Dark fiber as a rat sensor network

What distributed acoustic sensing could change in urban rodent control

Urban rats are hard to study in the places that matter most. They move through tunnels, utility corridors, sewers, building voids, and other spaces that are difficult to enter, difficult to observe, and difficult to monitor over time. That is one reason cities still rely so heavily on partial signals: traps, sightings, complaint data, occasional camera footage, and short field surveys.

Those methods can be useful, but they rarely produce continuous, high-resolution data on how rats actually move through urban infrastructure.

A 2025 paper in Communications Earth & Environment points to a very different future. Instead of trying to place more devices in more hard-to-reach spaces, the researchers used something cities already have: unused fiber-optic cable. With distributed acoustic sensing, or DAS, one strand of dark fiber can act like a long line of vibration sensors, turning hidden infrastructure into an environmental monitoring system.

In this study, that system was used to track Norway rat activity inside an inaccessible urban tunnel.

This pillar article explains what the study did, what it found, why it matters, and how the EUREKA team can read it as an early signal of a broader shift: urban rodent management moving from episodic observation toward continuous environmental intelligence.

Why this paper matters

The value of this paper is not only technical. It addresses one of the core gaps in urban rat work: the lack of long-duration, fine-scale, non-invasive monitoring in places where rats are most active. The authors frame the problem clearly. Traditional ecological monitoring in hidden urban habitats is labor-intensive, spatially limited, and often unable to capture continuous behavior.

DAS offers a way to observe movement across long stretches of infrastructure with meter-scale spatial resolution and high sampling rates, using fiber that is already in the ground. That matters because better control depends on better measurement. If the field wants to move beyond scattered sightings and short monitoring windows, it needs systems that can detect when rats are active, where they are active, how they move, and how their behavior changes with weather and local conditions.

This paper shows that such systems are no longer theoretical.

What distributed acoustic sensing does

Distributed acoustic sensing converts a fiber-optic cable into a long chain of sensors by measuring tiny changes in backscattered light along the fiber. Those changes reflect local vibrations. In other fields, DAS is already used for seismology, geothermal monitoring, structural health assessment, and traffic studies.

The novelty here is the application to urban wildlife, and specifically to rats in a tunnel environment. The idea is simple enough to picture. When a rat moves on or near the cable path, it produces small mechanical vibrations.

DAS records those vibrations across multiple sensing channels. The result is not a photograph of the animal, but a spatiotemporal signature: a trace of movement through space and time. From that trace, the researchers can estimate direction, travel distance, timing, activity duration, and even gait-related features.

What the study actually did

The research team deployed DAS on a 20-kilometer section of unused dark fiber within a telecommunication cable in Yudu County, Jiangxi Province, China.

Their analysis focused on an 810-meter segment of fiber inside an inaccessible urban tunnel, monitored over 39 days between late June and early August 2021. The system sampled vibrations at 500 Hz with 10-meter gauge length and channel spacing, producing a large continuous dataset.

Because a sensing system is only useful if the source is known, the team also validated what they were seeing. They used excitation experiments to locate the relevant tunnel channels precisely, field inspection to rule out geological explanations, and video footage to confirm that wild rats were entering the tunnel through a drainage pipe.

They then ran laboratory simulation experiments to match the vibration signatures from the field with rat locomotion on the cable system. That validation step is important. It means the paper is not just saying “something moved.” It built a chain of evidence connecting the observed signals to Rattus norvegicus moving in the tunnel environment.

At the same time, the authors are careful not to overclaim: the method captures vibroacoustic signatures, not full species identity under all conditions, and complementary methods may still be needed in more complex settings.

What the system revealed about rat movement

One of the strongest parts of the paper is its movement analysis. The DAS system captured distinct spatiotemporal signals corresponding to individual rat travel along the tunnel cable path.

Frequency analysis showed dominant components in the 6 to 10 Hz range. With the help of laboratory calibration, the authors inferred stride frequencies of roughly 3 to 5 Hz, consistent with known walking and trotting patterns in rats.

This is a big step up from the level of information that most urban rat programs usually have. A complaint log can say that rats were seen. A trap can say that a rat was present.

A DAS trace can say much more: that an individual entered, moved a certain distance, slowed, paused, changed gait, turned around, or left. The paper even describes behavioral sequences such as a rat entering the tunnel, trotting a short distance, pausing, and then reversing direction.

In other words, the system begins to turn hidden movement into readable behavior. That does not replace direct ecological observation, but it makes a much richer layer of observation possible in places where direct observation is hard or impossible.

What it revealed about social behavior

The study also captured interaction between two rats in enough detail to reconstruct a chase sequence. The system recorded pursuit, closing distance, reversal of direction, and later reversal of roles between the two animals.

During these pursuit phases, the pursued rat showed higher stride frequencies than the pursuer, which the authors interpret as faster movement during evasion. Solo movement, by contrast, was more often associated with lower stride frequencies and slower gaits.

That result matters because it shows this is not just a presence-detection tool. It is capable of registering dynamic behavior between animals. For urban ecology, that opens the door to questions that are usually very difficult to study in subterranean systems: spacing, pursuit, clustering, repeated use of the same corridor, and possibly changes in social activity under different environmental pressures.

