Google has announced it will be expanding its AI flood forecasting model to 100 countries now. It will also make their model forecasts available to researchers via an API along with their datasets to help them understand the impact of riverine floods historically. 

A blog post with the news shared that now the model will be covering 700 million people compared to 80 countries and a population of 400 million people previously. 

Google has also worked to improve their model so it is accurate with a seven-day lead time while the older version offered a five-day lead time. 

Users can access the hydrology forecasts via the API and know about the expected flood status for both urban areas as well as parts where data is limited. 

Interested users can sign up and join the waitlist to access the data. 

The model uses what it calls “virtual gauges” for locations in case data is scarce, which are simulation-based prediction systems that use geological and atmospheric factors to predict flooding. 

It and has added a new expert data layer on the Flood Hub with close to 250,000 forecast points of our Flood Forecasting model.