Amid the growing threat of floods globally, researchers at Google have developed a flash flood forecasting tool designed to help predict the risk of flash floods in urban areas up to 24 hours in advance.
The tool, available through Flood Hub, uses satellite data to generate Synthetic Aperture Radar (SAR) images. These images are then processed by algorithms to calculate and map flooded areas.
Google says the system relies on an artificial intelligence-driven approach that combines data from Google Maps and capabilities from Gemini to estimate the areas affected by flooding and produce rapid flood maps.
Through a methodology known as Groundsource, the company analysed decades of public reports containing flood data. The analysis identified more than 2.6 million historical flood events across over 150 countries.
“It then used Google Maps to determine precise geographic boundaries for each event to create a dataset focused on flash floods,” the company said.
However, Google cautions that the accuracy of the Flood Hub tool can sometimes be limited due to challenges in interpreting satellite imagery.
“It’s difficult to recognise flooded and non-flooded areas from a satellite image, so there may sometimes be errors,” the company noted.
In some cases, the platform may not display a flood map at all, particularly when satellite images are unavailable or when the images captured do not produce a qualified flood map.
For the flood mapping feature to function, Google says updates are made only when new satellite images become available.
“After an image is taken, it takes several hours until it’s in our systems. It can also take several days to produce a satellite image that we can use,” the company explained.
These data availability challenges could sometimes affect the reliability of the tool. Google adds that if the latest available satellite image is older than 72 hours, the platform stops displaying the flood map until a new image becomes available.
