Alert button
Picture for Simon Van Gennip

Simon Van Gennip

Alert button

ENIB

Deep learning for Lagrangian drift simulation at the sea surface

Nov 17, 2022
Daria Botvynko, Carlos Granero-Belinchon, Simon Van Gennip, Abdesslam Benzinou, Ronan Fablet

Figure 1 for Deep learning for Lagrangian drift simulation at the sea surface
Figure 2 for Deep learning for Lagrangian drift simulation at the sea surface
Figure 3 for Deep learning for Lagrangian drift simulation at the sea surface
Figure 4 for Deep learning for Lagrangian drift simulation at the sea surface

We address Lagrangian drift simulation in geophysical dynamics and explore deep learning approaches to overcome known limitations of state-of-the-art model-based and Markovian approaches in terms of computational complexity and error propagation. We introduce a novel architecture, referred to as DriftNet, inspired from the Eulerian Fokker-Planck representation of Lagrangian dynamics. Numerical experiments for Lagrangian drift simulation at the sea surface demonstrates the relevance of DriftNet w.r.t. state-of-the-art schemes. Benefiting from the fully-convolutional nature of Drift-Net, we explore through a neural inversion how to diagnose modelderived velocities w.r.t. real drifter trajectories.

Viaarxiv icon