Abstract:
Tropospheric amplitude scintillation has been reported as a significant cause of
degradation to satellite communication systems operating at Ku-Band frequencies using
small aperture antennas. Due to the dependence of amplitude scintillation on
meteorological and link factors, it experiences significant spatio-temporal variation
which must be statistically characterised for effective satellite link budgeting on any
Earth-space path. A detailed experimental study of tropospheric amplitude scintillation
on the Earth-space path over Akure, southwest Nigeria (7.17oN, 5.18oE, 358 m) was
carried out using radio beacon signals from EutelSat W4/W7 satellite with concurrent
measurement of meteorological parameters of temperature, pressure, humidity and
rainfall rate. Beacon measurement was performed using spectrum analyser (Tektronix
Y400 NetTek Analyzer) at 1-second integration time on the satellite downlink
frequency of 12.254 GHz and path elevation 036oE; while meteorological parameters
were measured at 1-minute integration time using Vantage Vue weather station
equipped with integrated sensor suite (ISS). Extensive analysis involving time series,
power spectral density, peak-to-peak excursion, cumulative distribution and probability
density functions of scintillation amplitude and intensity have been performed on
diurnal, seasonal and annual basis. Also, the dependence of scintillation on weather
parameters was investigated and a semi-empirical model developed for scintillation
analysis. Performance evaluation of the model was carried out through comparison with
existing scintillation models and statistical distribution functions. Sub-annual analysis
of scintillation amplitude and intensity was equally examined for worst month
characterization. Results show that a cut-off frequency of 0.27 Hz performed optimally
for scintillation extraction, and the existence of strong and weak scintillation regimes
were observed during both dry and rainy seasons. Peak-to-peak scintillation amplitude exceeded 1.35 dB during at least 1% of one-minute intervals in each year, while
variation of hourly scintillation intensity was well approximated by a lognormal
distribution, although Gamma and Generalized Extreme Value (GEV) distributions
were well followed also in most months. Developed location-based prediction models
for scintillation intensity, enhancement and fades performed excellently well with R2
value of at least 0.9. Although the developed scintillation models mirror ITU-R and
Karasawa models to a good extent; both and other temperate region – developed models
under-predicted scintillation phenomenon over this location, which is an indication that
scintillation effect is more severe in the tropics.