International Journal

on Marine Navigation

and Safety of Sea Transportation

Volume 5

Number 4

December 2011

441

1 INTRODUCTION

The reflected transmission signal can interfere with

the direct signal from the satellite to produce unac-

ceptable levels of signal degradation. In addition to

fading, signal degradations can include intersymbol

interference, arising from delayed replicas. The im-

pact of the impairments depends on the specific ap-

plication, namely in the case of typical Land MSC

(LMSC) links, all measurements and theoretical

analysis indicate that the specular reflection compo-

nent is usually negligible for path elevation angles

above 20o.

Moreover, for handheld terminals, specular re-

flections may be important as the low antenna di-

rectivity increases the potential for significant specu-

lar reflection effects. For Maritime MSC (MMSC),

LMSC and Aeronautical MSC (AMSC) system

links, design reflection multipath fading, in combi-

nation with possible shadowing and blockage of the

direct signal from the satellite, is generally the dom-

inant system impairment.

2 REFLECTION FROM THE EARTH’S

SURFACE

Prediction of the propagation impairments caused by

reflections from the Earth’s surface and from differ-

ent objects (buildings, hills, vegetation) on the sur-

face is difficult because the possible impairment

scenarios are quite numerous, complex and often

cannot be easily quantified. For example, the degree

of shadowing in LMSC satellite links frequently

cannot be precisely specified.

Therefore, impairment prediction models for

some complicated situations, especially for LMSC

links, tend to be primarily empirical, while more an-

alytical models, such as those used to predict sea re-

flection fading, have restricted regions of applicabil-

ity. Nevertheless, the basic features of surface

reflections and the resultant effects on propagating

signals can be understood in terms of the general

theory of surface reflections, as summarized in the

following classification:

1 Specular Reflection from a Plane Earth – Here,

this effect is less than or equal to the coefficient

for horizontal polarization. Thus the polarization

of the reflected waves will be different from the

polarization of the incident wave if the incident

Surface Reflection and Local Environmental

Effects in Maritime and other Mobile Satellite

Communications

S. D. Ilcev

Durban University of Technology (DUT), Durban, South Africa

ABSTRACT: This paper introduces the effects of surface reflections and local environmental as very im-

portant particulars for mobile and especially for Maritime Satellite Communications (MSC), because such

factors generally tend to impair the performance of satellite communications links, although signal enhance-

ments are also occasionally observed. Local environmental effects include shadowing and blockage from ob-

jects and vegetation near the Ship Earth Station (SES) and other mobile terminals. The advantages and disad-

vantages of those effects are discussed, the areas of surface reflection were examined and

the further

investigations of local environmental are provided. At this point, surface reflections are generated either in the

immediate vicinity of the SES terminals or from distant reflectors, such as mountains and large industrial in-

frastructures. Specific issues related to these challenges are concluded and a set of solutions is proposed to

maximize the availability of satellite communication capacity to the mobile user applications. The specific ef-

fects on propagation in the mobile environments are examined and explained important characteristics of the

Interference from Adjacent Satellite Systems, Specific Local Environmental Influence in MSC, different

Noise Contribution of Local Ships’ Environment, Blockages Caused by Ship Superstructures and Motion of

Ship’s Antenna.

442

polarization is not purely horizontal or purely ver-

tical. For example, a circularly polarized incident

wave becomes elliptically polarized after reflec-

tion.

2 Specular Reflection from a Smooth Spherical

Earth – Here, the incident grazing angle is equal

to the angle of reflection. The amplitude of the re-

flected signal is equal to the amplitude of the in-

cident signal multiplied by the modules of the re-

flection coefficient.

3 Divergence Factor – When rays are secularly re-

flected from a spherical surface, there is an effec-

tive reduction in the reflection coefficient, which

is a geometrical effect arising from the diver-

gence of the rays.

4 Reflection from Rough Surface – In many practi-

cal cases, the surface of the Earth is not smooth.

