Google has expanded its AI-driven flood forecasting system to reach 700 million people in 100 countries, aiming to strengthen preparedness in regions prone to riverine flooding. The enhanced system, powered by advanced AI, delivers real-time flood predictions to help communities better prepare for potential disasters.

In a recent blog post, Google shared details on this global expansion, which now includes an additional 300 million people beyond its previous coverage of 400 million across 80 countries. The growth is supported by significant improvements in AI, enabling Google to process vast amounts of labeled data to refine its flood prediction models.

The upgraded system incorporates a “frontier forecasting model” that enhances accuracy, allowing Google to predict floods up to seven days in advance, a notable improvement from the previous five-day forecast. This extended warning period is designed to give communities more time to act, potentially saving lives and reducing property damage.

To support researchers and organizations, Google is offering access to its flood data in two ways. First, Google’s datasets from its Google Runoff Reanalysis & Reforecast (GRRR) are now accessible to partners. Additionally, a new API allows seamless integration of Google’s flood forecasting data for further research and collaboration.

Through the API, users can access critical information about anticipated flood status and hydrology forecasts, even in areas with limited local data. Researchers and organizations interested in using this resource can join a waitlist to access Google’s flood data.

Google has also introduced a new data layer to its Flood Hub, featuring more than 250,000 virtual gauges worldwide. These gauges, using atmospheric and geological data, provide detailed insights into flood risks. While Google’s system models flood risks globally, only verified data from regions with reliable historical records are displayed.

By expanding its forecasting capabilities and enhancing data-sharing, Google aims for its flood prediction system to become a valuable tool for public authorities. The goal is to help governments develop effective emergency response plans to protect communities and infrastructure from the impacts of flooding.