AI Case Study

The US National Oceanic and Atmospheric Administration is using AI to improve real-time decision-making for high-impact weather

The US National Oceanic and Atmospheric Administration is leveraging machine learning to improve high-impact weather forecats. Better accuracy can help avoid and prevent disruptions and accidents and also exploit benefits such as the generation of renewable energy.


Public And Social Sector


Project Overview

"The US National Oceanic and Atmospheric Administration (NOAA) has recently been using machine learning more to improve their forecasts. A group of researchers from the NOAA found that “applying AI techniques along with a physical understanding of the environment can significantly improve the prediction skill for multiple types of high-impact weather.” High-impact weather includes events like severe thunderstorms, tornadoes, and hurricanes.

Their paper concluded these improvements have clear commercial applications stating, “AI methods extend easily to directly predicting impacts of high-impact weather, such as power generated by variable sources such as solar or wind, energy consumption in an area, or airport arrival capacity.”

One example the paper highlighted is that machine learning can provide more accurate hail forecasting. Hail cause billions of dollars damages every year. Even a modest improvement in hail warning could produce significant savings by getting individuals move their cars and themselves to safety. Provide these types of warnings for car insurance companies is one way IBM is commercializing their weather predictions." (techemergence)

Reported Results

"This paper raises the interesting question of the role of automated guidance in forecasts. While we have demonstrated that AI/data science techniques can be used to significantly improve forecasts in a va- riety of high-impact weather domains, it is not simply a matter of bringing these techniques to operations. The forecasters must be able to trust the forecast pro- duced by such techniques, as has been demonstrated in the HWT/PHI experiments (Karstens et al. 2016).

By studying the error characteristics of different machine-learning methods in high-impact weather situations, researchers and forecasters can identify when the automated guidance should be trusted and when it is more likely to struggle." (ametsoc)


"We tested three machine-learning methods: GBRT, RF, and elastic nets." (ametsoc)





"High-impact weather events, such as severe thunderstorms, tornadoes, and hurricanes, cause significant disruptions to infrastructure, property loss, and even fatalities. High-impact events can also positively impact society, such as the impact on savings through renewable energy. Prediction of these events has improved substantially with greater observational capabilities, increased computing power, and better model physics, but there is still significant room for improvement. (ametsoc)

"Government institutions are still the major players in weather forecasting, both as a source of raw data used by private companies and as a source of predictive models. The Department of Commerce found that in 2014 all federal spending on meteorological operations and research was $3.4 billion while private sector spending on weather forecasting reached only $1.7 billion." (techemergence)