29 June 2018

Impact and Importance of Hyperdiffusion on the Spectral Element Method: A Linear Dispersion Analysis

Hyperviscosity is demonstrably important for eliminating spectral gaps and improving the treatment of waves
Science

Hyperdiffusion is used in atmospheric models to eliminate spurious, unphysical noise that emerges from the way numerical methods represent the atmosphere.  This paper uses a theoretical analysis to compute the optimal amount of hyperdiffusion needed by atmospheric models using the spectral element method (such as E3SM's dynamical core).

Impact

Determination of an optimal hyperdiffusion coefficient allows us to minimize diffusive errors and computational cost while maximizing the accuracy of the model’s dynamical core.

Summary

This paper targets the calculation of a theoretically optimal hyperdiffusion coefficient for use in the DOE Energy Exascale Earth System Model (E3SM) non-hydrostatic spectral-element dynamical core.  The selection of such a coefficient would be effective at eliminating the spectral gap and improving the accuracy of the dynamical core.

Contact
Paul Ullrich
University of California at Davis (UC Davis)
Funding
Programs
Publications
Ullrich, PA, DR Reynolds, JE Guerra, and MA Taylor.  2018.  "Impact and Importance of Hyperdiffusion on the Spectral Element Method: A Linear Dispersion Analysis."  Journal of Computational Physics.  https://doi.org/10.1016/j.jcp.2018.06.035.