In 2003 scientists conducted a broad-band (0–80 kHz) study of various gun arrays at the Heggernes Acoustic Range near Bergen, Norway, an institution operated by the Norwegian, Danish, Dutch and German navies for noise measurements of military and civil vessels. The spatial dimensions and volumes of the source configurations were small, on average 10–20 dB lower than arrays used by the exploration industry. Additionally, the noise generated by the seismic vessel itself was recorded.
Ship-generated noise
Broad-band smoothed amplitude spectra from 38.2 1-gun array at 550m range, received by a hydrophone at 35m depth, compared to noise generated by the seismic vessel. Below 1 kHz the amplitude spectrum of the air-gun signal differs significantly from the vessel’s noise spectrum due to the low-frequency energy emitted by the air-gun, while above 1 kHz the amplitude spectrum of the air-gun signal coincides with the vessel’s noise almost completely. This indicates that the slow spectral level decay is mainly caused by ship-generated noise. Modified from Breitzke et al, 2008Most of the emitted source energy had frequencies below 150 Hz, far in excess of the low-frequency noise generated by the source vessel. Spectral source levels (measured in decibels) were highest below 100 Hz, dropping off continuously with range and frequency so that at 1 kHz they were approximately 40 dB, and at 80 kHz approximately 60 dB lower than the peak level.
Above 1 kHz, spectral levels agreed almost completely with the noise generated by the vessel, meaning that if low-level, high-frequency (>1 kHz) spectral components were emitted, they were masked by ship-generated noise. This is of particular importance for marine mammals with pronounced high-frequency hearing sensitivity like toothed or beaked whales. The results indicate that any high frequency signals created by these relatively small air-gun arrays are so weak that they would not disturb these marine mammals significantly more than normal ship traffic.
Air-gun high frequency signals
Photograph of an air-gun fired under water. Notice the four bubbles emerging from the ports of the gun. Source: Langhammer (1994)What causes high frequency signals from an air-gun? Here, we give you the current understanding.
The first and most obvious cause is that the rapid movement of the air escaping through the air-gun ports creates cavities in the water close to the gun. This effect is the same as cavities created on propellers, or those associated with turbulent flow in water, as shown on the photographs in the box on page 64. Since we believe that these cavities are created close to the source, it is reasonable to assume that the amount and strength of the cavities are dependent on the design of the air-gun. Hence, there might be differences between high frequency noise from air-guns produced by different manufacturers. It is also evident that the triggering mechanism, the so-called solenoid, which is an electrical coil sitting on each individual air-gun, creates high frequency noise.
Furthermore, the air-gun shuttle, which is the piston that pushes the air out of the gun, creates high frequency noise during the rapid movement and sudden stop of the shuttle. Finally, all air-guns jump as a result of the bubble movement, and this jumping will create mechanical shaking of the air-gun which again creates mechanical high frequency noise.
All these mechanisms occur close to or in the vicinity of each air-gun.
Cavitation due to ghosts?
If these two air-guns generate a peak pressure of 4 bars at 1m, the reflected signal from the sea surface measured at the hydrophone between the guns is approximately –1.3 bars, assuming linear theory. The hydrostatic pressure at the hydrophone position is 1.2 bars, which means that the total pressure – assuming that the pressure contribution from each source can be added linearly – is negative. Then, cavitation will occur. Source: Landrø (2000).Another, far more sophisticated and at the moment speculative cause for high frequency emissions from air-gun arrays is coupled to the effect of reflections from the sea surface. If two air-guns emit a strong signal of, for example, 4 bars, the pressure in the water between the two guns might approach zero since the reflected signals from the sea surface are negative.
In the example shown, we have assumed that the signals from the two air-guns can be linearly superimposed. However, as the water pressure approaches zero, non-linear effects will be more prominent, meaning that it is not straightforward to estimate exactly for which pressures cavitation will occur. Despite this, it is reasonable to assume that for compact and large air-gun arrays, we have a risk of cavitation formation in the area where the ghost reflections from several air-guns coincide in time and space.
This type of cavitation will be independent of the type of air-gun used, since it is simply a function of the geometry of the array. And this is the good news: if the major part of the high frequency signal generated by an air-gun array is generated by cavitation between air-guns, this effect can be eliminated simply by increasing the distance between the guns.
Plesset and Ellis showed in 1955 that it is indeed possible to generate cavities by acoustic stimulation, as shown on the figure in the box on page 62. It is therefore reasonable to assume that similar cavitation phenomena might also occur for an air-gun array.
Generally, the strength and length of this high frequency cavitation signal will therefore increase with the size and compactness of the air-gun array.
Field data observations
Smoothed frequency spectra of the source signature measured from the small array (black line) and the big array (red line).
Source far-field signatures from a single string array (black line) and a multi-string array (red line). Notice the high frequency noise appearing at approximately 0.08 seconds for the big array.Is this effect observed on air-gun signature data? Normally, the answer is no, since the sampling frequency is typically 250 Hz.
However, the figure below shows far-field signatures sampled at 60 kHz vertically below both a small single string air-gun array and a big multi-string array towed at 5m depth. The peak-to-peak amplitude for the small and big arrays are 37 and 62 bar-m, respectively (corresponding to 251 and 256 dB re 1 μPa). It can be seen that the amount of high frequency noise is negligible for the small array, while it is more prominent for the full array configuration. The most prominent part of the high frequency noise is centered around 0.08 seconds. Its amplitude level is approximately 1-2 bar-m (220-226 dB re 1 μPa).
A frequency spectrum comparison shows that there is a significant difference between the signatures of the two arrays. A deviation up to 15 to 20 dB is observed in the frequency band between 10 and 20 kHz. For frequencies above 30 kHz this deviation decreases to approximately 10 dB.
More measurements are required to quantify and assess which of the above-mentioned mechanisms for high frequency noise are most significant, and whether there are other effects that have not been discussed in this article.
We believe, however, that a thorough understanding of the causes for high frequency signals from air-gun arrays are important, both in order to assess their environmental impact and potentially to diminish these effects.
To assess this potential mechanism further we plan to conduct a number of experiments where we gradually increase the distance between the subarrays in the airgun array, and where we gradually displace the subarrays relative to each other in the sailing direction. If our predictions are correct, we should measure that the amount of cavitation and the associated high frequency noise gradually decrease.
Further discussion on this topic has been submitted to Geophysics for publication.
