How European research is turning submarine fibre-optic cables into seismic sentinels
Roughly 1.5 million kilometres of fibre-optic cable cross the world’s ocean floors. Built to carry data, this vast infrastructure is now being explored as something else entirely: a distributed instrument for observing the planet. Across Europe, a growing community of network operators, universities and research institutes is investigating how existing submarine cables, without compromising their primary role as data conduits, can double as sensors capable of detecting earthquakes, tsunamis and other seismic phenomena in real time.
Sparkle is one of the operators contributing infrastructure and historical data to this effort. But the story is really about a research collaboration that stretches across academia, industry and national geophysical agencies, and about what it can offer the wider research and education networking community.
Two ways to listen to the sea
Two complementary approaches are being studied. Distributed Fibre-Optic Sensing (DFOS) repurposes cables already in service, detecting minute changes in light transmission caused by environmental disturbances. Science Monitoring and Reliable Telecommunications (SMART) takes a different route, integrating dedicated sensors directly into new cable installations or subsea repeaters. Together, the two methods open two ways of building sensing capacity: retrofitting the network that already exists and designing sensing into the network still to come.
The Mediterranean as a research testbed
Since 2022, Sparkle has worked with Italy’s National Institute of Geophysics and Volcanology (INGV) and Google to test the seismic sensing potential of the Mednautilus cable system, which runs for more than 7,000 kilometres from Sicily across the eastern Mediterranean at depths of up to 4,000 metres. The same exploration has since extended to BlueMed, the newest cable connecting Italy with France, Greece and other Mediterranean countries.
The work now sits within ECSTATIC, a €4 million EU-funded project bringing together 14 academic and industrial partners, including the University of L’Aquila, the University of Padua and Nokia. Sparkle contributes years of historical cable data to what is effectively a second generation of this research. Current trials focus on the Tyrrhenian segment of BlueMed, between Genoa and Palermo, and on Mednautilus, with data processed and stored at Sparkle’s Network Operations Centre in Catania. The project’s aim is to validate new methodologies, based on interferometry and signal polarisation, combined with AI and machine learning, for seismic early warning, predictive maintenance and network integrity monitoring.
Independent academic work is now reinforcing these results. A study published in March 2026 analysed state-of-polarisation data collected from the Mednautilus cable between 2022 and 2024, applying machine learning models, including gradient boosting and deep autoencoders, to distinguish seismic signals from background noise. The study found that moderate-to-large earthquakes, of magnitude 5 or above, produced measurable polarisation anomalies, with the best-performing model reaching around 60% accuracy in classifying them. The result is preliminary, and it must be considered that the cable was not built for this purpose. Still, independent peer review offers the validation this line of research is looking for.
What the signals have already shown
The clearest results so far come from a different analytical approach: monitoring the state of polarisation of optical signals along the Catania–Haifa–Tel Aviv link, using an algorithm built around the Jones matrix. This method clearly detected the magnitude 5.8 earthquake that struck the Dodecanese Islands on 2 June 2025, and it also picked up a precursor signal ahead of the magnitude 8.8 megathrust earthquake that hit Russia’s Kamchatka Peninsula on 29 July 2025. Researchers involved in the work describe the technique as notably robust: it produces consistent results across different fibres within the same cable, regardless of random polarisation coupling or the direction signals travel.
Looking ahead
Fibre-sensing research still has a considerable distance to travel before it becomes a reliable, standardised early-warning tool. But the direction of travel is telling. Cable manufacturers are beginning to offer integrated sensing capabilities as standard, and climate scientists are increasingly interested in the technology’s potential to fill long-standing gaps in deep-sea temperature monitoring, where conventional sensors remain expensive and sparse.
What stands out most, though, is the shape of the collaboration itself: a network of universities, research institutes and telecommunications providers sharing infrastructure, data and expertise toward a common scientific goal. As fibre-optic sensing matures, the boundary between the networks that carry our data and the instruments that help us understand the planet may continue to blur, and the next steps will depend as much on open collaboration between research and industry as on the technology itself.
