Nos tutelles

CNRS

Nos partenaires

Rechercher




Accueil > FR > Actualités

Séminaire de Günter Steinmeyer

par Webmestre - publié le

« The rogue wave mystery »

Günter Steinmeyer, Max-Born-Institut, Berlin, Germany

Vendredi 6 Mai — 11h — P5 salle 172

In many physical systems, measurements underlie Gaussian statistics, which arise as a consequence of the central limit theorem (CLT). If all presuppositions for application of this theorem hold, one can reliably estimate the probability of events that exceed the standard deviation by a certain factor. For example, 2events still show up quite frequently in 0.5% of the samples ; 5events (like the recently claimed gravitational wave observation) are only expected to appear by chance with 10-12 probability. However, there are many systems known where we have no proof for applicability of the CLT. Probably the most prominent example for deviations from Gaussian statistics are ocean rogue waves. These waves exceed 4, and they appear at a rate exceeding the Gaussian prediction of 1 in 108 waves. There is an ongoing discussion on the origin of this anomaly, in particular whether the weak nonlinearity of surface gravity waves plays a decisive role or not. I will demonstrate that ocean rogue waves can be nearly perfectly explained by simple linear interference of waves on the two-dimensional ocean surface. The dimensionality of the ocean surface is one of the key differences to the CLT ; a second difference is assumption of a finite number of interfering waves. A further important aspect of extreme events is their predictability. Analyzing data sets from three different rogue-wave supporting systems, we found remarkable differences concerning the latter aspect. While it seems that ocean rogue waves are predictable, at least in principle, an effective forecast nevertheless appears futile as practical forecast times are limited to a few ten seconds at best.