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.

111 This chapter is the third of four toolbox chapters presenting potential strategies for improving bus speeds and reliability. The strategies presented in this chapter focus on physical changes to the roadway that can improve bus operations. Bus lanes are also a type of infrastructure strategy, but given the wide variety of ways they can be implemented, they are discussed separately in Chapter 8. Chapter 7 defines and discusses the following strategies: • Modifying speed humps, • Lengthening bus stops, • Bus shoulder use, • Red-colored pavement, • Curb extensions, • Boarding islands, and • Bus-only links. The introduction to Chapter 5 describes how each strategy section is organized. 7.1 Speed Hump Modifications Description Speed bumps and humps along bus routes are replaced with bus-friendlier versions. Purpose Speed bumps and humps that are relatively short (e.g., 3 to 6 ft long) force buses to slow to speeds (e.g., 15 mph or less) that are much slower than the street’s posted speed to avoid uncomfort- able (or dangerous, for standing passengers) jolts to passengers and damage to the bus’s suspension. Because buses accelerate more slowly than automobiles, they experience more delay from speed bumps. The bus acceleration and deceleration associated with speed humps also consumes more fuel and can create noise impacts to adjacent land users (TransLink 2002, TriMet 2005, BC Transit 2010). Replacing them with alternative designs that buses can traverse at the street’s posted speed, as well as avoiding install- ing new speed humps along bus routes, can avoid these impacts. C H A P T E R 7 Infrastructure Strategy Toolbox

112 A Guidebook on Transit-Supportive Roadway Strategies Applications Any short speed bump or hump along a bus route is a candidate for this strategy. Bus-friendlier alternative designs include: • Speed tables. Speed tables are longer (e.g., 22 ft), flat-topped elevations of the roadway surface that raise the entire vehicle wheelbase and can also be used in conjunction with raised pedestrian crosswalks; and • Speed cushions. Speed cushions are speed humps or speed tables that provide wheel cutouts that allow wider-wheelbase vehicles such as buses and emergency vehicles to pass without a bump while still reducing automobile speeds (National Association of City Transportation Officials [NACTO] 2013). Malmö, Sweden, has used a modified speed table design (pictured at the beginning of this section) on transit and emergency vehicle routes where the roadway surface is raised quickly on the entry side, similar to a speed table, but is then lowered gently back to grade on the departure side. Companion Strategies This strategy can be implemented by itself or as part of a package of transit-supportive road- way strategies along a street. Constraints Roadway agency design manuals may need to be updated to allow the use of alternative designs. Roadway agency traffic-calming policies may need to be updated to discourage the installation of new speed humps along bus routes. Benefits Replacing speed humps with bus-friendlier designs can retain the desired traffic-calming effect while improving bus passengers’ comfort, improving bus fuel economy (by avoiding the need to accelerate after the hump), and reducing noise impacts in the vicinity of the speed hump. Emergency vehicles will also benefit from bus-friendly speed hump designs. Cost Considerations • Planning and coordination costs. Moderate. Up-front coordination will be needed with the roadway agency to approve bus-friendlier traffic-calming designs or to develop policies discouraging or preventing speed hump use on bus routes or designated transit streets. Neigh- borhood outreach is suggested when changing an existing speed hump to a bus-friendlier design. Emergency responders may be supportive of speed hump changes that allow faster emergency vehicle response times. • Capital costs. Low to moderate costs to remove or replace the speed hump. • Maintenance costs. A bus-friendlier design may reduce pavement damage caused by buses decelerating and traveling over a speed hump. • Bus operations costs. Potential savings from reductions in travel time and improved fuel economy. • Other user costs. Removing an existing speed hump may result in higher traffic speeds along the street before and after where the hump was located, which has potential safety impacts.

Infrastructure Strategy Toolbox 113 Implementation Examples No U.S. or Canadian examples were identified in the literature. As noted previously, Malmö, Sweden, has implemented bus-friendlier speed tables. Implementation Guidance U.S. and Canadian transit agency design guidelines discourage the use of speed humps along transit routes for the reasons described in the Purpose section. Alternative traffic-calming strategies should be investigated first. When speed humps must be used, it is suggested that they: • Not be installed near bus stops since passengers may be moving to or from their seats during this time, • Provide as long a distance as possible (e.g., 22 ft) between the slope up and the slope down or be designed such that buses avoid the bump (e.g., a speed cushion), • Provide at least 600 ft between successive bumps, and • Be located so that buses can traverse them at a 90-degree angle (e.g., not near bus stops) (TransLink 2002, TriMet 2005). Additional Resources • Urban Street Design Guide (NACTO 2013)—Pages 54 and 55 describe speed tables and speed cushions, respectively. 7.2 Bus Stop Lengthening Description A bus stop’s length is increased to allow it to serve more (or longer) buses simultaneously. Purpose If more buses want to use a stop at one time than space exists to allow, the other buses have to wait in the street until space opens up at the stop. This delays both buses and general traffic. Matching a bus stop’s capacity to serve buses to the scheduled number of buses can minimize the potential for bus stop failure to occur. Applications Bus stop failure can occur for several reasons: (1) the number of buses scheduled to use the stop over the course of an hour exceeds the bus stop’s capacity, (2) the number of buses scheduled to use the stop over a short period of time exceeds the number of loading areas pro- vided, or (3) schedule irregularities and bus bunching result in more buses arriving at a time than the stop can accommodate. The second and third reasons can be addressed first through schedule adjustments and transit agency actions to improve bus reliability and do not neces- sarily require lengthening stops. The first reason cannot be addressed by scheduling alone and requires lengthening bus stops to provide more capacity, changing route patterns so that fewer buses use the stop, or both.

