Bringing Agentic Search to Earth Observation Data Discovery
This work addresses the problem of discovering relevant geoscience datasets and tools for domain experts, providing a practical deployed system with substantial retrieval improvements.
The authors built an agentic search system for NASA's Earth observation data that uses a knowledge graph and a neural scorer fine-tuned on 47k query-dataset pairs, achieving over 5x improvement in Recall@10 and MRR over baselines, with an additional 28% MRR gain from zero-shot LLM reranking.
NASA and its data centers hold thousands of geoscience datasets and tools like Worldview, Giovanni, the Science Discovery Engine, and Harmony. Finding the right one is hard even for domain experts. We present an agentic search system, deployed as a public service for the geoscience community, that takes a natural-language research query and returns the matching datasets and tools. We demonstrate that, in the era of large language models, the latent value of knowledge graphs (KGs) can be substantially amplified through agentic search. From the NASA Earth Observation Knowledge Graph (NASA EO-KG) we derive NASA-EO-Bench, an open benchmark of 47k query-dataset pairs (21k task-based queries). A neural scorer fine-tuned on NASA-EO-Bench beats cosine and BM25 baselines. Further combining it with BM25 via score fusion raises both Recall@10 (R@10) and MRR by over 5x. On top of this supervised pipeline, we add a zero-shot agentic reranking stage that, without any additional training, lifts MRR by 28% on a stratified N=200 subset, showing that LLM reasoning is complementary to supervised retrieval.