Computationally-Efficient Climate Predictions using Multi-Fidelity Surrogate Modelling

TL;DR

This paper presents a multi-fidelity surrogate modelling approach using Gaussian processes to produce accurate high-fidelity climate predictions at a fraction of the computational cost, demonstrated on regional temperature forecasts.

Contribution

It introduces a novel multi-fidelity Gaussian process model combining global and regional climate models for efficient climate prediction.

Findings

Achieved high-fidelity temperature predictions with 15.62°C² average error.

Reduced high-fidelity model evaluations to 6% of the region.

Significantly lowered computational costs compared to high-fidelity simulations.

Abstract

Accurately modelling the Earth's climate has widespread applications ranging from forecasting local weather to understanding global climate change. Low-fidelity simulations of climate phenomena are readily available, but high-fidelity simulations are expensive to obtain. We therefore investigate the potential of Gaussian process-based multi-fidelity surrogate modelling as a way to produce high-fidelity climate predictions at low cost. Specifically, our model combines the predictions of a low-fidelity Global Climate Model (GCM) and those of a high-fidelity Regional Climate Model (RCM) to produce high-fidelity temperature predictions for a mountainous region on the coastline of Peru. We are able to produce high-fidelity temperature predictions at significantly lower computational cost compared to the high-fidelity model alone: our predictions have an average error of $15.62^\circ\text{C}^2$ yet our approach only evaluates the high-fidelity model on 6% of the region of interest.