Quantitative resilience assessment in emergency response reveals how organizations trade efficiency for redundancy
Current frameworks to assess resilience are often based on low-dimensional models that describe a small number of inter-related components. However, for many infrastructures their resilience is a property that emerges from myriads of individual interactions. Here, we develop for the first time fully quantitative and data-driven indicators for several aspects that relate to resilient organizational behavior during emergency situations. We consider dynamic communication networks of emergency responders collected during the annual public emergency drill of the Hungarian National University of Public Service. Our key hypothesis is that the time-dependent network structure derived from these communication flows conveys information as to how redundant, vulnerable, and efficient individual organizations acted during the drill. The resulting indicators can be applied on two different levels—the level of the entire network and the level of individual participants.
To validate our framework, we show that when the personnel is prepared for a given event (e.g., critical weather conditions), effective communication within the organization is dominated by vertical flows of information. However, under surprise the affected units concentrated on horizontal communication, thereby creating bottlenecks and losses of efficiency in the network. Our results indicate a generic trade-off between efficiency, vulnerability and redundancy of communication networks in emergency response. Moreover, we can relate the origin of this trade-off to the specific way of how individual organizations adapted their communication behavior in times of crisis. We do so in a fully quantitative and data-driven way that is directly linked to real-world networks, infrastructures, and the people acting within them.
P. Klimek, J. Varga, A.S. Jovanovic, Z. Szekely, Quantitative resilience assessment in emergency response reveals how organizations trade efficiency for redundancy, Safety Science 113 (2019) 404–414