Making Predictions … So That NOAA Can


“GDIT increasingly uses ML and AI, alongside our knowledge of our customer environments, to predict where a system outage might occur."

Mike Cole
CTO, Federal Civilian

Using Machine Learning and Artificial Intelligence to Enable, Sustain and Improve Weather Forecasting

Weather predictions are about more than just rain and sunshine. They’re how we track storms, protect our citizens, monitor our climate, plan events and even forecast retail demand. Knowing what the future holds for us weather-wise is a critical part of the National Oceanic and Atmospheric Association’s (NOAA) mission. At GDIT, we not only understand that mission, but the satellite, ground and computer systems that enable it as well.

Incredibly, satellites deployed to space a generation ago are still delivering a majority of the weather data used to generate weather forecasts. These systems require careful upkeep and care. As they orbit Earth, an outage – even for a few moments – could mean missing a critical weather development and the opportunity to send out lifesaving warnings, and not getting another look at that spot on the globe for 24 hours.

Ensuring NOAA “never misses” requires deep understanding of its satellites and their components. It requires an understanding of the mission and the environment in order to deploy innovative solutions specifically for that mission and its systems. It requires applying the best of today’s technologies – like machine learning (ML), artificial intelligence (AI) and high-performance computing (HPC), to name a few – alongside the existing systems. And, perhaps most important for NOAA, doing it with zero disruptions or down time.

This is why GDIT increasingly uses machine learning (ML) and artificial intelligence (AI), alongside our knowledge of our customer environments, to predict where an outage might occur. We use performance data to identify trends and patterns (and anomalies) that tell us where to look. We look for failings and potential failings and we intercept them before they become issues or affect weather data collection and forecasting.

In other words, we predict so customers like NOAA can predict.

Right now, we’re working with our partners to build two of the world’s most powerful supercomputers, which will become NOAA’s new Weather and Climate Operational Supercomputing System (WCOSS). GDIT's new WCOSS supercomputers will enable the next generation of American weather forecast models, leading to more accurate and longer range forecasts supporting the National Weather Service's mission to provide forecasts and warnings to protect life and property and to enhance the American economy.

We’re constantly enhancing our ability to predict and prevent issues across IT environments for our customers, making our kind of predictions similarly better, faster and more accurate. This kind of comprehensive, full-circle support for the missions of GDIT’s customers positions us well to support them now and into the future.

In a time of expected budget cuts and re-appropriations amid the continued response to the COVID-19 pandemic, we’re proud to be a partner that can help agencies save money and work more effectively as well. Using AI/ML and HPC, as well as our robotic process automation tools (RPA), we can help customers like NOAA reduce labor costs, conserve staff-hours for more complex tasks and become more efficient at the same time.

By marrying our knowledge of the customer and its systems with our technology expertise, we’re able to help government agencies deliver on their critical missions.