Fast Summation on the Sphere with Applications to the Barotropic Vorticity Equation

Anthony Chen,Christiane Jablonowski

2024-01-15

Abstract:Fast summation refers to a family of techniques for approximating $O(N^2)$ sums in $O(N\log{N})$ or $O(N)$ time. These techniques have traditionally found wide use in astrophysics and electrostatics in calculating the forces in a $N$-body problem. In this work, we present a spherical tree code, and apply it to the problem of efficiently solving the barotropic vorticity equation.

Numerical Analysis,Atmospheric and Oceanic Physics,Computational Physics,Fluid Dynamics

What problem does this paper attempt to address?

This paper mainly discusses the problem of fast summation on a sphere and its application to the efficient solution of the Barotropic Vorticity Equation. Fast summation techniques are commonly used to reduce the computational complexity from O(N^2) to O(N log N) or O(N), which is often seen in N-body problems in astrophysics and electrostatics. The paper proposes a spherical tree encoding method, which takes into account the common spherical geometry properties in geophysics problems.

The work designs a spherical tree code to solve fluid dynamics problems in geophysics, particularly the time evolution of the Barotropic Vorticity Equation. This equation describes the motion of a two-dimensional incompressible and inviscid fluid under the Coriolis force. The paper uses the Lagrangian particle method and interpolating functions of spherical triangles for numerical simulations, and validates the method's effectiveness by comparing the error between direct summation and fast summation.

The paper also discusses different types of interactions, such as particle-particle, particle-cluster, cluster-particle, and cluster-cluster interactions, and introduces an algorithm based on dual-tree traversal for approximate computations on the sphere. Moreover, the precision and computational cost of fast summation are controlled by comparing different interpolation degrees and Multipole Acceptance Criteria (MAC).

Numerical test results show that the method performs well in both serial and parallel computing environments, and it is competitive compared to existing fast summation methods like the Barycentric Lagrange Tree Code. By evaluating the error of the method using the exact solution of the Rossby-Haurwitz wave, the paper demonstrates that as the interpolation degree increases, the error decreases but the computation time also increases.

Future research directions include applying this method to more complex Earth science problems, such as the computation of shallow water equations and self-attraction and loading (SAL), as well as running on graphics processing units (GPUs) to handle larger-scale problems.