When individuals contemplate fiber optic cables, it’s typically in reference to their application in telecommunications and internet connectivity. However, fiber optic cables — strands of glass or plastic that facilitate the transmission of light — can serve an alternative function: imaging the Earth beneath us.
Hilary Chang, a PhD student in the MIT Department of Earth, Atmospheric and Planetary Sciences (EAPS), recently utilized the fiber optic cable network at MIT to effectively image the ground under the campus using a technique known as distributed acoustic sensing (DAS). By leveraging existing infrastructure, DAS proves to be a resourceful and efficient means of comprehending ground composition, which is vital for evaluating seismic risks in certain areas and their vulnerability to earthquake damage.
“We managed to extract remarkably coherent waves from our surroundings, which allowed us to gather information about the subsurface,” remarks Chang, the principal author of a recent publication detailing her research, co-authored with EAPS Principal Research Scientist Nori Nakata.
Inactive fibers
The fiber optic system on the MIT campus, installed between 2000 and 2003, supports internal data transfer between laboratories and buildings, as well as external communications like the campus internet (MITNet). There are three primary cable hubs on campus from which lines extend into various buildings and underground, resembling a spiderweb.
The network designates a specific number of strands for each building, some of which are labeled “inactive fibers,” or cables that are not currently transmitting information. Each fiber hub on campus is equipped with redundant backbone cables to ensure that, in case of a failure, network transmission can seamlessly switch to the inactive fibers without interruption.
DAS can utilize existing telecommunications cables and surrounding wavefields to gather information about the materials they traverse, making it an invaluable tool for environments like urban settings or the ocean floor, where traditional sensors may not be applicable. Chang, whose research focuses on earthquake waveforms and the insights they can provide, chose to test this method on the MIT campus.
To gain access to the fiber optic network for her experiment, Chang contacted John Morgante, a manager in infrastructure project engineering at MIT Information Systems and Technology (IS&T). Having been with MIT since 1998 and participating in the initial fiber optic network installation project, Morgante was able to offer personal guidance on selecting a suitable route.
“It was fascinating to hear what they aimed to achieve with their testing,” states Morgante. While IS&T has previously collaborated with students on various projects involving the network, he noted that “in the physical plant domain, this is the first instance I can recall where we actually partnered on an experiment together.”
They opted for a pathway starting from a hub in Building 24, as it was the longest uninterrupted route located entirely underground; overhead cables running through buildings were unsuitable as they lacked grounding and were ineffective for the experiment. This route extended from east to west, commencing at Building 24, traversing beneath a portion of Massachusetts Ave., along segments of Amherst and Vassar streets, and concluding at Building W92.
[Morgante] was incredibly supportive,” says Chang, describing her collaboration as “a highly positive experience working with the campus IT team.”
Identifying the cables
After securing a rental on an interrogator, a device that emits laser pulses to detect ambient vibrations along the fiber optic cables, Chang and a group of volunteers gained special access to connect it to the hub in Building 24. They allowed it to operate for five days.
To verify the route and ensure the interrogator was functioning correctly, Chang performed a tap test, striking the ground with a hammer multiple times to capture the precise GPS coordinates of the cable. Conveniently, the underground pathway is marked by maintenance hole covers, which provided suitable locations for conducting the test. Furthermore, as she required a quiet environment to collect clear data, she had to conduct the test around 2 a.m.
“While I was striking near a dorm, someone yelled ‘shut up,’ probably because my hammering disturbed their sleep,” Chang recollects. “I felt bad.” Fortunately, she only needed to tap at a few locations and could interpolate positions for the remainder.
During daylight hours, Chang and her fellow students — Denzel Segbefia, Congcong Yuan, and Jared Bryan — conducted an additional test with geophones, another instrument designed to detect seismic waves, at Brigg’s Field where the cable passed beneath to compare the signals. For Chang, it was a delightful experience; when the data were gathered in 2022, the campus was gradually returning to normalcy from pandemic measures, with remote classes still occasionally in effect. “It was wonderful to have everyone on the field engaging in hands-on activities,” she shares.
Environmental noise
Once Chang obtained the data, she observed significant environmental activity in the waveforms, including the movement of cars, bicycles, and even the nightly train that passes along the northern boundary of the campus.
By identifying the sources of noise, Chang and Nakata were able to extract coherent surface waves from the ambient sounds and utilize the wave speeds linked to various frequencies to interpret the characteristics of the ground through which the cables passed. Firmer materials allow for quicker velocities, while softer materials impede them.
“We discovered that the MIT campus is situated on soft materials overlaying a comparatively hard bedrock,” states Chang, confirming previously known, albeit lower-resolution, insights about the region’s geology obtained through seismometer data.
Such information is crucial for areas prone to damaging earthquakes and other seismic threats, including the Commonwealth of Massachusetts, which has experienced earthquakes as recently as last week. Neighborhoods in Boston and Cambridge characterized by artificial fill during rapid urban development are particularly vulnerable due to the subsurface structure’s tendency to amplify seismic frequencies and cause structural damage. This non-intrusive technique for site characterization can aid in ensuring buildings meet the required standards for the appropriate seismic hazard level.
“Destructive seismic occurrences do take place, and we must be ready,” she asserts.