Investigating Motorcycle Crash Injury Severity along Kentucky’s Rural Two-Lane Roads Using Microscopic Detailed Real-Time Weather Data

Abstract

This study investigated factors influencing motorcycle crash injury severity along Kentucky’s rural two-lane undivided roadways using motorcycle-related crashes from 2019 to 2022 and roadway information from the Kentucky Transportation Cabinet (KYTC). Additionally, this study leveraged microscopic real-time weather data during one hour of the crash using the High-Resolution Rapid Refresh (HRRR) model [retrieved from the National Oceanic and Atmospheric Administration (NOAA)] to examine the impact of detailed real-time weather variables on motorcycle crash severity, which were not thoroughly explored in real-time in previous motorcycle studies. The mixed logit model (MXL) and its extensions, including MXL with heterogeneity in means (MXLHM) and correlated MXL with heterogeneity in means (CMXLHM), were employed to account for unobserved heterogeneity. The CMXLHM model provided a better fit and effectively captured unobserved heterogeneity compared to the other two models. The model results showed that the weather states of solar radiation ≤ 5    watts / m 2 , air temperature ≥ 70 ° F , and relative humidity ≤ 65 % increased the likelihood of severe motorcycle injury outcomes by 8.24%, 15.96%, and 26.14%, respectively. Furthermore, at-fault driver age and head-on and angle collisions were associated with increased likelihood of severe motorcycle injury outcomes, while higher hourly traffic volume was associated with reduced severe injury likelihood. Several recommendations were proposed to enhance motorcyclists’ safety along rural two-lane roadways, e.g., displaying real-time safety alert messages using dynamic message signs (DMSs) at high motorcycle crash locations whenever air temperature ≥ 70°F or relative humidity ≤ 65%.