Table 4.1. Methods to measure experience of the cyclist on rural highways.

• Urban street facilities.
• Off-street paths.

Bicycle and pedestrian operations, as counterparts to motor vehicle measures, have been included in all editions of the HCM. The concepts include delay, density, space, or capacity for bicycle and pedestrian modes. They have been expanded to bicycle flow parameters, capacity concepts, and delay. Since the HCM 2010, measures are included that consider how well bicycle facilities operate from a bicyclist’s perspective. These measures incorporate multiple factors that a roadway agency can influence, such as bicycle lane width, traffic volume, traffic speed, and pavement quality. The analysis is intended to reflect the perspective of the mainstream cyclist using a given facility.

BLOS was first proposed by Botma (1995) in the Dutch Manual on Bicycle Infrastructure. This measure of quality was developed adapting the HCM7 concept of LOS for bicycles. After considering mean speed, density, or percent of bicyclists following, the selected performance measure was hindrance, which used frequencies of events between bicyclists and other facility users to arrive at various LOSs.

Seminal works in the United States include the traveler perception model from Landis et al. (1997). They used 150 cyclists in actual urban traffic and roadway conditions, in a looped course of 30 road segments along collectors, arterials, and some local streets. Their model used the posted speed limit as a surrogate for actual motorized vehicle speeds and included the independent variables of trip generation intensity of the land use and frequency per mile of non-controlled vehicular access. A key finding was the importance of pavement surface conditions and striping of bicycle lanes for quality of service for bicyclists. In parallel, Allen et al. (1998) recommended the frequency of events (passings and meetings) for uninterrupted off-street bicycle facilities. Hindrance depended on bicycle traffic volume in the same and opposite directions. For on-street bicycle facilities, they suggest complementing the frequency of events with the impact of prevailing geometric and traffic conditions on the average and standard deviation of bicycle speeds.

Subsequent works included BLOS at arterials (Jensen 2007; Kang and Lee 2012), roundabouts (Jensen 2007), networks (Zolnik and Cromley 2007), and tools for identifying bicycle route suitability based on BLOS (Lowry et al. 2012; Callister and Lowry 2013).

This section provides a brief overview of the BLOS analysis methodologies in the HCM7. A more detailed overview of each methodology is contained in Part 3 of the Guide.

Multilane and Two-Lane Highways

As stated in the HCM7, “the calculation of BLOS on multilane and two-lane highways shares the same methodology, since multilane and two-lane highways operate in fundamentally the same manner for bicyclists and motorized vehicle drivers. Bicyclists travel much more slowly than the prevailing traffic flow and stay as far to the right as possible, including using paved shoulders when available. This similarity indicates the need for only one model” (TRB 2022). The model applies to bicycle operations in a shared lane, bicycle lane, or paved shoulder. The quality of bicycle flow, safety, and the bicycling environment are all considerations for these types of facilities. Complete details on the BLOS methodology and calculation procedures are provided in Chapter 15 of the HCM7.

BLOSs for multilane and two-lane highway segments are based on a BLOS score. The BLOS model uses a traveler perception index calibrated by using a linear regression model. The BLOS model fits independent variables associated with roadway characteristics to the results of a user survey

that rates the comfort of various bicycle facilities. The BLOS score is based, in order of importance, on five variables:

• Average effective width of the outside through lane.
• Motorized vehicle volumes.
• Motorized vehicle speeds.
• Heavy vehicle (truck) volumes.
• Pavement condition.

The resulting BLOS index computes a numerical LOS score, generally ranging from 0.5 to 6.5, which is stratified to produce a LOS A to F result by using Table 4.2.

Table 4.2. HCM7 BLOS score thresholds for multilane and two-lane highways.

Source: HCM7 Exhibit 12-31 and 15-7 (TRB 2022).

Required input data to calculate BLOS include lane and shoulder width, speed limit, number of directional through lanes, pavement condition (FHWA five-point scale), hourly motor vehicle demand, directional volume split, PHF, percent of heavy vehicles, and percent of segment with occupied on-highway parking.

Higher vehicle volumes, a greater proportion of trucks, higher vehicle speeds, and poor pavement quality, all act to decrease cyclist’s perceived comfort. LOS is highly sensitive to shoulder width and heavy vehicle percentage (modification of two LOS letters), and moderate sensible to lane width, pavement condition, and percent of segment with occupied on-highway parking (modification of one LOS letter).

The methodology was developed with data collected on urban and suburban streets (Petritsch et al. 2007) and is limited to the range of observed values, such as width of the outside through lane between 10 and 16 ft; shoulder width between 0 and 6 ft; motorized vehicle volumes up to 36,000 annual average daily traffic (AADT); posted speed limits from 25 to 50 mi/h; heavy vehicle percentage up to 2%; and pavement condition from 2 to 5 on the FHWA five-point pavement rating scale.

