Presentation Information

[AI-09]Traffic Impact Assessment of Earthquake-Induced Pavement Damage Using Connected Car Data and Machine Learning: The Case of the 2024 Noto Peninsula Earthquake

*Saki Tabata1, Takumi Asada1 (1. Muroran Institute of Technology (Japan))

Keywords:

Post-disaster restoration prioritization,connected car data,Spatiotemporal analysis,Travel speed,Wide-area earthquake impacts

A major earthquake struck Japan’s Noto Peninsula on January 1, 2024, severely affecting a wide area. The event triggered extensive pavement distress, including cracks and step-like deformations, which substantially disrupted general traffic, snow-removal operations, and recovery activities. In recent years, leveraging digital technologies and assessing societal impacts have become urgent challenges in earthquake disaster response; however, comparatively limited attention has been paid to pavement damage restoration and its mobility impacts. To address this gap, we conducted a spatiotemporal analysis of pavement damage and an evaluation of traffic impacts in the affected region using connected vehicle data that include both travel speed and road surface condition information.
By tracking changes before and after the event, we quantified how mobility deteriorated and how it recovered as restoration progressed. The results indicate that travel speeds dropped sharply immediately after the earthquake and then gradually rebounded; nevertheless, even two months after the event, speeds remained at approximately 80% of pre-earthquake levels in heavily affected areas. We further investigated the relationship between pavement condition and mobility using statistical modeling and machine learning. Both approaches consistently show that increases in earthquake-induced pavement damage significantly contribute to travel speed reductions, even after accounting for road and traffic-related factors.
These findings demonstrate the value of connected vehicle data for rapid, large-area screening of road surface damage and for evidence-based prioritization of restoration actions. More broadly, the proposed approach can reduce the burden of post-disaster field surveys and support effective recovery planning that explicitly considers traffic impacts, thereby improving the resilience of road network operations after large-scale earthquakes.