Network-Based Simulation of Flood Alert Dissemination Using GDACS and Synthetic Communication Logs

This study presents a simulation-based analysis of disaster alert dissemination using flood alerts from the Global Disaster Alert and Coordination System (GDACS) combined with a synthetic stakeholder communication network. The proposed framework models information flow across control centers, media outlets, community leaders, and citizens using the Independent Cascade Model. A directed graph was constructed from 300 simulated message exchanges, incorporating delay and reliability attributes. Results from ten simulation runs show significant variability in alert reach (19–49%) and propagation delays, with citizens experiencing the longest average delay (348 sec). Centrality and bottleneck-based analysis identified mid-network actors, particularly community leaders and media outlets, as critical to effective dissemination. These findings underscore the structural dependencies within early warning systems and highlight the need for resilient communication pathways in disaster response networks.