Urban Street Segments

The bicycle urban streets analysis methodology is contained in Chapter 18, Urban Street Segments. It relates to cyclists’ experiences at signalized intersections and the links between signalized intersections. The bicycle analysis methodology process is given in Exhibit 18-24. The methodology can be used to evaluate the service provided to bicyclists when sharing a lane with motorized vehicles or when traveling in an exclusive bicycle lane (TRB 2022).

LOS criteria are based on the BLOS index (HCM7 Exhibit 18-3), as summarized in Table 4.3.

Table 4.3. HCM7 BLOS score thresholds for urban segments (link-based).

Source: HCM7 Exhibit 18-3 (TRB 2022).

Required input data to calculate BLOS include bicycle lane width, paved outside shoulder width, striped parking lane width, median type, curb presence, number of access point approaches, pavement conditions, number of midsegment through lanes, midsegment motorized vehicle flow rate, heavy vehicle percentage, proportion of on-street parking occupied, pavement condition, and motorized vehicle midsegment running speed.

As for multilane and two-lane highways, higher vehicle volumes, a greater proportion of heavy vehicles, and higher speeds decrease perceived comfort. Furthermore, the presence of parking and pavement quality also decreases the LOS. LOS is highly sensitive to bicycle lane width and paved outside shoulder width (modification of two LOS letters) and moderately sensitive to outside through lane width and pavement condition (modification of one LOS letter).

Signalized Intersections

The bicycle signalized intersection analysis methodology is contained in Chapter 19, Signalized Intersections. The bicycle analysis methodology process is given in Exhibit 19-40 (TRB 2022). Intersection performance is evaluated separately for each intersection approach; the bicyclists

are assumed to travel in the street sharing a lane with motorized vehicles or when traveling in an exclusive bicycle lane and in the same direction as adjacent motorized vehicles. Table 4.4 shows the BLOS score thresholds for signalized intersections.

Table 4.4. HCM7 BLOS score thresholds for signalized intersections.

Source: HCM7 Exhibit 19-3 (TRB 2022).

The input data require traffic characteristics, such as the motorized vehicle demand flow rate, bicycle flow rate, and proportion of on-street parking occupied; geometric design data, such as number of lanes, street width, and width of cross-street elements; and signal control data, such as the cycle length, yellow change and red clearance, and duration of the signal phase that serves bicycles.

Higher vehicle volumes, reduced perceived separation from motorized vehicle traffic, shorter cross-street widths, and increased on-street parking reduce the LOS.

AWSC Intersections

There is no specific analysis procedure for bicycle mode. The procedures described to estimate motor vehicle delay can be applied to bicycles that queue with motor vehicles on AWSC approaches. Specific considerations could be applied because bicycles do not queue linearly at STOP signs, or the presence of a bicycle lane may reduce bicycle delay.

Roundabout Intersections

No methodology specific to bicyclists has been developed to assess the performance of bicyclists at roundabouts because limited data are available in the United States to support model calibration. A bicyclist may either navigate the roundabout in the same manner as a motorized vehicle or use sidewalks and crossings as would a pedestrian. If bicyclists are circulating in the same manner as motorized vehicles, their effect can be approximated by combining bicyclist flow rates with other vehicles by using a PCE factor of 0.5 for the bicycles. If bicyclists are navigating the roundabout in the same manner as pedestrians, their effect can be analyzed by using the methodology described previously for pedestrians.

Urban Street Facilities

The bicycle urban street facility analysis methodology is contained in HCM7 Chapter 16, Urban Street Facility. The bicycle analysis methodology process is given in Exhibit 16-13. The bicycle methodology aggregates the performance of the segments that make up the facility. In this regard, it considers the performance of each link and boundary intersection. The methodologies for evaluating the link and boundary (signalized) intersection are described in Chapters 18 and 19, respectively (TRB 2022). Table 4.5 summarizes the thresholds of the BLOS score.

Table 4.5. HCM7 BLOS score thresholds for urban segments (segment-based).

Source: HCM7 Exhibit 19-3 (TRB 2022).

The methodology can be used to evaluate the service provided to bicyclists when sharing a lane with motorized vehicles or when traveling in an exclusive bicycle lane.

The BLOS score for the segment is computed by using HCM7 Equation 18-46, which averages the BLOS score for the boundary intersection and the link and accounts for unsignalized conflicts along the link. BLOS for the link is weighted using the segment running time of through bicycles, while the BLOS score for the intersection uses the bicycle control delay (TRB 2022).

4.1.2 LTS Analysis Methodology

The LTS analysis methodology provides a set of criteria that indicates how stressful (or uncomfortable) a given facility is for cyclists. The LTS analysis methodology relies on simple rules based on data that are available for most agencies or are easy to obtain; it is gaining interest among practitioners and planners in the United States due to its simplicity.

