Infrasound and seismology in the Low Frequency Array LOFAR

Infrasound was first discovered after the destructive eruption of the Krakatoa in Indonesia (1883). Low frequency acoustic waves, caused by this eruption, appeared to have travelled around the globe up to four times (1).

These infrasonic waves were detected world-wide on traditional barographs as small air pressure fluctuations. Due to its low-frequency content, infrasound hardly experiences attenuation in the atmosphere and travels to thermospheric altitudes of over 100 km. In addition to seismology, infrasound was used to measure the occurrence of atmospheric nuclear tests and to estimate their yield (2). It was also in this period that the possibility to use infrasound as passive atmospheric probe started to be recognized( 3). This interest diminished after nuclear tests were confined to the subsurface under the Limited Test Ban Treaty (1963). 

Recently, the study of infrasound is experiencing a renaissance as it was chosen as a verification technique for the Comprehensive Nuclear-Test-Ban Treaty (CTBT), that opened for signing in 1996 (4). Infrasound science currently concentrates on source identification and passive remote sensing of the upper atmosphere.

The construction of the largest radio-telescope in the world in the northern part of the Netherlands and neighbouring countries, the Low Frequency Array (LOFAR), opened the possibility to co-locate geophysical sensors and realize an efficient multi-sensor network. The KNMI, Delft University of Technology and TNO, all partners in LOFAR, make use of the advanced LOFAR infrastructure to build-up an infrasound and seismological research network. This network consists of a temporary 80 element high density array, a permanent 30 element microbarometer array with an aperture of 100 km and, at the same locations, a 20 to 30 element seismological component. Here, we present the scientific background, goals and first results.

The physical characteristics of infrasound

Sound waves below 20 Hz are inaudible for the human ear. These sound waves are called infrasound. The lower limit of infrasound is controlled by the thickness of the atmosphere or of an atmospheric layer. For the troposphere, the acoustic cut-off period is roughly 5 minutes. For longer period waves gravity acts as restoring force, instead of the molecular relaxation for sound waves, and hence these are called gravity waves. These gravity waves propagate with typical wind speed velocities of 5 to 10 m/s. Infrasonic waves travel with the sound speed which is 340 m/s for air of 20°C. The amplitudes of infrasonic waves are small with respect to the ambient pressure and vary between milli-pascals (Pa) to tens of Pa.

The propagation of infrasound is controlled by the effective sound speed, which is a function of the temperature and wind along the source-receiver trajectory (5). Infrasonic waves will be refracted if vertical gradients in the effective sound speed exist. Waves will be bended back towards the earth's surface in case these gradients are strong enough. There are three regions in the atmosphere where strong wind and/or temperature gradients (may) exist that lead to turning infrasonic waves. (1) In the troposphere, in case of a temperature inversion near the surface or a strong jet stream around the tropopause at 10 km altitude. (2) In the stratosphere, due to the combined effect of a temperature increase with increasing altitude due to presence of ozone and strong seasonal winds around the stratopause at 50 km altitude. (3) In the thermosphere, where from 100 km and upwards the temperature strongly increases with increasing altitude due to direct influence of solar radiation on the molecules. As an example, the temperature and wind for a winter and summer atmosphere in De Bilt (the Netherlands) are shown in Figure 1.


Figure 1. Typical summer (red, 2007-07-01) and winter (blue, 2007-12-01) temperature and wind models for De Bilt, the Netherlands. The wind is split in a zonal (Zw), west to east, and meridional (Mw) component, south to north.

Figure 2 shows how infrasound refracts in the stratosphere and thermosphere in a retracing approach through the summer profiles of Figure 1. Multiple returns are predicted from both the stratopause (labelled as Is) and thermosphere (It). The polar vortex in the stratosphere is directed westwards during the northern Hemisphere summer and eastwards during winter, making the refractivity of the medium seasonal dependent. In other words, the atmosphere is an anisotropic medium. This anisotropy is also reflected in Figure 3. It phases are to be observed in all directions from the source, while Is phases only occur to the west. This signature on the earth surface changes as function of time of the day and geographical location. Therefore, Figure 3 is illustrative for the challenge in source identification from a verification point of view. But it also shows the enormous amount of information on the state of the upper atmosphere concealed in the surface based microbarometer recordings.

The general scientific challenges

A large amount of infrasound is continuously being recorded from a variety of man-made and natural sources. Anthropogenic sources include: explosions, nuclear tests, mining, military activities and supersonic flights. The latter is the cause of frequent reports in the Netherlands of felt tremors in buildings, similar to the sensation of an earthquake. Natural sources comprise: avalanches, oceanic waves, severe weather, sprites (lightning from cloud top to ionosphere), earthquakes, meteors, lightning, volcanoes and aurora. The measurement of infrasound is affected by noise due to wind and turbulence in the boundary layer. Therefore, infrasound is measured with arrays to increase the signal-to-noise ratio (SNR) through signal summation. Arrays are also used to estimate the direction of arrival of a wave and its propagation velocity. Typical sizes, i.e., apertures, are in the order of 100 to 1000 meters. Additional noise reduction at each array element is achieved by a wind barrier, a porous hose or pipe array with discrete inlets6). The recorded signals are thus a function of the state of the boundary layer and the upper atmosphere, which changes with time and geographical location. To unravel this complex picture, is the major challenge in source identification (7).

