CHAPTER 1

Introduction

1.1 Background

The safety analysis and modeling techniques used to evaluate the safety of bikeways identified several factors that have either a positive or negative effect on bikeway safety; however, due to data limitations, the safety analysis did not include discrete design features or contextual factors that likely impacted safety outcomes. Therefore, the research team also conducted a visual assessment of separated bicycle lane segments where clusters of crashes occurred to better understand potential contributing factors. The visual assessment focused on separated bicycle lanes because this bikeway type has been shown to have superior safety performance over other bikeway types and shared travel lanes due to constraining motorist interactions with bicyclists to defined crossings at streets, alleys, and driveways. Separated bicycle lanes are of interest to the engineering and planning profession because they also have the most variation in design elements used by implementation agencies, and design practices continue to evolve. Design elements that could be considered in midblock bikeway design are then discussed. These design suggestions are based on general safety principles, previous studies, existing guidance, and the research team’s visual assessment of sites where crashes occurred on separated bicycle lanes.

The design of conventional bicycle lanes and buffered bicycle lanes at midblock locations is relatively standardized, with variations in design practices limited to bicycle lane width and presence, type of motorist parking/loading, and the application of supplemental traffic control treatments such as colored pavement or traffic signs. Existing bikeway design guidance presents only minor differences in bikeway width, typically within a range of 4 ft to 7 ft wide, exclusive of buffers and gutter pans [National Association of City Transportation Officials (NACTO) 2011; AASHTO 2023]. Given the relatively modest variability in these bikeway types, the focus of guidance in this report for conventional bicycle lanes and buffered bicycle lanes is on design strategies in high crash spots (conflict zones).

Separated bicycle lanes provide the highest degree of protection and comfort compared to other on-street bikeway types because they limit motorist encroachment to defined crossing locations with physical barriers. However, the design of separated bicycle lanes at midblock locations varies considerably more than conventional and buffered bicycle lanes due to factors such as the selection of vertical elements, the lateral placement and spacing of vertical elements, where and how parking and loading are managed on a street, and how mixing zones and transitions are treated. Each of these factors can impact the comfort and safety of bicyclists and have not all been well researched.

A variety of electrically powered micromobility and power-driven mobility devices have entered the market that may be operated on bikeways. Like bicyclists, many of these users are uncomfortable operating on roadways in shared lanes with motor vehicle traffic, especially in areas with higher motor vehicle volumes and operating speeds. For the purposes of this research, these users are assumed to be present and operating on bikeways.

Chapter 2 of these guidelines presents general safety principles and strategies applicable to midblock locations of all on-street bikeways. Chapter 3 and Chapter 4 present common factors contributing to bikeway safety found in the safety analysis and visual assessment and provide practitioners with guidance or strategies to increase bikeway safety. Chapter 5 of this report explores policy and planning strategies that may also mitigate negative outcomes at midblock locations of bikeways.

1.2 Visual Assessment of Separated Bicycle Lane Crash Sites

The safety analysis identified a range of contextual or built environmental factors that research has found to impact bicyclist safety. While the safety analysis and modeling techniques used to evaluate the safety of bikeways in this report did not analyze these factors directly due to limitations in available data, the research team performed a limited visual assessment of all crash locations on or near separated bicycle lanes in four cities using an online map tool’s roadside photography feature to identify the presence of those factors and evaluate whether they may have contributed to bicyclist crashes (Images taken by the research team were used in the report and thus are not subject to copyright). The relatively small set of crashes (98 total) that occurred on separated bicycle lanes within the data set presented a unique opportunity to perform a desktop visual assessment of each crash location. In many instances, the research team had direct and personal knowledge of the streets in question, which was obtained through past site visits or direct involvement in the design of the bikeway. While the team did not review crash locations of other bikeway types, likely, many factors that were identified in the visual scans are also applicable to other bikeway types.

