Water scarcity, flooding, and shifting agricultural conditions are among the most pressing challenges facing communities across North America. Recent assessments show that extreme weather, particularly drought, continues to drive significant agricultural losses across the country, while flooding remains one of the nation’s most destructive and costly natural hazards, damaging infrastructure, disrupting economies, and disproportionately affecting vulnerable communities. These negative impacts increase the need for timely, granular insights to help decision-makers—from state water agencies to farmers to urban planners—anticipate risks and adapt.
NASA has long produced advanced hydrology and land-surface datasets, powering breakthroughs in drought early-warning systems, environmental planning, and environmental research. Yet despite their value, these datasets and the specialized tools required to navigate and interpret them remain difficult to access for many who could benefit most.
Together, NASA and Microsoft are working to make these insights more accessible with the use of AI agents powered by Microsoft Azure OpenAI Service and Microsoft Foundry. Building on the success of the Earth Copilot demonstration, researchers in NASA’s Hydrological Sciences Lab are developing a multi-agent copilot to make it easier for a wider range of scientists, researchers and policy makers to engage with and draw insights from the comprehensive, yet complex, datasets.
The Usability Gap
For many organizations, including state water agencies, emergency managers, agricultural planners, and even interdisciplinary research teams, the challenge is not access to data, but the expertise required to work with it. Hydrologic datasets contain rich, multidimensional information across dozens of variables, each with specific units, resolutions, processing workflows, and modeling assumptions. Understanding soil moisture patterns, drought anomalies, or flood indicators can require a deep background in hydrology and remote sensing.
This complexity slows down scientific analysis and limits the broader societal reach of hydrology data. Insights that could help communities prepare for drought, manage reservoirs, or anticipate flood risks often remain locked behind technical barriers. The question becomes: how do we preserve scientific rigor while making this information easier to explore, interpret, and apply?
A Copilot for Hydrology
To address this challenge, NASA and Microsoft collaborated a new multi-agent AI system that enables intuitive, plain-language interaction with hydrology datasets. Built on the NASA Earth Copilot architecture that the two organizations previously prototyped together, this solution extends the same principles—trusted science, human-centered AI, and scalable cloud integration—into the hydrology domain.
Instead of navigating complex documentation or writing custom code, users simply ask the system questions:
“How has soil moisture changed this year?”
“Which regions may be facing elevated flood risk?”
“Show me drought stress indicators across the southern plains.”
The copilot interprets the question, identifies the relevant hydrologic variables, retrieves authoritative explanations, runs the necessary geospatial queries, and presents results as maps, charts, and clear narrative descriptions. It feels less like searching a database and more like collaborating with a hydrologist who understands both the science and the user’s intent.
This approach dramatically lowers the barrier to working with NASA’s hydrology datasets, empowering a broader range of users—from planners and water managers to scientists and agricultural analysts—to quickly derive insights that previously required significant technical expertise.
NLDAS-3: A New Era of Water Intelligence
Central to this new capability is the North American Land Data Assimilation System Version 3 (NLDAS-3), one of NASA’s most advanced hydrology datasets. NLDAS-3 delivers hourly, 1-kilometer resolution data across North and Central America, spanning more than two decades of observations. By harmonizing satellite measurements with state-of-the-art land-surface and hydrology models, it presents a detailed, continental-scale view of the water cycle.
With NLDAS-3, users can examine changing precipitation and temperature patterns, track shifts in soil moisture and groundwater, monitor runoff and flood-relevant conditions, assess snowpack and snowmelt trends, and explore vegetation health and evapotranspiration across seasons and years. The dataset’s richness makes it invaluable for drought monitoring, flood risk assessment and emergency preparedness, agricultural planning, and water resource management efforts.
Most importantly, NLDAS-3 offers not just snapshots but continuous, high-resolution insights. Its consistency across regions and decades allows organizations to evaluate risk, compare conditions, and make informed decisions in a unified framework. By pairing NLDAS-3 with the AI copilot, insights could become more accessible to a much broader audience, turning scientific data into actionable intelligence at scale.
Azure AI Behind the Scenes
The hydrology copilot is powered by a modern Azure AI and data architecture purpose-built for scientific workloads. Azure Synapse Analytics curates and indexes hydrology datasets, including NLDAS-3, so they can be queried quickly and efficiently. This enables users to explore decades of environmental history or near real-time conditions in seconds.
Azure AI Search provides the semantic understanding needed for the system to interpret hydrology concepts and map them to the correct variables and workflows. By indexing NASA’s scientific documentation and metadata, it ensures that answers are grounded in authoritative knowledge.
These components run within Microsoft Foundry, which orchestrates the large language models, multi-agent workflows, and retrieval-augmented generation (RAG) pipelines. The result is a system that brings NASA’s hydrology science and Microsoft’s AI capabilities into a seamless, conversational experience.
A Call to Action
Water scarcity, flooding, and agricultural volatility will continue to shape the environmental and economic landscape of the 21st century. Meeting these challenges requires the ability to understand water conditions as they evolve—clearly, quickly, and with scientific precision.
NASA has created some of the most powerful hydrology datasets in existence. Microsoft, through Azure AI, is helping ensure those datasets can drive real-world impact by making them easier to explore, interpret, and apply.
This partnership marks an important step toward democratizing access to Earth science and equipping communities with the insights they need to prepare for the future. But it is only the beginning. As AI and cloud technologies advance, so too will our ability to transform scientific data into meaningful action.
To explore how Azure AI can help your organization harness NASA’s hydrology data for environmental resilience and water resource planning, connect with our team. Together, we can turn data into actionable intelligence—and build a more resilient future for all.
Microsoft