Estimating the benefits of early warning systems in reducing urban flood risk to people: a spatially explicit Bayesian model

Flood risk assessment usually focuses on damages to material objects (indirect tangible costs) and downplays the broader socio-economic aspects of flood-prone systems. Such aspects are crucial for an accurate assessment of risk to human receptors and of the benefits of non-structural measures. For example, an early warning system (EWS) that reduces the amount of direct tangible costs only partially could: (i) save lives (direct intangible costs); (ii) help avoid long-lasting trauma (indirect intangible costs); (iii) prevent post-disaster evacuation costs (indirect tangible costs). We present a methodology to assess flood risk to people, which integrates people’s vulnerability and ability to cushion hazards by coping and adapting. The study area covers the lower part of the Sihl valley (Switzerland) including the city of Zurich. Flood risk to people is modelled using a spatially explicit Bayesian network model calibrated on expert opinions (25 experts were involved). Risk to people is assessed in terms of: (1) likelihood of non-fatal physical injury; (2) likelihood of post-traumatic stress disorder; (3) likelihood of death. The model is used to estimate the benefits of improving an existing EWS, taking into account reliability, lead-time and scope. The proposed approach can: (1) improve flood cost estimation by extending its scope beyond direct and tangible damages; (2) complement quantitative and semi-quantitative data with subjective and local knowledge, improving the use of commonly available information; and (3) produce estimates of model uncertainty by providing probability distributions for all its outputs.