ABSTRACT
Floods and dam overtopping risks during strong rainfall preoccupy the authorities since a long time. To assure a good protection, extreme floods associated with very low probability-ties must be estimated. Their determination is usually based on the statistical extrapolation of data for long periods of time. The limits of this probabilistic method led to adaptation of the PMP-PMF (Probable Maximum Precipitation-Probable Maximum Flood) concept to Swiss conditions. This paper shows first that the PMP-PMF method applied without regard to the probability associated to initial conditions of the simulation may lead to exceptional flow values that may seem unrealistic. A solution is then proposed to estimate such probabilities for various initial conditions such as the rain-snow boundary or the wind direction and speed. These occurrence probabilities can then be included in the PMP-PMF calculations to lend more clarity to the likelihood of the extreme conditions causing the flood and allow the estimation of its return period.
