EarthLink: Interpreting Climate Signals with Self-Evolving AI Agents

2025-07-23

AI Summary

Earth system data exhibit massive scale, heterogeneous provenance, and structural complexity, while analytical tasks—particularly spatiotemporal overlay analyses—are becoming increasingly intricate, severely impeding scientific discovery efficiency. To address this, we propose EarthLink: the first interactive AI assistant tailored for Earth scientists. It introduces a self-evolving AI agent architecture that automates end-to-end research workflows—including hypothesis-driven study design, executable code generation, and multi-scenario analysis—while enabling interpretable climate signal decomposition and dynamic model refinement. EarthLink integrates natural language understanding, program synthesis, feedback-driven optimization, and multimodal analysis to ensure workflow transparency and auditability. Evaluated on core climate science tasks, it achieves performance comparable to early-career researchers; domain expert assessment confirms its scientific validity. Empirical results demonstrate substantial gains in analytical efficiency and result reproducibility, advancing human–AI collaborative scientific discovery.

Technology Category

• Humans and AI: Explainable AI (XAI) for Human Understanding
• Machine Learning: Efficient ML / Green AI
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Assisted, interactive, and conversational search
• Economics, Online Markets and Human Computation: Architectures and workflows that use LLMs for crowd work

Abstract

Modern Earth science is at an inflection point. The vast, fragmented, and complex nature of Earth system data, coupled with increasingly sophisticated analytical demands, creates a significant bottleneck for rapid scientific discovery. Here we introduce EarthLink, the first AI agent designed as an interactive copilot for Earth scientists. It automates the end-to-end research workflow, from planning and code generation to multi-scenario analysis. Unlike static diagnostic tools, EarthLink can learn from user interaction, continuously refining its capabilities through a dynamic feedback loop. We validated its performance on a number of core scientific tasks of climate change, ranging from model-observation comparisons to the diagnosis of complex phenomena. In a multi-expert evaluation, EarthLink produced scientifically sound analyses and demonstrated an analytical competency that was rated as comparable to specific aspects of a human junior researcher's workflow. Additionally, its transparent, auditable workflows and natural language interface empower scientists to shift from laborious manual execution to strategic oversight and hypothesis generation. EarthLink marks a pivotal step towards an efficient, trustworthy, and collaborative paradigm for Earth system research in an era of accelerating global change.

Problem

• Automating end-to-end Earth science research workflows
• Interpreting complex climate data with evolving AI
• Enhancing scientific discovery through interactive AI copilot

Innovation

• AI agent automates end-to-end research workflow
• Learns from user interaction via feedback loop
• Transparent workflows with natural language interface