An interpretation of TRiSK-type schemes from a discrete exterior calculus perspective

TL;DR

This paper interprets TRiSK-type numerical schemes for geophysical fluid dynamics as instances of discrete exterior calculus applied to Hamiltonian shallow water equations, offering a unified geometric framework and potential improvements.

Contribution

It provides a comprehensive geometric interpretation of TRiSK schemes using discrete exterior calculus, unifying existing methods and suggesting new operator choices for enhanced accuracy.

Findings

All known TRiSK schemes fit within the DEC framework.

Identifies operator choices that could improve scheme accuracy.

Suggests new combinations to address known issues like instability.

Abstract

TRiSK-type numerical schemes are widely used in both atmospheric and oceanic dynamical cores, due to their discrete analogues of important properties such as energy conservation and steady geostrophic modes. In this work, we show that these numerical methods are best understood as a discrete exterior calculus (DEC) scheme applied to a Hamiltonian formulation of the rotating shallow water equations based on split exterior calculus. This comprehensive description of the differential geometric underpinnings of TRiSK-type schemes completes the one started in \cite{Thuburn2012,Eldred2017}, and provides a new understanding of certain operators in TRiSK-type schemes as discrete wedge products and topological pairings from split exterior calculus. All known TRiSK-type schemes in the literature are shown to fit inside this general framework, by identifying the (implicit) choices made for various DEC operators by the different schemes. In doing so, unexplored choices and combinations are identified that might offer the possibility of fixing known issues with TRiSK-type schemes such as operator accuracy and Hollingsworth instability.