114 A Guidebook on Transit-Supportive Roadway Strategies Companion Strategies Bus stop lengthening may need to be considered when stops are consolidated (Section 5.2) since the increased passenger activity at the remaining stops will increase bus dwell times and thus reduce the number of buses that a bus stop can accommodate during an hour. Bus stops may also need to be lengthened when longer buses are introduced (Section 5.5). If a bus stop cannot be lengthened at its current location, it may need to be relocated (Section 5.1). Constraints Lengthening bus stops may result in a loss of on-street parking. It may not be feasible if drive- ways, alleys, or intersections are located close to the stop. If one stop requires lengthening, there is a good chance that other stops with the same number of loading areas, where buses dwell as long or longer, will also require lengthening. Benefits When bus stop failure occurs, the delay experienced by a bus can last up to the dwell time and subsequent traffic signal delay time of the buses already using the stop. Lengthening the bus stop will reduce the probability that failure occurs, thereby improving travel time variability. When the stop is offline (e.g., in the parking lane or at a pullout), general traffic delay and travel time variability will improve to the extent that bus stop failure is reduced. The TCQSM (Kittelson & Associates et al. 2013) can be used to estimate how often bus failure occurs, or AVL data or field measurements can be used to determine bus delay directly. Cost Considerations • Planning and coordination costs. Relatively low on a per-site basis. Lengthening a stop will need to be coordinated with the appropriate roadway agency. It is also desirable to engage adjacent property owners in advance about potential negative impacts to them (e.g., loss of parking). • Capital costs. Typically relatively low on a per-stop basis, consisting of moving parking signs and making any required ADA improvements such as a landing pad. The need for concrete paving at the bus stop to reduce bus-caused pavement damage may also be considered. Costs will be higher when curb lines or parking meters need to be moved to accommodate a longer stop. • Maintenance costs. No significant change in costs. • Bus operations costs. Reduces bus travel time variability. • Other user costs. For offline stops, reduced delay for motor vehicles that would otherwise be blocked by buses waiting in the travel lane to enter the stop. Potential loss of on-street parking. Implementation Examples Thirteen of 59 transit agencies responding to a survey reported increasing bus stop length to improve bus speeds (Boyle 2013). Implementation Guidance The TCQSM (Kittelson & Associates et al. 2013) can be used to estimate hourly bus stop capacity given a desired failure rate and provides recommended failure rates for different land use contexts (e.g., downtown, suburbs). If the calculated hourly bus stop capacity is adequate, yet bus stop failures occur significantly more often than the failure rate used in the calculation, this is a sign of either schedule reliability problems or scheduling issues over a short period of

Infrastructure Strategy Toolbox 115 time that cause too many buses to arrive at the stop at once. In these cases, the transit agency may wish to address the cause of the problem first (scheduling or reliability) rather than the symptom (bus stop failure). If the calculated hourly bus stop capacity is inadequate, the capacity of other bus stops along the street with the same or longer dwell times should also be checked since fixing the problem at one stop may simply move it to another stop, and it would be preferable to address all of the corridor or route’s capacity issues at one time. If capacity is inadequate, but lengthening stops is physically or politically infeasible, the transit agency may wish to consider skip-stop operations, where buses are divided into groups and assigned specific sets of stops to serve (e.g., every other existing stop). This is a form of stop consolidation (Section 5.2) where the number of stops served by a given route is reduced, although the physical number of stops is unchanged. The available bus stop capacity can thus serve a greater number of buses (up to nearly twice as many) compared to having all buses serve all stops. The TCQSM describes skip-stop operations and analyzing their capacity in detail. 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 bus stop capacity and describes skip-stop operations and their impact on bus capacity. • TCRP Project A-42, “Minutes Matter: A Guide to Bus Transit Service Reliability,” which began in 2015 with the objective of providing comprehensive guidance to transit agencies on ways to improve their bus reliability. 7.3 Bus Shoulder Use Description Buses are allowed to use roadway shoulders during peak periods. Purpose To avoid congestion in the general traffic lanes and thereby gain a speed advantage on general traffic. Applications A typical arterial roadway application is a suburban divided high- way with occasional signalized intersections. Minneapolis uses the following criteria to identify potential corridors: • Peak-period running speeds between intersections are regularly 35 mph or less for general traffic, or intersection approaches regularly have continuous queues during peak hours; • A minimum of six transit buses per day are likely to use the shoulder (i.e., are scheduled to operate during periods when congestion occurs); • The anticipated time savings must be at least 8 bus-minutes per mile per week; and • The shoulder must be at least 10 ft wide, although pinch points where buses merge back into regular traffic are permitted (Martin et al. 2012). Companion Strategies Bus shoulder use can be implemented in smaller sections as part of queue jump (Section 6.10) and queue bypass (Section 8.6) projects. Transit signal priority (Section 6.7) can be implemented in