Namely, when the surface is rough, the reflected

signal has two components: one is a specular

component, which is coherent with the incident

signal, while the other is a diffuse component,

which fluctuates in amplitude and phase with a

Rayleigh distribution.

5 Total Reflected Field – The total field above a re-

flecting surface is a result of the direct field, the

coherent specular component and the random dif-

fuse component.

6 Reflection Multipath – Owing to the existence of

surface reflection phenomena signals may arrive

at a receiver from multiple apparent sources.

Thus, the combination of the direct signal (line-

of-sight) with specular and diffusely reflected

waves causes signal fading at the receiver. The

resultant multipath fading, in combination with

varying levels of shadowing and blockage of the

line-of-sight components, can cause the received

signal power to fade severely and rapidly for

MES and is really the dominant impairment in the

Global Mobile Satellite Communications

(GMSC) service.

3 FADING IN MMSC AND AMSC SYSTEMS

DUE TO SEA SURFACE REFLECTION

Multipath fading due to sea reflection is caused by

interference between direct and reflected radio-

waves. The reflected radiowaves are composed of

coherent and incoherent components, namely specu-

lar and diffuse reflections, respectively, that fluctu-

ate with time due to the motion of sea waves. The

coherent component is predominant under calm sea

conditions and at low elevation angles, whereas the

incoherent becomes significant in rough sea condi-

tions. If the intensity of the coherent component and

the variance of the incoherent component are both

known, the cumulative time distribution of the signal

intensity can be determined by statistical considera-

tion.

In any event, a prediction model for multipath

fading due to sea reflection, however, was first de-

veloped for MMSC systems at a frequency near 1.5

GHz. Although the mechanism of sea reflection is

common for MMSC and AMSC systems, only with

the difference that fading characteristics for AMSC

are expected to differ from those for MMSC, this is

because the speed and altitude of aircraft so much

greater than those of ships. At this point, the effects

of refractions and scattering by the sea surface be-

come quite severe in case of MMSC and AMSC,

particularly where antennas with wide beam widths

are used.

The most common parameter used to describe sea

condition is the significant wave height (H), defined

as the average value of the peak-to-trough heights of

the highest one-third of all waves. Empirically, H is

related to the r.m.s. height (h

) by:

H = 4h

Hence, at 1.5 GHz the smaller-scale waves can be

neglected and the r.m.s. value of the sea surface

slopes appear to fall between 0.04 to 0.07 in the case

of wave heights less than 4 m.

Thus, with diminishing satellite elevation angle,

the propagation path increases, causing a decrease of

signal power at the Rx side. The noise level is ini-

tially constant; however, upon reaching some critical

value of the elevation angle, sea-reflection signals

appear at the Rx input, which begins to affect the

C/N value. To include the effect of multipath inter-

ference caused by sea-refracted signals, the recep-

tion quality would be more properly described by

the C/N plus M, where M is an interfering sea-

reflected signal acting as a disturbance. Thus, sea-

reflected signals differ in structure and can be divid-

ed into two categories:

1 Radio signals with the rapid continuous fluctua-

tions of amplitudes and phases and with a possi-

ble frequency shift due to the motion of small

portions of the specular cross-section relative to

the source of signals (noise or diffused compo-

nents).

2 Radiowaves with relatively slowly changing

phase close to the phase of the basic signal and

with an amplitude correlating with that of the

basic signal (specular component).

Consequently, within the overall specular cross-

section, an angle of arrival reflected radio signals

relative to the horizontal plane may be regarded as

constant and can be described by the following ex-

pression:

α = 90

– γ

where α = angle of radio signals arrival in accord-

ance with Figure 1. and γ = reflection angle. The

modulus of sea reflection factor for L-band signals is

443

within 0.8 and 0.9, which means that the amplitude

of the specular reflected signal is nearly the same as

that of the direct signal. As measurements have

shown, the noise component depends only upon an

elevation angle and a wave height. Decreasing the

elevation angle and increasing the wave height result

in an increase in the total amplitude of the noise,

which includes the noise component. At elevation

angles below 5

, the amplitude component reaches a

peak value and is no longer affected by the wave

height. Now an increase of the wave height causes

primarily more frequent variations in the noise com-

ponent. The corresponding deviation of C.N meas-

ured in 1 kHz bandwidth amounts from 4.5 to 5 dB.

