Road User Understanding of Bicycle Signal Faces on Traffic Signals (2020)

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.

CHAPTER 3 State of the Practice and Inventory Design Guidance Before IA-16 was approved, Thompson et al. (2013) summarized the state of the practice for bicycle signals, reviewing design guidance current at that time: Guide for the Development of Bicycle Facilities (AASHTO, 2012), California Manual on Uniform Traffic Control Devices (MUTCD) (Caltrans, 2012), Urban Bikeway Design Guide (NACTO, 2011), Traffic Signal Guidelines for Bicycles (Transportation Association of Canada (TAC), 2004), Design Manual for Bicycle Traffic (CROW, 2007) and the Manual of Uniform Traffic Control Devices for Canada, 2008 update (TAC, 2008). The review found similar design guidance on lens size, use of the bicycle symbol, and the use of optical shielding. There was variance in mounting heights, housing color, and signal timing parameters. The research also collected detailed information about bicycle signal installations such as the presence and color of backplates, signal housing color, lens size, the presence of visibility-limiting lens or louvers, near-side or far-side placement, mounting height, phasing, restriction on vehicle movements, and supplemental signage. After the release of IA-16, which provided much needed guidance on the use of bicycle symbols in the signal face, NACTO updated the Urban Bikeway Design Guide (2014), Caltrans updated its MUTCD to incorporate IA-16 (2018), and Massachusetts DOT released the Separated Bike Lane Planning and Design Guide (2015). These documents substantially reflect the guidance in IA-16. IA-16, NACTO, and the MassDOT guide require that the bicycle signal faces be placed to maximize visibility to bicyclists, minimize visibility to users of other modes, and encourage the use of visibility-limited bicycle signal faces. It is important to note that even with IA-16, jurisdictions must still submit a written request to the FHWA and comply with all provisions to use the bicycle symbol in the signal face as described in Section 1A.10 of the MUTCD. IA-16 states that the bicycle signal face should be placed at least three feet away (horizontally or vertically) from the nearest motor vehicle traffic signal. The MassDOT design guide recommends that the bike signal face should be located at the far side of the intersection within five feet of the edge of the bike lane, mounted right of the bike lane, and that the bike signal faces not be placed between the vehicle signal faces (MassDOT, 2015).IA-16 requires at least one signal face be provided for controlling bicycle movements. The primary bicycle signal faces must be either 8 or 12-inches, even if they are placed on the near side of the intersection. A secondary signal head is required near side when the primary signal head is located more than 120 feet upstream of the stop line and is recommended when the primary signal head is located more than 80 feet beyond the stop line. A 4-inch supplemental near-side signal face may be used. The required dimensions of the lens are similar to the Canadian guidance (200 or 300 mm) and the United Kingdom’s (200 mm primary, 90-110 mm for supplemental) (TAC, 2004; Department of Transport, 2016). For a signal not over the roadway, IA-16 requires that the bottom of the signal face must be seven feet above the ground or sidewalk. The 4-inch post-mounted signal must more than four feet and less than eight feet above the ground or sidewalk. 26 

