Crash Modification Functions for Passing Relief Lanes on Two-Lane Rural Roads. Issue 9 (September 2020)

• Record Type: Journal Article
• Authors: Persaud, Bhagwant; Anarkooli, Alireza Jafari; Almasi, Shahram; Lyon, Craig
• Is Part Of: Transportation research record. Volume 2674:Issue 9(2020)
• Page Start: 586
• Page End: 592
• Publication Date: 2020-09
• Subjects: Transportation -- Periodicals; Roads; Transport -- Périodiques; Routes -- Périodiques; Routes -- Conception et construction -- Périodiques; Roads; Transportation; 388.05
• DOI: 10.1177/0361198120930719
• Languages: English
• ISSNs: 0361-1981

Abstract

Passing relief lanes on two-lane rural roads provide passing opportunities that would otherwise be scarce where there are extensive no-passing zones, high opposing traffic volumes, or both. This paper addresses the safety effects of installing a passing lane or lengthening an existing one. It stands to reason that the effect of installing a passing lane will depend on the actual length of that lane. By extension, it is also reasonable to expect that the safety effects of lengthening an existing one will depend not only on the amount of the lengthening, but also on the original length. Yet, knowledge that can be applied to estimate these two sets of effects in a design process is lacking. The crash modification factors (CMFs) in the Highway Safety Manual (HSM) and in the CMF Clearinghouse for installing a passing lane are all single-valued, of the order of 0.75. And neither source provides CMFs for lengthening an existing passing lane. This paper seeks to address these voids by developing continuous crash modification functions (CMFunctions) for both sets of design decisions using Michigan, U.S., and Ontario, Canada, crash, geometric, and traffic data for passing lane and reference sections. Generalized linear modeling and full Bayes Markov Chain Monte Carlo (FB MCMC) simulation are used to develop cross-section regression models from which crash modification functions are derived and compared. The results are consistent with those from credible before-after studies, so are recommended for implementation in practice, in particular for HSM applications.