The LTS methodology has its foundation on the design principles for the Dutch cycling network. The Dutch design guidelines propose infrastructure features so that the resulting cycling

infrastructure enables “direct, comfortable journeys by bicycle in a safe and attractive (traffic) environment”; as the main possibility to compete with the car and increase bicycle mode share (de Groot and Centrum voor Regelgeving en Onderzoek in de Grond 2007; de Groot and CROW kenniscentrum voor verkeer 2016). The target user is the mainstream population, which includes cyclists of all ages and purposes, and has the cyclist as a measure for the design. Comfort, in terms of minimizing stress, anxiety, and safety concerns for the target design user, is one of the five design principles for Dutch cycling networks, together with cohesion, safety, attractiveness, and directness. Following this concept, LTS-based design criteria are developed according to the maximum tolerance for traffic stress that a certain type of bicyclist may have.

The Dutch cycling network is already a cycle-friendly infrastructure that is part of the “cycle culture” of the country. As a result, it does not provide indications of which design features produce which level of stress to users. To fill this gap, the Mineta Transport Institute started to operationalize the LTS theory into design criteria for the U.S. conditions (Mekuria et al. 2012; Furth et al. 2016). More details can be found at https://peterfurth.sites.northeastern.edu/level-of-traffic-stress.

Design criteria are applied for a given set of cyclist types. Usually, four categories are considered, the lowest stress level being 1 and the highest stress level being 4. The lowest stress level (LTS1) is usually associated with child cyclists while the highest stress level (LTS4) is only acceptable for those who are highly tolerant of stress. The categories follow the classification proposed by Geller in 2006 to describe the overall cycling population of Portland, Oregon (Geller 2006). He distinguished four classes of cyclists: “no way, no how,” “interested but concerned,” “enthused and confident,” and “strong and fearless.” The distinction deviates from actual cycling behavior and is aimed at designing cycling facilities that attract potential cyclists to increase cycling ridership.

The LTS methodology operationalized by Furth et al. (2016) considers the following descriptions:

• LTS1: tolerable to child bicyclists. Little attention to traffic is demanded from cyclists and the facilities are attractive for relaxing bike rides. In mixed traffic, there is a low-speed differential with motor vehicles and demands only occasional interactions with them.
• LTS2: tolerable to “interested but concerned.” “Demands more attention than might be expected from children and presents little traffic stress.”
• LTS3: tolerable to “enthused and confident.” “Offers cyclists and exclusive cycling zone requiring little negotiation with motor traffic, but in close proximity to moderately-high speed traffic.”
• LTS4: tolerable to “strong and fearless.” “Requires riding near to high-speed traffic, or regularly negotiating with moderate-speed traffic or making dangerous crossings.”

LTS-based design criteria were developed for the FHWA Bikeway Selection Guide (Schultheiss et al. 2019), AASHTO Guide for the Development of Bicycle Facilities (Onta 2018), Minnesota DOT Bikeway Facility Design Manual (Minnesota DOT 2020), Massachusetts DOT Separated Bike Lane Planning and Design Guide (Massachusetts DOT 2015), Oregon DOT Analysis Procedures Manual (Oregon DOT 2020), Washington State DOT Roadway Bicycle Facilities (Washington State DOT 2021), Montgomery County Bicycle Planning Guidance (@MontgomeryBikeGuide), and the Maryland DOT LTS Technical Memo (@MaryDOTBikeGuide). These references are summarized in Table 4.6. Figure 4.1 illustrates the three user types that the Bikeway Selection Guide from FHWA considers “interested but concerned” (for LTS2), “somewhat confident” (for LTS3), and “highly confident” (for LTS4). They depend on comfort level, bicycling skill and experience, age, and trip purpose; but no specific thresholds are provided.

Table 4.6. Bicyclist design user references.

Furth et al. (2016) set criteria to classify urban streets by LTS, following the seminal work of Mekuria et al. (2012) at the Mineta Transportation Institute. Criteria for LTS2 are based on the Dutch design standards, as “the Dutch standards were proven on a population basis to be acceptable to the general traveler” (Mekuria et al. 2007). The Dutch “cycle culture” has achieved a high proportion of bicycle use by providing a good cycle infrastructure, with equal age and gender distributions of users. Criteria for less traffic stress (LTS1) were provided by increasing the separation to motorized traffic and easier crossing, while criteria for more traffic stress (LTS 3 or 4) presented less separation to motorized traffic and/or higher speed differentials. Specifically, LTS3 criteria were targeted to “enthused and confident” cyclists who would cycle on arterials, and LTS4 criteria were mapped to cyclists who feel comfortable riding in mixed traffic at 35 mi/h or more, or next to traffic at higher speeds. Interestingly, the Maryland DOT LTS Technical Memo includes LTS5, as roadways on which cycling is not suitable and therefore prohibited by the roadway agency.

The criteria required easily obtainable data, such as speed limit, street width, on-street parking, bike lane blockage, number of through lanes, or average daily traffic (ADT), which are available to most planners. Specific input data by facility will be summarized in the following sections.

