Thanks to artificial intelligence, a new age in weather forecasting has begun. Google DeepMind’s AI research department is making significant progress in the field of weather prediction with its machine learning model. The model, known as GraphCast, claims to surpass 90% of benchmarks used by the most advanced weather prediction systems, providing accurate forecasts in just seconds.
The AI model has been trained on nearly 40 years’ worth of historical weather data, allowing it to increase its accuracy. It took four weeks and thirty-two computers to train the algorithm. However, the generated algorithm can now forecast weather up to ten days ahead of time in less than a minute on a single desktop computer.
GraphCast exhibits better accuracy compared to existing weather systems on 90% of 1,380 parameters. Its capabilities extend beyond regular weather forecasts, as it excels in predicting catastrophic weather occurrences such as excessive temperatures and tracking tropical storms.
The traditional weather forecasting process relies on supercomputers, which perform complex computations based on data from buoys, satellites, and weather stations. This method is expensive and time-consuming, with the most accurate weather forecasts being produced every six hours by the European Centre for Medium-Range Weather Forecasts in Italy.
GraphCast’s superiority in weather prediction stems from its ability to utilize previous meteorological data. The AI can detect patterns that may be difficult to perceive in equations, leading to more accurate and affordable weather predictions. It is reported that GraphCast’s energy efficiency is around 1,000 times less expensive than traditional weather forecasting techniques.
The application of AI weather forecasting extends to addressing global challenges such as climate change. Google is an active partner in the World Economic Forum’s AI Governance Alliance, ensuring responsible development and implementation of this technology. GraphCast can play a crucial role in predicting extreme weather events caused by climate change, aiding efforts to combat global warming.
With technology evolving rapidly, AI weather forecasting brings a promising future for accurate and timely weather predictions. The GraphCast model’s ability to process data quickly and efficiently, surpassing traditional methods, opens doors for improved decision-making and preparedness. As scientists continue to innovate in this field, the world can benefit from more precise forecasts that help mitigate the impacts of weather-related events.