What it revealed about activity timing

The 39-day dataset also showed strong circadian patterning. Rats in the tunnel were clearly nocturnal, with a bimodal activity pattern: one activity peak in the hours after sunset and another in the hours before dawn, building toward sunrise.

The system recorded activity bouts and long travel distances, with some individual bouts extending to about 1480 meters.

This kind of continuous temporal record is one of the paper’s greatest strengths. Many urban rat studies rely on snapshots. This one produces a timeline. That allows the authors to move beyond “rats are nocturnal” and begin asking when activity intensifies, how long bouts last, and what environmental variables shift those patterns.

What it revealed about weather

The most operationally interesting result may be the weather analysis. Using multivariable regression and lagged environmental variables, the authors found significant negative associations between rat activity and cloud fraction, wind speed, and temperature.

Their preferred model used a three-hour lag, suggesting that environmental conditions affected rat activity with some delay rather than only in the exact same hour. The interpretation is cautious but useful. Clearer nights may have changed visibility conditions. Stronger winds may have affected foraging behavior.

Higher temperatures likely mattered because ambient temperatures often exceeded the rat thermoneutral zone identified in the paper. Even though the tunnel offered some buffering, rats still needed to forage outside and move through a broader environmental context.

For management, this is where the paper becomes more than a sensing demonstration. It suggests that rodent activity is not only spatially patterned. It is dynamically responsive to the urban environment, including weather. That means better monitoring could eventually support better timing of interventions, better prediction of activity surges, and better interpretation of why some nights look “quiet” and others do not.

Why this is different from traditional monitoring

Trap-based monitoring, camera placement, bait take, and complaint systems are all useful, but each has a narrow window. Traps and bait points can trigger neophobia. Cameras are limited by line of sight and location. Complaint systems depend on people seeing and reporting activity. DAS changes the starting point.

It monitors continuously without direct handling, without entering the habitat repeatedly, and without needing to place a visible device in the animal’s path. The paper argues that this gives DAS several advantages: long-duration monitoring, high spatial and temporal resolution, low disturbance to animals, and the possibility of automated alerts when activity changes.

That last point is especially relevant to urban management. A sensing layer is most useful when it can move from passive recording to practical detection of anomalies and changes.

What the study does not solve yet

The paper is strong partly because it does not pretend to solve everything. The authors name several limitations directly. DAS cannot identify individual animals. It cannot always distinguish similar species without additional methods. Signal detection depends on proximity to the cable and on the physical coupling between cable and environment.

Urban background noise can also make signal separation difficult. Those limitations matter for interpretation. This is not yet a universal rat sensor for every city block and every buried cable.

It is a proof that existing fiber infrastructure can be repurposed for ecological sensing under real field conditions, and that the resulting data can be biologically meaningful. That is already a substantial result.

The paper also points toward the next technical layer: multimodal systems. The authors discuss combining DAS with other sensing methods, such as distributed temperature sensing, machine-learning-based signal classification, environmental DNA, or thermal imaging. In other words, dark fiber is not the whole answer. It is a strong backbone for a larger sensing stack.

What this means for EUREKA

For the EUREKA team, the paper is important less as a specific recipe than as a direction of travel. It shows that urban rat monitoring is starting to move into the same world as other forms of infrastructure sensing: continuous, indirect, large-scale, and computational.

That aligns with EUREKA’s broader logic of turning scattered rodent evidence into structured intelligence. The immediate lesson is not that every pilot site needs dark fiber.

The immediate lesson is that future-ready rodent programs should be designed to accept multiple sensing layers: complaints, logs, field observations, cameras, device outcomes, and eventually infrastructure-derived signals. A project that treats each of those as isolated data streams will learn slowly. A project that treats them as one shared observation system will learn much faster.

There is also a more strategic lesson. This paper moves the conversation away from “How do we find more rats with more labor?” toward “How do we turn existing infrastructure into a sensor network?” That is exactly the kind of systems thinking EUREKA can help organize.

The team does not need to overclaim current capability to benefit from this paper. It only needs to place it correctly: as evidence that the sensor layer of urban rodent ecology is arriving.

Where this could lead

If dark fiber and related sensing methods become practical at city scale, the implications are large. Hidden infrastructure could become observable. Repeated nighttime movement could be measured instead of guessed. Activity changes could be tied to weather, season, construction, sanitation disruptions, or intervention timing.

Management could become more adaptive because baseline activity would be known more clearly. That does not mean the field becomes fully automated. It means human teams get a better instrument panel.

PMPs still interpret sites. Public health teams still set priorities. City agencies still make decisions. But they would do so with better temporal and spatial intelligence than complaint counts alone can provide.

Closing

This paper is one of the clearest examples in the current literature of urban rat ecology becoming a sensing problem as much as a trapping problem. By repurposing dark fiber as a line of vibration sensors, the authors showed that an inaccessible tunnel could be monitored continuously, that rat movement and social interaction could be reconstructed from those signals, and that activity patterns could be linked to changing weather conditions.

For EUREKA, that makes the paper more than a technology story. It is a signal that the next generation of urban rodent work will depend on connected observation systems, not just isolated field notes.

Better tools will matter. Better exclusion will matter. Better housing will matter. But better measurement may be the layer that finally allows all the others to be compared, timed, and improved.