Google Unveils AI-Driven Flood Forecasting Service

Google Research has introduced an AI-based technology that enhances flood forecasting worldwide.

The corporation has implemented the service on the Flood Hub platform. This platform is designed to deliver timely alerts to users via Google Search, Maps, and notifications on Android devices.

By training global machine learning models on various weather and geographical data, the system can predict floods even in areas where necessary information is scarce.

Over the past year, Google Flood Hub’s activities have expanded to 80 countries, providing access to reliable natural disaster alerts to 460 million people.

The service covers regions in Africa, Asia, Europe, as well as parts of South and Central America. The accuracy of the alerts has reached the level of forecasts in the United States and Canada.

Testing of the technology began in 2017 with a pilot project in the Patna region of India. Through further research, Google improved its machine learning models, surpassing traditional hydrological simulations.

The forecasting is based on a large-scale AI system that processes a multitude of public meteorological data and information about physical watersheds. It generates predictions up to seven days in advance with high accuracy, providing time for population evacuation and disaster preparedness.

To achieve maximum effectiveness, Google collaborates with the scientific community, governments, and humanitarian organizations. Partnership with the World Meteorological Organization allows for the demonstration of the most accurate results.

The company believes that by enhancing forecasts with neural networks, it is possible to significantly reduce flood damage and save many lives.

In 2022, American scientists announced the MuSiKAL artificial intelligence system, which predicts floods in the Gulf of Mexico region.