An explosion occured in an office in the center of Utrecht (The Netherlands) during the night of 2001, May 31. Large explosions are known to generate infrasound, or inaudible sound (i.e. frequencies lower than 30 Hz). Infrasound from the explosion was recorded by the De Bilt infrasound array (DBN). DBN consists of six microphones in an array with a diameter (or aperture) of 70 meters, see figure 1.

Figure 1: The configuration of the De Bilt infrasound array (DBN). DBN consists of six microphones in an array with an aperture of 70 meters. The vertical axis aligns with a North-South direction.

Figure 2 shows the recordings over the infrasonic wave travelling over DBN. The energy is identified by the increase in amplitude around 28 seconds. The differential traveltimes of the wave over the instruments already resolves a western direction as probable source location.

Figure 2: Recording of an infrasonic wave travelling over DBN. The six microphones show a coherent recording around 28 seconds from the time axis zero time which is 23h19m14.1s GMT on 2001, May 31. The signal comes from the West since microphone 01 is the first to record the sound wave.

The signal is analysed in the frequency domain. Coherency values are plotted in figure 3 as a function of time and frequency. The coherency shows a significant increase around 28 seconds. The dominant frequency of the signal is 4.2 Hz since this is the frequency where maximum coherency is found.

Figure 3: The coherent wave clearly stacks up to high coherency values around 28 seconds (time axis is the same as in figure 2). The cohereny finds is maximum spectral energy around a frequency of 4.2 Hz.

A frequency-wavenumber analysis is conducted to locate the source. Figure 4 shows a maximum in power spectral amplitude towards the West of DBN. The exact direction or back azimuth is 252 degrees.

Figure 4: Localizing the coherent energy around 28 seconds at 4.2 Hz resolves the exact direction of the energy (or back azimuth) being 252 degrees with respect to the North.

The resolved back azimuth, DBN and source location are plotted in figure 5. The small discrepancy between calculated and real location is caused by scattering of the energy. Scattering of energy can be caused by objects located along the path of the infrasonic energy. The azimuthal discrepancy can also be caused by cross winds along the energy's trajectory through the atmosphere. Calculations show that this effect is negligble at this small range (i.e. 0.003 degrees).

Figure 5: The explosion's location in the street "Achter St. Pieter" given by the red star. The red dot is the infrasound array DBN, located somewhat South of the KNMI. The resolved back azimuth of 252 degrees is given by the red vector.

Infrasonic wave propagation depends on the wind and temperature structure of the atmosphere. The effective sound speed, as contoured in the lower frame of figure 6, incorporates these effects. The rays along which the infrasonic waves have travelled are plotted in white and determined by this velocity structure. The top frame of figure 6 shows the traveltimes for rays reaching the surface. It took the energy around 11 seconds to reach DBN, with an estimated distance between source and receiver of 3.7 km.

Figure 6: The possible paths of the infrasonic waves through the atmosphere. The velocity with which the rays travel is a function of wind ad temperature of the atmosphere. The effective sound speed velocity, given in the lower frame, is the projection of this velocity on the direction between source and receiver. The top frame gives the traveltime for the rays reaching the surface. It took the energy around 11 seconds to reach DBN, with an estimated distance between source and receiver of 3.7 km.