Engineers at Rice University have created a deep learning computer system that taught itself to predict extreme weather events up to five days in advance with minimal current weather information.
The self-learning capsule neural network uses an analog method of weather forecasting that computers made obsolete in the 1950s.
Studying hundreds of pairs of maps, each one shows surface temperatures and air pressures at 5km heights and shows those conditions several days apart.
The training includes scenarios that produced extreme weather such as extended hot and cold spells that can lead to deadly heat waves and winter storms.
Once trained, the system could examine maps it had not previously seen and make five-day forecasts of extreme weather with 85% accuracy.
According to Pedram Hassanzadeh, co-author of a study about the system, the system could serve as an early warning system for weather forecasters and as a tool to learn more about atmospheric conditions.
Day-to-day weather forecasts have improved since the introduction of numerical weather predictions (NWP), but they cannot reliably predict extreme weather events such as deadly heatwaves.
Hassanzadeh, an assistant professor of mechanical engineering and earth, environmental and planetary sciences at the university said, “It may be that we need faster supercomputers to solve the governing equations of the numerical weather prediction models at higher resolutions.
“But because we don’t fully understand the physics and precursor conditions of extreme-causing weather patterns, it’s also possible that the equations aren’t fully accurate, and they won’t produce better forecasts, no matter how much computing power we put in.”
Taking a different approach, deep learning allows the computer to be trained to make humanlike decisions with being programmed for them.
Hassanzadeh said, “It didn’t matter that we don’t fully understand the precursors because the neural network learned to find those connections itself. It learned which patterns were critical for extreme weather, and it used those to find the best analog.”
He added that the goal is to extend forecasts beyond 10 days and although it is unlikely to replace NWP, it could be a useful guide.
Hassanzadeh said, “Computationally, this could be a super-cheap way to provide some guidance, an early warning, that allows you to focus NWP resources specifically where extreme weather is likely.”