Investigating the Impacts of Lifestyle, Attitudes and Built Environment on Recreational Biking in a Coastal City
Roya Etminani-Ghasrodashti, University of Texas, ArlingtonShow Abstract
Mohammad Paydar, Pontifical Catholic University of Chile
Ali Ardeshiri, Amirkabir University of Technology
Recreational biking in a coastal city is explored through investigating individuals’ lifestyles, attitudes, and the effects of built environmental features on frequent biking. Data were collected from 426 residents of Kish Island, an Iranian city which has been provided with well-designed, dedicated biking lanes. Using confirmatory factor analysis, lifestyle patterns have been extracted from leisure activities which are mostly engaged by both tourists and residents. Additionally, attitudes were evaluated by exploring individuals’ preferences through selecting various biking facilities. Results from ordinal regression analysis revealed that residents with an active and beach-oriented lifestyle have a higher tendency toward recreational biking. Among other classifications of attitudes, landscape design made the highest contribution to recreational biking. Furthermore, findings demonstrated that network connectivity does not significantly influence recreational biking. Although findings from this study suggest that identifying lifestyle patterns and cyclist attitudes in coastal cities can enhance the probability of recreational biking among residents as well as tourists, they can also be of interest to planners and policy makers in terms of enhancing the frequency of biking in other regions with similar situations, namely, being a tourist destination and having a well-organized infrastructure in terms of biking paths and other required biking facilities that support recreational as well as other types of biking behavior.
User Preferences for Bicycle Infrastructure in Communities with Emerging Cycling Cultures
Calvin Clark, Georgia Institute of Technology (Georgia Tech)Show Abstract
Patricia Mokhtarian, Georgia Institute of Technology (Georgia Tech)
Giovanni Circella, University of California, Davis
Kari Watkins, Georgia Institute of Technology (Georgia Tech)
Non-motorized travel modes, particularly cycling, are experiencing a resurgence in many U.S. states as well as other countries. Still, most studies focus on bicyclists’ behaviors in areas with strong bicycling cultures. This paper discusses the findings of a survey (N=1,178) deployed in six communities in Alabama and Tennessee, where cycling is not (yet) popular nor widely adopted. The analysis includes linear regression models built on respondents’ reactions to images of bicycling infrastructure and their perceptions of being comfortable, safe, and willing to try cycling on the displayed roadway type. Findings indicate a preference for more separated bicycle infrastructure types along with options that exclude on-street parking. Segmented models indicate that, compared to potential cyclists, the preferences of regular utilitarian cyclists can vary more than those of recreational/occasional cyclists. Results from this study provide useful insights into ways to maximize the return on investments, and design bike infrastructure that can attract patronage and be most successful in areas lacking a substantial bicycling population.
Less Than Human?: The Dehumanization of Cyclists Predicts Self-Reported Aggressive Behavior Toward Them
Alexa Delbosc, Monash UniversityShow Abstract View Presentation
Farhana Naznin, Monash University
Nicholas Haslam, Melbourne School of Psychological Sciences
Narelle Haworth, Queensland University of Technology
Cycling provides many benefits to individuals and society, yet in many countries attitudes toward cyclists are largely negative. Public and humorous references to violence against cyclists is not uncommon and a significant minority of cyclists report harassment and aggression. We hypothesize that these hostile attitudes and behaviors are caused, in part, by the dehumanization of cyclists among some individuals. Dehumanization refers to any situation where people are seen or treated as if they are less than fully human. This paper presents a pilot study applying two validated dehumanization measures to a road user group for the first time. We found that the dehumanization measures were internally consistent, showed good discriminant validity (compared to general attitudes to cyclists) and were associated with self-reported aggression toward cyclists. The findings suggest that dehumanization is a concept that deserves further exploration in contexts where cyclists are a minority group. If we can put a human face to cyclists, we may improve attitudes and reduce aggression directed at on-road cyclists. This could result in a reduction in cyclist road trauma or an increase in public acceptance of cyclists as legitimate road users
Travel Mode Attitudes, Urban Context, and Demographics: Do They Interact Differently for Bicycle Commuting and Cycling for Other Purposes?
Jie Gao, Universiteit Utrecht Faculteit GeowetenschappenShow Abstract View Presentation
Dick Ettema, Universiteit Utrecht
Marco Helbich, Universiteit Utrecht Faculteit Geowetenschappen
Carlijn Kamphuis, Universiteit Utrecht
This study examined whether interactions between travel mode attitudes, urbanization level and socio-demographics were different for bicycle commuting and cycling for other purposes. Data were obtained from the 2014 wave of the Netherlands Mobility Panel (MPN), in total, 2,673 respondents who recorded at least some trips on the day surveyed. In total, four outcomes were constructed, two of them concerned commuting-related cycling (i.e., include trips to and from work and education): any commuting-related bicycle usage (yes vs. no), and average cycling duration (in hours per weekday). Likewise, two similar outcome variables concerning cycling for other proposes were constructed. These outcomes were analyzed by means of Tobit regression models (cycling duration) and binary logistic models (any bicycle usage). Attitudinal factors towards different travel modes, i.e. bus, car, cycling, and train were constructed by means of factor analysis. Regarding commuting-related cycling, the results showed that a positive attitude towards cycling was positively related with bicycle commuting duration, but this association was less strong among those with a positive attitude towards bus use. Having a positive cycling attitude had a weaker association with both commuting cycling usage and duration in those who not always have a car available. Regarding cycling for other purposes, cycling attitude had a stronger positive association with cycling duration among residents of very highly urbanized area, compared to residents of less urbanized areas. The available evidence, though limited, suggests that targeting attitudes can have a measurable impact on bicycling.
Understanding Factors Influencing Small to Mid-Size City Bikeshare Ridership: A Direct Ridership Analysis
Richard Mucci, Kentucky Transportation CabinetShow Abstract
Shraddha Sagar, University of Kentucky
Gregory Erhardt, University of Kentucky
Nick Stamatiadis, Kentucky Transportation Center
Bike share systems in large cities serve majority of bike share riders in the United States. This has led to much of the research on bike sharing focusing on ridership in large cities. A significant analysis on factors affecting bike share ridership in small to mid-size cities has been left out. This study quantifies the factors affecting bike sharing and develop a direct ridership model predicting the monthly station-level bike share ridership in small to mid-size cities. The study investigates the effect of demographic and employment characteristics near bike share stations for bike share ridership in Boulder, Colorado, Chattanooga, Tennessee, and Columbus, Ohio. A direct ridership model is used to estimate how station-level bike sharing ridership correlates with many factors, such as population, employment, income, education, built environment around a station, and weather. The models estimated using data in large cities typically find employment and population to be large influencers, which likely means bicyclists are using them mostly to commute to work. The model estimated using small to mid-size city data finds that bike share ridership is influenced the most by attractions being nearby, such as parks. The results suggest that bike share riders in smaller cities use the bikes for recreational purposes, rather than avoiding congestion along their commute. The developed direct ridership model is better suited to predict ridership, and identify the best location for a new bike share station in a small to mid-size city.
Maturing Urban Cycling: Comparing Barriers and Motivators to Bicycle in Lisbon, Portugal
Rosa Felix, Instituto Superior TecnicoShow Abstract View Presentation
Filipe Moura, FUNDEC
Kelly Clifton, Portland State University
Cities with low cycling maturity (LCM) are cities with a small cycling modal share and little cycling infrastructure. Despite the increasing public interest in cycling as travel mode, LCM cities are still common in the western world. This research explores the motivators and deterrents to bicycle in Lisbon (Portugal), a city with a cycling modal share below 1%, and compares the perceived barriers to cycling between cyclists and non-cyclists, as well as the triggers and motivators to start cycling within the same groups. Results from a survey, with different questions to cyclists and non-cyclists (n=1079), showed that both groups considered the issues related to the perception of safety, effort, the lack of a safe cycling network, and bicycle ownership as important barriers to cycling in Lisbon. We conclude that non-cyclists’ perceived barriers are similar to the barriers cyclists had before they changed behavior. In contrast, the self-reported triggers that induced cyclists to take up cycling are not similar to the expectations that non-cyclists have of what would, or could, change their behavior. Nevertheless, the non-cyclists’ barriers and motivators towards cycling are very related. We analyzed the triggers for cycling for different generations of cyclists, taking into consideration specific public policies and infrastructure investments that promote bicycling. We find that the triggers for cycling vary over time periods, and it is expected that they also change as cities transition along cycling maturity levels. This research and conclusions support the design of policies in order to increase cycling levels in LCM cities.
Cycling Across Canada: Understanding the Importance of Infrastructure, Climate, and Socioeconomics on Cycling Mode Change Over Time
Kevin Manaugh ANF10, McGill UniversityShow Abstract
Meghan Winters, Simon Fraser University
This study examined census data along with topographic, weather and built form variables to understand how cycling rates have changed over a 20 year period in Canadian metropolitan regions. We examined factors associated with increased cycling, as measured by journey -to-work data, for 2808 census tracts in 40 metropolitan areas. Overall, Canada has seen an increase in cycling mode share to work from 1.1% to 1.5%. At the level of cities, large metropolises such as Montreal, Toronto, Quebec, Vancouver, and Calgary have seen the largest increases in bike modal share. These cities have large amounts of dedicated cycling infrastructure. At the census tract level, we uncover more nuances related to population and socio-economic factors. Cycling is growing faster in areas with more dedicated cycling infrastructure, in areas that are seeing increases in median income and in areas with fewer children. However, these demographic factors are less important than street layout and cycling infrastructure variables in explaining the variance in cycling across census tracts. While some snowy and hilly areas do see increasing rates of cycling, overall areas with warmer winter temperatures and less variation in elevation across the region show higher rates of cycling. Further, 57% of the census tracts which were in the top decile of cycling in 1996 remain in the top decile in 2016, pointing toward a consistency in relative rates through time. This study offers insight into the importance of dedicated cycling infrastructure in order to increase cycling in Canadian urban regions.
Biking the First Mile: Exploring a Cyclist Typology and Potential for Cycling to Transit Stations by Suburban Commuters
Raktim Mitra, Ryerson UniversityShow Abstract
James Schofield, Ryerson University
Regional commuter rail has become an important means of travelling to urban employment centres across North America, but planners are faced with the challenge of connecting commuters from their origin or destination locations to a train station. Cycling may be an efficient and low-cost way of taking these transit access trips. However, cycling behaviour of rail commuters, particularly in a suburban context, remains understudied. This research examined perceptions of cycling and current cycling behaviour of 257 transit users from three suburban commuter rail stations in the Toronto region, Canada. Using a cluster analysis approach, four distinct cyclist types were identified, namely: recreational cyclists (29%), all-purpose cyclists (10%), safety-conscious occasional cyclists (33%), and facility-demanding occasional cyclists (28%). Differences between these groups included different mode choice motivations, tolerance for adverse weather conditions, comfort bicycling in various hypothetical traffic/infrastructure conditions, and current frequency of cycling for transportation and recreational purposes. The safety-conscious group had a higher percentage of women compared to other groups. Overall, 32.5% of regional transit users would be interested in cycling more often to rail stations. A higher proportion of recreational cyclists (compared to other groups) were “interested first-mile cyclists”, while the safety-conscious group had a significantly greater proportion of “uninterested” respondents. With careful planning of bicycle infrastructure and awareness campaigns targeting perceptions of cycling, there is much potential for cycling to accommodate a greater proportion of transit access trips in suburban communities, reducing demand for automobile parking at transit stations.
Bicycling in Walkable Environments: Does Wakability Promote Bicycling as Much as It Promotes Walking?
Or Caspi, Rutgers, The State University of New JerseyShow Abstract
Michael Smart, Rutgers, The State University of New Jersey
The walkability score has been developed to assess the suitability of built environments for walking. However, many planners use it in a more general manner to assess and promote ‘active transportation’. Previous studies found that environments can be suitable for pedestrians but uncomfortable for bicyclists. In this study, we assess whether walkable environments influence bicycling in the same manner as they influence walking. We have used data from the EPA and U.S. Bureau of Census to examine the walkability index’s influence on walking and bicycling commute rates, using a Zero-Inflated Negative-Binomial regression model. We have found that walkability influences walking and bicycling rates at about the same degree. However, the walkability index’s components influence the active transportation modes differently. Our findings suggest that overly dense and diverse streets and environments seem to be less comfortable for bicycling as they are for walking.
