When will it rain? Google's new AI will provide precise weather updates and even measure wind speed..

You no longer need to worry about checking the weather. Google has introduced WeatherNext 3, its latest AI model for weather forecasting. It utilizes data from geostationary satellites to generate high-resolution forecasts on an hourly basis. Based on a live review by Brightband, Google DeepMind and Google Research have described WeatherNext 3 as "the most advanced and accurate global weather model to date."

While older AI weather models learned from forecasts generated by traditional systems, this new model learns directly from real-time weather data. Google states that WeatherNext 3 will now power weather information across Google Search, Gemini, Google Maps, Google Maps Platform, and Google Earth Engine.

WeatherNext 3 Provides Accurate Predictions
The most significant upgrade in WeatherNext 3 lies in the level of detail it provides. This Google model can accurately predict key variables like temperature and humidity at a 5-kilometer resolution. Additionally, surface-level data is available at a 10km resolution, while atmospheric factors—such as wind speed—can be tracked at a 25km resolution.

Overall, according to Google, it delivers a "global weather picture that is approximately five times clearer than the previous model, WeatherNext 2." That earlier model provided forecasts on a 25km grid with six-hour intervals. Notably, WeatherNext 3 can refresh its data every hour, making it much easier to track rapidly changing weather conditions such as storms, cyclones, and sudden rainfall.

The most significant improvement in this model stems from its data source. Older numerical models involved a lag of up to six hours, whereas WeatherNext 3 utilizes a comprehensive network of live global geostationary satellite data. The company states that this provides the model with a "continuously updated, live view of the weather." This enables it to generate a new forecast every hour based on the latest satellite data.

The model also utilizes data from ground-based weather stations to account for local factors such as mountains, valleys, and coastal areas. This could prove highly beneficial for regions like Latin America, Africa, and Asia-Pacific, where obtaining accurate, high-resolution forecasts was previously difficult due to the high cost of supercomputer models.

**Model capable of accurate predictions**
Google trained this model using NASA’s IMERG rainfall data and its own global precipitation database. The company states that initial testing showed performance improvements of up to 60% compared to IMERG, 30% compared to MRMS, and 10% compared to rain gauge readings.

Google has also designed WeatherNext 3 to support solar and wind energy planning. The model can accurately estimate wind speed at a height of 100 meters, cloud conditions, and solar radiation. This will help wind and solar project operators determine their potential power generation output.

WeatherNext 3 is now being integrated directly into Google’s apps and services. The company notes that when users plan one or two days in advance, they will see rainfall forecasts that are "up to 50% more accurate." This will be particularly useful in regions where obtaining reliable weather information was previously challenging.

Disclaimer: This content has been sourced and edited from NDTV India. While we have made modifications for clarity and presentation, the original content belongs to its respective authors and website. We do not claim ownership of the content.