Chapter 2

Attenuation Due to Trees:

Static Case

Table of Contents
2 Attenuation Due to Trees: Static Case ___________________________________ 2-1
2.1 Background _________________________________________________________ 2-1
2.2 Attenuation and Attenuation Coefficient at UHF___________________________ 2-2
2.3 Single Tree Attenuation at L-Band ______________________________________ 2-4
2.4 Attenuation through Vegetation: ITU-R Results ___________________________ 2-5
2.5 Distributions of Tree Attenuation at L-Band and K-Band ___________________ 2-6
2.6 Seasonal Effects on Path Attenuation ____________________________________ 2-8
2.6.1
2.6.2
2.6.3

Effects of Foliage at UHF _________________________________________________ 2-8

Effects of Foliage at L-Band ______________________________________________ 2-10
Effects of Foliage at K-Band ______________________________________________ 2-10

2.7 Frequency Scaling Considerations______________________________________ 2-11

2.7.1
2.7.2
2.7.3

Scaling between 870 MHz and L-Band______________________________________ 2-11

Scaling between 1 GHz and 4 GHz _________________________________________ 2-12
Scaling between L-Band and K-Band _______________________________________ 2-14

2.8 Conclusions and Recommendations_____________________________________ 2-15

2.9 References _________________________________________________________ 2-16

Table of Figures
Figure 2-1: LMSS propagation path shadowed by the canopies of one or two trees in which the
attenuation path length is relatively well defined. .......................................................................... 2-1
Figure 2-2: Low elevation propagation through a grove of trees giving rise to ambiguity in
attenuation path length.................................................................................................................. 2-2
Figure 2-3: Attenuation coefficients as described by the ITU-R for both "short paths (square
points) and "long path" (solid and dashed lines and diamond point at 10 GHz) scenarios. .............. 2-6
Figure 2-4: Cumulative distributions at L-Band (1.6 GHz) and K-Band (19.6 GHz). The cases
considered are: (A) K-Band Pecan in leaf, (B) K-Band Magnolia (evergreen), (C) K-Band
Pecan without leaves, (D) L-Band Pecan in leaf, (E) L-Band Pecan without leaves, (F) K-Band
unobstructed line-of-sight ............................................................................................................. 2-7
Figure 2-5: Configuration showing the approximate dimensions of the Callery Pear tree and the
relative location of the receiver. All dimensions are expressed in meters....................................... 2-8
Figure 2-6: Static tree attenuation versus elevation angle at 870 MHz for the Callery Pear tree
configuration in Figure 2-5. Triangles represent the full-foliage case, diamonds the no-foliage
case.............................................................................................................................................. 2-9
Figure 2-7: Comparison of measured (solid curves) and predicted (dashed) attenuation distributions
at 19.6 GHz corresponding to foliage and non-foliage cases. ....................................................... 2-12

2-i

Figure 2-8: Ratio of Attenuations versus frequency using different frequency scaling criteria
normalized to 1 GHz. The solid curve corresponds to (2-7), the dashed curve to (2-8) and the
dot-dashed curve to (2-9)............................................................................................................ 2-13
Figure 2-9: Frequency scaling of attenuation distribution of measured L-Band (1.6 GHz) to S-Band
frequencies using different criteria. Also shown is measured S-Band distribution........................ 2-13
Figure 2-10: Cumulative distributions at L (1.6 GHz; curve A) and K (19.6 GHz curve B) derived
from measurements. Curves C, D, and E are the frequency scaled distributions (L to K)
derived employing (2-9), (2-7), and (2-8), respectively................................................................ 2-15

Table of Tables
Table 2-1: Summary of Single Tree Attenuations at f = 870 MHz. ........................................................ 2-3
Table 2-2: Attenuation coefficient and average attenuations at 1.6 GHz of the different tree types.......... 2-5
Table 2-3: Median and 1% attenuation and attenuation coefficients at K- and L-Band. ........................... 2-7

