A Guidebook on Transit-Supportive Roadway Strategies (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.

66 This chapter is the second of four toolbox chapters presenting potential strategies for improv- ing bus speeds and reliability. The strategies presented in this chapter require the participation of the roadway agencies responsible for traffic control devices, and sometimes other agencies, but are generally less infrastructure-intensive than the strategies presented in the following two chapters. This chapter defines and discusses the following strategies: • Allowing buses to make movements (e.g., left turns) prohibited to other vehicles, • Restricting the ability of other vehicles to make turns, • Yield to bus, • Passive traffic signal timing adjustments, • Phase reservice, • Traffic signal shadowing, • Transit signal priority, • Transit signal faces, • Bus-only signal phases, • Queue jumps, • Pre-signals, • Traffic signal installed specifically for buses, and • Traffic control enforcement. The introduction to Chapter 5 describes how each strategy section is organized. 6.1 Movement Restriction Exemption Description Buses are allowed to make movements (e.g., left turns, right turns, proceed straight ahead) that are prohibited for other vehicles. Purpose Turning movements may be prohibited for a number of rea- sons, including: • A signalized intersection has insufficient capacity to provide a left-turn phase for general traffic; • A street has insufficient space to provide a left- or right-turn lane, and vehicles waiting to make turns would excessively delay vehicles behind them; C H A P T E R 6 Traffic Control Strategy Toolbox

Traffic Control Strategy Toolbox 67 • Roadway agency access management policies exist that divert left-turn movements to signal- ized intersections; • Boulevard-type street treatments with raised landscaped medians exist that prevent left turns; and • Allowing vehicles to make the turn could generate undesired through traffic within a neighbor- hood or district. At the same time, allowing buses to make these movements may allow a more direct routing that would save travel time or provide bus service closer to passengers’ origins and destinations. Applications Applications are in locations where the most direct bus routing is not feasible because of a turn prohibition for traffic operations reasons (e.g., delays to through traffic, cut-through traf- fic prevention), as opposed to prohibitions for safety reasons (e.g., previous intersection crash experience) or one-way street patterns, and where street widening is infeasible. Companion Strategies At unsignalized intersections, if the turn prohibition is due to a lack of gaps in the opposing traffic, and a traffic signal exists a relatively short distance downstream of the intersection, traffic signal shadowing (Section 6.6) may be an option for creating a gap. When a turn lane is provided for bus use only, red-colored pavement (Section 7.4) may be considered to help reinforce the bus-only message. Periodic enforcement (Section 6.13) may be required to maintain motorist respect for the traffic control. At signalized intersections, bus-only signal phases (Section 6.9) are suggested to be considered in conjunction with a turn exemption, potentially supplemented with transit signal priority (Section 6.7), transit signal faces (Section 6.8), or both. Exemptions from right-turn requirements are commonly used with queue jump (Section 6.10) and queue bypass (Section 8.6) strategies. Constraints • The turn prohibition may have been installed because of safety concerns that would also apply to bus movements. • The more frequent the bus service, the greater the potential delay to other traffic. • Neighborhoods may fear that allowing buses into the neighborhood will encourage other vehicles to make the turn illegally or will open the door to eventually allowing all traffic to use the street. • A formal exception to a roadway agency’s access management policy may need to be requested. Benefits Buses will save travel time equivalent to the difference in travel time via the existing routing and the travel time possible via the proposed routing. The magnitude of the benefit is highly site- specific but can be estimated by a traffic analysis. Cost Considerations • Planning and coordination costs. Moderate. A traffic engineering study will be required to evaluate the traffic and bus operations impacts of the proposed change and to evaluate poten- tial safety issues. A formal request for an exception to the roadway agency’s access management

68 A Guidebook on Transit-Supportive Roadway Strategies policy may be required. If infrastructure is being modified (e.g., creating a median opening for a bus-only turn lane), design plans will need to be developed. • Capital costs. Low to moderate, depending on the specific site characteristics. Some sites may require only replacing the existing signs; other sites may require changes to pavement mark- ings, traffic signal heads, or curb lines and medians. • Maintenance costs. Low incremental costs typically. If a painted bus lane treatment (Sec- tion 7.4) is used, the paint may need to be restored more frequently than other pavement markings. • Bus operations costs. Lower costs due to shorter distances traveled as well as potential savings from the travel time reduction achieved. • Other user costs. Depending on how the strategy is implemented, other traffic may not be affected (e.g., the bus waits in its own turn lane at an unsignalized location) or may experi- ence additional delay (e.g., extra time required to serve the bus left-turn phase at the signal, delay waiting behind a bus waiting for a gap in traffic to make its turn). The magnitude of these delays would be determined through the traffic engineering study. Implementation Examples OC Transpo has installed bus-only left-turn lanes at key intersections where there is insufficient capacity to serve automobile left turns. At an intersection where right turns would be blocked by pedestrians, right turns are prohibited, but buses on a route that turns right are allowed to make the turn. At a T-intersection with a two-lane approach (left-turn lane and right-turn lane), buses are allowed to make a left turn from the right-turn lane as a form of a queue bypass (Section 8.6). A “Bus Excepted” plaque on the lane-usage sign is used to indicate the allowed bus use. Portland, Oregon, provides a bus-only left-turn lane at a busy, complex intersection where there is insufficient capacity to serve automobile left turns. Implementation Guidance It is suggested that the transit agency first discuss the reason(s) behind the existing turn pro- hibition with the roadway agency since this will help focus the scope for a traffic engineering study to determine the feasibility of making a change for buses. Delay impacts are suggested to be quantified on both a vehicle-delay and person-delay basis. If the turn prohibition was implemented as part of a neighborhood traffic-calming program, it is suggested to meet with the neighborhood to identify potential concerns with allowing bus access into the neighborhood, present the benefits of doing so, and identify potential mitigation needs, including helping pay for enforcement if it turns out to be necessary. The community’s experience with traffic-calming measures can be relevant in assessing whether motorists can be expected to generally respect the traffic control and thus whether enforcement will be required. If a jurisdiction has no experience with signing-only violation rates for traffic-calming or transit- preferential strategies, it would be prudent to plan in advance for an enhanced level of enforce- ment if the violation rate turns out to be unacceptably high to stakeholders. See also Section 6.13. Additional Resources • Access Management Application Guidelines (Dixon et al., no date)—a companion to TRB’s Access Management Manual (Williams et al. 2014) that provides guidance on and case studies of incorporating multimodal considerations, including bus transit, into roadway agencies’ access management programs. Not yet published. • Highway Capacity Manual 2010 (Transportation Research Board 2010)—analytical methods for estimating the delay associated with turns at signalized and unsignalized intersections.

Traffic Control Strategy Toolbox 69 6.2 Turn Restrictions Description One or more existing general traffic turning movements at an intersection are prohibited. Purpose Turning movements at intersections can cause delay for buses and other intersection users when: • No turn lane is available, and vehicles wishing to continue straight must wait for a vehicle to make its turn before they can proceed; • Protected left-turn phases (i.e., left-turn arrows) are provided at a traffic signal since each additional signal phase adds additional lost time—time unusable by vehicles at the start of green and during a portion of the yellow and all-red intervals—and therefore delay (Urbanik et al. 2015); and • Turning volumes are small relative to through volumes, and the road space used by a turn lane could be more efficiently used to serve other bus or other general traffic movements, thus reducing overall delay. Selectively prohibiting turning movements can free up time or roadway space for use by buses and traffic in general. Applications Intersections where relatively low turning volumes share a lane with through traffic, and where the turning traffic experiences relatively high delays (e.g., waiting for a gap in oncoming traffic [left turns], waiting for pedestrians to finish crossing a crosswalk [right turns]), are good candidates for this strategy. In these cases, relatively few motorists will be inconvenienced, while many other roadway users will benefit. In order to free up sufficient intersection capacity to keep the intersection operating acceptably following construction of a bus lane, bus lane projects (Section 8.1) that take a travel lane may require turning movement prohibitions at intersections currently operating near capacity. Intersections where the queue in a turn lane spills over into the adjacent through lane, and the turn lane cannot be lengthened, and intersections experiencing high crash rates due to turning movements are also potential candidates for this strategy. Turning prohibitions can be implemented full time or only during peak periods. Companion Strategies To avoid the need for an indirect routing, buses can be exempted from the turn prohibition (Section 6.1). To eliminate delays to buses caused by right-turning vehicles, the strategy can be applied with queue jumps (Section 6.10) and queue bypasses (Section 8.6). It may also be needed for contraflow (Section 8.8) and reversible (Section 8.9) bus lanes to prevent potential conflicts between buses and vehicles turning across the bus lane. Right turns by large vehicles (e.g., trucks) may need to be prohibited if curb extensions (Section 7.5) would reduce the avail- able turning area too much. Right-turn prohibitions help make curbside bus lanes (Section 8.2) operate more effectively. Enforcement (Section 6.13) may be required to ensure that only buses make the restricted turn.

70 A Guidebook on Transit-Supportive Roadway Strategies Constraints A key consideration is the availability of a suitable alternate route for the diverted traffic. Suitability criteria can include traffic operations, safety, compatibility with adjacent land uses, and diversion distance. The displaced turns may be accommodated as left turns or U-turns at upstream or downstream intersections, by alternative intersection forms (e.g., jug-handle inter- sections), or by requiring traffic to travel around the block, making three right turns (AASHTO 2014). Guide signs may be necessary to communicate the desired diversion route(s) to motorists or to advise of the turn prohibition in advance of the intersection. Benefits The magnitude of the delay benefit from a turn prohibition is highly site-specific but can be esti- mated through a traffic analysis. Archived AVL data can be used to identify locations along a route where buses experience delay and determine the magnitude of those delays (Furth et al. 2006). Turn prohibitions can also produce safety benefits. For example, New York City’s evaluation of Select Bus Service on Webster Avenue in the Bronx identified that left-turn prohibitions at selected intersections not only helped traffic operations, they also addressed issues with left- turning crashes and conflicts between turning vehicles and pedestrians in the crosswalk (New York City DOT and MTA-NYCT 2014). To the extent that crashes are reduced, travel time vari- ability due to crash-caused congestion can also be reduced. Prohibiting right turns eliminates conflicts between turning vehicles and bicyclists and pedestrians. Cost Considerations • Planning and coordination costs. Moderate. A traffic engineering study will be required to evaluate the traffic and bus operations impacts of the proposed change, including for the streets and intersections likely to be used by diverted traffic, and to evaluate how inter section safety could be affected. If it appears that a turn prohibition may induce neighborhood cut-through traffic as motorists seek out alternate routes, additional planning for mitigation measures may be required along with conducting outreach to the neighborhood. • Capital costs. Typically relatively low, depending on the specific site characteristics. Some sites may simply require posting turn-prohibition signs; others may require new guide signs, changes to pavement markings, or removal of left-turn traffic signal heads. • Maintenance costs. Typically minor impacts associated with maintaining the extra signs that are required. • Bus operations costs. Potential savings from reductions in travel time and travel time variability. • Other user costs. At a minimum, this strategy will normally reduce delay to through traffic at the intersection and, depending on how it is implemented, may also benefit other traffic move- ments at the intersection. The reduced delay will typically more than offset any increased travel time experienced by the diverted traffic in cases where through traffic volumes are high rela- tive to the diverted traffic volumes. In addition, crashes associated with the prohibited turning movement should be greatly reduced, although crashes along the diversion route may go up due to the increased traffic volume. In addition, intersection delays along the diversion route may increase as a result of the increased traffic volume caused by diverted traffic. Implementation Examples A total of nine of 59 agencies responding to a survey had implemented left-turn, right-turn, or (in one case) through-movement restrictions (Boyle 2013).

Traffic Control Strategy Toolbox 71 Implementation Guidance A traffic engineering study will be necessary to quantify the impact of the proposed change on bus and general traffic operations and on safety. Delay impacts are suggested to be quantified on both a vehicle-delay and person-delay basis. AASHTO (2014) recommends prohibiting left turns when vehicles turning left would need to share a lane with through traffic because they “reduce capacity about 50 percent, delay through- vehicles, and tend to increase crashes.” Additional Resources • Guide for Geometric Design of Transit Facilities on Highways and Streets (AASHTO 2014)— Section 5.3.2.2 addresses turning movement controls. • Highway Capacity Manual 2010 (Transportation Research Board 2010)—analytical methods for estimating the delay associated with turns at signalized and unsignalized intersections. • Highway Safety Manual (AASHTO 2010)—analytical methods for estimating the effect of traffic control and roadway geometry changes on crashes. 6.3 Yield to Bus Description Motorists are required by law, or are encouraged through bus- mounted signs, to let buses back into traffic when they are signal- ing to exit a bus stop. Purpose To reduce the reentry delay experienced by buses that have finished serving passengers but then need to wait for a gap in traffic to continue on their route. Applications Yield-to-bus strategies are implemented as agency-wide measures. In states and provinces with yield-to-bus laws, the law specifies how motorists are to be notified—typically via a sticker mounted on the left rear of the bus or through a flashing or illuminated sign mounted on the left rear of the bus. Some states and provinces only apply the law to lower-speed roadways (i.e., speed limits under 35 mph). In locations without yield-to-bus laws, “Please Yield,” “Thanks for the Brake,” or a similar slogan is used on a sticker or rear advertising panel to encourage motor- ists to let buses back in (King 2003). Companion Strategies The strategy is implemented as a stand-alone measure, with periodic enforcement (Sec- tion 6.13) desirable. Installing curb extensions (Section 7.5) achieves the same purpose. Constraints To be enforceable, state legislation needs to be passed to require motorists to yield to buses exiting stops. The experience of transit agencies that take advantage of yield-to-bus laws has been that they are rarely enforced and that they might see larger benefits if they were more regularly

