Séminaires

Predicting ocean waves in real-time

Ocean waves are usually predicted in a meteorological sense, from hourly to weekly time scales. In this talk, we cover the much lesser-known topic of real-time ocean wave forecasting, which consists in predicting accurately the second-by-second motion of the sea surface, a ship or another floating structure, up to a few minutes in advance. Such real-time forecasts would be highly desirable to mitigate risks and costs, associated with the commissioning and maintenance of marine renewable energy systems.
We will first focus on ocean wave predictability : Given available observations of the sea surface, where, and when can the wave motion be predicted accurately ? We will show that the commonly employed model for “deterministic” prediction zones is unreasonably optimistic for realistic wave fields. Instead, a classical Gaussian wave field representation allows the rigorous calculation of “probabilistic” prediction zones, defined with respect to a given uncertainty criterion. Those prediction zones can have relatively simple physical interpretation. We then show experimental evidence of the validity of our wave predictability model.
Real-time wave forecasting implies that waves are also measured in real time and, as we will see, over large distances around the point of interest, which represents another scientific and technical challenge. We will thus briefly describe ongoing experimental and numerical work at IFP Energies Nouvelles, on remote wave measurements using sensors such as LiDARs and X-band marine radars.