116 A Guidebook on Transit-Supportive Roadway Strategies conjunction with shoulder operation. Bus stops are typically located on right-turn channelization islands (Section 7.6) at signalized intersections. Periodic enforcement efforts (Section 6.13) may be required to ensure that only authorized vehicles use the shoulder. Constraints The shoulder must be sufficiently wide to permit buses to operate. TCRP Report 151: A Guide for Implementing Bus On Shoulder (BOS) Systems (Martin et al. 2012) suggests a minimum shoulder width of 10 ft, based on successful freeway and arterial implementations in the United States and Canada. Occasional pinch points where the shoulder narrows, such as on bridges, may be tolerated, but buses will need to merge into the adjacent general traffic lane at those points. The shoulder must be capable of supporting the weight of the number of buses expected to use it. Shoulder use by buses may require changes to state traffic laws, and individual projects may require exceptions to roadway agency design standards. Bus shoulder operation on roadways with designated bicycle facilities or routes on the shoulder is not suggested. Benefits The potential benefit will depend on how long the general traffic lanes are congested and how far that congestion extends. Travel time surveys of three directional arterial corridors in the Minneapolis region indicated average time savings of 1.5 to 2.5 min plus reductions in travel time variability. For safety reasons, buses are typically limited to traveling 10 to 15 mph faster than general traffic while using shoulder lanes, which limits how much of a travel time benefit can be achieved. Passengers perceive the travel time savings (and sometimes perceive greater savings than actually occur) and also perceive improved travel time reliability (Martin et al. 2012). Shoulder operation on an arterial expressway in Calgary, Canada, saves buses up to 15 min during peak periods (Jordan et al. 2010). Cost Considerations • Planning and coordination costs. High. The suitability of the corridor will need to be evaluated with respect to safety and operations at intersections and access points, shoulder width, and pavement strength. New signs will need to be developed the first time a shoulder facility is implemented by a roadway agency, and exceptions to roadway agency design policy will need to be documented (or the policy updated with provisions for shoulder use). Coor- dination with law enforcement will be required. Public outreach is desirable, particularly for the first implementation in an area. Training will be needed for the bus drivers who will use the facility. • Capital costs. Moderate to high. At a minimum, new signs will be required along the length of the corridor. Other potential costs may be for relocating or upgrading bus stops and pedestrian access routes to those stops, relocating bicycle facilities in the corridor, widening the shoulder, strengthening the shoulder, and constructing pullouts for enforcement activities. • Maintenance costs. More frequent sweeping may be required to ensure the shoulder is free of debris. • Bus operations costs. Reduces bus travel times and travel time variability. Because shoulder bus use is often implemented on long-distance commuter bus routes operating on suburban highways, buses frequently make only one peak-period trip, and the time savings may therefore not translate into using the bus for additional trips (Martin et al. 2012). There may be additional fuel economy savings related to avoiding stop-and-go traffic. • Other user costs. In general, there is no impact to other roadway users (Martin et al. 2012).

Infrastructure Strategy Toolbox 117 Implementation Examples As of January 2011, shoulder operation on arterial roadways had been implemented in the following locations in the United States: • Minneapolis region, Minnesota (many corridors); • Burtonsville, Maryland (U.S. 29); • Kenmore, Washington (SR 522); and • Mountainside (U.S. 22) and Old Bridge (U.S. 9), New Jersey (Martin et al. 2012). Shoulder operation is also used on a section of Crowchild Trail in Calgary, Alberta, that experiences queues up to 1.5 km (0.9 miles) long during peak periods (Jordan et al. 2010). Implementation Guidance TCRP Report 151 provides comprehensive guidance on the planning, design, and operational considerations associated with implementing bus shoulder operations. Additional Resources • TCRP Report 151: A Guide for Implementing Bus On Shoulder (BOS) Systems (Martin et al. 2012)—comprehensive implementation guidance, accompanied by case studies of successful implementations. 7.4 Red-Colored Pavement Description All or selected segments of a bus lane are indicated with red- colored pavement as a supplement to the normal bus lane signs and striping. Purpose To improve the conspicuity of the bus lane and thereby reduce the number of bus lane violations by unauthorized vehicles. Applications This strategy can be considered anywhere a roadway lane is reserved exclusively or primarily for buses. The greater the number of buses using the bus lane, the greater the impact of bus lane viola- tions on bus operations and thus the greater the potential benefit offered by the strategy. The colored pavement can be applied solely at the start of a lane (e.g., to guide turning vehicles away from the bus lane), only in the sections where only buses are permitted (e.g., to indicate where vehicles may enter the lane to make right turns), or for the full length of the lane, including sections where other vehicles are permitted by law to briefly enter the lane (e.g., to enter or cross the lane to make a right turn, to stop to immediately pick up or drop off passengers); however, it should be applied consistently within a jurisdiction. Companion Strategies Red-colored pavement can be used in combination with turn lanes designated exclusively for buses (Section 6.1), queue jump lanes designated exclusively for buses (Section 6.10), bus-only

118 A Guidebook on Transit-Supportive Roadway Strategies links (Section 7.7), and most types of bus lanes (Section 8.1) except shared bus and bicycle lanes (Section 8.3). Constraints At the time of writing, the use of red-colored pavement for transit was expected to be permitted in the next edition of the MUTCD, which is anticipated to be published in 2017. Until such time that it appears in the MUTCD (or the FHWA issues an Interim Approval for its use), roadway agencies in the United States will need to submit an experimentation request to FHWA, and have it approved, to be able to use this strategy. An experimentation request template is provided in Appendix D. New York City’s testing of different forms of red coloring treatments found no treatment that lasted more than one year on Portland cement concrete surfaces (Carry et al. 2014). Benefits Red-colored pavement would be expected to reduce the number of bus lane violations. A study of 61 bus lane segments in New York City found no significant difference in the occurrence of obstructions (other roadway users legally or illegally entering the lane) when comparing lanes with white bus lane pavement markings only to those lanes supplemented with red coloring (Safran et al. 2014). However, red coloring was highly correlated with interior bus lanes, while lack of red color- ing was highly correlated with curbside bus lanes, and interior bus lanes showed significantly lower obstruction rates. In addition, the same study found the bus driver used red lanes 52% more often than non-red lanes, indicating a greater degree of bus driver confidence in red-colored lanes being unobstructed. More research is needed to quantify the contribution of red-colored pavement to reductions in bus lane violations and reductions in bus delay and travel time variability. Cost Considerations • Planning and coordination costs. High at present, due to the need for FHWA experimenta- tion and the corresponding documentation. Once incorporated into the MUTCD, the strategy would require roadway agencies to develop policies on when to apply red-colored pavement. A moderate level of planning and coordination would be needed when applying the strategy to existing bus lanes, but a relatively small increment of additional planning would be needed when incorporating the strategy into a new bus lane project. • Capital costs. Moderate to high since a large surface area may need to be colored. • Maintenance costs. Will increase maintenance costs since the coloring will need to be reapplied periodically. • Bus operations costs. Expected to help reduce bus travel times and travel time variability, but little quantitative information was available at the time of writing. • Other user costs. May reduce the likelihood that drivers will inadvertently drive in a bus lane. May make bus lanes easier to enforce due to the extra conspicuity provided by the red coloring. Implementation Examples As of early 2015, U.S. implementations included New York City, San Francisco, Chicago, and Seattle. This strategy is used in many cities internationally. Implementation Guidance The proposed MUTCD language indicates that travel lanes “used by public transit vehicles and other modes” (e.g., shared bus and bicycle or bus and taxi lanes) “should not use red-colored

