Discussing issues in uncertainty quantification. The case of geohazard assessments

By making explicit the modeling choices and assumptions made, we analyze some issues in quantifying uncertainty using geohazard models. Under the often condition of very limited data, a major problem is constraining the many parameters involved. We conclude that, despite the availability of recently developed sophistications, the quantification based on these ideal parameterized models can hardly be justified since, e.g., they will only reflect some aspects of the uncertainty involved. This calls for more insightful approaches which are yet to be developed.