Bayesian Analysis of Traffic Accidents and Weather Conditions

Abstract

Background: This study examines the relationship between weather conditions and traffic accidents in Korea. <strong>Methods: We employ a Bayesian analysis to estimate these effects. Results: Given that precipitation levels in summer are less than those in winter, we find that snow has the largest impact on traffic accidents. Thus, the highest relative risk seems reasonable. The city with the lowest average rank is Jinju, which is located in a basin and experiences weather that is influenced by the continental climate of Korea. The annual precipitation in the city is 1,490 mm, and it is classified as an excess rainfall region. In particular, it often experiences foggy weather conditions in the morning because of its proximity to Jinyang Lake. However, the effect of precipitation is similar to that of clear weather, and the fixed effect of fog is negative. Therefore, these weather conditions appear not to have a significant effect on traffic accidents in Jinju. Conclusion: Overall, the results are consistent with the common perception that severe weather conditions aggravate traffic conditions and increase the probability of accidents.