Cyclists’ Perceived Infrastructure Service Quality and Enjoyment: A SEM Approach
Jose Vallejo-Borda, Universidad de Los AndesShow Abstract
Daniel Rosas-Satizabal, IDOM CONSULTING ENGINEERING ARCHITECTURE SAU
Alvaro Rodriguez-Valencia, Universidad de Los Andes
The way that the service provided by cycling infrastructure is perceived by cyclists has important implications on system transportation planning. This paper seeks to better understand bicycle users’ perceptions of the elements integral to and surrounding the infrastructure and the way in which these perceptions relate to each other. A literature review focused on studies and guides that use a service-first perspective to measure infrastructure performance found that user perceptions are not frequently considered when trying to calculate or predict the way in which users perceive a service. For this reason, to the aim of this paper is to contribute to this debate by providing tools that highlight the importance of the sensory experiences of bicycle users. Through intercept surveys and the use of latent variable methodologies (CFA and SEM) a causal model was developed, making it possible to understand the way in which users perceive enjoyment when using their bicycle as well as their perception of the service provided, be it the infrastructure or other elements surrounding it. Additionally, the paper demonstrates the importance of building convenient cycling infrastructures in order to generate greater enjoyment and provide a better service for bicycle users.
Ride This Way?: The Impact of a Bicycle Skills Training Intervention on Bicycle Uptake
Stephanie Sersli, Simon Fraser UniversityShow Abstract
Nicholas Scott, Simon Fraser University
Meghan Winters, Simon Fraser University
Cities seek strategies to encourage more people to bicycle for active transport. One strategy that has shown some promise is bicycle skills training courses, but studies are few and rarely assess longer-term changes. We collaborated with a bicycle advocacy organization delivering adult bicycle courses in Metro Vancouver to assess the impact that courses have on bicycling uptake. This paper compares changes in overall cycling, as well as transportation (commuting, errands) and leisure bicycling over time (baseline-3 months) between course participants and a comparison group. We collected data in 2016 and 2017 through online questionnaires administered at baseline, 1, and 3 months post-course and used generalized linear mixed models to assess changes in bicycling between intervention and comparison groups. We enrolled 135 course and 43 comparison participants. Course participants increased overall bicycling at 1 month but this was not sustained. By contrast, comparison participants’ bicycling decreased. Adjusted models reflected different trajectories for course and comparison participants for commuting (RR= 1.22, 95% CI: 1.00, 1.48) and leisure bicycling (RR= 1.22, 95% CI: 1.02, 1.47). Men had higher rates of using bicycles for commuting and leisure, and participants living in the city of Vancouver had higher rates for commuting. Bicycle courses address individual-level barriers to bicycling, such as skills, knowledge, and confidence. Our results suggest courses can encourage modest increases in bicycling, but sustained increases could be difficult. Bicycle courses should be combined with environmental and other means of support to achieve greater impact on bicycling.
Rules of the Road: Compliance and Defiance Among the Different Types of Cyclists
Nick Chaloux, McGill UniversityShow Abstract View Presentation
Ahmed El-Geneidy, McGill University
While cycling has become a more attractive option to commuters in many North American cities recently, significant apprehension remains around its safety. While risks experienced by cyclists are diverse, the idea that they are due to scofflaw cyclists – cyclists who regularly ignore the rules of the road – remains prevalent. Improving cycling safety requires countering this idea, and therefore an understanding of how different cyclists act under the existing rules. Using a survey of 1,329 cyclists in Montreal, Canada, this study generates a typology of cyclists based on cycling motivations and behaviors and conducts comparisons based on their responses to four cycling rule-breaking scenarios. Our study shows that all cyclist types contravene traffic laws in similar ways, and 0.6% of respondents consistently follow traffic laws. Breaking the law is often considered the safest option by respondents, which reflects a disconnect between the safety goals of traffic laws and the reality on the streets based on the perspectives of different cyclist types. While cyclist types may act similarly in response to existing laws, they still respond uniquely to different policies aimed at increasing rule adherence. Targeted interventions aimed at educating young cyclists, improving dedicated infrastructure, and prioritising cycling traffic can increase rule compliance across all types. Through our study planners, policy makers, and law enforcement can improve cycling safety by better understanding the behaviour and rationale taken by cyclists.
Measuring Bicycle Ridership Seasonality Relative to Associated Factors
Nicholas Fournier, University of Massachusetts, AmherstShow Abstract
Eleni Christofa AHB25, University of Massachusetts, Amherst
Mark Hamin, University of Massachusetts, Amherst
Harsh winters experienced in North American present a formidable barrier to promoting sustained year-round bicycle ridership. To mitigate this, many cities and agencies are implementing a growing number of creative methods to increase bicycle ridership, such as educational campaigns, year-round bike-shares, innovative infrastructure and facilities, and economic policies. However, quantitatively measuring the impact of these efforts is difficult without normalizing the ``seasonality'' of bicycle ridership. The lack of such quantitative measures is often an obstacle for efforts geared towards increasing bicycle ridership, making justification for such efforts difficult. This paper proposes a framework for measuring seasonality with a model for demand using parametric sinusoidal function to provide a normalized seasonality factor for demand, called alpha. This model is fitted through least-squares optimization, enabling the interpolation of missing data points (e.g. winter closure of bike-share stations). The proposed model and framework is applied to the Bluebike bike-share system in Boston, Cambridge, Brookline, and Somerville, Massachusetts, analyzing a variety of factors as infrastructure, land use, demographics, and employment statistics. The outcome of this research is a basic framework that can be used for analyzing bicycle seasonality to provide insights for planning, engineering, and decision-making.
Exploring Spatial Variation of the Bikesharing Ridership: A Study Based on Semi-Parametric Geographically Weighted Regression
Li Pu, Southwest Jiaotong UniversityShow Abstract
Xiaojia Zhang, National United Engineering Laboratory of Integrated and Intelligent Transportation
Ziwen Ling, University of Tennessee, Knoxville
Hongtai Yang, Southwest Jiaotong University
As an important part of urban public transport system, bike sharing systems are adopted by many cities due to its contribution to energy saving and mitigation of traffic congestion. Understanding factors that influence bike sharing ridership and accurate estimation of ridership at different locations play an important role in determining location of stations and could provide reference for making policies to increase bike sharing ridership. This study divides the ridership into three types: trip production of members, trip attraction of members, and trips of 24-hour pass users, and explored factors that influence the three types of ridership. Previous studies assume the relationship between predicting variables and response variables are the same across the study area. We test this assumption by employing semi-parametric geographically weighted regression (S-GWR) model to fit the data and found that the relationship between some predicting variables and response variable are local while other relationships are global. Results show that S-GWR models have better goodness-of-fit than OLS models and can eliminate the autocorrelation in the residuals, which is present in the OLS models. As a result, spatially varying relationship between ridership and influencing factors should be considered when designing bike sharing system.
Bikeshare Users on a Budget: A Trip Chaining Analysis of Bikeshare User Groups in Chicago
Siyue Yang, City University of Hong KongShow Abstract
Candace E. Brakewood, University of Tennessee, Knoxville
Virgile Nicolas, Cerema - Département Laboratoire de Saint-Brieuc
Jake Sion, Veolia Transportation Services, Inc.
This analysis focuses on a smartphone app known as “Transit” that is used to unlock shared bicycles in Chicago. Data from the app were utilized in a three-part analysis. First, Transit app bikeshare usage patterns are compared to system-wide bikeshare utilization using publicly available data. The results reveal that hourly usage on weekdays generally follows classical peaked commuting patterns; however, daily usage reached its highest level on weekends. This suggests that there may be large numbers of both commuting and recreational bikeshare users. The second part aims to identify distinct user groups via cluster analysis; the results reveal six different clusters: (1) commuters; (2) utility users; (3) leisure users; (4) infrequent commuters, (5) weekday visitors; and (6) weekend visitors. The group unlocking the most shared bikes (45.58% of all Transit app unlocks) was commuters, who represent 10% of Transit app bikeshare users. The third part proposes a trip chaining algorithm to identify “trip chaining bikers.” This term refers to bikeshare users who return a shared bicycle and immediately check out another to avoid paying extra usage fees for trips over 30 minutes. The algorithm reveals that 27.3% of Transit app bikeshare users exhibit this type of “bike chaining” behavior, presumably to avoid paying additional usage fees. However, this varies substantially between user groups; notably, 66% of Transit app bikeshare users identified as commuters made one or more bike chaining unlocks. The implications are important for bikeshare providers to understand the impact of pricing policies, particularly to encourage turn-over of bicycles.
Predicting Station Locations in Bikesharing Systems Using a Proposed Quality-of-Service Measurement: Methodology and Case Study
Huthaifa Ashqar, Booz Allen Hamilton, Inc.Show Abstract View Presentation
Mohammed Elhenawy, Virginia Polytechnic Institute and State University
Hesham Rakha, Virginia Polytechnic Institute and State University
Leanna House, Virginia Polytechnic Institute and State University
Bike-sharing systems (BSSs) operators tend to spend a great amount of time and effort to satisfy users. Accurately measuring the quality-of-service (QoS) of each station in a BSS will advance this mission. Moreover, measuring the QoS and using it to study the spatial dependencies in a BSS allows operators to better manage the system. The traditionally-known QoS measurement reported in the literature is based on the proportion of problematic stations, which are defined as those with no bikes or docks available to users. We investigated the traditionally-known QoS measurement, and it was found neither exposes the spatial dependencies between stations nor does it discriminate between stations in a BSS. This study proposes a novel QoS measurement, namely Optimal Occupancy that captures the impact of heterogeneity of bike-sharing systems (BSSs) and reflect the spatial dependencies between the stations. Optimal Occupancy is defined as the ratio of the total time a station is functional during a given interval to the length of the interval, in which it also redefines problematic stations. We applied geo-statistics to explore the spatial configuration of Optimal Occupancy variations and model variograms for spatial prediction. Results revealed that the Optimal Occupancy is beneficial for operators, would result in better prediction of the QoS at nearby locations, and can be used to predict candidate spots for new stations in an existing BSS. For example, the proposed QoS for Station 50 was improved after adding a new nearby station, increasing QoS from 0.52 to 0.84 for Monday of July.
Portraying and Differentiating Profiles and Preferences of Casual Users and Registered Members of Capital Bikeshare
Shruthi Kaviti, George Mason UniversityShow Abstract
Mohan Venigalla, George Mason University
Kimberly Lucas, District Department of Transportation
(Previous submission of the same paper didn't include word count.) Even though casual users of bikeshare account for a large share of ridership and revenue at public bikeshare systems in North America, very little is known about the characteristics and preferences of casual users and how they compare to registered members. The primary objectives of the study include identifying the similarities and differences between registered members and casual users by such characteristics as demographics, usage, and indicated preferences; and examining and modeling pricing preferences of bikeshare users. An intercept survey was conducted to obtain demographic information, bikeshare usage and various preferences of Capital Bikeshare (CaBi) users in the metro Washington DC area. The survey data was validated against the data from an existing member survey with large sample using goodness of fit tests. Survey participants reported that the single-trip fare (STF) and annual membership paid at once are their preferred pricing options and a combination of STF, 24-hour pass, and annual membership with monthly installments as their favorable pricing model. Two logistic regression models, namely, a formative model for user type as Model 1 and casual user fare choice model as Model 2 were developed. Results indicated that, when compared to casual users, registered members are more likely to be White, have higher income and reside in the metro DC area. Casual users make fewer bikeshare trips and are less sensitive to the service (as measured by station density) compared to members. Logistic regression results among the casual users demonstrated that STF purchasers are less likely to be white and more likely to be the metro DC area residents compared to the 24-hour pass users. Gender, age, and income distribution do not appear to influence casual fare product choice. Results from this study are useful in policy-making, planning, and operations for bikeshare systems.
How to Make Dockless Bikeshare Good for Cities: Curbing Oversupplied Bikes
Yuanjie Tu, University of WashingtonShow Abstract
Peng Chen, Tongji University
Xu Gao, University of California, Irvine
Jiawen Yang, Peking University
After experiencing many years of shrinking in bike mode share, the birth of dockless bike has incurred a large-scale bike renaissance since 2016. However, the quick expansion of dockless bikeshare is coincident with the serious oversupply of bikes and chaos of parking on the streets. Predicting the right level of dockless bike use is essential to maintain the order of the road space. This study aims to control the number of dockless bikes in each neighborhood. With data obtained from a smartphone app, MoBike, this study examines factors associated with dockless bikeshare. A generalized additive mixed model is applied to investigate associations between dockless bike count density and various factors. The results are: (1) floor area ratio, which represents density, is positively associated with dockless bike count density; (2) percentages of residential, industrial and green spaces, and degrees of mixed land use, are positively related to dockless bike count density; (3) densities of primary and secondary roads are positively related to dockless bike count density, while the density of intersections is negatively associated with dockless bike count density; (4) females and children are less likely to ride dockless bikes; (5) dockless bikes are more often used during peak hours, on sunny or cloudy days, and on weekdays. To promote dockless bike use, the right level supply for different weather, time and location conditions should be identified, encouraging dense urban development in establishing a friendly biking environment, and improving street connectivity to reduce the impedance of biking in intersection dense areas.
