Concerns about noise in the community date back to the dawn of recorded history. Then, after centuries of relatively little activity, scientific interest grew during the 17th century and social concerns were again voiced during the 19th century. Many of the wave-propagation mechanisms relevant outdoors were understood at least qualitatively by the late 1800s. Today, knowledge of sound propagation phenomena is of great economic and social importance because of environmental and other concerns. Reality is more complicated than geometrical spreading above flat ground. Some grounds are acoustically hard like concrete, and others soft as snow. Corresponding reflection coefficients are complex and vary with angle. Grounds may not be flat, leading to shadow zones or alternatively multiple reflections at the ground. Gradients of wind or temperature refract waves either upwards (upwind or in a temperature lapse) or downwards (downwind or in a temperature inversion). Atmospheric turbulence causes fluctuations in the acoustical effects. Many of these features mutually interact. Measured sound pressure levels owe as much to near-surface weather and to ground shape and impedance as to acoustical factors such as source and receiver heights and their separation. This tutorial paper emphasizes field measurements and simple physical interpretations.

A general review is presented of most areas of sound-propagation outdoors that are of interest for the control of community noise. These areas are geometrical spreading, atmospheric absorption, ground effect, (near horizontal propagation in a homogenous atmosphere close to flat ground), refraction, the effect of atmospheric turbulence, and the effect of topography (elevation, hillsides, foilage, etc.) The current state of knowledge in each area is presented and suggestions made concerning research activities, applications of existing research, and practical problems which arise in the prediction of noise levels.

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.

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.

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 is however penetrated at low frequencies by a ground wave, to an extent that depends principally on distance and ground impedance, and at higher frequencies by interference between direct and ground-reflected waves to an extent that depends also on source and receiver heights. These phenomena, well established at ranges up to about 1000 ft and in some aspects to over 3000 ft, have been extrapolated theoretically to the order of 10 000 ft so that simple effects of topography and meteorology can be added to show how reflection or refraction acts in conjunction with ground impedance to result in penetration of the shadow region. Subject Classification: [43]28.40; [43]20.55.