Infrastructure Strategy Toolbox 119 pavement.” The proposed language also indicates that red-colored pavement “shall be applied only in lanes, areas, or locations where general-purpose traffic is generally prohibited to use, queue, wait, idle, or otherwise occupy the lane area or location where red-colored pavement is used” and that “regulatory signs shall also be used when it is determined that other vehicles will be allowed to enter the lane to turn or bypass queues” (NCUTCD 2014b). As proposed, this language seems to indicate that red-colored pavement could be used in sections where specified classes of vehicles (e.g., right-turning vehicles) are allowed limited access to the lane but not in sections where general traffic has unrestricted use of the lane at all times nor for long sections of a lane that allow shared use by other modes. Red-colored pavement is considered supplemental to the signs and pavement markings (e.g., solid white stripe, bus-only markings) that are required to enforce bus lanes. As pavement markings are allowed for part-time bus lanes with signs indicating the times the bus lane is in effect, it follows that red-colored pavement would also be allowed for part-time bus lanes. The use of red-colored pavement may depend on local laws governing bus lanes. If other vehicles are permitted to use the lane to make a right turn at the next driveway or cross street, then non- transit vehicles may be present at any point along the bus lane, and using red coloring for the entire lane may be appropriate to discourage through-traffic use. On the other hand, if vehicles are only allowed to enter or cross the lane at designated points (e.g., where a right-turn lane begins), then ending the coloring at that point would provide a visual cue to direct motorists to the desired location for them to enter or cross the lane. To date, no research has been done on the comparative effectiveness of different red-colored pavement applications (e.g., for a short distance after an access point or for the full length of the lane, not where the bus lane becomes a right-turn lane). Carry et al. (2014) describe the results of durability and skid-resistance testing of different colored pavement treatments. They found that “red epoxy-based street paint, an epoxy with red aggregate product, and a red asphalt concrete-based micro surface performed well across all field and laboratory tests.” They found no treatment that lasted more than 1 year when applied to Portland cement concrete surfaces. Additional Resources • Guide for Geometric Design of Transit Facilities on Highways and Streets (AASHTO 2014)— Section 5.5.6.4.2 describes colored pavement applications. • Red Bus Lane Treatment Evaluation (Carry et al. 2014)—paper describing New York City’s efforts to test the durability and skid resistance of different types of colored pavement treatments. 7.5 Curb Extensions Description Curb extensions (bus bulbs, bus nubs) extend the curb and sidewalk out to the edge of the parking lane. Purpose This strategy allows buses to stop in the travel lane and thereby avoid delay waiting for a gap in traffic (reentry delay) when leav- ing the stop. When used at intersections, it reduces the pedestrian crossing distance, which reduces pedestrian exposure to traffic con- flicts. At signalized intersections, it also reduces the time required

120 A Guidebook on Transit-Supportive Roadway Strategies to serve pedestrian movements, which may allow a shorter traffic signal cycle length, which can also reduce bus delay. Applications Curb extensions are particularly suited to areas with high-density development, where the percentage of people moving through the corridor as pedestrians or in transit vehicles is relatively high in comparison with the percentage of people moving in automobiles. On-street parking is a prerequisite since curb extensions are constructed within the area used by the parking lane. Companion Strategies Curb extensions can be used in combination with interior bus lanes (Section 8.4). Yield-to-bus laws (Section 6.3) are another way of tackling the problem of reentry delay. Constraints Curb extensions affect street drainage patterns, and drainage may need to be reworked to prevent water ponding issues. When used at intersections, they reduce the turning radius available for larger vehicles, which may require restrictions on right turns or moving the side-street stop bar away from the intersection to provide more room for larger turning vehicles. If bicycle facilities exist, consideration will need to be given to how to route bicycles around stopped buses (see Appendix C) (Fitzpatrick et al. 2001). The ability to match the roadway and sidewalk cross- slopes so that a low-floor bus’s wheelchair ramps can deploy at an ADA-acceptable slope should be carefully considered (TriMet 2010). Benefits The potential benefit from curb extensions 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). However, as with any strategy involving spot locations, the sum of the delay benefit for a route or along a street will generally be less than the sum of the individual bus stop delay benefits since the delay saved at one stop is sometimes lost at a downstream traffic signal. See Section 4.4 for more information. A study of curb extensions along a four-lane street in San Francisco found that average bus speeds within the block where the stop was located improved by an average 0.2 to 2.2 mph, while average vehicle speeds within the block improved by an average 4.5 to 8.5 mph. Over the length of a 2,400-ft section of the street containing seven bus stops and six traffic signals, average bus speeds improved by 0.5 mph, while average vehicle speeds improved by 3 to 7 mph. Reasons given for the improvement in vehicle speeds are that 48% to 72% of buses would stop partly in the travel lane (i.e., not pull up to the curb) prior to the construction of curb extensions and thus disrupted traffic flow, and that buses would sometimes use both travel lanes when exiting the stop (Fitzpatrick et al. 2001). Following the construction of curb extensions, bus stopping pat- terns would have become more predictable to motorists, and buses would not have encroached on the second lane. A simulation study of the same corridor, where buses did pull to the curb in the before case, found smaller effects: speeds through the corridor improved by 0 to 2 mph for both vehicles and buses (Fitzpatrick et al. 2001).

