Linear and nonlinear multivariate statistical analyses are applied to determine how the types of accidents that occur on heavily used freeways in Southern California are related to both the flow of traffic and weather and ambient lighting conditions. Traffic flow is measured in terms of time series of 30-second observations from inductive loop detectors in the vicinity of the accident prior to the time of its occurrence. Results indicate that the type of collision is strongly related to median traffic speed and to temporal variations in speed in the left and interior lanes. Hit-object collisions and collisions involving multiple vehicles that are associated with lane-change manoeuvres are more likely to occur on wet roads, while rear-end collisions are more likely to occur on dry roads during daylight. Controlling for weather and lighting conditions, there is evidence that accident severity is influenced more by volume than by speed. (Author/publisher)
Abstract