Calibration of Highway Safety Manual Predictive Methods for Rural Local Roads

Xiao Qin, Zhi Chen, Kimberly VachalTransportation 2014

Abstract

Establishing performance-based safety goals and objectives becomes more attainable with the Highway Safety Manual (HSM). However, the safety performance functions (SPFs) in the HSM may not be accurate as they are not calibrated to local conditions. In addition, each SPF and crash modification factor (CMF) assumes a set of base site conditions which may not be realistic for local roadways. Although calibration procedures are available in HSM Part C Appendix A, they should be refined or modified to accommodate local data availability and roadway, traffic, and crash characteristics. It is also necessary to determine a set of base conditions applicable to local highways. This paper presents the application of the HSM for rural local two-lane two-way highway segments in South Dakota (SD). The calibration was based on three-year (2009-2011) crash data from 657 roadway segments constituting more than 750 miles of roadways. The calibration process includes establishing new base conditions, developing SPFs, converting CMFs to base conditions as well as substituting default values with state-specific values. Five models have been developed and compared based on statistical goodness-of-fit and calibration factors. Results show that the jurisdiction-specific crash type distribution for CMFs can be drastically different from what is presented in the HSM. The HSM method without modification underestimates SD crashes by 35 percent. The method based on SPFs developed from a full model has the best performance. This study provides important guidance and empirical results regarding how to calibrate HSM models.
BibTeX
@inproceedings{qin2014calibration,
  author = {Xiao Qin and Zhi Chen and Kimberly Vachal},
  title = {{Calibration of Highway Safety Manual Predictive Methods for Rural Local Roads}},
  booktitle = {Transportation},
  year = {2014},
}

Loading related content...

Figure