Blog

Jun 25, 2025

Detection of aircraft noise using distributed acoustic sensing with a buried telecommunication cable | npj Acoustics

npj Acoustics volume 1, Article number: 2 (2025) Cite this article 2155 Accesses 1 Citations 1 Altmetric Metrics details Buried distributed acoustic sensor (DAS) arrays are known to detect and monitor

npj Acoustics volume 1, Article number: 2 (2025) Cite this article

2155 Accesses

1 Citations

1 Altmetric

Metrics details

Buried distributed acoustic sensor (DAS) arrays are known to detect and monitor various sources of seismic waves, both natural and anthropogenic. But can buried DAS detect sources in the atmosphere such as aircraft? There is some evidence of aircraft signal in DAS data, but more research is needed to understand the detection capabilities and estimate flight parameters. A four-week continuous DAS recording with a gauge length of 10 m using a telecommunication cable buried beneath the Curtin University campus in Perth, Western Australia shows distinct signatures of aircraft in the frequency range of 70 to 120 Hz at distances of up to 2.5 km. However, the system detected only propeller but not jet aircraft, which are known to emit sound at higher frequencies and thus require a smaller gauge length. The findings reveal a potential for independent monitoring of aircraft traffic using existing telecommunication infrastructure.

Detection of flying aircraft using acoustic sensors presents a less expensive alternative to traditional radar systems1. One option for such detection is distributed acoustic sensing (DAS), a rapidly developing technology for detecting and recording local vibrations using Rayleigh scattering of laser pulses in optical fibres2,3,4,5. DAS measures temporal variations of linear strain at multiple points along the fibre. The DAS measurements are performed by connecting a specialised optical device called interrogator to one end of an optical fibre. The interrogator sends laser pulses into the fibre and measures strain or strain rate by detecting temporal changes in the interference patterns between optical pulses backscattered by adjacent elements of the fibre. Unlike traditional sensors such as geophones, hydrophones, and seismometers, DAS provides so-called distributed measurements, whereby each segment of the optical fibre becomes a sensor. The effective length of the sensing segment (called gauge length) can vary between tens of centimeters to tens of meters, and is defined by a trade-off between the signal-to-noise ratio and spatial resolution of the measurement6. The sensing segments may overlap, so that spacing between channels is often smaller than the gauge length, providing useful redundancy. Some interrogators operate with a fixed gauge length, while more advanced ones allow the gauge length to be set by the user.

DAS can detect and monitor various sources of seismic and acoustic energy, both natural (proximate, regional and remote earthquakes, ocean storms, marine mammals) and anthropogenic (mine blasts, vehicle and pedestrian traffic)7,8,9,10,11. Many of these sources generate seismic waves, and thus can be detected with any seismic sensors, including DAS. An intriguing question is whether buried DAS cables (which are not necessarily tightly coupled to the ground) can detect aircraft noise generated in air. Previously DAS was employed to image reflectors deep below the sea bottom using air guns deployed in sea water12. However, detecting sound generated in the air using buried sensors poses a different challenge. Buried point sensors, such as microphones, are known to record airborne sound; these measurements are sometimes used to infer properties of porous soil13,14. However, the use of DAS to detect aircraft noise may be particularly important for independent monitoring of air traffic by leveraging vast existing fiber-optic telecommunication networks.

Previous studies show that aircraft noise has a distinct signature detectable by geophones and microphones on land15,16,17,18 and hydrophones placed under water19. Yet, most of these technologies require installation of sensors dedicated to monitoring. Detecting such noise using a buried fibre-optic cables may present an attractive alternative as it can take advantage of an existing telecommunication infrastructure without a need for dedicated sensors. Previous investigations show that some aircraft can be detected while others cannot20. We explore detection of aircraft noise by a buried DAS cable and estimation of the detection range, aircraft position and speed.

To explore the possibility of detecting aircraft noise, we recorded DAS data continuously for four weeks in November 2023 using a telecommunication fiber-optic cable buried beneath the Curtin University Campus in Perth, Western Australia, about 10 km from Perth International Airport21. Previously, the possibility of detecting ground motion caused by aircraft noise in the area was confirmed using geophones22.

A typical spectrogram (after denoising) of DAS data acquired on the Curtin University campus is displayed in Fig. 1 and shows the distinct Doppler pattern discussed in Methods.

The position of the aircraft at 3:07 UTC is right above the DAS cable (as shown in Fig. 2). The dashed line over the spectrogram shows the variation of the apparent frequency with time for an airplane at the distance of 600 m from the receiver traveling along a straight-line flight path with ground speed of 365 km/h.