Figure 2. Ray tracing through the summer profiles of Figure 1, towards the west and east. Refractions back to the earth's surface occur from the stratosphere (Is) and thermosphere (It). The polar vortex is directed to the west during the northern Hemisphe

High frequency seismological arrays are similar in layout to those used in infrasound. The co-location of infrasound and seismic equipment is beneficial for both techniques. The processing techniques developed in infrasound and seismology can be interchanged. Such techniques are applied in data processing and signal detection. Furthermore, atmospheric waves can interact with the solid earth. The interaction between both media is a field of recent research interest.

The surface based microbarometers can also be used as a passive probe for the upper atmosphere (higher than 30 km) with the large amount of sources continuously present. Actual recordings of the basic properties, like wind and temperature, of the upper atmosphere are sparse. Meteorological balloons reach an altitude of roughly 35 km but lack spatial and temporal coverage. Rocket sondes can reach the upper atmosphere but also lack coverage. Satellites have global coverage but have a limited vertical resolution and are difficult to validate for stratospheric altitudes. Therefore, most information currently depends on numerical weather prediction model characteristics. Infrasound can validate such models, and even information on a finer temporal and spatial scale is expected to be retrieved. Such information is very welcome for future atmospheric research. The troposphere and stratosphere have long been considered two isolated layers, split by an impermeable tropopause. The influence of the stratosphere on our daily weather and climate has recently been firmly established, showing that processes in the upper atmosphere do couple to the troposphere8).

Figure 3. Ray tracing through the summer profiles in all directions up to a distance of 600 km. The bounce points of the rays, in blue, on the earth, in gray, are plotted for a source located in the centre at an altitude of 0 km. Blue colours refer to str

The microbarometers not only sense infrasound but also gravity waves. The amplitudes of tens of pascal associated with these internal waves fall within the dynamic range of the sensors. Gravity waves are not represented in climate models while they play an important role in many atmospheric processes and, therefore, this leads to uncertainties in climate predictions.

Technical details

The geophysical application within LOFAR, i.e., GEO-LOFAR, consists of seismological and infrasound equipment. The astronomical application will be realized on antenna fields where infrastructure, like power and high-speed Internet, is being established. GEO-LOFAR will make use of this infrastructure. The acquired data will be gathered at Groningen University from where they are distributed to the GEO-LOFAR partners, which are TU Delft, TNO and KNMI. The infrasound contribution consists of a High Density Infrasound Array (HDIA) and a Large Aperture Infrasound Array (LAIA).

A six element infrasound array (EXL) was realized at LOFAR's Initial Test Station near Exloo. EXL has an aperture of 250 meter and contributed to the discovery of exceptional fast infrasonic phases observed after the explosion of an oil-depot in the UK (10,11) (see Figure 4). The array also showed its value in the detection of lightning, by combining infrasound and electro-magnetic measurements within LOFAR (12,13).

Figure 4. Regular (Is3) and exceptionally fast (Is3f) infrasonic phases, i.e., forerunners, observed over the Netherlands after the explosion of an oil-depot in the UK. The forerunners are caused by a very strong polar vortex wind, at 40 km altitude, reac

HDIA has been a temporary deployment of 80 acoustic pressure and vector sensors (of type Microflown) in an 80x80 meters array. The aim was to characterize the acoustic and noise field on a small spatial scale. HDIA's size was comparable to the size of an analog wind filter used at each array element in a conventional array (see Figure 5). Furthermore, the use of a vector sensor enables direction finding with a small instrument instead of a large array. A first analysis of the collected data looks promising (14).

LAIA will consist of 30 microbarometers, mostly co-located with seismic sensors, in an array with an aperture of 100 km. Recordings from LAIA will provide excellent research opportunities through its unique layout. Only one comparable array has been operational, which was used to analyze the spatial coherence of short period acoustic-gravity waves, with period from 30 to 50 seconds. The Large Aperture Microbarograph Array (LAMA) functioned in the 1960s in the US to detect atmospheric waves from nuclear tests (15).

In order to also sense gravity waves, the KNMI-microbarometer (KNMI-mb) has been adapted to periods of 1000 seconds, as lower cut-off. The earlier version of the KNMI-mb had a cut-off at 500 seconds and was specially designed to measure acoustic waves. The influence of temperature on the differential KNMI-mb is of main concern when lowering the frequency response. Temperature stability is ensured by properly insulating the instrument's fault and by its subsurface mounting. A detailed study of the amplitude and frequency response has been carried out (16).

