Causes and diagnostics of internal tide non-stationarity
| Quand ? |
Le 07/06/2022, de 11:00 à 12:00 |
|---|---|
| Où ? | Salle des thèses |
| Participants |
Nicolas Grisouard |
| Ajouter un événement au calendrier |
vCal
iCal
|
Internal tides are oceanic internal waves that oscillate approximately at tidal frequencies. However, their scattering by the turbulent oceanic eddy field is somewhat time-dependent. “Approximately” and “somewhat” are unfortunately the operative words: our fundamental understanding of interactions between eddies and internal waves is still far from being able to make predictions in realistic settings. On the other hand, satellite altimeters, which are our most reliable measurements to track internal tides globally, suffer from a sampling that is too coarse in time to capture the tidal oscillations. Consequently, oceanographers can only fit perfectly periodic signals through very incomplete times series, and completely miss the rest, i.e., the part of the signal whose local phase is “somewhat” time-dependent.
In this talk, I will start by presenting the theoretical underpinnings of wave scattering by eddies. I will then present numerical experiments of tidal wave scattering by barotropic balanced vortices. This will allow me to propose a scaling law for the degree of scattering. This new law unifies, quantifies, and completes our current understanding of scattering. On the other hand, these idealized experiments will not immediately help us, tackling the realistic flows that next-generation satellite altimeters will observe. In the last part of this talk, I describe how our group trained and tested a deep-learning algorithm to produce snapshots of the wave’s signature based on a raw snapshot containing both eddies and waves. I will conclude by explaining the limitations of our method, and our current efforts to overcome them.