A DSS for risk management on drinking water infrastructures during extreme events

The safety of drinking water infrastructures is fundamental for economic, social and sanitary reasons, and should be guaranteed both during ordinary service and in case of emergencies. Besides verifying and preventing deterioration and ageing, the response of the system to extreme events should be also carefully analyzed. As a matter of fact, depending on the level of preparedness that water system authorities have adopted, the restoration of system functionality may require days, weeks, or even months. Water supply systems are vulnerable towards several hazardous events, which can be mainly classified as natural (such as earthquakes, hurricanes, volcanic eruptions, landslides, fires...) or anthropic (both intentional and accidental, such as pollution, operational mistakes, black-outs etc.). Referring to the potential consequences of such events on the system, physical damages consisting in breakage or malfunctioning of one or more elements of the network, should be distinguished from water contamination. A research activity is being developed by the Water Research Institute of the National Research Council (Istituto di Ricerca sulle Acque del Consiglio Nazionale delle Ricerche IRSA-CNR), supported by the Italian Department of Civil Protection (Dipartimento della Protezione Civile - DPC), with the aim of defining a strategic Decision Support System (DSS) for efficient and coherent decision-making in case of threats involving drinking water infrastructures. The DSS is based on Bayesian Belief Networks (BBNs), a semi-quantitative probabilistic tool particularly useful for managing emergency situations, characterized by time shortness and information uncertainty. It should be mainly used for detecting potential shortcomings of the system during emergencies, but also for helping water authorities in defining priorities of action, even with reference to ordinary management procedures. In the following, the methodological approach adopted is firstly presented, with specific reference to the main features of BBNs. Then, the structure of the methodology is described in synthesis. At last, the applicability of the tool is discussed, referring to a couple of real case studies developed with the cooperation of Acquedotto Pugliese S.p.A., an Italian water authority.