The Project’s overall goal, that also coined its acronym REPORT-DL ("Rapid Earthquake Phase analysis of Ocean-bottom, Regional and Teleseismic events with Deep Learning"), is to improve the assessment of earthquake related hazards worldwide. A thorough analysis of the huge amount of data acquired over the last decades, by seismological observatories worldwide, is hoped to allow the identification of the earth’s faintest and most unusual signals. Therefore, improving the thus derived earthquake distribution, which conveys information about the stress condition in the earth’s crust and therefore allows the geohazard assessment.

To optimise said identification process, Scientists are adapting state of the art AI procedures commonly used in image and natural language processing to recognise and time recorded signals. Due to this automatization of signal recognition, large quantities of archived data or even live stream Data can be scanned for seismic signals in a quick and systematic manner. The main proponent is the KIT (Karlsruhe Institute for Technology).

GEOMAR is involved in the algorithm’s implementation for ocean bottom seismometers, whose records not only show little earthquakes, but also the movement of the ocean floor as well as signals from the water column, such as wave motions or ship and whale noises.

For more information on REPORT-DL