Google has used artificial intelligence to improve global flood forecasting and was able to accurately predict some floods up to seven days in advance by training machine learning models. Google offers accurate flood forecasting in 80 countries.

One of the reasons why it can be difficult to predict floods ahead of time is the fact that most rivers do not have what is known as a flow meter. These metrics help provide relevant data such as information on rainfall and physical catchment area. However, the tech giant was able to circumvent this problem by using its ML technology with all available river data and applying the ML model to watersheds where data is missing.

The company’s findings were published in the journal Nature, which further confirms how impressive this achievement is. For context, Nature is a leading scientific journal that publishes peer-reviewed research.

The ultimate goal of the technology is to scale the accuracy of flood forecasting to a global level, even in areas where local data is not available. Google has been able to forecast floods for more than 80 countries through its Flood Hub. It also delivers alerts in Google Search, Maps and Android notifications.

Source: Android Authority