Except for the 4-inch signal heads, IA-16 suggests that the illumination and light distribution for the 8-inch and 12-inch signal heads should be similar to the motor vehicle signal heads. The interim approval permits the use of backplates without the presence of any legends. A supplemental sign (R10-10b, see Figure 10) noting the signal face for bicycles is required. IA-16 defines the meaning of the bicycle symbol indications as the same as vehicular indications. Turning after stopping is allowed with the RED BICYCLE indication, except for bicyclists positioned left of the adjacent motor vehicle traffic, who are prohibited from turning right on red, and bicyclists positioned to the right of the motor vehicle traffic on the same approach are prohibited from turning left on red. A YELLOW BICYCLE is used to indicate the change interval. A steady GREEN BICYCLE indication is displayed when traffic is allowed to proceed in any direction that is lawful and practical, provided that bicyclists are not in conflict with any simultaneous motor vehicle movements, including turning movements, and the bicycle movements are not modified based on turn prohibition signs, pavement markings, separate turn signal indications, and other control devices. Implied in the IA-16 definition is that a bicyclist interprets the GREEN BICYCLE indication as allowing any movement from the lane controlled by bicycle symbol (including a left turn from a bicycle lane positioned to the right of all traffic lanes or a right turn from a bicycle lane positioned to the left of all traffic lanes). In ruling 9(09)-47(I), FHWA clarified that the intent of IA-16 is to limit the use of a bicycle signal face to operations where the bicycle movement is “protected from any simultaneous motor vehicle movement at the signalized intersections (FHWA, 2014).” At many intersections, compliance with this provision requires the installation of fully protected left and right turns across the bicycle facility. Most often, separate turning lanes are also required. Note that NACTO also requires restricting right turns on red across the bicycle facility if the signal is used to separate through bicycles from right-turning traffic. IA-16 further requires that bicycle turning movements can only be prohibited with the use of arrows (i.e., movement prohibition signs are not sufficient). NACTO only states that the use of arrows should be considered. To use a bicycle symbol in the signal face that does not comply with the provisions of IA-16 requires that a jurisdiction submit and obtain approval through the “Request to Experiment (RTE)” process described in Section 1A.10 of the MUTCD (FHWA, 2009). All of the active Requests to Experiment involving IA-16 filed at the time of this project involve exceptions to the requirement of protection from any simultaneous motor vehicle movements. Evanston, IL, and Boston, MA, are experimenting with a GREEN BICYCLE allowing permissive right turns across the bicycle facility at multiple intersections in these locations. Minneapolis, MN, and Newark, DE, are experimenting with a FLASHING YELLOW BICYCLE to indicate a permissive bicycle movement. St. Paul, MN, is experimenting with both the FYA for vehicles and FLASHING YELLOW BICYCLE. The interim approval also describes the provisions for the layout of the bicycle symbol and the signal faces (see Figure 11). The approval requires that the bicycle symbol in the Standard Highway Signs (FHWA, 2004) be used for the bicycle signal indications and the symbol be positioned horizontally and face to the left. The bicycle signal faces themselves may be placed horizontally or vertically, in the same order as motor vehicle applications. The use of arrows in the bicycle signal faces is allowed in situations when it is necessary to prohibit certain turning movements by bicyclists due to conflicts with motor vehicles. While circular and bicycle signal indications are not allowed to be used on the same traffic signal face, arrow and bicycle signal indications can be used together. 27 

Source: MUTCD Figure 11. Typical Arrangements of Signal Sections in Bicycle Signal Faces Inventory of Intersections with Bicycle Signals To describe the state of the practice, the research team conducted an inventory of intersections with bicycle signals limited to those signals with the bicycle symbol in the face. The research team assembled the inventory of 511 intersections from a variety of sources in the following sequence: • An initial list of 411 intersections with bicycle signals in the U.S., dated January 2019, was obtained from the NCUTCD’s bicycle technical committee. The information had a varying level of detail for each intersection. • The PSU team’s prior bicycle signal inventory (Monsere et al., 2013) had 18 intersections that were not on the NCUTCD’s list. • All the current Requests to Experiment (RTE) with bicycle signal faces were obtained from FHWA. These applications describe 33 intersections (some of which were on the other lists). • An online survey to collect location information for additional bicycle signals was sent by email to the Association of Pedestrian and Bicycle Professionals (APBP) discussion group, TRB’s Traffic Signal Systems Committee and Bicycle Committee, and we posted a request on social media (Twitter) asking respondents to identify locations with bicycle signals not in our inventory. The Twitter post received 4,612 impressions and 196 engagements. The survey and Twitter post generated 142 new intersections (and removed 60 locations from the original NCUTCD list which were initially planned signals in New York but were never installed). As part of the cataloging process, each of the locations from all sources was mapped and signal face verified either in Streetview or by the jurisdiction in their survey response. Any intersections where the bicycle signal did not contain the bicycle symbol in the face were removed from the list (i.e., some signal faces were for controlling bicycle traffic but did not use symbols). The complete list of intersections is presented in Appendix A. 28 

