Deltares, an independent institute for applied research in the field of water and subsurface, presented a new method to evaluate flood extent and permanent water by combining several remote sensing data sets in Google Earth Engine, Landsat, and Shuttle Radar Topography Mission.

With this new method, flood maps all over the world (with details up to 30 meters) can be produced, going back up to 1999. The research, that was performed by Deltares researcher Gennadi Donchyts, was presented at the 2015 American Geophysical Union fall meeting in San Francisco in December.

In recent years Deltares has been a heavy user of Google Earth Engine, which eliminates the need to manually download and process satellite data, allowing researchers to process large amounts of satellite data faster than before.

Deltares has developed an algorithm that makes it is possible to extract the required information from satellite data without heavy downloads or local calculation capacity. Deltares combines this information with knowledge about the flow of water over the terrain and OpenStreetMap open GIS data. The storage and calculation are arranged by Google through its Earth Engine.  The research was performed based on two locations: The Mekong River delta and Murray-Darling basin in Australia.

Next step: global level

This research received funding from the FP7 project eartH2Observe, the SERVIR Mekong project, and is performed in cooperation with the Delft University of Technology. The next steps will be to collect data from other satellites to increase the accuracy of the maps and extend the algorithm on a global level.

“The research shows that it will now be possible, with only one single application, to look at any location in the world (zoomed in until 30 meters) and see where it was flooded in the last 15 years,” said project coordinator Jaap Schellekens of Deltares.

“The flood maps we create are not based on single images, but on a time series of images. Once a person has access to Google Earth Engine, the scripts and the supporting data sets the process of creating a maps is a simple matter of zooming to the area of interest using the Google interface and setting the period for analysis,” he said. “This will create the maps on the fly. A lot of time (and cost) can be saved.”