72 A Guidebook on Transit-Supportive Roadway Strategies enforced. Whether a suggestion or a legal requirement, an outreach campaign may be necessary to raise public awareness of the issue. Benefits Any benefit from a yield-to-bus strategy will occur at locations where the bus stops out of the traffic lane since buses normally do not have to wait for other traffic when they stop in the traffic lane (unless a queue exists). The potential benefit from yield to bus is high since reentry delay can range from 1 to 12 s, depending on traffic volumes, at bus stops well away (i.e., at least ¼ mile downstream) from traffic signals, and can be considerably higher at stops at or near signals (Kittelson & Associates et al. 2013). A Florida study recorded average reentry delays of over 30 s at some bus stops (Zhou et al. 2011). In practice, little documented benefit has been found, although some agencies have reported success with long-standing public campaigns. One study found that six of 16 agencies respond- ing to a survey felt that yield to bus had helped schedule reliability somewhat (King 2003). However, an observational study in Florida found almost no instances of motorists yielding to buses except when traffic was queued; that study did note that, unlike other states, Florida did not require transit agencies to conduct public relations campaigns prior to using yield-to-bus stickers (Zhou et al. 2011). Both studies found a transit agency and bus operator preference for electronic signs over stickers or decals. Motorists appear to be most willing to yield when traffic speeds are low (25 mph or less), with compliance increasing as speeds decrease to stop-and-go conditions. Transit agencies have generally not reported any issues with increased accidents related to buses pulling into traffic following the implementation of yield-to-bus laws (King 2003). Cost Considerations • Planning and coordination costs. Moderate to high. If a yield-to-bus law is desired and does not currently exist, some effort will be necessary to convince the state legislature to pass such a law. Transit agency experience has been that, whether a request or a legal requirement, yield to bus is more effective when accompanied by a public awareness campaign; some agencies have spent up to $250,000 to $350,000 on such campaigns (King 2003). • Capital costs. Low on a per-vehicle basis, particularly when stickers are used, but add up when the entire bus fleet is equipped. • Maintenance costs. Electronic yield signs will require extra maintenance. • Bus operations costs. Potential savings from reductions in travel time. Although not quanti- fied to date, travel time variability may increase if no one yields prior to yield to bus and some begin to yield following yield to bus. • Other user costs. Motorists who yield (and those behind them) will experience small delays as they allow buses back into traffic. Implementation Examples As of 2011, seven states (California, Colorado, Florida, Minnesota, New Jersey, Oregon, and Washington) and two Canadian provinces (British Columbia and Quebec) had passed yield-to- bus laws (Zhou et al. 2011). Implementation Guidance It is not necessary to have a yield-to-bus law in place to see benefits from a public awareness campaign; a “Thanks for the Brake” campaign that started in Vancouver more than 35 years ago and was adopted across British Columbia is reported to have “been highly successful in

Traffic Control Strategy Toolbox 73 nurturing a more friendly and courteous environment between bus operators and motorists” (King 2003). Bus drivers blink their four-way lights or give a wave out their window as thanks when motorists let them back in. The literature reports that yield-to-bus laws are generally not enforced by the police; there- fore, transit agencies may need to consider funding occasional enforcement efforts, combined with public awareness campaigns, to see meaningful benefits from yield-to-bus laws. Additional Resources • TCRP Report 165: Transit Capacity and Quality of Service Manual, 3rd Edition (Kittelson & Associates et al. 2013)—Chapter 6 provides analytical methods for estimating the magnitude of reentry delay. • TCRP Synthesis 49: Yield to Bus—State of the Practice (King 2003)—review of transit agency experiences with yield to bus. 6.4 Passive Traffic Signal Timing Adjustments Description Existing signal timing plans are optimized to reduce delay for traffic in general on inter section approaches used by buses or for buses specifically. Since the signal timing is followed whether or not a bus is present, the adjustments are considered to be passive. Purpose Signal timing that worked well when the timing plans were origi- nally developed may become less effective over time due to a variety of reasons (Urbanik et al. 2015): • Changes in traffic volumes, • Changes in traffic patterns (e.g., length of the peak periods, vehicle mix), • Changes in roadway geometry (e.g., new turn lane, relocated bus stop), and • Changes in pedestrian volumes (e.g., resulting from new development in the area). Therefore, reviewing existing signal timing is an activity that roadway agencies should under- take on a periodic or ongoing basis, although they may not always do so due to a lack of resources or other reasons. Optimizing traffic signal timing is done to achieve desired roadway agency goals such as minimizing the number of stops or traffic signal delays experienced by vehicles traveling along a street. Changes that result in better operations for automobiles may also benefit buses, although good signal timing for automobiles is not necessarily good signal timing for buses. Signal timing can also be adjusted specifically to benefit buses. Some of these changes, such as shorter cycle lengths or more green time for the approaches used by buses, will also improve operations for many other roadway users. Other changes, such as signal timing designed to allow buses to progress, may benefit some modes and disbenefit others. Applications Typical passive traffic signal timing adjustments that benefit buses include: • Signal retiming in a corridor to reduce delay to through traffic, including buses; • Reducing intersection cycle lengths to reduce the amount of delay experienced by buses when they do have to stop for a red signal; and

74 A Guidebook on Transit-Supportive Roadway Strategies • Allocating more green time to approaches used by buses (which can potentially include minor-street approaches and left-turn lanes). Companion Strategies Passive traffic signal timing adjustments can be implemented in conjunction with other sig- nal timing strategies that react to the presence of a bus, including phase reservice (Section 6.5), transit signal priority (Section 6.7), bus-only signal phases (Section 6.9), queue jumps (Sec- tion 6.10), and pre-signals (Section 6.11). Constraints The amount of signal time that can be reallocated to approaches served by buses will be con- strained by the amount of time required to serve vehicles on other approaches (dependent on traffic volumes and the number of lanes) and by the minimum time required to serve pedestrian movements (dependent on the crossing width and the minimum pedestrian walk times specified in the MUTCD and roadway agency policy). In corridors where the signals are coordinated (i.e., operate as group, allowing traffic move- ments to be synchronized between intersections), a common cycle length will be used. Making a change to one intersection’s cycle length will normally require all of the other intersections’ cycle lengths to be changed identically. (A potential exception is an intersection that can operate acceptably with a cycle length that is half that of the other coordinated intersections, thus pro- viding a green twice as often as the other intersections—an operation known as double cycling.) Thus, the operations of all the coordinated intersections will need to be considered when con- sidering changing the cycle length (Urbanik et al. 2015). Benefits The potential benefit from signal timing adjustments will depend on the quality of the existing timing and intersection- and corridor-specific conditions that include traffic volumes, traffic pat- terns, vehicle mix, and traffic signal spacing, among others. Transit signal priority benefits reported in the literature tend to be greater in cases when traffic signals had not been recently retimed, suggesting that retiming signals on a regular basis is an action that will provide benefits both to transit and general traffic. NCHRP Report 812: Signal Timing Manual (Urbanik et al. 2015) describes tools available for timing signals and determining the effects on multimodal operations. In general, a shorter cycle length will reduce delay for buses and pedestrians, as long as the intersection can continue to operate under capacity. Delay experienced by vehicles and bicyclists may go up or down. Allocating more green time to intersection approaches used by buses will reduce delay for motorized vehicles, bicyclists, and (potentially) pedestrians on those approaches, but will increase delay to motorized vehicles and bicyclists on the approaches whose green time is shortened. (Side-street pedestrian delay will only increase if its walk time is reduced or the cycle length increased, but often only the minimum walk time is provided, to allow the signal to stop serving the side street when no side-street vehicles remain to be served). Timing traffic signals to allow buses to progress, accounting for typical dwell times at stops, can significantly reduce the traffic signal delays and extra stops experienced by buses but will likely increase the delay experienced by motorized vehicles on the intersection approaches used by buses. It is thought that bicyclists may also benefit from bus signal progression in some cases since the two modes have similar average speeds in urban (non-downtown) conditions, but research is lacking in this area.

Traffic Control Strategy Toolbox 75 Cost Considerations • Planning and coordination costs. Moderate for a corridor, requiring collecting traffic demand data at each intersection (including for buses, bicyclists, and pedestrians) along with existing signal timing, using an appropriate set of tools to develop an initial timing plan and to evaluate the effects of that plan, and finally, implementing and adjusting the final plan in the field. Bus operators can be a good source of information about intersections or corridors where signal timing improvements may be useful. • Capital costs. No change. • Maintenance costs. No change. • Bus operations costs. Potential savings from reductions in traffic signal delay. Timing signals to allow buses to progress can also help reduce bus travel time variability. • Other user costs. Potential changes in delay (both up and down) for other intersection users, as discussed in the Benefits section; however, the outcome should be an overall improvement on a person-delay basis. Implementation Examples TCRP Synthesis 110 reports that 14 of 59 transit agencies responding to a survey had experi- enced traffic signal timing changes. Seven of these agencies measured the impact on bus speeds, with three experiencing 5% to 10% increases in speeds, three experiencing 0% to 5% increases in speeds, and one experiencing a decrease in speeds after the traffic signal cycle length was increased. Three of the 14 transit agencies reported that they had a formal process in place with one or more roadway agencies to raise potential signal timing issues and that the roadway agencies generally made changes when feasible. San Francisco has changed traffic signal timing to allow buses to progress on two corridors; Ottawa times its traffic signals to allow buses to progress through downtown (Boyle 2013). OC Transpo staff evaluate intersection operations to identify whether shorter signal cycles or more green time for bus movements can be accom- modated. Many transit signal priority implementations reported in the literature included traffic signal timing optimization (see TCRP Web-Only Document 66). Implementation Guidance It is suggested that transit agencies develop both an internal process for identifying and report- ing signal timing issues that affect bus operations (e.g., bus operator reporting) and a formal process with their roadway agency partners for submitting those issues for investigation and potential action. Transit agencies may wish to consider proposing (and potentially funding) a pilot project on a major bus corridor to make signal timing improvements for buses, including evaluating the effects on bus operations and other roadway users. A successful pilot project can lead to increased attention on the part of the roadway agency to considering bus operations needs when retiming traffic signals. Signal progression for buses is a potential strategy for high-passenger-volume corridors where a net person-delay benefit may be feasible. It can also be considered in communities that wish to prioritize non-automobile traffic. Additional Resources • Manual on Uniform Traffic Control Devices (FHWA 2009)—Section 4E.05 addresses pedes- trian signal timing requirements.

76 A Guidebook on Transit-Supportive Roadway Strategies • NCHRP Report 812: Signal Timing Manual, 2nd Edition (Urbanik et al. 2015)—presents traf- fic signal and signal timing concepts, provides guidance on developing signal timing plans, and describes tools for timing signals and estimating the impacts of signal timing plans. 6.5 Phase Reservice Description A traffic signal phase is served twice during a traffic signal cycle—for example, a left-turn phase that is served both at the start and the end of the green phase for through traffic. Purpose Serving a phase twice per cycle minimizes the time a bus has to wait to be served and thereby reduces bus travel time variability. It accom- modates varying bus arrival times at a traffic signal (e.g., caused by varying dwell times at an upstream stop) better than serving a phase only once per cycle. Applications Potential applications for phase reservice to benefit buses are: • Serving peak-direction major-street bus left turns during peak periods on roadways that have highly directional traffic flows (e.g., mostly inbound toward downtown in the morning); • Serving major-street bus left turns or bus movements on minor-street or driveway approaches (e.g., serving transit centers or park-and-rides) to the traffic signal during periods of low to moderate volumes on the major street; • As a substitute for double cycling (see Section 6.4, Constraints) at minor intersections, when half the normal cycle length would not produce good operations; and • Serving special bus phases (Section 6.9) or queue jumps (Section 6.10) (Urbanik et al. 2015, Corby et al. 2013). Phase reservice can be made conditional on the presence of a bus or a predetermined number of vehicles. Companion Strategies Phase reservice can be considered in conjunction with special bus phases (Section 6.9) or queue jumps (Section 6.10). Constraints Phase reservice is an advanced feature that may not be provided by the intersection’s current traffic signal controller. This strategy requires that underutilized green time be available within the traffic signal cycle that can be used to reserve a phase (e.g., relatively low traffic volumes in the opposite direction of travel when reserving a left-turn phase).

Traffic Control Strategy Toolbox 77 Benefits Bus delay will be reduced, as will the delay experienced by other vehicles sharing the inter- section approach. The amount of time saved will be site-specific, but average movement delay reductions in the range of 10 to 30 s have been reported in the literature (e.g., Corey et al. 2013, Lavrenz et al. 2015). Bus travel time variability will also be reduced, although this is more dif- ficult to quantify without a fairly extensive simulation model that includes upstream and down- stream intersections. Cost Considerations Many costs are dependent on whether a new signal controller would be required. • Planning and coordination costs. Relatively low for an intersection equipped with a suitable controller, requiring collecting traffic demand data at each intersection (including for buses, bicyclists, and pedestrians) along with existing signal timing, evaluating the effects of phase reservice on intersection operations, and implementing in the field. Moderate when a new signal controller is required. • Capital costs. Potentially no change if a suitable controller already exists; moderate if a new controller will be required. Some additional vehicle detection capability may be required to implement phase reservice conditionally, which entails relatively low costs. • Maintenance costs. No change unless a new controller is required and it is the roadway agency’s first advanced controller, in which case staff training will be required. • Bus operations costs. Potential savings from reductions in traffic signal delay and bus travel time variability. • Other user costs. Phases whose green times are shortened to provide phase reservice will experience greater vehicle delay. A net vehicle-delay benefit will be more likely to occur as traffic volumes served by the reserved phase increase and as traffic volumes served by the phase(s) with reduced green time decrease (Corey et al. 2013). Pedestrian crossing delay will increase on the crosswalk conflicting with a reserved left-turn phase if its walk time is reduced. Implementation Examples Ottawa has applied conditional phase reservice. When two to three cars or a bus occupy a left- turn lane, the left turn is served twice within the cycle, both as a leading left turn and as a lagging left turn. This strategy was already being used for non-transit applications (clearing queues of cars), so no special negotiating was needed with the city transportation department to use it for buses, subject to the normal checks that there was sufficient capacity available to accommodate the extra interval. City staff have not observed any driver expectancy issues with the use of this treatment. It is only used during the morning peak period (6 a.m. to 9 a.m.). Copenhagen, Denmark, uses phase reservice at a few intersections to provide queue jumps before and after parallel traffic is served. Implementation Guidance In a transit context, this strategy has greatest potential application to signalized intersections where buses turn left. Additional Resources • NCHRP Report 812: Signal Timing Manual (Urbanik et al. 2015)—Sections 9.2.3 and 12.3.1.4 address phase reservice.

