Earth System

Earth system science aims to understand the complex interactions within and between Earth's components (atmosphere, oceans, land, ice) to improve predictions of climate and weather. Current research heavily utilizes machine learning, employing architectures like convolutional neural networks, vision transformers, and normalizing flows, to analyze high-dimensional datasets (e.g., Earth System Data Cubes) and build surrogate models for computationally expensive Earth system models. This work focuses on improving forecast accuracy, particularly for extreme events, enhancing model interpretability through explainable AI, and addressing challenges like data scarcity and model robustness in a changing climate. These advancements have significant implications for climate change mitigation and adaptation strategies, as well as for various sectors reliant on accurate weather and climate predictions.