Injury Severity of Multivehicle Crash in Rainy Weather

Abstract

As part of the Wisconsin road weather safety initiative, the objective of this study was to microscopically assess the factor effects on the severities of multivehicle-involved crashes on high-speed roadways during rainfall utilizing a sequential logistic regression approach. Research began by considering interstate freeways in Wisconsin. Weather-related factors considered in the research included estimated rainfall intensity, water film depth, temperature, wind speed and direction, and the car-following distance at the time of crash. With each crash observation, weather data were obtained through the three most adjacent weather station locations and the inverse-squared distance method. Nonweather factors such as roadway geometries, traffic conditions, collision manners, vehicle types, and driver and temporal attributes were also considered. Sequential logistic regression was applied to predict multivehicle crash severities in ascending (forward) and descending (backward) orders, respectively. The final model was selected on the basis of a combination of model performance, parameter significance, and prediction accuracies. The backward sequential logistic regression model produced the most desirable results for predicting crash severities in rainy weather in which deficiency of car following, wind speed, the first harmful spot, vehicle types, temporal, and at-fault driver-related actions at the crash moment were found to be statistically significant. These findings can be used to provide quantitative support of road weather safety improvements via weather warning systems, highway infrastructure enhancements, and traffic management.