Method for Data Collection The research team developed a list of data elements to be collected about each intersection, approach, and bicycle signal face. Initially, the research team proposed gathering information from each jurisdiction; however, a large percentage of the locations identified had Google Streetview images available. Thus, nearly all of the data was collected by visual inspection of Google Streetview. The data was collected by six research assistants at PSU and OSU following a brief training session. The training was conducted online and reviewed the document containing detailed instructions (Appendix B) and explained how to code the data in a shared spreadsheet. Most data were collected directly from the Streetview image. Visibility distance from the stop line was measured using Google Maps. Mounting heights, separation from other signal faces was measured using on open source imageJ software. The software estimates distances in images based on a known reference distance. In the pilot testing of the data collection, these measurements were compared with previously field measured data or design plans for three locations. Data were within one foot of known distances for field measured or plan documents. Finally, the research team identified a set of intersections for data collection by multiple coders to check for accuracy and agreement. The approach data was cleaned and validated following the data collection. Results For brevity, the results of the data collection are summarized for the primary bicycle signal face by intersection and approach in this report unless noted. Locations of Intersections Using Bicycle Symbol Signal Faces The locations of all 511 signals were mapped and verified. Table 1 shows a summary of locations and Figure 12 maps these locations. The states with the most installations are New York (156), California (70), Illinois (40), Washington (51), Oregon (33) and Texas (26). The large cities in those states are the primary locations for these installations including New York City, NY (154), Seattle, WA (51), Chicago, IL (32), Portland, OR (25), San Francisco, CA (24), Long Beach, CA (18), Los Angeles, CA (17) and Austin, TX (16). The cities of Denver, CO (14), Atlanta, GA (17), Lincoln, NE (10), and Boston, MA (12) also have a number of installations. Most other jurisdictions have a small number of installations. As shown on the map, with the exception of Atlanta, GA, Houston, TX, and Austin, TX, our inventory did not identify many locations in the southeast part of the country. 29 

Table 1. Intersections with Bicycle Signal Faces by State and Jurisdiction State City Intersections State City Intersections AZ Phoenix 1 MI Detroit 1 Tucson 1 MO Kansas City 1 CA Davis 2 St Louis 1 Long Beach 18 MT Missoula 1 Los Angeles 17 NC Charlotte 1 Mountainview 2 NE Lincoln 10 Palo Alto 1 NY Buffalo 1 Redondo Beach 3 Ithaca 1 Sacramento 1* New York City 154 San Diego 1 OH Cleveland 2 San Francisco 24 Columbus 7 San Jose 1 Xenia 1 CO Boulder 1 OR Ashland 1 Denver 14 Bend 1 Fort Collins 3 Clackamas Co. 1 DC Washington DC 8 Dundee 1 DE Newark 7* Eugene 2 FL Tampa 1 Portland 25 GA Atlanta 17 Salem 2 IA Des Moines 2 PA Philadelphia 2 IL Aurora 3 Pittsburgh 1 Chicago 32 SC Spartanburg 2* Evanston 5 TX Austin 16 IN Indianapolis 2 Houston 10 MA Arlington 1 UT Bluffdale 1 Boston 12 Salt Lake City 3 Cambridge 1 South Jordan 4 Lexington 1 VA Alexandria 3 Newton 1 WA Seattle 51 MN Minneapolis 7 WI Madison 7 St Paul 7 *planned installations 30 

Source: Google Fusion Maps, 2019 Figure 12. Map of Intersections with Bicycle Signal Faces 31 

Installation Year Installation year was either determined by data provided to the team via the survey, other documents, or by using historical views in Google Streetview to estimate the installation year. For some locations, only a range could be determined by this method (e.g., 2014-2017). Installation year was determined for 410 intersections and a range estimated for another 44 locations. Figure 13 shows the installation year for the 410 locations only (planned installations and ranges are not included). As apparent in the figure, the number of intersections with bicycle signals has been steadily increasing. The increasing trend aligns with the release of the NACTO Urban Bikeway Design Guide in 2011, which presented bicycle signals as tools, and IA-16’s release in December 2013. Figure 13. Installation Year of Bicycle Signals General Context Many signals are used in bicycle corridors and the same type of design is repeated. A total of 383 intersections (75%) in the sample were part of a bicycling route with the use of bicycle signals at many intersections along the corridor. Many of these corridors have separated bicycle lanes (two-way or one-way) or incorporate a multiuse path, which requires traffic control separation for bicycles for safe operation. At the intersection-level, the bicycle traffic was categorized as one-way, two-way or mixed (a one-way facility on one leg and a two-way facility on the other). In the inventory, signals on two- way bicycle facilities are 49% of the sample. Because of their two-way operation, these often require the use of separate signals for bicycles at locations with turning traffic. 32 