Infrastructure Strategy Toolbox 121 When used at intersections, curb extensions shorten the pedestrian crossing distance, thus reducing the amount of time pedestrians are exposed to conflicts with other road users and, potentially, the amount of time that other road users are delayed by pedestrians crossing. At signalized intersections, the reduced crossing distance results in less flashing “Don’t Walk” time; this time can be used for longer walk times (reducing pedestrian delay) or to decrease the traffic sig- nal cycle length (if minimum pedestrian phase lengths determine the length of side-street phases), potentially resulting in reduced delay for all intersection users. A review of state DOT design manu- als found that while most did not include sections covering transit-supportive roadway strategies, a number provided the option of using curb extensions as a strategy to benefit pedestrians. A study of curb extensions in San Francisco found that curb extensions provided both a better bus stop waiting environment (in terms of the space available per waiting passenger) and better adjacent sidewalk flow by giving bus passengers a place to wait other than the sidewalk (Fitzpatrick et al. 2001). The added space can be used to add bus stop amenities such as bus shelters, thereby reducing the potential for waiting passengers to congregate in front of businesses’ display windows and doors. Curb extensions can increase the amount of on-street parking provided since the parking lane can be continued up to the start of the bus stop. Without a curb extension, parking needs to be prohibited before or after the stop (or both) to give buses the opportunity to maneuver from the travel lane to the curb and vice versa (Kittelson & Associates et al. 2013). Cost Considerations • Planning and coordination costs. Moderate. A traffic analysis may be needed to evaluate the impact of the curb extension on vehicular traffic. Civil engineering plans will need to be developed to address street drainage modifications. Outreach to adjacent businesses is suggested, particularly when installing shelters that may block views of businesses’ signs from the street (Fitzpatrick et al. 2001). • Capital costs. Moderate, with the largest portion of the cost involving drainage changes and, potentially, utility relocations (Danaher 2010). There will be added costs if bicycle facilities are to be relocated around the bus stop. • Maintenance costs. No significant change expected. • Bus operations costs. Reduces bus travel time and travel time variability. • Other user costs. On streets with one lane of travel per direction, will tend to increase vehicular delay, with the extent of the delay dependent on bus frequencies, dwell times, traffic volumes, and whether the stop is located at a signalized intersection (because traffic might need to stop anyway). On streets with two or more travel lanes per direction, one study found a decrease in motorized traffic delay (Fitzpatrick et al. 2001). Implementation Examples A survey of 52 transit agencies found that 25% had implemented at least one curb extension (Danaher 2010). Portland, Oregon; San Francisco, California; Seattle, Washington; and Vancouver, Canada, are among the cities that have widely implemented curb extensions and developed formal guidelines on their use (Fitzpatrick et al. 2001). Implementation Guidance Conditions supportive of installing curb extensions include: • Presence of full-time curbside parking; • Near-side or midblock stop locations;

122 A Guidebook on Transit-Supportive Roadway Strategies • Relatively low traffic speeds (35 mph or less); • Low to moderate traffic volumes (<500 vehicles per hour per lane in the same direction); • Two or more travel lanes in the direction of travel, to allow passing (desirable but not essential); • Relatively high sidewalk or crosswalk usage or relatively high passenger volumes using the stop (e.g., is sidewalk flow or access to adjacent businesses affected by passengers waiting on the sidewalk?); and • Relatively low right-turning volumes, particularly larger vehicles such as trucks and buses (Danaher 2010, Fitzpatrick et al. 2001). Conditions requiring special attention include complex drainage issues, streets with bicycle facilities, and intersections where the right-turning traffic volume might require a right-turn lane (Danaher 2010). Far-side locations on streets with only one travel lane in the direction of travel are not recommended due to the potential for queues behind the bus to block the intersection. A traffic analysis, based on expected bus frequencies, average dwell times, vehicular volumes, and estimated improvements in pedestrian crossing delay, can determine the typical level of queuing and vehicle delay that would be expected as a result of buses serving a stop with a curb extension. The TCQSM can be used to estimate the reentry delay saved by buses and their passen- gers. Appendix C of this guidebook provides guidance for accommodating bicycles at bus stops. Additional Resources • TCRP Report 65: Evaluation of Bus Bulbs (Fitzpatrick et al. 2001)—implementation examples and application guidance. • Guide for Geometric Design of Transit Facilities on Highways and Streets (AASHTO 2014)— Section 5.2.2.2 provides guidance on curb extensions. The AASHTO guide’s list of situations that are not recommended for transit is more auto-centric than this guidebook’s; conditions that are not listed here (e.g., only one lane available in the direction of travel, low pedestrian volumes) may warrant analysis but should not by themselves disqualify a location from consideration. As with any other type of transit-supportive roadway strategy, the benefit provided to transit passengers and operators should be weighed against the cost of implementing the strategy, including disbenefits to other roadway users. Section 7.1.4.3 describes the benefits of curb extensions for pedestrians. 7.6 Boarding Islands Description Bus stops on raised concrete islands within the roadway. Purpose Boarding islands are a supporting strategy that allows bus stops to remain at intersections when another strategy is implemented. Applications Figure 9 illustrates three potential applications for boarding islands. For simplicity in presenting the concept, bicycle facilities are not shown but can be incorporated into the design as dis- cussed in the Implementation Guidance section. Source: © 2015 Google