2-ii

Chapter 2
Attenuation Due to Trees: Static Case

2.1 Background
A typical scenario in which fading occurs is depicted in Figure 2-1, which shows a
vehicle receiving satellite transmissions. The vehicle, which has an antenna mounted on
its roof, is presumed to be at a distance of 10 to 20 m from the roadside trees, and the
path to the satellite is generally above 20 in elevation. The antenna is to some extent
directive in elevation such that multipath from lower elevation (i.e., near zero degrees and
below) is filtered out by the antenna gain pattern characteristics. Although there may
exist multipath contributions at various azimuths, shadowing from the canopies of one or
two trees give rise to the major attenuation contributions. That is, the signal fade for this
case is due primarily to scattering and absorption from both branches and foliage where
the attenuation path length is the interval within the first few Fresnel zones intersected by
the canopies.

Figure 2-1: LMSS propagation path shadowed by the canopies of one or two trees in
which the attenuation path length is relatively well defined.

This geometry is in contrast to the configuration in which the transmitter and receiver are
located near the ground and propagation takes place through a grove of trees as shown in

2-2

Propagation Effects for Vehicular and Personal Mobile Satellite Systems

Figure 2-2. The attenuation contribution for this configuration is a manifestation of the
combined absorption and multiple scattering from the conglomeration of tree canopies
and trunks. For this case, an estimation of the attenuation coefficient from attenuation
measurements requires knowledge of the path length usually estimated to be the grove
thickness. This thickness may encompass a proportionately large interval of nonattenuating space between the trees. Hence attenuation coefficients as derived for groves
of trees may underestimate the attenuation coefficient vis--vis those derived for path
lengths intersecting one or two contiguous canopies for LMSS scenarios as shown in
Figure 2-1. This chapter deals primarily with characterizing the path attenuation through
tree canopies pertaining to the scenario of Figure 2-1, although attenuation coefficients
associated with that of Figure 2-2 are briefly characterized in Section 2.4.

Figure 2-2: Low elevation propagation through a grove of trees giving rise to ambiguity
in attenuation path length.

Static measurements of attenuation due to isolated trees for LMSS configurations have
been systematically performed at UHF, L-Band, S-Band and or K-Band by Benzair et al.
[1991], Butterworth [1984a; 1984b], Cavdar et al. [1994], Vogel and Goldhirsh [1994;
1993; 1986], Vogel et al. [1995], Ulaby et al. [1990], and Yoshikawa and Kagohara
[1989].

2.2 Attenuation and Attenuation Coefficient at UHF

For those cases in which shadowing dominates, the attenuation primarily depends on the
path length through the canopy, and the density of foliage and branches in the first
Fresnel region along the line-of-sight path. The receiver antenna pattern may also
influence the extent of fading or signal enhancements via the mechanism of multipath
scattering from surrounding trees or nearby illuminated terrain. An azimuthally omnidirectional antenna is more susceptible to such multipath scattering than a directive
antenna.
Nevertheless, the authors found through measurements and modeling
considerations for LMSS scenarios, the major fading effects are a result of the extent of
shadowing along the line-of-sight direction.
In Table 2-1 is given a summary of the single tree attenuation results at 870 MHz
based on the measurements by the authors [Vogel and Goldhirsh, 1986; Goldhirsh and
Vogel, 1987] who employed remotely piloted aircraft and helicopter transmitter
platforms. The attenuations were calculated by comparing the power changes for a
configuration in which the receiving antenna (on the roof of a van) was in front of and

Attenuation Due to Trees: Static Case

2-3

behind a particular tree. The former and latter cases offered non-shadowed and
maximum shadowing conditions, respectively, relative to the line of sight propagation
path from the transmitter on the aircraft to the stationary receiver. During each flyby, the
signal levels as a function of time were expressed in terms of a series of median fades
derived from the 1024 samples measured over one second periods. The attenuation
assigned to the particular flyby was the highest median fade level observed at the
measured elevation angle. It may be deduced that the motion of the transmitter aperture
and the receivers sampling rate of 1024 samples per second resulted in more than 200
independent samples averaged each second. This sample size is normally adequate to
provide a well-defined average of a noisy signal. The individual samples from which the
median was derived over the one-second period were observed to fluctuate on the
average 2 dB about the median due to the influence of variable shadowing and
multipath.
Table 2-1: Summary of Single Tree Attenuations at f = 870 MHz.
Tree Type