Motivation for Bicycle Signal For each intersection, the research team assigned a primary motivation for the bicycle signal. As many designs serve multiple purposes this was an iterative process. After first describing the motivation in detail, the team condensed coding to arrive at descriptions that could be grouped in categories. Table 2 describes the typical installations of bicycle signal faces. Note that 14 locations were unique applications that could not be grouped and are not presented in the table. Further, the research team could not code 29 locations due to missing Streetview images or documentation. Finally, the motivation for an installation could not be determined for an additional 13 intersections. The table is ordered by the most frequent category. 33 

Table 2. Summary of Typical Applications of Bicycle Signals Category Number Brief Description of Typical Location Photo of Typical Application Bicycle Lane to 129 This design used bicycle signals to separate through the Left of a Left- bicycle movements from left-turning vehicles. For Turn Lane these intersections, bicycle lanes (either one-way or two-way) are located on the left side of the road to the left of a left-turn lane. Long Beach, CA. Source: Google StreetView, 2019 Bicycle Lane to 41 This design used bicycle signals to separate through the Right of a bicycle movements from right-turning vehicles. For Right-Turn Lane these intersections, bicycle lanes (either one-way or two-way) are located on the right of a right-turn lane. Los Angeles, CA. Source: Google StreetView, 2019 34 

Category Number Brief Description of Typical Location Photo of Typical Application Two-Way Bicycle 69 In this configuration, bicycle signals were used for the Lane on One- contra-flow bicycle-traffic direction. Bicycle traffic in Way Street the same direction as motor vehicle traffic on the one- way street can be controlled by the vehicle signals, unless there is a need to control turning conflicts (these configurations were categorized in one of the turn-lane categories). Chicago, IL. Source: Google StreetView, 2019 Two-Way Bicycle 41 In this configuration, bicycle signals were used for Lane on Two- both traffic directions and contra-flow bicycle-traffic Way Street unless there is a need to control turning conflicts (these configurations were categorized in one of the turn-lane categories). St Paul, MN. Source: Google StreetView, 2019 35 

Category Number Brief Description of Typical Location Photo of Typical Application Multiuse Path 36 Where multiuse paths cross roadways, bicycle traffic Crossings signals are used to provide a better indication of the crossing time for bicycles (rather than using the pedestrian timing). Seattle, WA. Source: Google StreetView, 2019 Bicycle-Only 17 Bicycle signals were used to provide bicycle-only Connections to connections to specific facilities such as parks or Parks, Train median bicycle lanes. There are a variety of Stations, or configurations as these are somewhat unique designs. Center Bike The image shows a bicycle-only crossing to an Lanes intersection island which connects to the downtown transit mall. Denver, CO. Source: Google StreetView, 2019 36 

Category Number Brief Description of Typical Location Photo of Typical Application Connection to a 15 The connection to the start of two-way facilities is Two-Way Bicycle another typical application of bicycle signals. The Lane signals provide a means to transition from one-way operations to two-way. The sample image shows the connections to the start of the two-way separated bicycling lane in that allows contra-flow bicycles to depart from two-way bicycle lane and connect to a traditional one-way facility. Charlotte, NC. Source: Google StreetView, 2019 Contra-Flow 14 Another common use of the bicycle signal was to Bicycle Lane accommodate a contra-flow bicycle movement. In most cases, the bicycle signal face is only visible to the bicyclist. The sample image, however, shows an intersection where vehicles must turn left but cyclists are allowed to proceed through. Boulder, CO. Source: Google StreetView, 2019 37 