78 A Guidebook on Transit-Supportive Roadway Strategies 6.6 Traffic Signal Shadowing Description A bus wishing to turn left at an unsignalized intersection triggers a call for a left-turn phase at a nearby downstream intersection, thereby creating a gap in traffic that the bus can use to turn left. Purpose When opposing traffic volumes are sufficiently high, and right- turning traffic from driveways or the downstream cross street fill available gaps before a bus can use them, buses may experience significant delays waiting to turn left at an unsignalized location. Applications Left turns from a major street or left turns out of a minor street or driveway at an unsignalized intersection. Examples include those at transit centers, park-and- ride lots, and shopping centers. Figure 2 illustrates the process for a left turn from a major street into a cross street; a similar process can be used for turns from a minor street into a major street. In Figure 2 (1), a bus arrives at the unsignalized intersection and is blocked from making a left turn by oncoming traffic. The bus is detected in the left-turn bay (e.g., using a transponder or video detection, or a normal loop detector if only buses are allowed to make the turn) and a call is placed for the left-turn phase at the downstream intersection, even when no vehicles are waiting to make the left turn. In Figure 2 (2), the left-turn call is served, stopping the flow of oncoming traffic in the process. Right turns on red need to be prohibited on the cross street at the traffic signal to ensure that a gap is formed. The right-turn-on-red prohibition can be permanent or can be implemented only when needed by activating a blank-out sign. In Figure 2 (3), the gap has reached the unsignalized intersection and the bus can make its turn. Companion Strategies This strategy can be used in conjunction with turn restrictions (Section 6.2); for example, a bus-only turn into a transit center or park-and-ride lot. It can also be combined with transit signal priority (Section 6.7) to serve the left-turn phase sooner than usual. Traffic signal shadowing is considered only as a strategy because the MUTCD does not cur- rently provide warrants for traffic signals installed specifically for buses (Section 6.12). In most cases, a traffic signal would be the more straightforward option and could also serve other needs, such as providing a safer pedestrian crossing opportunity on busy streets that have a long dis- tance between traffic signals. Constraints This strategy requires a nearby downstream traffic signal that provides a protected left-turn phase (i.e., left-turn arrow), the ability to prohibit right turns on red from the cross street at that traffic signal, no or few driveways between the traffic signal and the unsignalized intersection, and a means of detecting buses. The ability to serve the left-turn phase early is constrained by the requirement to provide a minimum pedestrian crossing time on the conflicting crosswalk. Source: © 2015 Google

Traffic Control Strategy Toolbox 79 Benefits The delay benefit to buses is site-specific and would need to be determined from a traffic engineering study. Cost Considerations • Planning and coordination costs. Moderate. A traffic engineering study is needed to deter- mine whether shadowing is the most appropriate strategy and what the impacts would be. If the strategy is determined to be feasible, design plans would need to be developed for the detection connection to the downstream traffic signal. • Capital costs. Moderate. A means will be needed to detect the bus, and those detections will need to be communicated to the signal controller at the downstream traffic signal. • Maintenance costs. Small increase in costs related to the detection system. • Bus operations costs. Potential savings from reductions in traffic signal delay and bus travel time variability. BU S BU S BUS ① ② ③ Figure 2. Example of traffic signal shadowing.

80 A Guidebook on Transit-Supportive Roadway Strategies • Other user costs. Delay may be increased for traffic from the opposite direction if no left- turning vehicles would have needed to be served or when the left-turn phase is served earlier than normal. Prohibiting right turns on red may increase delay for that movement if gaps in traffic and pedestrian crosswalk activity would have otherwise permitted right turns on red to occur. Calling a left-turn phase early may increase pedestrian delay on the conflicting cross- walk if the crosswalk’s walk time is reduced as a result. Implementation Examples TriMet uses traffic signal shadowing at the Barbur Transit Center in Portland, Oregon. The signalized intersection includes the bus entrance to the transit center, while the unsignalized intersection is the bus exit. When buses need to leave, the left-turn phase is called at the upstream traffic signal, creating a gap in northbound traffic that right-turning buses can use immediately. If a gap also happens to exist in southbound traffic, left-turning buses can complete their turn immediately; otherwise, they can pull into a center two-way left-turn lane and wait for a gap in southbound traffic before proceeding. No right-turn-on-red prohibition is required in this instance because the transit center driveway is a one-way entrance to the transit center from the traffic signal. Calgary and Edmonton, Canada, use a form of traffic signal shadowing at certain inter sections with half signals (i.e., where pedestrian crosswalks are signalized at an intersection, but cross- street traffic is stop-controlled). When a bus arrives at the intersection, the pedestrian crossing phase is called, whether or not pedestrians are present, creating a gap in traffic that the bus can use to turn onto the main street. However, this approach is not currently permitted by the MUTCD since half signals are not allowed. Implementation Guidance Before pursuing traffic signal shadowing as an option, first consider whether other solutions to the problem are feasible. Would a traffic signal be warranted at the location for general traffic reasons (e.g., the pedestrian volume, coordinated signal system, or crash experience warrants)? Can the bus route use a different set of streets to avoid the unsignalized intersection? The following characteristics make a site a potential candidate for a traffic signal shadowing treatment: • Traffic volumes that create substantial delay for turning buses; • Presence of a nearby traffic signal with a left-turn phase that can create a gap in traffic; • Ability for the traffic signal controller to distinguish buses from other turning traffic; • No driveways or only low-volume driveways located between the traffic signal and the loca- tion where bus turns occur so that other vehicles do not fill the gap that is created; • Low pedestrian and bicycle activity so that these road users in most cases do not prevent buses from using the gap that is created; and • For left turns from a cross street onto a major street, existence of a two-way left-turn lane or similar refuge area that buses can use as needed to complete their turns in two stages. The Transportation Association of Canada guidelines suggest using, in situations where traf- fic signal shadowing might be used, a “traffic signal required by transit” (Section 6.12) installed solely to serve transit needs that might not be justified by general traffic needs (Corby et al. 2013). The guidelines note that not all Canadian jurisdictions may permit a traffic signal for transit purposes and also note that some Canadian jurisdictions permit the use of half signals that serve both pedestrian and transit needs. As discussed in Section 6.12, at the time of writing,

Traffic Control Strategy Toolbox 81 a provision for traffic signals required by buses was being considered for inclusion in a future edition of the MUTCD. Additional Resources • Manual on Uniform Traffic Control Devices (FHWA 2009)—information on how and where traffic signals, transit signals, and pedestrian hybrid beacons may be used (Parts 4 and 8). 6.7 Transit Signal Priority Description Traffic signal timing is altered in response to a request from a bus so that the bus experiences no or reduced delay passing through the intersection. Purpose Traffic signal delay contributes significantly to slower bus speeds and greater travel time variability, particularly with greater numbers of traffic signals encountered along a route and with longer cycle lengths (i.e., longer waiting times) at those signals. TSP strategies are designed to reduce traffic signal impacts on bus travel speeds and travel time variability. Applications Transit signal priority can be applied in several ways: • Green (phase) extension. If a bus is detected close to the end of the green phase for the bus’s intersection approach, the green phase is extended to allow the bus to pass through the inter- section, thus allowing the bus to avoid a lengthy delay waiting for the next green. Depend- ing on how TSP is implemented and the bus detection capabilities provided, the length of the extension can be a fixed amount for every bus, or the extension can be ended when it is detected that the bus has cleared the intersection. • Red truncation (early green). If a bus is detected stopped at the intersection, conflicting phases (e.g., the side-street green) are ended early, and a green is provided to the bus’s approach sooner than would have occurred otherwise, thus reducing the amount of delay the bus experiences. • Phase insertion. A special phase is provided to serve the bus when it is detected. This applica- tion is typically used in conjunction with turn lanes serving buses only (e.g., a left turn into a transit center—see Section 6.1, Movement Restriction Exemption), special bus phases serving nonstandard movements (Section 6.9), and queue jumps (Section 6.10). • Sequence change. The order in which phases are served is altered to serve the bus sooner than would occur otherwise. For example, a lagging left-turn phase can be switched to a leading left-turn phase when a bus is detected. • Phase skipping. Phases are skipped (i.e., not served) so that service can return to the phase serving the bus more quickly (Urbanik et al. 2015). Depending on the capabilities of the traffic signal controller, more than one of these applica- tions can be implemented at a given intersection.

82 A Guidebook on Transit-Supportive Roadway Strategies Two key aspects of each of these TSP applications are that (1) signal timing changes are made within the context of the normal signal timing plan, and (2) priority may not necessarily be granted to a bus making a request. A related strategy, transit signal preemption, was used early on as a transit-preferential strategy. However, because preemption interrupts the normal signal timing plan to immediately serve a specific request for service, it interrupts any traffic signal coordination that might be provided, can cause significant delays to other intersection users, and can terminate pedestrian phases while pedestrians are still crossing the intersection. As a result, preemption is generally used today only in conjunction with railroad crossings, drawbridges, and emergency vehicles. Priority is typically used to serve buses, trucks, and other preferred vehicles since it maintains signal coordination and provides minimum pedestrian crossing time, thereby producing much smaller impacts on intersection operations (Urbanik et al. 2015). Chapter 10 of the NCHRP Report 812: Signal Timing Manual (Urbanik et al. 2015) provides details about how each TSP application can be implemented; there are various ways to generate a request for service, detect bus locations, and process service requests. The general process is described as follows, and is illustrated in Figure 3: • Upstream detection. Either the traffic signal system detects the bus’s presence directly, or the bus is aware of its position and sends a request for service at an appropriate time. • Transition selection. The signal controller or the central system receives the request and decides how or whether to serve it, depending on when in the traffic signal cycle the request is received and whether higher-priority requests have been received. If the request is granted, the controller or central system decides how best to implement it, based on a number of fac- tors (e.g., cycle length, traffic detected on other approaches, provision of minimum pedestrian phase lengths) that can potentially be programmed by the roadway agency. Source: Urbanik et al. (2015). Figure 3. Transit signal priority process.

Traffic Control Strategy Toolbox 83 • Timing transition. Changes to the signal timing are implemented to serve the request as soon as safely and operationally feasible. • Dwell stage. The priority request continues to be served until (1) the bus is detected at a downstream location or (2) the maximum time allotted for serving the priority request has expired. Other phases that do not conflict with the priority phase can also be served. • Recovery. The signal timing is adjusted to restore normal operations as soon as safely and operationally feasible. Depending on how TSP is implemented, normal operations can resume as soon as the next cycle; in other cases, it may take several cycles. Requests for priority can be made, prioritized, and granted based on a variety of conditions, including: • Schedule (e.g., late buses are granted priority; on-time or early buses are not); • Bus status (e.g., in service or out of service, door open or door closed, on route or off route); • Direction of travel (e.g., peak direction or off-peak direction); • Passenger load (e.g., buses with more passengers can be prioritized); • Level of priority (e.g., BRT or frequent-service routes versus low-frequency routes, emer- gency vehicles versus buses); and • Number of cycles since the previous granting of priority (if more than one cycle is needed for the intersection to recover to normal operations) (Urbanik et al. 2015). The degree to which a traffic signal controller can prioritize requests (if at all) depends on the controller’s capabilities; advanced controllers provide much more flexibility. When buses initiate priority requests themselves, some of the intelligence can be placed onboard the bus (e.g., bus status and schedule adherence information) and used in deciding whether to make a priority request. Companion Strategies The potential need for stop relocations (Section 5.1) should be considered when implement- ing TSP since some applications work better with some stop locations than with others (e.g., green extension with far-side stops); see the Implementation Guidance subsection for more information. TSP can also be combined with most signal-related strategies, including passive signal timing adjustments (Section 6.4), traffic signal shadowing (Section 6.6), bus-only signal phases (Section 6.9), turn lanes serving buses only (Section 6.1), queue jumps (Section 6.10), and pre-signals (Section 6.11). Constraints Implementing TSP can be a significant capital investment because traffic signal controllers may need to be upgraded, bus detection capabilities at intersections may need to be improved, and (depending on the type of system implemented) equipment may need to be provided onboard a portion or all of the bus fleet. A key requirement for TSP to be successful is that buses actually be able to reach the inter- section to take advantage of it. If an intersection approach operates over capacity, it may make matters worse to adjust the signal timing when the bus is blocked by other vehicles since the bus cannot get to the intersection and may not be granted priority again until the signal recovers from the first granting of priority. Pre-signals (Section 6.11) can be considered to manage queues at an intersection and give buses unrestricted access to the intersection while continuing to serve as many other vehicles on the approach as possible, given the over-capacity conditions. As with other strategies that reduce signal delay at individual intersections, the overall effec- tiveness of TSP is typically less than the sum of the individual intersection delay savings because