Effects of Docked Bikesharing Fleets, Urban Land Uses, and Socio-Demographic Characteristics on Dockless Bikesharing User Demand
Xinwei Ma, Southeast UniversityShow Abstract View Presentation
Yanjie Ji, Southeast University
Yufei Yuan, Delft University of Technology
Niels van Oort, Delft University of Technology
Serge Hoogendoorn, Delft University of Technology
The emergence of dockless bikes has revolutionized the traditional docked bike-sharing market and led to changes in people's travel behavior. To better understand the usage of dockless bike-sharing, this study aims at examining the influence of the docked bike-sharing fleets, socio-demographic factors and land use factors on user demand of dockless bike-sharing. To this end, the geographically and temporally weighted regression (GTWR) model was adopted as its ability to simultaneously incorporate spatial and temporal heterogeneity into dockless bike-sharing data analysis, which outperforms the ordinary least squares (OLS) regression or geographically weighted regression (GWR) significantly in terms of model fit. Then, the spatial and temporal variations of estimated coefficients were visualized and analyzed. Results show that docked bike-sharing fleets promote dockless bike-share usage, especially during rush hours and in the suburban areas. Places with high density of road and bus/metro/docked bike-sharing stations impede dockless bike-sharing usage, especially during rush hours on weekdays. Entertainment POI (Point of Interest) promote the usage in the urban areas during afternoon periods at weekends. Car ownership yet significantly impedes dockless bike-sharing use during morning rush hours. Also, car owners and the elderly groups prefer to use dockless bike-sharing in the areas where bike infrastructure is well-developed. This study can offer meaningful implications for policy makers and dockless bike-sharing companies to improve dockless bike-sharing systems.
Exploring the Social, Spatial, and Temporal Performance of Bikesharing: A Case Study of DIVVY
Christopher Smith, DePaul UniversityShow Abstract
MD Mehedi Hasan, Western Michigan University
Numerous studies over the past two decades have found clear evidence that vibrant communities are inextricably linked with opportunities for active and/or non-motorized transportation. A synergetic force working within the broader movement of active transportation is the emergence, widespread diffusion and expansion of public bicycle sharing systems (BSS). Such systems-which make bicycles available to the general public on an as-needed basis-have undergone several refinements over the past five decades and, in recent years, have dramatically changed the ecology of urban and, increasingly, suburban transport. This study first characterizes the three phases of Chicago's Divvy system, paying special attention to service and performance gaps and distributions relative to equity. It then develops a series of statistical models designed to identify both community- and station-level factors that best explain variations in Divvy system usage at the station level. The study finds that a number of variables have significant correlations with bikeshare usage at the station level. Neighborhood racial and ethnic diversity, proportion of condominium units, job accessibility to public transit and the average diurnal diversity of Divvy trips for example, are all strongly and positively correlated with total annual station trips whereas percentage unemployed, average distance to Divvy stations and percent of residential foreclosures are negatively correlated.
A First Look: Comparison of Users and Usage Patterns of Dockless and Docking-Station-Based Bikeshare Systems in Washington, D.C.
Ralph Buehler, Virginia Polytechnic Institute and State UniversityShow Abstract View Presentation
Fanglan Chen, Virginia Polytechnic Institute and State University
John Cole, Virginia Polytechnic Institute and State University
John Hicks, Virginia Polytechnic Institute and State University
Andrew Devereux, Virginia Polytechnic Institute and State University
Hazel Ventura, Virginia Polytechnic Institute and State University
In 2017, dockless bikeshare systems launched in U.S. cities. Little is known about the demographics of dockless bikeshare riders and the geographic reach of these systems. Dockless bikeshare can potentially reach different geographic areas compared to docking-station based systems because bicycles do not need to be returned to docking stations. Moreover, dockless bikeshare can potentially reach different users, based on lower user cost, larger geographic coverage, and different pricing structures. This paper provides an initial comparison of rider demographics and geographic reach of five dockless systems and the docking station based Capital Bikeshare (CaBi) system in Washington, DC. Using an intercept survey of dockless bikeshare riders (n=49), results from the CaBi membership survey (n=5,848), and GIS data from CaBi (n=983,854) and the dockless systems (n=152,548) this paper explores user demographics, motivations for riding, and trip alternatives; as well as daily distributions of trips and trip-start and trip-path patterns. Between October 2017 and January 2018 CaBi had 6 times as many trips as all dockless providers combined. Compared to dockless bikeshare, a greater share of CaBi trips was concentrated in the morning and afternoon peaks. Dockless and CaBi systems show many similarities in user demographics, motivations to ride, and geographic areas of system usage. However, ridership of the dockless systems seems to be more diverse than CaBi ridership with greater shares of minority, lower income, and female riders. In addition, compared to CaBi, usage of dockless systems seems to be proportionally less concentrated in the center of the city.
Understanding the Difference in Travel Patterns Between Docked and Dockless Bikesharing Systems: A Case Study in Nanjing, China
Xinwei Ma, Southeast UniversityShow Abstract View Presentation
Yufei Yuan, Delft University of Technology
Niels van Oort, Delft University of Technology
Yanjie Ji, Southeast University
Serge Hoogendoorn, Delft University of Technology
The co-existence of dockless and traditional docked bike-sharing systems presents new opportunities for sustainable transportation in cities all over the world, both serving door to door trips and access and egress to and from transit. To compare travel patterns of these two systems, we explored the GPS data of a dockless bike-sharing scheme and the smart card data of a docked bike-sharing scheme in the city of Nanjing, China over the same time period. In order to obtain information from different perspectives, such as user perception and opinions, an intercept survey on bike-sharing mode choice was conducted. A mode choice model was estimated to reveal the effects of personal information, user perception and experience on bike-sharing usage. Results show that dockless bike-sharing systems have a shorter average travel distance and travel time but a higher use frequency and hourly usage volume compared to docked bike-sharing systems. Trips of docked and dockless bike-sharing on workdays are more frequent than those on weekends, especially during the morning and evening rush hours from 7:00-9:00 and 17:00-19:00, respectively. As to the factors influencing travelers’ mode choice, results show that retirees, enterprise staff and users with E-bikes are less likely to use docked sharing-bikes than dockless bikes. In contrast, high-income travelers and people who are highly sensitive to discounts, internet technology and online payment service are more likely to use the dockless bike-sharing. Finally, policy implications are discussed for cities to improve the performance of docked and dockless bike-sharing systems.
Modeling the Factors Influencing the Activity Spaces of Bikeshare Around Metro Stations: A Spatial Regression Model
Yuchuan Jin, Southeast UniversityShow Abstract
Jianbiao Wang, Southeast University
Mingjia He, Southeast University
Xinwei Ma, Southeast University
Jiao Ye, Southeast University
Metro-bikeshare integration is considered a green and efficient travel model. To better understand bikeshare as a feeder mode to metro, this study explored the factors that influence the activity spaces of bikeshare around metro stations. First, metro-bikeshare transfer trips were recognized through matching the bikeshare smart card data and metro smart card data. Then, Standard Deviation Ellipse (SDE) was used for the calculation of the metro-bikeshare activity spaces. Moreover, an ordinary least squares (OLS) regression and a spatial error model (SEM) are established to reveal the effects of social-demographic, travel-related and built environment factors on the activity spaces of bikeshare around metro stations, and the SEM outperforms OLS significantly in terms of model fit. Results show that the average metro-bikeshare activity space on weekdays is larger than that on weekends. The proportion of local residents promotes the activity space on weekends, while high density of road and metro impedes the activity space on weekdays. Proportion of age group 1(below 18) significantly impedes activity space throughout the week, while that of age group 4(between 45 and retirement age) promotes the activity space on weekdays. Additionally, with the job density increasing, the activity space becomes smaller significantly throughout the week. Also, both on weekdays and weekends, the closer to the CBD, the smaller the activity space will be. This study can offer meaningful implications for policy makers and and city planners to make the bikeshare distribution more reasonable.
How Have Travelers Changed Mode Choices for First/Last Mile Trips After the Introduction of Bicycle-Sharing Schemes: An Empirical Study in Beijing, China
Aihua Fan, Beijing Jiaotong UniversityShow Abstract View Presentation
Xumei Chen, Beijing Jiaotong University
In recent years, there has been rapid development in bicycle-sharing schemes in China. Moreover, such schemes are considered promising solutions to the first/last mile problem. This study investigates the mode choice behaviors of travelers before and after the introduction of bicycle-sharing schemes for first/last mile trips. Individual and demographic attributes, travel characteristics, and details of the built environment are collected to model travel choice behaviors. Travel choice models for first/last mile trips before and after the introduction of bicycle-sharing schemes are determined using a multinomial logit model. It also analyzes the differences in choice behavior between the young and other age groups. The findings show that shared bicycles have become the preferred mode choice, while travelers preferred walking before the implementation of bicycle-sharing schemes. Gender, bicycle availability, and travel frequency were the most significant factors before the implementation of bicycle-sharing schemes. However, after implementation, distance from origin to boarding station dramatically affects mode choices for first/last mile trips. When shared-bicycles are available, the mode choices of middle-aged group depend mainly on gender and distance from origin to boarding station. All factors are not significant for the young and aged groups. More than 80% of public transport travelers take walking and shared bicycles as feeder modes. The proposed models and findings contribute to a better understanding of travelers’ choice behaviors and to the development of solutions for the first/last mile problem.
Station-Level Bikesharing Demand Prediction Based on Graph Convolutional Neural Network Model
Lei Lin, PARC ResearchShow Abstract
Weizi Li, University of North Carolina, Chapel Hill
Srinivas Peeta, Purdue University
This study proposes a novel Graph Convolutional Neural Network with Data-driven Graph Filter (GCNN-DDGF) model that can learn hidden heterogeneous pairwise correlations between stations to predict station-level hourly demand in a large-scale bike-sharing network. Two architectures of the GCNN-DDGF model are explored; GCNNreg-DDGF is a regular GCNN-DDGF model which contains the convolution and feedforward blocks, and GCNNrec-DDGF additionally contains a recurrent block from the Long Short-term Memory neural net-work architecture to capture temporal dependencies in the bike-sharing demand series. Furthermore, four types of GCNN models are proposed whose adjacency matrices are based on various bike-sharing system data, including Spatial Distance matrix (SD), Demand matrix (DE), Average Trip Duration matrix (ATD), and Demand Correlation matrix (DC). These six types of GCNN models and seven other benchmark models are built and compared on a Citi Bike dataset from New York City which includes 272 stations and over 28 million transactions from 2013 to 2016. Results show that the GCNNrec-DDGF performs the best in terms of the Root Mean Square Error, the Mean Absolute Error and the coefficient of determination (R2), followed by the GCNNreg-DDGF. They outperform the other models.
A Heuristic Algorithm for Rebalancing Large-Scale Bikesharing Systems Using Multiple Trucks
Mohammed Elhenawy, Queensland University of TechnologyShow Abstract View Presentation
Youssef Bichiou, Virginia Polytechnic Institute and State University
Hesham Rakha, Virginia Polytechnic Institute and State University
City bikes and bike sharing systems (BSSs) are one solution to reducing traffic-related CO2 emissions. These systems feature a multitude of bike stations scattered around a city where users can borrow bikes from one station and return them to the same or a different station. However, this action may create an unbalanced system, namely, some stations will have excess bikes and other with a low number. In this paper, we propose a solution to address this issue and satisfy expected demand patterns. We unveil an algorithm that provides a delivery truck with a road to follow. The truck will go from one station to another and satisfy individual station demands. Achieving this objective in an optimal manner (i.e., finding the shortest Hamiltonian cycle) is an NP-hard problem. However, for small instances, an exact solution can be obtained. The heuristic algorithm presented can deliver optimal and/or near optimal solutions at substantially lower computational costs, eventhough the algorithm is not parallel and the different subtours are solved in a sequential manner. It has three steps. The first step consists of dividing a large BSS network into “p” subnetworks, where p is also the number of available trucks. Each truck will serve a subnetwork. Then the redistribution within each subnetwork is modeled as a cooperative game and the deferred acceptance algorithm is used to construct a good initial tours. Finally, an improvement to the tours is performed using the 2-opt algorithm. The proposed algorithm was tested using large benchmark instances. The results show promising performance in terms of solution quality and computational time. The proposed algorithm can be used for real-time BSS re-balancing given its computational efficiency.
