How do dolphins and other toothed whales echolocate in a bubble cloud?

T G Leighton


One might say sound is 'vision' underwater, as far as the sonar of man is concerned. The same applies to the echolocation of toothed whales, such as the orca (opposite, with Prof. Leighton).

Part (I) described the refraction and scattering of the feeding calls of humpback whales within bubble nets produced by these creatures.

However many more species exploit bubble clouds to feed underwater. A superb example can be found in the BBC/Discovery series 'The Blue Planet'. Stills taken from episode one of that series (Figure 1, below) show common dolphin exploiting bubble nets to catch fish.

Not only that, but diving gannets create a wealth of bubble plumes which must complicate the acoustic environment, not only generating loud noises as each plume is formed, but also scattering the ocean sounds (including dolphin signals). This could well startle the fish and cause them to school, aiding feeding as described in Part (I). However whether this is intentional or uinintentional, is unknown.












 Figure 1. A series of images taken from The Blue Planet, BBC. (a) Common dolphins herd sardines with bubble nets. (b) A dolphin starts to release a cloud of bubbles (arrowed) from its blowhole. A moment later (c) the dolphin (1) swims on, leaving behind the expanding cloud (2). Other dolphins (incl. 3) enter the frame. (d) The sardines school within a wall of




bubbles that they are reluctant to cross, whilst (e) gannets dive into the sardine shoal to feed, folding their wings just before entry (arrowed). (f) On diving, a gannet (1) entrains a bubble plume (2). Plumes a few seconds old (3, with an older 4) have spread. (g) An aerial view shows hundreds of tight bubble plumes beneath airborne gannets. (h) A Bryde's Whale joins the feed. It surfaces with open mouth, which it then closes, sardines spilling from it. Images copyright of the BBC The Blue Planet (BBC). The accompanying book to the series is: Byatt, A., Fothergill, A.,  Holmes, M., and Attenborough, Sir David, The Blue Planet, BBC Consumer Publishing (2001).  



To dolphins, sonar in a sense provides their 'vision' underwater. Yet such dense bubble clouds as they produce would be impenetrable to  human sonar equipment. This raises the interesting dilemma. When they generate bubble clouds to hunt, are dolphins really nullifying their most spectacular sensory apparatus, in an environment so visually complex that successful echolocation would be very beneficial? Are we to accept that, in making such bubble nets, the dolphins are 'blinding' their most impressive sensory capability? Or have the dolphins adapted their sonar to work in this environment?


We do not really know if dolphin sonar is operating when they hunt using bubble nets, but if they did, what could they be doing to ensure their sonar does work? How might dolphins manage to echolocate in such environments as these bubble nets, or the bubbly surf zone? Dolphin sonar has remarkable capabilities (it is currently the only system which the US Navy has that is capable of finding small mines buried in the seabed), despite the fact that the dolphin hardware appears to be relatively mediocre compared to some man-made devices (Au, W. W. L. "The Dolphin Sonar: Excellent capabilities in spite of some mediocre properties," Eds. M.B. Porter, M. Siderius, and W. Kuperman, 2002, American Institute of Physics, Melville, New York (in press), 2004).


If we propose that the dolphins do not accept blinding of their sonar systems when they hunt with bubble nets, then, given that the hardware at their disposal is mediocre compared to human sonar which would not operate under such circumstances, a strong candidate explanation is that the dolphins are thinking in a way different to human interpreters of sonar. So how might a dolphin think to get their sonar to work in bubble clouds?

One speculative explanation is based on the fact that the interpretation of man-made sonar signals in bubbly water, and the theory which predicts that dolphin sonar should not work in intense bubble clouds, is based on a theory which assumes that the bubbles are undergoing linear steady-state pulsations in response to continuous-wave driving fields (the same is true of the attenuations calculated in Part (I)). However the sonar systems of dolphins and humans usually exploit acoustic pulses, and at the short ranges typical of the problem of finding fishes in bubble clouds, can generate amplitudes sufficiently great to produce nonlinear oscillations in bubbles. If the dolphin can mentally process the sonar echoes to exploit this nonlinearity, its ability to detect linearly-scattering targets within nonlinearly oscillating bubble clouds could be greatly enhanced.