84 A Guidebook on Transit-Supportive Roadway Strategies in some cases the bus will simply arrive early at the next traffic signal and wait (see Section 4.4). As discussed in the Benefits section, in some instances no significant travel time benefit has been achieved on a corridor basis. In other cases, TSP has been quite successful at improving bus speeds. Therefore, because this strategy’s potential costs are significantly higher than those of most other transit-supportive roadway strategies, prior to committing to implement TSP, it is particularly important to evaluate the potential benefit of it in the context of the signal controller capabilities that will be provided, the traffic signal spacing, bus stop locations, and other corridor characteristics. Granting signal priority to buses may require changes in state traffic laws or administrative rules. Specific applications, such as phase skipping, might not be permitted by state or local policy. Benefits A challenge with evaluating implemented TSP projects is that they are usually implemented in conjunction with other transit-supportive roadway strategies, and it is difficult from field studies to separate out the impact of TSP relative to (for example) signal timing plan optimization or bus stop relocation unless each strategy is implemented sequentially. Therefore, simulation is often relied on to evaluate the specific impact of TSP on bus travel times. TCRP Project A-39 simulated the effects of a number of strategies, including TSP, at an iso- lated intersection. With a near-side stop, providing up to 10 s of green extension or red trunca- tion (TSP) in the peak direction reduced average bus delay by 3, 3, and 10 s and average vehicle delay on the approach by <1, 1, and 2 s, with intersection v/c ratios of 0.5, 0.8, and 1.0, respec- tively. With a far-side stop, TSP reduced average bus delay by 3, 4, and 6 s and average vehicle delay by <1, 1, and 3 s for the same v/c ratios. Implementing TSP in both directions resulted in similar or slightly worse results for bus delay and generally substantially worse results for vehicle delay. The best package of strategies was to move a near-side stop to the far side and implement TSP in one direction, which provided average time savings of 6 to 15 s for buses and 1 to 3 s for vehicles on the same intersection approach. Cesme et al. (2015) found similar results via simula- tion, with 4- to 12-s average bus delay savings with 10 s of green extension or red truncation, and 6 to 18 s of savings with 15 s of green extension or red truncation. However, when TCRP Project A-39 simulated providing up to 10 s of green extension and red truncation along a 1.3-mile-long corridor with nine traffic signals, average bus travel times through the corridor were only reduced by 9 s in the best-case scenario (TSP applied at all signals), or an average of 1 s per intersection. When TSP was implemented at just the three moderate-volume intersections (those with v/c ratios of between 0.6 and 0.9), corridor travel times were reduced by about 7.5 s, or an average of 2.5 s per intersection. Implementing TSP at moderate-volume intersections and at all intersections resulted in 8% and 9% reductions in travel time variability, respectively (Ryus et al. 2015). Turning to actual transit agency implementations of TSP, Smith et al. (2005) present eight case studies of TSP. In six cases, bus travel times through a corridor decreased by 9% to 16%; in the other two cases, no significant travel time reductions were observed, but travel time vari- ability decreased. Gardner et al. (2009) provide similar data for 12 international cities. In eight cities, average bus travel times decreased by 4% to 19% (and up to 24% in one specific corridor). In four cities, there was no significant change in bus travel times, but travel time variability decreased in three of those cities. Reductions in travel time variability were also observed in four other cities that did not quantify changes in bus travel times. Studies of TSP in Portland, Oregon (Kimpel et al. 2005, Smith et al. 2005, Koonce et al. 2006) have found limited travel time savings from TSP, but did find travel time variability improvements.

Traffic Control Strategy Toolbox 85 The reasons why TSP does not always provide travel time benefits have not been well-quantified to date, but potential reasons include: • Too few priority calls, whether due to too-restrictive conditions (e.g., high thresholds for being behind schedule) or incorrect programming of the priority logic in the signal controller; • No change made to bus schedules to take advantage of potentially faster travel times, thus locking the scheduled travel time in place but reducing the number of late buses; • Locations for detecting the bus located inappropriately for the selected TSP application (e.g., providing priority for buses still serving passengers at a near-side stop); • Too much traffic congestion, so that buses could not take advantage of any priority granted; • Not enough traffic congestion, so that buses experienced relatively little delay at signals prior to implementing TSP; and • Signal spacing too frequent, so that time saved at one intersection was spent waiting at a downstream intersection, with no net change in travel time (Gardner et al. 2009, Albright and Figliozzi 2012, Feng et al. 2015, Ryus et al. 2015). Both simulation results and actual implementations have found that TSP typically reduces delay for traffic on the intersection approaches used by buses (typically the major-street approaches) and produces negligible to minor increases in side-street delay (Smith et al. 2005, Ryus et al. 2015). Cost Considerations • Planning and coordination costs. Moderate to high. Traffic demand data need to be collected for each intersection (including for buses, bicyclists, and pedestrians) along with existing signal timing, the logic for granting priority needs to be determined (if not already developed by prior implementations), an initial timing plan needs to be developed and evaluated, and the final plan needs to be implemented and adjusted in the field. If TSP is a new strategy for the roadway agency or the transit agency, the TSP infrastructure will need to be planned and designed. Performing a simulation study of corridor operations with and without TSP is suggested. • Capital costs. Variable, depending on how TSP is to be implemented and how much of the required infrastructure already exists (e.g., the roadway agency has already installed advanced signal controllers for other reasons), but is typically high when starting from scratch. • Maintenance costs. Variable, depending on how TSP is to be implemented and how much of the required infrastructure already exists. Roadway agency costs will likely increase; transit agency costs will also increase when some of the TSP infrastructure is placed on buses. There will also be staff training costs associated with introducing TSP to a jurisdiction. • Bus operations costs. Bus travel time variability is typically reduced. Depending on how TSP is implemented (e.g., schedules adjusted or not) and corridor-specific conditions, there may also be a reduction in bus travel times. • Other user costs. Vehicular traffic on the approaches served by TSP will typically experience a small average delay reduction due to TSP, while vehicular traffic on the approaches not served by TSP will typically experience a negligible to small delay increase. If the cycle time and pedes- trian walk times are not changed, pedestrian delay will be unchanged. Phase skipping can significantly increase vehicular and pedestrian delay for those approaches that are skipped. Implementation Examples TSP has been implemented on a large scale in Houston, Texas; Sacramento and Los Angeles, California; and Portland, Oregon, among other cities. It has also been implemented on one or more corridors (including many BRT routes) in numerous other cities; Smith et al. (2005) identified at least 24 transit agencies that had implemented TSP at the time of their research.

86 A Guidebook on Transit-Supportive Roadway Strategies Implementation Guidance As discussed in the Benefits section, TSP does not always provide a significant corridor-level travel time benefit. There it is suggested that the transit agency first investigate lower-cost, quicker-to-implement strategies that may provide as great or greater benefits than TSP. In par- ticular, stop relocations (Section 5.1), stop consolidations (Section 5.2), and passive traffic signal timing adjustments (Section 6.4) offer good potential for time savings. An exception to this guidance would be when a transit agency is developing a BRT project; in this case, TSP becomes one component of a major investment project. Interviews conducted for TCRP Project A-39 showed that roadway agency staff were often concerned about the possible impact of TSP on traffic and pedestrian operations. Successful approaches that transit agencies took to overcome these perceptions included: • Commissioning a traffic analysis to evaluate corridor operations with TSP in place, • Lending spare TSP equipment to roadway agency staff to experiment with in their signal shops, and • Taking roadway agency staff on study trips to meet with their peers in cities with similar TSP implementations to learn from their experiences. Fire departments may also be concerned that providing signal priority for buses may interfere with priority operations for emergency vehicles; in these cases, ensuring that the system is capa- ble of prioritizing emergency vehicles over buses can overcome these concerns. If traffic signals do not currently have the ability for emergency vehicle preemption, then upgrading traffic sig- nals to serve both buses and emergency vehicles can generate stakeholder support for the project. Interviews conducted for TCRP Project A-39 showed that successful first experiences with TSP paved the way for an expansion of its use on subsequent projects. The following characteristics make a corridor more suitable for TSP: • Peak-period intersection v/c ratios of between 0.6 and 0.9, such that intersections operate below capacity but with sufficient traffic demand that buses experience significant delays at traffic signals; • High existing transit ridership or the potential for higher ridership with service improvements; • Sufficient bus volumes to justify the investment, but not so high that TSP would be called nearly every cycle (in which case, passively adjusting the signal timing to provide an equivalent benefit would be more appropriate); in general, corridors with at least four buses per hour per direction; and • Primarily far-side bus stops or stops that can be relocated to the far side (AASHTO 2014, Ryus et al. 2015). Note that these characteristics are guidelines and not hard-and-fast rules. For example, inter- sections with relatively low v/c ratios may still benefit from TSP if the street is wide and mini- mum pedestrian crossing times dictate a long cycle length, with the result that buses experience considerable delay when they must stop for a red light at times when cross-street pedestrians are present. If traffic signals are already TSP-capable (for example, to serve emergency vehicles), then the relatively low incremental cost of implementing TSP may justify doing so at all signals in the corridor or on corridors with lower bus volumes. Other considerations related to TSP are: • Providing TSP only in the peak direction provides similar bus delay savings but better general traffic delay savings than providing it in both directions. • Conditional application of TSP allows the system to be used more effectively—for example, by only granting priority to late buses, by prioritizing routes at intersections with bus service on both streets, and by prioritizing buses on the basis of passenger loads.

Traffic Control Strategy Toolbox 87 • Adjust the schedule to take advantage of any time savings provided by TSP; otherwise, giving priority only to late buses will lock the existing bus travel time in place. In this respect, TSP is well-suited for routes with headway-based schedules since conditional TSP can be configured to maintain bus headways while still providing a speed benefit. • TSP implementations that change the order in which pedestrian crosswalks are served (e.g., phase skipping, sequence change) can be confusing for visually impaired pedestrians. Acces- sible pedestrian signals are suggested to be used to indicate to these pedestrians when their crosswalk is being served. Note that the U.S. Access Board’s Proposed Guidelines for Pedes- trian Facilities in the Public Right-of-Way (2011) would require accessible pedestrian signals to be installed in any event at an existing signal when the signal controller and software are altered, which is often the case when TSP is implemented. Additional Resources • Guide for Geometric Design of Transit Facilities on Highways and Streets (AASHTO 2014)— Section 5.8.1 covers transit signal operations for transit priority. • Manual on Uniform Traffic Control Devices (FHWA 2009)—Section 4D.27 addresses priority control of traffic signals. • NCHRP Report 812: Signal Timing Manual (Urbanik et al. 2015)—Chapter 10 addresses traffic signal priority in general and transit signal priority specifically, from the signal timing standpoint. • Transit Signal Priority (TSP): A Planning and Implementation Handbook (Smith et al. 2005)— presentation of a systems engineering approach to planning, designing, and implementing TSP, suggested roles and responsibilities for stakeholders in a TSP system, and case study implementation examples. 6.8 Transit Signal Faces Description Special traffic signal faces (displays) used for controlling bus, streetcar, or light rail operations. Purpose Transit signals can help reduce the possibility that road users will mis- interpret regular traffic signals designed to control transit vehicles as apply- ing to them, leading to potential conflicts. Applications The MUTCD (Section 4D.27) identifies the following applications for transit signal faces for buses when engineering judgment indicates that using these signal faces in place of standard red/yellow/green signal faces would reduce road user confusion: • Public transit buses in queue jump lanes, and • Control of bus rapid transit in mixed traffic and bus lanes (FHWA 2009). MUTCD Figure 8C-3 illustrates the various transit signal faces. According to the MUTCD, the go signal can be used in a flashing mode to indicate “prepare to stop” when two faces are used, or a triangle symbol can be used when three faces are used. A diagonal bar indication can be used to indicate that a turn should be made (FHWA 2009). In general, transit signal faces are potentially applicable in situations where transit vehicles are allowed to move at different times than parallel traffic.

88 A Guidebook on Transit-Supportive Roadway Strategies Companion Strategies Transit signal faces are used to support other strategies that involve giving buses a head start on other traffic or moving from potentially unexpected locations. These strategies include bus-only signal phases (Section 6.9), queue jumps (Section 6.10), pre-signals (Section 6.11), traffic signals installed specifically for buses (Section 6.12), queue bypasses (Section 8.6), median bus lanes (Section 8.7), contraflow bus lanes (Section 8.8), and reversible bus lanes (Section 8.9). Constraints State traffic laws and the state supplement to the MUTCD may not allow the use of transit signal faces, may not allow the use of particular faces (e.g., triangles), or may require additional elements not specified in the national MUTCD, such as “Bus Signal” signs. Benefits Transit signal faces can reduce the potential for crashes that can occur when motorists or other road users misinterpret a standard signal display meant only for buses as being a green indication for them. Cost Considerations • Planning and coordination costs. Low to moderate. The first implementation in a jurisdic- tion will likely require a higher level of coordination with the roadway agency responsible for traffic signals, particularly if a signal controller upgrade is needed to support the bus signal phase and signal faces. Outreach to the police (about the meaning of the signals) and the public (that buses may move in advance of other traffic) is also suggested the first time transit signal faces are used in a jurisdiction. Subsequent implementations will likely require coordi- nation only with the roadway agency. • Capital costs. Moderate, but in a fairly large range covering signal heads and wiring only to a signal controller upgrade. • Maintenance costs. Low additional cost to maintain the extra signal equipment. • Bus operations costs. No direct impact but helps support other strategies designed to provide benefits. To the extent that they reduce road user confusion, the signals may provide a safety benefit relative to using shielded standard signal heads. • Other user costs. No direct impact, but used in conjunction with other strategies, may pro- duce other user costs or benefits. Implementation Examples Transit signal faces have been implemented in conjunction with a number of BRT projects employing median bus lanes, such as in Cleveland, Ohio; Eugene, Oregon; and Orlando, Florida; and for a number of bus-only signal phase and queue jump applications, such as in Calgary, Alberta; Las Vegas, Nevada; and San Francisco, California. Implementation Guidance Assuming that state and local laws, rules, and policies permit their use, transit signal faces are suggested for consideration in conjunction with any of the strategies listed in the Companion Strategies section.

