2018/08/13 by Matthias Zacharuk, Zacharuk, Matthias, Stamen Dolaptchiev +5
Computer Science · Earth and Planetary Sciences · Physics and Astronomy · #35Q35 #60J70 #86A10 #FOS: Physical sciences #Fluid Dynamics (physics.flu-dyn) #Image and Signal Denoising Methods #Meteorological Phenomena and Simulations #Model Reduction and Neural Networks #Oceanographic and Atmospheric Processes #Seismic Imaging and Inversion Techniques
paper · pdf · doi:10.48550/arxiv.1808.05467
openalex publication_date 2018/08/13 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
We address the question of parameterizing the subgrid scales in simulations\nof geophysical flows by applying stochastic mode reduction to the\none-dimensional stochastically forced shallow water equations. The problem is\nformulated in physical space by defining resolved variables as local spatial\naverages over finite-volume cells and unresolved variables as corresponding\nresiduals. Based on the assumption of a time-scale separation between the slow\nspatial averages and the fast residuals, the stochastic mode reduction\nprocedure is used to obtain a low-resolution model for the spatial averages\nalone with local stochastic subgrid-scale parameterization coupling each\nresolved variable only to a few neighboring cells. The closure improves the\nresults of the low-resolution model and outperforms two purely empirical\nstochastic parameterizations. It is shown that the largest benefit is in the\nrepresentation of the energy spectrum. By adjusting only a single coefficient\n(the strength of the noise) we observe that there is a potential for improving\nthe performance of the parameterization, if additional tuning of the\ncoefficients is performed. In addition, the scale-awareness of the\nparameterizations is studied.\n