Figure 5. The array layout from the High Density Infrasound Array (HDIA). Acoustic pressure and vector sensors (Microflown) are used. The axes are scaled to show the inner rings of the array more clearly. The diameter of the outer ring is approximately 80

An experimental seismological array has been installed at the EXL location which recorded induced earthquakes that occurred in producing gas fields in the region. A test of real time event detection and processing, using e.g. techniques developed for infrasound processing, was carried out successfully (17). The build-up of a larger array near Annerveen will allow real-time detection and possible identification of induced seismic events.


The high-quality infrastructure provided by LOFAR enables the deployment of large infrasound and seismic installations, in terms of spatial and temporal coverage. Infrasound measurements on such a scale are unique in the world. This uniqueness will provide excellent research opportunities on subjects like source identification and passive acoustic remote sensing of the upper atmosphere (18,19).

The combined measurement of elastic waves in the solid earth and acoustic waves in the atmosphere enables a so-called seismo-acoustic analysis. Specific sources, like explosions and oceanic waves, generate both seismic and infrasonic waves. The infrasonic signature of such a source highly depends on the state of the atmosphere. On the other hand, the seismic waveform is more or less similar throughout the seasons. A seismo-acoustic analysis may reveal the influence of the atmosphere on the detection and localization capability of infrasound as monitoring technique and will also allow for acoustic remote sensing.

Furthermore, infrasound not only travels through the atmosphere but also couples to the solid earth as air-coupled Rayleigh wave or ground-coupled air wave. In this case, the propagation velocity of Rayleigh waves approaches the sound speed. As seismic and infrasound observations are done at the same site, the atmospheric contribution to seismic noise can be quantified.

Future studies with the LOFAR data will be carried out and show the enormous potential of large scale sensor networks.


Symons, G.J., 1888. The eruption of Krakatoa and subsequent phenomena. Trübner & Co., London, UK, 487pp.
Posey, J.W. and A.D. Pierce, 1971. Estimation of nuclear explosion energies from microbarograph records. Nature, 232, 253.
Donn, W.L. and D. Rind, 1971. Natural infrasound as an atmospheric probe. Geoph. J. R. Astr. Soc., 26, 111-133.
Dahlman, O., S. Mykkeltveit and H. Haak, 2009. Nuclear Test Ban. Springer, Dordrecht, the Netherlands, 250pp.
Drob, D.P., J.M. Picone and M.A. Garcés, 2003. The global morphology of infrasound propagation. J. Geoph. Res., 108, D21, 4680, doi:10.1029/2002JD003307
Hedlin, M.A.H., B. Alcoverro and G. D'Spain, 2003. Evaluation of rosette infrasonic noise-reducing spatial filters. J. Acoust. Soc. Am., 114, 1807-1820.
Evers, L.G. and H.W. Haak, 2001. Listening to sounds from an exploding meteor and oceanic waves. Geoph. Res. Lett., 28, 41-44.
Ineson, S. and A.A. Scaife, 2009. The role of the stratosphere in the European climate response to El Niño. Nature Geoscience, 2, 32-36.
Gossard, E.E. and W.H. Hooke, 1975. Waves in the atmosphere. Elsevier, Amsterdam, the Netherlands, 456pp.
Evers, L.G. and H.W. Haak, 2007. Infrasonic forerunners: Exceptionally fast acoustic phases. Geoph. Res. Lett., 34, L10806, doi:10.1029/2007GL029353
Ottemöller, L. and L.G. Evers, 2008. Seismo-acoustic analysis of the Buncefield oil depot explosion in the UK. Geoph. J. Int., 172, 1123-1134.
Holleman, I, H. Beekhuis, S. Noteboom, L. Evers, H. Falcke and L. Bähren, 2006. Validation of an operational lightning detection system. 19th International Lightning Detection Conference, Tucson, USA, 12pp.
Assink, J.D, L.G. Evers, I. Holleman and H. Paulssen, 2008. Characterization of infrasound from lightning. Geoph. Res. Lett., 35, L15802, doi:10.1029/2008GL034193
Van Zon, T., L.G. Evers, R. van Vossen and M. Ainslie, 2009. Direction of arrival estimates with vector sensors: First results of an atmospheric infrasound array in the Netherlands. Proc. 3d Int. Conf. on Underwater acoustic measurements: Technologies and results, Nafplion, Greece, 6pp.
Mack, H. and E. A. Flinn., 1971. Analysis of the spatial coherence of short-period acoustic gravity waves in the atmosphere. Geoph. J. R. Astr. Soc., 26, 255-269.
Mentink, J. and L.G. Evers. The response of the KNMI microbarometer. J. Acoust. Soc. Am., in preperation.
Hazel, G.-J. van den, 2008. Detection and processing 3C signals from a small scale seismic array, MSc thesis, Utrecht University, 68pp.
Godin, O.A., 2006. Recovering the acoustic Green's function from ambient noise cross correlation in an inhomogeneous moving medium. Phys. Rev. Lett., 97, doi:10.1103/PhysRevLett.97.054301
Wapenaar, K., 2006. Nonreciprocal Green’s function retrieval by cross correlation. J. Acoust. Soc. Am., 120, EL7-13.17.

Niet gevonden wat u zocht? Zoek meer achtergrond artikelen