Aardvark, a new AI weather prediction model from the University of Cambridge, promises to revolutionize how weather forecasts are made, delivering faster and more accurate predictions than ever before.
Weather forecasting is on the brink of a monumental shift thanks to a groundbreaking AI model named Aardvark. Researchers from the University of Cambridge, alongside the Alan Turing Institute, Microsoft Research and the European Centre for Medium Range Weather Forecasting (ECMWF), have developed this innovative model, promising faster and more accurate weather predictions while using significantly less computing power.
The research is published in the journal Nature.
Traditionally, weather forecasts are generated via a complex multi-stage process that relies on bespoke supercomputers, taking hours to deliver results.
However, Aardvark has streamlined this cumbersome pipeline into a singular machine learning model that processes data from satellites, weather stations and various sensors to produce global and local forecasts within minutes on a standard desktop computer. This efficiency could potentially transform weather prediction practices worldwide.
Using only 10% of the input data required by existing systems, Aardvark has outperformed the U.S. National Global Forecast System (GFS) on multiple variables and matches the U.S. Weather Service’s forecast accuracy, which traditionally incorporates data from numerous weather models analyzed by expert forecasters.
One of the standout features of Aardvark is its adaptability. Its simple design allows it to be easily customized for specific industries or regions, whether predicting temperatures for African agriculture or wind speeds for European renewable energy companies. This flexibility starkly contrasts with traditional systems that require years of work by large research teams to develop tailored models.
Aardvark’s potential is particularly transformative for developing countries, which often lack the expertise and computational resources necessary for conventional weather prediction systems.
“Aardvark reimagines current weather prediction methods offering the potential to make weather forecasts faster, cheaper, more flexible and more accurate than ever before, helping to transform weather prediction in both developed and developing countries,” Richard Turner, a professor of machine learning in Cambridge’s Department of Engineering and a lead researcher for weather prediction at the Alan Turing Institute, who led the research, said in a news release.
First author Anna Allen from Cambridge’s Department of Computer Science and Technology noted the broader application possibilities: “These results are just the beginning of what Aardvark can achieve. This end-to-end learning approach can be easily applied to other weather forecasting problems, for example hurricanes, wildfires and tornadoes. Beyond weather, its applications extend to broader Earth system forecasting, including air quality, ocean dynamics and sea ice prediction.”
Moreover, Turner emphasized the collaborative effort: “Aardvark would not have been possible without decades of physical-model development by the community, and we are particularly indebted to ECMWF for their ERA5 dataset which is essential for training Aardvark.”
Matthew Chantry, strategic lead for machine learning at ECMWF, expressed the significance of the collaboration: “It is essential that academia and industry work together to address technological challenges and leverage new opportunities that AI offers. Aardvark’s approach combines both modularity with end-to-end forecasting optimization, ensuring effective use of the available datasets.”
Bishop from Microsoft Research AI for Science highlighted the collaborative spirit: “Aardvark represents not only an important achievement in AI weather prediction but it also reflects the power of collaboration and bringing the research community together to improve and apply AI technology in meaningful ways.”
Reflecting on the societal impact, Scott Hosking of The Alan Turing Institute said, “Unleashing AI’s potential will transform decision-making for everyone from policymakers and emergency planners to industries that rely on accurate weather forecasts. Aardvark’s breakthrough is not just about speed, it’s about access. By shifting weather prediction from supercomputers to desktop computers, we can democratize forecasting, making these powerful technologies available to developing nations and data-sparse regions around the world.”
Moving forward, the Alan Turing Institute aims to establish a new team under Turner to explore deploying Aardvark in the global south and integrate it into broader efforts for high-precision environmental forecasting.
Source: University of Cambridge