5 – Adaptations for Acoustic Communication in Birds: Sound Transmission and Signal Detection

1.

Sound transmission was measured in open fields, mixed decidous forest with and without leaves and coniferous forest in Dutchess County, New York. Attenuation of white noise and pure tones was measured between one microphone close to a loudspeaker and another microphone 100 m away, at the same height. Graphs of excess attenuation (E.A. in dB/100 m) against frequency were obtained at 0.15, 1, 2, 5, and 10 m above the ground. An analysis of variance was conducted to estimate effects of height, frequency and habitat.




2.

Height and frequency affect sound transmission more than habitat. With a sound source close to the ground (15 cm and 1 m) all frequencies were more attenuated than at greater heights. The patterns of E.A. as a function of sound frequency were basically similar in all habitats. At all source heights the lower the frequency the better the sound carried, with the exception that close to the ground, sounds below 2 kHz were excessively attenuated. Comparing open field and forest, trees improved transmission of frequencies below 3 kHz, especially close to the ground.




3.

Some general trends can be predicted for maximization of transmission distances of animal sounds in these habitats. For an animal vocalizing higher than 1 m above the ground, the lower the frequency the further the sound travels. Close to the ground, low frequencies are again preferred for maximization of transmission distances, but the frequencies must be pitched above a range of attenuated, low-pitched sounds, the limits of which vary to some extent with habitat, creating the ‘sound window’ of Morton. This ‘window’ of least-attenuated frequencies, only occurring close to the ground, tends to be pitched somewhat lower in forest than in open habitats. However, for an animal producing sounds in the habitats tested, perch height and sound frequency are more important than the habitat in determining how far the sound will carry.

Sound transmission and its significance for animal vocalization

Summary1.Attenuation of white noise and pure tones from 350 Hz to 10 kHz was measured at three secondary forest sites in Panama at different stages of maturity. Graphs of excess attenuation (E.A.) versus frequency were obtained near ground level, and at heights of 1, 2, 5, 10, and 12 m in each habitat.2.The pattern of E.A. vs. frequency was similar for all habitats. For all heights other than ground level and 1 m the lower the frequency the better the sound carried. Sounds below 2 kHz were attenuated by a ground effect if the source was 1 m or less from the ground. Consequently there was a minimum of E.A. in all three habitats between 500 Hz and 2 kHz at ground level and 1 m.3.Relevance of the data to assessment of the role of environmental variables in the natural selection of vocalization for long-distance transmission is discussed. Two facts mitigating against Morton's sound “window” as an explanation for lower frequencies in songs of forest as opposed to open country birds are presence of similar “windows” in both habitats and restriction of windows to a zone close to the ground in most habitats.

Sound transmission and its significance for animal vocalization: II. Tropical forest habitats

SYNOPSIS. Acoustic signals transmitted over large distances differ significantly from those emitted by the signaler. Acoustic signals degrade in amplitude, spectral and temporal structure as they propagate through theenvironment. A great deal of work on acoustic communication is aimed at understanding the selective forces imposed by the environment on animal signals. I will discuss the physical constraints the environment puts on acoustic communication, and then discuss similarities in communication by anurans and insects that relate these environmental constraints to their signaling systems. Lastly, I show how changes in signals during propagation relate to changes in signal perception during phonotaxis, and thus, how propagation relates to mate choice and sexual selection

From Sender to Receiver: Propagation and Environmental Effects on Acoustic Signals

On the basis of experience accumulated over the past few years, a revision has been made in the currently used Taylor-Pelmear scale for the staging of Raynaud's phenomenon in persons exposed to vibration from hand-held tools, while retaining as much as possible of the well-established advantages of the scale for research and its proved usefulness for clinical and medicolegal purposes. The 0T and 0N stages of symptoms have been omitted, together with the parallel disability scale. A separate staging for neurological disorders connected with the syndrome was proposed and accepted at the workshop "Symptomatology and Diagnostic Methods in the Hand-Arm Vibration Syndrome," held in Stockholm in 1986. The criteria descriptions have been changed so as to minimize their reliance on seasonal factors. The new staging system--a stage 0 and four stages (1-4) with attacks of cold-induced Raynaud's phenomenon--clearly defines the differences in the descriptions of the stage criteria in order to improve their clinical usefulness. A numerical scoring based on the extent and distribution of finger blanching was not, however, introduced, whereas a score based on the number of affected fingers on each hand was proposed, considered, and accepted.

