Anemoi: An Agentic Framework for Weather Science

Sumanth Varambally, Marshall Fisher, Jas Thakker, Yiwei Chen, Zhirui Xia, Ruijia Niu, Yasaman Jafari, Veeramakali Vignesh Manivannan, Zachary Novack, Luyu Han, Srikar Eranky, Salva Rühling Cachay, Taylor Berg-Kirkpatrick, Duncan Watson-Parris, Yian Ma, Rose Yu

Keywords: Agents, Large Language Models, Weather Science, Code Generation

TL;DR: We created an AI weather assistant that combines language models with specialized weather analysis tools, enabling scientists to explore meteorological data through natural conversation rather than complex code.

Abstract: Foundation models for weather science are pre-trained on vast amounts of structured numerical data and outperform traditional weather forecasting systems. However, these models lack language-based reasoning capabilities, limiting their utility in interactive scientific workflows. Large language models (LLMs) excel at understanding and generating text but cannot reason about high-dimensional meteorological datasets. We bridge this gap by building a novel agentic framework for weather science. Our framework includes a Python code-based environment for agents (\texttt{AnemoiWorld}) to interact with weather data, featuring tools like an interface to WeatherBench 2 dataset, geoquerying for geographical masks from natural language, weather forecasting, and climate simulation capabilities. We design \texttt{Anemoi}, a multi-turn LLM-based weather agent that iteratively analyzes weather datasets, observes results, and refines its approach through conversational feedback loops. We accompany the agent with a new benchmark, \texttt{AnemoiBench}, with a scalable data generation pipeline that constructs diverse question-answer pairs across weather-related tasks, from basic lookups to advanced forecasting, extreme event detection, and counterfactual reasoning. Experiments on this benchmark demonstrate strong promise for LLM agents to help weather scientists reason about meteorological data more effectively.

Archival Option: The authors of this submission do *not* want it to appear in the archival proceedings.

Submission Number: 79