Investigation of Contributing Factors to Travel Demand of Free-Floating Bikesharing: A Geographically Weighted Regression Approach
Chengcheng Xu, school of transportation, southeast universityShow Abstract View Presentation
Yuxuan Wang, Southeast University
Chen Wang, Southeast University
Pan Liu, school of transportation, southeast university
This study aimed to investigate the factors contributing to the travel demand of free-floating bike sharing system. Data were collected from 87 traffic analysis zones (TAZs) in the downtown area of the Nanjing City, China. To account for the spatial heterogeneity, the geographically weighted Poisson regression (GWPR) models were developed to link trip frequency of free-floating bike sharing with land use characteristics, road density and travel mode split. The performance of the GWPR was compared to that of the traditional generalized linear model (GLM). The results showed that the GWPR outperforms the GLM in predicting the demand of free-floating sharing bike. The GWPR results indicated that shopping centers, leisure places, tourist places, restaurant, institution, public transit percentage, walking percentage, and road network density are the main contributing factors. The comparison results suggested the contributing factors to free-floating bike sharing demand are the same between weekdays and weekends. But the degree of variable effects is different. Moreover, the results of visualization analysis indicated that the effects of candidate variables are quite different across TAZs. These findings provide useful insights in improving the operational performance of free-floating bike sharing system.
Do Monetary Incentive Measures Affect Users’ Illegal Parking of Station-Free Shared Bikes?: A Case Study in Nanjing, China
Yao Fan, Southeast UniversityShow Abstract View Presentation
Yuchuan Jin, Southeast University
Liangpeng Gao, Southeast University
Yanjie Ji, Southeast University
Xinwei Ma, Southeast University
The recent development of shared-bike systems in China has brought not only convenience for users but also great pressure on bicycle parking management. Although research on shared-bike allocation has been conducted to address this problem, there is limited empirical evidence regarding the effectiveness of monetary incentives from the perspective of users. In this paper, the conformity effects among shared-bike users in behavioural decisions related to bicycle parking were modelled and estimated within a log-log regression model that was econometrically identified by using data collected in a face-to-face survey with 322 respondents who had just illegally parked a shared bike. The empirical results indicated that with the increase in walking distance caused by a shift in parking areas, various external factors, such as individual characteristics and trip characteristics, would have positive or negative effects on the desired amount of incentive cash. In addition, as expected, in the parking-stressed areas an appropriate incentive can encourage shared-bike users to spontaneously park their shared bikes in an unsaturated area near their destinations.
Dockless in Sydney: The Rise and Decline of Bikesharing in Australia
Capucine Heymes, Universite de LyonShow Abstract View Presentation
David Levinson, University of Sydney
In mid-2017, dockless, (or stationless) bikesharing appeared on the streets of Sydney. The birth of dockless bikesharing, its evolution as well as its consequences, and use habits are studied with review of policies and field investigations. It is found that bicycle use in Sydney is less than hoped for, vandalism is high, regulations unfavourable, and thus, the conditions for successful bikesharing are not met.
Station-Level Analysis for Bikeshare Activity in Disadvantaged Communities
Xiaodong Qian, University of California, DavisShow Abstract
Miguel Jaller, University of California, Davis
Bikeshare systems are rapidly becoming more prevalent around the world. This popularity means that an increasing number of people can enjoy the convenience of cycling and the associated physical health benefits without actually owning a bike (or having access to their own bikes). However, low-income populations and people of color are not highly represented users of bikeshares. Currently, there is limited research to estimate bikeshare ridership in these communities. This research fills that gap by analyzing current utilization rates of bikeshare systems among disadvantaged populations. This study develops a Negative Binomial regression model to estimate bikeshare ridership in disadvantaged areas using data from Chicago’s bikeshare system (Divvy). The results show that bikeshare stations in disadvantaged communities have approximately 2,380 annual trips less than in other areas on average (an approximate 32% reduction from the average generated trips in all stations). Among factors influencing bikeshare trips, employment rate has the highest positive marginal effect. Additionally, this research analyzes the bikeshare trip utilization between annual member and 24-hour pass users in disadvantaged areas. The proportion of trips by subscribers is significantly lower in disadvantaged communities than in other areas. Interestingly, residents in disadvantaged communities tend to make longer bikeshare trips once they are already annual members.
The Effects of Neighborhood Disadvantage on Temporal and Spatial Patterns of Bikeshare Use
Jueyu Wang, University of MinnesotaShow Abstract View Presentation
Greg Lindsey, Humphrey School of Public Affairs
Policy-makers and operators of bike share programs need information about factors that affect member use to manage systems effectively and to address questions of equity of access that have been raised in many urban areas. Using 2017 data on trips taken by 30-day and annual members of the Nice Ride Bike Share system in Minneapolis-St. Paul, we estimate multi-level, mixed-effect regression and multinomial logistic models to analyze the effects of neighborhood disadvantage on, respectively, (1) the frequency and duration of weekday and weekend trips by members and (2) the temporal and spatial patterns of their trips. We use principal components analysis to establish a standardized, composite index of neighborhood disadvantage (SNDCI). We use k-means clustering procedures to establish two temporal and four spatial patterns of member trips. Our results show that, after controlling for station accessibility, nearby bike infrastructure, the built environment, gender, and age, members who reside in disadvantaged neighborhoods use bike share more frequently, take longer trips, and have more dispersed spatial patterns associated with more origin-destination pairs. Our findings have implications for efforts to serve members in disadvantaged neighborhoods and illustrate the need for more detailed surveys of members to obtain additional information about individual characteristics associated with travel behavior.
Widening the Gap?: Area-Level Associations Between Bicycle Commuting, Sociodemographic Advantage, and Bike Lane Access in 22 U.S. Cities
Lindsay Braun, University of Illinois at Urbana-ChampaignShow Abstract
While past research has found bike lanes to be positively associated with cycling, few studies have considered whether these associations vary by sociodemographic characteristics such as race, ethnicity, and socioeconomic status (SES). Similarly, while bike lane investment is often seen as a method for reducing sociodemographic disparities in cycling and health, limited research has examined whether cycling disparities are, in fact, lower in places with bike lanes. To address these gaps, I examined cross-sectional associations of bicycle commuting mode shares (dependent variable) with sociodemographic advantage and bike lane presence (independent variables) in 22 U.S. cities, using block groups as the unit of analysis. Sociodemographic characteristics included six block group-level measures of race, ethnicity, and SES. I used multilevel mixed-effects regression models with interaction terms to assess (1) whether associations between bike lane presence and bicycle commuting were moderated by each sociodemographic characteristic and (2) whether disparities in bicycle commuting (i.e. higher bicycle commuting mode shares in block groups with greater sociodemographic advantage) were moderated by bike lane presence. Across all sociodemographic characteristics, (1) associations between bike lanes and bicycle commuting were stronger among more advantaged block groups and (2) disparities in bicycle commuting were more pronounced among block groups that contained bike lanes. These findings suggest that bike lane investment could have the unintended consequence of widening the gap in cycling between more- and less-advantaged places. An expanded conceptualization of “access” that addresses non-infrastructure barriers to cycling is needed if bike lane investment is to effectively advance social equity goals.
Analysis of Bicyclist Race and Ethnicity from Eight Travel Surveys in the United States
Sarah Howerter, University of VermontShow Abstract View Presentation
Dillon Fitch, University of California, Davis
Lisa Aultman-Hall, University of Vermont
The limited use of the bicycle, whether for commuting or other trips, in the United States results in few bicyclist observations within the large random transportation surveys that might be used to better understand factors associated with differential levels of bicycling including associations with racial identity. Moreover, specialized bicycle surveys, such as those undertaken with clubs or workplaces, often fail to include diverse racial groups. Prior studies on race and bicycling are limited in number and show unclear trends. Relationships between racial identity and bicycling may be due to socio-cultural mechanisms or simply other confounding factors, especially income. Compilation of eight travel surveys in this project to increase sample size was complicated by the multiple methods used to measure race and limited bicycling questions. The process to tabulate and obtain common race classifications is described. In five of the surveys, the measurement of bicycling used questions beyond commute mode and trips undertaken on the travel day. This improved our ability to subdivide the sample by race. Consistently across surveys, White and mixed-race individuals bicycled more. Blacks bicycled less, but other trends with respect to racial identity were limited. Cross regional patterns suggested other factors, such as climate and culture, had a similar effect on people of different races and that racial identity may be less important than these other factors. With this newly compiled large dataset, multivariate modeling can be pursued to more fully consider race with other potentially confounding factors.
Assessing Bicyclists’ Perceived Safety, Risk, and Comfort: State of the Literature
Erin Robartes, University of VirginiaShow Abstract View Presentation
T. Donna Chen, University of Virginia
Expanding bicycle mode share can be achieved by two means: 1) increasing frequency and distance of current bicyclists’ cycling activity, and 2) attracting more travelers to choose bicycling. Previous synthesis studies have examined the impact of infrastructure characteristics on measured rates of bicycling, which serve to better understand preferences of current bicyclists. This synthesis paper focuses on studies related to travelers’ perceptions of bicycle safety, comfort, and risk, and measures of roadway suitability for bicycling (level of traffic stress and bicycle level of service), as a way to better understand preferences of both bicyclists and non-cyclists. This work assesses the different research methods, compares the nuances between the descriptors (safety, risk, comfort, level of service, level of stress) and reviews measured results. The comprehensive literature review reveals much overlap in the findings of literature related to these topics, including greater perceptions of comfort and safety with lower motor vehicle volumes and speeds, wider bike and automobile lane widths, absence of on-street parking, and greater physical separation between automobiles and bicycles. In terms of methodological approach, bicyclists’ feelings of safety, risk, and comfort have traditionally been measured using stated preference surveys. Recently, these studies have been enhanced by use of video clips and simulations, cell phone apps, instrumented bicycles in naturalistic experiments, and virtual reality. These trends point to a future where the ability to better replicate a wider variety of real-world settings as a key factor in obtaining more nuanced results regarding bicyclists’ perception of safety, risk, and comfort.
Industry Stakeholder Perspectives on the Adoption of Electric Bicycles in British Columbia
Saki Aono, University of British ColumbiaShow Abstract View Presentation
Alexander Bigazzi, University of British Columbia
Electric-assist bicycles (e-bikes) are an emerging mode of transportation that offers a sustainable alternative to automobile use in urban areas. Past research on e-bike adoption has focused on user perspectives. Understanding other stakeholder perspectives is also essential to implementing effective e-bike policy. The objectives of this research are to identify alignments and misalignments in perspectives on e-bike adoption across industry stakeholders in British Columbia (BC), including e-bike retailers, manufacturers, cycling coalitions, and government agencies, and to provide recommendations for e-bike policy that account for those perspectives. An online survey was distributed to industry stakeholders to examine perceived barriers to adoption, expected impacts of adoption, and effects of policy on adoption. Questions about regulations were discriminated among five e-bike types: pedal-assist, throttle-assist, scooter-style, electric recumbents, and enclosed electric recumbents. Results indicate strong agreement among industry stakeholders that scooter-style e-bikes require separate and additional regulation from other types of e-bikes and from existing regulation in BC. In contrast, there was misalignment in the expected mode shift resulting from e-bike adoption, with government agencies least optimistic about diversion of automobile trips. Industry stakeholders broadly agreed on the need for speed regulation and viewed higher speeds as one of a least-important benefits of e-bikes, which contrasts with past research on user perspectives. Policy recommendations include reclassifying scooter style e-bikes, rebate or tax programs to reduce e-bike costs, further research on optimal e-bike speed limits, and continued support for improvements in general cycling infrastructure (a top priority for industry and user stakeholders).
Electric Kick Scooters on Sidewalks in Virginia But Not in California: A Review of How States Regulate Personal Transportation Devices
Asha Weinstein Agrawal, San Jose State UniversityShow Abstract
Kevin Fang, Sonoma State University
Ashley M. Hooper, University of California, Irvine
In recent years, communities have observed fast growth in the number and use of “personal transportation devices” (PTDs) such as Segways, hoverboards, skateboards, kick scooters, and electric kick-scooters. PTDs provide mobility for people making short trips who want to travel faster than walking speeds. Greater use of PTDs has the potential to benefit both individual travelers and communities as a whole, providing inexpensive mobility options that can help communities reach their environmental and congestion-relief goals. Despite the long list of possible benefits, however, incorporating PTDs into communities is not without challenges, most notably thorny regulatory and facilities management questions. For this paper, we reviewed in detail the state vehicle codes from U.S. states and territories to identify all direct or indirect regulations on PTDs of all types. The findings show that, overall, no widely-established state regulations clearly and effectively regulate where and how PTDs should be used in order to manage the risk of collisions while still making PTDs a viable, convenient transportation option. Other key findings include: 1) states regulate devices with similar characteristics differently; 2) many states have very few PTDs regulations; and 3) states are inconsistent in terms of where PTDs should be ridden.
Do They Block the Way in San Jose?: Where Do Riders Park Dockless, Shared Electric Scooters?
