Understanding changes in water, weather, and land conditions is essential for managing droughts, floods, farming, and ecosystems. NASA’s new North American Land Data Assimilation System Version 3 (NLDAS-3) helps by providing detailed, hourly information at a 1-kilometer scale across North and Central America, starting from 2001 to today. This dataset combines satellite data with computer models to estimate important variables like rainfall, temperature, soil moisture, snow, vegetation health, and groundwater.

But because NLDAS-3 is large and complex, it can be hard for users outside the climate or data science fields to access and understand. To make it easier, we built an AI-powered assistant—or “copilot”—that lets users explore and ask questions about the data using natural language. This copilot, developed with Microsoft Azure tools and NASA’s Earth Copilot system, helps people find the right information quickly and apply it to real-world problems. We show how this system works and give examples of how it can support better decision-making for things like drought monitoring and land management.