Resources

Driven by the potential global impact of reliable flood forecasting on human lives, we’ve developed specialized resources to support National Meteorological and Hydrological Services (NMHSs), governments, Intergovernmental Organizations (IGOs), and Non-Governmental Organizations (NGOs) in alignment with the Early Warnings for All initiative. By working closely with these partners, our goal is to ensure that agencies and communities have access to life-saving information to remain safe, prepared, and resilient during crises.

The following tools provide immediate access to current conditions and forecast data to support emergency management and public safety:

• Flood Hub: A platform for surfacing real-time forecasts for riverine floods up to 7 days in advance and flash floods in densely populated areas up to 24 hours in advance. You can read about our scientific publications here.
• Email notifications: A dedicated service to receive timely updates and alerts regarding flood risks straight to your inbox in countries of your interest.
• Flood Hub help center: A comprehensive resource for understanding the platform's features and utilizing its data effectively.
• Our Models: Read more about how the models work.

For governments and organizations requiring automated data ingestion or deep historical analysis, the following resources are available:

• API Access: Request programmatic access to real-time flood forecasts to integrate directly into local systems. Currently, the use of the flood forecasting API is primarily limited to non-commercial use. Please note, a Google account is required in your application to access the API. Information provided by the API is licensed under CC BY 4.0.
• Google Runoff Reanalysis & Reforecast (GRRR) Dataset: Access global river discharge estimates from 1980 to 2023.
• Inundation History Dataset: Explore historical data on how often each 128-meter pixel in the world has been wet between 1999 and 2020, derived from satellite imagery. The dataset is available under a CC-BY-4.0 license.

We support the global hydrology community through open-source datasets and research tools to improve predictive modeling:

• Open Sourced Model Code: To help further protect vulnerable communities affected by floods, we have open-sourced our hydrology model code on GitHub for others to use and build upon. To get started with implementation, check out this introductory tutorial on navigating the model code. The model code is available under an Apache 2.0 license.
• Caravan: An open-source global streamflow dataset and community resource designed for large-sample hydrology research and training AI prediction models. Several countries are currently represented in this dataset, including (but not limited to): Australia, Brazil, Canada, Chile, Denmark, Mexico, the UK, the US, and much of central Europe. Moreover, the World Meteorological Organization-supported Global Runoff Data Center contributed streamflow data for 25 additional countries. We encourage countries to contribute to this open source dataset to improve hydrological abilities.
• Caravan MultiMet: An extension of the Caravan dataset that enriches streamflow data with weather forecasts for more robust modeling.
• Groundsource Flash Flood Dataset: Groundsource is an open global dataset of 2.6 million historical flood events derived from news articles spanning more than 150 countries.

• Public Alerts: Learn more about how we surface authoritative alerts on our platforms using the Common Alerting Protocol.
• Wildfires: Read about our wildfire boundary tracking model available in 30+ countries and about how we’re supporting breakthroughs in wildfire detection through FireSat.
• WeatherNext 2: Explore our state-of-the-art mid-range forecasting models, including GenCast and GraphCast.
• Cyclones Prediction: Explore AI-powered tropical cyclone forecasting. Access real-time predictive datasets, available for visualization and download via Weather Lab.
• Open Buildings: Discover large-scale building footprints detected via satellite imagery.
• Android Earthquake Alerts System: Read more about how we provide early warnings by partnering with regional seismic networks in the U.S. and utilizing a global, crowdsourced network of Android phone accelerometers to detect vibrations and alert users before shaking begins.