Kevin Fang, Sonoma State UniversityShow Abstract
Asha Weinstein Agrawal, San Jose State University
Jeremy Steele, Mineta Transportation Institute San Jose State University
John Joseph Hunter, Mineta Transportation Institute San Jose State University
Ashley M. Hooper, University of California, Irvine
Dockless, shared, electric kick-scooters systems have been springing up in cities across the country since late 2017. These shared scooters have proven popular with riders, attracting investment capital and increasing expansion of shared scooters to additional cities. However, shared scooters have attracted sometimes heated opposition. One of the major points of contention is over how shared scooters are parked, with complaints that shared scooters clutter sidewalks and block pedestrian access, particularly for the disabled. To evaluate the degree to which shared scooters are “well-parked” or not, we photographed and analyzed 530 parked scooters in San Jose, California. The study found that the shared scooters were for the most part parked inoffensively. While about three quarters are parked on sidewalks, the vast majority are parked out of the way of pedestrian flow on the edge of sidewalks or in already-obstructed “street furniture zones.” Furthermore, less than two percent of the shared scooters observed were blocking accessibility infrastructure for the disabled, doors, or more than half the width of a sidewalk, standards that the City of San Jose requires for parked bicycles.
Does One Bicycle Facility Type Fit All?: Evaluating the Stated Usage of Different Types of Bicycle Facilities Among Cyclists in Quebec City, Canada
Marie-Pier Veillette, Oc TranspoShow Abstract View Presentation
Emily Grise, University of Alberta
Ahmed El-Geneidy, McGill University
For cities wishing to foster a strong culture of cycling, developing a network of safe and efficient bicycle infrastructure is paramount, yet not a straightforward task. Once transport professionals have selected the optimal location for a new bicycle facility, determining the optimal facility type is imperative to ensure that new infrastructure encourages cycling trips and increases the safety of cyclists. The present study presents a nuanced approach to evaluating cyclists’ usage of various types of bicycle facilities. To achieve this goal, we employed survey data of cyclists in Quebec City, Canada, to study how many cyclists reported using a particular bicycle facility in the survey against their reasonable access to those reported facilities. To account for different preferences, behaviour and motivations among cyclists, we segmented our study sample into six types of cyclists. Finally, regression modeling is employed to predict the stated usage of three facility types present in the study area (recreational path, bi-directional protected lane and painted lane), while controlling for access to this path, cyclist type, and personal and neighbourhood characteristics. Results indicate that if a cyclist has access to each facility type on their commute, they are most likely to use a recreational path on their commute, followed by a painted bicycle lane. Respondents with access to a bi-directional lane are no more likely to report using this facility than respondents without access. Overall, this study is intended to encourage a dialogue between cyclists and transport practitioners to uncover the factors contributing to effective bicycle infrastructure.
Rethinking Streets for People on Bikes: An Evidence-Based, Visual Guide of Completed Street Retrofits
Marc Schlossberg, University of OregonShow Abstract
Roger Lindgren, Oregon Institute of Technology
Cities of all sizes are increasingly interested in bicycle transportation and are pursuing a new range of infrastructure from protected bike lanes to bikeshare systems. There are many excellent street design manuals that provide good reference material, technical specifications, and case study information for local professionals to use. Yet, it is also difficult for many cities to retrofit streets for bicycle transportation because such decisions are inherently political, including outcomes based on public input on street projects. Despite the political nature of street retrofit decisions, most design guides are generally targeted only toward transportation professionals, often get too detailed and thus overwhelming to broader audiences, and sometimes focus on the hypothetical street re-design rather than evidence from completed projects. In 2014, a research team at the University of Oregon released Rethinking Streets: An Evidence Based Guide to 25 Street Transformations to provide guidance to the full range of community stakeholders usually engaged in street projects: engineers, planners, politicians, urban designers, and the general public. At TRB 2019, we would like to debut the second iteration - Rethinking Streets for Bikes – a free, publicly available, multi-stakeholder, visually accessible guide of high quality, completed, bicycle transportation projects from across the U.S. (expected release date: December 2018). Bicycle transportation system changes are being adopted all across the country, but local officials have few documented guidebooks to help them think about how to retrofit streets for people on bikes based on completed projects using best practices. Rethinking Streets for Bikes fills this gap.
Operational Evaluation of Advisory Bike Lane Treatment on Road User Behavior in Ottawa, Canada
Ali Kassim, Carleton UniversityShow Abstract
Alex Culley, Transportation Services City of Ottawa Ontario Canada
Shawn McGuire, City of Ottawa
The City of Ottawa has been investigating design solutions to facilitate the addition of cycling facilities, while maintaining parking, to roadways with limited right-of-ways. A pilot project to install Advisory Bike Lanes was initiated. The purpose of this study is to determine how new pavement markings (advisory bike lanes) influence cyclist and motorist interactions and positioning, especially with respect to the distance between motorists and cyclists when passing. The study presents a before/ after evaluation of two contrasting pavement indications. Video data were collected in two phases (pre- and post- treatment) where each phase consisted of two different days. A number of safety performance parameters were used to assess whether safer conditions existed after the new treatment was installed: [i] the lateral distance between the motor vehicle and cyclist [ii] the lateral distance between the cyclist and curbside edge/cyclist and buffer edge line and [iii] the speed of the cyclists and motor vehicle. The findings indicate that the advisory bike lanes created more favorable conditions for cyclist safety and for motor vehicle compliance. These findings are: [i] motorists passed cyclists with a greater lateral separation distance, [ii] cyclists positioned themselves further from parking edge line and rode in the middle of the bike lane, [iii] motor vehicle travelling speed decreased (the 85th percentile speed is decreased by 5.2% after the installation of the advisory bike lane), and [iv] average cyclist speed increased (the average cyclist speed is increased by 7.7% after the installation of the advisory bike lane).
Bicycle Infrastructure and Commercial Rents in San Francisco: A Hedonic Regression Analysis
Raleigh McCoy, Metropolitan Transportation Commission (MTC)Show Abstract View Presentation
Much work has been done to examine the influence of bicycle infrastructure on the market for urban residential real estate, but the same attention has not been paid to the commercial real estate market. This paper seeks to understand the relationship between proximity to bicycle infrastructure and commercial rents through an empirical analysis of publicly available secondary rent data. A hedonic regression model is applied to a sample of commercial properties listed for rent in San Francisco, California. Characteristics of the property, surrounding neighborhood, and zoning designation are included in the model to isolate the effect of proximity to bicycle infrastructure. The paper finds that proximity to bicycle infrastructure of any kind is not associated with rent per square foot for commercial properties, though characteristics of the property and the surrounding neighborhood are significant. The main takeaway is that neither the accessibility bonus nor the perceived disamenities (e.g., removal of parking, disruption of access by those arriving by automobile) provided by proximity to bicycle infrastructure appear to be capitalized within the commercial real estate market.
Assessment of Local, State, and Federal Barriers to Implementing Bicycle Infrastructure: A Virginia Case Study
Erin Robartes, University of VirginiaShow Abstract View Presentation
Emily Chen, University of Virginia
T. Donna Chen, University of Virginia
Peter Ohlms, Virginia Department of Transportation
Bicycle infrastructure can increase comfort of bicyclists, grow bicycle mode share, and reduce bicycle-vehicle crashes. With worldwide goals to increase non-motorized vehicle travel and decrease traffic fatalities, successful bicycle infrastructure implementation is a key component to a safer and more comfortable bicycling environment. Yet, transportation planners and engineers often face difficulties in implementing bicycle infrastructure. This study examines the state of bicycle infrastructure and the barriers to implementation encountered in Virginia. A two-stage survey was deployed to transportation planners, engineers, and other government personnel at the town, city, county, regional, and state levels to collect information about their experiences with bicycle infrastructure implementation. Survey results indicate that the majority (77%) of jurisdictions intended to improve bicycle infrastructure in the next five years. However, despite this anticipated progress, those surveyed cited substantial barriers towards actually implementing the infrastructure. Lack of funding was the most commonly cited barrier, with 61% of respondents reporting this barrier at all levels of government. At the local level, lack of public support was cited as the most common barrier. Yet, less than half of the respondents indicate their jurisdiction has an established process for residents to provide input for bicycle infrastructure improvements. Additionally, 21% of respondents mentioned right-of-way acquisition as a difficulty. In summary, this study presents a framework for assessing the state of bicycle infrastructure development and identifies specific barriers at various government levels, allowing policy makers and transportation planners to target barriers specific to their governing unit, for more effective bicycle infrastructure implementation.
An Assessment of Bicycle Commuting Activity Using Strava and Implications for Developing Infrastructure
Scott Kelley, University of Nevada, RenoShow Abstract View Presentation
Klenke Chrissy, University of Nevada, Reno
Amy Fitch, University of Nevada, Reno
Nathan Bergrin, The Davidson Academy
Alexander Hacker, University of Nevada, Reno
Liam Bean, University of Nevada, Reno
Miguel Aguilera, University of Nevada, Reno
Cameron Krek, University of Nevada, Reno
The development of dedicated infrastructure for cyclists, including designated lanes, paint, and signage can encourage higher levels of bicycle commuting. Empirical data on cyclist activity is valuable when recommending locations for such interventions, especially in areas where it is currently sparse. In this study, we consider bicycle commuting to the University of Nevada, Reno (UNR) campus, which lies in the center of rapidly-growing Reno, Nevada. There is university-wide interest in developing a more robust bicycling infrastructure, since the campus currently lacks it. As a priority first step, this study aims to assess the nature of current UNR bicycle commuter activity, including how commuters travel through the campus and how they access it from the surrounding community. To address this, we recruited 54 members of the UNR community to record all bicycle commuting activity for two weeks in April 2018 using the popular fitness application Strava. Observed routes are compared with on-campus suggested travel paths, city bicycle lanes, steep terrain, high-traffic roads, and shortest path use using GIS analysis. Results show that intracampus commuting is uncommon, and that while commuters do often follow suggested on-campus travel paths, only 25% of observed commuting in the surrounding community occurs along city bicycle lanes. A logit model demonstrates that shortest path routes are also more likely to avoid steep terrain and less likely to use city bicycle lanes. We discuss the implications of these findings and provide recommendations on future bicycle infrastructure development on the UNR campus and the surrounding community.
Utilization of Secure Bicycle Parking Rooms in Multi-Unit Residential Buildings
Cail Smith, University of British ColumbiaShow Abstract View Presentation
Alexander Bigazzi, University of British Columbia
Existing research supports the importance of high-quality bicycle parking facilities for cycling promotion but does not provide quantitative data on utilization in residential buildings. Secure bicycle parking rooms in large developments are important for cycling policy in cities such as Vancouver, Canada, where 42% of households live in apartments in multi-unit buildings. A better understanding of how bicycle parking and storage spaces in these buildings are used can help develop guidelines that support residents choosing to cycle. The objective of this study was to provide quantitative information on the utilization of secure bicycle parking rooms in multi-unit residential buildings for university staff near a large post-secondary institution. Counts were made to quantify the number of bicycles in secure parking rooms used over time in three sample buildings, and residents were surveyed to investigate perceptions, preferences, and bicycle parking demand. Even meeting current guidelines with approximately 1.5 spaces per unit of secure bicycle parking capacity, there is heavy bicycle parking congestion in the study buildings with overall occupancy of 99%. Around 1/3rd of the bicycles were used within the first week of the study, increasing steadily to 2/3rd after 9 weeks. Most respondents with bicycles (65%) regularly store them in locations other than the bicycle parking rooms, indicating a high amount of latent demand for bicycle parking in this context. Policy recommendations include consideration of higher bicycle parking capacity in development standards, and provision of different types of bicycle parking for frequent, low-barrier access versus long-term storage.
Toward Agent-Based Microsimulation of Cyclist Following Behavior: Estimation of Reward Function Parameters Using Inverse Reinforcement Learning
Hossameldin Mohammed, University of British ColumbiaShow Abstract View Presentation
Tarek Sayed, University of British Columbia
Alexander Bigazzi, University of British Columbia
Reward functions are a key component in developing agent-based microsimulation models. The objective of this research is to estimate reward function parameters for cyclists in following interactions with other cyclists on bicycle paths. Decisions of cyclists (acceleration and direction) in following interactions are modeled as a finite state Markov Decision Process, in which the reward function describing the desired state of the cyclist is unknown. Two algorithms of imitation learning using Inverse Reinforcement Learning (IRL) are evaluated to estimate reward function parameters: Feature Matching (FM) and Maximum Entropy (ME) IRL. The algorithms are trained on 1297 cyclist trajectories in following interactions extracted from video data using computer vision, and then validated using a separate set of 349 trajectories. The estimated reward function parameters indicate how cyclists weigh the five state features in the reward function: speed, speed difference from leading cyclist, lateral position in path, lateral distance from leading cyclist, and longitudinal distance from leading cyclist. Following cyclists tend to prefer intermediate values of longitudinal and lateral distance to leading cyclists. Cyclists also prefer high speeds, with low speed difference from the leading cyclist and low deviation from the center of the path. Implementation of the reward functions derived from the FM and ME algorithms correctly predicted 58% and 67%, respectively, of the observed cyclist decisions (acceleration and direction) in the validation data set. This research is a key step toward developing operational bicycle traffic microsimulation models with applications such as facility planning and bicycle safety modeling.