LTS analysis methodology has been applied to evaluate bicycle network connectivity (Furth et al. 2016), detours (Furth et al. 2016), and propose bicycle facilities based on clearance, motorized vehicle speeds, and traffic volumes (Minnesota DOT 2020; Kittelson & Associates and Toole Design Group 2014; Massachusetts DOT 2015; Onta 2018; Schultheiss et al. 2019; Oregon DOT 2020; Washington State DOT 2021).

Multilane and Two-Lane Highways

The LTS analysis methodology has been applied to propose recommendation criteria for facility type in rural applications (Minnesota DOT 2020; Schultheiss et al. 2019; Oregon DOT 2020; Washington State DOT 2021). Segregation from main traffic and shoulder width are important considerations to accommodate rural bicyclists based on the expected traffic volume and traffic speed on the rural highway. For LTS rural applications, rural bicyclists are likely more stress-tolerant than their urban counterparts. Therefore, the design-based criteria are usually developed to accommodate either somewhat confident cyclists (LTS3) or highly confident (LTS4).

The FHWA Bikeway Selection Guide proposes shared lanes or paved shoulders to accommodate cyclists (see Figure 10 of that guide) (Schultheiss et al. 2019). The stress of cyclists depends on the following:

• Presence of shoulder and shoulder width.
• Operating speed (or posted speed limit).
• Traffic volume.
• Percentage of heavy vehicles.

Criteria are based on motorized vehicle speeds (25 to 60 mi/h) and traffic volumes (500 to 20,000 veh/day) (Figure 4.2). Shared lanes are preferred on highways with less than 1,000 vehicles per day. Paved shoulders from 4 to 10 ft wide are preferred as volume and posted speed limit increase. It is also noted that a wide shoulder or a separated pathway is preferred if the percentage of heavy vehicles is greater than 5%.

Some state DOTs developed a similar framework. In the state of Washington, bicycle facility needs should be accounted for within the design if the state highway is identified as a bike route,

intersects with an existing route, or if bicycle users are a modal priority in the route. It is indicated that bicycle users are usually accommodated on the shoulder and shoulder improvements include widening to a minimum of 4 ft and maintaining the shoulder usable, either by improving roadside maintenance or removing surface obstacles. If shoulder rumble strips are present, provide for at least 4 ft of usable shoulder between the rumble strip and the outside edge of the shoulder. If a guardrail or barrier is present, increase the dimension to 5 ft of usable shoulder. Different facility types are recommended depending on traffic volume and vehicle speeds for interested but concerned cyclists and for confident cyclists.

Minnesota DOT (2020) proposes paved shoulder width, shared lanes, shared-use paths and wide outside lanes based on traffic volume, number of lanes, and vehicle speeds for rural applications (see Table 4.2). Shared lanes are recommended for very low traffic volume (500 veh/day or less) and low speed (speed limit of 40 mi/h or less), while wide outside lanes are allowed for low traffic volume (1,000 veh/day or less) and speed limits below 40 mi/h. Highways with high-speed limits (45 mi/h or more) and high traffic volume (10,000 veh/day or more) should consider the use of shared-use paths. Paved shoulders are preferred in other combinations of traffic volume and vehicle speeds, with increasing width (4 to 8 ft) as volume and or speed increases.

In the case of Oregon DOT (2020), it is indicated that rural bicyclists may tolerate higher levels of stress than their urban counterparts and it uses a different nomenclature for rural (RLTS) and urban (BLTS) applications. The prefix (“R” or “B”) in the nomenclature highlights the environmental difference, as an urban BLTS 2 is not the same experience as a rural BLTS R2. For rural applications with posted or operating speeds less than 45 mi/h, urban application methods can be used. For highways with posted or operating speeds equal to or higher than 45 mi/h, specific criteria for RLTS R1 to RTLS R4 are provided. Criteria depend on the following:

• Paved shoulder width.
• Daily bi-directional (combined) volumes.

Rural Interstate highways are coded as BLTS R4 as posted speeds could be up to 70 mi/h and shoulder width equal to 10 ft.

Washington State initially considered setting LTS-based criteria for rural bicyclists. However, this approach was abandoned because the LTS methodology appeared less suited to rural contexts compared to population centers (Washington State DOT). The baseline assumption was that “the population center method would apply to users of all ages and abilities, and the rural high-speed method would apply to the subset of the population willing to travel along high-speed roadways between population centers.”

Furth et al. (2016) do not propose criteria for rural applications.

Urban Street Segments

The LTS analysis methodology has been applied to propose recommendation criteria for facility type in urban applications (Schultheiss et al. 2019). The target user in urban, urban core, suburban, and rural town contexts is the interested but concerned group. The level of stress of bicyclists in urban areas depends on the following:

• Facility type (shared lane, bike boulevard, bike lane with optional buffer, separated bike lane, shared-use path).
• Operating speed (or posted speed limit).
• Traffic volume.

