Bringing Agentic Search to Earth Observation Data Discovery
Abstract
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 framework for geoscience data discovery that takes a natural-language research query and returns 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 to over 5x the unadapted cosine baseline. On top of this supervised pipeline, a zero-shot reranking stage lifts MRR by 16%, significant under a paired bootstrap, with no additional training, and autonomous web and arXiv tool use adds a further gain, showing that LLM reasoning is complementary to supervised retrieval.
Related
- Agentic Large Language Model Workflows for Earth Observation: Tool Orchestration, Foundation-Model Adaptation, and the Evaluation Gap
- CLIVeg: Contrastive Image-Vegetation Alignment for Botany-Aware Earth Observation Representations
- Limits and promises of earth observation foundation models in predicting multi-trophic soil biodiversity
- Earth observation and causal inference in assessing protected area effectiveness
- A perspective on the interpretability of poverty maps derived from Earth Observation
BibTeX
@misc{yu2026bringing,
title = {{Bringing Agentic Search to Earth Observation Data Discovery}},
author = {Minghan Yu and Youran Sun and Chugang Yi and Yixin Wen and Haizhao Yang},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2607.02387},
eprint = {2607.02387},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2607.02387},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Bringing Agentic Search to Earth Observation Data Discovery
AU - Minghan Yu
AU - Youran Sun
AU - Chugang Yi
AU - Yixin Wen
AU - Haizhao Yang
PY - 2026
DA - 2026/09/24
PB - arXiv
DO - 10.48550/arxiv.2607.02387
UR - https://arxiv.org/abs/2607.02387
AB - 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 framework for geoscience data discovery that takes a natural-language research query and returns 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 to over 5x the unadapted cosine baseline. On top of this supervised pipeline, a zero-shot reranking stage lifts MRR by 16%, significant under a paired bootstrap, with no additional training, and autonomous web and arXiv tool use adds a further gain, showing that LLM reasoning is complementary to supervised retrieval.
ER -