Listening to the seafloor with optical fibers

Measurements of seismic ground motion at the seafloor provide critical insights into earthquake and tsunami hazards, deep-Earth structure, plate tectonics, submarine volcanism, and interactions between the ocean and solid Earth. Yet ocean-bottom seismometers (OBSs) and observatories are exceedingly scarce: Although about 70% of Earth’s surface is covered by water, less than 1% of the global network of permanent broadband seismic stations is installed at the seafloor (see “Deploying seismometers where they’re needed most: Underwater ,” Physics Today online, 24 May 2019).

The emerging field of fiber-optic seismology offers a promising new paradigm for ocean-bottom instrumentation: distributed sensing rather than point sensing. Submarine fiber-optic cables for intercontinental telecommunications and power transmission traverse the global oceans and can be harnessed for distributed sensor networks. As seismic and ocean waves stretch and compress optical fibers at the seafloor, the light traveling through them encodes valuable information. With distributed acoustic sensing (DAS) and other new fiber-seismic methods, that information can be exploited for geophysical monitoring (see Physics Today, March 2018, page 24 ). This Quick Study explains how.

What is DAS?

The individual optical fibers in a telecommunications-grade cable contain small, inherent density fluctuations that locally perturb the index of refraction and cause Rayleigh scattering. As a pulse of light propagates down the cable, each scattering point reflects a minuscule amount of light back along the fiber. The accumulation of light from numerous scattering points constitutes a measurable backscatter trace. From that trace, one can map the arrival time of light to its scattering position in the fiber.

Whereas many DAS implementations exist, the general principle for all of them is similar. The method starts with an interrogator unit, the instrument that regularly transmits laser pulses into an optical fiber—typically at a rate of 1–10 kHz—and the corresponding backscatter trace for each pulse is recorded by a photodetector. By comparing consecutive traces, one can extract the change in phase of backscattered light from one pulse to the next as a function of distance along the fiber. The number and distribution of scatterers is predetermined by the fiber’s manufacturing process, so changes in the phase of backscattered light are caused only by changes in the optical path length between scatterers. Consequently, any internal deformation in the fiber can be calculated from the differential phase by a simple linear relationship.

The result is a two-dimensional data set of the strain or strain rate measured at thousands of locations along the fiber and thousands of times per second. Think of DAS as analogous to a dense linear array of conventional seismometers, except that each channel measures the distributed deformation over a finite distance instead of the particle motion at a single point. To date, DAS networks on land have recorded natural and human-generated earthquakes, detected icequakes on glaciers, imaged subsurface geologic structures, characterized soils for geotechnical analysis, and even monitored urban traffic patterns. Now scientists are turning to the oceans as the next frontier in fiber-optic seismology.

From point sensing to distributed sensing networks

The first pioneering effort to put seismometers on the seafloor was carried out by Maurice Ewing aboard the Atlantis in September 1935. After dropping a 1280-m-long string of seismic instruments through roughly 5 km of water to the bottom of the Atlantic Ocean, Ewing and colleagues detonated several explosives in an attempt to generate enough seismic waves to image the oceanic crust. Subsequent efforts to bring seismic instruments to the seafloor were delayed by the outbreak of World War II. It wasn’t until the 1960s that the need to monitor international nuclear-testing programs reignited interest in offshore seismic observatories. That interest led to the fundamental blueprint for OBSs that is still used today.

The majority of modern broadband OBSs are deployed over a period of one to two years. During that time, they operate autonomously: Using power from battery packs dropped to the seafloor, OBSs record their data to onboard storage. At the end of a deployment, an acoustic transponder prompts an OBS to return to the surface, where it is recovered and reconfigured for a new deployment. A benefit of the approach is that multiple research groups leverage the same pool of instruments to study diverse seafloor environments.

