A Foundation Model for the Earth System

TL;DR

Aurora is a large-scale foundation model for the Earth system that significantly improves prediction accuracy across various domains while reducing computational costs, making Earth system forecasts more accessible.

Contribution

The paper introduces Aurora, a novel foundation model trained on extensive Earth system data, enabling accurate, efficient, and adaptable predictions across multiple environmental domains.

Findings

Aurora outperforms existing operational forecasts in air quality, ocean waves, and cyclone tracking.

Aurora achieves high accuracy with substantially lower computational costs.

Fine-tuning Aurora is feasible for diverse applications with modest computational resources.

Abstract

Reliable forecasts of the Earth system are crucial for human progress and safety from natural disasters. Artificial intelligence offers substantial potential to improve prediction accuracy and computational efficiency in this field, however this remains underexplored in many domains. Here we introduce Aurora, a large-scale foundation model for the Earth system trained on over a million hours of diverse data. Aurora outperforms operational forecasts for air quality, ocean waves, tropical cyclone tracks, and high-resolution weather forecasting at orders of magnitude smaller computational expense than dedicated existing systems. With the ability to fine-tune Aurora to diverse application domains at only modest computational cost, Aurora represents significant progress in making actionable Earth system predictions accessible to anyone.