Google has announced the launch of a new AI-driven tool for forecasting precipitation via Search, marking a first for Africa. By utilizing Google Research’s advanced MetNet model, the company aims to improve the accuracy of weather forecasts in data-limited regions.
New Capabilities for Precipitation Forecasting
Google's MetNet model uses satellite and ground observations to produce state-of-the-art precipitation forecasts. This capability allows for global precipitation prediction every 15 minutes within a 5km radius, processed in under a minute. Such technology is crucial for Africa, where the number of radar facilities is significantly lower compared to other parts of the world.
Leveraging AI in Data-Sparse Conditions
The Google AI team in Accra has been working on addressing the data scarcity issue in Africa. By employing AI and satellite observations, the model can fill data gaps without the need for ground radar. This is Google’s first model incorporating a multitude of satellite observations, overcoming previous regional limitations.
Flood Forecasting with Google
Google has also incorporated AI into flood forecasting, utilizing hydrologic and inundation models to predict affected areas. The Flood Hub platform covers over 80 countries, providing forecasts and alerts to 460 million people via Google Search, Maps, and Android notifications.
The introduction of Google’s MetNet model and Flood Hub platform promises more accurate weather and flood forecasts, which is particularly vital for regions with insufficient traditional observation methods. This is a significant step forward in climate analysis and quality of life improvements in Africa.