The visual assessment identified contextual and design factors present across multiple crash sites, some of which are included in the literature review findings of the associated conduct of research report, NCHRP Web-Only Document 414, which can be found by searching the National Academies Press website (nap.nationalacademies.org) for NCHRP Web-Only Document 414: Safety Evaluation of On-Street Bicycle Facility Design Features. Additional contextual factors that were not identified in the literature review but have been noted in other crash research included the presence of roadways with downhill grades and proximity to pedestrian generators and alley access points. Suggestions for applying existing guidance, modifying existing guidance, or creating new guidance for design elements that may mitigate the presence of contextual factors or lead to improved design decisions are discussed in detail in Chapter 3.

1.2.1 Contextual Factors

The following contextual factors related to crash locations were consistently identified across the four cities reviewed, and are ranked from the highest number of crashes to the lowest number of crashes:

• Roadways with downhill grades (24 crashes; 23 in Seattle, two in Austin, none in Philadelphia or Minneapolis). Previous research in Seattle found that downhill approaches (greater than 4 percent) to intersections may increase the potential for left-hook bicyclist crashes [Seattle Department of Transportation (DOT) 2016]. The crashes within separated bicycle lanes analyzed in this study all occurred at driveway access points to commercial parking lots or garages. Design elements that warrant review include bikeway width (Section 4.1), sight distance (Section 4.3), and on-street parking restrictions (Section 4.3.1). Contextual elements that warrant review include roadway grades (Section 3.1). In some cases, a different bikeway selection may need to be made (Section 2.2).
• Commercial parking garage driveway (22 crashes; 19 in Seattle, two in Austin, one in Minneapolis, none noted in Philadelphia). Twenty crashes were located on streets where

• on-street parking is permitted in the vicinity of the garage entrance/exit. Design elements that warrant review include sight distance (Section 4.3), on-street parking restrictions (Section 4.3.1), driveway design (Section 4.5), and application of traffic control devices (Section 4.4). Contextual elements that warrant review include driveway volumes (Section 3.2) and frequency of access points (Section 3.4).
• Proximity to pedestrian generators (12 crashes involving pedestrians; eight in Austin, four noted in Seattle, none in Philadelphia or Minneapolis). The eight crashes involving pedestrians in Austin occurred at one location on Guadalupe Street where a bikeshare station is near a transit stop. This location is also at a main access point to the University of Texas at Austin where there are high volumes of pedestrians crossing between the University to access adjacent commercial destinations, bikeshare, and transit. The crashes in Seattle occurred at one location with on-street parking and a transit stop. Design elements that warrant review include bikeway width (Section 4.1), on-street parking restrictions (Section 4.3.1), lighting (Section 4.8), sight distance (Section 4.3), and application of traffic control devices (Section 4.4). Contextual elements that warrant review include pedestrian generators (Section 3.3). Additional design elements not addressed in this report include bikeshare siting and transit stop placement and design.
• Alley access points (seven crashes; five in Seattle, two in Minneapolis, none noted in Philadelphia or Austin). Except for one alley in Minneapolis that was in a residential single-family area with large setbacks, the remaining six crashes were in denser land use settings where multi-story buildings abut the sidewalks and alley edges squarely without sight-line setbacks. On-street parking is allowed at only two of the locations. Design elements that warrant review include driveway design (Section 4.5), sight distance (Section 4.3), on-street parking restrictions (Section 4.3.1), alley design considerations (Section 4.6), and application of traffic control devices (Section 4.4). Contextual elements that warrant review include frequency of access points (Section 3.4).
• Commercial surface parking driveway (seven crashes; none noted in Philadelphia). Two crashes were located on streets where on-street parking is permitted in the vicinity of the driveway to access the parking lot. Design elements that warrant review include sight distance (Section 4.3), on-street parking restrictions (Section 4.3.1), driveway design (Section 4.5), lighting (Section 4.8), and application of traffic control devices (Section 4.4). Contextual elements that warrant review include frequency of access points (Section 3.4).
• Gas station/convenience store driveway (six crashes, all one site in Austin; none noted in other cities). Design elements that warrant review include driveway design (Section 4.5), buffer design (Section 4.2), lighting (Section 4.8), and application of traffic control devices (Section 4.4). Contextual elements that warrant review include frequency of access points (Section 3.4).
• Residential single-family driveway (three crashes, all in Seattle; none noted in other cities). Design elements that warrant review include sight distance (Section 4.3), on-street parking restrictions (Section 4.3.1), lighting (Section 4.8), and application of traffic control devices (Section 4.4).