However, this pattern was only observed for a small proportion of the passing aircraft. A detailed case by case analysis in conjunction with Flightradar24 data shows that these signals correspond to turboprop (propeller) airplanes only, and not to any of the jet aircraft, which are far more numerous in the area. All turboprop commercial airplanes passing within 2.5 km from the cable were detected. The noise from jet aircraft has a higher frequency range, usually around 600 Hz19,23. This corresponds to a wavelength of about 0.5 m. Such sound cannot be detected with a gauge length of 10 m, as a DAS sensor would act as a low-pass spatial filter averaging out signals with spatial periods much smaller than the gauge length.

In that case, the central frequency of sound emitted by the aircraft, f0, distance d between the sensor and the aircraft and aircraft speed v can be estimated by fitting the curve defined by eqs. (1) to (4) to the variation of the apparent frequency with time as shown in Fig. 1. For the airplane noise shown in Fig. 1, this fitting gives f0 = 83 Hz, d = 600 m and v = 365 km/h. The speed given by Flightradar24 at that point (Fig. 2) is 197 kts or 364.8 km/h, virtually the same as estimated by fitting. The distance d = 600 m is 10% larger than the altitude of 541 m. However, the airplane is unlikely to be exactly above the specific DAS channel shown in Fig. 1. For a lateral distance of, say, 200 m between the nearest location of the plane and the DAS channel, the total source-receiver distance would be \(\sqrt{54{1}^{2}+20{0}^{2}}=576\) m, or 4% less than the distance estimated from the spectrogram. The time of the nearest position of the airplane is within 0.5 seconds of the timing given by Flightradar24. More precise analysis requires mapping exact coordinates of DAS channels and exact position of the aircraft. This will be a subject of future work.

Flight path of the airplane corresponding to the Doppler effect signature shown in Fig. 1 ©Fightradar24.

As mentioned in Introduction, DAS has many channels along the fibre-optic cable. This opens a potential for estimating the flight path by muti-channel triangulation. Figure 3 shows spectrograms of 10-minute recordings of two DAS channels (178 and 330), which contain the sound signatures from the same airplane. The slopes of the apparent frequency versus time are clearly different, indicating different distances to the aircraft. The fitting using eqs. (1) to (4) gives f0 = 90 Hz and v = 290 km/h for both channels and d1 = 450 m and d2 = 600 m for channels 178 and 330, respectively. These distances provide an opportunity to estimate the position of the aircraft if precise coordinates of each channel are known. Note also that the noise patterns for the two channels are very different, possibly due to differences in ambient noise level or in the coupling of the cable to the ground.

The dashed lines over the spectrogram show the best fits of the variation in apparent frequency with time for an airplane traveling along a straight-line flight path.

Equation (4) assumes that the airplane is traveling along a straight line with a constant speed. This approximation is reasonable for cruising part of the flight path but not in the relative proximity to an airport. Figure 4 shows an example of a spectrogram where the variation of the apparent frequency with time is very different from that given by eqs. (1) to (4). Note, however, that eqs. (1) – (3) are valid for a flight path of any shape. For fitting purposes, it is useful to divide the flight path (a certain vicinity, say 5 km) into circular arcs. The variation of the apparent frequency with time can be divided into three segments. The first and last segments look similar to the ones corresponding to straight line paths, while the middle segment is clearly different: the apparent frequency increases with increasing time. Note that when the instantaneous aircraft speed is perpendicular to the aircraft–receiver line, there is no Doppler shift and hence f = f0. It follows that, if the flight path is a circular arc whose center coincides with the projection of the receiver onto the plane of the path, the apparent frequency is constant (and equal f0). This condition is fulfilled, for instance, when the arc is horizontal and its radius R equals the projection D of the source–receiver line onto the horizontal plane. It follows that the apparent frequency decreases with time when the flight path is an arc with R > D and increases with time when R < D.

The dashed line over the spectrogram shows the best fit of the variation in apparent frequency with time for an airplane traveling along a three-segment flight path shown in Fig. 5.

A rapid increase of f with time in the middle segment of the aircraft response in Fig. 4 indicates that here the radius of curvature is considerably smaller than D. Thus, we model the flight path of this airplane by a circular arc connecting two straight-line segments. By varying the curvature of the arc, we find a combination of the three segments that fits the spectrogram in Fig. 4 reasonably well (dashed line). The corresponding flight path consisting of two straight-line segments connected by a circular arc with a curvature radius of 1800 m shown in three different colours is displayed in Fig. 5. This trajectory is similar to the flight path given by Flightradar24 data in Fig. 6.