The Stockholm Workshop scale for the classification of cold-induced Raynaud's phenomenon in the hand-arm vibration syndrome (revision of the Taylor-Pelmear scale)

Following earlier work by Chessell [J. Acoust. Soc. Am. 62, 825–834 (1977)] it is shown that his single‐parameter theory can be used to predict the measured transmission spectra between a source and receiver located above ground surfaces having a wide range of acoustic impedance—or effective flow resistivity. Surfaces behaving essentially as locally reacting range from new‐fallen snow, effective flow resistivity σ=10–30 cgs rayls, through grass‐covered ground, σ=150–300 rayls, to mature asphalt, σ=30 000 rayls. The thermal‐conduction and viscous boundary layer of the surface limits the effective flow resistivity of even the hardest and most impervious surface to the range 105–106 rayls, depending on frequency: this value is appropriate to evaluate the complex reflection coefficient of the paint‐sealed surface of a thick slab of reinforced concrete.

Effective flow resistivity of ground surfaces determined by acoustical measurements

There is an extensive body of theory, and some laboratory measurements, of sound propagation over a surface of finite impedance. There are also reliable measurements of outdoor sound propagation in near‐horizontal directions over the ground. In an attempt to relate these more closely, we have made carefully controlled measurements at ranges from 1 to 1000 ft, in most cases over grass‐covered flat surfaces, to demonstrate the several phenomena that are involved. These phenomena depend on, and conversely provide a means of estimating, the values of ground impedance for waves at near‐grazing angles of incidence. Such values obtained for grass‐covered surfaces are in reasonable agreement with each other and with values obtained by conventional means at other angles of incidence. It is suggested that simple but accurate predictions of noise levels can be made by assuming that an excess attenuation due to finite ground impedance would always exist in a certain shadow region near the ground. This shadow region i...

Outdoor sound propagation over ground of finite impedance

Line‐of‐sight measurements of the log‐amplitude and phase fluctuations of pure tones between 250 and 4000 Hz propagated over distances between 2 and 300 m in the turbulent atmosphere close to the ground are compared quantitatively with simple theory using simultaneously measured meteorological variables. The theory is based on the assumption of homogeneous and isotropic turbulence and approximates the availability of eddy sizes in the source region of turbulence by a Gaussian spectrum. In particular the transverse or mutual coherence function (the coherence in a plane perpendicular to the direction of propagation) and the coherence in the direction of propagation which we call the longitudinal coherence, are also calculated and compared with the measurements. When the measured mean square phase fluctuations are compared with the theory using the meteorological measurements, good agreement is obtained. However the measured mean square log‐amplitude fluctuations are in general substantially smaller than predicted and, in addition, show clear evidence of saturation. The distance to saturation is shown to correspond to the longitudinal coherence length. Because of this behavior of the amplitude fluctuations both the transverse and longitudinal coherences are essentially a function of the phase variance only.

Line‐of‐sight propagation through atmospheric turbulence near the ground

The mean sound levels resulting from the interference between direct waves and those reflected from the ground are strongly influenced, especially at frequencies near interference minima, by fluctuations in phase and amplitude of the sound waves induced by propagation through atmospheric turbulence. Since it was found experimentally that the correlation length (∠1.1 m) of the meteorological fluctuations is comparable to the separation between the interfering sound paths, previous theoretical work by Ingard and Maling [J. Acoust. Soc. Am. 35, 1056–1058 (1963)] has been extended to allow for partial covariance between the two waves. The theory has been further extended to use the calculations of fluctuations in phase and amplitude of spherical waves, and to include the explicit calculation of the fluctuating acoustical index of refraction from the fluctuating values of temperature and wind velocity. Measurements (1–6 kHz) have been made of the interference spectrum at 15, 30, and 45 m from a point source 1....

Effects of atmospheric turbulence on the interference of sound waves near a hard boundary

The sound field due to a point source behind a barrier on ground of finite impedance has been calculated from five theories that differ mainly in their theoretical approach to diffraction and the model for ground impedance. These predicted values for the sound field have been compared with results measured outdoors using plywood barriers on different combinations of hard and soft ground. Each of these theories allows for interference due to differences between several paths of propagation, determined by the geometry of the source, receiver, barrier, and ground. One of these theories that shows good agreement with measurements, has been extended to calculate the sound spectrum level behind a barrier due to an incoherent line source, and further, to calculate the overall or A‐weighted sound level for a known source spectrum. Results suggest that there is a significant effect, due to the presence of the ground, that is much greater than that due to absorptive properties of the barrier. Results also predict sound level reductions that differ from predictions using well‐known barrier theories (most noticeably a smaller insertion loss): these differences can be of the order of 10 dB(A) depending on geometry, source spectrum, and acoustical condition of the ground.

J. E. Piercy

Papers

Effective flow resistivity of ground surfaces determined by acoustical measurements

Outdoor sound propagation over ground of finite impedance

Line‐of‐sight propagation through atmospheric turbulence near the ground

Effects of atmospheric turbulence on the interference of sound waves near a hard boundary

Noise reduction by barriers on finite impedance ground