Understanding Extreme Geohazards: The Science of the Disaster Risk Management Cycle European Science Foundation Conference November 28 to December 1, 2011, Sant Feliu de Guixols, Spain

Session 2 “Preparing for the extreme: quantifying the propabilities and uncertainties of extreme hazards.”

Discussion: “How can science better address the uncertainties of the extreme events?”

Carl Bonnevie Harbitz

• The extreme events are also the least likely ones, but still most often dominate the risk. Hence, the probabilities are associated with large uncertainties. Will the simple "scenario based approach" give the same (or better) results than probabilistic assessments?
• Searching the risk related to extreme events with limited data bases often calls for "engineering judgment". Will we then need probabilistic models?
• Not all hazards are convenient for probabilistic approaches, e.g. landslide data normally do not provide equally good statistics as earthquakes while the local geological/geotechnical settings and the release mechanism are crucial.
• For submarine landside tsunamis the uncertainties are related to the intensity/magnitude of the trigger, the response in terms of the landslide, and the tsunami generation, propagation, and inundation. Hence, for tsunamis and probably most other hazards not only the frequency/intensity of the scale needs to be analyzed, but also the error propagation. This calls for random source distributions and stochastic approaches.
• Some hazards are ill-behaving (e.g. tsunamis having a very nonlinear relation between metric and consequence) and others are more well-behaving (e.g. storm surge). Where is the challenge?
• How relevant are combined effects and cascading events (multi-hazards)?
• How reliable is past experience in view of the climate changes?
• How mature are the vulnerability models to be applied in a risk assessment?

Ivan Wong

Probabilistic hazard analysis is the approach used to incorporate the uncertainties in characterizing natural phenomena such as earthquakes to predict the resulting hazards including those from extreme events. A successful probabilistic hazard analysis captures the center, body, and the range of technical interpretations that the larger informed technical community would have if they were to conduct the study. A major obstacle however is that experts tend to underestimate uncertainties. They tend to act as proponents of models rather than being objective evaluators. As experts, we must recognize that the earth is a complex system and that our knowledge is far from complete. I contend that if probabilistic hazard analyses had been properly conducted in northern Japan, efforts to mitigate the hazards from extreme events such as the 2011 Tohoku-Oki earthquake would have proved to be more effective.