Using AI for Earth and Ocean Observation Using Telecommunication Fiber Optic Cables
SECOND CALL
APPLICATION ID: ALL13
What we are looking for:
We seek innovative research proposals based on AI methods to utilize ground motion data acquired on fiber optic cables for multiple applications, such as earthquake detection, tsunami warning, identification of ships, marine mammals, and other acoustic signals.
The context:
DAS (distributed acoustic sensing) is a new type of instrumentation that allows to convert existing fiber optic cables into thousands of sensors of ground deformation, temperature, etc. The use of this technology in existing submarine telecommunication is an inexpensive way of acquiring high-resolution continuous seismic and oceanographic data from the seafloor, which were previously unavailable and provides a wealth of information on ocean processes.
The problem to address:
DAS is a recent technology and therefore methodologies to process their data are still being developed. This is both a great opportunity for innovative research and a challenge. And additional difficulty arises from the large volumes of data obtained in DAS acquisitions (typically 10s of TB per experiment).
Objectives:
- Develop methodologies based on AI (including but not limited to Deep Learning) to detect submarine earthquakes using DAS data, and asses its potential for tsunami generation.
- Develop methodologies to identify other natural and anthropogenic processes that generate acoustic signals recorded on DAS, such as: internal waves and tides, marine mammals, ship traffic, submarine landslides, gas emissions from the seafloor, etc.
Expected Outcomes:
- Build and train Machine Learning models to detect submarine earthquakes and other oceanographic processes that generate acoustic signals.
- Implement these models in operational environments for existing submarine cables in the Iberian Peninsula and Canary Islands.