The temporary operational framework, however, limits the capabilities of OBSs relative to permanent terrestrial seismic networks. Only a handful of large (above magnitude 7.0) earthquakes occur worldwide over the span of a deployment, and typical OBS arrays contain fewer than 30 instruments. Seismic waves from smaller, more frequent earthquakes travel shorter distances through Earth, so the number of seismic ray paths traversing deep Earth that can be obtained by OBS arrays is small. Furthermore, because the data are trapped at the seafloor for the duration of the experiment, temporary OBSs cannot be incorporated into real-time systems that provide early warnings of earthquakes and tsunamis.

Over the past two decades, significant work has gone into developing alternative sensing networks that circumvent those challenges. One of the most successful approaches has been the establishment of cabled observatories in which seismometers installed at the seafloor are connected to land by a dedicated cable that provides unlimited power to the seismometer instruments and transmits real-time data back to operators. The cost of installing and maintaining such observatories, however, renders them impractical for global-scale applications.

Fiber-optic seismology offers a fundamentally different framework for monitoring the ocean bottom. A single DAS interrogator unit is connected to a seafloor fiber-optic cable at its landing station, and the laser pulse illuminates the fiber with thousands of effective sensors that record the vibrations of the seafloor in real time. Because the instrument itself is located on land, the deployment duration is unlimited, and data telemetry occurs at the speed of light. Both advantages allow DAS to be integrated into earthquake and tsunami early warning systems. Although the acquisition range of DAS systems is presently limited to signals no more than 100 km from the interrogator, other complementary fiber-seismic methods that are under development, such as polarization-based sensing, can span thousands of kilometers across entire ocean basins.

The distributed nature of ocean-bottom DAS measurements motivates a new approach to seismic data processing. Conventional OBS networks are sparse, with stations often spaced greater than 10 km apart. Those networks use multicomponent instrumentation—that is, three orthogonally oriented seismometers and a pressure gauge—to differentiate oceanic signals from solid-Earth signals and infer the properties of the seismic wave field. On the other hand, DAS networks are dense (1–10 m channel spacing), but they record only a single measure of seafloor motion at each sensing location. DAS networks are thus inherently multiscale and spatially coherent. They render data that are well suited for analysis by image-based algorithms, which remove noise from the data, separate wave-field components, and automate earthquake detection.

Bridging seismology and oceanography

Over the past four years, seismologists have designed ocean-bottom DAS arrays at diverse sites—from a wind farm in the Belgian North Sea to Monterey Canyon offshore of California. Those experiments have detected earthquakes ranging from a tiny magnitude 1.9 about 90 km away to a massive magnitude 8.2 whose source was more than 16 000 km away, identified previously unrecognized offshore fault zones through use of scattered waves, and mapped the velocity structure of the shallow crust by employing tomography.

Perhaps the most interesting findings pertain not to the seafloor but to the oceans. Pressure perturbations and thermal forcing from waves and tides can deform fiber-optic cables at the seafloor and be recorded by DAS. Since the 1930s, seismologists have known that ocean waves are the principal source of Earth’s quiet hum, called ambient seismic noise or microseism. But the generation mechanism remains a subject of considerable debate largely because of the absence of in situ observations. Three independent studies published in 2019 all observed surface gravity waves in the ocean with ocean-bottom DAS and demonstrated a clear link between the ocean waves and ambient seismic noise recorded on the same arrays.

Other oceanographic observations reported to date include ocean swell waves from distant storms; tidal bores and internal wave breaking; wave–current interaction, which allows DAS to measure ocean currents; and hydroacoustic signals from ship traffic, distant earthquakes, and marine mammal vocalizations. Identifying and separating such diverse signals, which are often indistinguishable on conventional OBS networks, is made possible by the dense and distributed nature of ocean-bottom DAS arrays.

Although fiber-optic seismology is still a nascent field, the success of early experiments has brought the geophysics community much closer to realizing the vision of permanent, telemetered seismographic and oceanographic monitoring across the oceans.