Modeling crash frequency and severity using historical traffic and weather data: truck involved crashes on I-80 in Iowa

As congestion grows along roadways in the country, it is important to see how this will affect crashes on America’s highways. I-80 in Iowa is a major trucking corridor for transferring goods between the east and west coasts and carries an increasing volume of freight trucks on the road. The recent ability to record detailed speed and volume data over Iowa’s road system presents a new opportunity to examine whether congestion and slowdown affect the occurrence and severity of crashes along I-80. This study examines the use of INRIX speed data, Wavetronix radar data and RoadWeather Information Systems [RWIS] data on I-80 in Iowa to model freight truck crashes. A random-parameter Poisson regression model is used to examine how speed, weather and roadway characteristics affect the frequency of crashes along different segments. An ordered probit model examines how these factors affect the severity of injuries in truck crashes. In general, lower speeds and congestion were associated with more frequent crashes (taking into account the vehicle-miles travelled) of lower severity. High speed, low congestion periods are more often associated with fewer, but more severe, crashes.