Figure 1. Geometry of Sea Reflection of Satellite Radio Signals

Courtesy of Book: “Global Mobile Satellite Communications”

by D.S. Ilcev

The specular component that appears at the Rx

input together with the direct signal causes fading in

the direct signal due to both the minor difference be-

tween their phases and the slow change of the pa-

rameters of the reflected signals. The ratio of the di-

rect to specular reflected signal can be described as:

C/M = (C + G

) – [C – G

(α

+ ε)

]

where C = direct signal; M = specular reflected sig-

nal power from the sea; G

= maximum gain of the

receive SES antenna pointing towards the satellite; ε

= elevation angle and α, as is shown in Figure 1. In

addition, keeping accuracy sufficient for practical

purposes, the previous relation gives:

C/M = β

C/N

+ [G

– G

(α

+ ε)

]

where β

C/N

= deviation of C/N ratio. With decreasing

elevation angle, the C/M diminishes monotonically,

except for the elevation angle range of 5

to 8

, with-

in which a rise in C/N is observed. This is obviously

due to the fact that at the said angles the difference

in path between the direct and the specular signal

becomes negligible, so that conditions appears close

to the summation of the similar signals at the receiv-

er input. An increase of the C/N plus M ratio is ob-

served simultaneously due to reaching a peak value

of amplitude in the noise (diffused) component. In

fact, experimental measurements show that as the

elevation angle decreases from 10

to 1

, the mean

C/N plus M diminishes from 22–24 dB to 17–18 dB,

with the deviation increasing from 1.5–2 dB to 4.5–5

dB.

4 MULTIPATH FADING CALCULATION

MODEL FOR REFLECTION FROM THE SEA

The amplitude of the resultant signal at the SES ter-

minal, being the sum of the direct wave component,

the coherent and the incoherent reflection compo-

nents, has a Nakagami-Rice distribution (see ITU R

P.1057). The cumulative distribution of fading de-

pends on the coherent-to-incoherent signal intensi-

ties. For example, in the case of rough sea conditions

at 1.5 GHz, the coherent reflection from the sea is

virtually non-existent and the coherent signal is

composed only of the direct component. Therefore,

the fading is determined by the Carrier-to-Multipath

ratio (C/M), i.e., the power ratio of the direct signal

and multipath component caused by incoherent re-

flection. The maximum fade depth (Φ

max

) occurs

when the coherent multipath signal is in anti-phase

with the direct signal, given by:

max

= – 20 log (1 – A

) [dB]

where A

= amplitude of the coherently reflected

component. The value decreases rapidly with in-

creasing wave height, elevation angle and RF. In

practice, due to the vertical motion of the ship an-

tenna relative to average sea surface height, the max-

imum fade value will seldom occur. By adding Φ

max

and Φ

(p) as signal fade due to the incoherent com-

ponent in the function of time percentage (p), a prac-

tical estimate of the combined fading effects of the

coherent and incoherent multipath signal for sea

conditions is obtained:

Figure 2. Estimates of Coherent Reflection and Multipath

Power

Courtesy of Book: “Mobile Antenna Systems Handbook” by

K. Fujimoto and J.R. James

= Φ

max

+ Φ

(p)

444

The maximum fade value due to the coherent

component will not occur constantly because of the

vertical motion of the ship antenna relative to aver-

age sea surface height; therefore, the estimate using

this equation seems to give the worst-case value. In

practice, for low elevation angles (les than 10

) at

around L-band frequencies, the maximum fading oc-

curs when the significant sea wave height is between

1.5 and 3 m, where the coherent reflected compo-

nent is negligible. Accordingly, the dependence of

fading depth on wave height in this range is relative-

ly small.