Suggested model of a flight path of the airplane represented by the spectrogram in Fig. 4.

A flight path of the aircraft with spectrogram of Fig. 4. ©Fightradar24.

In the present study, detection of all aircraft was done by visual inspection of spectrograms and thus involved subjective judgement. Analysis of these detections provides an opportunity to obtain a rough estimate of the detection range. From the analysis of the spectrogram corresponding to the curved path (Fig. 4) in conjunction with the map of the flight path in Fig. 6 shows that the sound is still clearly detectable from the aircraft after it has completed the turn towards south-east. This straight-line segment of the flight path is nearly 5 km from the (nearest edge of) the DAS cable. This is probably close to the upper bound of the detection range estimates. However, this segment of the spectrogram, although quite bright, is relatively short and does not show any levelling out towards the asymptote as the airplane flies further away; this suggests that detection using this segment alone would be ambiguous. On the other hand, the straight-line segment before the turn does show a very slight levelling out of the spectrogram (with decreasing time). This segment of the flight path is 2.5 km from the cable; this distance can be regarded as a reasonable estimate of the detection range using a single channel. Multichannel (array) data processing should increase the range. The above analysis corresponds to turboprop passenger aircraft such as Saab340B. The detection range should depend on the model of the aircraft.

Figure 7 shows a spectrogram of a DAS record from a fiber-optic cable laid on the bottom of an outdoor swimming pool. The spectrogram shows rapid decreases and increases in apparent frequency, probably caused by sharp maneuvers of small training aircraft. These aircraft (but not necessarily all of their maneuvers) have also been identified in Flightradar24 data (not shown). This confirms that aircraft can also be detected in DAS data acquired on fibre-optic cables laid underwater.

A spectrogram of a segment of DAS data recorded with a fiber-optic cable on the bottom of an outdoor pool.

The spectrograms of sound emitted by an aircraft traveling at a constant speed along a straight line as recorded by geophones show a narrow-band signal whose frequency decreases with time as the aircraft is passing in the vicinity of the receiver15,16. This distinct frequency shift is associated with the Doppler effect: the sound frequency of a noise source is higher than the intrinsic source frequency when that source is approaching the sound sensor and vice versa. Thus, the intrinsic frequency emitted by the aircraft is shifted in proportion to its apparent speed with respect to the position of the receiver.

For an airplane travelling with a speed v and emitting sound with a central frequency f0, the variation of the apparent frequency f at a stationary receiver with time \({t}^{{\prime} }\) is given by15,16

where c is sound velocity in air, t is the time when the signal was emitted by the aircraft, \({t}^{{\prime} }\) the time when it reached the receiver, and α the angle between the aircraft path and the straight line connecting the airplane to the receiver. The angle α is related to the shortest difference d between the receiver and the aircraft by

In eq. (2), s = − sign(t)a where a is the projection of the straight line between the aircraft and the receiver onto the tangent to the flight path at time t (t = 0 corresponds to the time when the aircraft is the closest to the receiver if it were travelling along the tangent while maintaining the constant speed v it had at time t), see Fig. 8. The times t and \({t}^{{\prime} }\) at the source and receiver are related by

The source–receiver geometry in the plane containing the tangent to the flight path and the receiver on the ground

For a straight flight path with a constant speed v,

The characteristic Doppler effect signature in the spectrogram can be used to estimate the intrinsic frequency emitted by the aircraft, distance between the aircraft position and the sensor, and aircraft’s ground speed.

The 1.5 km-long fibre-optic cable is buried beneath the campus grounds at a depth of less than 1 m with a maximum spread of ~500 m in both north–south and east–west directions (Fig. 9). The DAS data were acquired with a time-sampling rate of 1 ms, channel spacing along the cable of 1 m and a gauge length of 10 m. Since the altitude of airplanes in the area is at least a few hundred meters, the 10-m long DAS elements can be regarded as point sensors.

The scale is shown as white text in the lower-right corner. ©GoogleEarth.

Within the period of DAS recording, aircraft flight data from Flightradar24.com (a global ADS-B crowd service24) were analysed to identify aircraft that passed within 3 km from the cable. For each such aircraft, a spectrogram was computed for segments of 5 to 10 minutes in duration around the time the aircraft was the closest to the cable. Note that the university campus environment has many sources of acoustic and seismic noise, and the coupling of the cable to the ground might be variable. Since the cable was deployed for telecommunication purposes rather than for fibre-optic sensing, details of cable deployments such as precise burial depth or the coupling to the ground are not available to the researchers.