There are many potential ways by which nonlinearities could be exploited to enhance detection in this way. One example is outlined below.




Figure 2. Schematic of a proposed ‘Twin Inverted Pulse Sonar’, whereby the scattering from a linear scatterer (such as a fish or a mine), and scattering from nonlinear scatterers (such as bubbles) can be enhanced and suppressed relative to one another (see text; ). The schematic bubble radius and time histories are similar to those found on page 6 of the project entitled "An acoustic diagnostic for lithotripsy"). The linear signals (the driving sonar field and the scatter from the 'target') are shown in yellow; the even powered nonlinearities (from the bubble scatter) are shown in blue; and the processing instructions are shown in pink.

Let us say the problem is to detect a linearly scattering target (the 'target') which is difficult to detect because it is immersed in a cloud of bubbly water. Such a target might be a fish in a dolphin bubble net  - even with a swim bladder, the fish would produce ostensibly linear scatter from dolphin echolocation because the gas is driven so far from resonance.  Alternatively, it might consist of a military mine which is a hazard to landing craft because it is hidden from sonar by breaking waves. 

Consider if the emitted sonar signal were to consist of two high amplitude pulses, one having reverse polarity with respect to the other (Figure 2, top line).  Linear reflection from the solid body is shown in Figure 2b(i). The bubble generates nonlinear radial excursions (Figure 2a(i)) and emits a corresponding pressure field (Figure 2a(ii)). Whilst the pressure emitted by the bubble may contain linear and odd-powered nonlinearities, it is the even powered (e.g. quadratic) nonlinearities which will be insensitive to the sign or the driving pulse, and hence which can be used to enhance the scatter from the target over that from the bubbles. It is these quadratic (and high even-powered components) which will be discussed in Figure 2, and below.

Normal sonar would not be able to detect the signal from the solid (Figure 2b(i)), as it is swamped by that from the bubbles (Figure 2a(ii)). If however the returned time histories are split in the middle and combined to make a time history half as long, enhancement and suppression occurs. If the two halves of the returned signals are added, the even-powered nonlinear components of the scattering from the bubble are enhanced (Figure 2a(iii)), whilst the scatter from linearly scattering target is suppressed (Figure 2b(ii)). This can be used to enhance the scatter from biomedical contrast agents. If however the two halves of each returned signal are subtracted from one another,  the even-powered nonlinear components of the scattering from the bubbles is suppressed (Figure 2a(iv)) whilst the reflections from the solid body are enhanced  (with the usual constraints imposed by increased signal-to-noise ratio) (Figure 2b(iii)).

If echolocation is the equivalent of vision underwater, then switching from linear to nonlinear sonar in bubble clouds might find analogy with driving through fog. 'Linear headlamps' would provide the familiar backscatter from the fog, making detection of targets difficult (analogous to the intense sonar backscatter from bubbles). However switching to nonlinear sonar might be equivalent to turning on 'nonlinear headlamps' in a car, which backscatter far less from the fog and so make driving easier.

The hypothesis is therefore, is this the way dolphins think?

For associated journal publications, click on the links below:

Leighton, T.G. From seas to surgeries, from babbling brooks to baby scans: The acoustics of gas bubbles in liquids, Invited Review Article for International Journal of Modern Physics B, 18(25), 2004, 3267-314

Leighton T G, ‘Nonlinear Bubble Dynamics And The Effects On Propagation Through Near-Surface Bubble Layers,’ High-Frequency Ocean Acoustics, Eds. M.B. Porter, M. Siderius, and W. Kuperman, 2005, American Institute of Physics, Melville, New York (in press) (2004)

Leighton T G, Meers S D and White P R, Propagation through nonlinear time-dependent bubble clouds, and the estimation of bubble populations from measured acoustic characteristics. Proceedings of the Royal Society A, 460(2049) 2521-2550, 2004



(Page last updated by T. G. Leighton, 26 August 2004)

© T G Leighton 2004