Improving Met Office Weather and Climate Forecasts with Bespoke Multigrid Solvers

TL;DR

This paper presents a custom multigrid solver that significantly accelerates the linear system solutions in the Met Office's atmospheric models, leading to faster forecasts and cost savings.

Contribution

A bespoke multigrid solver was developed, reducing linear solve time by half and improving overall forecast speed and efficiency.

Findings

Linear system solution time halved with the new solver

Global forecast production increased by 10-15%

Estimated annual cost savings of GBP 300k

Abstract

At the heart of the Met Office climate and weather forecasting capabilities lies a sophisticated numerical model which solves the equations of large-scale atmospheric flow. Since this model uses semi-implicit time-stepping, it requires the repeated solution of a large sparse system of linear equations with hundreds of millions of unknowns. This is one of the computational bottlenecks of operational forecasts and efficient numerical algorithms are crucial to ensure optimal performance. We developed and implemented a bespoke multigrid solver to address this challenge. Our solver reduces the time for solving the linear system by a factor two, compared to the previously used BiCGStab method. This leads to significant improvements of overall model performance: global forecasts can be produced 10%-15% faster. Multigrid also avoids stagnating convergence of the iterative scheme in single precision. By allowing better utilisation of computational resources, our work has led to estimated annual cost savings of GBP 300k for the Met Office.