different conclusions as to whether enhanced decision sight distance is needed in advance of a particular decision point. To serve as a controlling criterion for design, revised criteria for decision sight distance would need to be applied more objectively, and there would need to be a dem- onstrated relationship of the revised decision sight distance criteria to safety. Without more definitive design criteria and research quantifying the traffic operational and safety effects of those criteria, designation of decision sight distance as a controlling criterion for geometric design does not appear appropriate. Spacing between crossroad ramp terminals at interchanges and the nearest access point has an obvious, but unquantified, relationship to traffic operations and safety. While there are no quantified traffic operational or safety relationships, there is ample anecdotal evidence from many locations that traffic operational or safety concerns may arise where access points are located close to interchange ramp terminals. However, in the absence of quantitative relationships, and because this is both an access control and an intersection design issue, spac- ing between crossroad ramp terminals and the nearest access point does not appear appropriate as a controlling criterion for design. All other ideas for potential controlling criteria considered by the research team were rejected as being clearly related to intersection design, roadside design, or access control.