The amplitude level of the coherent component

decreases rapidly with increasing sea wave height,

elevation angle and frequency. Figure 2. (A) illus-

trates the relationship between coherent reflection

and significant wave height. Namely, estimates of

amplitude of the coherent component for an omnidi-

rectional antenna as a function of a significant wave

height for low elevation angles are illustrated; the

frequency is 1.5 GHz and polarization is circular.

Thus, the incoherent component is random in both

amplitude and phase, since it originates from a large

number of reflecting facets on the sea’s waves. The

amplitude of this component follows Rayleigh dis-

tribution and the phase has a uniform distribution.

Since the theoretical model concerning the inco-

herent components is not suitable for engineering

computations using a small calculator, simpler pre-

diction models are useful for the approximate calcu-

lation of fading. Such simple methods for predicting

multipath power or fading depth have been recently

developed. Thus, in Figure 2. (B) is presented the re-

lationship between multipath power and elevation

angle for different antenna gains. Although fading

depth depends slightly on sea surface conditions,

even if the incoherent is dominant, the simple model

is useful for a rough estimate of fading depth.

Fading depth, which is a scale of intensity of fad-

ing, is usually defined by the difference in decibels

between the direct wave signal level and the signal

level for 99% of the time. The fading depth can be

approximated by a 50% to 99% value for fading

where the incoherent component is fully developed.

Large fading depths usually appear in rough sea

conditions, where the incoherent component is dom-

inant. Thus, Figure 3. (A) shows the fading depth es-

timated by the simple method for antenna not ex-

ceeding for 99% of the time and the corresponding

C/M ratio for circular polarization at 1.5 GHz band

under the condition of significant wave heights from

1.5 to 3 m. The antenna gains of 24, 20, 15 and 8 dB

are functions of elevation angle with a fully-

developed incoherent component. The calculation is

based on the theoretical method, where the shaded

area covers the practical range of the sea wave slope,

which depends on fading depth in rough sea condi-

tions. Values estimated by this simple method give

the mean values of those given in Figure 3. (B).

On the other hand, as the theoretical model is not

suitable for engineering computations using a small

calculator, these simple prediction models are really

useful for the approximate calculation of fading or

interference. Such simple methods for predicting

multipath power or fading depth have been devel-

oped by Sandrin and Fang [1986] and by Karasawa

and Shiokawa [1988] for MMSS and Karasawa

[1990] for AMSS.

Furthermore, the frequency spectral bandwidth of

temporal amplitude variations enlarges with increas-

ing wave height and elevation angle. Figure 3. (B)

shows the probable range of –10 dB spectral band-

width (which is defined by the frequency corre-

sponding to the spectral power density of –10 dB

relative to the flat portion of power spectrum) of L-

band multipath fading obtained by the theoretical

fading model as a function of the elevation angle

under the usual conditions of MMSC; namely, sig-

nificant wave height of 1 m to 5 m, ship speed of 0

to 20 knots and rolling conditions of 0 to 30

Figure 3. Estimates of Fading Depth and Spectral Bandwidth

Courtesy of Book: “Mobile Antenna Systems Handbook” by

K. Fujimoto and J.R. James

5 OTHER ESTIMATIONS OF FADING FOR

MMSC AND AMSC SYSTEMS

The error pattern in digital transmission systems af-

fected by multipath fading is usually of the burst

type. Accordingly, a firm understanding of the fade

duration statistics of burst type fading is required.

Mean value of fade duration (Φ

) and fade occur-

rence interval (Φ

) for a given threshold level as a

function of time percentage, can be estimated from

the fading spectrum. A simple method for predicting

the mean value from the –10 dB spectral bandwidth

is available as a theoretical fading model.

Predicted values of (Φ

) and (Φ

) for 99% of the

time at an elevation angle from 5 to 10

are 0.05 to

0.4 sec for (Φ

) and 5 to 40 sec for (Φ

). The proba-

bility density function of (Φ

) and (Φ

) at any per-

centages ranging from 50% to 99% approximates an

exponential distribution.

