SAFETY EVALUATION OF TWO-LANE TO FOUR-LANE CONVERSIONS IN WISCONSIN
Abstract
For this study, 12 two-lane to four-lane conversions constructed in Wisconsin within the last decade were analyzed. Five years of before crash data and a range of one to five years of after crash data (depending on the construction year) were collected, as well as geometric and volume data. A simple before-and-after analysis was completed to analyze specific types of injury crashes and manners of collision for each two-lane to four-lane conversion. An Empirical Bayes (EB) analysis was also used to determine the expected average crash volume per year to evaluate the safety benefits of the conversion.
Using a simple before-and-after crash analysis, average number of crashes per year was found to have reduced by between 7 to 82 % after the two-lane to four-lane conversion for all 12 projects. Five out of the 12 project sites reviewed reduced in all types of injury crashes and property damage crashes after the conversion. Three additional projects reduced in all types of injury crashes and property damage crashes after the conversion except in fatal injury in which there were no recorded crashes before or after the conversion. The remaining four projects showed an increase in fatal, incapacitating, non-incapacitating, or possible injury crashes per year. All projects, however, reduced in property damage only crashes per year. In all manners of collision considered, three projects showed a reduction in average crash volume per year. One additional project reduced in all types of injury crashes and property damage crashes after the conversion except in head on collision and sideswipe in the same direction crashes in which the recorded crashes before and after the conversion remained unchanged. The remaining eight projects showed an increase in angle, head on collision, no collision with another vehicle, rear end, or sideswipe in the same direction crashes per year. However, all projects reduced in sideswipe in the opposite direction crashes per year. The increase in crashes usually occurred at specific intersections within a project within a particular year and therefore was not thought to reflect the effect of the two-lane to four-lane conversion on the safety of the entire roadway section. Also, as four out the 12 projects had less than five years of comparable after conversion crash data, the percent increase or decrease of average crash volume per year may be different for those four projects if five years of after conversion crash data is later used.
EB analysis was performed using the Highway Safety Manual (HSM) Rural 2-Lane Road spreadsheet to compare the observed average number of crashes per year to the expected average number of crashes per year if the two-lane to four-lane conversion had not been implemented. The expected average crash volume per year accounted for changes in traffic volumes between the before and after conversion periods. The HSM spreadsheet was populated with geometric and crash data before the conversion, and traffic volume data after the conversion. Without the conversion, expected average number of crashes per year increased as traffic volume increased. Therefore, when compared to the observed average number of crashes per year after the conversion, all the projects showed that expected average number of crashes per year was higher. The EB analysis proved that the two-lane to four-lane conversion resulted in a reduction in average number of crashes per year of between 10 and 85 %. One out of the 12 projects showed that expected average number of crashes per year still reduced without the conversion but a further reduction of approximately 10 to 11 % was obtained after the conversion. Effects of geometric considerations such as lane width, shoulder width and type, percent grade, intersection type, and number of turn lanes on safety from the EB analysis proved to be difficult to accurately determine as lane width, shoulder width and shoulder type remained unchanged within each project and so there was no data within the project to make the comparison while percent grade, intersection type, and number of turn lanes changed constantly with traffic volume within each project and so it was difficult to isolate the effect of the individual geometric characteristic on safety. However, percent grade was observed to have minimal effect on the expected average number of crashes per year while intersections with higher number of turn lanes, and signalized or four-leg stop control types had higher average number of crashes per year.
Overall, conclusion from both analyses was that two-lane to four-lane conversions are a safety benefit and result in average number of crashes per year reduction of as much as 85 %.