Category Number Brief Description of Typical Location Photo of Typical Application Diagonal 11 Bicycle signals are used to provide for a diagonal Crossing crossing of an intersection. All of the locations identified had two-way bicycle traffic at both or one side of the diagonal, often connecting a shared-use path. The image shows a diagonal crossing of a regional bicycle trail. Clackamas County, OR. Source: Google StreetView, 2019 Bicycle Left Turns 10 Bicycle signals were used to provide a separate bicycle left-turn phase from the right either in a jug- handle style pocket, from a two-stage turn box, or other waiting areas. The image shows the queuing area for cyclists to wait to complete a left-turn movement (bicycle signal is just visible on the far side of the intersection). San Francisco, CA. Source: Google StreetView, 2019 38 

Category Number Brief Description of Typical Location Photo of Typical Application Bicycle-Only 9 Bicycle signals were used to control bicycle-only Crossing for movements for a median two-way bicycle lane. The Median Two-Way intersections were in South Jordan, UT (shown in Bicycle Facility image), Portland, OR, and New York City, NY. South Jordan, UT. Source: Google StreetView, 2019 Connection to a 8 Bicycle signals were used to make bicycle-only Multiuse Path connections to multiuse paths from bicycle lanes. The image shows the connections to the start of a two-way separated bicycling lane near the Google campus that allows bicycles to connect to the Hetch Hetchy multiuse path. Mountainview, CA. Source: Google StreetView, 2019 39 

Category Number Brief Description of Typical Location Photo of Typical Application BL to the Right of 8 As part of an RTE with bicycle signal faces, the City of Shared Minneapolis installed bicycle signals to control bicycle Thru/Right traffic adjacent to a through and right-turn lane. The experiment involves a five-section head and a flashing yellow bicycle symbol displayed during the vehicle green. The image shows one of the intersections near the completion of construction. Minneapolis, MN. Source: Google StreetView, 2019 Bike Only Thru 8 Bicycle signals were used to provide crossings at Crossing with intersections where motor vehicle through movements Restricted Motor (or turns) are restricted but bicycle movement is Vehicle allowed. Some of these are at locations with a half- Movements signal treatment and appear on “bicycle boulevard” street crossings. An additional three intersections were identified where the bicycle signals were used in combination with a pedestrian hybrid beacon. Fort Collins, CO. Source: Google StreetView, 2019 40 

Type of Phasing Operation The type of phasing for the bicycle signal cannot generally be obtained from an image or photo, so there are fewer intersections where this information is available. The phasing information was obtained via the survey, RTE documentation, or the team’s knowledge of some locations. Table 3 summarizes the phasing information for 173 intersections with bicycle signals. While most signals operate with an exclusive movement for bicycles, there are some that operate concurrently. At the locations where phasing information is available, 67% of the bicycle signals are used for exclusive bicycle movements where the bicycles do not conflict with vehicle movements. Most of the RTE sites requested permission to experiment at locations with conflicting vehicular traffic by allowing the bicycle traffic to proceed concurrently with the vehicular traffic. In Minneapolis, a Flashing Yellow Bicycle symbol is used with a four-section head to indicate permissive movements while in Boston, MA, the green bicycle symbol is used. A number of installations prior to IA-16 feature concurrent phasing. Bicycle signals have also been used to provide leading intervals via leading bicycle interval (LBI) (6%) or a Split LBI (14%). In an LBI, all parallel traffic is held during the lead interval, following which the operation reverts to concurrent timing and turning vehicles must yield to bicycles. In the Split LBI, the through and bicycle traffic is allowed to proceed during the lead interval, while the turning traffic is held. After the lead interval, the operation reverts to concurrent timing, where the turning vehicles have to yield to bicycles. At the NYC locations with the Split LBI, an FYA is used for the turning vehicular movements during the permissive phase. Table 3. Number of Intersections by Phasing Type Phasing operation Number of for bicycle Percent Intersections movement Concurrent 22 13% Exclusive 116 67% LBI 11 6% Split LBI 24 14% Total 173 100% 41 

