Session: 02-04-01 Extreme and Freak Waves (Co-hosted with the Ian Young Honoring Symposium)

Paper Number: 103020

103020 - Are Rogue Waves Predicable From Field Measurements? 

Rouge waves continue to endanger ships and offshore infrastructure. Indeed, the loss of several supercarriers and injuries to crew and passengers have been linked to the occurrence of such freak waves. Hence, reliable rogue wave forecasts could significantly reduce the risk for operations on the ocean. To forecasts extreme waves, their formation needs to be understood. Hence, scientists have investigated rogue waves and have found many possible mechanisms generating them. Popular mechanisms for rogue wave generation include random superposition, spatially focusing and nonlinear focusing via for example, the Benjamin-Feir instability. However, under realistic conditions, the role of each mechanism remains unclear.

Different formation mechanisms have direct implications on the predictability of extreme waves events. On the one hand, random superposition can lead to the formation of extreme waves. Therein, waves are modelled as harmonic functions with uniformly distributed phases. Then, rogue waves are generated through constructive inference of multiple wave trains. In this setting, extreme waves are unpredictable. Indeed, the phase information is crucial for prediction of rogue waves. On the other hand, the Benjamin-Feir instability develops over long-time spans and, hence, rogue waves generated by it are predicable. However, in general, this instability is observed in idealized scenarios. Thus, the relevance of this mechanism in an actual ocean remains unclear. This contrast between two prominent generation mechanisms for rogue waves, is illustrative of the need to address a basic question. Are rogue waves predictable at all?

In this talk, the authors ask the questions if rogue waves are predictable from field measurements. Forecasting rogue waves from past data, implicitly assumes a relationship between the past sea state and the future occurrence of a rogue wave. If this functional relationship exists, this relationship can be approximated with universal function approximators, such as neural networks. To this end, the authors compile a large database of buoy measurements with and without extreme waves. This data serves as basis to fit universal function approximators for predicting future rogue waves.

Presenting Author: Thomas Breunung University of Maryland, College Park

Presenting Author Biography: Studied Mechanical Engineering in Germany, Singapore and US

PhD in Mechanical Engineering from ETH Zurich, Switzerland

Currently Post-Doc at the University of Maryland, College Park

OMAE2022 Outreach Fellow

Authors:

Thomas Breunung University of Maryland, College Park
Balakumar Balachandran University of Maryland, College Park