On September 3rd, Google released its WeatherNext 3 model, which is capable of predicting wind speed, cloud cover, sunlight, temperature, and moisture. It sports a 5-kilometer resolution for variables and 1-hour update, and that’s a significant evolution from the 25-kilometer resolution and 6-hour update of the previous model.
The Significance of an Accurate AI Weather Forecasting Model
The companies in the energy sectors are dealing with increasingly harder-to-predict systems when it comes to solar and wind energy. Each forecast suggests how the grid operators should act. For example, if they predict lower winds, they must buy energy.
Every underestimation or overestimation leads to costs, and those costs are not at all small. An accurate AI weather forecasting model could help the energy sector lowering the costs of forecasting errors.
The New Google WeatherNext 3
Other competitive companies are already selling their forecast to the energy market players. The advantage Google has is reach. The model is part of the Earth Engine, Maps, and Search. Thus, the spread Google has, along with the 1-hour updates, put WeatherNext 3 into the top-tier of the world’s weather forecasting models.
Even more, Google WeatherNext 3 drives its forecasts from real observations, not from simulations as WeatherNext 2 and other similar models did. The new architecture takes live images from geostationary satellite and uses weather stations readings for training. You can read more about it here.
Google and AI Vs. Energy Sector — The Paradox
As we speak, the data centers scattered around the world, of tech and AI companies, are among the most significant energy consumers. Google itself contributed a lot to that. However, the majority of these high-tech companies switched to regenerable energy. And now, with a more accurate AI weather forecasting model, such as WeatherNext 3, it can give back something more to the energy sector.