Station-Level Demand Forecasting in a Public Bicycle Sharing System Using Station Activity Based on Random Forest
Young-Hyun Seo, Seoul National UniversityShow Abstract
Jaemin Hwang, Seoul National University
Seung-Young Kho, Seoul National University
Dong-Kyu Kim, Seoul National University
The public bicycle sharing (PBS) system is one of the modes of transportation that can help to relieve several urban problems, such as traffic congestion, air pollution, and high oil prices. Because users can rent and return bicycles anytime and anywhere a station is located, inventory imbalances can occur. Therefore, to prevent system failures, such as not being able to rent a bicycle, the operator must establish an appropriate repositioning strategy. For efficient relocation, accurate demand forecasting must be done first. In this study, we use station activity information to predict the station-level demand for public bicycles. In addition to temporal factors and meteorological factors, we use the number of rentals and returns at the station one to three hours before the prediction as predictors. We predict the demand for public bicycles using the random forest, and we use the evaluation index to compare the results and improve our predictions. The analysis shows that the accuracy of our predictions increases by 2.3% when we use the lag information of the station. By the stock variation of the station, the stations are classified into four clusters, i.e., unstable, short, overstocked, and stable, and forecasting demand should be done more frequently for unstable stations. This study can be used to calculate a safety stock of the station when repositioning bicycles by improving the predictability of demand, which can lead to operational efficiency and ultimately reduce costs.
Modeling Instantaneous Cyclist Acceleration and Deceleration Behavior
Ahmed Ghanem, Virginia Polytechnic Institute and State UniversityShow Abstract View Presentation
Hesham Rakha, Virginia Polytechnic Institute and State University
Cycling has gained more acceptance as a sustainable mode of transportation that can provide an excellent solution for short-distance transfers for several reasons. Typically, cycling involves traveling in less-congested conditions, reduces the traveler's carbon footprint, and improves the traveler's lifestyle. In recent years, bike sharing systems (BSSs) have been introduced in many cities, and there is a growing need to further incorporate bicycles into traffic planning. Many of the widely used microscopic traffic simulation tools have now been extended to model bicycle traffic in addition to vehicular traffic. However, the accuracy and reliability of these frameworks depends mainly on understanding cyclist behavior, but to date there have been few studies in the literature that provide such models. In this paper, we used cycling Global Positioning System (GPS) data collected from 10 people (3 females and 7 males) to develop a dynamics-based cycling acceleration model that captures cyclist aggressiveness. We augmented the model by calibrating the maximum power for average cyclists. We also developed a model that captures cyclist deceleration behavior. The results show that the acceleration model can estimate the cyclist's pedaling input with a root-mean-square error (RMSE) of less than 21% in most cases. The results also show that the deceleration model can estimate cycling deceleration with an RMSE of 12%.
Improving Bicycle Route Choice Set Generation Using Route Complexity in GPS Traces
Niels Wardenier, Not applicableShow Abstract View Presentation
Luk Knapen, UHasselt
Thomas Koch, Vrije Universiteit
Elenna Dugundji, Vrije Universiteit
Everyday route choices made by bicyclists are known to be more difficult to explain than vehicle routes, yet prediction of these choices is essential for guiding infrastructural investment in safe cycling. In this paper we study how the concept of route complexity can help generate plausible choice sets in the demand modeling process. Defining route complexity to be the number of shortest path segments that form a complete route, we characterize the routes bicyclists take in large set of GPS traces gathered voluntarily by persons traveling to everyday activities at work, school, friends, etc. Subsequently we analyze how predicted bicycle routes from a route choice set generator compare to the GPS observations in the field. Finally we propose two sampling methods to improve the output from the choice set generator.
A Method to Identify and Visualize Barriers in a Low-Stress Bike Network
Theja Putta, Northeastern UniversityShow Abstract View Presentation
Peter Furth, Northeastern University
Low-stress bike networks are often disconnected, with gaps or barriers that make travel between two points impossible without riding on high stress roads. Barriers can also force long detours that people are not willing to make. While existing methods of low-stress bike network analysis have been used to point out some barriers, a method is needed to systematically identify and draw barriers to assist in network planning. Such a method is developed, taking only the low-stress network as an input, and yielding a set of polylines that indicate barriers to bicycling. Applications in Arlington, Virginia and Boston show how it detects what might otherwise have been hidden barriers. The method also successfully highlights critical low stress links that breach what would otherwise have been a far longer barrier.
Bikeability in Basel
Elena Grigore, ETH ZurichShow Abstract View Presentation
Norman Garrick, University of Connecticut
Raphael Fuhrer, ETHZ - Swiss Federal Institute of Technology
Kay Axhausen, Eidgenossische Technische Hochschule Zurich
“Bikeability” is becoming increasingly relevant in the field of transport- and urban planning. However, it is often unclear how bikeability is defined, let alone how it can be modeled. The goal of this project is to develop a quantitative method to model bikeability. A case study area in the city of Basel, Switzerland has been selected for assessing the model. Here “bikeability” is understood as the ability and convenience to reach important destinations by bike, based on the travel distance weighted by the perceived safety, comfort and attractiveness of the streets and intersections along the routes. The underlying assumption is that cyclists try to minimize the distance traveled and maximize the perceived safety, comfort and attractiveness of their route of choice. Unlike most of the previous bikeability assessements that we have reviewed, our method uses existing route choice studies to identify attributes for quantifying cycling quality, which presumably results in a model, which more accurately reflects real life behavior. Many relevant attributes are included in this work that have not been captured by previous models, such as high curbs of tram stops, tram tracks and the turn direction at intersections. The method is suitable for several applications in urban planning, such as the identification of locations that need improvement and the comparison of various planning measures. The current model covers conventional bikes being used by commuting cyclists. However, the method can be used for E-bikes and non-commuting cyclists by applying the appropriate input values.
Low-Stress Bicycling Connectivity: Assessment of the Network Build-Out in Edmonton, Canada
Laura Cabral, Toole DesignShow Abstract View Presentation
Amy Kim, University of Alberta
Manish Shirgaokar, University of Colorado, Denver
Studies have shown that a network of safe, connected, and direct facilities increase urban cycling levels. During summer 2017, the city of Edmonton, Canada, constructed nearly 20 kilometers of protected bicycle lanes on its core neighborhood streets. In this paper, we evaluate the low-stress connectivity improvements afforded by this network build-out. We first classify streets and trails according to the Level of Traffic Stress (LTS) framework, which we adapt to the metric system. Using only LTS 2 network links, posited to be adequate for most adults, we apply three analyses. First, we draw “Bikeshed” maps, which show areas of connectivity around seven central destinations. Our comparison before and after the build-out points to a better integration of the network, with previously separate bikesheds overlapping and allowing uninterrupted low-stress travel to more destinations. This analysis also allows us to identify several central neighborhoods which are disconnected due to remaining high-stress links. Second, we generate roughly three-hundred hypothetical origins located in the central neighborhoods of the city. Reflecting the improved bikeshed integration, we observe a four-fold increase in connected origin-destination pairs. Finally, we find small reductions in trip lengths between connected pairs for some of the trips that were possible before the build-out. However, important detours are still necessary to remain on an exclusively low-stress network when compared to the shortest path using the full network, regardless of LTS. Our proposed assessment methodology is straightforward and most relevant for cities in the initial stages of bicycle network development.
Capacity, Capacity Drop, and Relation of Capacity to the Path Width in Bicycle Traffic
Maria Wierbos, Delft University of TechnologyShow Abstract
Victor Knoop, Delft University of Technology
Flurin Hänseler, Delft University of Technology
Serge Hoogendoorn, Delft University of Technology
Bicycle usage is encouraged in many cities because of its health and environmental benefits. As a result, bicycle traffic increases which leads to questions on the requirements of bicycle infrastructure. Design guidelines are available but the scientific substantiation is limited. This research contributes to understanding bicycle traffic flow by studying the aggregated movements of cyclists before and after the onset of congestion within the setting of a controlled bottleneck flow experiment. The paper quantitatively describes the relation between capacity and path width, provides a qualitative explanation of this relation by analyzing the cyclist configuration for different path widths, and studies the existence of a capacity drop in bicycle flow. Using slanted cumulative curves and regression analysis, the capacity of a bicycle path is found to increase linearly with increasing path width. A steady drop in flow rate is observed after the onset of congestion, indicating that the capacity drop phenomenon is observed in bicycle traffic. The results presented in this paper can help city planners to create bicycle infrastructure that can handle high cyclist demand.
Operational Impacts of Protected-Permitted Right Turn Phasing and Pavement Markings on Bicyclists Performance During Conflicts with Right Turning Vehicles
Masoud Ghodrat Abadi, California State University, SacramentoShow Abstract
David S. Hurwitz, Oregon State University
Conflict between bicycles and right-turning vehicles on the approach to signalized intersections is a critical safety concern. To understand the operational implications of protected-permitted right-turn signal indications in conjunction with pavement markings on bicyclist performance, a full-scale bicycling simulator experiment was performed. Velocity and lateral position of bicyclists were evaluated during conflicts between bicycles and right-turning vehicles. A mixed factorial design was considered. Two within-subject factors were analyzed: the signal indication for right-turning vehicles with five levels (circular red, circular green, solid red arrow, solid green arrow, and flashing yellow arrow), and the pavement markings in the conflict area with two levels (white lane markings with no supplemental pavement color and white lane markings with solid green pavement applied in the conflict area). Additionally, the influence of gender as a between-subject variable was considered. Forty-eight participants (24 female) completed the experiment. Signal indications and pavement markings had statistically significant effects on bicyclist velocity and lateral position, but these effects varied at different factor levels. Additionally, amid the conflicts, male participants were found to have higher velocity than female participants. This difference was not influenced by engineering treatments. The results provide guidance to transportation professionals about how traffic control devices could be applied to conflict areas on the approach to signalized intersections.
Analysis of Bicycle Headway Distribution, Saturation Flow, and Capacity at a Signalized Intersection Using Empirical Trajectory Data
Yufei Yuan, Delft University of TechnologyShow Abstract View Presentation
Bernat Goñi-Ros, Delft University of Technology
Mees Poppe, Delft University of Technology
Winnie Daamen, Delft University of Technology
Serge Hoogendoorn, Delft University of Technology
Predicting the bicycle flow capacity at signalized intersections is crucial for urban infrastructure design and traffic management. However, it is also a difficult task due to the large heterogeneity in cycling behavior and several limitations of traditional capacity estimation methods. This paper proposes several methodological improvements, illustrates them using high-resolution trajectory data collected at a busy signalized intersection in the Netherlands, and investigates the influence of key variables of capacity estimation. More specifically, we show that the (virtual) sublane width has a significant effect on the shape of the headway distribution at the stop line. Furthermore, we propose a new method to calculate the saturation headway (a key variable determining capacity), which excludes the cyclists initially located close to the stop line using a distance-based rule instead of a fixed number. We also show that the saturation headway is quite sensitive to the sublane width. Moreover, we propose a new, empirically-based method to identify the number of sublanes that can be accommodated in a given cycle path (another key variable). This method yields considerably lower estimates of the number of sublanes than traditional methods, which rely solely on the (available) cycle path width. Finally, we show that methodological choices such as the sublane width and the method used to estimate the number of sublanes have a considerable effect on capacity estimates. Therefore, this paper highlights the need to define a sound methodology to estimate bicycle flow capacity at signalized intersections and proposes some steps to move towards that direction.
Optimal Speed Advice for Cyclists Using a Roadside Sign at Signalized Intersections with Uncertainty in Traffic Light Timing
Azita Dabiri, Delft University of TechnologyShow Abstract
A. Hegyi, Delft University of Technology
Bernat Goñi-Ros, Delft University of Technology
Signalized intersections are one of the most common source of inconvenience for cyclists. The aim of this paper is to develop an approach that helps cyclists to meet their cycling preferences (regarding, e.g., the energy they spend, and their preference to avoid any unnecessary stop) while passing the intersections. The uncertainty in the traffic light's timing is considered explicitly in the approach, which makes the approach applicable for intersections with traffic-responsive signals. The suggested approach provides cyclists with optimal and personalised speed advice. The advice is communicated to the cyclists through a roadside sign located upstream of the intersection. It is assumed that the roadside sign can measure the speed of the approaching bike and can also communicate with the traffic light to get the traffic light's state and give advice accordingly. To consider the cyclist's behaviour when they disregard the advice as they get close to the intersection, the problem is separated in two parts and formulated as a Markov Reward Process combined with a Markov Decision Process and stochastic dynamic programming is used to solve the corresponding optimization problem. The approach is generic in terms of the underlying process model, and the objective function. The results of an illustrative case study shows how much improvement, in terms of the cyclist's average number of stops, and average energy consumption, could be achieved by use of the suggested approach in a simulated intersection. We also investigate how the location of the sign may effect the performance of the approach.
Research on E-Bike Conversion Factors at the Signalized Intersection of the Urban Road
Hongwei Li, Hohai UniversityShow Abstract
Xin Zhong, Hohai University
Yunyue Zhou, Hohai University
With the rapid development of China's economy and society and the accelerating process of urbanization, e-bikes have gradually become one of the main travel tools in urban road traffic because of their advantages of speed, convenience and high efficiency. Therefore, e-bikes need to be converted into standard models, so that the mixed traffic volume under various road traffic conditions is comparable. By analyzing the field survey data of two signalized intersections in Nanjing, this paper obtains that the effective driving area can effectively characterize the influence of e-bikes on the driving of vehicles and the speed and traffic volume are the main influencing factors of the number of traffic conflict events. On this basis, this paper establishes two calculation models of e-bike conversion factor based on the effective driving area of vehicles and the number of traffic conflict events. By comparing the calculation results of the two models and other scholars, the vehicle conversion factors of the e-bikes in the left, straight and right directions of the urban road signal intersections are 0.28, 0.33, 0.22 respectively.
Macroscopic Characteristics of Bicycle Traffic on the Bike Lane of the Brooklyn Bridge
Rasha Hassan, McMaster UniversityShow Abstract
Mohamed Hussein, McMaster University
Tarek Sayed, University of British Columbia
This study investigates the macroscopic characteristics of bike traffic along the pedestrian/bike path located in the middle of the Brooklyn Bridge in New York City. Video data were collected at four different locations along the path. The four locations represent two different biking environments. For the first three locations, the bike path was separated from the adjacent pedestrian path using a longitudinal marking line. The fourth location was an access ramp where pedestrians and cyclists share the path without any clear separation between them. Computer vision techniques were applied to extract the trajectories of road users from the video data and automatically obtain bike speed, flow and density. The study provides a distribution for biking speed along the bridge and investigates the effect of different factors such as longitudinal grade, bike path width, and the biking environment on the biking speed. Furthermore, the study investigated the relationship between the bike flow and speed at the three locations where bikes and pedestrians are separated. The adjacent pedestrian flow was found to have a significant effect on biking speed when pedestrian flow exceeds four pedestrians/m/minute. Additionally, the impact of pedestrian flow and bike directional split on bike flow rate was also addressed. At the fourth location, where bikes shared the path with pedestrians, the study developed a relationship between the shared space density and biking speed.
Is That Move Safe?: A Case Study of Cyclist Movements at Intersections with Cycling Discontinuities
Matin Nabavi Niaki, Ecole Polytechnique de MontrealShow Abstract
Nicolas Saunier, Ecole Polytechnique de Montreal
Luis Miranda-Moreno, McGill University
Cyclist safety deals with methods to analyse safety and case studies to better understand the factors that lead to cyclist crashes. Surrogate measures of safety (SMoS) are being used as a proactive approach to identify severe interactions that do not result in an accident and interpreting them for a safety diagnosis. While most cyclist studies adopting SMoS have evaluated interactions by counting the total number of severe events per location, only a few have focused on the interactions between general directions of movement e.g. through cyclists and right turning vehicles. However, road users perform maneuvers that are more varied at a high spatiotemporal resolution such as a range of sharp to wide turning movements. These maneuvers (motion patterns) have not been considered in past studies as a basis for analysis to identify, among a range of possible motion patterns in each direction of travel, which ones are safer and which are more likely to result in a crash. In this study, cyclist motion patterns are obtained from two cycling network discontinuity and two control sites in Montréal. A probabilistic SMoS method is adopted to obtain cyclist-vehicle interactions and compute their time-to-collision. The Kruskal-Wallis and Kolmogorov–Smirnov tests are used to compare the TTC distribution between motion patterns in each site and between sites with and without a discontinuity. Results show that interactions are more severe and less safe, at both locations with a cycling network discontinuity and that cyclists following different movements have statistically different levels of safety.
Exploring the Impacts of Traffic Characteristics on Bicycle–Vehicle Conflicts
Sirisha Kothuri, Portland State UniversityShow Abstract
Edward Smaglik, Northern Arizona University
Brendan Russo, Northern Arizona University
Cristopher Aguilar, Northern Arizona University
Emmanuel James, Northern Arizona University
Nolan Levenson, New York City Department of Transportation
Peter Koonce, City of Portland, Oregon
In urban locations, intersections are areas where a variety of modes converge, thus leading to an increased potential for conflicts. A common crash type involving bicycles at intersections is the “right/left-hook” where a right/left-turning vehicle collides with a through bicyclist. While geometric treatments and signal control strategies have been used to mitigate right-left/hook conflicts, agencies often face questions about optimal treatments and when to use these treatments at intersections. This exploratory study aims to fill that gap by exploring the safety impacts of split LBI and mixing zone treatments to reduce conflicts between bicycles and turning vehicles at intersections using surrogate safety measures with video observations. The surrogate safety measure post encroachment time (PET) was used to classify conflicts by severity based on the magnitude of the value. Next, Poisson and negative binomial models were estimated to observe the relationship between conflicts, bicycle and motor vehicle volumes. The results revealed that through bicycle and turning motor vehicle volumes are significant predictors of right/left-hook conflicts between bicycles and motor vehicles at signalized intersections, which could allow for targeted implementation of safety measures at candidate locations identified through volume analysis. Providing guidance for improving bicycle safety at intersections could increase the attractiveness of this mode for potential new cyclists.
Bicycle Crash Types on National, State, and Local Levels: A New Look
Libby Thomas, University of North Carolina Highway Safety Research CenterShow Abstract View Presentation
Krista Nordback, UNC Highway Safety Research Center
Rebecca Lauren Sanders, Arizona State University
This paper presents an overview of prevalent bicyclist crash types in the U.S. providing insights for practitioners that may be useful in planning safer networks and taking other proactive and risk-based approaches to treatment. The study compares fatal bicyclist crash types from national data with serious injury and all severity bicyclist collisions from the state of North Carolina (NC) and the City of Boulder, Colorado. Overall bicyclist fatalities in the U.S. are more prevalent in urban areas (69%) than rural areas (29%). While the majority of all-severity crashes are at intersections, most fatal and disabling injury bicyclist crashes occur at non-intersection locations, including nearly one-third of bicyclists who died from collisions involving overtaking motorists. Top intersection crash types across national fatal, and all severity crashes in NC and Boulder include bicyclist failed to yield and motorist turning across a bicyclist path. However, many of the top all-severity types in the two jurisdictions differ from the top fatal crash types. These comparisons provide a fresh look at bicyclist crash type trends and have potential importance with respect to planning safer networks for Vision Zero communities, since a key finding is that locations and crash types most prevalent among fatal and serious injuries may differ from the most prevalent types for all severity crashes. Findings could be useful to agencies lacking their own resources for risk-based assessment, but also suggest it may be important to analyze higher severity crash types and jurisdiction-specific data when possible.
Evaluating the Safety Impacts of Green Bike Lanes in Suburban Communities
Jeffrey LaMondia, Auburn UniversityShow Abstract View Presentation
Jacob McGhee, Auburn University
Mitchell Fisher, Auburn University
Fernando Cordero, Auburn University
A study of the effects on bicyclist and driver behavior when painting a bicycle lane green was conducted using video footage along East Thach Avenue in Auburn, Alabama. Whereas most previous studies regarding the implementation of green bicycle lanes focused on primarily urban areas, this study aimed to investigate the implications in a suburban community in east Alabama. Before and after data regarding vehicle speed, vehicle lateral location, and bicyclist lateral location was collected and two sample t-tests were used to analyze the differences between each set of data. Analysis showed that with the addition of green paint to existing bicycle lanes, vehicle speeds dropped off statistically significantly and motorists who were previously giving small amounts of passing space moved further away when passing a bicyclist. In addition, drivers were who were previously driving in or near the bike lane when bicyclists were not present moved further towards the center of their lane. Together, these results indicated that the introduction of green paint to existing bike lanes in a suburban area had positive impacts on bicyclist-vehicle interactions.
Cycling Risks and Needs Perception of Different Skill Levels and Age Groups
Fadi Alhomaidat, Western Michigan UniversityShow Abstract
Valerian Kwigizile, Western Michigan University
Jun-Seok Oh, Western Michigan University
While the percentage of people cycling for transportation rose during the last decade, still only 1% of all U.S. workers use a bicycle for commuting (1). It was found that cyclists’ discomfort regarding the risk of cycling near traffic remains a significant hurdle to foster cycling. This study examines the impact of age, gender, and skill level on the ability to identify hazards/needs while cycling. An online survey was disseminated in May 2015. The survey included three main questions on risk/need factors related to cycling: infrastructure-related, traffic-related, and facility-related risk/need factors. Mean score analysis and an ordered probit model were used to analyze the differences of risk/need perception among bicyclists with different ages, genders, and skill levels. The analysis found potholes, aggressive drivers, and overgrowing vegetation are the main risk/need factors for infrastructure, traffic, and facility, respectively. The analysis also found that there is a relationship between cyclists’ skill levels and given risk/need factors. More intermediate cyclists were influenced by risk/need factors when riding roadways with facility-related issues than experienced cyclists. Beginner cyclists were more likely to be influenced by infrastructure-related risk/need factors than other those at skill levels. This study will help clarify potential issues contributing to understanding the differences in perceived risks/needs among age groups, genders, and skill levels, in an effort to promote cycling safety and particularly in efforts to increase cycling for transportation. Policy-makers and city planners should consider differences in risk/need perception to improve the safety of riding to promote cycling.
Exploring Street Noise and Bicycle Safety: Initial Evidence from Austin, Texas, and the Washington, D.C., Capital Area
Greg Griffin, University of Texas, San AntonioShow Abstract View Presentation
Steve Hankey, Virginia Polytechnic Institute and State University
Ralph Buehler, Virginia Polytechnic Institute and State University
Boya Dai, Texas A&M Transportation Institute
Huyen Le, Virginia Polytechnic Institute and State University
Chris Simek, Texas A&M Transportation Institute
Studies show the relationship of many environmental factors with the safety of bicycle transportation, including street infrastructure, urban densities, safety-in-numbers, and others. However, no work to date includes the relationship of street-level noise with the safety of vulnerable road users. This study deploys bicycle-mounted smartphones with apps recording A-weighted decibels with GPS points to explore this issue for the first time in Austin, Texas, and the Washington, DC Capital Area. Our initial exploration of results shows inconsistent results between the cities, with no direct relationship between street noise and exposure-normalized crash rates. However, when considering infrastructure and nearby bicycle commute share rates with street noise, our model in the Washington, DC Capital Area explained over 87% of the variation in crash risk. This approach to street noise data collection invites other explorations of the relationship of street noise to vulnerable road safety to improve future guidance for transportation planning and engineering.
Drivers’ Behavior When Passing Cyclists Riding in Line in Two-Lane Rural Roads
Pérez-Zuriaga Ana María, Universitat Politècnica de ValènciaShow Abstract View Presentation
Sara Moll, Universitat Politècnica de València
GRISELDA LÓPEZ, Universitat Politècnica de València
Alfredo Garcia, Universitat Politècnica de València
The interaction between motor vehicles and bicycles, especially during passing maneuvers, is one of the main causes of cyclists’ fatalities at two-lane rural roads. This maneuver is even more challenging when the number of cyclists in the group increases. This research studied drivers’ behavior when passing one, two and four cyclists riding in a single file. Data collection was based on four instrumented bicycle riding along four road segments with different traffic and infrastructure characteristics. Collected data were process and 533 overtaking maneuvers were considered in the study. It is mainly focused on motor vehicle speed and lateral clearance, since they are the most influencing variables on overtaking road safety. Observed drivers’ behavior was different when overtaking a cyclist riding alone and when overtaking a group. Their developed speed was higher when overtaking only one cyclist and the lateral clearance was lower. Besides, results showed up that overtaking speed increases when the lane width does and that the painted-shoulder safety countermeasure is not as adequate as expected, since overtaking speeds are higher and lateral clearances lower. These results may be the base for road design improvement considering cyclists presence.
Reexamination of Bicycle/Motor Vehicle Crash Typologies
Timothy Wright, Dunlap and Associates, Inc.Show Abstract View Presentation
Richard Blomberg, Dunlap and Associates, Inc.
Dennis Thomas, Dunlap and Associates, Inc.
