Google’s DeepMind, a new version of its AI weather forecasting model. The company promises that it will provide “more efficient, more accurate and higher-resolution global weather predictions.” For this purpose, it should be able to provide accurate forecasts for up to two weeks, including temperature, pressure and wind information.
According to the researchers, it should also be able to better predict tropical storm tracks. This means that predictions of a storm’s path must be accurate up to three days. The previous model predicted the storm only up to two days in advance.
The model also brings hourly forecasts into the mix. Overall, Google says this new model is eight times faster than the previous version. This will help businesses like energy traders make more accurate decisions, ,
“It gives you more detailed predictions,” said DeepMind AI researcher Aqib Uddin. “Many other industries are very interested in these one-hour steps. It helps them make more accurate decisions. Their goal is, how can they make their business more weather-resilient?”
As noted, the improvements here stem from a new approach to climate models. Older methods required the use of machine learning models that were built for image and video generation. Repeated processing is required to ensure accurate results. The new model requires only one processing step, which also reduces reliance on expensive AI computing systems.
AI may not exist, despite what proponents want you to believe, but it exists. Generally newer models, even those that rely on supercomputers.
However, these models are not perfect, as Google has also acknowledged that WeatherNext 2 will have difficulty predicting outdoor rain and snowfall events. This is due to gaps in the training data. “This is a limitation of our prediction, but we are working on improving it,” said DeepMind research scientist Ferran Alet. bloomberg,
AI weather forecasting is becoming a crowded field. Companies like, and all are contributing to the ring in some way or the other.