Traffic Control Strategy Toolbox 89 Additional Resources • Manual on Uniform Traffic Control Devices (FHWA 2009)—Section 4D.27 addresses the use of light rail transit (LRT) signals for bus transit; Section 8C.11 describes LRT signals, their operation, and their installation requirements. 6.9 Bus-Only Signal Phases Description A traffic signal phase included in the traffic signal cycle to serve bus movements that cannot be served, or are not desired to be served, concurrently with other traffic. Purpose Bus-only signal phases help support other strategies, such as queue jumps (Section 6.10) and bus lanes (Section 8.1), by allow- ing buses to make nonstandard movements at an intersection. Without such signals, some transit-supportive roadway strategies might not be feasible (e.g., queue jumps), while others would be less effective (e.g., ending a bus lane a block or two early to give buses time to move across traffic lanes to a standard left-turn lane). Applications One typical application allows a bus turning movement from a nonstandard location, such as making a left turn from a right-side bus lane, making a right turn from a left-side bus lane, or movements to and from a median bus lane. Another typical application is to give buses a head start on parallel traffic, such as with a queue jump (Section 6.10). Figure 4 illustrates the BU S ONLY BUS ① ② ③ ONLY BUS BUS ONLY BUS Figure 4. Example of bus-only signal phase operation.

90 A Guidebook on Transit-Supportive Roadway Strategies operation of a bus-only signal phase used to allow buses to make a left turn from a right-side bus lane. Companion Strategies The need for transit signal faces (Section 6.8) should be considered when implementing bus-only signal phases. These phases can be used in conjunction with movement restriction exemptions (Section 6.1), TSP (for example, to call the special bus phase sooner) (Section 6.7), queue jumps (Section 6.10), pre-signals (Section 6.11), bus-specific signals (Section 6.12), queue bypasses (Section 8.6), median bus lanes (Section 8.7), contraflow bus lanes (Section 8.8), and single-lane reversible bus lanes (Section 8.9). Constraints State or local laws may not allow bus turns from nonstandard locations or allow the use of transit signal faces (Section 6.8). The signal controller needs to have an unused phase available to serve the bus-only phase. Bus turning radii will need to be checked, particularly for a right turn from a left-side lane, and it may be necessary to set the stop bar for the general traffic lanes back from the intersection to create sufficient space for a bus to make its turn. As the time required to serve the bus phase will be taken from other traffic movements, traffic operations will need to be evaluated to make sure that the signal will still operate acceptably with the addition of the extra phase. See the Implementation Guidance section for potential alternatives when one of these con- straints makes a bus-only signal phase infeasible. Benefits Bus-only signal phases are typically a support strategy and make another strategy feasible or allow another strategy to be used to maximum effectiveness. When used to serve turning move- ments from unconventional locations, they may reduce travel time or travel time variability, depending on the level of traffic congestion and challenges faced by buses to weave through traffic to position themselves to make a turn from a conventional location. The potential benefit is highly site-specific and would need to be determined by a traffic analysis. Cost Considerations The cost considerations associated with transit signal faces (Section 6.8) may also be applicable. • Planning and coordination costs. Low to moderate. The first implementation in a jurisdic- tion will likely require a higher level of coordination with the roadway agency responsible for traffic signals, particularly if a signal controller upgrade is needed to support the bus signal phase and signal heads. Public outreach may also be needed to minimize the risk of increased crashes resulting from other roadway users reacting incorrectly to the new signal operation, particularly during the first year of operation. • Capital costs. None to moderate, depending on whether bus detection infrastructure exists or would be installed for another strategy and whether a signal controller upgrade is needed. Accessible pedestrian signals may also be required. • Maintenance costs. No direct impact. • Bus operations costs. When used to serve unconventional turning movements, the strategy may reduce delays associated with buses weaving across traffic lanes to a location where a conventional turning movement can be made. • Other user costs. The time required to serve a bus-only signal phase will likely increase delay for at least some other vehicles using the intersection.

Traffic Control Strategy Toolbox 91 Implementation Examples • San Francisco, California. A bus-only signal phase was used to allow buses to make a right turn from a left-side bus lane on a one-way street. The combination of the three-block bus lane and bus-only signal phase saved buses 1.5 min of travel time compared to waiting in traf- fic to make a conventional right turn and also reduced travel time variability, as measured by the standard deviation of travel times, by more than half (Mirabdal and Thesen 2002). After the closure of the old Transbay Terminal, the bus line was rerouted and the bus-only signal phase taken out of service since buses no longer turned at that location. At the time of writing, the strategy was being considered at other locations on a case-by-case basis. • Eugene, Oregon. A bus-only signal phase was used to allow buses to make a right turn from a left-side reversible bus lane on a one-way street. After the route realignment, the bus-only phase supports a queue jump operation, providing buses with a head start on other traffic so they can merge across the street to enter a right-side bus lane that begins one block downstream. • Richmond, British Columbia, Canada. A bus-only signal phase was used to provide priority for many buses exiting the former Airport Transit Centre. The phase served both left-turning buses (toward Vancouver International Airport) and through bus movements (toward down- town Vancouver). After the Canada Line rail extension opened to the airport, the transit center was closed and the bus-only signal phase taken out of service. Implementation Guidance Special bus phases are a potential option when bus turning movements need to be made from unconventional locations. Designs may need to take into consideration conditions where other intersection users need to be warned about the unconventional movement (e.g., “Bus Signal” signs, accessible pedestrian signals, a special sign depicting the bus maneuver, dotted pavement markings), and the conditions listed in the Constraints section will need to be checked and potentially addressed prior to proceeding. Use with Median Bus Lanes If buses must leave a median bus lane at a signalized intersection, a special bus phase will be needed to serve them. When not all buses will turn and bus volumes are relatively high, it may be desirable to provide a separate bus turn lane. A turn lane allows both a through bus phase and a right- or left-turn bus phase to be provided. Late-arriving turning buses wait in the turn lane until the next cycle, while through buses can continue to be served at the same time as parallel general traffic through movements. In this way, turning buses do not delay through buses, and the length of the special bus phase can be minimized, reducing its impact on other traffic. How- ever, constrained median widths may make it impractical to provide turn lanes, particularly at intersections where busway stations will also be located. An alternative to a special bus phase, particularly when turning bus volumes are low, is to start the median bus lane’s bus phase at an upstream intersection before parallel traffic receives a green signal, thus giving buses a gap in traffic that allows them to move out of the bus lane and into the correct general traffic lane for their turn at the downstream intersection. Another alternative is to provide a midblock “slip-ramp” exit from the median bus lane that allows buses to merge into general traffic (AASHTO 2014). Use with Right-Side Bus Lanes In most cases, a pre-signal or an upstream queue jump can provide the necessary gap in traf- fic that would allow buses to merge across lanes to use a standard left-turn lane to make their turn. However, there may be times (e.g., a high-volume passenger generator best served by a

92 A Guidebook on Transit-Supportive Roadway Strategies near-side bus stop, an over-capacity left-turn movement) where it would be desirable to allow buses to make a left turn from a right-side bus lane. In these cases, a special bus phase could be considered. Use with Left-Side Bus Lanes Similar to right-side bus lanes, a pre-signal or upstream queue jump can often address the need to move buses from one side of the street to the other to prepare to make a turn. However, there may be times (e.g., a high-volume passenger generator on the left side of a one-way street, traffic congestion in the block preceding the right turn) when it would be desirable to allow buses to make a right turn from a left-side bus lane. In these cases, a special bus phase could be considered. Additional Resources • Guide for Geometric Design of Transit Facilities on Highways and Streets (AASHTO 2014)— Section 5.5.5.2 discusses the potential need for a special bus phase (possibly in conjunction with a separate bus right-turn lane) when a high volume of buses make right turns from a left-side bus lane. Section 5.6.2.2 discusses the potential need for a special bus phase in con- junction with median bus lane operations. 6.10 Queue Jumps Description Buses (or in some applications, buses and right-turning vehi- cles) are given an opportunity to move ahead of queued through- vehicles at a signalized intersection and, in many cases, to proceed into the intersection in advance of the through traffic. Purpose Queue jumps are a combination infrastructure and traffic con- trol strategy. First, buses are provided the opportunity to bypass any queue of vehicles that might exist at a traffic signal. Second, a special signal phase allows the buses to depart the intersection ahead of other through-vehicles and, thus, jump the queue. If a near-side bus stop exists at the intersection, and if buses are ready to proceed at the start of the green, buses arriving on red do not have to wait for the queue in front of them to clear to access the bus stop, and they do not have to wait for a gap in traffic when departing the stop. If a far-side bus stop or no bus stop exists at the intersection, buses arriving on red bypass the queue and get through the intersection sooner than they would have otherwise. Applications Figure 5 illustrates three typical ways that a queue jump can be developed. The Implementa- tion Guidance section provides more specific design and implementation guidance. • Shared right-turn lane. A right-turn lane is provided that is longer than the queue length in the through lanes, allowing buses and right-turning vehicles free access to the lane (Figure 5a). Right-turning vehicles stopped in front of the bus can block the bus’s access to a near-side bus

Traffic Control Strategy Toolbox 93 stop; therefore, to avoid the potential for buses stopping twice (and needing to wait an extra cycle length), bus stops should be either located at the far side or prior to the start of the right- turn lane. If right-turning volumes are high, providing a separate right-turn lane and placing the bus stop on a right-turn channelization island (Section 7.6) may be a better option. • Short bus lane. A short bus lane is provided that is longer than the queue length in the through lanes, allowing buses free access to the lane (Figure 5b). As right turns from the lane to the left of a bus lane are not recommended due to possible conflicts between right-turning vehicles and buses using the queue jump (AASHTO 2014), this option is best suited for inter- sections without right-turn movements. These include T-intersections where the minor street approaches from the left, intersections with one-way streets approaching from the right, and intersections with low right-turn volumes that can be shifted to an upstream or downstream intersection. The bus stop can be located at the near side at the stop bar or one bus length prior to the stop bar (depending on how buses will be detected to activate the queue jump phase), or can be located at the far side. • Shoulder bus lane. This type of queue jump lane is operationally similar to a short bus lane, except that buses are allowed to use the shoulder when approaching the intersection (Fig- ure 5c). The shoulder needs to have been constructed to support regular bus use. This type of lane is potentially applicable on suburban arterial streets and highways that use a shoulder instead of a curb and gutter on the edge of the roadway. Figure 6 illustrates one potential way a queue jump in a right-turn lane can operate, assuming that the pedestrian push button has been pressed on the parallel crosswalk. In step 1, the bus is detected at the bus stop and the queue jump phase is called. In step 2, the queue jump phase is activated. A green arrow is provided for the right-turn lane, allowing any right-turning vehicles ONLY BUS ONLY BUS ONLY BUS BU S BU S BU S (a) RIGHT-TURN LANE (b) SHORT BUS LANE (c) BUS SHOULDER USE RIGHT LANE TURN RIGHT EXCEPT BUS SHOULDER Figure 5. Illustrative examples of queue jump approaches.

94 A Guidebook on Transit-Supportive Roadway Strategies in front of the bus to clear out of the way so the bus can proceed into the intersection. In step 3, the bus has cleared the intersection, and parallel traffic and the parallel crosswalk phases are served. If no pedestrians need to be served on the crosswalk, the right-turn phase can continue as long as the through green phase is served. If a bicycle lane is provided on the roadway, it would be located to the left of the right-turn lane and served at the same time as the parallel through traffic. Other options for accommodating right-turning traffic are discussed in the Implementa- tion Guidance section. When short bus lanes and shoulder lane queue jumps are employed, parallel pedestrians and bicycles can be allowed to start moving at the same time as buses since there are no other con- flicting movements when right turns are prohibited. If no bus stops are provided at the intersec- tion, bus drivers can decide whether to use a queue jump lane, depending on whether they arrive at the intersection on a green or red signal. Figure 7 illustrates a queue jump from a short bus lane into a far-side bus pullout. In this case, it is not necessary to provide an advance green to buses because they stay in their own lane and thus do not conflict with the parallel through traffic, and right turns are prohibited. The short bus lane prior to the traffic signal provides the time savings for the bus by allowing it to bypass the queue waiting at the signal. The parallel pedestrian movement is served at the same time as buses and general traffic. See Appendix C for concepts for accommodating bicycles in the vicinity of the bus stop. Companion Strategies Queue jumps can be used in combination with bus stop relocations (Section 5.1), move- ment restriction exemptions (Section 6.1), right-turn restrictions (Section 6.2), phase reservice BU S ① ② ③ BU S BU S RIGHT LANE TURN RIGHT EXCEPT BUS RIGHT LANE TURN RIGHT EXCEPT BUS RIGHT LANE TURN RIGHT EXCEPT BUS Figure 6. Illustrative example of queue jump operation from a right-turn lane with a near-side stop.

