Bayesian networks to quantify the reliability of a debris flow alarm system

Warning and alarm systems are part of an integrative approach to risk management for natural hazards that supplement protective measures such as rock fall nets, dams and galleries. To integrate warning and alarms systems as standard measures in a risk management approach, their reliability should be quantified. In this paper, selected methods are applied to quantify the reliability of an active threshold-based debris flow alarm system. The reliability is defined as the ability of the system to detect dangerous debris flow events, to issue alarms in a timely manner and to avoid false alarms. Bayesian networks are applied to probabilistically model the considered alarm system and to calculate its overall reliability. The final system reliability is expressed in terms of the receiver operator characteristics, which allow the identification of the optimal trade-off between the probability of detection and the probability of false alarms.