Is city traffic damaged by torrential rain?

• Published in: Physica A 2018-08, Vol.503, p.1073-1080
• Main Authors: Guo, Shengmin, Wu, Ruoqian, Tong, Qingfeng, Zeng, Guanwen, Yang, Jian, Chen, Long, Zhu, Tongyu, Lv, Weifeng, Li, Daqing
• Format: Article
• Language: English
• Subjects: Network robustness; Percolation theory; Traffic data science; Traffic reliability
• ISSN: 0378-4371; 1873-2119
• DOI: 10.1016/j.physa.2018.08.044

Extreme weather, such as torrential rain, could lead to severe damage to transportation. Many studies have been proposed focusing on the influence of extreme weather on the traffic flow properties. However, the robustness of whole dynamic traffic networks under extreme weather is rarely addressed. Particularly, little attention has been paid to the question whether and how the local destruction of roads is aggregated into a degradation of global traffic operation. Based on real-time traffic data, here we apply percolation analysis on traffic networks and find that the torrential rain can lead to different effects on different levels: on the network scale, the traffic percolation threshold as an indicator for city traffic reliability is stable against weather perturbation, while a portion of roads at the microscopic level is significantly influenced and forming local cluster isolated from the main functional network. This may be due to the fact that torrential rain and other extreme weather condition will not only generate the damaged roads in the supply end, but also reduce the traffic demand correspondingly. Our research suggests the traffic percolation may reflect the nature of relation between local flow and global flow, which can help to design corresponding management strategies. •Two quantitative indicators of traffic percolation were used in the city traffic.•The traffic percolation threshold is stable against weather perturbation.•Some roads are seriously affected and forming local cluster under torrential rain.