Attenuation (dB)

Attenuation Coefficient (dB/m)

Largest

Average

Largest

Average

Burr Oak*

13.9

11.1

1.0

0.8

Callery Pear

18.4

10.6

1.7

1.0

Holly*

19.9

12.1

2.3

1.2

Norway Maple

10.8

10.0

3.5

3.2

Pin Oak

8.4

6.3

0.85

0.6

Pin Oak*

18.4

13.1

1.85

1.3

Pine Grove

17.2

15.4

1.3

1.1

Sassafras

16.1

9.8

3.2

1.9

Scotch Pine

7.7

6.6

0.9

0.7

White Pine*

12.1

10.6

1.5

1.2

Average

14.3

10.6

1.8

1.3

RMS

4.15

2.6

0.9

0.7

The first column in Table 2-1 lists the trees examined where the presence of an asterisk
corresponds to results of measurements at Wallops Island, VA in June 1985 (remotely
piloted aircraft), and the absence of the asterisk represents measurements in central MD
in October 1985 (helicopter). During both measurement periods, the trees examined were
approximately in full foliage conditions. The second and third columns labeled Largest
and Average represent the largest and average values, respectively, of attenuation (in
dB) derived for the sum total of flybys for that particular tree. The fourth and fifth
columns denote the corresponding attenuation coefficients derived from the path length
through the canopy. The path length was estimated from measurements of the elevation
angle, the tree dimensions, and the relative geometry between the tree and the receiving
antenna height. The dependence of the attenuation on elevation angle is described in
Section 2.6. We note that the attenuations from Pin Oak as measured at Wallops Island
(with asterisk) is significantly larger than that measured in central Maryland (without

2-4

Propagation Effects for Vehicular and Personal Mobile Satellite Systems

asterisk) because the former tree had a significantly greater density of foliage over
approximately the same path length interval. This result demonstrates that a description
of the attenuation from trees for LMSS scenarios may only be handled employing
statistical processes.
Butterworth [1984b] performed single tree fade measurements at 800 MHz
(circularly polarized transmissions) at seven sites in Ottawa, Canada over the path
elevation interval 15o to 20o. The transmitter was located on a tower and receiver
measurements were taken at a height of 0.6 m above the ground. Measurements were
performed from April 28 to November 4, 1981 covering the period when leaf buds started
to open until after the leaves had fallen from the trees. A cumulative distribution of
foliage attenuation readings covering a 19 day period in June 1981 was noted to be
lognormal, where the fades exceeded 3 and 17 dB for 80% and 1% of the measured
samples, respectively. The median attenuation was approximately 7 dB with an
approximate median attenuation coefficient of 0.3 dB/m (24 m mean foliage depth). The
average attenuation coefficient of Butterworth is smaller than those measured by the
authors in central Maryland and Virginia. The disparity between these results may be
due to differences in the methods of averaging, the heights of the receiver and the
interpretation of the shadowing path length as previously described.

2.3 Single Tree Attenuation at L-Band

Single tree attenuation measurements at 1.6 GHz were conducted in Turkey between
April and September of 1993 by Cavdar et al. [1994]. The transmitter was placed atop a
building and the receiver antenna was located on top of a mobile unit which positioned
itself at different locations in the shadowed region of a number of trees.
Table 2-2 gives the total path attenuation and attenuation coefficients for a series of
trees. The average attenuation and attenuation coefficient are similar at UHF (Table 2-1)
to their respective values at L-Band (Table 2-2), although their respective RMS values
are significantly different.

