Estimation of Safety Effectiveness of Composite Shoulders on Rural Two-Lane Highways

Abstract

A paved shoulder is regarded as an effective safety improvement to reduce crashes. It is believed that there is a diminishing safety benefit for each additional increment of paved shoulder over a certain width. Thus, there might be greater systemwide safety benefits from paving longer roadway segments with a composite shoulder than from paving shorter roadway segments with a full-width paved shoulder. The objective of this study was to determine the safety benefits of composite shoulders, such as a small paved shoulder combined with turf on the outside. This approach was part of the Kansas Department of Transportation's effort to find practical improvements to maximize the benefits of shoulders relative to the input costs. Of the 8,300 mi (13,358 km) of rural two-lane highways in Kansas, approximately 25% are equipped with composite shoulders consisting of 3 ft (0.9 m) of pavement, with the remainder being turf. The safety effectiveness of these shoulders was studied with a combined empirical Bayes method. Three safety performance functions were used to create Kansas-specific crash modification factors for composite shoulders compared with segments with no or unpaved shoulders. It was found that upgrading narrow unpaved shoulders to composite shoulders could reduce shoulder-related crashes by up to 61% and fatal and injury crashes by 31%. On the basis of these results, 20-year projections were developed that estimated the safety effectiveness that could be achieved through the implementation of these safety improvements.