Diagnosing the Effects of Weather Factors on Cold Region Highway Traffic Operation with Two Layers of Nonparametric Statistical Analysis

Abstract

Highway agencies use significant financial resources to operate traffic monitoring programs. Conventional traffic monitoring practices are operated only during weather-permissible periods, leading to inefficient traffic monitoring caused by restricted sample sizes. None of the previous studies considered the possibility of executing a traffic monitoring program throughout the entire year including wsevere winter weather conditions. The objective of this research was to diagnose the influence of snowfall and temperature on traffic volume changes according to vehicle class by employing two phases of statistical approaches and to propose unique traffic investigation programs for a different type of vehicle. This research used huge vehicular records archived for 6 years at the weigh-in-motion site operated in Leduc City, Alberta, Canada. Three vehicle classes classified from the raw data were used in the analysis. Three classes were total traffic, passenger cars, and truck traffic. Both layers of the statistical approach reported that passenger vehicles are impacted by climate elements. However, the truck was not significantly affected by snowfall and temperature. The results of this investigation will support highway agencies in developing evidence-based programs for traffic counting and road maintenance in winter.