Traffic Control Strategy Toolbox 95 (Section 6.5), and transit signal priority (Section 6.7). A bus-only signal phase (Section 6.9) is often employed as part of the queue jump and may be indicated using transit signal faces (Sec- tion 6.8). Red pavement coloring (Section 7.4) is a potential support strategy for short bus lanes. Pre-signals (Section 6.11) and queue bypasses (Section 8.6) are related strategies. Constraints A constraint common to any queue jump configuration is the need for a sufficiently long lane to allow buses unimpeded access to the lane under most circumstances. AASHTO (2011) recommends that 1.5 to 2 times the average peak-period queue length be used in designing turn lane storage lengths, which approximate 85th- and 95th-percentile queues, respectively. However, simulation studies by Cesme et al. (2015) found that bus delay was relatively insensi- tive to the queue jump lane length when buses arrived randomly, with designs accommodating a 35th-percentile queue resulting in only a 1.3-s reduction in the delay benefit at a v/c ratio of 0.9 relative to a 95th-percentile queue design; smaller delay reductions were observed at lower v/c ratios. Nevertheless, the lack of available right-of-way or the cost of extending or constructing a queue jump lane may constrain a transportation agency’s ability to implement a queue jump. A traffic analysis can determine queue length percentiles (both current and future design year con- ditions should be considered) as an input for determining the required lane length, and can also determine the probability that an arriving bus would not be able to access the queue jump lane. BUS STOP ONLY BUS BU S BUS STOP ONLY BUS BU S ① ② Figure 7. Illustrative queue jump operation from a short bus lane with a far-side stop.

96 A Guidebook on Transit-Supportive Roadway Strategies Another potential constraint is the ability to provide sufficient time during the traffic signal cycle to the queue jump, particularly in the case of shared right-turn lanes. The potential benefit of queue jumps grows as traffic volumes increase, but the ability to reallocate time within the signal cycle diminishes. The Highway Capacity Manual (HCM) or analytical techniques can be used to determine how much time on average would be required to serve the queue jump lane and, subsequently, how the intersection would operate with the revised signal timing. If right- turning volumes are too high to be served with a protected phase, they would need to be served as a permitted movement from their own lane to the right of a bus lane, which again raises the question of available right-of-way and the potential need to reconstruct curbs and sidewalks to add the extra width. In these cases, it is not necessary to widen the full length of the queue jump lane, just the portion closest to the intersection (see Section 7.6 on boarding islands). Traffic laws may need to be revised to allow buses to continue straight from a right-turn lane. Benefits When through-traffic queues are long (due to high traffic volumes or long traffic signal cycle lengths) and right-turning traffic is low or non-existent, the queue bypass aspect of queue jumps can potentially save buses significant amounts of time. As a rule of thumb, the time savings is equivalent to 2.25 s times the number of vehicles queued in the right-hand through lane when the bus arrives at the intersection minus 2.5 s times the number of vehicles queued in the queue jump lane when the bus arrives. However, if the bus arrives on green and proceeds to a far-side stop, no time savings result. Therefore, the actual delay benefit depends on when in the traffic signal cycle buses and other traffic arrive at the intersection. The advanced-green aspect of queue jumps saves buses no additional time since the time is typically taken from the parallel through traffic. Instead, the through traffic experiences extra delay because the traffic signal serves them later than they would have been served otherwise. A simulation analysis conducted under TCRP Project A-39 found that, for a queue jump designed for 95th-percentile queues, the tested queue jumps produced no significant change in bus travel times at v/c ratios of 0.5 for both near- and far-side stops and a v/c ratio of 0.8 for far-side stops. For near-side stops, buses experienced average delay reductions of 1.5 and 7 s for v/c ratios of 0.8 and 1.0, respectively, while parallel through traffic experienced average delay increases of 1 and 2 s per vehicle, respectively. For far-side stops, bus delay was reduced an aver- age of 2 s at a v/c ratio of 1.0, while parallel through traffic experienced average delay increases of 3 s per vehicle. A simulation study by Cesme et al. (2015) found reductions in average bus delays of between 2 and 9 s for v/c ratios ranging from approximately 0.5 to 0.9 for a queue jump designed for 95th-percentile queues and no right turns allowed. If the queue jump is developed by restricting right turns so that they occur only during a protected right-turn phase, pedestrians benefit from the reduced number of interactions with right-turning traffic (i.e., none from the queue jump lane, but potentially right turns on red from the side street approach). Cost Considerations The cost considerations associated with transit signal faces (Section 6.8) and bus-only signal phases (Section 6.9) may also be applicable. • Planning and coordination costs. Low to moderate. An analysis of the effect of the queue jump on intersection operations and users is recommended. The first implementation in a jurisdiction will likely require a higher level of coordination with the roadway agency. Outreach to the police (about any special traffic regulations such as bus use of the right-turn lane) and

Traffic Control Strategy Toolbox 97 the public (about queue jump operation generally) is suggested the first time a queue jump is implemented in a jurisdiction. Subsequent implementations will likely require coordination only with the roadway agency. • Capital costs. Moderate to high. Sign costs are low, but construction and potential right- of-way acquisition costs to construct or lengthen the queue jump lane are potentially high. There are also signal-related costs previously discussed as part of related strategies. Accessible pedestrian signals may be required. • Maintenance costs. Potential low to moderate added cost to maintain extra signs, extra pave- ment area, and detection equipment. • Bus operations costs. Relatively small time savings benefits, except when the intersection operates close to capacity. • Other user costs. Small increase in average delay to parallel through traffic if time is taken from its phase to provide an advanced green for buses. More substantial potential increase in delay to right-turning traffic if right turns are restricted to a protected right-turn phase only or if right turns are relocated to another intersection. Implementation Examples A survey of 52 North American transit agencies found that 27 reported having installed at least one queue jump or queue bypass lane. Of these, Ottawa, Ontario (8), Halifax, Nova Scotia (8), and King County Metro (Seattle, Washington [6]) were notable for the number of installations (Danaher 2010). A shoulder lane queue jump exists on southbound U.S. 202 near Wilmington, Delaware (Martin et al. 2012). Implementation Guidance General Guidance The following characteristics make a signalized intersection more suitable for a queue jump: • Near-side stop is desired for non-operational purposes (e.g., to facilitate passenger transfers, to serve near-side land uses), or no bus stop is provided at the intersection; • Ending point of a bus lane, existing shoulder suitable for buses, or existing right-turn lane; • No or low (e.g., a few vehicles per cycle) right-turning traffic; • Low pedestrian usage of the parallel crosswalk (if one exists); and • High peak-period intersection v/c ratio (e.g., 0.7 or greater), but sufficiently below capacity that green time can be taken to serve the queue jump phase. Bus Stop Location Not having a bus stop at the intersection provides the most flexibility in selecting a specific queue jump application and is compatible with any of the queue jump applications depicted in Figure 5. Near-side stops work best in conjunction with short bus lanes or shoulder lanes. They also can work in a shared right-turn lane application when the right turns are channelized and the bus stop can be placed on the channelizing island (see Section 7.6). When right turns have an exclu- sive lane but are not channelized, placing the bus stop at the stop bar is not recommended unless it is highly likely that buses can access the stop without stopping twice (for example, because of a queue jump at the upstream signal); otherwise, buses risk having to wait through another signal cycle before they can depart (and also block right-turning traffic behind them). Placing the bus stop elsewhere in the right-turn lane is also not recommended since it encourages right-turning traffic to cut in front of buses, leading to potential conflicts with both buses and bicycles. Placing a stop prior to the start of the right-turn lane is an option, but this may be inconveniently far from the intersection for passengers.

98 A Guidebook on Transit-Supportive Roadway Strategies Far-side stops also work best in conjunction with short bus lanes or shoulder lanes and elimi- nate the requirement for a special queue jump phase since buses do not need to merge back into traffic immediately. Far-side stops can be used in conjunction with shared right-turn lanes but may not provide any operational benefit or may even provide a disbenefit. A simulation study found that a disbenefit resulted with the following combinations of right-turning vehicles and pedestrians in the parallel crosswalk: • 100 or more right-turning vehicles per hour, with 300 or more pedestrians per hour; • 200 or more right-turning vehicles per hour, with 150 or more pedestrians per hour; or • 300 or more right-turning vehicles per hour, with nearly any pedestrian volume (Cesme et al. 2015). Queue Jump Lane Length The desirable queue jump lane length is one that is at least as long as the 85th- or 95th-percentile peak-period queue in the through lane. However, shorter lengths may function adequately, and accepting a shorter lane length may be a more cost-effective solution than providing the desir- able length. A simulation study found that the bus delay benefit was reduced by no more than 1.3 s when a 35th-percentile length was used, compared to a 95th-percentile length when buses arrived randomly (Cesme et al. 2015). Conditions at the upstream stop will help determine when a bus is likely to arrive at the inter- section relative to other through traffic. If buses are likely to arrive ahead of other traffic (e.g., due to an upstream queue jump) or randomly, shorter queue jump lane lengths will likely oper- ate satisfactorily. If buses are likely to arrive behind other traffic (e.g., due to an upstream bus pullout), queue jump lane lengths closer to the desirable length will likely be necessary. Options for Serving Right-Turning Traffic If right-turning traffic cannot be diverted to another intersection, then short bus lanes and shoulder lanes are not options since it is undesirable to have traffic turn right from the lane to the left of the queue jump lane (AASHTO 2014). Options for serving right-turning traffic, in order of desirability, are: • Separate, channelized right-turn lane. The bus stop (if any) is placed on a channelizing island and is served by a short bus lane accessed from the right-turn lane (see Section 7.6). The time required for the queue jump phase is minimized since only buses need to be served. Pedestri- ans have to cross to the channelizing island to reach the main intersection crosswalks, which is less convenient and potentially more time-consuming for them compared to not having a channelizing island. Right turns can begin to be served when the queue jump phase is served. • Protected (i.e., green arrow) right-turn phase with the queue jump phase. Right-turning vehicles and buses are served first (if buses need to use the intersection to merge back into the through lane) or start simultaneously with through traffic (if buses continue to a far-side stop). If a pedestrian call needs to be served on the parallel crosswalk, right turns are stopped at the end of the queue jump phase and the crosswalk is served while parallel through traf- fic continues to be served. If no pedestrian call is waiting, right turns continue to be served through the end of the parallel through-traffic phase. Pedestrians benefit from the absence of conflicting right-turning traffic (although right turns on red may still occur at the opposite end of the crosswalk). This option requires the longest queue jump phase length in order to accommodate the potential of the bus being at the back of the right-turn queue; therefore, it may not be feasible with more than a few right-turning vehicles per cycle. • Permitted (i.e., circular green) right-turn phase. This is an option with far-side stops, but if even moderate right-turn volumes exist, it can result in greater bus delay than simply allowing

Traffic Control Strategy Toolbox 99 the bus to use the through lane to access the far-side bus stop. It can also be used with near- side stops if one is willing to accept that the bus will often stop twice when accessing the stop. In the latter case, a short queue jump phase can be provided after the parallel through traffic phase ends for use if the bus is ready to depart. • Protected/permitted right-turn phase with a queue jump phase. This option first provides a protected right-turn/queue jump phase to clear out any right-turning vehicles in front of the bus. At the end of the queue jump phase, the parallel through traffic and parallel crosswalk phases begin. Right turns can still be made during this time, but vehicles must yield to pedes- trians in the crosswalk. This option reduces vehicular right-turn delay relative to a protected- only option but is least desirable from a pedestrian standpoint since the pedestrian phase starts after right-turning traffic. This situation can lead to late-arriving, right-turning motor- ists not realizing there are now pedestrians in the crosswalk; the potential conflict is similar to that experienced with protected/permitted left turns. Additional signs, such as “Turning Vehicles Yield to Pedestrians” (MUTCD sign R10-15) may be necessary if this option is used. Additional Resources • Guide for Geometric Design of Transit Facilities on Highways and Streets (AASHTO 2014)— Section 5.3.2.3 provides AASHTO’s recommendations for queue jumps (note that AASHTO’s usage of the terms queue jump and queue bypass are the opposite of this guidebook and other TCRP publications). 6.11 Pre-Signals Description A traffic signal for one direction of a street, coordinated with a traffic signal at a downstream intersection, that is used to control the times when particular vehicles may approach the intersection. Purpose In a transit context, pre-signals are used at the end of a bus lane to give buses priority access to the intersection when constraints make it infeasible to continue the bus lane all the way to the inter- section. They can also be used to manage queues on the inter- section approach—for example, when a side street or driveway is regularly blocked by queues extending back from the traffic signal. Applications A significant challenge when implementing bus lanes is the loss of vehicular capacity that results at signalized intersections if a general-purpose lane is converted to bus-only use. Because traffic signals meter the amount of traffic that can pass through an intersection, it is often possible for a roadway to have sufficient capacity between traffic signals to convert a lane to bus-only (or bus-plus-right-turn) use but not have sufficient capacity to provide a bus lane at the traffic signal. Pre-signals address this issue by moving buses to the head of the line at traffic signals (thus mini- mizing bus delay) while maximizing the amount of roadway space that can be used for general traffic movement at the intersection (thus using the intersection as efficiently as possible). When the pre-signal is properly located and timed relative to the main intersection signal, the transit benefit can be achieved with no loss of intersection capacity and negligible delay to general traffic.

100 A Guidebook on Transit-Supportive Roadway Strategies There are three main applications of pre-signals: • Virtual bus lane. In this application (Figure 8), queues on a congested approach to a traffic signal extend well back from the signal, and it may take several traffic signal cycles before a vehicle can get through the intersection once it joins the back of the queue. A pre-signal is used both to manage queues at the intersection—metering only as much traffic to the intersection as can be served in the approach’s green interval—and to provide a virtual bus lane between the pre-signal and intersection that is clear of other traffic at the time the bus needs to use it. In this application, the pre-signal serves as a signalized queue bypass. • Merge assist. In this application, the right-of-way used by the bus lane is needed for other purposes downstream—for example, for a right-turn lane at the downstream intersection or for curbside parking. A traffic signal assists buses in merging into the adjacent general traffic lane in front of other traffic. This application can also be used to assist buses in reentering traf- fic from an offline bus stop (e.g., a bus pullout), such as in the photograph at the beginning of this section. BU S BU S BU S BU S ONLY BUS BU S BEFORE AFTER ① AFTER ② AFTER ③ AFTER ④ ONLY BUS ONLY BUS ONLY BUS Main Intersecon Pre-signal Virtual Bus Lane Reservoir Area Figure 8. Examples of virtual bus lane and weave-assist applications for pre-signals.