445

1 Simple Prediction Method of Fading Depth – Ac-

cording to theoretical analysis and experimental

results made by the mentioned researchers in Ja-

pan, the lowest elevation angle Earth-to-space

path at 1.5 GHz RF band satisfies the energy con-

servation law: [Power of coherent component] +

[Average power of incoherent component] ~

Constant. If this expression is satisfied, the max-

imum incoherent power can be estimated easily

by calculating the coherent power at u = 0. Oth-

erwise, for a more accurate estimation, small

modifications of some parameter dependencies

are necessary. The modified procedure has been

adopted in P.680 for MMSC and P.682 ITU-R

Recommendations for AMSC. Figure 4. (A)

shows a scattergram of measured and predicted

fading depths (i.e., fade for 99% of the time rela-

tive to that for 50%) in the case of MMSC sys-

tems between measured data and predicted values

derived from the simple calculation method with

the same conditions. In this figure, Φ

dp2

are values

from the method set out in ITU-R P.680, while

dp1

are those from an alternative procedure of

the prediction method for scattering angles. It is

evident that the values given by these methods

agree well with the experimental values although

the methods are rather approximate.

Figure 4. Scattergram and Altitude Dependence of Fading

Depth Courtesy of Handbook: “Radiowave Propagation Infor-

mation for Predictions for Earth-to-Space Path Communica-

tions” by ITU

Figure 5. Typical LMSC Propagation Environment

Courtesy of Book: “Mobile Satellite Communications” by S.

Ohmori and other

In Figure 4. (B) is shown the altitude dependence

of signal fade depth not exceeded for 99% of the

time vs. antenna height on board ships or aircraft.

This experiment was obtained from measure-

ments with a helicopter together with the calcu-

lated values from the simple estimation method of

the solid line and the theoretical model of the

shaded region in the figure. From the figure it can

be seen that the simple prediction method agrees

well with both the theoretical model and meas-

ured data even in the case of the AMSC system.

2 Fading Spectrum – In system design, particularly

for digital transmission systems, it is important

not only to estimate the fading depth but also to

know the properties of temporal variation, such as

the frequency power spectrum. For MMSC sys-

tems, theoretical analyses were carried out in Ja-

pan and all parameters affecting the spectrum

such as wave height, wave direction, ship’s direc-

tion and velocity, path elevation angle and anten-

na height variations due to ship’s motion (rolling

and pitching) were taken into account. In general,

spectrum bandwidth is broader with increasing

wave height, elevation angle, ship velocity and

the relative motion of the ship borne antenna. The

dependence of the spectral shape on antenna po-

larization and gain is usually very small. Moreo-

ver, since the speed of aircraft is significantly

higher than that of ships, the fluctuation speed of

multipath fading in AMSC is much faster than

that in MMSC, depending on the flight elevation

angle measured from the horizontal plane. The

calculated –10 dB spectral bandwidth is between

20 and 200 Hz for elevation angles of 5

to 20

for flight elevation angles 0

to 5

at a speed of

1,000 km/h.

6 FADING IN LMSC SYSTEM DUE TO SIGNAL

BLOCKAGE AND SHADOWING

Recently, in the USA, Canada, Australia, Japan and

Europe, domestic LMSC and GMPSC services have

started. The main purpose of these systems is to ex-

tend MSC voice services to rural/remote areas where

terrestrial/cellular services are not provided. In a

typical urban environment, line-of-sight for cellular

systems sometimes is not available due to blockage

by buildings and other structures, when a MES can

receive many waves reflected from these structures

and conduct communication link using these signals.

