Dynamic assessment of postdisaster road network vulnerability using crowdsourced traffic data

Abstract

Accurate identification of the spatial and temporal dynamic impacts of disasters on road networks is crucial for making informed decisions regarding postdisaster emergency response and recovery. In this paper, we propose a multidimensional vulnerability assessment framework that utilizes time series crowdsourced data. The framework enables a spatiotemporal analysis of traffic flows at the link level, allowing us to identify the primary factors that contribute to vulnerability after a disaster. The results show that there is a significant variation in link vulnerability over time, with the overall network gradually returning to a normal state. We also observe a global correlation in link vulnerability across the spatial aspect, although the autocorrelation effect is only present among directly contiguous links. Furthermore, metrics affecting link vulnerability varied during different periods of response. Graph-based metrics of road network structure, population density, and road type are key factors influencing postearthquake vulnerability.