Number and Location of Bicycle and Vehicular Signal Heads per Approach Table 4 summarizes the number of bicycle signal faces on each approach that was sampled, cross-tabulated by the number of vehicle heads that were visible in the direction of bicycle travel. Overall, 221 of the 432 approaches (51%) presented two bicycle signal faces, typically in a far- side/near-side arrangement. A total of 204 approaches used a singular bicycle face. Finally, seven intersections were identified with three bicycle signal faces. As shown, 98 of the bicycle signal faces were presented for bicycle-only movements and no vehicle signal heads were visible. A total of 79 approaches had the minimum number of vehicle heads (two). The remainder (255) had three or more vehicle signal heads per approach. Table 4. Number of Approaches by Bicycle and Vehicular Signal Heads Number of vehicle signal heads per approach Number of bicycle signal heads per approach 0 2 3 4 5 6 Total 1 53 48 63 30 8 2 204 2 42 30 31 106 10 2 221 3 3 1 1 2 7 Total 98 79 95 138 18 4 432 Visibility Distance, Lens Diameter and Number of Bicycle Signal Heads Using the satellite maps and the Google measuring tool, the visibility distance from the stop line for bicycle traffic to the primary bicycle signal face was measured. For presentation in the table, distances were grouped by the IA-16 recommendations: <=80 feet, >80 feet and <=120 feet, and >120 feet. Appendix B pg B-5 shows an example of the measurement distance protocol. For each approach, the lens size of the bicycle signal was estimated when a sufficient quality Streetview image was present. While 4-inch lenses are allowed and known to be in use, these heads are difficult to spot in Streetview images and none were identified during the data collection. Table 5 presents the visibility distance category tabulated by lens size and the number of bicycle signal faces on the approach. Overall, the majority of bicycle signal lenses in the sample is 8 inches (77%). There is some jurisdictional consistency in lens size as design choices tend to be replicated. Note that IA-16 requires a supplemental signal face when the visibility distance is greater than 120 feet and suggests (“should”) for distances greater than 80 feet. As shown in Table 6, only five intersections have visibility distance more than 120 feet without a supplemental near- side head. A total of 108 approaches (33%) are in the recommended distance (>=80 but less than 120 feet). 42 

Table 5. Number of Approaches by Visibility Distance, Lens Size, and Bicycle Signal Heads Number of Number of Approaches by Visibility Distance Categories Lens bicycle diameter signal heads per approach <=80 >80 and <=120 >120 Total 1 19 36 4 59 12 inch 2 14 22 4 40 3 1 - - 1 12 inch Total 34 58 8 100 1 67 72 1 140 8 inch 2 80 89 8 177 3 6 - 6 8 inch Total 147 167 9 323 Total (12 inch 181 225 17 423 and 8 inch) Note: While 4-inch lens are allowed and known to be in use, these heads are difficult to identify in Streetview images and none were identified during the data collection. Placement of Bicycle Signal Head and Mounting Height The data collection process identified the placement of the primary bicycle face relative to the bicycle travel lane. The placement was coded as either over the roadway or over the sidewalk or path, and then whether it was to the left, center or right of the bicycle lane. As described in the data collection protocol, mounting height was estimated from the bottom of the bicycle signal face to the ground to/from the edge of backplate or housing, rounded to the nearest foot. These measurements were obtained from scaling the Google Streetview photos after first obtaining a measurement in the horizontal plane. Measurements were made only when a suitable view could be obtained. As shown in Table 6, the most common mounting location for the bicycle signal face is off the roadway, over a sidewalk or path. These signals were equally placed to the left or right of the bicycle lane driven by the context of each location. Only 60 bicycle signal faces were mounted over the roadway surface, primarily centered over the bicycle lane. Mounting heights that were observed are reasonable (the 25-foot maximum heights are observed for signals in Seattle on span wire support streetcar catenary). Use of Arrows used in Bicycle Signal Face The inventory only identified five locations (Palo Alto, CA, South Jordan, UT, St. Paul, MN, and San Francisco, CA) where arrows were used in conjunction with the bicycle symbol faces. Given the limitations of the data collection, arrows could only be confirmed if visible or assumed by the arrangement of faces. Any four- or five-section bicycle signal heads were considered to include arrow indications. Use Distinguishing Signal Housing or Backplate Color The use of a distinct color for the bicycle signal housing or backplate may help drivers distinguish it from a vehicle signal head. However, the majority of the primary bicycle signal faces 43 

