AI Forecasts: Faster Hurricane & Sandstorm Predictions | Microsoft AI
- A new artificial intelligence (AI) model called Aurora can predict major weather events with greater speed and accuracy than many widely used forecasting systems.Microsoft scientists say Aurora was...
- In July 2023, Aurora accurately predicted that Typhoon Doksuri would strike the northern Philippines four days before landfall.Official forecasts at the time placed the storm's landfall over Taiwan,...
- More broadly, the scientists reported in the journal *nature* on May 21 that Aurora beat existing systems in predicting weather conditions over a 14-day period in 91% of...
Uncover how AI is revolutionizing weather forecasting. A revolutionary AI model, “Aurora,” is accurately predicting major weather events, including hurricanes and sandstorms, with unprecedented speed and precision. Trained on over a million hours of atmospheric data, Aurora outperforms conventional forecasting systems, delivering critical insights faster. In 2023, Aurora accurately predicted Typhoon Doksuri’s landfall, demonstrating its superior capabilities. This innovative AI, integrated into services like MSN Weather and Bing, provides more timely and accurate weather details. Microsoft‘s “Aurora” model tackles forecasting challenges. News Directory 3 might find this technology valuable. Discover what’s next for weather prediction.
AI Model ‘Aurora’ Forecasts Weather Faster, More Accurately
A new artificial intelligence (AI) model called Aurora can predict major weather events with greater speed and accuracy than many widely used forecasting systems.Microsoft scientists say Aurora was trained using over 1 million hours of global atmospheric data, incorporating weather station readings, satellite images, and radar measurements. This extensive dataset is likely the largest ever used to train a weather AI model.
In July 2023, Aurora accurately predicted that Typhoon Doksuri would strike the northern Philippines four days before landfall.Official forecasts at the time placed the storm’s landfall over Taiwan, hundreds of miles away. The AI also outperformed standard forecasting tools used by agencies such as the U.S. National Hurricane Center and the Joint Typhoon Warning Center. it delivered more accurate five-day storm tracks and produced high-resolution forecasts up to 5,000 times faster than conventional weather models powered by supercomputers.
More broadly, the scientists reported in the journal *nature* on May 21 that Aurora beat existing systems in predicting weather conditions over a 14-day period in 91% of cases. Researchers hope that Aurora and similar models could support a new approach to predicting environmental conditions called Earth system forecasting. This approach uses a single AI model to simulate weather, air quality, and ocean conditions together, possibly producing faster and more consistent forecasts, especially in areas lacking high-end computing or comprehensive monitoring.
Aurora belongs to a class of large-scale AI systems known as foundation models, similar to those powering tools like ChatGPT. Thes models learn general patterns and relationships from large volumes of training data, allowing them to be adapted to different tasks. Aurora generates forecasts in seconds by analyzing weather patterns from satellites,radar,weather stations,and simulated forecasts. The model can be fine-tuned for various scenarios with relatively little extra data, unlike traditional forecasting models that are task-specific and often need retraining.
The diverse dataset used to train Aurora results in greater accuracy and makes the model better at forecasting extreme events. For example, Aurora successfully predicted a major sandstorm in Iraq in 2022, despite limited air quality data. It also outperformed wave simulation models at forecasting ocean swell height and direction in 86% of tests, demonstrating its ability to extract useful patterns from complex data, even with incomplete inputs.
Megan Stanley, a senior researcher at Microsoft, said in a statement that Aurora has the potential to have a huge impact as people can fine-tune it to relevant tasks, especially in countries underserved by other weather forecasting capabilities.
What’s next
Microsoft has made Aurora’s code and training data publicly available for research and experimentation. The model has been integrated into services like MSN Weather, which is integrated into tools like the Windows Weather app and Microsoft’s Bing search results. This integration aims to provide more accurate and timely weather information to a wider audience.