Joint effects of access and geometry on two-lane rural highway safety in British Columbia

Abstract

This paper examines the safety impact of access type, access density, traffic volume, and road geometry on two-lane rural highways in British Columbia. Documented accident to access relationships were reviewed and found to be in poor agreement with observed accident experience within the Province. The nontransferability of accident to access relationships, the lack of studies of traffic and road geometry interaction effects, and the need for a tool to predict the impact of access on highway safety led to the development of the models described in this paper. A comprehensive and disaggregate data base was constructed for some 560 km of the provincial primary arterial highway network. Highway access was classified into four groups: public road intersection, business access, private access, and roadside pullout. Various geometric data, traffic volume, and posted speed limit were included in the data base. Different measures of accidents were chosen to develop the models, including accident rate, accident frequency, severe (fatal plus injury) accident rate, and severe accident frequency. The statistical analyses indicated that when viewed as a disaggregated set of multivariate elements, all access types were significantly correlated with accidents. The joint effects of access and geometry, such as the combination of private access and horizontal curvature, intensify the effect of access on accidents. As a result, relative weightings of all access types in terms of their impact on safety were empirically derived, with business access weighted as 0.5 and private access weighted as 0.10 of public road intersections. Key words: highway access, geometry, accidents, rural highways, two-lane highways, multivariate regression.