Attenuation Due to Trees: Static Case

2-5

Table 2-2: Attenuation coefficient and average attenuations at 1.6 GHz of the different
tree types.
Tree Type

Average
Attenuation (dB)

Attenuation
Coefficient
(dB/m)

Willow

10.45

1.1

Pine

18.0

1.8

Linden

9.1

1.4

European Alder

7.0

1.0

Acacia

6.75

0.9

Poplar

3.5

0.7

Elm

9.0

1.2

Hazelnut

2.75

1.1

Maple

16.25

1.25

White Spruce

20.1

1.75

Laurel Cherry

12.0

2.0

Plane

16.9

1.35

Fir

12.75

1.5

Fruit

9.6

1.2

Average

11.0

1.3

RMS

5.1

0.35

2.4 Attenuation through Vegetation: ITU-R Results

The ITU-R [1994] designates attenuation through vegetation in terms of ground-toground measurements over paths of approximately 100 m or more, in woodland, forest or
jungle with antenna heights of 2-3 m above the ground with only part of the ray passing
through the foliage as designated in Figure 2-2. Attenuation corresponding to this
scenario is referred to as long path. A short-path attenuation scenario is also
designated corresponding to short ground-to-ground or slant-path measurements through
the foliage of individual trees with foliage depths of no more than 10-15 m as shown in
Figure 2-1. Figure 2-3 gives the attenuation coefficient in dB/m versus frequency for
both the short-path (squares) and long-path (solid and dashed curves and diamond at
approximately 10 GHz). As mentioned, the long-path results show attenuation
coefficients significantly smaller than the short path results. Short-path attenuation
coefficients between 1 and 2 dB/m are generally indicated at frequencies between
approximately 1 and 4 GHz consistent with the results of Table 2-1 and Table 2-2.

Propagation Effects for Vehicular and Personal Mobile Satellite Systems

2-6
1E+1 8

Attenuation Coefficient (dB/m)

6
4
2

1E+0 8

Short Path Data

6
4
2

1E-1 8
6
4
2

Long Path

1E-2 8
6
4

1E-3
1E+1

6 8

1E+2

6 8

1E+3

6 8

1E+4

6 8

1E+5

Frequency (MHz)
Figure 2-3: Attenuation coefficients as described by the ITU-R for both "short paths
(square points) and "long path" (solid and dashed lines and diamond point at 10 GHz)
scenarios.

2.5 Distributions of Tree Attenuation at L-Band and K-Band

Static tree attenuation measurements at L- (1.6 GHz) and K-Band (19.6 GHz) were
executed in Austin, Texas [Vogel and Goldhirsh, 1993; 1994]. The trees sampled were
Pecan (deciduous) and Magnolia (evergreen). The L-Band measurements were
performed in December 1990 and July 1991 during which times the sampled Pecan tree
was without foliage and in full foliage, respectively. The transmitter was placed atop
a 20 m tall tower and the receiving antenna was mounted on a motorized positioner
placed within the geometric shadow zone of each tree. The antenna was moved slowly
over a horizontal distance of several meters, and the received power was sampled every
0.1 s for about 100 s. The K-Band measurements were performed in March and May,
1993 employing the same approximate geometry. These months also correspond to
periods in which the Pecan tree was without leaves and in full foliage, respectively. The
receiving antenna was hand-held and moved horizontally over a distance of several
meters, first in the shadow of the same Pecan tree, and then in the shadow of the nearby
Magnolia tree. Quadrature detector receiver voltages were sampled at a 1000 Hz rate for
several minutes. The path lengths within the Pecan and Magnolia crowns were on
average 9 m and 4.5 m, respectively. The clear line-of-sight reference signal levels were
determined for all cases by moving the receiver to an equi-distant position where the
signal path was unobstructed.

Attenuation Due to Trees: Static Case

2-7

Resulting cumulative fade distributions are given in Figure 2-4 for cases in which
the canopy optically shadowed the line-of-sight path between the transmitter and the
receiver (curves A through E). Also given, as a reference, is the distribution for the
unobstructed K-Band line-of-sight case (curve F). In Table 2-3 is given a summary of the
total attenuation and attenuation coefficient for the median (50%) and 1% cases for the
different frequencies and tree scenarios characterized in Figure 2-4. The elevation angles
for the K and L-Band measurements relative to the Pecan tree were approximately 26
and 30, respectively.
Table 2-3: Median and 1% attenuation and attenuation coefficients at K- and L-Band.
Total Fade (dB)

Attenuation Coefficient (dB/m)

Tree Condition

Percentage

Clear LOS

Median

0.5

1%

2.6

Bare Pecan
Pecan in leaf

L-Band

K-Band

L-Band

K-Band

Median

10.3

6.9

1.1

0.75

1%

18.4

25.0

2.0

2.8

Median

11.6

22.7

1.3

2.5

1%

18.6

43.0

2.1

4.8

Magnolia

Median

19.6

4.4

(evergreen)

1%

39.6

8.8

100

9
8
7
6

Percentage of Locations >Abscissa

5
4
3

A
B
10

9
8
7
6
5

D
E

F
1
0

10

15

20

25

30

35

40

Fade (dB)

Figure 2-4: Cumulative distributions at L-Band (1.6 GHz) and K-Band (19.6 GHz). The
cases considered are: (A) K-Band Pecan in leaf, (B) K-Band Magnolia (evergreen),
(C) K-Band Pecan without leaves, (D) L-Band Pecan in leaf, (E) L-Band Pecan without
leaves, (F) K-Band unobstructed line-of-sight

Propagation Effects for Vehicular and Personal Mobile Satellite Systems

2-8

2.6 Seasonal Effects on Path Attenuation

2.6.1

Effects of Foliage at UHF

It is observed from the configuration of Figure 2-1, that at smaller elevation angles, the
path length through the canopy will increase. Likewise, it is expected that the attenuation
will also increase assuming the path cuts through the canopy. Single tree attenuation
measurements for 870 MHz at different elevation angles were performed in centralMaryland in October 1985 and in March 1986 for a Callery Pear tree employing a
helicopter as the transmitter platform [Goldhirsh and Vogel, 1987]. The approximate
geometry of the tree and relative location of the receiving antenna is given in Figure 2-5.
Figure 2-6 shows the corresponding results for the two seasons during which the tree was
in full foliage (October) and without leaves (March). Also shown are the linear fits for
each of the sets of data points.

45
To Helicopter
6.6
30

Antenna
Position

13.8

2.4

2
11.2
17.6

Figure 2-5: Configuration showing the approximate dimensions of the Callery Pear tree
and the relative location of the receiver. All dimensions are expressed in meters.

Attenuation Due to Trees: Static Case

2-9

20
18

Tree Attenuation (dB)

16
14
12
10
8
6
4
2
0
0

10

15

20

25

30

35

40

45

Path Elevation Angle (deg)

Figure 2-6: Static tree attenuation versus elevation angle at 870 MHz for the Callery Pear
tree configuration in Figure 2-5. Triangles represent the full-foliage case, diamonds the
no-foliage case.

The linear best fits in Figure 2-6 may be described as follows:

for 15o < <40o (full foliage)
A( ) = 0.48 + 26.2

(2-1)

for 15o < <40o (bare tree)

A( ) = 0.35 + 19.2

(2-2)

where is in degrees and A( ) is in dB. We derive from (2-1) and (2-2) over the given
angular range at 870 MHz:
for 5 dB < A(no foliage) < 14 dB,
A( full foliage) = 1.35 A(no foliage)

(2-3)

which states that for the static case, the maximum attenuation contribution (in dB) for the
full-foliage case is 35% larger than the corresponding attenuation for the bare tree case.

Propagation Effects for Vehicular and Personal Mobile Satellite Systems

2-10

This results suggests that the predominant attenuation at 870 MHz arises from the tree
branches via the mechanism of absorption and the scattering of energy away from the
receiver. The conclusion that the wood part of the tree is the major contributor to
attenuation has also been substantiated at UHF for the mobile case (Chapter 3). The
above results are limited to elevation angles above approximately 15. At smaller angles,
the path may pass through the bottom part of the canopy. More complicated scenarios
may result in which there may be multiple tree effects as depicted in Figure 2-2 or terrain
blockage may arise. The lower angle limit for mobile scenarios is broached in Chapter 3.
2.6.2

Effects of Foliage at L-Band

Upon analyzing the data points at equal probability levels for the L-Band distributions
with and without foliage (curves D and E of Figure 2-4), we observed that the percent
difference in fades between the foliage relative to the no-foliage cases for the Pecan tree
ranged from approximately 15% at 70% probability to 1% at 1% probability with an
average percent difference of approximately 7% (average fade difference = 0.8 dB). The
linear least square relation with a standard error of 0.1 dB relating the attenuation with
foliage to that with no foliage is given by,
for 9 dB < A(no foliage) < 18 dB
A( full foliage) = 2.33 + 0.9 A(no foliage)

(2-4)

Here again, we observe that the major contribution due to attenuation is the wood part of
the tree. Since the tree configurations, tree types, and foliage path lengths were different
for the Callery Pear sampled at UHF and the Pecan tree sampled at L-Band, no inference
should be made as to the frequency relationships pertaining to foliage versus no-foliage
fading.
2.6.3

Effects of Foliage at K-Band

Curves B and C in Figure 2-4 give the distributions for the attenuation with and without
foliage, respectively. The following formulation relates the attenuation for the foliage
case versus the no-foliage case:
for 5 dB < A(no foliage) < 25 dB (static case)
A( foliage) = a + bA(no foliage) c

(2-5)

where the coefficients a, b, and c are given by

a = 0.351
b = 6.8253
c = 0.5776

(2-6)

Although the above formulation was derived from a mobile run of a street lined with a
high density of Pecan trees in Austin, Texas, it appears to give predictions (dashed
curves) which agree quite well with the static runs (solid curves) as shown in Figure 2-7.

Attenuation Due to Trees: Static Case

2-11

The range of no-foliage attenuations shown in the line preceding (2-5) corresponds to the
static attenuation measurement range depicted in Figure 2-7.

2.7 Frequency Scaling Considerations

2.7.1

Scaling between 870 MHz and L-Band

Ulaby et al. [1990] measured the attenuation properties at 50 elevation associated with
transmission at 1.6 GHz through a canopy of red pine foliage in Michigan at both
horizontal and vertical polarizations. The path length through the canopy was
approximately 5.2 m and the average attenuations measured at horizontal and vertical
polarizations were 9.3 and 9.2 dB. Their measurements gave rise to an average
attenuation coefficient of approximately 1.8 dB/m. Combining this result at L-Band with
the average value of 1.3 dB/m at UHF given in Table 2-1, the frequency scaling
formulation found applicable between UHF (870 MHz) and L-Band (1.6 GHz) assuming
a full foliage scenario is given by
A( f 2 ) = A( f1 )

f2
f1

(2-7)

where A(f1 ) and A(f2 ) are the equal probability attenuations expressed in dB at the
indicated frequencies between 870 MHz and 1.6 GHz. This expression was found also to
be applicable for frequencies between UHF and S-Band for mobile scenarios as discussed
in Chapter 3.

Propagation Effects for Vehicular and Personal Mobile Satellite Systems

2-12

Percentage of Distance Fade > Abscissa

100

9
8
7
6
5
4
3

Predicted
2

10

9
8
7
6
5
4
3

Foliage

No Foliage

1
0

10

15

20

25

30

35

40

45

Fade (dB)
Figure 2-7: Comparison of measured (solid curves) and predicted (dashed) attenuation
distributions at 19.6 GHz corresponding to foliage and non-foliage cases.

2.7.2

Scaling between 1 GHz and 4 GHz

Benzair et al. [1991] performed static attenuation measurements on a mature deciduous

tree in full foliage at a series of frequencies between 1 and 4 GHz at an elevation angle of
approximately 45. They found the attenuation coefficient to obey the following
expression.
For 1 GHz f 4 GHz
MEL = 0.79 f 0.61

(2-8)

where f is the frequency in GHz and MEL (mean excess loss) represents the mean
attenuation coefficient in dB/m. This expression (normalized to 1 GHz) is plotted in
Figure 2-8 (dashed curve) with other frequency scaling curves. In Figure 2-9, the various
frequency-scaling criteria are also applied to the measured L-Band distribution and
compared with the measured S-Band distribution as measured by Vogel et al. [1995].
The three frequency scaling techniques are shown to give similar results over the
indicated frequency ranges.

Attenuation Due to Trees: Static Case

2-13

3.0

Ratio of Attenuations

2.5

2.0

1.5

1.0

Formulation
Ratio to 0.5

0.5

Ratio to 0.61
Exponential

0.0
1

Frequency (GHz)

Figure 2-8: Ratio of Attenuations versus frequency using different frequency scaling
criteria normalized to 1 GHz. The solid curve corresponds to (2-7), the dashed curve to
(2-8) and the dot-dashed curve to (2-9).

Percentage of Distance > Abscissa

100

9
8
7
6
5
4

Frequency Scaling Case

Square Root
Ratio to 0.61
EERS Scaling

3
2

Measured L-Band
10

9
8
7
6
5
4
3

Measured S-Band
2

1
0

8 10 12 14 16 18 20 22 24 26 28 30 32

Fade (dB)

Figure 2-9: Frequency scaling of attenuation distribution of measured L-Band (1.6 GHz)
to S-Band frequencies using different criteria. Also shown is measured S-Band
distribution.

Propagation Effects for Vehicular and Personal Mobile Satellite Systems

2-14
2.7.3

Scaling between L-Band and K-Band

A frequency scaling formulation which extends between L- and K-Band for the static
case and also applicable for the mobile case has been derived by the authors [Vogel and
Goldhirsh, 1993] and is given by
0.5
0.5

1
1

A( f 2 ) = A( f1 ) exp b
f 2
f1

(2-9)

where
b = 1.5

(2-10)

and where A(f1 ) and A(f2 ) are the attenuations in dB at frequencies f1 and f2 (in GHz).
Figure 2-10 shows a comparison of L- (1.6 GHz; curve A) and K-Band (19.6 GHz; curve
B) cumulative fade distributions calculated from static measurements of a Pecan tree in
full leaf. Curves C, D, and E are the frequency scaled distributions (L to K) derived
employing (2-9), (2-7), and (2-8), respectively.
The transmitters at L- and K-Bands were placed atop a 20 m tower and the receiver
systems were located within the geometric shadow zone of the tree where the elevation
angle was approximately 30o [Vogel and Goldhirsh, 1993]. The vertical scale represents
the percentage of optically shadowed locations over which the abscissa fade was
exceeded. Curve C represents the application of (2-9), D makes use of (2-7) and E was
derived applying (2-8) on the L-Band curve at equal probability values. The frequency
scaling formulation shows agreement to within a few dB for percentages between 2% and
20% for the static case. The formulations (2-7) and (2-8) are shown to give excessive
fades and are not applicable at K-Band. In the paper by Vogel and Goldhirsh [1993], a
multiplying constant of b = 1.173 was given as being applicable at the median fade level
(i.e., P = 50%). The constant as given by (2-10) is the suggested value for the static case
as it gives good agreement over the dominant part of the fade distribution curve B in
Figure 2-10. Hence, (2-9) is also the same frequency scaling formulation for the mobile
scenario case.

Attenuation Due to Trees: Static Case

2-15

Percentage of Distance > Abscissa

100

9
8
7
6
5

E
D

4
3

B
A
10

9
8
7
6
5
4
3
2

1
0

10

15

20

25

30

35

40

45

50

55

60

Fade (dB)
Figure 2-10: Cumulative distributions at L (1.6 GHz; curve A) and K (19.6 GHz curve
B) derived from measurements. Curves C, D, and E are the frequency scaled
distributions (L to K) derived employing (2-9), (2-7), and (2-8), respectively.

2.8 Conclusions and Recommendations

1. The average single tree attenuation at UHF (870 MHz) is 10.6 dB (2.6 dB RMS)
(Table 2-1).
2. The average single tree attenuation at L-Band (1.6 GHz) is 11 dB (5.1 dB RMS)
(Table 2-2).
3. The median single tree attenuation at K-Band (20 GHz) is 23 dB (Table 2-3).
4. The recommended frequency scaling formulation pertaining to single trees in full
foliage at frequencies between UHF (870 MHz) and K-Band (20 GHz) is given by

1 0.5 1 0.5
A( f 2 ) = A( f1 ) exp b
f 2
f1

(2-11)

Propagation Effects for Vehicular and Personal Mobile Satellite Systems

2-16

where A(f1 ), A(f2 ) are the respective equal probability attenuations (dB) at
frequencies f1, f2 (in GHz).
5. The dominant contributor to attenuation is the wood part of the tree at frequencies
between UHF (870 MHz) and S-Band (4 GHz). For example, foliage has been found
to introduce approximately 35% additional attenuation at UHF (Equation (2-3) and
15% at L-Band (Figure 2-4).
6. At K-Band (20 GHz), the wood and leaf parts of the tree are both important showing
increases due to foliage ranging from 2 to 3 times the attenuation (Figure 2-7).

2.9 References
Benzair, B., H. Smith, and J. R. Norbury, [1991], Tree Attenuation Measurements at
1-4 GHz for Mobile Radio Systems, Sixth International Conference on Mobile Radio
and Personal Communications, 9-11 December, London, England, pp. 16-20. (IEE
Conference Publication No. 351).
Butterworth, J. S. [1984a], Propagation Measurements for Land-Mobile Satellite
Systems at 1542 MHz, Communication Research Centre Technical Note 723,
August. (Communication Research Centre, Ottawa, Canada.)
Butterworth, J. S. [1984b], Propagation Measurements for Land-Mobile Satellite
Services in the 800 MHz, Communication Research Centre Technical Note 724,
August. (Communication Research Centre, Ottawa, Canada.)
Cavdar, I. H., H. Dincer, and K. Erdogdu [1994], Propagation Measurements at L-Band
for Land Mobile Satellite Link Design, Proceedings of the 7th Mediterranean
Electrotechnical Conference, April 12-14, Antalya, Turkey, pp. 1162-1165.
Goldhirsh, J. and W. J. Vogel [1987], Roadside Tree Attenuation Measurements at UHF
for Land-Mobile Satellite Systems, IEEE Transactions on Antennas and
Propagation, Vol. AP-35, pp. 589-596, May.
ITU-R [1994] (International Telecommunication Union, Radio Communications Study
Groups), Propagation Data Required for the Design of Earth-Space Land Mobile
Telecommunication Systems, Recommendation ITU-R PN.681-1, International
Telecommunication Union, ITU-R Recommendations, 1994 PN Series Volume,
Propagation in Non-Ionized Media, pp. 203-204.
Ulaby, F. T., M. W. Whitt, and M. C. Dobson [1990], Measuring the Propagation
Properties of A Forest Canopy Using A Polarimetric Scatterometer, IEEE
Transactions on Antennas and Propagation, Vol. AP-38, No. 2, pp. 251-258, Feb.
Vogel, W. J. and J. Goldhirsh [1994], Tree Attenuation at 20 GHz: Foliage Effects,
Presentations of the Sixth ACTS Propagation Studies Workshop (APSW VI),
Clearwater Beach, Florida, November 28-30, pp. 219-223. (Jet Propulsion
Laboratory Technical Report, JPL D-12350, California Institute of Technology,
Pasadena, California.)
Vogel, W. J., G. W. Torrence, and H. P. Lin [1995], Simultaneous Measurements of Land S-Band Tree Shadowing for Space-Earth Communications, IEEE Transactions
on Antennas and Propagation, Vol. AP-43, pp. 713-719, July

Attenuation Due to Trees: Static Case

2-17

Vogel, W. J. and J. Goldhirsh [1993], Earth-Satellite Tree Attenuation at 20 GHz:

Foliage Effects, Electronics Letters, Vol. 29, No. 18, 2nd September, 19, pp. 16401641.
Vogel, W. J., and J. Goldhirsh [1986], Tree Attenuation at 869 MHz Derived from
Remotely Piloted Aircraft Measurements, IEEE Transactions on Antennas and
Propagation, Vol. AP-34, No. 12, pp. 1460-1464, Dec.
Yoshikawa, M. and M. Kagohara [1989], Propagation Characteristics in Land Mobile
Satellite Systems, 39th IEEE Vehicular Technology Conference, pp. 550-556, 1-3
May.