Microsoft Unveils New Machine Learning Tool to Improve Weather Forecasting
Physics-based weather forecasts, and even newer AI models, typically become unreliable after about two weeks. Microsoft claims it has solved this problem using machine learning. The company's researchers say they have found a new, cost-effective method for accurate weather prediction two to six weeks in advance, allowing farmers and emergency responders to […]
Posts in this thread will be added to your daily email summary and home page feed.
New open source software could make weather forecasting with artificial intelligence more affordable.
Posts from this author will be added to your daily email digest and home page feed.
is a senior science reporter covering energy and the environment with more than a decade of experience. He is also the host of Hell or High Water: When Disaster Hits Home, a podcast from Vox Media and Audible Originals.
Weather forecasts based on old physics and even new AI models tend to become less reliable after two weeks, a problem Microsoft claims to have solved using machine learning.
The company's researchers say they have found a new, cost-effective way to more accurately predict weather two to six weeks in advance - allowing farmers and emergency responders to plan with greater confidence. The findings were published today in the journal Nature, and the tool is now open source software.
Microsoft joins Google, Nvidia and other companies developing artificial intelligence tools for weather forecasting. Instead of creating a new stand-alone climate model, Microsoft's latest research in this area uses machine learning to correct biases in existing models. The researchers hope this is a more accessible way for resource-constrained weather agencies to improve accuracy using artificial intelligence.
This is a more accessible way for resource-constrained weather agencies to improve accuracy using artificial intelligence
While physics-based modeling is a resource-intensive endeavor that runs on supercomputers, and generative AI relies on powerful GPUs, Lester McKee, a senior research fellow at Microsoft New England, says the new tool can run on just a CPU. "I have high hopes for both increased productivity of weather forecasting resources and greater reliability of forecasts, especially for extreme weather," Mackey tells The Verge.
Physics-based and AI-based models face similar problems when forecasting weather for two weeks or more. Traditional models try to simulate the complex physics of Earth's atmosphere, while AI models look for patterns in historical data to make predictions. In both approaches, models rely on approximations to fill data gaps, whether from regions with sparse weather stations or complex atmospheric phenomena that scientists are still working to understand. The ways these models compensate for those gaps in the data can become a form of systematic bias. When making more predictions in time, the resulting errors compound, making those predictions more accurate.
Microsoft calls its method Probabilistic Bias Correction (PBC). This tool uses machine learning to identify systematic bias in a model and make corrections. One of the examples Mackey gives is climate change mitigation. Greenhouse gas emissions from fossil fuels are warming the planet, but a climate model might make predictions based on the historical record that could push it toward colder temperatures than the world is experiencing today.
The researchers found that PBC improved the accuracy of the leading physics-based meteorological model by the European Center for Medium-Range Weather Forecasts (ECMWF) for 98% of precipitation and 92% of temperature targets. The machine learning framework also doubled the accuracy of ECMWF's new artificial intelligence forecasting system in the study's analysis. Before the paper was published, ECMWF had already awarded the authors' research on PBC in the Weather Quest artificial intelligence competition.
The new paper on PBC, published by researchers from Microsoft, MIT, ECMWF and other institutions, "is a good example of new ideas being developed for the benefit and attention of the community," Matthew Chantry, ECMWF's strategic lead for machine learning, said in an emailed comment to The Verge.
"As always, turning ideas into reality [in] an operational process is the most challenging step," added Chantry.
The amount of computation and infrastructure required for weather forecasting often puts technological advances out of reach for many agencies in economically developing countries. However, the authors of the paper think that PBC can be different. "One of the things that's really exciting to me about models like PBC is how light they are, how easy they are to train, that scientists without massive resources can work on trying to configure [these] models and build them," says Genevieve Flaspohler, co-author of the paper.
Flospoller worked with Mackey on the PBC as an intern at Microsoft and is now executive director of the nonprofit Rhiza Research, which is working with the Kenya Meteorological Service — funded by the Gates Foundation — to see if artificial intelligence can help people get more accurate early warnings. PBC is one of the methods they are testing ahead of a more intense rainy season during a record-breaking El Niño.
I tried Apple's HomeView smart display
Everything Apple released at its "Welcome Home" event
The Tesla Roadster could have remained vaporized
A free daily digest of the news that matters most.
By providing your information, you agree to our Terms of Use and Privacy Policy. We use vendors who may also process your information to help provide our services. This site is protected by reCAPTCHA and Google's Privacy Policy and Terms of Service apply.