Flood warnings from various information sources are important for individuals to make evacuation decisions during a flood event. In this study, we develop a general opinion dynamics model to simulate how individuals update their flood hazard awareness when exposed to multiple information sources, including global broadcast, social media, and observations of neighbors' actions. The opinion dynamics model is coupled with a traffic model to simulate the evacuation processes of a residential community with a given transportation network. Through various scenarios, we investigate how social media affect the opinion dynamics and evacuation processes. We find that stronger social media can make evacuation processes more sensitive to the change of global broadcast and neighbor observations, and thus, impose larger uncertainty on evacuation rates (i.e., a large range of evacuation rates corresponding to sources of information). For instance, evacuation rates are lower when social media become more influential and individuals have less trust in global broadcast. Stubborn individuals can significantly affect the opinion dynamics and reduce evacuation rates. In addition, evacuation rates respond to the percentage of stubborn agents in a nonlinear manner, i.e., above a threshold, the impact of stubborn agents will be intensified by stronger social media. These results highlight the role of social media in flood evacuation processes and the need to monitor social media so that misinformation can be corrected in a timely manner. The joint impacts of social media, quality of flood warnings, and transportation capacity on evacuation rates are also discussed.

No abstract

No abstract

This paper proposes a distributed continuous-time epidemic model, called networked SIWS (Susceptible-Infected-Water-Susceptible) model, for an SIS type waterborne disease spreading over a network of multiple groups of individuals sharing a water source. A sufficient condition is obtained for the healthy state, at which all individuals are not infected and the water is not contaminated, to be globally asymptotically stable. The effects of the shared water source on the disease spreading are analyzed through the comparison of the basic reproduction number with the networked SIS model without water and demonstrated via simulations.