How Google’s Weather Forecast AI Can Replace Meteorologists – Times of India

of Google DeepMind Developed an AI-powered weather forecasting model, GraphCast, which can make a 10-day forecast in one minute. According to research by scientists, GraphCast Exceeded the accuracy of traditional weather forecasting technologies with a 90% verification rate.
The GraphCast weather forecast software uses the two most recent Earth weather conditions, including variables from the current time and 6 hours ahead, to predict the weather conditions six hours and 10 days ahead.
Google boasts over a million global network points for hyperlocal data, and GraphCast only needs two pieces of information to make a prediction.
The study claims this represents a turning point in weather forecasting, allowing for cheaper, more accurate and more accessible forecasts that can be tailored to specific applications. This, in turn, can help individuals and industries make better weather-based decisions.
The GraphCast model was compared to the High Resolution Forecast (HRES) of the European Center for Medium-Range Weather Forecasts, which is considered the gold standard system. In some cases, GraphCast was found to be 99.7% more accurate.
The tool is open source, allowing anyone to use it, and ECMWF is already experimenting with it.
Google said the system can detect extreme weather events before they arrive. During training, GraphCast was able to predict cyclone movement more accurately than HRES. In September, Hurricane Lee was predicted to make landfall in Nova Scotia Nine days before arrival, compared to three days before traditional models.
The model has the ability to predict severe weather events such as tropical cyclones and extreme heat waves that occur in different regions. What's more, the algorithm can be retrained with more recent data, which scientists believe will make the tool more accurate in predicting changes in weather consistent with climate change.
Google is looking into how it can incorporate GraphCast into its products. The National Oceanic and Atmospheric Administration (NOAA) has been working to develop models that will give more accurate readings during severe weather events and, more importantly, predict the intensity of storms.