are the same color as vehicular signal heads (84%). A total of 61 primary signal faces were identified that used a different color for the bicycle signal housing or backplate in Denver, CO, Long Beach, CA, Portland, OR, and Minneapolis, MN. Finally, slightly more than half (56%) of the primary bicycle signals selected did not use a backplate. Table 6. Number of Approaches by Placement of Bicycle Signal and Mounting Over Roadway Bicycle Signal Mounting Height (ft) Left, Center (OR) or Over Number of or Right of Sidewalk/Path Approaches Bicycle Lane (OS/P) Minimum Average Maximum Center 51 10 18 25 OR Left 4 14 19 23 Right 5 16 16 17 OR Total 60 10 18 25 Left 188 5 11 22 OS/P Right 184 6 11 19 OS/P Total 372 5 11 22 Total 432 Presence of R10-10b sign A total of 231 of the primary signal faces (53%) were identified with the R10-10b sign present. Sign dimensions were not confirmed though it was apparent that smaller dimensioned signs were in use at some locations. The placement of the R10-10b “Bicycle Signal” sign, required by IA-16, can be challenging at some intersections due to pole and space limitations. An additional 18 signal faces in Long Beach, CA, were accompanied by backplates with “BIKE SIGNAL” lettering on the signal housing and without the actual accompanying sign (an image shown in Table 2). Many of the signals in Denver incorporate the R10-10b text and legend in a horizontal arrangement in the backplate (also shown in Table 2). Presence of Louvers or Visibility-Restricting Device on Bicycle Signal Face Louvers or visibility-restricting devices are used on bicycle signals to prevent motorists from seeing the bicycle specific symbol indications from other lanes. The presence of these restrictions is difficult to detect in Google Streetview. Only six of the primary signal faces (1%) were identified in the sample as having louvers or visibility-restricting device. Horizontal and Vertical Separation to Nearest Vehicular Signal Face The horizontal and vertical distance between the far-side primary bicycle signal face and the nearest vehicular signal face were estimated from/to the edge of backplate or housing of each 44 

signal head, rounded to the nearest foot. If the signal heads were adjacent, a measurement of <1ft was entered. The horizontal offset was measured between the edge (either the signal housing or the backplate) of the bicycle signal face to the nearest motor vehicle signal face. The protocol for the vertical measurement is not specified in IA-16. For these data, the distance was measured from the top edge of the bicycle signal face to the bottom edge of the motor vehicle signal. These measurements were obtained from scaling the Google Streetview photos, after first obtaining a measurement in the horizontal plane. Measurements were made only when a suitable view could be obtained. The research team collected these measurements since the placement of the bicycle signal face in the vicinity of the motor vehicle signal may contribute to driver confusion. Further, IA-16 suggests that a bicycle signal face be separated vertically or horizontally from the nearest motor vehicle traffic signal face for the same approach by at least three feet. Table 7 shows the measured estimates of horizontal and vertical separation. A total of 53 (19%) of the bicycle signals measured in our sample had less than the recommended horizontal and vertical separation, respectively (separation < 3ft). To simplify the presentation, the measured offsets were grouped into three categories: <=3 feet, >3 feet and less <= 8 feet and > 8 feet. The detailed measurements of the horizontal and vertical offsets are plotted together in Figure 14. Table 7. Horizontal and Vertical Placement of Bicycle Signal Face from Nearest Motor Vehicle Face Horizontal Vertical Separation Category Separation Category <=3 ft. >3 and <=8 ft. >8 ft. Total <3 53 7 - 60 >3 and <=8 15 13 - 28 >8 77 98 7 182 Total 145 118 7 270 45 