Infrastructure Strategy Toolbox 123 Figure 9(a) shows a boarding island on a right-turn channelizing island where buses stop in the travel lane to serve the stop. The island needs to be wide enough to provide at least the 8-ft by 5-ft clear area required by the ADA next to where the front door of a bus would stop. Not shown in the illustration, but also potentially needed, are bollards to protect the boarding area from errant vehicles and pedestrian fencing or similar barriers to limit pedestrian access to areas, as wanted. Figure 9(b) shows a boarding island (in conjunction with a queue jump) located on a larger right-turn channelizing island. It is similar to the concept shown in Figure 9(a) but provides more passenger waiting area and allows buses to stop in a short bus lane. Figure 9(c) shows a boarding island in the interior of the roadway and served by a short bus lane. The configuration shown in the illustration could support buses transitioning into a median bus lane (Section 8.7) beyond the intersection or buses making a left turn at the intersection. A similar configuration could support a left-side bus lane on a one-way street (Section 8.5). Space permitting, it would be possible to provide a conventional left-turn lane for general traffic to the left of the bus lane or on the right side of the boarding island; in the latter case, a bus-only signal phase (Section 6.9) would be required for buses departing the stop. The bus stop itself would be configured similarly to median bus stops on bus rapid transit lines by using a ramp connecting the platform to the crosswalk. As with the other boarding island configurations, the island would need to be at least 8 ft wide to provide the minimum required ADA clear area, and pedestrian fencing or other similar barriers may need to be considered. A bus stop could also be provided on a large-enough right-turn channelizing island on the far side of an intersection. In this case, a short bus lane would be preferred so that buses could stop out of the traffic lanes without other vehicles possibly stopping behind them and blocking the intersection. Design considerations with this application include managing the area where cross-street right-turning traffic enters the main street and managing conflicts with buses exiting the bus stop. BU S BU S ONLY BUS ONLY BUS BU S (a) RIGHT-TURN CHANNELIZATION (b) RIGHT-TURN CHANNELIZATION PLUS BUS LANE (c) INTERIOR BOARDING ISLAND Figure 9. Illustrative boarding island configurations.

124 A Guidebook on Transit-Supportive Roadway Strategies Companion Strategies This strategy supports bus-only signal phases (Section 6.9—for example, a bus left turn from a right-side lane), queue jumps (Section 6.10), most forms of bus lanes (Section 8.1), and other strategies that can be used in combination with queue jumps and bus lanes. For example, a short bus lane could be highlighted with red pavement coloring (Section 7.4). Constraints Sufficient space needs to be available on the island to provide the minimum required ADA clear area for each bus loading area provided at the stop. Potential sight-distance issues created by a bus shelter or stopped buses are suggested to be considered when placing bus stops on right- turn channelization islands. Right-turn channelization islands large enough to accommodate a bus stop are more likely to be found in suburban areas where right-of-way may be less constrained and where roadway designs provide larger vehicle turning radii. Benefits When the location of passenger generators or other considerations suggest the need for a near-side stop, it can be difficult to find a suitable location when a right-turn lane is provided. Right-turning traffic will delay buses trying to access a bus stop located at the stop bar, while a stop located prior to the start of the right-turn lane may require significant extra walking distance for most passengers and thereby discourage ridership. Placing a stop on a boarding island can allow the stop to be located at the intersection without buses experiencing the negative effects of the right-turning traffic. Unless buses are equipped with doors on both the right and left sides, boarding islands are required when bus stops are to be provided along bus lanes located on the left side of the street. In either case, the boarding island itself provides no special benefit; rather, it makes other strate- gies feasible while maintaining good access to bus service. Cost Considerations • Planning and coordination costs. Low to moderate for right-turn channelizing islands, depending on how much modification the island will require; sight distances will need to be evaluated. Moderate to high for channelizing islands elsewhere in the roadway due to the likely need to realign other travel lanes, but may be incorporated as part of a larger bus lane or roadway improvement project. • Capital costs. Low (for existing right-turn channelization islands that require no modification) to high (new island construction and changes to the roadway). Pedestrian fencing, bollards, and MUTCD object markers may be required. Concrete paving at the bus stop to reduce bus-caused pavement damage may also be needed. • Maintenance costs. Unchanged (for existing right-turn channelization islands that require no modification) to moderate increases (to replace or repair damaged bollards, fencing, etc.) • Bus operations costs. No direct change. • Other user costs. May reduce bus-caused delays to right-turning traffic when used in combi- nation with a channelizing island. Implementation Examples • San Francisco. Examples of ADA-compliant boarding islands serving left-side bus lanes are Bush Street at the near side of Battery Street, Folsom Street at the far side of 2nd Street,

Infrastructure Strategy Toolbox 125 Fremont Street at the near side of Market Street, and Beale Street at the far side of Howard Street. San Francisco is also constructing ADA-compliant interior boarding islands at some street-running light rail stops. Market Street has a number of examples of pre-ADA street car-era interior boarding islands. • New York. An interior boarding island is used to serve a bus stop on the left side of White Plains Road at the Gun Hill Road subway station in the Bronx. At the time of writing, an interior boarding island was being considered for Third Avenue at East 57th Street in Manhattan. This island, in conjunction with an offset bus lane in the third lane from the right curb, would allow buses to avoid heavy right-turning traffic in the two right lanes and allow a stop to be placed within a six-block section of Third Avenue currently lacking stops (New York City DOT 2014). • Atlanta. Bus stops are provided on right-turn channelizing islands in conjunction with queue jumps at several locations along Memorial Drive in suburban Atlanta. • Copenhagen, Denmark. Two interior bus islands serve northbound buses along Øster Farimagsgade at Sølvgade in central Copenhagen, allowing buses to avoid significant bicycle and right-turning traffic while continuing to provide a stop at the intersection. At the street’s northern end, a similar arrangement is used to create a near-side bus stop prior to the bus route turning left. Implementation Guidance General Considerations Conditions supportive of installing a boarding island on a right-turn channelizing island include those discussed in this section. • Suburban locations, which are more likely to have right-turn channelizing islands, due to greater potential right-of-way availability and higher-speed roadway design, compared to urban and downtown environments. • Sufficient space on the island to accommodate the ADA-required clear area at the bus stop, passenger waiting area, bus shelter (if warranted by passenger volumes), and waiting areas for pedestrians using the crosswalks leading off the island. • Passenger generator or transfer opportunities that suggest the need for a near-side stop. • Desire to provide a queue jump (Section 6.10), bus-only signal phase (Section 6.9), or other near-side transit-preferential treatment. • Ability to accommodate bicycle facilities that may be present on the street. • For right-turning traffic, ability to address potential sight-distance issues caused by bus shelters or stopped buses. • For a far-side channelizing island, space to provide a short bus lane and the ability to manage potential merging conflicts. Interior boarding islands are necessary supporting infrastructure when bus stops are desired to be provided along interior, left-side, or median bus lanes. They need to be wide enough to provide the ADA-required clear area at the bus stop and need to provide an accessible route connecting to a pedestrian crosswalk leading away from the island. With all types of boarding islands, consider the need to provide pedestrian fencing or similar barriers to control pedestrian movements, bollards to protect passenger waiting areas from errant vehicles, and the roadway agency’s requirements for marking, signing, and striping raised islands in the roadway. Bicycle Considerations Boarding islands can be designed to accommodate bicycle traffic. When the boarding island is also a right-turn channelization island, the first consideration is managing the conflict between