Shared lanes or bike boulevards are preferred for low operating speeds (or posted speeds) and low traffic volume. As traffic volume and speed increase (up to 6,000 veh/day and 30 mi/h),

bike lanes with buffers are preferred. For other conditions, a separated bike lane or shared-use path is recommended.

Washington State DOT (2021), Minnesota DOT (2020), and Kittelson & Associates and Toole Design Group (2014) propose similar criteria for facility selection in urban areas. They are based on operating speed and traffic volume.

Furth et al. (2016) propose LTS segment criteria based on facility type: bike lane or mixed traffic. Oregon DOT developed the criteria to be applied in Oregon based on the seminal work of Mekuria et al. (2012). For bike lanes, LTS criteria depend on the following:

• Prevailing or posted speed.
• Number of lanes.
• Bike lane width.
• Presence of adjacent parking lane.
• Bike lane blockage.

For Oregon DOT (2020), it is noted that existing bike lanes that have less than 4 ft of useable width should be considered as mixed traffic. Further, poor pavement conditions in the bike-traveled portion of the roadway should have the BLTS increased by one level. Frequent drainage grates, requiring movement into the travel lanes may also increase the BLTS level. The Washington State DOT (2020) adapted the Oregon DOT (2020) LTS criteria to their local conditions. Network scoring tables depend on the following:

• Prevailing or posted speed.
• Number of lanes.
• Presence of bike lane.
• If bike lane is present, bike lane width.

The 20 mi/h speed limit threshold for roadways was included to account for roadways with unmarked centerlines. The speed limit threshold of 20 mi/h is intended to identify low-speed roadways that are consistent with LTS 1 conditions.

Furth et al. (2016) set LTS criteria for mixed traffic based on the following:

• Speed limit.
• Number of through lanes per direction.
• ADT.
• Presence of marked centerline.

Only roadways with no marked centerline, AADT lower or equal to 3,000 veh/day and speed limit up to 25 mi/h were classified as LTS1 (http://www.northeastern.edu/peter.furth/wp-content/uploads/2014/05/LTS-Tables-v2-June-1.pdf).

Oregon DOT (2020) uses the same variables and also distinguishes by functional class (arterial, collector, local). Local streets generally have LTS1 or LTS2, and arterials LTS3 or more. Only arterials with prevailing speeds equal to or lower than 20 mi/h could be suited for LTS2. It is indicated that BLTS should be increased by one level when poor pavement conditions in the bike-traveled portion of the roadway are present. Poor pavement conditions are defined as having potholes, large cracks, heavy raveling with loose aggregate, or spalling. Gravel and other debris are common in the shoulders and may limit their use.

Signalized Intersections

Furth et al. (2016), Oregon DOT (2020), and Washington State DOT (2020) indicate that signalized crossings are assumed to pose no traffic stress to cyclists because the signal provides a protected way across. Therefore, LTS1 is assigned.

Engineering judgment is required to assign higher stress levels at signalized intersections; as, at certain locations, the LTS can be increased. Examples provided by Oregon DOT (2020) include locations that have difficulties with triggering the signal detection, intersections that may not have the proper striping, locations that have push-button accommodations for bicyclists, and crossing locations where bicyclists use the crosswalk like pedestrians.

At intersection approaches, LTS could be increased with the following:

• Presence of right-turn lane(s).
• Length of right-turn lane.
• Turn lane configuration (bike lane shifts versus bike lane continues straight).

Washington State DOT (2020) does not propose LTS criteria for signalized intersections.

Unsignalized Intersections

Unsignalized intersection crossings can pose significant challenges for bicyclists, particularly in situations where there are multiple lanes or high speeds, and the intersection must be crossed in one phase. Furth et al. (2016) and Oregon DOT (2020) propose LTS for unsignalized intersections depending on the following:

• Width of cross street.
• Speed limit of cross street.
• Presence or absence of median refuge.

The Oregon DOT method for unsignalized intersection crossing without a median refuge was adapted to Washington State’s local conditions. The scores are generally higher, consistent with Washington State DOT policy. The LTS for unsignalized intersections in the Washington State DOT (2020) is based on the following:

• Speed limit of cross street.
• Total lanes crossed (in both directions).

It is worth noting that bicycle over/underpasses should be considered as separate facilities where the crossing is safe and accessible to cyclists and, therefore, do not generate stress to cyclists. LTS 1 is assigned to overpasses and underpasses.

Roundabout Intersections

Stress while crossing a roundabout differs if the crossing follows a path around the roundabout or is carried out in mixed traffic. For paths around a roundabout, Furth et al. (2016) and Oregon DOT (2020) indicate that the only traffic stress arises from crossings and depends on

• Non-tangential or tangential entry or exit lane.
• Type of entry/exit being crossed (single lane, dual lane).

