Road Lighting Effects on Bicycle and Pedestrian Accident Frequency: Case Study in Montreal, Quebec, Canada

• Published in: Transportation research record 2016, Vol.2555 (1), p.86-94
• Main Authors: Nabavi Niaki, Matin S., Fu, Ting, Saunier, Nicolas, Miranda-Moreno, Luis F., Amador, Luis, Bruneau, Jean-François
• Format: Article
• Language: English
• ISSN: 0361-1981; 2169-4052
• DOI: 10.3141/2555-12

Although vehicle, bicycle, and pedestrian flows are considerably lower in general during the nighttime, a higher number of accidents than expected occur during this time. A highly influential factor is the lack of visibility at nighttime. Several studies have shown the negative effects of the lack of visibility on bicycle and pedestrian accident frequency and injury severity at nighttime. However, these studies considered only the presence or absence of light, which was not sufficient to evaluate road user safety. Only a limited number of studies in this field actually measured nighttime road illuminance levels. This study relied on the collection of road illuminance data on road links during the nighttime in downtown Montreal, Quebec, Canada, through the use of an illuminance sensor mounted on a scooter. Pedestrian and bicycle accident frequencies were analyzed separately with the use of the negative binomial model. Unexpectedly, the result showed that an increase in road lighting was associated with more bicycle and pedestrian accidents, which might have been explained by the decision to add or increase the amount of lighting at locations in which accidents occurred. The presence of a bike facility and arterial roads was associated with a decrease in bicycle accident occurrence. For pedestrians, the number of lanes per link and the pedestrian flow were associated with an increase in nighttime accident frequency, while the vehicle flow was associated with a decreasing number of accidents. The study called for more investigation of the precise relationship between safety and the amount of light provided by road lighting.