126 A Guidebook on Transit-Supportive Roadway Strategies bicycles and right-turning traffic, typically by transitioning bicycle traffic to the left of the right-turn lane. NACTO (2012) provides several potential concepts. The next consideration is managing the bicycle–bus conflict. Options include: • If the island is large enough, creating a channel through the island for the bicycle facility or raising the bicycle lane to the level of the island. In either case, the bicycle facility would separate the bus stop platform area from the remainder of the island. To minimize bicycle–pedestrian conflicts, the parallel crosswalk could be set back from the intersection by locating it to the right of the bicycle facility (as seen from the bicyclist point of view). • If sufficient space exists, create a short shared bus and bicycle lane wide enough to allow bicycles to pass stopped buses. • Continue an exclusive bicycle lane through the bus stop using a dotted lane marking to indicate that buses can cross into the bicycle lane to serve the stop. • If no bicycle facility exists, shared-lane markings (sharrows) could be used to guide bicyclists through the bus stop area. Additional Resources • Guide for Geometric Design of Transit Facilities on Highways and Streets (AASHTO 2014)— The guide discusses the potential for placing a bus stop on a right-turn channelization island if the island “is long and wide enough” (Section 5.1.1.2.1). The guide also suggests the pos- sibility of providing right-side island platforms for left-side bus lanes on two-way streets by shifting the bus lanes into the median at bus stops and splitting the stops between the two sides of an intersection to reduce the total width required (Section 5.5.5.1). • TCRP Web-Only Document 66: Improving Transportation Network Efficiency Through Imple- mentation of Transit-Supportive Roadway Strategies—Appendix A presents the results of a literature review on boarding islands. 7.7 Bus-Only Links Description Bus-only links (bus gates, bus-only crossings, bus sluices) are short sections of roadway connecting public streets that can only be used by transit vehicles and other authorized vehicles (e.g., emergency vehicles). Purpose Bus-only links are typically used to provide direct bus access to areas where general traffic is not desired. Bus-only links: • Provide bus, pedestrian, and bicycle access between neighbor- hoods with limited street connectivity by design; • Maintain bus access through a neighborhood after a neighbor- hood traffic management program is implemented; • Allow buses to make turns that are prohibited to general traffic (Section 6.1) due to cut-through traffic concerns or capacity constraints; • Prioritize non-automobile traffic on a street by using a short bus-only link to eliminate through traffic while maintaining local traffic access on either side of the link; and • Provide bus access to activity center areas (e.g., city centers, university campus areas) where private vehicles are restricted.

Infrastructure Strategy Toolbox 127 Applications Bus-only links can be enforced in several ways: • Signs and pavement markings only. Signs, or a combination of signs and pavement markings, prohibit general traffic but allow transit vehicles. This is a typical treatment for bus-only links allowing buses to make turns prohibited to general traffic. It has also been used for other bus-only link applications in communities where motorists generally respect traffic control devices. • Gates. A variety of gates and movable barriers have been used internationally to allow bus access while preventing access by private vehicles. The gates open when an authorized vehicle is detected (e.g., using a transponder or a transmitter). Examples include parking lot–style gates, swinging gates, rolling gates, and descending bollards. The gates are supplemented with appropriate signs and pavement markings. These devices physically restrict access into or between selected areas to buses only, but maintenance can be an issue. • Automobile traps. These are self-enforcing barriers that physically prevent automobile passage while permitting buses and other wider or higher vehicles (e.g., fire trucks) to pass through the link. Examples include pits in the roadway designed to trap automobile wheels and raised blocks that catch the undercarriage of an automobile. Similar types of barriers were used as part of early traffic-calming programs in some U.S. cities to allow fire truck access between closed street segments while preventing through automobile traffic (Smith and Appleyard 1980). Traps have generally fallen out of favor as a traffic-calming treatment in the United States but continue to be used internationally for bus-only links. • Photo enforcement. If local laws permit, photo or video enforcement is an option for enforcing bus-only links without resorting to gates or traps. Bus-only links that provide access between neighborhoods or into activity centers typically have provisions for pedestrian and bicycle access as well. Companion Strategies Bus-only links can be used to provide more-direct bus routes within suburban areas (Sec- tion 5.3). One application is to allow buses to make turns prohibited for other vehicles (Section 6.1). Red-colored pavement (Section 7.4) is an option for highlighting roadway sections open only to buses. Enforcement measures (Section 6.13) are suggested for links that do not use physical means to prevent access by unauthorized vehicles. Constraints Installing bus-only links in established neighborhoods may raise neighborhood concerns about unauthorized usage, new routes into the neighborhood for criminals, or the potential to open the link to general traffic at some point in the future. While the use of gates is permitted by the MUTCD, descending bollards are not discussed and would likely require an experimentation request to the FHWA if proposed for a roadway open to public travel, while pit trap treatments would likely raise liability issues in a U.S. context. Raised “undercarriage preventers” will likely be ineffective with pickup trucks, sport-utility vehicles, and other higher-slung vehicles, and may block police or fire chief cars (Smith and Appleyard 1980). Benefits Bus-only links support more-direct bus routings, which can reduce the time required for a bus to travel a route or can expand the area served by a bus route within a given cycle time. In suburban areas, they may allow minimal bus service to be provided to areas that could not