Traffic Control Strategy Toolbox 101 • Weave assist. In this application, buses need to exit a bus lane to turn left at a downstream intersection. The pre-signal provides a gap in traffic that allows buses to weave from a right- side bus lane to the left-turn lane. Figure 8 illustrates one possible configuration for a pre-signal. In the before case, the approach lanes to the intersection operate over capacity, and buses experience the same delay as other vehicles. In the after case, the right lane is converted to a bus-only lane, while the left lane(s) are used by general traffic. One more general-purpose lane is provided after the pre-signal than before the pre-signal, using the space that had been occupied by the physical bus lane, thus maximizing vehicle throughput at the intersection. The pre-signal is preferably installed on the approach at a location that ensures that all of the vehicles that pass the pre-signal can be served on the next green at the intersection. The back of the queue extends farther back than previously, but the same number of vehicles pass through the signal each cycle and a bus lane allows buses to bypass the queue. In step 1 of the after case, the pre-signal for general traffic has turned red, and the final vehicles that passed the pre-signal are entering the intersection as its signal turns yellow. At the end of this step, the reservoir area between the pre-signal and the intersection is clear of vehicles. Next, in step 2, the pre-signal for bus traffic changes to “go,” allowing buses to bypass the queue and proceed to the intersection stop bar (or left-turn lane) unimpeded. In step 3, the bus pre-signal changes to “stop,” the general traffic pre-signal changes to green, and general traffic is allowed to fill the reservoir area between the pre-signal and the intersection. Finally, in step 4, the inter- section traffic signal turns green, allowing the queued traffic to proceed, while at the same time the pre-signal stops additional vehicles from entering the reservoir area. Companion Strategies A bus lane (Chapter 8) is a prerequisite for employing a pre-signal. It is suggested that the bus lane be controlled by transit signal faces (Section 6.8) providing a bus-only signal phase (Sec- tion 6.9). Transit signal priority (Section 6.7) can potentially be applied both at the pre-signal and the downstream signal. Queue jumps (Section 6.10), where priority is provided at the signalized intersection, and queue bypasses (Section 8.6), which do not use traffic signals, are related strategies. Constraints Pre-signals are a support strategy for bus lanes, and therefore the constraints generally appli- cable to bus lanes (Section 8.1) also apply to pre-signals. Properly implemented, they do not affect the signal timing or approach throughput at the downstream intersection; however, the presence of side streets or driveways may require locating the pre-signal at a less-optimal loca- tion that may affect intersection throughput. Their main impact lies in relocating the queue from the intersection to an area farther upstream from the intersection. When the pre-signal helps facilitate the conversion of a general-purpose lane to a bus lane, the queue in the remaining general-purpose lanes will become longer (up to twice as long). Benefits When an intersection operates over capacity, a pre-signal’s ability to allow buses to bypass the queue can result in substantial time savings; the combination of a bus lane and pre-signal in York, United Kingdom, was reported to save buses 4 to 12 min per trip during peak periods

102 A Guidebook on Transit-Supportive Roadway Strategies (Hodge et al. 2009). The magnitude of the potential time savings will depend on the severity of the congestion. When an intersection operates under capacity, the main benefit is facilitating bus movements into or across the general traffic lane(s). Pre-signals will also tend to reduce bus travel time variability. Cost Considerations • Planning and coordination costs. Moderate. A traffic analysis will be needed to identify the optimal location for the pre-signal. A signal timing plan will need to be developed for the pre-signal and coordinated with the downstream signal. Planning and coordination costs associated with transit signal faces (Section 6.8) and special bus phases (Section 6.9) will also be applicable. Outreach to the stakeholders normally interested in bus lane projects (Sec- tion 8.1) is also suggested. • Capital costs. Moderate to high, involving obtaining and installing the traffic signal equipment for the pre-signal. Depending on how close the pre-signal is to the downstream signal, the signal faces for the approach at the downstream signal may need to be replaced with visibility- limited signal faces that are only visible once motorists pass the pre-signal. • Maintenance costs. Moderate, involving added costs for operating and maintaining the pre-signal. • Bus operations costs. Potential savings from reductions in travel time and travel time variability. • Other user costs. Properly implemented, there will be no change in general traffic delay. However, the presence of driveways, unsignalized intersections, or other queue management concerns may require locating the pre-signal in a less-optimal location that does affect general traffic delay. Implementation Examples At the time of writing, no U.S. or Canadian installations were known to exist, but many installations have been documented elsewhere in the world. In particular, London has had con- siderable experience with pre-signals, with at least 21 signals in operation by mid-1998 (Beswick 1999). Other installations documented in the literature are Manchester and York, United King- dom; Frederiksberg, Lyngby, and Svendborg, Denmark; Zurich, Switzerland; Melbourne and Brisbane, Australia; and Wellington, New Zealand. Implementation Guidance General Guidance The presence of a bus lane or an extended bus pullout is a prerequisite for considering a pre- signal. The pre-signal should operate full time unless there are overriding reasons not to do so (Beswick 1999). To obtain maximum benefit for buses, locate bus stops either immediately prior to the pre-signal or on the far side of the intersection. Virtual Bus Lane Applications • Pre-signals providing a virtual bus lane are well-suited for the critical intersection(s) along a bus route where as many roadway lanes as possible are needed to serve through traffic. These are the intersections with the highest demand-to-capacity ratios. • It is frequently not necessary to continue the bus lane past the critical intersection since the intersection limits the amount of traffic that can enter the next block, and bus operation in mixed traffic downstream of the critical intersection may operate without problems.

Traffic Control Strategy Toolbox 103 • If the bus lane is restarted downstream of the intersection, the right lane will operate as an auxiliary through lane and will likely not be fully utilized. Procedures in NCHRP Report 707: Guidelines on the Use of Auxiliary Through Lanes at Signalized Intersections (Nevers et al. 2011) can be used to estimate the amount of traffic that will use the right lane. Merge-Assist Applications Pre-signals providing a merge-assist function can be considered for locations where: • Policy needs (e.g., providing on-street parking for a commercial node along a street), geo- metric constraints (e.g., narrowed right-of-way), or traffic operation needs (e.g., providing a right-turn lane at the next intersection) dictate ending a bus lane; or • Buses have difficulty reentering traffic from a midblock stop. Weave-Assist Applications Pre-signals providing a weave-assist function can be considered for locations where buses need to exit a bus lane to turn left at a downstream intersection. MUTCD Compatibility and Requirements As part of the research behind the development of this guidebook, the research team corre- sponded with FHWA about the need for an experimentation request when applying pre-signals. The response from the FHWA staff person responsible for the MUTCD’s traffic signals material was that no experimentation request would be necessary and that bus movements would prefer- ably be controlled by transit (LRT) signal heads (see TCRP Web-Only Document 66). The MUTCD already provides provisions for pre-signals in a railroad crossing context (FHWA 2009, Section 8C.09), for using priority control of traffic signals to assign priority right- of-way “to specified classes of vehicles at certain non-intersection locations,” including transit operations (FHWA 2009, Section 4D.27), and for using transit signal heads to control public transit buses using “queue jumper lanes” (FHWA 2009, Section 4D.27). In addition, the use of pre-signals coordinated with a main intersection is similar to the use of supplemental signals with alternative intersection forms such as displaced left-turn intersections. Pre-Signal Placement and Timing For virtual bus lane applications, particularly when the main intersection operates at or near capacity, the pre-signal ideally would be located far enough away from the main intersection that all of the vehicles that can be served during the green interval during peak periods can be stored between the pre-signal and the intersection. Furthermore, during peak periods, the pre-signal ide- ally would be timed to turn green such that the entire area between the pre-signal and the inter- section can fill with vehicles by the time the main signal turns green. During off-peak periods, when traffic volumes are lower and intersection efficiency is less important, the pre-signal could be timed to progress vehicles through to the main intersection without forcing traffic to stop twice. Locating the pre-signal more than the ideal minimum distance from the main intersection is generally not an issue since there will simply be some empty space between the pre-signal and the back of the queue from the main signal. On the other hand, locating the pre-signal less than the ideal minimum distance from the intersection will result in lower intersection efficiency because the pre-signal will deliver fewer vehicles to the main intersection than can be served at the intersection toward the end of the approach’s green interval. With merge-assist applications, the number of general-purpose lanes at the pre-signal and the main intersection is the same, and therefore the pre-signal location is more flexible than in a

104 A Guidebook on Transit-Supportive Roadway Strategies virtual bus lane application. The pre-signal timing should allow vehicles to progress through to the main intersection without forcing traffic to stop twice. In a weave-assist application, the number of general-purpose lanes downstream of the pre- signal may be greater than those upstream (e.g., if all buses will turn left and the bus lane is no longer needed), in which case the placement considerations are similar to a virtual bus lane application, particularly if the approach operates at or near capacity. Otherwise, if the bus lane continues past the pre-signal (e.g., some buses continue straight instead of turning left), the placement considerations are similar to those of a merge-assist application, with the additional consideration that buses will need sufficient roadway space to weave over to the left-turn lane. Similar to other types of traffic signals, pre-signals should not be installed at an unsignalized intersection or within 100 ft of one. Consider the need for visibility-limited signal faces at the downstream intersection (FHWA 2009). Queue Management Considerations The location of the pre-signal and the expected back of the queue relative to upstream drive- ways and intersections requires special consideration. Access management measures (e.g., clos- ing or consolidating access points) may be needed if queues from the pre-signal regularly block access points; in a worst case, access point blockage could pose a fatal flaw to installing pre- signals. The Access Management Manual (Williams et al. 2014) provides guidance on potential access management strategies, and bus lane and pre-signal installation could be considered in conjunction with an overall access management plan for a corridor. At the same time, pre- signals can provide an access management benefit when access points are located close to a traffic signal, cannot be readily moved or closed, and are frequently blocked by stopped traffic. Pedestrian Considerations If pedestrian volumes warrant a signalized pedestrian crossing, a pre-signal could be installed in conjunction with a midblock pedestrian crossing; this type of treatment has been docu- mented internationally (e.g., Beswick 1999, Greater Manchester Public Transport Authority 2007). Although pedestrian jaywalking at pre-signals has not been identified as an issue in the international literature, evaluating the potential for jaywalking at a potential pre-signal site may result in possible countermeasures being identified, such as “No Pedestrian Crossing” signs, landscaping, or railings. Bicycle Considerations The same considerations described in Appendix C that generally apply to bicycle facilities shared with or adjacent to bus lanes also apply to pre-signals. In addition, if the bicycle facility type changes downstream of the pre-signal (e.g., from shared bus/bike to general-purpose lane), pavement markings may be required downstream of the pre-signal to direct bicyclists and to warn other road users about the presence of bicyclists. Unless the pre-signal is installed in com- bination with a signalized pedestrian crosswalk, it should not be necessary for bicycles traveling in their own lane to have to stop at the pre-signal. Options for addressing bicycle movements include (subject to local laws and policies): • Where an exclusive bicycle facility is provided, a bicycle signal head could be provided (allowed by an FHWA Interim Approval, but still requires a formal request to FHWA until the MUTCD is updated); or • Directing bicycles from the roadway onto a short section of cycle track or shared-use path that bypasses the signal. This is likely the only feasible option for shared bus/bike lane operation since bicycle signals cannot be used for shared-lane applications, bicycles in the shared lane should not be controlled by the vehicular signal (as buses would be blocked by stopped bicy- cles), and bicycles cannot be controlled by a transit signal. Although used in some European

Traffic Control Strategy Toolbox 105 countries, signs exempting bicycles from the traffic signal indications would be inconsistent with the meaning of traffic signal indications provided in the MUTCD. Use with Transit Signal Priority In many of the installations documented in the literature, pre-signals have been combined with forms of TSP at both the pre-signal and the intersection. In these installations, the pre- signal typically stops general traffic as soon as an approaching bus is detected (i.e., buses preempt the pre-signal), allowing the bus to proceed without stopping, while priority (red truncation or green extension, as appropriate) is employed at the downstream intersection. This approach has been shown internationally to have a net person-delay benefit (e.g., Koumara et al. 2007, Guler and Menendez 2013), but significant delays (e.g., 1 to 2 min) occur to motorists during the traf- fic signal cycles when buses arrive because the pre-signal delivers traffic to the intersection less efficiently in this form of operation. This guidebook’s suggestion is to not use priority or preemption at either the pre-signal or main intersection during times when the approach operates over capacity, due to their impact on intersection operations. Preemption of the pre-signal (if not also a pedestrian crossing sig- nal), priority at the pre-signal, priority at the main intersection, or a combination of these, can be considered at other times, but it should be recognized that changing the timing of the pre-signal may make the downstream signal operate less efficiently. Additional Resources • Guide for Geometric Design of Transit Facilities on Highways and Streets (AASHTO 2014)— Section 5.5.7.5 describes the weave-assist application of pre-signals, described as an “advance stop bar for bus left turns.” • Manual on Uniform Traffic Control Devices (FHWA 2009)—Section 4D.27 addresses priority control of traffic signals for bus transit, and Section 8C.09 discusses pre-signals in a railroad crossing context. • TCRP Web-Only Document 66: Improving Transportation Network Efficiency Through Imple- mentation of Transit-Supportive Roadway Strategies—Appendix A presents the results of an international literature review on pre-signals. 6.12 Traffic Signal Installed Specifically for Buses Description An intersection that is signalized primarily to serve bus move- ments rather than general traffic. Purpose Buses may experience significant delays making turns at an unsignalized intersection along a major roadway, but the minor- street traffic volumes may not be sufficient to meet the MUTCD’s volume-based traffic signal warrants. Applications Typical locations where a traffic signal specifically for buses might be considered are (1) unsignalized intersections where buses turn left onto or from a busy major street, (2) transit center and park-and-ride entrances or exits, and (3) off-street busway crossings of public roadways.

