Assessing Submarine Earthquake Detectability with Fiber-Optic Cables in the Canary Islands

Nowadays, fiber-optic telecommunication cables are serving a multitude of purposes beyond their conventional role. One such application is earthquake detection, which is particularly advantageous in submarine environments, where data acquisition is inherently more challenging and costly. Leveraging Distributed Acoustic Sensing (DAS) technology, these cables undergo a remarkable metamorphosis, transforming into a network of seismic sensors spaced just meters apart, and spanning remote and inaccessible environments with ease.

Nevertheless, the quality of DAS data relies on varios factors that surpass the interrogator’s performance, such as the seafloor topography and cable characteristics and coupling. Equally critical is the development of robust software capable of discerning earthquake waves amidst a cacophony of noise and a broad spectrum of signals in the submarine environment.

With this aim in mind, this study utilizes DAS data from a telecommunications fiber-optic cable linking the islands of Tenerife and Gran Canaria within the Canary Islands region. Our dataset spans approximately 2 months in 2020 and encompasses readings from both cable ends, each equipped with an interrogator, providing coverage of approximately 60 kilometers on each side. Situated between these islands lies a submarine volcano, which exhibits seismic activity nearly every day. Additionally, the recent installation of new land stations has enabled to observe an increase in seismicity to the east of Gran Canaria. Hence, this dark-fiber could enhance our ability to monitor the volcano and to accurately locate the source of this newfound seismic activity.

Leveraging the pre-trained PhaseNet-DAS model, we detect P and S waves of seismic events and compare our findings with those published by the National Geographic Institute (IGN) earthquake bulletin. Through comparative analysis of both cable ends, we ascertain their earthquake detection capabilities, delineating sensitivity levels and identifying cable segments with optimal event detection and minimal noise interference.