If there is no adequate alternative path or sidewalk to use around the roundabout, bicyclists need to use the vehicle lane under mixed traffic conditions through the roundabout. Both Furth et al. (2016) and Oregon DOT (2020) criteria are based on the following:

• Number of circulating lanes.
• ADT (sum over all entry legs).

Washington State DOT (2020) does not propose LTS criteria for roundabouts.

4.1.3 Bicycle Compatibility Index Analysis Methodology

The Bicycle Compatibility Index (BCI) was developed from the perspectives of bicyclists and incorporates “the variables that bicyclists typically use to assess how compatible a roadway

segment is for bicycle travel” (Harkey 1998). It uses a six-point scale to evaluate comfort level ratings and BLOS criteria.

Initially developed for urban and suburban segments, Jones and Carlson (2003) adapted to rural conditions in Nebraska. An online survey was used to distribute video recordings of rural highways in Nebraska under different traffic conditions (vehicle and truck traffic volume) and geometric conditions (with and without shoulder). Three groups of bicyclists were defined: experienced commuter, experienced recreational, and casual recreational. The results indicate that the overall comfort level rating depended on two variables:

• Shoulder width.
• Flow rate of heavy vehicles.

Other variables, such as intersection density, observed 85th percentile speed, traffic volumes, or number of lanes, were not statistically significant.

Comfort level ratings were developed for all cyclists and by cyclist type. Experienced and casual riders provided different mean comfort level ratings (i.e., experienced riders were more comfortable than casual riders for the same conditions).

4.2 BLOS Research Recommendations

This section is intended to (1) provide recommendations for future research needs to improve/enhance the BLOS analysis methodologies for rural facilities in the HCM7; and (2) point readers to appropriate bicycle analysis method(s), HCM7 or other, for common analysis scenarios. This material is based on the results of the two surveys conducted as part of this project.

Two online surveys were conducted to identify potential revisions/improvements to the HCM7 LOS methodology and/or the AASHTO Green Book design elements to help agencies better evaluate the impact of traffic, roadway, and control characteristics on BLOS.

• The practitioner online survey was designed to solicit input from state DOTs to identify, based on expert opinions, which elements are more sensitive to rural bicyclists’ perception of shared and exclusive bicycle lanes and how rural bicycle design users can be classified if needed. Expert opinions from 52 individuals with experience in designing bicycle facilities and/or evaluating bicycling operational conditions were collected and analyzed (Moreno et al. 2023).
• The user online survey was designed to identify the drives and constraints on rural cycling, either for commuting or leisure, for different user types. The results suggested which variables are more relevant to users to cycle (or not) on rural highways. A total of 1,650 individuals entered the survey, and 1,032 complete responses were collected and analyzed.

In the first survey, practitioners were asked to (1) select which of the bicycle operations quality assessment methodologies they apply, (2) indicate their concerns regarding bicycle indicators and to what extent they make use of different data sources for bicycle analyses, (3) select how sensitive they consider that rural cyclist perceptions of quality of service would be to 52 different elements, and (4) rank different scenarios (geometric, traffic, control) according to their importance to assess the quality of service for the rural cyclist. Sociodemographic attributes, employment data, and cycling behavior were also collected.

In the second survey, users were asked to (1) select whether or not to cycle in different rural settings with six varying attributes, (2) rank different elements according to how important they were when riding their bike, (3) indicate how comfortable they would be cycling in given urban scenarios, and (4) select which statement best defined them as a cyclist. Sociodemographic attributes and cycling behavior were also collected. The elements identified by the practitioners’ survey were used to design the user survey to reduce the number of elements presented to users.

The results of the practitioner survey were used to design the second survey. Specifically, the rural scenarios presented to users were chosen based on the six most important elements to rural cyclists in assessing the quality of service that practitioners identified.

Refer to NCHRP Web-Only Document 392 for the complete description of the surveys, the analysis of results, and the main conclusions (Washburn et al. 2024).

The conclusions drawn from the surveys are limited to the designed questionnaires and the collected responses. The practitioner survey collected 52 expert opinions. The extrapolation of these results should be cautiously undertaken because the sample size was limited. The user survey targeted individuals who cycle on rural highways. Therefore, the extrapolation of these results to the overall cycling population or urban cyclists in specific geographical areas should be undertaken with caution.

4.2.1 Proposed Revisions to HCM7 LOS for Rural Highways

The main conclusions of the two qualitative assessments regarding the HCM7 analysis methodology are as follows:

• Bicyclists believed that the most important element for them to cycle on a rural highway was related to lateral clearance, paved shoulder width, or lane width. The presence of a shoulder had the highest impact on the odds of cycling in a given scenario. Furthermore, the strength of this effect was different by context class: rural areas without shoulders were less likely to be selected compared to rural town areas without shoulders. This result validates the practitioner survey’s conclusions, in which respondents indicated that bicyclists would have different sensitivities to context class in shared lanes than in paved shoulders. Responses also expressed that shoulders are rideable when clear of debris, gravel, or obstacles and that rumble strips make it challenging to navigate the shoulder when debris is present.
• The analyses of user responses revealed a high sensitivity to speed and automobile traffic volume. These results are consistent with the practitioner survey, previous studies, and design guidelines, which indicate that the two main criteria for selecting specific bicycle infrastructure and its width are ADT and posted speed limit. Cyclists could tolerate high speed limits if automobile traffic is meager, reinforcing the importance of a combined analysis of speed and traffic volume.
• Pavement conditions and context classification (or activity-volume-related elements) were equally important for users and experts. While bicyclists ranked them with lower relative importance than clearance, speed, or traffic volume, the choice experiments revealed a significant negative impact of deteriorated pavement and rural context on the cycling odds in a given scenario. While the differences between acceptable and good pavement conditions were substantial, the differences between rural town and suburban contexts were less pronounced compared to those between rural and rural town contexts.
• There was a division of opinions about the influence of terrain and environment-related variables. A significant portion of respondents considered the environment to be among the most crucial factors, while an equal number of users stated that it was the least significant factor. A consensus among respondents was the ranking of other cyclists as the least essential element. Terrain-related elements, such as elevation, grade, or sequence of curves, ranked low. However, scenarios with higher grades presented in the images yielded higher odds of being cycled on. Scenarios with high grades were more appealing to cyclists, suggesting that high grades could have been associated with more interesting scenery in rolling terrain and, therefore, less monotonous to cycle.
• For rural scenarios, “interested but concerned” cyclists were less likely to select to cycle than the other more confident groups. There were no apparent differences in preference between “enthused and confident” cyclists and “strong and fearless” cyclists.

Based on these conclusions, the proposed revisions to the HCM7 analysis procedure are as follows:

• The HCM7 bicycle analysis procedure should be revisited for rural highways because the perceptions of cyclists on rural highways differ from those on urban segments. BLOS scores by context classification (rural town, rural) and facility type (shared lanes, paved shoulder) should be distinct because their sensitivities to variables differ. For that, subjective comfort levels should be collected at least based on shoulder width, speed, automobile traffic volume, presence of heavy vehicles, pavement conditions, and context classification. Other aspects include maintenance of paved shoulders, grades, or intersections. The grade should be represented with caution because it could be misinterpreted by scenery quality. Furthermore, it is recommended to repeat rephrased questions to improve results fidelity.
• The HCM7 analysis procedure should account for data limitations. Sensitivity analysis of the input data could reduce the number of inputs to those most significant and, preferably, available to most agencies.
• The consideration of environmental factors on BLOS should be revisited because non-motorized users are more sensitive to weather and lighting conditions than their motorized counterparts. However, only a share of the most confident cyclists cycle under adverse environmental conditions, and weather data could be challenging to collect.
• Different LOS thresholds could be considered for different bicycle users if bicycle user classification for operational analyses becomes relevant. The most essential characteristics for bicyclist classification have been identified, but specific criteria and thresholds still need to be defined. Furthermore, the target user for rural highways still needs to be selected.

4.2.2 Proposed Revisions to LTS Analysis Methodology

The main conclusions of the two qualitative assessments regarding the LTS analysis methodology are as follows:

• The distribution of cyclist types for rural conditions differed substantially from the overall cycling population. The largest segment was for “enthused and confident” cyclists, compared to “interested but concerned” in urban applications. It follows the criteria from FHWA, which identified “confident” cyclists as the rural target users in rural applications. Many “interested but concerned” cyclists cycled on rural highways.
• Stated cyclist type classification (“no way, no how,” “interested but concerned,” “enthused and confident,” “strong and fearless”) aligned with declared comfort levels at different urban cycling environments. Compared to the “interested but concerned” group, “enthused and confident” cyclists showed increased levels of comfort, while “strong and fearless” cyclists declared the highest comfort levels. The most critical elements were facility type and traffic volume.
• For rural scenarios, “interested but concerned” cyclists were less likely to select to cycle than the other two groups; however, there were no apparent differences between “enthused and confident” cyclists and “strong and fearless” cyclists.
• Interactions between cyclist type and scenario elements were tested but were not statistically significant. It indicated that the effect of the elements did not depend on cyclist types but rather that more confident cyclists had, in general, a more significant likelihood of selecting to cycle in a given scenario. Only “strong and fearless” and “enthused and confident” cyclists were less challenged by terrain.
• All-year-round cyclists presented different sensitivities to which elements they considered the most important while riding, with relatively lower importance of clearance, pavement quality, environment, and terrain.

