Google unveils GraphCast, an “unprecedented” AI weather forecaster
The forecast for this new weather program looks good.
Google DeepMind, the search engine’s AI-focused lab, has launched a groundbreaking weather forecasting model that aims to deliver 10-day forecasts in under a minute – “with unprecedented accuracy”.
GraphCast outperforms traditional and high-tech weather models by at least 90% because it focuses on historical weather data. according to a research paper on the AI model released on Tuesday.
Google promotes hyperlocal data with more than a million global grid points. GraphCast requires two pieces of information: the state of the weather now and six hours earlier, as the weather is predicted six hours in the future and up to ten days in advance.
“We believe this represents a turning point in weather forecasting,” the Science article says, “helping to open new avenues to strengthen the breadth of weather-dependent decision-making by individuals and industries by making cheap forecasts more accurate and become more accessible.” and suitable for specific applications.”
The model was tested against the gold standard High RESolution Forecast (HRES) system of the European Center for Medium-Range Weather Forecasts.
GraphCast was found to be 99.7% more accurate in some cases.
The ECMRWF is I’m already experimenting with the toolthanks to open source coding that anyone can use.
The system is also described by Google as a breakthrough in detecting extreme weather events before they happen.
During training, it was able to predict cyclone motion more accurately than HRES. In September, it was also predicted that Hurricane Lee would reach Nova Scotia nine days before landfall.
According to Google, traditional models took three additional days.
Additionally, GraphCast excels at predicting flooding using atmospheric data on rainfall patterns.
What sets this interface apart from traditional approaches is that it does not run algorithms and mega-equations like those of supercomputers. Instead, deep learning is used.
“Deep learning offers a different approach: using data instead of physical equations to build a weather forecasting system,” wrote GraphCast team member Remi Lam. “GraphCast is trained on decades of historical weather data to learn a model of the cause-and-effect relationships that determine how Earth’s weather evolves from the present into the future.”
Meanwhile, Google recently released a new, separate 24-hour forecast model called MetNet-3 this also delivers incredible accuracy.
“The groundbreaking use of AI in weather forecasting will benefit billions of people in their everyday lives,” Lam added. “But our broader research isn’t just about predicting the weather – it’s about understanding the broader patterns of our climate.”
The use of AI in weather forecasting is described by experts as a “quiet revolution” that is significantly changing the industry. Science.org reported.
“It’s very, very exciting to know that we can make global forecasts that are skillful and really cost-effective,” Maria Molina, an AI-focused research meteorologist at the University of Maryland, told the outlet on Tuesday.
And that’s just the beginning, says Christopher Bretherton, an atmospheric scientist at the Allen Institute for AI.
The institute’s 40 years of training data could be instrumental in catalyzing new models on which future AI will be trained, says Bretherton, “and then running them 100 times faster.”
