Joint Modeling of Traffic Incident Duration Components (Reporting, Response, and Clearance Time): A Copula-Based Approach

Abstract

The current study develops a trivariate framework that accommodates for inherent dependencies across the various components of incident duration. A unique ordered response model structure (sometimes referred to as grouped ordered response model) is introduced for modeling three durations—reporting time, response time, and clearance time—in this study. Further, as opposed to employing a simulation oriented multivariate model approach, the study proposes and estimates a copula-based methodology that allows for a closed-form probability computation. The approach is the first application of this model for incident duration analysis. The proposed copula framework is estimated to identify factors affecting incident duration components from a host of characteristics including incident characteristics, traffic conditions, roadway, and environmental characteristics. The data for the analysis are obtained from the 2015 Florida Department of Transportation incident database. The model estimates were also augmented by conducting policy analysis by generating a three-dimensional representation of incident frequencies as a function of reporting, response, and clearance time.