Parametric and Non-Parametric Ocean Swell Dynamics Modeling in Particle-In-Cell
Please login to view abstract download link
An extension of the Particle-in-Cell for Efficient Swell (PiCLES) wave modeling framework (Hell et al., 2024) is proposed. PiCLES uses a cost-efficient, mixed Lagrangian-Eulerian algorithm and is currently valid for wind-sea modeling. This work aims at extending it to ocean surface swell systems. To achieve this, two methods are proposed. The first is a non-parametric ensemble approach to sample the energy spectrum with individual particles. In particular, we study how this model spreads an initial energy spectrum in space and time and how to constrain it to match observations. Particle degeneracy emerges from this method, thus, sequential importance resampling techniques are considered to prevent it and to control how the discrete distribution, i.e. the particle ensemble, can match the true distribution after a time step. The possibility of data assimilation before remeshing is considered, which allows to harmonize the weights with observations. Tests are performed and compared to analytical solutions driven by the expected diffuse geometric optic behavior. Along with the non-parametric ensemble approach, a parametric method is explored that drastically reduces the number of particles by allowing each particle to carry more information. References : Hell, M., Chapron, B., and Fox-Kemper, B. (2024). A particle-in-cell wave model for efficient sea-state and swell estimates in earth system models - PiCLES.
