PULSE Lab Awarded a NOAA GOMO Pilot for Multi-Agent AI Ocean Data Analysis

We are pleased to share that the NOAA Global Ocean Monitoring and Observing (GOMO) Program has selected the PULSE Lab for one of the eight pilot projects in its FY26 Innovation Program on Artificial Intelligence and Data Management, announced on 30 September 2026.

Our project

A Multi-Agent System for Ocean Data Analysis and Knowledge Generation

• PI: Olmo Zavala-Romero (Department of Scientific Computing)
• Co-Investigators: Gordon Erlebacher (Department of Scientific Computing) and Keshav Raja (Center for Ocean-Atmospheric Prediction Studies)

GOMO funded eight pilots this year, with a combined expected investment of about $1.5 million that includes partner funding from the U.S. Integrated Ocean Observing System (IOOS) Program and the Uncrewed Systems (UxS) Operations Center.

What we will build

NOAA and its partners collect large amounts of ocean data from floats, ships, stations and satellites. Turning those data into maps, charts and summaries usually requires specialised programming. This project will build an AI system that takes a plain-language question about ocean data and returns an answer together with the evidence behind it.

The system coordinates specialised agents. An orchestrator interprets the question and plans the analysis. A data-access agent finds and retrieves the relevant data through ERDDAP, a widely used service for finding and downloading scientific datasets. An analysis agent computes statistics, trends and diagnostics, and a visualization agent produces maps, time series and profiles. Every answer records where the data came from, which dates and locations were used, which quality checks were applied, and what limitations apply, so that results can be reviewed and reproduced.

The first demonstrations will use Argo float observations distributed through NOAA's Observing System Monitoring Center and regional observations from GCOOS, the coastal ocean observing system for the Gulf. They will address questions about ocean temperature, salinity, sampling coverage, recent conditions in the Gulf, and gaps in the data. Alongside the system, the team will develop reusable Argo and Gulf workflows, benchmark tests, documentation, and guidance for transitioning the tools to NOAA use.

Why it matters

The aim is to make ocean observations easier to use, easier to review, and faster to turn into products that support NOAA science, data management, visualization, education and decision-making.

The full announcement, including the other seven pilots, is on the GOMO website.