Over the past century, we have learned how to predict the weather by understanding the science that governs Earth's atmosphere and leveraging enough computing power to generate global forecasts. But in the past three years alone, AI models from companies like Google, Huawei, and Nvidia that use historical weather data have issued predictions that rival those made with traditional forecasting methods.
In this video, we discuss the promise and challenges of these new models built on artificial intelligence rather than numerical forecasting, especially the ability to predict extreme weather events.
Additional materials:
- Here is the paper that describes the model mentioned in the video:
- google graph cast
- huawei pangu weather
- Nvidia's FourCastNet
- Announcing the ERA5 dataset released by the European Center for Medium-Range Weather Forecasts in 2020.
- We interviewed Aaron Hill via email for this video. Hill is involved in developing responsible AI for environmental science through AI2ES.
- Google has also developed a weather forecasting model called Nowcasting. This is already incorporated into weather products, especially for short-term precipitation forecasting.
- If you would like to learn more about the history of the development of weather forecasts, we recommend the following: weather forecast machine Written by Andrew Blum.
This video and all of Vox's videos can be found on YouTube.
This video is sponsored by microsoft Copilot for Microsoft 365. Although Microsoft has no editorial influence over our videos, their support makes these videos possible.