1.2.2 Design Factors

The following design factors were consistently identified across the four cities reviewed, and are ranked from the highest number of crashes to the lowest number of crashes:

• Significant gaps between vertical elements (56 crashes; none in Minneapolis). The factor observed across the highest number of crash sites in the visual assessment was the presence of significant gaps in vertical elements due to the frequency and width of access points such as driveways and alleys. These gaps resulted in large conflict zones, which increased a bicyclist’s exposure to motorists and, in some cases, effectively downgraded the bicycle facility from a separated bicycle lane to a buffered bicycle lane along significant portions of the bikeway. Design elements that warrant review include bikeway width (Section 4.1), buffer design (Section 4.2), sight distance (Section 4.3), lighting (Section 4.8), on-street parking restrictions

• (Section 4.3.1), and application of traffic control devices (Section 4.4). The inability to maintain physical separation within a separated bicycle lane may warrant selecting a different bikeway type as discussed in Section 2.2.
• Two-way bikeway on a one-way street (25 crashes; 13 in Seattle, 10 in Austin, two in Minneapolis, none noted in Philadelphia). Twelve crashes were located on streets where on-street parking is restricted. The remaining 13 were located on 2nd Avenue in Seattle at parking garage driveway locations. Design elements that warrant review include on-street parking restrictions (Section 4.3.1), commercial lot and garage driveway design (Section 4.5), and application of traffic control devices (Section 4.4)
• Narrow bikeway or buffer (17 crashes; 14 in Austin, three in Seattle, none in Minneapolis or Philadelphia). All locations were adjacent to on-street parking. A relatively narrow bikeway and/or buffer design in separated bicycle lanes were noted in these crashes, many of which involved pedestrians near transit stops or crossing a bicycle lane from a parking spot. Design elements that warrant review include bicycle lane width (Section 4.1), buffer design (Section 4.2), sight distance (Section 4.3), lighting (Section 4.8), on-street parking restrictions (Section 4.3.1), and application of traffic signs (Section 4.4.). The inability to provide adequate bikeway or buffer width may warrant selecting a different bikeway type as discussed in Section 2.2.
• Unmarked or unclear driveways (21 crashes; all cities). Nine crashes were located on streets where on-street parking is permitted. Design elements that warrant review include sight distance (Section 4.3), driveway design (Section 4.4), on-street parking restrictions (Section 4.3.1), mixing zones or undefined transitions (Section 4.7), lighting (Section 4.8), and application of traffic control devices (Section 4.4).
• Mixing zones or undefined transitions (16 crashes; all cities). These locations exhibited a pattern of having multiple driveways, large driveways, or an intersection near the crash location that reduced the ability of vertical elements to remain on the separated bicycle lane and added complexity to the transition of the separated bicycle lane to a conventional bicycle lane or shared lane. Six crashes were located on streets where on-street parking is permitted. Design elements that warrant review include sight distance (Section 4.3), on-street parking restrictions (Section 4.3.1), driveway design (Section 4.5), lighting (Section 4.8), mixing zones or undefined transitions (Section 4.7), and application of traffic control devices (Section 4.4).

The largest number of crashes in the visual scan were associated with sites where a combination of contextual and design factors was present at the crash location or were located within close proximity to the crash location. A sample of these locations is shown in Figure 1.