Based on these conclusions, the revisions to the LTS analysis methodology are as follows:

• The bicyclist classification criteria should differ for rural highways and urban environments. On the one hand, rural bicyclists may be more aware and are likely to tolerate higher stress levels than urban cyclists. On the other hand, rural bicyclists may not be used to cycling in urban low-speed environments, where traffic disruptions may be more frequent and, therefore, may generate higher stress levels. More research is needed to identify their main characteristics and how they relate to rural bicycling infrastructure.
• The target user for rural highways should be “enthused and confident.” For some locations where general cyclists are expected, the target user should be “interested but concerned” as a significant share of “interested but concerned” still cycled on rural highways.
• Further research could relate different bicycle user classifications to their comfort level in different environments. A combination of stated cyclist type and seasonal variation may better capture bicyclists’ attitudes and comfort levels on rural highways.
• The LTS analysis methodology should be revisited for rural highways. LTS criteria should account for clearance and traffic volume and include the effect of speed, pavement quality, and terrain. Sensitivity to elements was equal across all groups, and the more confident groups had, in general, a more significant likelihood of selecting to cycle in a given scenario. More research is required to validate LTS-based design criteria with user comfort perception.

4.3 Guidance on Current Methods for Rural Highways

4.3.1 Adaptation to Rural Conditions

The HCM7 methodology for multilane and two-lane highways was adopted from the methodology for arterials. Given the unique challenges faced by rural cyclists, such as limited infrastructure and higher speeds on rural roads, the HCM7 analysis methodology provides a first assessment of their safety and accessibility needs. However, it should be further adapted to rural conditions.

The effect of segregating bicycle and automobile flow may become more critical. Further, some elements may not apply to rural conditions (such as on-street parking), and rural corridors may be composed of several sections in rural and/or rural town contexts, where traffic operations are affected by the new functionalities of the road in rural town contexts. Additionally, the HCM7 method does not distinguish between bicyclists riding in bicycle lanes or sharing the road with motorized vehicles. For that, the current HCM7 analysis procedure should be used with caution.

The LTS analysis methodology distinguishes between bicycle lanes and mixed-use. While this methodology has proven useful in evaluating the stress levels experienced by cyclists in urban environments, its effectiveness in rural settings has yet to be fully explored. Only Oregon DOT has provided an adaptation of LTS-based criteria for rural settings based on simple input data available for most agencies.

Nevertheless, the LTS analysis methodology must still be validated with user perceptions. Notable exceptions were the work of Ferenchak et al. (2020), who indicated that the LTS score correlates well with biking allowance for children, and the work of Dill and McNeil (2016), who collected the percentage of respondents who were comfortable or very comfortable at different urban cycling infrastructure scenarios, with unclear trends by cyclist type. Further, how to assess cyclists’ comfort level in rural settings still needs to be determined.

Future research should also consider the overall suitability of a roadway for cycling. Not all facility types are suitable for cycling, and the managing roadway agency can prohibit bicycle access.

4.3.2 Data Requirements

Analysis methodologies should account for data limitations. Data availability was the highest concern for practitioners regarding bicycle indicators, followed by insufficient staff to estimate the indicators and expensive input data.

Complex traveler perception methods, such as the HCM7 or BCI, require specific data that may only be available for some agencies or companies. The HCM7 analysis procedure is highly sensitive to heavy vehicle volume and moderately sensitive to pavement quality and shoulder width. These two parameters may be complicated to obtain for large networks.

Specifically, the HCM7 method is the one that requires more input data, with 10 different attributes compared to the 2 attributes required in the LTS analysis methodology from Oregon DOT (paved shoulder width and daily bi-directional volume) or the 4 attributes in the HCM7 simplified analysis methodology from Oregon DOT (number of through traffic lanes per direction, bike lane or paved shoulder present, posted speed limit, and unsignalized conflicts).

Sensitivity analysis of the input data could be performed to reduce the number of inputs to those most significant and, preferably, readily available to most agencies.

4.3.3 Applications

The recommendation of using one method over another depends on the application. High-level applications, such as regional transport planning or transportation system plans, may not require a specific analysis of the facility but rather an overview of which users it would serve to identify critical locations where the bicycle network should be improved. For such applications, the LTS analysis procedure is preferred. The HCM7 analysis procedure should not be used, as data availability would be limited.

Simplicity, transparency, and easiness of application are the main advantages of the LTS analysis methodology. Nevertheless, the methodology must be more sensitive to infrastructure countermeasures (e.g., increased bike lane width or improved pavement quality) or traffic conditions (e.g., reduced percentage of heavy vehicles). Furthermore, how to define bicyclist types for rural conditions and who may be the target users need to be clarified.

For more specific applications with an increased project complexity and detail level, the HCM7 analysis procedure is preferred. This method is sensitive to more variables; therefore, a specific countermeasure’s impact is covered (e.g., increasing bicycle lane width). The LTS analysis methodology could also be used, although the results may be less sensitive to improvements beyond changes to facility type, automobile running speed, or automobile traffic volume.

This page intentionally left blank.