446

Figure 6. Loss due to Knife-Edge Diffraction

Courtesy of Handbook: “Radiowave Propagation Information

for Predictions for Earth-to-Space Path Communications” by

ITU

Figure 7. Geometry of Knife-edge Diffraction Phenomena

Courtesy of Handbook: “Radiowave Propagation Information

for Predictions for Earth-to-Space Path Communications” by

ITU

The LMSC system can be expected to use the di-

rect line-of-sight signal from a satellite because of

the high elevation angles. When the line-of-sight is

blocked by any obstacle, MSC is not available, but

using path diversity the link is available because of

overlapping of two or more signal from adjacent sat-

ellites. At present, path diversity from separate satel-

lites is rarely used for GEO MSS, but Non-GEO

MSS have an inherent capability to exploit diversity,

because the number of satellites is large providing

path diversity. At this point, Figure 5. presents a

typical propagation environment for scattering, mul-

tipath fading, shadowing, diffusion, ect.

Therefore, to design LMSC system, one needs in-

formation about the propagation statistics of multi-

path fading and shadowing. A vehicle runs at a dis-

tance of 5 to 20 m from roadside obstacles using an

omnidirectional antenna, which has azimuthally uni-

form gain but elevation directivity, or a medium or

high-gain antenna with automatic tracking capabil-

ity. Thus, signal blockage and shadowing effects oc-

cur when an obstacle, such as roadside trees, over-

passes, bridges, tunnels, utility poles, high buildings,

hills or mountains, impedes visibility to the focus of

satellite. This results in the attenuation of the re-

ceived signal to such an extent that transmissions

meeting a certain quality of service may not be pos-

sible. At any rate, in the shadowing environments

the presence of the trees will result in the random at-

tenuation of the strength of the direct path signal.

Hence, the depth of the fade is dependent on a num-

ber of parameters including tree type, height, as well

as season due to the leaf density on the trees.

Whether a VES is transmitting on the left or right-

hand side of the road could also have a bearing on

the depth of the fade, due to the line-of-sight path

length variation through the tree canopy being dif-

ferent for each side of the road. In fact, fades of up

to 20 dB at the L-band may be presented due to

shadowing caused by roadside trees. This shadowing

by roadside trees cannot occur on modern highways

because they are free of trees, only sometimes can

shadowing appear by tunnels, very big construc-

tions, bridges and mountains or hills in narrow pas-

sages.

1 Tree Shadowing – Attenuation due to trees near-

by the roads arises from absorption by leaves and

blockage by trunks and branches. Absorption by

leaves is a function of the type and size of leaves

and the water content therein. Blockage due to

trunks is primarily a function of their size. In ad-

dition to attenuation of the direct signal, trees also

cause an incoherent component due to signals re-

flected and diffracted off the tree surfaces.

The overall attenuation from different types of

fully foliated trees varies from 10.6 to 14.3 dB

and the attenuation coefficient is from 1.3 to 1.8

dB/m. Measurements were conducted with MES

Rx in a rural environment. Based on these aver-

age values, a frequency scaling law for the atten-

uation coefficient has been derived [Goldhirsh

and Vogel] by:

= a

√f

[dB/m]

where a

and a

= attenuation coefficients at fre-

quencies f

and f

[GHz], respectively. Hence, the

range of variation for a

at 1.5 GHz is from 0.5 to

1.7 dB/m. Moreover, trees without foliage attenu-

ate less and the reduction in attenuation appears

to be proportional to the total attenuation experi-

enced when the tree is fully foliated. The received

strength of the direct signal behind a tree will de-

pend on the orientation of the signal path with re-

spect to the tree. The amount of absorbing matter

lying along the path will determine the degree of

attenuation and hence, on average, the length of

the signal path through the tree shield can be con-

sidered a major factor in determining the signal

level. The path length is a function of the eleva-

tion angle and the distance between the receiver

and the tree. At this point, the average attenuation

behind an isolated tree can be estimated as the

product of the attenuation coefficient and the path

length through the tree. The path length through

the tree canopy will depend on its shape and the

orientation of the signal path within the canopy.

Depending on the type being considered, the tree

canopy may be modeled as any one of the shapes.