106 A Guidebook on Transit-Supportive Roadway Strategies At the time of writing, the MUTCD did not provide a traffic signal warrant specifically for bus operations. However, some of the MUTCD’s signal warrants not related to traffic volumes may be applicable to the intersection and could potentially be used to justify a traffic signal that would also benefit bus operations. These warrants are: • Warrant 4, Pedestrian Volumes. A traffic signal may be warranted with pedestrian volumes crossing the major street of as low as 107 pedestrians per hour in each of 4 h of the day (depending on major-street traffic volumes), or as low as 75 pedestrians per hour when the major street posted or 85th-percentile speed exceeds 35 mph or the city population is under 10,000. A signal may also be warranted with peak-hour pedestrian crossing volumes of as low as 133 or 103 pedestrians per hour (depending on the major street speed and city population). • Warrant 6, Coordinated Signal System. A traffic signal may be warranted if the nearest traffic signals are at least 1,000 ft away and an engineering study determines that traffic platooning on the major street will be improved with the installation of a signal. • Warrant 7, Crash Experience. A traffic signal may be warranted if (1) the intersection has expe- rienced at least five crashes within the last 12 months of types that could be corrected by traffic signal control, (2) other alternatives have failed to reduce the incidence of crashes, and (3) the 8-h traffic volume or the pedestrian volume warrant is met at a reduced level (FHWA 2009). At the time of writing, the National Committee on Uniform Traffic Control Devices had approved text for a proposed new MUTCD chapter on busway grade crossings for FHWA’s con- sideration for the next edition of the MUTCD. If adopted as written, bus-specific signals would be permitted “at busway grade crossings and at intersections where buses operate in mixed traffic in conjunction with standard traffic control signals where special bus signal phases are used to accommodate turning bus vehicles or where additional bus clearance time is desirable” (NCUTCD 2014a). However, until such time that language allowing signals specifically for buses is included in the MUTCD, roadway jurisdictions would need to submit an experimentation request to FHWA (see Appendix D) and have it approved to be able to use such signals. Some Canadian provinces allow the use of a transit signal for transit purposes. Some Canadian jurisdictions permit the use of half signals that serve both pedestrian crossing and transit needs (Corby et al. 2013); however, half signals are not allowed by the MUTCD. Companion Strategies Traffic signals specifically for buses could be implemented using transit signal faces (Sec- tion 6.8) to control bus movements and would typically be used in conjunction with bus-only signal phases (Section 6.9). Transit signal priority (Section 6.7) could also potentially be pro- vided. Traffic signal shadowing (Section 6.6) may be an alternative strategy if a signal specifically for buses is not feasible. Constraints One key potential constraint is regulatory—the ability to justify the traffic signal on the basis of an existing MUTCD warrant or by receiving an experimentation request. Some roadway jurisdictions only consider specific MUTCD warrants as a matter of policy and might need to change their policies to allow use of warrants not based on traffic volumes. Another potential constraint is the effect of a new traffic signal on roadway operations. If the major street currently provides good traffic progression and the installation of a new signal would disrupt that progression, the roadway agency is unlikely to be in favor of a signal at the proposed location. If the signal is primarily for the benefit of the transit agency, the roadway agency may seek to have the transit agency bear the cost of operating and maintaining the signal.

Traffic Control Strategy Toolbox 107 A third potential constraint is the possible effect of a new signal on roadway safety. U.S. experi- ence with signalized crossings of off-street busways with other roadways has been that a number of bus–vehicle crashes have occurred at these locations (Diaz and Hinebaugh 2009), particularly in the first year after installation. Driver expectancy issues may be at the root of the crashes, whether from not expecting two non-coordinated signals in short succession or from becoming accustomed to not having to stop for the signal due to relatively low bus headways. Therefore, a traffic signal installed specifically for buses would preferably have sufficient turning bus, pedes- trian, and minor-street traffic volumes to require major street traffic to stop during most signal cycles so that major street drivers anticipate the potential need to stop as they approach the signal. Benefits Traffic signals specifically for buses are typically installed to address issues with buses turn- ing left onto, turning left from, or crossing major streets and experiencing substantial delays doing so. A traffic signal could reduce bus travel time and travel time variability, but the specific benefits are highly site-specific and would need to be determined by a traffic engineering study. A traffic signal provides a new signalized pedestrian crossing opportunity, thereby improving pedestrian mobility in the area. A traffic signal installed in Calgary, Alberta, on a divided road- way stops traffic in the opposing direction only when a left-turning bus arrives, saving buses up to 90 s compared to waiting for a gap in traffic to make their turns (Jordan et al. 2010). Cost Considerations • Planning and coordination costs. Moderate to high. A traffic analysis would be needed to evaluate the intersection’s current operations (including evaluating the effects on traffic pro- gression) and how they would change with the presence of a new signal. An experimenta- tion request entails additional study requirements (see Appendix D). If the signal would stop major street traffic infrequently, additional traffic control measures (e.g., signs) and motorist outreach programs may need to be considered (Diaz and Hinebaugh 2009). • Capital costs. High—to install a new traffic signal and potentially make ADA-related improve- ments such as curb ramps if not already provided. • Maintenance costs. Moderate, involving added costs for operating and maintaining the signal. • Bus operations costs. Potential savings from reductions in travel time and travel time variability. • Other user costs. Will likely increase delay to general traffic. Implementation Examples Canadian examples include those in Halifax, Nova Scotia (Corby et al. 2013); Calgary, Alberta (Jordan et al. 2010); Edmonton, Alberta; and Vancouver, British Columbia. In the United States, busway crossings have been signalized in conjunction with the South Dade Busway in Miami, Florida, and the Orange Line in Los Angeles, California (Diaz and Hinebaugh 2009). The Lymmo BRT line in Orlando, Florida, includes a signalized bus exit from a parking garage used in part as a park-and-ride facility. Implementation Guidance Before pursuing a traffic signal option, first consider whether rerouting buses to avoid the intersection is a feasible option. If not, an engineering study will be required to evaluate the need for a signal, the impacts of the signal on all roadway users, and potential impacts (positive and negative) on roadway user safety. If a signal would not be warranted on the basis of current

108 A Guidebook on Transit-Supportive Roadway Strategies MUTCD warrants, an experimentation request would need to be prepared and approved by the FHWA to allow its use in the United States until such time that the MUTCD permits the use of signals specifically for buses. Note that a standard FHWA condition of approval is that the jurisdiction agrees to remove the installation (in this case, the signal) if the FHWA determines that the experiment is unsuccessful. Additional Resources • Manual on Uniform Traffic Control Devices (FHWA 2009)—Chapter 4C discusses traffic con- trol signal needs studies. A section of FHWA’s MUTCD website describes the steps involved in the experimentation process: http://mutcd.fhwa.dot.gov/condexper.htm. 6.13 Traffic Control Enforcement Description Automated or manual techniques to enforce traffic laws essen- tial for the successful operation of certain transit-supportive road- way strategies. Purpose The benefits of certain strategies can only be realized if other motorists comply with the traffic control devices used to provide preferential treatment to transit vehicles. Enforcement efforts provide a consequence when motorists do not comply with the device indications or regulations, and these efforts thereby make it more likely that the devices will be respected and the strategy will be effective. Applications Typical enforcement activities relate to: • Enforcing turning movement restrictions for non-transit vehicles (Section 6.1, Section 6.2); • Enforcing yield-to-bus laws (Section 6.3); • Enforcing non-transit vehicle usage of bus lanes (Section 8.1), queue jump lanes (Sec- tion 6.10), and bus-only links (Section 7.7); and • Enforcing parking and stopping restrictions associated with bus lanes (Chapter 8). Enforcement can take place through traditional enforcement efforts involving parking enforce- ment staff or law enforcement officers. If permitted by local laws, photo or video enforcement can be an effective way to enforce bus-only links and bus lanes. Companion Strategies See the Applications section for a list of typical strategies requiring enforcement. Design- ing transit-supportive roadway strategies to be self-evident or self-enforcing to the extent possible, such as by employing painted bus lanes (Section 7.4), can reduce the need for active enforcement.

Traffic Control Strategy Toolbox 109 Constraints Enforcing traffic laws that affect bus operations may be a lower priority for the local police department than enforcing laws that affect traffic safety or addressing community crime issues. State and local laws may need to be changed to permit the use of photo or video enforcement for transit-related purposes. Benefits Enforcement maximizes the benefit of transit-supportive roadway strategies that require other roadway users to respect the traffic controls that provide preferential treatment for buses. With- out enforcement, the investments made in implementing these strategies may not pay off, and support for implementing other strategies in the future may be reduced. A Transport for London study (2006) of automated bus lane enforcement found that through enforcement, bus lane vio- lations had been reduced by 85% and bus delays in bus lanes reduced by 15%. New York City credits automated bus lane enforcement as one of the factors behind a 15% to 23% improvement in bus speeds on three BRT routes (New York City DOT 2012), with the bus lanes themselves and stop consolidation being other major factors. Cost Considerations • Planning and coordination costs. High. A transit agency’s own police force, if one exists, may be able to conduct enforcement, or coordination with local police departments may be necessary (or both). Part-time bus lanes (Section 8.1) will require coordination with towing companies to ensure that the lanes are clear of parked vehicles when in operation. For photo and video enforcement, agency staff time will be required to work with state legislators and local jurisdictions to authorize the use of automated enforcement, and public outreach will be needed to inform motorists about the new enforcement techniques. Depending on the type of strategy being enforced, other stakeholders may include business owners (for parking restrictions), neighborhood organizations (for bus-only links and turn restriction exemp- tions), district attorneys, and traffic court judges (AASHTO 2014). • Capital costs. Potentially none (traditional enforcement) to high on a per-site basis (auto- mated enforcement). AASHTO (2014) suggests the potential need to incorporate enforce- ment areas (e.g., extended-length pullouts) into bus lane projects. • Maintenance costs. Potentially none (traditional enforcement) to moderate (added costs to maintain camera equipment). Traffic control devices (e.g., signs and markings) will need to be adequately maintained for rules to be enforceable. • Bus operations costs. For traditional enforcement, the costs will depend on how often enforcement activities are undertaken and who performs them (e.g., local police departments, who may wish to be reimbursed for their costs, or transit agency police, who may need addi- tional staff to add traffic control enforcement efforts to their existing duties.) For automated enforcement, there may be significant costs associated with processing violations, and these may be able to be recouped from the collected fines, depending on how the authorizing law is written. • Other user costs. Strict enforcement of parking and delivery activities may affect local resi- dents and businesses. Implementation Examples The New York State legislature granted New York City DOT and MTA the ability to install bus lane enforcement cameras on specified SBS routes. As of 2012, New York City DOT had

110 A Guidebook on Transit-Supportive Roadway Strategies installed cameras at 20 locations on three bus routes. Between April 2011 and March 2012, the cameras, in total, recorded approximately 6,000 bus lane moving violations per month, of which 14% were challenged, with 17% of the challenges being upheld, which is equivalent to 2% of all violation notices issued. As of 2012, the system had accrued about $2.6 million in capital costs and $860,000 in operating costs, which were offset by over $7.5 million in collected fines. Additionally, MTA installed in-bus cameras on six buses on one SBS route as a pilot project to record parking violations in bus lanes. A parking violation was determined to occur when the same vehicle was photographed by successive camera-equipped buses. The agencies considered the enforcement program to be a success in terms of covering its cost and in contributing to improved bus speeds and improved passenger perceptions of service reliability (New York City DOT 2012). In 2007, the California legislature granted San Francisco the ability to conduct a pilot test of video enforcement of bus lane parking violations through 2011, which was subsequently extended through 2015. Video cameras were installed on 30 buses, with the footage reviewed by two parking control officers. Over 3,000 citations were issued in 2011, resulting in over $300,000 in fines. No information was provided about the cost of the program (SFMTA 2012). Implementation Guidance Enforcement begins by clearly informing motorists of the presence of the traffic control through clearly visible signs and pavement markings. These measures help reduce bus lane and other traffic control violations by inattentive motorists. Regular enforcement efforts, in combination with sufficiently high fines, are necessary to deter willful violators. Posting the fine amount has been shown to be effective in reducing viola- tions. In the absence of automated enforcement, enforcement efforts will need to be targeted to specific facilities at specific times (AASHTO 2014). This approach allows a large percentage of the motorists who regularly violate the traffic control at those locations to be caught, and it is visible—and thus potentially effective as a deterrent—to other roadway users. Automated enforcement allows many bus facilities to be continually monitored and can recoup enforcement costs if agencies are allowed to keep some or all of the fines collected. To be effec- tive in discouraging repeat offenses and to preserve a public perception that the enforcement is being conducted to allow buses to operate as efficiently as possible as opposed to being a revenue generator, violation notices should be sent out as soon as possible after the violation so that recipients can still remember what they were doing at the time. It is particularly important from a public relations standpoint to make sure that cameras are placed where there is no question that the law is being violated (a particular issue where vehicles are allowed to enter the bus lane to make right turns) and that the system can differentiate between legal and illegal bus lane uses. Additional Resources • Guide for Geometric Design of Transit Facilities on Highways and Streets (AASHTO 2014)— Section 5.5.7.5 describes the weave-assist application of pre-signals, described as an “advance stop bar for bus left turns.” • 2012 Bus Lane Camera Enforcement Update Report (New York City DOT 2012)—this report summarizes how New York implemented its enforcement program, including its outreach and education efforts.