Graph Cast: A Revolutionary Machine Learning for Weather Forecasting
Google DeepMind, an AI firm, has designed a machine-learning model called Graph Cast that is able to make more accurate weather forecasts in minutes than conventional models in hours. It has been identified as the leading AI model in the field.
Numerical weather predictions, which are based on physical principles, are expensive and energy-intensive to run.
Graph Cast uses past and current weather data to predict the future state of global weather. It can make predictions up to 10 days in advance in under a minute and outperforms the ECMWF's NWP model.
It was trained using weather estimates from 1979 to 2017 and can accurately predict the weather in advance. It has also been found to be useful in predicting several weather events.
Graph Cast predicted the state of 5 weather variables close to the Earth’s surface, such as the air temperature 2-metres above the ground, and 6 atmospheric variables, such as wind speed, further from the Earth’s surface.
Therefore, several tech companies have developed machine-learning models that are energy-efficient and make rapid predictions from past and current weather data but Graph Cast outperforms both AI and conventional approaches in most global weather forecasting tasks.
However, machine-learning models may amplify biases in their training data and require a lot of energy for training, although less than NWP models. It could take two to five years before the public can make decisions based on AI forecasts.
Comments