Fault tree analysis for urban flooding

摘要

Traditional methods to evaluate flood risk generally focus on heavy storm events as the principalncause of flooding. Conversely, fault tree analysis is a technique that aims at modelling allnpotential causes of flooding. It quantifies both overall flood probability and relative contributionsnof individual causes of flooding. This paper presents a fault model for urban flooding and annapplication to the case of Haarlem, a city of 147,000 inhabitants. Data from a complaint register,nrainfall gauges and hydrodynamic model calculations are used to quantify probabilities of basicnevents in the fault tree. This results in a flood probability of 0.78/week for Haarlem. It is shownnthat gully pot blockages contribute to 79% of flood incidents, whereas storm events contributenonly 5%. This implies that for this case more efficient gully pot cleaning is a more effectivenstrategy to reduce flood probability than enlarging drainage system capacity. Whether this is alsonthe most cost-effective strategy can only be decided after risk assessment has beenncomplemented with a quantification of consequences of both types of events. To do this will benthe next step in this study.