Home 
Forward
3.5 Attenuation
Effects of Foliage
3.5.1 K-Band Effects
Measurements made in Austin, Texas during February and May when the trees
were without and with leaves, respectively, enabled a "foliage’’ adjustment
model to be developed. The formulation relates equal probability attenuation
(dB) corresponding to "foliage" and "no-foliage" cases at K-Band and is
given by
|
|
|
(3-16)
|
and where
|
(3-17)
|
|
(3-18)
|
This formulation was derived from "no-foliage"
and "foliage" mobile measurements made during February and May of 1994,
respectively, in Austin, Texas during the ACTS campaigns. The time-series
signal level characteristics for these two cases are shown in Figure
3-17 and Figure 3-18, where the measurements
were made along a one kilometer segment of a street heavily populated with
Pecan trees. Shown are the minimum, maximum, and average signal levels
over one second periods for a sampling rate of 1 KHz. The corresponding
cumulative fade distributions are shown in Figure
3-19 where the dashed and solid curves correspond to the "no-foliage"
and "foliage" cases. The direction of travel for these runs was approximately
orthogonal to the satellite pointing direction. The optical blockage to
the satellite during the full foliage period was estimated to be well in
excess of 55%. Performing a least square fit associated with equal probability
levels of the attenuation for the two curves in Figure
3-19, the formulation (3-15) was derived. A comparison
of the measured and predicted levels describing the foliage versus no foliage
fades is given in Figure 3-20. The predicted
curve (dashed) is shown to agree to within a fraction of a dB up to fades
(with foliage) of approximately 38 dB.
Figure 3-17: Minimum, maximum, and average fades over 1 second interval
at 20 GHz for a tree-lined run in Austin, Texas during February 6, 1994.
Deciduous trees (Pecan) were devoid of leaves.
Figure 3-18: Minimum, maximum, and average fades over 1 second intervals
at 20 GHz for a tree-lined run in Austin, Texas during May 2, 1994. Trees
(Pecan) were in full foliage.
Figure 3-19: Cumulative distributions at K-Band (20 GHz) for foliage and
no-foliage runs in Austin, Texas.
Figure 3-20: Measured and predicted levels of "foliage" versus "no-foliage"
fades at 20 GHz.
Figure 3-21
represents an independent validation of the above foliage formulation for
a set of static measurements at f = 19.6 GHz in Austin,
Texas pertaining to a Pecan tree in full foliage and the same tree devoid
of leaves [Vogel
and Goldhirsh, 1993a]. The resulting distributions calculated from
measurements correspond to the left and right solid curves, respectively.
The dashed curves represent the predicted levels using the formulation
given by (3-15). That is, the dashed curve on the
right is the predicted level of A(Foliage) where equal probabilities of
the measured A(No Foliage) as given by the left solid curve was injected
into (3-15). The left dashed curve represents the
predicted levels of the A(No Foliage), where the measured levels of A(Foliage)
as given by the right solid curve values were injected into (3-15).
The formulation (3-15) generally produces agreement
to within 1 dB or smaller over most of the indicated static probability
range.
Figure 3-21: Independent validation of "no-foliage" versus "foliage" prediction
formulation (3-15) employing static measurements
in Austin, Texas at 20 GHz. Solid curves represent measurements made during
different seasons and dashed curves are the predicted levels.
3.5.2 UHF (870 MHz)
In Chapter
2 it was shown that a 35% increase in the dB attenuation was experienced
at 870 MHz when comparing attenuation from trees having no foliage and
those having foliage (winter versus summer) for the static measurement
case. This case corresponded to a configuration in which the vehicle was
stationary and the propagation path intersected the canopy. Seasonal measurements
employing a helicopter as the transmitter platform were also performed
by the authors for the dynamic case in which the vehicle traveled along
a tree-lined highway in Central Maryland (Route 295) along which the propagation
path was shadowed over approximately 75% of the road distance [Goldhirsh
and Vogel, 1987; 1989]. Cumulative
fade distributions obtained from measurements in October 1985 (trees with
leaves) and in March 1986 (trees without leaves) are shown in Figure
3-22. The results demonstrate the following:
At f = 870 MHz, 
:
 .
|
|
Equation (3-19) states that over the percentage range
from 1% to 80% of the seasonal cumulative distributions, there is an average
increase of 24% (at equal probability values) in the dB fade of trees with
leaves relative to trees with no leaves. The dB error associated with the
above formulation over this percentage range for the curves in Figure
3-22 is less than 0.2 dB. The percentage fade increase (seasonal) for
the dynamic case (24%) is less than that for the static case (35%) because
the former case represents a condition in which the optical path is always
shadowed, whereas the dynamic case has associated with it measurements
between trees.
The question arises "Why is there a small difference between the "foliage"
and "no-foliage" distributions at UHF (e.g. 24% change), whereas at K-Band,
there is a large change between the distributions for these two scenarios
(e.g., two to three times)?" This question may be answered as follows:
The major contributor due to tree attenuation at the UHF frequency (wavelength
of approximately 35 cm) for trees without leaves are the branches, since
the separation between contiguous branches along the path are generally
smaller than 35 cm. That is, since the branch separation is generally smaller
than a wavelength at UHF, no substantial difference exists between the
"leaf" and "no-leaf" cases as both scenarios result in attenuating environments.
On the other hand, because the K-Band measurements have an associated wavelength
of 1.5 cm, the separation between branches for the "no leaf" case is generally
larger than this dimension resulting in a smaller relative fading condition.
On the other hand, the "full blossom" case is highly attenuating at K-Band
because of the continuous blockage caused by the high density of leaves.
Figure 3-22: Cumulative fade distributions at L-Band (1.5 GHz) in central-Maryland
during various seasons.
Home
Forward