230 Chapter 9
9.2.3.2 Absorption from Trees and Shrubs
When dealing with issues related to outdoor noise prop-
agation one may need to predict the anticipated noise
reduction that can be expected from vegetation. In this
instance, some of the general attributes of a tree-barrier
such as height and width can be modeled from geome-
try, but others like leaf density, wind resistance, or dif-
fraction effects from trunks may prove very difficult to
describe either analytically or geometrically. In this
instance an empirical model that fits experimental data
to polynomial equations based on statistical regression
is the most appropriate^19 to yield the sound pressure
level at various distances from a sound source while
taking into account tree height, width of the tree barrier,
wind velocity, and tree type. An example of such an
equation is presented below, and it is shown to give
excellent (±1 dB) accuracy between predicted and
observed levels for distances extending 150 ft to 400 ft
from the source that is assumed to be truck noise on an
interstate highway. The receiver is shielded from the
traffic by a belt of conifer trees planted along the
interstate.
(9-17)where,
LdB is the predicted sound level behind the tree belt,
H is the height of the tree belt, expressed in feet,
W is the width of the tree belt, expressed in feet,
V is the wind velocity component in the direction of the
sound propagation, expressed in mph,
D is the distance from the receiver to the tree belt.
Other equations are available for different sources
and different types of trees. In this class of empirical
models, no attempt is made to support the equation by
analytical expressions but this does not affect the
usefulness or the accuracy of the model.
9.2.4 Hybrid Models
As the name implies hybrid models use a combination
of techniques to yield results and the choice of the tech-
nique may be based on a specific need such as fast out-
put, accuracy, range of applicability, etc. A hybrid
model can combine the inherent accuracy of the image
method for the determination of reflection arrival time
in the specular case, with an adaptive beam-tracing
approach when diffusion is required, and may also
incorporate some BEM computations for complicated
materials wherever required. A hybrid model can also
rely on empirical approaches to provide a confidence
factor for results obtained from physical scale models or
from statistical approaches.An example of hybrid techniques can be found in
models that are aimed at assessing outdoor noise propa-
gation.^20 In this instance, the objects that are in the path
of the sound waves are typically buildings or large
natural obstacles and can be considered to be much
larger than the wavelength, except for the lowest
frequencies, and as such, the geometrical acoustics
assumptions apply very well; as such, the image method
is very appropriate to compute reflection paths between
obstacles. On the other hand one cannot ignore the fact
that outdoor noise may contain a lot of low frequencies
and that diffraction effects will take place; in this
instance the model must use an appropriate description
of diffraction such as the one presented in Chapter
4—Acoustical Treatment of Rooms and the model may
also be refined from empirical data table to represent
complicated sources such as car traffic, aircraft, and
trains since point source assumptions become invalid
and the sources are also moving. Figs. 9-20 and 9-21
shows the type of noise prediction maps that can be
obtained from such a model; in the first instance the
noise sources are a combination of street traffic and
large mechanical systems, and the model takes into
account the diffraction effects of various buildings. In
the second instance, the model is used to assess the
difference in expected noise levels between different
types of pavements (asphalt vs. concrete) based on
traffic data on a segment of road that is surrounded by
residences.Hybrid models are also found in construction appli-
cations, where they combine analytical techniques
based on specific equations with databases of test
results obtained in the laboratory and in the field. As an
example, a simple model could be developed using the
Mass Law in order to predict the sound transmission
between two spaces and yield an estimate of the Sound
Transmission Class (STC) of the partition, however, the
results would not be very useful because they would
extensively be influenced by the construction technique
and the presence of flanking paths. With a model that
takes into account the construction techniques of the
partition,^21 the results are much more accurate and
provide the designer with valuable insight on the weak
links of the construction as they pertains to noise
transmission.LdB 81.65 0.2257– H–0.0229W
0.728V–0.0576D= +