We’ve discovered the best way to predict climate over the previous century by understanding the science that governs Earth’s ambiance and harnessing sufficient computing energy to generate international forecasts. But in simply the previous three years, AI fashions from firms like Google, Huawei, and Nvidia that use historic climate knowledge have been releasing forecasts rivaling these created via conventional forecasting strategies.
This video explains the promise and challenges of those new fashions constructed on synthetic intelligence quite than numerical forecasting, notably the flexibility to foresee excessive climate.
Additional studying:
- Here are the papers that describe the fashions talked about within the video.
- Google’s GraphCast
- Huawei’s Pangu-Weather
- Nvidia’s FourCastNet
- Here is the announcement of the ERA5 dataset, launched by the European Centre for Medium-Range Weather Forecasts in 2020.
- We interviewed Aaron Hill over e-mail for this video. Hill is concerned in growing accountable AI for environmental science through AI2ES.
- Google has additionally developed a climate forecasting mannequin known as Nowcasting, which is already embedded in its climate merchandise particularly for short-term precipitation forecasts.
- If you’re considering studying extra concerning the historical past of how we developed climate forecasting, I’d advocate The Weather Machine by Andrew Blum.
You can discover this video and all of Vox’s movies on YouTube.
This video is sponsored by Microsoft Copilot for Microsoft 365. Microsoft has no editorial affect on our movies, however their assist makes movies like these doable.