We analyse the DAS data by computing spectrograms for 1000 s windows with a 50% overlap. Computed spectrograms were subjected to low-frequency noise removal using a median filter in two steps: first for vertical stripes (broadband noise bursts) and then for horizontal stripes (persistent narrow-band noise). For each step, this was done by estimating a low-frequency trend and then subtracting it from the spectrogram.

To detect aircraft using cables laid underwater, the experiments on the Curtin University campus were complemented with DAS measurements using a 25 m fibre-optic cable laid at the bottom of a small outdoor swimming pool. The experiment site is located in close proximity (within 3 km) of Jandakot Airport, Australia’s largest general aviation airport by aircraft movements located about 10 km south of the Perth city centre. The DAS measurements were performed with a gauge length of 10 m, channel spacing of 4 m, and time sampling interval of 1 ms.

Our results show that propeller airplanes are detectable in spectrograms of DAS data by their characteristic Doppler signature, which can be recorded on fibre-optic cables buried underground or laid under water. Parameters and trajectory of the aircraft can be estimated using multiple DAS channels. The detection range using a single channel is about 2.5 km.

Detecting noise from jet aircraft requires a smaller gauge length, which is achievable with modern DAS instrumentation and will be a subject of future work. Other important directions of future research are automatic detection of aircraft in DAS data as well as multichannel (array) data analysis and beamforming aimed at increasing the detection range and detailed characterisation of the flight path and parameters.

The advantages of monitoring of air traffic with DAS are relatively low cost, the possibility of leveraging existing telecommunication infrastructure and a potential for beamforming using a dense array of sensors.

The datasets analyzed in the current study are available from the corresponding author on reasonable request.

The computer code for this study is not publicly available for proprietary reasons but might be provided on a case by case basis.

Damarla, T. & Kirkpatrick Alberts, W. C. Helicopter tracking using acoustic arrays. In 17th International Conference on Information Fusion (FUSION) (Salamanca, Spasin, 2014).

Daley, T. M. et al. Field testing of fiber-optic distributed acoustic sensing (das) for subsurface seismic monitoring. Lead. Edge 32, 699–706 (2013).

MATH Google Scholar

Parker, T., Shatalin, S. & Farhadiroushan, M. Distributed acoustic sensing - a new tool for seismic applications. First Break 32 https://www.earthdoc.org/content/journals/10.3997/1365-2397.2013034 (2014).

Hartog, A. An Introduction to Distributed Optical Fibre Sensors (CRC Press, 2017).

He, Z. & Liu, Q. Optical fiber distributed acoustic sensors: A review. J. Lightwave Technol. 39, 3671–3686 (2021).

ADS MATH Google Scholar

Dean, T., Cuny, T. & Hartog, A. H. The effect of gauge length on axially incident p-waves measured using fibre optic distributed vibration sensing. Geophys. Prospecting 65, 184–193 (2017).

ADS MATH Google Scholar

Glubokovskikh, S. et al. Downhole distributed acoustic sensing provides insights into the structure of short-period ocean-generated seismic wavefield. J. Geophys. Res.: Solid Earth 126, e2020JB021463 (2021).

ADS Google Scholar

Glubokovskikh, S., Shashkin, P., Shapiro, S., Gurevich, B. & Pevzner, R. Multiwell fiber optic sensing reveals effects of CO2 flow on triggered seismicity. Seismological Res. Lett. 94, 2215–2230 (2023).

Google Scholar

Lindsey, N. J. et al. Fiber-optic network observations of earthquake wavefields. Geophys. Res. Lett. 44, 11,792–11,799 (2017).

MATH Google Scholar

Zhu, T., Shen, J. & Martin, E. R. Sensing earth and environment dynamics by telecommunication fiber-optic sensors: an urban experiment in Pennsylvania, usa. Solid Earth 12, 219–235 (2021).

ADS Google Scholar

Debens, H. A. et al. Whale detection and microseismic monitoring via das using submarine telecommunications cables - a case study from the nws, western australia. Aust. Energy Producers J. 64, S481–S486 (2024).

MATH Google Scholar

Wilk-Lopes, M. et al. PP and PS imaging from fiber optic acquisition (S-DAS): a North Sea case study https://www.earthdoc.org/content/papers/10.3997/2214-4609.202436006 (2024).

Sabatier, J. M. et al. In situ measurements of soil physical properties by acoustical techniques. Soil Sci. Soc. Am. J. 54, 658–672 (1990).

ADS MATH Google Scholar

Moore, H., Attenborough, K., Rogers, J. & Lee, S. In-situ acoustical investigations of deep snow. Appl. Acoust. 33, 281–301 (1991).

