This report describes a comprehensive study that aims to identify ways to reduce the frequency and severity of bicycle crashes in Florida. The objective is achieved through a detailed analysis of the roadway, behavioral, and spatial factors associated with bicycle crashes. An extensive literature review was first conducted. The review focuses on the methods to identify bicycle hot spots and findings on bicycle crash causes, crash contributing factors, and potential countermeasures. A descriptive trend analysis was then performed based on a total of 26,036 bicycle crashes that occurred during 2011-2014. A spatial analysis using ArcGIS was then performed to identify the top five bicycle crash hot spots in each Florida Department of Transportation (FDOT) district. These hot spots together experienced a total of 2,954 bicycle crashes during the four-year analysis period. Police reports of these bicycle crashes were reviewed in detail to identify specific bicycle crash types, their crash contributing factors and potential countermeasures. Macroscopic spatial analysis was performed to model the relation between demographic, socio-economic, roadway, traffic, and bicycle activity data at the census block group level and bicycle crash frequencies in Florida. Finally, a cross-sectional analysis was performed to develop Florida-specific Crash Modification Factors (CMFs) for bicycle crashes for different roadway segment and intersection facility types. The review summarized existing studies in the following four areas: (1) risk factors that affect the frequency and severity of bicycle crashes; (2) bicycle crash causes, patterns, and contributing factors; (3) network screening methods used to identify and prioritize bicycle hot spots; and (4) safety performance of the most commonly implemented engineering countermeasures. Researchers preferred to differentiate the risk factors affecting bicycle safety for intersections and mid-block locations due to the obvious variability in the operational characteristics. Roadway traffic, geometric, and socio-economic variables were investigated to determine their impact on bicycle crash frequency and severity. Spatial analysis, especially the use of ArcGIS, has evolved as an effective tool to better understand and model bicycle crash frequencies. Moreover, spatial analysis using ArcGIS was found to be the most commonly used network screening approach. Several studies, however, used a combination of different methods to identify and rank bicycle high crash locations. In addition to the typical bicycle infrastructure such as bicycle lanes and bicycle slots, researchers have investigated the impact of several other roadway characteristics, including shared path width and separation, shoulder type, shoulder width, etc., on bicycle safety. One of the main challenges observed in improving bicycle safety is the lack of bicycle exposure data. Unlike traffic volumes, bicycle volumes are scarcely available, if at all. Researchers addressed this limitation by using surrogate measures of bicycle exposure such as number of transit stops in a region, population, etc. Statewide bicycle crash patterns and causes were identified based on a total of 26,036 bicycle crashes that occurred during 2011-2014. The descriptive trend analysis was based on temporal, environmental, bicyclist-related, crash location-related, and vehicle-related factors. The effect of roadway geometric features on the frequency and severity of bicycle crashes was also studied using data from 9,884.3 miles of non-limited-access state roads in Florida, which experienced a total of 10,546 bicycle crashes during the four-year analysis period. Some of the key findings include: * Bicycle fatal crashes accounted for 5.6% of all traffic fatal crashes, while they constituted only 1.9% of total crashes. * The majority of bicycle crashes occurred on urban roadways; only 1.2% of all crashes that occurred on state roads occurred in rural areas. In terms of crash severity, 16.9% of all bicycle crashes that occurred on rural facilities resulted in fatalities while only 2.5% of those that occurred on urban facilities resulted in fatalities. * Nighttime bicycle crashes resulted in more fatalities compared to daytime crashes. * Crashes involving elder bicyclists ([older or equal] 65 years) resulted in more fatalities compared to crashes involving younger bicyclists (< 65 years). * Crashes involving male bicyclists resulted in more fatalities compared to crashes involving female bicyclists. * Over 10% of all bicyclists involved in crashes who were under the influence of alcohol were killed, and a high 27.6% of all bicyclists involved in crashes who were under the influence of drugs were killed. * Crashes involving bicyclists using helmets or protective pads were less severe compared to those involving bicyclists using reflective clothing or lighting. * Although bicyclists were frequently hit while cycling on the sidewalk, these crashes resulted in very few fatalities. * Crashes involving bicyclists cycling along the roadway against traffic were found to be more severe compared to those involving bicyclists cycling along the roadway with traffic. * In terms of bicyclist’s action at the time of the crash, failure to yield right-of-way was the most frequent contributing cause, resulting in about 15% of total crashes. * Among all types of vehicles, passenger cars were found to result in relatively less severe crashes. Medium and heavy trucks resulted in more severe crashes; a relatively high 14.5% of all crashes involving medium and heavy trucks were fatal. A spatial analysis using ArcGIS was performed to identify the top five bicycle hot spots in each FDOT district. Police reports of all the 2,954 bicycle crashes that occurred at these hotspots were reviewed in detail to identify specific bicycle crash types and patterns. Some of the key findings from the police report review include: * Drivers were at-fault in 45.7% of the crashes, while bicyclists were at-fault in 30.2% of the crashes. * Crashes involving at-fault bicyclists resulted in a greater percentage of fatal crashes compared to those involving at-fault drivers. * Signalized intersections experienced a greater proportion of bicycle crashes compared to unsignalized locations. * Locations with bicycle lanes experienced a smaller proportion of fatal crashes compared to locations without bicycle lanes. * Crossing the street was found to result in a greater proportion of fatal crashes compared to riding along the roadway. * Crashes involving bicyclists riding along the roadway facing traffic resulted in a greater proportion of fatal crashes compared to crashes involving bicyclists riding along with vehicles. * Crosswalk locations, although experienced a high frequency of bicycle crashes, experienced a relatively low proportion of fatal crashes. The crash pattern analysis identified the following four major bicycle crash types: * Motorist turns right while bicyclist is crossing the street * Motorist turns left facing bicyclist * Bicyclist rides out at intersection * Motorist drives out at stop sign In addition to these crash types, the following bicycle crash contributing factors and scenarios were also observed frequently: * Inadequate street lighting * Unconventional intersection geometry * Traffic violations by motorists and bicyclists * Bicyclists sideswipe vehicles * Driveways near intersections * U-turn manoeuvres by bicyclists and motorists * Bicyclists hit the door of parked vehicle * Bicyclists ride opposite to the traffic Several engineering and education countermeasures were recommended for these crash types and scenarios. Engineering countermeasures, including signal optimization, turn restrictions, and sign and pavement marking improvements, could improve the overall safety situation for bicyclists. Agency-wide education campaigns on the laws pertaining to bicyclists and extensive driver education campaigns that focus on driver compliance with bicyclist right-of-way laws and stricter enforcement could improve bicycle safety. Bicycle crash trends are quite distinctive and are dependent on land use, existing bicycle infrastructure, socio-economic factors, etc. The impact of these factors on bicycle crash frequencies was therefore studied using spatial analysis. A macro-level spatial analysis was performed to determine the relation between bicycle crashes and independent variables, including demographic and socio-economic factors, roadway and traffic characteristics, and bicycle activity, while accounting for the effect of spatial correlation among census block groups. Separate models were developed for total and F+S bicycle crashes. (Author/publisher)
Samenvatting