ABSTRACT
Most community noise problems require information about the attenuation of A-weighted noise. Full wave numerical models, such as the Fast Field Program (FFP) and the Parabolic Equationmethod (PE) have been developed to compute the sound field in a complexoutdoor environment. However, evaluationofA-weighted mean square pressure by these numerical methods demands considerable computational resources. For some applications, it may be sufficient to assume that the atmosphere is vertically stratified so that an effective sound speed gradient can be used to replace wind and temperature gradients. With this assumption a ray-trace approach is convenient and, for some situations, it is adequate. In this section we investigate the analytical basis for ray tracing and in the following section we compare ray-trace and full-wave predictions. Ray tracing assumes an effective sound speed profile due to wind and temperature gradients and sums the contributions from all rays that are computed to pass through a chosen receiver. Such rays between source and receiver are known as eigenrays. For sources close to the ground, it is necessary to take into account any ground reflections. Rather than use plane wave reflection coefficients to describe these ground reflections, a better approximation is to use spherical wave reflection coefficients (see Chapter 2). Such an approach has resulted in a heuristic modification of the Weyl-Van der Pol formula (2.40) [1, 2]. It has been shown [3, 4] that if the rays have not passed through turning points and there is a single reflection at the ground, the resulting formulation represents the first term of an asymptotic solution of the full wave equation and is valid at short ranges. Furthermore, not only has it been demonstrated numerically that the ray-trace solution agrees reasonably well with other numerical schemes (see section 11.4), but there are also experimental data [e.g. 5] that agree tolerably well with ray-trace predictions based on a linear sound speed profile.
