Safety evaluation of design alternatives to separate freeway-opposing traffic

Objectives Different design alternatives used to separate freeway-opposing traffic may produce varying safety levels. State-of-the-art design guidelines have provided guidance in respect to a number of median features such as width, slope, and barrier installation. Guidance provided has been based on benefit-cost procedures; however, because project costs may significantly vary across different countries/jurisdictions, findings from benefit-cost procedures may be meaningless on a broader geographical scope. The objective of this study is to quantify the safety level of a number of design alternatives commonly used to separate opposing freeway traffic based solely on annualized crash costs. Methods The safety performance of 14 design alternatives were assessed using the Roadside Safety Analysis Program (RSAPv3). Results A test-level 3, low-tension cable barrier installed in the middle of a wide, hazard-free median was found to be the safest design alternative. Road sections containing: i) cable barriers were found to be safer than those containing w-beam guardrails, ii) low-tension cable barriers installed in the middle of the median were found to be safer than those containing high-tension cable barriers installed on one side of the median, regardless of median width, iii) wide, hazard-free medians were found to be safer than those containing WB guardrails, and iv) concrete barrier installations were found to be only safer than those containing unshielded medians, except when the median was 20 meters wide and hazard-free. Conclusions This paper not only provides evidence that some of the design alternatives often installed in real-world scenarios are not the safest, but it also discusses how some of these alternatives may not be the most cost-effective either. These findings make this study relevant and timely, as designers and policy/decision makers should always seek to maximize safety while optimizing the allocation of limited public funds.