Kristie Johnson, National Highway Traffic Safety Administration (NHTSA)
This research involved analyzing police crash reports to examine prevailing types of bicycle/motor vehicle crashes and the behaviors of each operator that precipitate them. By acquiring and typing bicycle/motor vehicle crash reports using an established typology, researchers could determine if the distribution or nature of crashes had changed meaningfully since they were last assessed over 20 years ago by the same methodology. The study acquired a sample of 500 crashes each from California, Florida, Maryland, Minnesota, North Carolina, and Utah to match the sample origin and size of the earlier study, and typed them using Manual Accident Typing (MAT) and Pedestrian and Bicycle Crash Analysis Tool (PBCAT) typing methods. A comparison of the distributions of crash types between the earlier study and the present sample did not uncover any new types beyond the existing schema. Also, the frequencies of the various crash types showed good correspondence between the present and earlier samples. One notable difference involved a reduction in crashes for which inadequate information was available to determine more detail than the basic crash class or group. The greater precision in type determination appeared to be the result of more detailed police crash reports likely due to computerized reporting tools and possibly better law enforcement awareness of the important information to report for a bicycle/motor vehicle crash. Overall, the relative proportions of bicycle/motor vehicle crash types appear to have remained stable, and no new types emerged.
An Investigation into the Varying Effects of Factors Contributing to Injury Severity of Different Bicyclist Age Groups in Bicycle–Vehicle Crashes
Dibakar Saha, Florida Atlantic UniversityShow Abstract View Presentation
Priyanka Alluri, Florida International University
Albert Gan, Florida International University
This study aims to understand the varying effect of factors contributing to the injury severity of bicyclists of different age groups in bicycle-vehicle crashes. An examination of bicyclist crash-related injury by age groups could provide insights on programming more effective educational and safety campaigns focusing on bicyclists of specific age. Using four years of bicycle-vehicle crash data from Florida, injury severity models were developed for four age groups of bicyclists: very young (6-19 years), young (20-44 years), middle-aged (45-64 years), and old (65 years or above). Several crash, geometric, environmental, temporal, vehicle, bicyclist, and driver characteristics were examined. The number of significant variables and their effects on the bicyclist injury severity were different by age groups. The variables, including crash type, lighting condition, vehicle type, driver’s inappropriate action, alcohol and drug influence, and use of safety gear are found to have varying effects on the injury severity levels of different bicyclist age groups. Specific suggestions to implement age-specific safety programs are provided.
Effects of Bicycle Passing Distance Law on Drivers’ Behavior
Ahmad Feizi, Western Michigan UniversityShow Abstract
Majid Mastali, Tetra Tech, Inc.
Ron Van Houten, Western Michigan University
Jun-Seok Oh, Western Michigan University
Valerian Kwigizile, Western Michigan University
This paper identifies the effect of passing distance laws on drivers’ passing behavior and perception using an instrumented bicycle and a driver survey. Bicycle passing was measured in a 25-hour naturalistic field experiment using video recording and an ultrasonic distance measuring device. In order to evaluate the effect of passing distance laws, various jurisdictions with 3-foot passing law, 5-foot passing law, and without a passing law were examined. The experiment required a bicyclist to ride the instrumented bicycle in 2-lane and 3-lane roads to capture the distance from the bicycle to the overtaking motor vehicle. An Ordered Probit model was adopted to describe the relationship between a discrete dependent variable (passing distance) and independent variables. The results demonstrated that overtaking distances in locations with 5-foot passing law were significantly more than other areas. It is also found that roads with paved shoulders, wider travel lanes, and more lanes contributed to greater passing distances. We also found that passing distance was closer on roads with shared lane markings (sharrows) or higher truck concentrations. The survey data collected in locations with different passing laws illustrated that drivers overestimate the distance that they usually pass bicyclist. These results should be useful to transportation engineers, policymakers, and legislators who intend to provide efficient designs of road infrastructure to better accommodate bicycles.
Using Open Data to Assess Cyclist Safety in London
Daniel Collins, Imperial College LondonShow Abstract View Presentation
Daniel Graham, Imperial College London
This study develops a predictive model for cycling collisions in London. Specifically, the effects of bus lanes, parking/loading facilities and multi-lane roads on the risk of cycling collisions are considered. To the best of the authors’ knowledge, this is the first such predictive collision model which develops covariates to measure the characteristics of different types of road infrastructure within zones. A kernel density estimator is used to identify 90 collision hotspots. Each hotspot is populated with information regarding the highway infrastructure within it. A multiple linear regression model tests for the statistical significance of the infrastructure variables. Bus lanes, multi-lane roads and 30mph speed limits are found to impact cycle collision counts while junction density has the largest impact on collision density. 20mph speed limits impact collision counts to a lesser degree than 30mph, indicating potential safety improvement from reducing speed limits. One-way roads are found to reduce the risk of collisions along with the provision of priority junctions. This infers that other junction types, such as roundabouts and signalized junctions, present higher collision risk. The models produce conflicting results on parking/loading provision. The models are expanded to include socio-demographic variables, such as population and employment. The combined model offers no performance improvement over the infrastructure-only model, although a potential link between public transport provision and reducing cycle collisions warrants further investigation.
Analysis of Cyclist Perception and Behavior on Two-Lane Rural Roads Through an Online Survey
GRISELDA LÓPEZ, Universitat Politècnica de ValènciaShow Abstract View Presentation
Igancio Martínez, Universitat Politècnica de València
Francisco Javier Camacho-Torregrosa, Universidad Politecnica de Valencia
Alfredo Garcia, Universitat Politècnica de València
In Spain, many recreational bicyclists use two-lane rural roads for leisure and fitness. This fact produces an impact on road facilities, in terms of both road safety and operation. Knowing the perception and behavior of cyclists on these roads under mixed traffic conditions can provide insight about road performance and safety. This paper describes the methodology and the results of an on-line survey of cyclists at national level. The purpose of the survey was to investigate the effects caused by the presence of cyclists on the roads, both individually and in groups. For this, survey is organized in several sections including general questions, type of user profile, knowledge of the traffic regulation, the perception of safety, possible measures that could improve the roads, and the preferences about specific roadway designs. The survey was responded by 523 cyclists, being most of them users of a racing bike. Lateral distance of 1.5 m is perceived by 72.1% of cyclists as enough. Regarding improvement of a segment, most of cyclists, individually or in groups, marked shoulder with a greater width as the best measure to improve the coexistence of cyclists and drivers. Two-lane rural roads with independent frontage road is the preferred roadway design by cyclists from a safety point of view. Results from this study may inform policy makers on how to successfully, and equitably, integrate cycling with vehicular traffic in two-lane rural roads in Spain. Offering possible solutions in terms of general criteria for planning, design and maintenance of roads.
Proactively Informing Targeted Safety Efforts in San Francisco, California: Developing a Safety Performance Function to Predict Intersection Cyclist Injuries
Megan Wall Shui, City and County of San FranciscoShow Abstract View Presentation
Mia Lei, San Francisco Department of Public Health, Program on Health, Equity and Sustainability
Megan Wier, San Francisco Department of Public Health
Bo Lan, UNC Highway Safety Research Center
Libby Thomas, University of North Carolina Highway Safety Research Center
Jennifer Ziebarth, Fehr & Peers
Rebecca Plevin, University of California, San Francisco
Catherine Juillard, University of California, San Francisco
We aimed to develop a proactive approach to identifying intersections in San Francisco, California at risk for future cyclist injuries. To do this, we developed a safety performance function (SPF) for intersection cyclist injuries. We utilized data from San Francisco’s Transportation-Related Injury Surveillance System (TISS), which combines police reported traffic injuries with hospital recorded injuries, for more comprehensive assessment of transportation-related injuries. Injury data was joined to each intersection, along with independent variables on traffic and cyclist volumes, roadway characteristics, land uses, and concentrations of vulnerable populations. We used Conditional Random Forest analysis to narrow our list of independent variables and negative binomial regression to develop the SPF. Our findings demonstrate positive associations between cyclist injury and bicycle and motor vehicle volumes, bike facilities, higher lane counts, and denser poverty rates in the surrounding area. The SPF was also used to rank intersections based on SPF predicted injuries, empirical-Bayes (EB) estimated injuries, and potential for safety improvement (PSI). We found that nearly all of our top 100 intersection locations, ranked by SPF, EB, or PSI, fell on our Vision Zero High Injury Network (HIN), presently used as our main tool for safety improvement prioritization, or on the bike network. The predicted locations thus provide an opportunity to prioritize safety improvements on and off the HIN. In addition to the application of model results using current conditions, the SPF could be used to predict injuries for future conditions, when model parameters are estimated as a part of transportation planning analyses.
What Are Vulnerable Road Users’ Perceptions and Expectations on Autonomous Vehicles?
Praveena Penmetsa, Alabama Transportation InstituteShow Abstract View Presentation
Emmanuel Adanu, University of Alabama
Dustin Wood, University of Alabama
Teng Wang, Texas A&M University
Steven Jones, University of Alabama
Public perceptions play a crucial role in wider adoption of Autonomous Vehicles (AVs). This paper aims to make two contributions to the understanding of public attitudes toward AVs. First, we explore opinions regarding the perceived benefits and challenges of AVs among vulnerable road users – in particular, pedestrians and bicyclists. Second, the paper evaluated whether interaction experiences with AVs influence perceptions among vulnerable road users. To explore this, we examined survey data collected by Bike PGH, a Pittsburgh based organization involved in programs to promote safe mobility options for road users. Analysis of the data revealed that respondents with direct experience interacting with AVs reported significantly higher expectations of the safety benefits of the transition to AVs than respondents with no AV interaction experience. This finding did not differ across pedestrian and bicyclist respondents. The results of this study indicate that as the public increasingly interacts with AVs, their attitudes toward the technology is more likely to be positive. Thus, this study recommends that policy makers should provide the opportunities for the public to have interaction experience with AVs. The opportunities can be provided through legislation that allows auto manufacturers and technology industries to operate and test AVs on public roads. This interactive experience will positively affect people’s perceptions and help in wider adoption of AV technology.
How Technology Can Affect the Demand for Bicycle Transportation: The State of Technology and Projected Applications of Connected Bicycles
John MacArthur, Portland State UniversityShow Abstract View Presentation
Michael Harpool, Portland State University
The term “connected vehicle (CV)” refers to vehicles equipped with devices which enable wireless communication between internal and external entities, supporting vehicle-to-vehicle (V2V), vehicle-to-infrastructure (V2I) and vehicle-to-everything (V2X) communications. The widespread deployment of CVs will address a range transportation challenges related to safety, mobility, and sustainability. Recent research efforts on connected bicycles have focused on the uses and limitations of the state-of-the-art technologies, safety implications, the reliability of various communication modes, and consumer adoption. Existing research focuses on either technologies that utilize data received from sensors and the internet to govern devices attached to the bicycle (situational sensing) or two-way communication. While there has been some mention of how these technologies may encourage an increase in bicycling through enhanced safety, the research is sparse and there is a lack of discussion on how connected bicycles can address other barriers to bicycling. This paper will provide context into the societal needs of bicycling and the current strategies utilized to increase the bicycle mode share, a cohesive review of existing and prototyped connected bicycle technologies and discuss their potential to mitigate barriers to bicycling and better accommodate the needs and desires of diverse riders. We will then explore the limitations and benefits of one-way and two-way communications, the potential of bicycle-to-infrastructure technologies, and the future needs and expected pathways of connected bicycle technologies.
Exploring Associations Between Non-Motorized Traffic and Episodic Area-Wide Air Pollution in Northern Utah
Patrick Singleton, Utah State UniversityShow Abstract
Curtis Knight, Utah State University
Dayton Crites, Cache County
Despite well-known linkages between transportation and air pollution, there is relatively little evidence of the impact of episodes of poor air quality on travel behaviors, particularly the use of non-motorized modes. Air pollution mitigation policies that rely on mode shifts may be ineffective during these episodes given personal conflicts between altruism (using a non-motorized mode) and risk aversion (reducing one’s exposure to air pollutants). This study examines associations of air pollution (PM2.5 and ground-level ozone concentrations) and weather (precipitation, temperature) with non-motorized traffic counts, while controlling for seasonal and weekly temporal variations, using 18 months of daily time series data in Logan, UT. Higher concentrations of PM2.5 were associated with reductions in non-motorized counts, but this association was only large and significant at unhealthy levels (air quality index > 150), when around a 50% reduction in non-motorized traffic was expected. This reduction was greater than the impact of a quarter inch of precipitation and slightly more than the impact of temperatures less than 10°F. No association was found for ozone, but concentrations were comparably low during the study period. The study’s key finding—severe air pollution deters walking and bicycling—offers implications for public health analyses and transportation strategies during such events. Future work should examine the generalizability of this result in other contexts and investigate how and why such travel behavior changes occur.