Repository logo

Development of Safety Performance Functions for High-Speed Roadways in Saskatchewan


Saskatchewan experienced the highest rate of collision-related fatalities among all the Canadian provinces and territories in 2011. To implement measures to reduce collisions on roadways, highway safety professionals require a tool to quantify road safety. There are currently no safety performance functions (SPFs) for high-speed roadways in Saskatchewan, which limits the ability of highway safety professional to quantify and evaluate safety as part of the decision-making process. SPFs are an important component of the Highway Safety Manual’s (HSM 2010) systematic safety management process, which evaluates the safety of roadways and the effectiveness of roadway safety improvements. SPFs are mathematical models that relate roadway features, such as traffic volume, geometry, and traffic control, to the observed collisions. They are validated by performing regression analyses. The objective of this study was to use local roadway features and collision data (2007 to 2011) to develop a set of local SPFs to predict the number of collisions on Saskatchewan high-speed roadways including interchanges. GIS maps were used to obtain roadway configuration data (basic freeway segments, ramps, weaving sections, etc.), which were correlated to existing collision data for those same roadways. The negative binomial (NB) distribution was used to derive the SPFs written in the programming language R. Statistical goodness of fit (GOF) tests were performed to identify the best-fitting SPFs, and the study produced 24 statistically significant models (eight roadway configurations and three levels of collision severity). The SPFs developed in this study can play a vital role in improving the planning and decision-making processes in roadway safety engineering and help to reduce the number of fatalities on roadways in Saskatchewan.



Highway safety, Collision prediction models, Safety performance functions (SPFs), High-speed roadways, Interchanges, Freeways between interchanges, Freeways within interchange system, Ramp influence areas, Weaving Sections, Ramps, Ramp terminals, Negative Binomial model.



Master of Science (M.Sc.)


Civil and Geological Engineering


Civil Engineering


Part Of