Safety Effect of U-Turn Conversions in Tehran: Empirical Bayes Observational Before-and-After Study and Crash Prediction Models

Abstract

Directional midblock median openings with left-turn lanes and loons were installed on major highways in Tehran, Iran, to prohibit direct left turns from major and minor roadways and to prevent through movements of minor roadways. Therefore, traffic lights were removed and intersections were converted to U-turns. The safety effect of this treatment was estimated by an empirical Bayesian method that indicated there was an increase of approximately 13.22% in crashes. Moreover, accident prediction models for crash types and frequency were developed to examine the geometry and traffic factors contributing to the crashes at U-turn conversions and to explain the reason some sites showed a decrease in the number of crashes while others showed a large increase. The results of data analysis showed that the increase in weaving length and median opening length resulted in the decrease in angle crashes. Also, the increase in the turn radius resulted in the decrease in angle and total crashes. Total crashes and crash types increased with the average daily traffic (ADT) increase on major roadways and U-turns. Consequently, the result of safety analysis with close examination of each site demonstrated that it is necessary to mention geometric specifications in order to improve the efficiency of the treatment. Therefore, it can be concluded with no reasonable doubt that with a certain range of traffic volume and right geometric specification, the treatment may be safe.