Spatiotemporal characteristics of traffic accidents on a high-volume expressway with closely spaced interchanges

OBJECTIVES: The rapid saturation of expressway capacity in economically developed city clusters has intensified traffic conflicts and accident risks, yet accidents on high-volume expressways with closely spaced interchanges remain understudied. This study aims to clarify the spatiotemporal distribution patterns of such accidents and to identify the locations and characteristics of accident-prone segments. METHODS: A total of 30,265 valid records from the Changhu Expressway in Dongguan, China, covering January 2019 to December 2023, were analyzed; these comprised 19,315 traffic accidents and 10,950 traffic incidents. Descriptive statistics were used to examine accident types and involved vehicle types by year and 15-min interval. Pearson correlation analysis assessed contemporaneous associations between accident counts and high-flow-event frequency, and a spatiotemporal overlap-rate indicator was used to identify accident-prone segments. RESULTS: Rear-end collisions were the predominant accident type (85.3%), with trucks disproportionately involved during 02:00-04:00. Accidents showed four daily peaks, whereas high-flow events showed two. High-flow-event peaks preceded the morning accident peak, whereas accident peaks preceded high-flow-event peaks in the afternoon. The ends of acceleration lanes, areas preceding divergence points, downhill sections, and continuous-gradient segments contained notable accident concentrations. Using the spatiotemporal overlap rate, 14 accident-prone segments were identified. CONCLUSIONS: On the studied high-volume expressway, traffic accidents showed pronounced spatiotemporal clustering. Accident counts were positively associated with high-flow-event frequency in most analyzed direction-period combinations, while interchange weaving areas and variable-gradient sections contained notable accident concentrations. The findings may inform time- and location-specific crash-prevention measures on similar facilities, but their transferability should be evaluated using exposure and severity data from additional corridors.