Google announces artificial intelligence that can predict floods one week in advance and save 460 million people in 80 countries from flood damage.

Nearly 1.5 billion people around the world, equivalent to 20% of the total population, are at risk of severe damage from floods, and floods cause damage to the global economy estimated at $50 billion (about 7.5 trillion yen) annually. It will reach. AI research by Google that can provide seven days early warning of floods, which are difficult to predict in advance, has been published in the academic journal Nature.

Global forecasting of extreme flooding in unmetered watersheds nature
https://www.nature.com/articles/s41586-024-07145-1

How Google is using AI to improve global flood forecasting
https://blog.google/technology/ai/google-ai-global-flood-forecasting/

Using AI to Expand Global Access to Reliable Flood Forecasts – Google Research Blog
https://blog.research.google/2024/03/using-ai-to-expand-global-access-to.html

Artificial intelligence can help predict floods when traditional methods struggle
https://phys.org/news/2024-03-ai-traditional-methods-struggle.html

Floods due to climate changeIntensificationHowever, there are many rivers around the world that do not have flow meters installed, making it very difficult to predict flood damage, especially in developing countries.

Below is a graph showing the amount of flow meter data on the vertical axis and the GDP of each country on the horizontal axis. It is clear that there is an inverse relationship between the data available within each country and GDP, which means that poor countries have less data that can be used to predict floods.

To use the power of technology to predict natural disasters, a research team led by Gray Nearing of Google Research trained an AI model using data collected from 5,680 flow meters around the world between 1980 and 2023.

Google's AI, developed in this way, can scale flood forecasts using flood nowcasts to 0 days in advance, or on average 5 days before the current day, and can forecast up to 7 days in advance.

Below, forecasts made 0 days in advance by traditional forecasting software Global Flood Awareness System (GloFAS) are shown in blue, and forecasts made by Google AI 0 to 7 days in advance are shown in orange. Google's AI was able to make predictions 5 days in advance with similar accuracy to GloFAS predictions 0 days in advance.

This AI enables Google to send flood forecasts in 80 countries around the world where 460 million people live, and forecast information can be sent to Google Search, Google Maps, Android notifications, etc.Flood axisIt is provided through this.

“There is still significant room for improvement in global flood forecasting and early warning systems. This is extremely important for the well-being of all people around the world,” the researchers wrote in their paper.