Figure 14. Plot of Signal Face Mounting Offsets Summary Table 8 is a summary of the inventory. Ten of these locations are currently planned for installation in 2019-2020. The location (intersection street names and latitude/longitude) was coded for all intersections. Google Streetview images, showing the bicycle signal face, were available for 441 intersections (86% of the sample). Phasing for the bicycle movements was obtained for 173 intersections from survey responses, agency contacts, and research team’s knowledge database. In addition, the research team coded a primary motivation for the installation of the bicycle signal for 469 intersections (92% of the sample) to provide insight into the typical installations even if phasing information was not available. The installation year of the bicycle signal was coded for 80% of the sample. Detailed data for the signal face on an approach were collected for a subset of the sample (361 intersections). These observations included bicycle signal mounting heights and offsets from vehicular signals estimated from the Streetview images for 348 intersections (68% of the sample). 46 

Table 8. Summary of Inventory of Bicycle Symbol in the Signal Face Number of Percent of Total Data Item Intersections Sample Location (latitude/longitude) 511 100% Google Streetview Image 441 86% Motivation for Bicycle Signal 469 92% Bicycle Phasing 173 34% Installation Year (Exact) 410 80% Detailed Approach Data 361 69% Mounting Height and Offset from Motor Vehicle 348 68% Signals for Primary Bicycle Signal Face Key findings of the inventory are: • The states with the most installations of bicycle signal faces are New York (156), California (70), Illinois (40), Washington (51), Oregon (33) and Texas (26), with the large cities in these states being the primary adopters. • The trends show that bicycle signals have been increasingly used as a tool in the development of bicycling networks. The adoption of design guidance for separated bicycle facilities, and especially two-way facilities, and IA-16 have likely contributed to the trend in the U.S. • A total of 75% of the bicycle signals in the sample were part of a bicycling route with the use of bicycle signals at many intersections along the corridors. • Bicycle signals are key tools in the design of two-way bicycle facilities. Nearly 49% of the sample are bicycle signals on two-way bicycle facilities. • Bicycle signals are also used to facilitate safe bicycle movements when the bicycle lane is placed either to the left of a left-turn lane (31%), or right of a right-turn lane (10%). Other significant motivators include facilitating bicycle movements when a two-way bicycle lane is placed on either a one-way street (16%), or two-way street (10%), and crossings for multiuse paths (9%). • About 67% of the bicycle signals in the sample have exclusive bicycle movements, with the remainder operating concurrently with compatible movements. • The majority of the bicycle signal installations use two signal heads (51%) but single far- side heads are also common (47%). For the same approach, our sample indicates that either four (32%), three (22%), or no (23%) motor vehicle heads are commonly present. • The majority of the bicycle signal lens in the sample are 8 inches (77%) – though this is because so many signals are in New York City, which uses 8-inch lens. Excluding these, approximately 62% are 8 inches. • Visibility distances to the far side bicycle signal vary by intersection context. Generally, most designs include supplemental heads when visibility distances exceed 120 feet. Four-inch near-side heads were not observed but are difficult to see in our data collection protocol. Most of the bicycle signals are mounted off the roadway (over a sidewalk or path). • At this time, only a small number of bicycle signals have been installed with arrows for communicating allowed bicycle movements. 47 

• A majority of the primary bicycle signal faces are the same color as vehicular signals head (84%) and 56% of the primary bicycle signal installations did not use a backplate for the bicycle signal. • The R10-10b sign was observed at 53% of the primary bicycle signal installations. • Only six bicycle signal faces were found with visibility-restricting devices. This is a challenging data element to collect by our method. • Only 19% of the signal faces had less than the recommended three feet of horizontal and vertical separation distances in the interim approval from the nearest vehicular signal. There are some limitations to the data collection protocol. Data was collected for the primary approach that contained the bicycle signal face at all locations, where the Streetview image was available. In most locations, data was also collected for other approaches at the intersection that had bicycle signal faces. However, at some locations, it is possible that the data collectors may have missed collecting information on additional approaches due to limitations of the Streetview images. Finally, the accuracy of measuring mounting height, horizontal, and vertical separation using the imageJ software was validated against a small set of known distances, and the measurements obtained should be viewed as estimates. 48