How Google uses its AI to predict floods
Google uses AI technology, including Flood Hub and Groundsource, to predict riverine and urban flash floods globally. The initiative aims to provide early warnings to keep people safe from natural disasters.
Why it matters
Early flood alerts enable individuals to evacuate and aid organizations to deliver life-saving support before disasters strike. This can reduce food insecurity and help families protect their assets.
The details
- Riverine floods are predicted up to seven days and urban flash floods 24 hours ahead.
- Groundsource analyzed 5 million news reports to create a dataset of 2.6 million events.
- The system provides predictions for over 2 billion people across 150 countries.
Show entities and relationshipsHide entities and relationships
In this article
Organizations
Technologies
Companies
People
Key connections
Google owns Groundsource
Google owns Hydrologic Model
Google owns Inundation Model
Flood Hub uses Hydrologic Model
Flood Hub uses the Hydrologic Model to forecast river water flow.
Show 9 more connectionsShow fewer connections
Flood Hub uses Inundation Model
Flood Hub uses the Inundation Model to predict which areas a river flood will affect.
Flood Hub uses Groundsource
Flood Hub integrates the Groundsource dataset to power its urban flash flood predictions.
Groundsource uses Gemini
Groundsource used Gemini to read over 5 million news reports about flooding to build its historical flood event dataset.
Deborah Cohen works at Google Research
Deborah Cohen leads the Flood Forecasting work on Google Research's Climate Crisis Resilience team.
Flood Hub is built with Artificial Intelligence
Flood Hub uses AI to generate flood prediction alerts.
Give Directly uses Flood Hub
Give Directly uses Flood Hub to understand where floods will occur and who is at risk.
Give Directly uses Flood Forecasting API
Give Directly used the Flood Forecasting API to deliver cash to people in Kogi, Nigeria before flood waters rose.
Give Directly is related to Nigeria
Give Directly delivered early cash assistance to flood-affected communities in Kogi, Nigeria.
Google piloted an early flood forecasting model in India in 2018.
Related events
Google announces Groundsource AI methodology for flood data
Get the weekly recap
The stories like this one, picked and explained — once a week, straight to your inbox.