128 A Guidebook on Transit-Supportive Roadway Strategies otherwise support bus service due to the out-of-direction travel required. They can also support neighborhood traffic management programs by preserving bus access while eliminating routes for cut-through traffic. Other stakeholders that may be supportive of bus-only links are: • Emergency responders (e.g., fire, police, ambulance), as these links help reduce response times; law enforcement may also see them as supportive of police activities (e.g., ability to surround a block to catch a suspect in hiding) (Smith and Appleyard 1980); • School districts, which can plan more-direct school bus routes; • Neighborhood residents, who are provided with new options for recreational walking and bicycling routes on low-volume streets; and • Bicycle advocacy groups, if a proposed link offers an opportunity to expand a community’s bicycle network. Cost Considerations • Planning and coordination costs. Low (when installed by policy as part of new subdivisions or when incorporated into a larger traffic management project) to moderate (when proposed for an existing neighborhood as a stand-alone project). For gate applications, coordination will be required with other authorized users to make sure their vehicles are equipped with a means of opening the gate (e.g., a transmitter). • Capital costs. Low (signing and marking only) to moderate (other types of treatments). • Maintenance costs. Gate applications are subject to mechanical failures and vandalism. Due to the need to maintain bus service, it is important to have staff available to immediately respond to gate failures, and it may be necessary to leave the gate open if it cannot be repaired immediately. Calgary experiences approximately one stuck vehicle per month at one of its pit trap–type links. • Bus operations costs. Can significantly shorten bus routes in suburban areas, allowing the same bus route to serve adjacent neighborhoods. When used as part of a traffic management program, delays caused by other traffic may be reduced as a result of lower traffic volumes. • Other user costs. Typically no impact (since no link was provided previously). When used as part of a traffic management program, it is the program itself that creates impacts to other users; the bus-only link serves to preserve access for buses. Implementation Examples Calgary, Canada, uses bus-only links to provide transit bus, paratransit vehicle, school bus, and emergency vehicle access between adjacent subdivisions that have no public street connection. Early links used a pit-type trap, supplemented by several warning signs, to prevent through-vehicle traffic. Some of these links have subsequently been retrofitted with parallel gate-controlled access points for use if a vehicle becomes stuck in the trap. Newer installations only use gates. As of 2014, Calgary had 10 gates in active use by transit (three pit-only, two pit-and-gate, and five gate-only). Two other pit-and-gate systems have been installed for future bus use and are currently used by emergency vehicles. Three other gates are designated in local development plans for future construction if needed for transit use (Calgary Transit, no date). Ottawa, Canada, has installed bus-only left-turn lanes at key intersections where there is insufficient capacity to serve general left-turning traffic but it is desired to provide direct bus routings. Bus-only links are used to connect some neighborhoods that have limited street con- nectivity to allow bus routes to penetrate neighborhoods rather than go around them. These streets are controlled only by signs, but OC Transpo staff believe the violation rate is low.

Infrastructure Strategy Toolbox 129 Portland, Oregon, has installed bus-only left-turn lanes at a couple of locations. One site pro- vides access to the 5th Avenue transit mall; the other site provides a direct routing for buses at a complex intersection where there is insufficient capacity to directly serve automobile left turns. Other international examples include the Copenhagen, Denmark, region (at least 18 links); Sorø, Denmark; London, United Kingdom; and Delft and The Hague, Netherlands. Implementation Guidance The following considerations apply to bus-only links: • Bus-only links can be considered when there is a need to provide transit service with the most direct routing possible without encouraging additional motorized vehicle traffic. • Access should be provided to other public service users that would benefit from the link, such as emergency responders and school buses. • The design of the link should clearly indicate to motorists that it is not for use by general traffic. Signs, pavement markings, entrance design, placement, and passive and active enforcement measures contribute to communicating this message. • The entrances to a link are preferably placed at locations, such as intersections, where motorists can change their travel direction to avoid the link or can continue straight past the entrance. Midblock locations are more likely to experience violations as well as problems with vehicles blocking access to or from the link while making a three-point turn. • Links are preferably designed to accommodate pedestrians and bicycles except when connecting to facilities where pedestrians and bicyclists are prohibited. • Unless previous experience (for example, with neighborhood traffic calming) indicates a potential for violations, signing and marking could be adequate. However, the potential for enforcement should be planned for and integrated into the design if the need arises. • In the United States, parking lot–style gates and sliding gates are the most feasible options for physically restricting access to authorized vehicles. Ongoing maintenance needs should be considered during planning, and an operations plan should be developed for addressing situations when a gate will not open or is blocked. • Trap-type treatments, whether raised or lowered, are not suggested for U.S. applications due to a lack of support for them in U.S. design standards. Additional Resources • Manual on Uniform Traffic Control Devices (FHWA 2009)—Section 2B.68 addresses the use of gates on public roadways. • Guide for Geometric Design of Transit Facilities on Highways and Streets (AASHTO 2014)— Section 5.5.7.1 discusses bus-only links, while Section 5.5.7.3 discusses bus-only turn lanes. It recommends that “normal practice” should be used in the design of the link, and the signs should clearly indicate that the link is for authorized vehicles only. If there is a risk of a high violation rate, the guide suggests (1) additional and larger signs, (2) traffic signal control, (3) physically gating the roadway, or (4) photo or video enforcement. • TCRP Web-Only Document 66: Improving Transportation Network Efficiency Through Imple- mentation of Transit-Supportive Roadway Strategies—Appendix E presents the results of an international literature review on bus-only links.