MATH Google Scholar

Meng, H. & Ben-Zion, Y. Characteristics of airplanes and helicopters recorded by a dense seismic array near anza california. J. Geophys. Res.: Solid Earth 123, 4783–4797 (2018).

ADS MATH Google Scholar

Fang, G., Li, Y. E. & Barak, O. The seismic aircraft footprint: Probing near surface and tracking aircraft https://doi.org/10.1190/segam2020-3426206.1 (2020).

Damarla, T. R. & Ufford, D. Helicopter detection using harmonics and seismic-acoustic coupling. In Carapezza, E. M. (ed.) Unattended Ground, Sea, and Air Sensor Technologies and Applications X, 6963, 69630W https://doi.org/10.1117/12.776899 (SPIE, 2008).

Wilson, T. C., Petrin, C. E. & Elbing, B. R. Infrasound and low-audible acoustic detections from a long-term microphone array deployment in oklahoma. Remote Sensing 15 https://www.mdpi.com/2072-4292/15/5/1455 (2023).

Erbe, C. et al. Underwater noise from airplanes: An overlooked source of ocean noise. Mar. Pollut. Bull. 137, 656–661 (2018).

MATH Google Scholar

Fang, G., Li, Y., Zhu, T. & Barak, O. Characteristics of airplane-induced ground motion recorded by three-component (3-C) geophone and distributed acoustic sensing (DAS) in cities https://ui.adsabs.harvard.edu/abs/2020AGUFMNS015..11F (2020).

Tertyshnikov, K. et al. Where on Curtin University Campus is the dark fibre? First Break 2021, 1–5 (2021).

Google Scholar

Dean, T. & Hasani, M. A. Seismic noise in an urban environment. Lead. Edge 39, 639–645 (2020).

MATH Google Scholar

Konopka, W., Pawlaczyk-Luszczyńska, M. & Śliwińska Kowalska, M. The influence of jet engine noise on hearing of technical staff. Med Pr. 65, 583–592 (2014).

Google Scholar

Lindahl, F. How Flightradar24 conquered the skies - and why the pope’s travel plans cause a surge in traffic. The Telegraph https://www.telegraph.co.uk/travel/travel-truths/flightradar24-what-how-works-fredrik-lindahl/ (2017). Accessed: 2024-12-04.

Download references

The authors acknowledge the financial support from the Australian Department of Industry, Science and Resources for the 2021 Global Innovation Linkage (GILIII000114) grant and the sponsors of the Curtin Reservoir Geophysics Consortium. The authors thank Silixa and ASN for providing DAS equipment. The authors are grateful to Gang Fang of Shandong University and Yunyue Elita Li of Purdue University for insightful discussions.

Centre for Exploration Geophysics, School of Earth and Planetary Sciences, Curtin University, GPO Box U1987, Perth, 6845, WA, Australia

Boris Gurevich, Roman Isaenkov, Konstantin Tertyshnikov, Mikhail Vorobev & Roman Pevzner

Centre for Marine Science and Technology, School of Earth and Planetary Sciences, Curtin University, GPO Box U1987, Perth, 6845, WA, Australia

Christine Erbe, Alexander N. Gavrilov & Evgenii Sidenko

Search author on:PubMed Google Scholar

Search author on:PubMed Google Scholar

Search author on:PubMed Google Scholar

Search author on:PubMed Google Scholar

Search author on:PubMed Google Scholar

Search author on:PubMed Google Scholar

Search author on:PubMed Google Scholar

Search author on:PubMed Google Scholar

B.G. conceived of the idea of this study, performed the initial analysis of DAS- and flight-radar data, and wrote the first draft of the paper; R.P. conceptualised how the study could be carried out; R.P., K.T., A.N.G., E.S., and C.E. conceptualised and carried out DAS data acquisition; R.P., R.I., C.E., and M.V. developed algorithms for data analysis; M.V. and E.S. performed detailed data analysis and generated all of the graphics; A.N.G. edited the manuscript. B.G. and M.V. did most of the revisions. B.G. wrote the response to the reviewers.

Correspondence to Boris Gurevich.

The authors declare no competing interests.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

Gurevich, B., Isaenkov, R., Erbe, C. et al. Detection of aircraft noise using distributed acoustic sensing with a buried telecommunication cable. npj Acoust. 1, 2 (2025). https://doi.org/10.1038/s44384-025-00007-8

Download citation

Received: 25 September 2024

Accepted: 22 February 2025

Published: 27 March 2025

DOI: https://doi.org/10.1038/s44384-025-00007-8

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative