You are on page 1of 15

Progress In Electromagnetics Research B, Vol.

26, 53–67, 2010
A NEW RAIN ATTENUATION CONVERSION TECH-
NIQUE FOR TROPICAL REGIONS
A. Y. Abdulrahman, T. A. Rahman, S. K. A. Rahim
and M. R. Ul Islam
Wireless Communication Centre (WCC)
Fakulti Kejuruteraan Elektrik
Universiti Technologi Malaysia (UTM)
81310, Skudai-Johor, Malaysia
Abstract—Rain attenuation is one of the most crucial factors to
be considered in the link budget estimation for microwave satellite
communication systems, operating at frequencies above 10 GHz. This
paper presents a mathematical model for converting terrestrial rain
attenuation data to be used for satellite applications at Ku-band.
In the proposed technique, the ITU-R P 618-9, together with a
combination of ITU-R P 530-12 and the revised Moupfouma model
have been adopted for satellite and terrestrial rain attenuation
predictions, respectively. The model has been used for transforming
the measured rain attenuation data of some DIGI MINI-LINKS
operating at 15 GHz in Malaysia, to be used for MEASAT 2
applications. It was found that the model predictions are fairly
reasonable when compared with direct beacon measurements in
Malaysia and similar tropical locations. The model will provide a
relatively accurate method for transforming the measured terrestrial
rain attenuation to be used for satellite applications; and therefore
substantially reduce the cost of implementing Earth-satellite links in
some tropical regions that have sufficient rain attenuation data for the
terrestrial links.
1. INTRODUCTION
In the design of Earth-space links for communication systems,
several factors must be put into consideration. These factors
include absorption in atmospheric gases; absorption, scattering and
Received 21 June 2010, Accepted 17 September 2010, Scheduled 27 September 2010
Corresponding author: A. Y. Abdulrahman (abdulrahman.yusuf@yahoo.com).
54 Abdulrahman et al.
depolarization by hydrometeors such as clouds and precipitation [1].
All these effects must be considered as they are the causes of
signal impairment for both terrestrial and Earth-satellite systems;
with rainfall being the most crucial especially at frequencies above
10 GHz. The effect of rainfall is more severe in tropical regions which
are characterized by heavy rainfall intensity and presence of large
raindrops [2, 3]. Raindrop size distribution changes with geographical
location and it can strongly influence rain specific attenuation and,
consequently, total rain attenuation.
The rain-attenuation analyses are important for the study of the
rain fade characteristics, which is a useful piece of information in the
link budget estimation for predicting the expected outage as a result of
rain attenuation on a microwave link [4, 5]. Long-term rain attenuation
data are usually available for the terrestrial links in most tropical
regions and so there is the need to convert them for use in satellite
applications. This has become imperative because of the fact that
these regions do not have sufficient earth-satellite data, which is useful
in the preliminary design of earth-satellite microwave links.
There are two broad classes of rain attenuation predictions on
any microwave link: The analytical models which are based on
physical laws governing electromagnetic wave propagation and which
attempt to reproduce the actual physical behavior in the attenuation
process; and the empirical models which are based on measurement
databases from stations in different climatic zones within a given
region [6, 7]. The aim of this paper is to present a mathematical model
for converting the rain-induced attenuation data on the terrestrial links
to be used for satellite applications at Ku-band, based on the ITU-R
recommendations ITU-R 618 -9 [8], ITU-R 530-12 [9] and the revised
Moupfouma model [10]. The results may also be extrapolated to be
used in other tropical regions that have similar situation.
2. A HANDFUL OF SOME EXISTING CONVERSION
TECHNIQUES AND MODELS FOR RAIN
ATTENUATION
Various conversion methods have earlier been developed for different
applications in rain attenuation predictions for both terrestrial and
satellite communication systems. These include the conversion
of rain rate integration time from higher time to a shorter one;
and the frequency scaling techniques for converting the measured
rain attenuation at one frequency to the equivalent value at other
frequencies for the same site.
However, no comprehensive research works have previously been
Progress In Electromagnetics Research B, Vol. 26, 2010 55
done on converting terrestrial rain attenuation data to be used
for satellite applications. The only published research works in
the literature are those of Singlair et al. [11, 19], who proposed a
transformation method for converting terrestrial rain attenuation time-
series into satellite attenuation time-series, based on terrestrial rain
attenuation measurement. The model is given by:
A
s
(t) =
k
s
L
s
1 +L
s
/d
0
_
A
T
(t)(1 +L
T
/d
0
)
k
T
L
T
_
α
s
α
T
(1)
where k
s
, k
T
, α
s
and α
T
are ITU-R recommended parameters whose
values depend on frequency band, polarization, temperature and
elevation, A
s
(t) is the transformed rain attenuation time series of the
satellite link, A
T
(t) is the measured rain attenuation time series of the
terrestrial link, L
s
and L
T
are the slant path and the physical length
of the satellite and terrestrial links, respectively; and d
0
is the ITU-R
reduction factor given by:
d
0
= 35e
(−0.015R
0.01
)
(2)
The major limit of the model being that the terrestrial rain
attenuation data were measured in Hungary, a temperate region
with rain rates less than 40 mm/h. This implies that the model
is unarguably unsuitable for tropical climates like Malaysia which
experiences heavy rain rates, as high as 120–130 mm/h. Rain rate is
the most crucial factor in any rain attenuation predictions. Therefore,
it has been considered as the principal input in the new proposed
conversion technique, as indicated in Equation (22). Also, the proposed
technique has been validated using the terrestrial rain attenuation data
measured in Malaysian tropical climate.
3. DEVELOPMENT OF THE PROPOSED RAIN
ATTENUATION CONVERSION TECHNIQUE
The development consists of three stages; first is the calculation of rain
attenuation over the satellite link using the ITU-R P. 618-9 and second
is the prediction of terrestrial rain attenuation, using a combination of
ITU-R P. 530-12 and the revised Moupfouma model. The third stage
involves derivation of the proposed transformation method, by finding
the ratio of the two attenuation values at the same rain rate, since
both links are sited in the same location.
3.1. Calculation of Rain Attenuation for the Satellite Link
The step-by-step procedure for calculating rain attenuation cumulative
distribution function (CDF) over the satellite link is given below:
56 Abdulrahman et al.
Step 1: Freezing height during rain, H
R
(km) is calculated from
the absolute values of station’s latitude and longitude, both in degrees,
as follows [12]:
H
R
= 0.36 +h
0
(3)
where h
0
is the 0

C isotherm height above mean sea level (km) and its
value can be obtained from the isotherm chats of the Recommendation
ITU-R P.839-3.
Step 2: The slant-path length, L
S
(km), below the rain height is
calculated as:
L
S
=
H
R
−H
S
sin θ
(4)
where H
S
is the station altitude, above the sea-level, in km and θ
(degrees) is the elevation angle.
Step 3: The horizontal projection of the slant path length, L
G
is
evaluated from:
L
G
= L
S
cos θ (5)
Step 4: The point rain rate, R
0.01
(mm/h), exceeded for 0.01% of
an average year may be obtained from one- minute integration rain rate
data, if it is available. An alternative method is by applying the Chebil
model, which makes use of long-term mean annual accumulation, M,
at the location under study; and is given by [13]:
R
0.01
= αM
β
(6)
where α and β are regression coefficients. This model was shown to
provide good approximation of the measured data. The regression
coefficients are defined as:
α = 12.2903 and β = 0.2973 (7)
We have found that the recommendation ITU-R P.837-5 is not
suitable for estimating R
0.01
in tropical Malaysian climate, because
it substantially underestimates the measured rain rate values [14, 21].
For instance, the ITU-R P.837-5 has predicted a value of 105 mm/h
for R
0.01
at UTM; whereas both direct measurement and Chebil’s
approximation of Equation (6) have given a value of approximately
125 mm/h for the same location.
Step 5: The specific attenuation, γ
s 0.01
(dB/km), for 0.01% of
time is calculated from:
γ
s0.01
= k
s
R
α
s
0.01
(8)
The parameters k
s
and α
s
depend on frequency, rain temperature,
raindrop size distribution, and polarization. Their values can be
obtained from ITU-R P.838-3 [15].
Progress In Electromagnetics Research B, Vol. 26, 2010 57
Step 6: The horizontal path adjustment reduction factor, r
h0.01
,
for 0.01% of the time is also calculated as:
r
h0.01
=
1
1 + 0.78
_
L
G
γ
0.01
f
−0.38[1 −exp −2L
G
]
(9)
Therefore, the horizontally adjusted rainy slant path length is
calculated from:
L
h0.01
=
_
L
G
r
h0.01
/ cos θ, for ρ > θ
(H
R
−H
S
)/ sin θ, for ρ ≤ θ
(10)
where
ρ = tan
−1
_
H
R
−H
S
L
G
r
h0.01
_
(11)
Step 7: The vertical reduction factor, r
v0.01
, for 0.01% of the time
is also given by:
r
v0.01
=
1
1+

sin θ
_
31 (1−exp(−θ/[1+σ]))

L
G
γ
0.01
f
2
−0.45
_ (12)
where
σ =
_
36 −|ϕ|, for |ϕ| < 36

0, for |ϕ| ≥ 36

(13)
Step 8: The effective path length through rain, L
eff
(km) is finally
obtained by multiplying the horizontally adjusted slant path by the
vertical reduction factor of Equation (10); that is,
L
eff
= L
h0.01
r
v0.01
(14)
Step 9: The predicted attenuation exceeded for 0.01% of an
average year is obtained from:
A
s0.01
= γ
s0.01
L
eff
(15)
Step 10: The predicted attenuation to be exceeded for other
percentages, p of an average year may be calculated from the value
of A
s0.01
by using the following:
A
s%p
= 0.12 ∗ A
s0.01
∗ p
(−(0.546+0.043∗log 10(p)))
(16)
58 Abdulrahman et al.
3.2. Calculation of Rain Attenuation for the Terrestrial Link
Step 1: The specific attenuation for the terrestrial link is given by:
γ
T0.01
= k
T
R
α
T
0.01
(17)
Again, parameters k
T
and α
T
depend on frequency, rain temperature,
raindrop size distribution, and polarization. Their values can be
obtained from ITU-R P.838-3, as earlier mentioned in step 5 of
section 3.1.
Step 2: For the terrestrial link, path reduction factor is calculated
using the revised Moupfouma’s model given by [10]:
δ(R
0.01
, L
T
) = exp
_
−R
0.01
1 +ζ(L
T
) ∗ R
0.01
_
(18)
where δ(R
0.01
, L
T
) is the reduction factor for 0.01% of the time
observed on any terrestrial microwave radio path of length, L
T
(km),
such that the equivalent propagation path, L
eq
is:
L
eq
(R
0.01
, L
T
) =L
T
∗ exp
_
−R
0.01
1+ζ(L
T
) ∗ R
0.01
_
with ζ(L
T
) = −100 for any L
T
≤ 7 km (19a)
and
L
eq
(R
0.01
, L
T
) =L
T
∗ exp
_
−R
0.01
1+ζ(L
T
)∗ R
0.01
_
with ζ(L
T
)=
_
44.2
L
T
_
0.78
for any L
T
>7 km (19b)
Step 3: The predicted attenuation exceeded for an average year is
then obtained from the following:
A
T0.01
= γ
T0.01
L
T
δ(R
0.01
, L
T
) (20)
The predicted attenuation to be exceeded for other percentages,
p = 0.01%, of an average year may be calculated from the value of
A
T0.01
by using the Equation (16). In this paper we have deliberately
chosen not to use the ITU-R path reduction factor of Equation (2)
for the terrestrial link. This is due to its unsuitability for the tropical
regions.
3.3. Data Collection for Terrestrial Links
The received relative intermediate frequency (IF) rain attenuation
data, sampled every second, were collected from seven 15 GHz oper-
ational microwave links of DIGI Telecommunications. Meteorological
Progress In Electromagnetics Research B, Vol. 26, 2010 59
stations were installed at the particular locations of the investigated
links, in order to be able to determine the relationship between the
rain attenuation and rain intensity.
Four year precipitation data were collected at Universti Teknologi
Malaysia, (UTM), Malaysia, Skudai campus and the data have been
used to investigate the DIGI link located at Johor Bahru. The rain
rate data were collected from January 2003 to December 2006, using
Casella rain gauge and thirty months using OSK rain gauge. Both
gauges are tipping bucket type and having 0.5 mm sensitivity. Casella
records the total rainfall occurring in each minute without recording
non-rainy events; therefore the rain rate is recorded as an integral
multiple of 0.5 mm/min, or 30 mm/h. In contrast, OSK records the
actual tipping time up to decimal of seconds.
The campus is located at Johor, southern part of Malaysia
peninsula close to Singapore. Malaysia is located in the South-Eastern
part of Asia, and falls in the zone P of the ITU-R rainfall rate
climatic zone with annual average accumulation as high as 4184.3 mm.
The Malaysian climate is tropical, and is characterized by uniform
temperature, high humidity and heavy rainfall which arises mainly
from the maritime exposure of the country. The heaviest amounts of
rainfall are noted to occur in the last two months of the year; that
is November and December. The uniform periodic changes in the
wind flow patterns result in seasons’ classifications as: The south-west
monsoon, the north-east monsoon and the two shorter inter-monsoon
seasons.
The specifications for the MINI-LINK are given in Table 1, while
the measured rain attenuation data at UTM, Johor Bahru are shown
in Table 2.
4. MATHEMATICAL DERIVATION OF THE
CONVERSION FACTOR
First it is logical to assume that the terrestrial rain attenuation
is greater than that of the satellite, under the same conditions of
rainfall intensity. This fact is justified as follows [15, 16]: Satellites in
geostationary orbit are 35,800 km above the earth, and since rain only
forms in the troposphere, which extends seven miles above the earth, a
signal traveling through a rain cell will experience attenuation during
only a small portion of its transmission path. Therefore, terrestrial
microwave transmissions are more susceptible to the effects of rain
attenuation because their signal paths are entirely in the troposphere,
and the signal may pass through an entire rain cell.
60 Abdulrahman et al.
Table 1. Specifications of the 15 GHz DIGI MINI-LINKS.
Links
Location
Hop

Length
(km)

Frequency
Band

(GHz)

Maximu
m
Transmit
Power
(dBm)
BER
(2x2 Mbs)
Receive
Threshold
(dBm)
Antennas for both
Transmit & Receive Side
Size
(m)
Gain
(dBi)
Johor
Bahru
5.83

14.8

+18.0 −84.0 0.6

37.0

Kuala
Lumpur-I
3.96

14.8

+18.0

84.0

0.6

37.0

Taiping 3.48 14.8 +18.0 84.0 0.6 37.0
Butterworh 11.33 14.8 +18.0 84.0 0.6 37.0
Alor Star 4.85 15.3 +18.0 84.0 0.6 37.0
Temerloh 5.36 14.8 +18.0 84.0 0.6 37.0
Kuantan 1.45 14.8 +18.0 84.0 0.6 37.0
10
−6






Table 2. Measured rain attenuation of Johor Bahru MINI-LINK.
Percentage
of time (%)
0.001 0.003 0.01 0.03 0.1 0.3 1.0
Rain
attenuation
(dB)
111.22 79.25 51.90 33.65 19.87 11.72 6.24
Therefore, the conversion factor C
%p
is more conveniently
expressed as the ratio of terrestrial rain attenuation to that of the
satellite’s. Thus dividing Equation (20) by Equation (15) would yield
the following expression:
C
%p
=
A
T%p
A
S%p
(21)
which gives
C
%0.01
=
k
T
L
T
r
d0.01
k
s
L
eff
[R
0.01
]

T
−α
S
)
(22)
According to Equation (22), the conversion factor C
%p
is proportional
to R
0.01
since all other parameters are constant; and its variation over
different rain rates, at 0.01% of the time, is presented in Figure 1.
Progress In Electromagnetics Research B, Vol. 26, 2010 61
It can be clearly seen from Figure 1 that the conversion factor
increases with an increasing rain rate. Also since the latter is a function
of p% of the time, further analyses have therefore revealed that the
former also is functionally related to p%, as presented in Figure 2.
It is glaring that the conversion factor is approximately 2.0159,
at 0.01%, which corresponds to a rain rate of 125 mm/h, as already
indicated in Figures 1 and 2. It is therefore evidently implied that
either the rain rate R
0.01
or the probability level p could be used as
the input in the proposed rain attenuation conversion factor C
%P
.
Figure 1. Variation of conversion factor with respect to rain rate.
Figure 2. Variation of conversion factor with p% of the time in an
average year.
62 Abdulrahman et al.
4.1. Applications of the Proposed Conversion Method
Equations (21) and (22) have been used for converting the predicted
and measured yearly rain attenuation for the 15 GHz DIGI MINI-
LINKS in Malaysia, to be used for MEASAT 2 applications.
The predictions are expected to provide a rough estimate of rain
attenuation to be expected on earth-satellite links in the Malaysian
tropical region. Figure 3 shows the comparison between predicted
satellite attenuation for both measured and predicted terrestrial rain
attenuation.
As clearly indicated in Figure 3, the corresponding satellite
attenuation values could be obtained from the measured terrestrial
rain attenuation, by direct application of the Equations (21) and (22).
For instance, the measured attenuations for the 15 GHz DIGI MINI-
LINK, at 0.01% and 0.1%, are nearly 52 dB and 20 dB, respectively;
whereas the corresponding satellite attenuation values are 25.8 dB and
9.9 dB, respectively.
However, if there are no terrestrial rain attenuation data, then the
predictions of Equation (20) may be used as the input in Equation (21)
for deriving the equivalent satellite values, as shown in Figure 3.
Although the weakness being that the inherent errors associated with
such predictions would naturally be inherited by the transformed
satellite values. According to Figure 3, the predicted values of
terrestrial rain attenuation, at 0.01% and 0.1%, are nearly 54.7 dB
and 21 dB, respectively; while the corresponding satellite attenuation
values are 27.1 dB and 10.4 dB, respectively.
Even though it is noted in both cases that the transformed values
for the satellite link are about 50% of the terrestrial rain attenuation
for all percentages, %p of the time, the terrestrial rain attenuation
predictions are a bit over estimated, compared to the actual measured
Figure 3. Comparison of terrestrial and satellite rain attenuation.
Progress In Electromagnetics Research B, Vol. 26, 2010 63
value. These prediction errors would naturally corrupt the transformed
satellite values. In this paper, the errors are generally insignificant (less
than 5%); and so they are overlooked. For instance, at 0.01% and 0.1%
of the time, the prediction errors in the terrestrial rain attenuation
are 2.7 dB and 1.0 dB, respectively; while the corresponding errors
inherited by the satellite attenuation values are 1.3 dB and 0.5 dB,
respectively.
Table 3 shows the site locations, propagation parameters and the
measured rain attenuation values, at 0.01% and 0.1% of the time, for
the earth-satellite links that were compared with the proposed model
predictions.
The comparison between the model’s estimates of Equations (21)
and (22) and measured Earth-satellite attenuation data (MEASAT
2 & SUPERBIRD-C [17]) are provided in terms of Complementary
Cumulative Distribution Functions (CCDFs) of attenuation as shown
in Figures 4 and 5.
The proposed model predictions appear to be in good agreement
with MEASAT 2 and SUPERBIRD-C at lower rain rates from 1.0% to
0.1%, when the rain attenuation of the terrestrial link is in the range of
6.0 dB to 21 dB. Also, the model predictions appear to be reasonable
at medium rain rates from 0.1% to 0.01%, which correspond to a range
of 21 dB to 54.7 dB of terrestrial attenuation. However, the predictions
generally overestimate the measured Earth-satellite attenuation data
at higher rain rates, from 0.01% to 0.001%.
4.2. Validation of the Model
According to Recommendations ITU-R P.311-13 [20], the rain
attenuation prediction models for Earth-satellite links are determined
for exceeding time percentages in the range 0.001% to 0.1%. Therefore
percentage errors between measured Earth-satellite attenuation data
A
m
(dB) and the model’s predictions A
p
(dB) are calculated for each
Table 3. Characteristics of the Earth-satellite measurement sites.
Earth-Station Altitude
(m)
Frequency
(GHz)
Elevation
(degrees)
Polari
zation

(mm/h)
Rain attenuation
(dB)
0.1 % 0.01 %
MEASAT 2
UTM-Johor, Malaysia
1.45 N, 103.75
0
E

1.0

12.594

70.0

V 125 9.8 25.0
SUPERBIRD-C
USM, Malaysia
51.7 N, 100.4 E

57.0

12.255

40.1

V 130 8.98 23.5
0
0
0
R
0.01
Locations
64 Abdulrahman et al.
exceeding time percentage of interest on the microwave radio link, as
follows:
E
i
=
A
p
i
−A
m
i
A
m
i
∗ 100% (i = 1 to N) (23a)
if: |A
p
i
−A
m
i
| < 1 then E
i
= 0 (23b)
The standard deviation, σ
e
i
of the error distribution can be defined
from:
σ
e
i
=
¸
¸
¸
_
1
N
N

i=1
e
2
i
−(µ
e
i
)
2
(24)
where µ
e
i
is the mean square error for each exceedance time percentage;
and is given by:
µ
e
i
=
1
N
N

i=1
e
i
(25)
The mean square error and standard deviation are then used to
calculate the Root Mean Square, D
e
i
(RMS); which is defined as
follows:
D
e
i
=
_

e
i
)
2
−(σ
e
i
)
2
¸
1/2
(26)
Table 4 compares parameters µ
e
i
, σ
e
i
and D
e
i
for the proposed
prediction model with measured Earth-satellite attenuation data
(MEASAT 2 and SUPERBIRD-C).
According to the evaluation procedures adopted by the Recom-
mendations ITU-R P.311-13, a lower standard deviation σ
e
i
and a lower
Figure 4. Comparison between measured and predicted attenuation
for MEASAT 2-UTM.
Progress In Electromagnetics Research B, Vol. 26, 2010 65
Figure 5. Comparison between measured and predicted attenuation
for SUPERBIRD-C, USM.
Table 4. Percentage errors and RMS comparison.
Parameters

Earth-satellite
links
Time Percentages ( p %)
0.001 0.003 0.01 0.03 0.1 0.3 1.0
Proposed model 18.46 13.16 8.62 5.59 3.30 1.95 1.04
ITU-R model −26.99 −19.23 −12.59 −8.16 −4.82 −2.84 −1.51
Proposed model 1.62 1.15 0.76 0.49 0.29 0.17 0.09
ITU-R model 3.42 2.44 1.60 1.04 0.61 0.36 0.19
Proposed model 18.53 13.20 8.65 5.61 3.31 1.95 1.04
ITU-R model 27.21 19.38 12.69 8.23 4.86 2.86 1.52
µ
σ
D
ei
ei
ei
RMS value D
e
i
for the whole range or for the majority of time per-
centages of interest suggest a high accuracy of the proposed model.
5. CONCLUSIONS
The conversion technique has been developed based on the ITU-R P
618-9, together with a combination of ITU-R P 530-12 and the revised
Moupfouma model, for the satellite and terrestrial rain attenuation
predictions, respectively.
The proposed method has been used for converting the measured
and predicted rain attenuation data exceeded for an average year for
15 GHz DIGI MINI-LINKS in Malaysia; and the results have been
found to be comparable with direct measurements taken in Malaysia
66 Abdulrahman et al.
and similar tropical climates like Singapore and Nigeria [2, 18].
The model is therefore recommended for converting terrestrial rain
attenuation data to be used for earth-satellite applications at Ku-
band in the Tropical Malaysian Peninsula and other climates that have
similar situations.
REFERENCES
1. Freeman, R. L., Radio System Design for Telecommunication, 3rd
edition, A Wiley Interscience Publication, John Wiley & Sons Inc.,
2007.
2. Ojo, J. S., M. O. Ajewole, and S. K. Sarkar, “Rain rate and
rain attenuation prediction for satellite communication in Ku and
Ka bands over nigeria,” Progress In Electromagnetics Research B,
Vol. 5, 207–223, 2008.
3. Moupfouma, F. and L. Martin, “Modelling of the rainfall rate
cumulative distribution for the design of satellite and terrestrial
communication systems,” Int. J. of Satellite Comm., Vol. 13,
No. 2, 105–115, 1995.
4. Ajayi, G. O., “Some aspects of tropical rainfall and their effect
on microwave propagation,” International Journal of Satellite
Communications, Vol. 8, No. 3, 163–172, May–June 1990.
5. UI Islam, M. R., T. B. A. Rahman, S. K. B. A. Rahim, K. F. Al-
tabatbaie, and A. Y. Abdulrahman, “Fade margins prediction for
broadband fixed wireless access (BFWA) from measurements in
tropics,” Progress In Electromagnetics Research C, Vol. 11, 199–
212, 2009.
6. Crane, R. K., “Rain attenuation models: Attenuation by clouds
and rain,” Propagation Handbook for Wireless Communication
System, 225–280, CRC Press, USA, 2003.
7. Ramachandran, V. and V. Kumar, “Invariance of accumulation
time factor of ku-band signals in the tropics,” Journal of
Electromagnetic Waves and Applications, Vol. 19, No. 11, 1501–
1509, 2005.
8. “Propagation data and prediction methods required for the design
of Earth-space telecommunications systems,” Recommendation
ITU-R P.618-9, ITU-R P Sers., August 2007.
9. “Propagation data and prediction methods required for the
design of terrestrial line-of-sight systems,” Recommendation ITU-
R P.530-12, ITU-R P Sers., February 2007.
10. Moupfouma, F., “Electromagnetic waves attenuation due to rain:
A prediction model for terrestrial or L.O.S SHF and EHF radio
Progress In Electromagnetics Research B, Vol. 26, 2010 67
communication,” J. Infrared Milli Terahz Waves, Vol. 30, 622–
632, 2009.
11. Singliar, R. and J. Bito, “Comparison of satellite and terrestrial
rain attenuation statistics based on terrestrial measurement,” ESA
Propagation Workshop, Noordwiik, Netherlands, November 2005.
12. “Rain height model for prediction methods,” Recommendation
ITU-R P.839-3, ITU-R P Sers., Int. Telecomm. Union, Geneva,
2001.
13. Chebil, J. and T. A. Rahman, “Development of 1 min rain rate
contour maps for microwave applications in Malaysia peninsula,”
Electronics Letts., Vol. 35, 1712–1774, 1999.
14. Final Reports on Rain Attenuation Studies for Communication
Systems Operating in Tropical Regions, Wireless Communication
Research Laboratory, Universiti Teknologi Malaysia, October 31,
2000.
15. “Specific attenuation model for rain for use in prediction
methods,” Recommendation ITU-R P.838-3, ITU-R P Sers.,
March 2005.
16. Matricciani, E. and A. Pawlina Botani, “Rain cell size statistics
inferred from long term point rain rate: Model and result,”
Proc. Third Ka Band Utilization Conf., 299–304, Sorrento, Italy,
September 15–18, 1997.
17. Mandeep, J. S. and J. E. Allnutt, “Rain attenuation predictions
at ku-band in south-east Asian countries,” Progress In Electro-
magnetics Research, Vol. 76, 65–74, 2007.
18. Lee, Y. H., J. X. Yeo, and J. T. Ong, “Rain attenuation on satellite
to ground link for beacon signal,” 27th International Symposium
on Space Technology and Science, Singapore, 2009.
19. Singliar, R., J. Bit´o, J. Din, and A. R. Tharek, “Comparison
of predicted attenuation of satellite rain attenuation distribution
in Malaysia and Hungary,” Proceedings of the 16th International
Czech-Slovakia Scientific Conference Radioelektronika, 246–249,
Brastislava, Solak Republic, April 25–26, 2006.
20. “Acquisition, presentation and analysis of data in studies of
tropospheric propagation,” Recommendation ITU-R P.311-13,
ITU-R P Sers., October 2009.
21. Mandeep, J. S., N. Y. Yann, and S. I. S. Hassan, “Case study
of the rain attenuation at Ka band,” Journal of Electromagnetic
Waves and Applications, Vol. 22, 1517–1525, 2008.

no comprehensive research works have previously been . These include the conversion of rain rate integration time from higher time to a shorter one. The results may also be extrapolated to be used in other tropical regions that have similar situation. However. and the frequency scaling techniques for converting the measured rain attenuation at one frequency to the equivalent value at other frequencies for the same site. The rain-attenuation analyses are important for the study of the rain fade characteristics. All these effects must be considered as they are the causes of signal impairment for both terrestrial and Earth-satellite systems. which is useful in the preliminary design of earth-satellite microwave links. This has become imperative because of the fact that these regions do not have sufficient earth-satellite data. depolarization by hydrometeors such as clouds and precipitation [1]. The effect of rainfall is more severe in tropical regions which are characterized by heavy rainfall intensity and presence of large raindrops [2. based on the ITU-R recommendations ITU-R 618 -9 [8]. 2. The aim of this paper is to present a mathematical model for converting the rain-induced attenuation data on the terrestrial links to be used for satellite applications at Ku-band. 5]. consequently.54 Abdulrahman et al. 3]. 7]. and the empirical models which are based on measurement databases from stations in different climatic zones within a given region [6. which is a useful piece of information in the link budget estimation for predicting the expected outage as a result of rain attenuation on a microwave link [4. Long-term rain attenuation data are usually available for the terrestrial links in most tropical regions and so there is the need to convert them for use in satellite applications. Raindrop size distribution changes with geographical location and it can strongly influence rain specific attenuation and. with rainfall being the most crucial especially at frequencies above 10 GHz. total rain attenuation. A HANDFUL OF SOME EXISTING CONVERSION TECHNIQUES AND MODELS FOR RAIN ATTENUATION Various conversion methods have earlier been developed for different applications in rain attenuation predictions for both terrestrial and satellite communication systems. ITU-R 530-12 [9] and the revised Moupfouma model [10]. There are two broad classes of rain attenuation predictions on any microwave link: The analytical models which are based on physical laws governing electromagnetic wave propagation and which attempt to reproduce the actual physical behavior in the attenuation process.

polarization. based on terrestrial rain attenuation measurement. 530-12 and the revised Moupfouma model. a temperate region with rain rates less than 40 mm/h. 19]. AT (t) is the measured rain attenuation time series of the terrestrial link. the proposed technique has been validated using the terrestrial rain attenuation data measured in Malaysian tropical climate. by finding the ratio of the two attenuation values at the same rain rate. who proposed a transformation method for converting terrestrial rain attenuation timeseries into satellite attenuation time-series. 3. 2010 55 done on converting terrestrial rain attenuation data to be used for satellite applications. Calculation of Rain Attenuation for the Satellite Link The step-by-step procedure for calculating rain attenuation cumulative distribution function (CDF) over the satellite link is given below: αs . The model is given by: ks Ls AT (t)(1 + LT /d0 ) αT (1) As (t) = 1 + Ls /d0 kT LT where ks . as high as 120–130 mm/h. Also. αs and αT are ITU-R recommended parameters whose values depend on frequency band. Vol. 618-9 and second is the prediction of terrestrial rain attenuation.Progress In Electromagnetics Research B. 3. The only published research works in the literature are those of Singlair et al.01 ) (2) The major limit of the model being that the terrestrial rain attenuation data were measured in Hungary.1. using a combination of ITU-R P.015R0. it has been considered as the principal input in the new proposed conversion technique. Ls and LT are the slant path and the physical length of the satellite and terrestrial links. respectively. This implies that the model is unarguably unsuitable for tropical climates like Malaysia which experiences heavy rain rates. and d0 is the ITU-R reduction factor given by: d0 = 35e(−0. 26. The third stage involves derivation of the proposed transformation method. [11. first is the calculation of rain attenuation over the satellite link using the ITU-R P. Rain rate is the most crucial factor in any rain attenuation predictions. kT . as indicated in Equation (22). temperature and elevation. DEVELOPMENT OF THE PROPOSED RAIN ATTENUATION CONVERSION TECHNIQUE The development consists of three stages. since both links are sited in the same location. As (t) is the transformed rain attenuation time series of the satellite link. Therefore.

γs 0. Step 5: The specific attenuation. as follows [12]: HR = 0.839-3. both in degrees. in km and θ (degrees) is the elevation angle. below the rain height is calculated as: HR − HS LS = (4) sin θ where HS is the station altitude.36 + h0 (3) where h0 is the 0◦ C isotherm height above mean sea level (km) and its value can be obtained from the isotherm chats of the Recommendation ITU-R P.01 = αM β (6) where α and β are regression coefficients. and is given by [13]: R0. and polarization.2973 (7) We have found that the recommendation ITU-R P.01 (mm/h).01 = ks R 0.01% of time is calculated from: αs γs0. which makes use of long-term mean annual accumulation.837-5 is not suitable for estimating R0. An alternative method is by applying the Chebil model. rain temperature. For instance. exceeded for 0. Step 3: The horizontal projection of the slant path length.838-3 [15].56 Abdulrahman et al.2903 and β = 0. M .01 at UTM.01 (8) The parameters ks and αs depend on frequency.01 in tropical Malaysian climate. above the sea-level. R0.837-5 has predicted a value of 105 mm/h for R0. Their values can be obtained from ITU-R P.01% of an average year may be obtained from one. whereas both direct measurement and Chebil’s approximation of Equation (6) have given a value of approximately 125 mm/h for the same location. Step 2: The slant-path length. the ITU-R P. at the location under study. Step 1: Freezing height during rain. LS (km). raindrop size distribution. . LG is evaluated from: LG = LS cos θ (5) Step 4: The point rain rate. The regression coefficients are defined as: α = 12. 21].01 (dB/km). HR (km) is calculated from the absolute values of station’s latitude and longitude. if it is available. This model was shown to provide good approximation of the measured data. because it substantially underestimates the measured rain rate values [14. for 0.minute integration rain rate data.

rh0.01 = √ LG γ0. 26. Leff (km) is finally obtained by multiplying the horizontally adjusted slant path by the vertical reduction factor of Equation (10).01 .01% of an average year is obtained from: As0. for ρ > θ Lh0. for |ϕ| ≥ 36◦ (13) Step 8: The effective path length through rain.01% of the time is also given by: 1 √ (12) rv0.01 1+ sin θ 31 (1−exp(−θ/[1+σ])) −0.12 ∗ As0.546+0.01% of the time is also calculated as: 1 (9) rh0.01 by using the following: As%p = 0. for 0. p of an average year may be calculated from the value of As0.01 = LG γ0.45 f2 where σ= 36 − |ϕ|.01 / cos θ. rv0. Leff = Lh0. for 0. the horizontally adjusted rainy slant path length is calculated from: LG rh0.01 = γs0.78 − 0.01 1 + 0.01 = (10) (HR − HS )/ sin θ.01 ∗ p(−(0. Vol. for |ϕ| < 36◦ 0.01 Leff (15) Step 10: The predicted attenuation to be exceeded for other percentages. that is.01 (11) Step 7: The vertical reduction factor.01 (14) Step 9: The predicted attenuation exceeded for 0. 2010 57 Step 6: The horizontal path adjustment reduction factor.043∗log 10(p))) (16) .01 rv0.Progress In Electromagnetics Research B.38[1 − exp −2LG ] f Therefore. for ρ ≤ θ where ρ = tan−1 HR − HS LG rh0.01 .

01 . Meteorological .2 with ζ(LT ) = LT for any LT > 7 km (19b) Step 3: The predicted attenuation exceeded for an average year is then obtained from the following: AT 0.01 (17) Again.01 . LT ) = exp −R0. In this paper we have deliberately chosen not to use the ITU-R path reduction factor of Equation (2) for the terrestrial link. and polarization.01 = γT 0. Their values can be obtained from ITU-R P.58 Abdulrahman et al.01%. This is due to its unsuitability for the tropical regions. were collected from seven 15 GHz operational microwave links of DIGI Telecommunications.01 .01 (18) where δ(R0.01 1 + ζ(LT ) ∗ R0.01 LT δ(R0. LT ) = LT ∗ exp −R0.78 (19a) and Leq (R0. LT ) is the reduction factor for 0.01 by using the Equation (16). such that the equivalent propagation path.01 0. parameters kT and αT depend on frequency.01 = kT R0.01% of the time observed on any terrestrial microwave radio path of length. sampled every second.838-3. raindrop size distribution. Calculation of Rain Attenuation for the Terrestrial Link Step 1: The specific attenuation for the terrestrial link is given by: αT γT 0. path reduction factor is calculated using the revised Moupfouma’s model given by [10]: δ(R0. 3. 3. Step 2: For the terrestrial link. LT (km). Data Collection for Terrestrial Links The received relative intermediate frequency (IF) rain attenuation data. rain temperature.01 1+ζ(LT )∗ R0. of an average year may be calculated from the value of AT 0.01 1+ζ(LT ) ∗ R0. Leq is: Leq (R0.2.01 . LT ) (20) The predicted attenuation to be exceeded for other percentages.1.3. p = 0.01 with ζ(LT ) = −100 for any LT ≤ 7 km −R0. as earlier mentioned in step 5 of section 3.01 . LT ) = LT ∗ exp 44.

800 km above the earth. terrestrial microwave transmissions are more susceptible to the effects of rain attenuation because their signal paths are entirely in the troposphere. In contrast. and since rain only forms in the troposphere. under the same conditions of rainfall intensity. The campus is located at Johor. 2010 59 stations were installed at the particular locations of the investigated links. southern part of Malaysia peninsula close to Singapore.5 mm sensitivity. 4.5 mm/min. This fact is justified as follows [15. a signal traveling through a rain cell will experience attenuation during only a small portion of its transmission path. Vol.3 mm. The Malaysian climate is tropical.Progress In Electromagnetics Research B. Skudai campus and the data have been used to investigate the DIGI link located at Johor Bahru. therefore the rain rate is recorded as an integral multiple of 0. 16]: Satellites in geostationary orbit are 35. and falls in the zone P of the ITU-R rainfall rate climatic zone with annual average accumulation as high as 4184. high humidity and heavy rainfall which arises mainly from the maritime exposure of the country. Johor Bahru are shown in Table 2. (UTM). . Four year precipitation data were collected at Universti Teknologi Malaysia. The uniform periodic changes in the wind flow patterns result in seasons’ classifications as: The south-west monsoon. that is November and December. Casella records the total rainfall occurring in each minute without recording non-rainy events. and is characterized by uniform temperature. using Casella rain gauge and thirty months using OSK rain gauge. the north-east monsoon and the two shorter inter-monsoon seasons. while the measured rain attenuation data at UTM. or 30 mm/h. Therefore. MATHEMATICAL DERIVATION OF THE CONVERSION FACTOR First it is logical to assume that the terrestrial rain attenuation is greater than that of the satellite. Both gauges are tipping bucket type and having 0. which extends seven miles above the earth. Malaysia is located in the South-Eastern part of Asia. The specifications for the MINI-LINK are given in Table 1. in order to be able to determine the relationship between the rain attenuation and rain intensity. The heaviest amounts of rainfall are noted to occur in the last two months of the year. 26. and the signal may pass through an entire rain cell. Malaysia. The rain rate data were collected from January 2003 to December 2006. OSK records the actual tipping time up to decimal of seconds.

001 0.01 [R0.45 14.0 − 84.0 37.01% of the time.90 33.6 0.0 − 84.87 11.6 0. the conversion factor C%p is proportional to R0.0 +18.01 = AT %p AS%p (21) kT LT rd0. at 0. is presented in Figure 1.8 15.0 − 84.01 since all other parameters are constant.01 ](αT −αS ) ks Leff (22) According to Equation (22). Measured rain attenuation of Johor Bahru MINI-LINK.0 +18.0 +18.48 4.33 Table 2. and its variation over different rain rates.6 0.8 14.003 0.6 0.65 19.0 − 84.0 +18.8 14. .6 0.72 6.0 111.01 0.83 3.0 37.60 Abdulrahman et al.0 37. the conversion factor C%p is more conveniently expressed as the ratio of terrestrial rain attenuation to that of the satellite’s. Specifications of the 15 GHz DIGI MINI-LINKS.0 − 84.3 14.03 0. Table 1.36 1. Thus dividing Equation (20) by Equation (15) would yield the following expression: C%p = which gives C%0.6 37.0 37.3 1.24 Therefore.25 51. Percentage of time (%) Rain attenuation (dB) 0.8 +18.0 37.96 3.8 14.0 − 84.22 79.0 37.8 14.0 0. −6 Antennas for both 10 BER Transmit & Receive Side Hop Frequency m Links (2x2 Mbs) Band Length Transmit Location Size Gain Receive (GHz) (km) Power (m) (dBi) Threshold (dBm) (dBm) Maximu Johor Bahru Kuala Lumpur-I Taiping Alor Star Temerloh Kuantan 5.0 − 84.0 Butterworh 11.6 0.0 +18.0 +18.1 0.85 5.

It is therefore evidently implied that either the rain rate R0. as presented in Figure 2. Figure 2. Also since the latter is a function of p% of the time.Progress In Electromagnetics Research B. Variation of conversion factor with p% of the time in an average year. It is glaring that the conversion factor is approximately 2. Vol. which corresponds to a rain rate of 125 mm/h. further analyses have therefore revealed that the former also is functionally related to p%.0159. Figure 1. Variation of conversion factor with respect to rain rate. as already indicated in Figures 1 and 2. 2010 61 It can be clearly seen from Figure 1 that the conversion factor increases with an increasing rain rate.01%.01 or the probability level p could be used as the input in the proposed rain attenuation conversion factor C%P . 26. . at 0.

According to Figure 3. Figure 3 shows the comparison between predicted satellite attenuation for both measured and predicted terrestrial rain attenuation. the corresponding satellite attenuation values could be obtained from the measured terrestrial rain attenuation. the measured attenuations for the 15 GHz DIGI MINILINK.01% and 0. .7 dB and 21 dB. if there are no terrestrial rain attenuation data. The predictions are expected to provide a rough estimate of rain attenuation to be expected on earth-satellite links in the Malaysian tropical region.1 dB and 10. 4. Comparison of terrestrial and satellite rain attenuation. as shown in Figure 3.4 dB. However.01% and 0. Applications of the Proposed Conversion Method Equations (21) and (22) have been used for converting the predicted and measured yearly rain attenuation for the 15 GHz DIGI MINILINKS in Malaysia. respectively. respectively. then the predictions of Equation (20) may be used as the input in Equation (21) for deriving the equivalent satellite values. while the corresponding satellite attenuation values are 27. %p of the time. the terrestrial rain attenuation predictions are a bit over estimated. respectively. Even though it is noted in both cases that the transformed values for the satellite link are about 50% of the terrestrial rain attenuation for all percentages. compared to the actual measured Figure 3.62 Abdulrahman et al. the predicted values of terrestrial rain attenuation. As clearly indicated in Figure 3.1%. whereas the corresponding satellite attenuation values are 25. to be used for MEASAT 2 applications. at 0. respectively.8 dB and 9. Although the weakness being that the inherent errors associated with such predictions would naturally be inherited by the transformed satellite values. by direct application of the Equations (21) and (22).1.9 dB. are nearly 54. are nearly 52 dB and 20 dB. For instance.1%. at 0.

respectively.1% of the time.7 dB of terrestrial attenuation. 103. These prediction errors would naturally corrupt the transformed satellite values. the model predictions appear to be reasonable at medium rain rates from 0. Malaysia 1.01 % 9. Characteristics of the Earth-satellite measurement sites.0% to 0.001% to 0.7 dB and 1. at 0.0 12. the predictions generally overestimate the measured Earth-satellite attenuation data at higher rain rates.5 dB. at 0. the errors are generally insignificant (less than 5%).0 57.1 V 130 8. Validation of the Model According to Recommendations ITU-R P. Malaysia 51. 2010 63 value. Table 3 shows the site locations. from 0.594 70.1 % 0.001%. while the corresponding errors inherited by the satellite attenuation values are 1.01% to 0.01% and 0. 26. propagation parameters and the measured rain attenuation values. Earth-Station Locations MEASAT 2 UTM-Johor. 100.4 0 E Altitude Frequency Elevation Polari R 0. In this paper.3 dB and 0. For instance. Vol.75 0 E SUPERBIRD-C USM.Progress In Electromagnetics Research B.5 . when the rain attenuation of the terrestrial link is in the range of 6.01% and 0.1% of the time.0 12. the rain attenuation prediction models for Earth-satellite links are determined for exceeding time percentages in the range 0. 4.1%.2.0 dB to 21 dB. The proposed model predictions appear to be in good agreement with MEASAT 2 and SUPERBIRD-C at lower rain rates from 1.255 40. which correspond to a range of 21 dB to 54.8 25. Also. However.0 V 125 Rain attenuation (dB) 0.01%. the prediction errors in the terrestrial rain attenuation are 2.98 23. and so they are overlooked. The comparison between the model’s estimates of Equations (21) and (22) and measured Earth-satellite attenuation data (MEASAT 2 & SUPERBIRD-C [17]) are provided in terms of Complementary Cumulative Distribution Functions (CCDFs) of attenuation as shown in Figures 4 and 5. Therefore percentage errors between measured Earth-satellite attenuation data Am (dB) and the model’s predictions Ap (dB) are calculated for each Table 3. respectively.1%.45 0 N.7 0 N.01 (m) (GHz) (degrees) zation (mm/h) 1.0 dB. for the earth-satellite links that were compared with the proposed model predictions.1% to 0.311-13 [20].

as follows: Api − Ami ∗ 100% (i = 1 to N ) (23a) Ei = Ami if: |Api − Ami | < 1 then Ei = 0 (23b) The standard deviation. σei of the error distribution can be defined from: σei = 1 N N e2 − (µei )2 i i=1 (24) where µei is the mean square error for each exceedance time percentage. . Dei (RMS). and is given by: N 1 ei (25) µei = N i=1 The mean square error and standard deviation are then used to calculate the Root Mean Square.64 Abdulrahman et al. According to the evaluation procedures adopted by the Recommendations ITU-R P. σei and Dei for the proposed prediction model with measured Earth-satellite attenuation data (MEASAT 2 and SUPERBIRD-C). Comparison between measured and predicted attenuation for MEASAT 2-UTM. which is defined as follows: 1/2 (26) Dei = (µei )2 − (σei )2 Table 4 compares parameters µei . a lower standard deviation σei and a lower Figure 4. exceeding time percentage of interest on the microwave radio link.311-13.

03 5. Comparison between measured and predicted attenuation for SUPERBIRD-C.19 1.61 8.01 8.30 0.04 1.09 0.003 0.16 −4.86 0. 5.04 5.21 19. 2010 65 Figure 5.3 1.59 −8.95 2.23 −12. USM. Table 4.38 12. Vol.23 0.69 RMS value Dei for the whole range or for the majority of time percentages of interest suggest a high accuracy of the proposed model.42 1.Progress In Electromagnetics Research B.001 0.62 3.15 2.44 0.0 1.86 0.29 0.60 8.62 0. and the results have been found to be comparable with direct measurements taken in Malaysia .61 3.49 1. together with a combination of ITU-R P 530-12 and the revised Moupfouma model.36 1.95 1.52 ITU-R model Proposed model 18.59 0.1 3.84 −1. 26.46 13.76 1.16 ITU-R model Proposed model σei Dei −26. CONCLUSIONS The conversion technique has been developed based on the ITU-R P 618-9. Parameters Earth-satellite links µ ei Time Percentages ( p %) 0.51 1.31 4.53 13. The proposed method has been used for converting the measured and predicted rain attenuation data exceeded for an average year for 15 GHz DIGI MINI-LINKS in Malaysia. Percentage errors and RMS comparison.65 0.17 0.82 −2.20 ITU-R model 27. respectively. for the satellite and terrestrial rain attenuation predictions.99 −19.04 Proposed model 18.

530-12. Kumar. 5. 3. Vol. and A.” International Journal of Satellite Communications.” Recommendation ITU-R P. 2008. No. Ojo.” Progress In Electromagnetics Research B.” Progress In Electromagnetics Research C. 1501– 1509. F.. Altabatbaie. S.” Recommendation ITUR P. 199– 212. Ajayi. 9.. 11. 18]. “Fade margins prediction for broadband fixed wireless access (BFWA) from measurements in tropics.O. of Satellite Comm.” Journal of Electromagnetic Waves and Applications. 6.. S. Vol. CRC Press. 2. L. and similar tropical climates like Singapore and Nigeria [2. 2007. J. No. “Electromagnetic waves attenuation due to rain: A prediction model for terrestrial or L.66 Abdulrahman et al. Freeman. M. “Rain rate and rain attenuation prediction for satellite communication in Ku and Ka bands over nigeria. Sarkar. 163–172. 2003. 19.. F. “Propagation data and prediction methods required for the design of Earth-space telecommunications systems. Ramachandran. M. K. USA. Rahman. 11. and S. B. K. 2. 105–115..” Propagation Handbook for Wireless Communication System. 7. “Invariance of accumulation time factor of ku-band signals in the tropics. May–June 1990. 1995. 10.. 2009. Radio System Design for Telecommunication.. ITU-R P Sers. Vol. Moupfouma. Vol. No. Crane. F. “Modelling of the rainfall rate cumulative distribution for the design of satellite and terrestrial communication systems. 8.. T. O. UI Islam. The model is therefore recommended for converting terrestrial rain attenuation data to be used for earth-satellite applications at Kuband in the Tropical Malaysian Peninsula and other climates that have similar situations. B. R. “Propagation data and prediction methods required for the design of terrestrial line-of-sight systems. 2005. A Wiley Interscience Publication. August 2007. 3. J. G. K. R. Vol.S SHF and EHF radio . “Some aspects of tropical rainfall and their effect on microwave propagation. Ajewole. Y.” Int. 13. Abdulrahman. A. 225–280. John Wiley & Sons Inc. R. O. REFERENCES 1.. and L.. Rahim. ITU-R P Sers. 5. “Rain attenuation models: Attenuation by clouds and rain. 8. K. February 2007. 3rd edition. A. and V. Martin. V. 207–223.618-9. Moupfouma. 4.

S..” ESA Propagation Workshop.” Recommendation ITU-R P. 2010 67 11. September 15–18. Singapore.. Yann. H. Vol. “Rain height model for prediction methods. 18. “Comparison of satellite and terrestrial rain attenuation statistics based on terrestrial measurement.” Recommendation ITU-R P. 17. 65–74. 622– 632. Third Ka Band Utilization Conf. and A. Universiti Teknologi Malaysia. J. 26. Vol.” Progress In Electromagnetics Research.838-3. S. Wireless Communication Research Laboratory. Sorrento. Union. Yeo. Vol. Solak Republic. “Rain cell size statistics inferred from long term point rain rate: Model and result.” Recommendation ITU-R P. Geneva. October 31. and J. Vol. Hassan.” Journal of Electromagnetic Waves and Applications.” J. J. Pawlina Botani. J. 1517–1525.” Electronics Letts.311-13. Y. 2001. . 22. E. 2009. X. Netherlands. Infrared Milli Terahz Waves.. 21. Mandeep. 15. Rahman. 2008..Progress In Electromagnetics Research B. “Rain attenuation predictions at ku-band in south-east Asian countries. 14. April 25–26. and J. presentation and analysis of data in studies of tropospheric propagation. 12.839-3. “Acquisition. 19. Allnutt. and T. communication. 246–249. Tharek.. 30. T. “Rain attenuation on satellite to ground link for beacon signal.” 27th International Symposium on Space Technology and Science.. 76. 20. 1997. 13. R. November 2005. Final Reports on Rain Attenuation Studies for Communication Systems Operating in Tropical Regions. J. 2007. Ong. and J. and A. 1712–1774. 2009. N. ITU-R P Sers. I. Matricciani. ITU-R P Sers. “Comparison o of predicted attenuation of satellite rain attenuation distribution in Malaysia and Hungary. 16. 35. October 2009. Brastislava. R. March 2005. Mandeep.. Lee. Italy. Y. ITU-R P Sers. 1999. E. Chebil. Singliar. “Development of 1 min rain rate contour maps for microwave applications in Malaysia peninsula. 299–304. Noordwiik.” Proceedings of the 16th International Czech-Slovakia Scientific Conference Radioelektronika. R. Din. “Specific attenuation model for rain for use in prediction methods. 2000. J. Telecomm. Singliar.” Proc. Int. “Case study of the rain attenuation at Ka band. Bito. S. J. A. and S.. Bit´. 2006. Vol.