A three-site rain-fade space-diversity measurements experiment at 20 GHz has been in near continuous operation since September 1, 1994. Two receiver sites (at the Applied Phys. Laboratory ~APL\ of The Johns Hopkins University and COMSAT Laboratories) are located in Central Maryland, and the other is located in Virginia. The APL and COMSAT locations are separated by 33 km. The Virginia site is separated from APL by 44.5 km and from COMSAT by 30.5 km. Receivers at each of the sites measure the 20-GHz CW signal level from the radiating beacon onboard the geostationary Advanced Communications Technology Satellite (ACTS). Because of the lateral variability of rain, the likelihood is diminished that intense rain cells will simultaneously intersect all three Earth-satellite paths. There is also a reduced likelihood that intense rain will simultaneously intersect two Earth-satellite paths associated with different two-site scenarios. Hence, a substantially smaller rain-fade margin is required when the sites operate in a diversity mode (e.g., connected together such that the largest signal is used) vis-a-vis uncoupled single terminal operation. In this effort, we examine the efficacy of employing three-site and two-site space-diversity systems to reduce the required fade margin. Single- and joint-terminal rain-fade distributions are calculated for both the three-site and different combinations of two-site scenarios. Diversity gains, which give a measure of the reduced fade margin, are determined from these distributions. The distributions and diversity gains are also individually determined for two six-month periods of the year for which precipitation is predominantly convective or stratiform, respectively. Comparisons between measured diversity gains and those estimated using the model of the radiocommunications sector of the International Telecommunications Union are made.

Cumulative fade distributions were derived from measured transmissions at 20 GHz emanating from the geostationary Advanced Communications Technology Satellite (ACTS) and at 12 GHz from the television broadcasting digital satellite system (DSS) over the one-year period September 1, 1995-August 31, 1996. The transmissions were acquired at two collocated receivers at the Applied Physics Laboratory of The Johns Hopkins University (central Maryland). Since both geostationary satellites are positioned within 1/spl deg/ from one another, the geometric pointing parameters at the receiver locations are approximately coincident (e.g., 38/spl deg/ elevation angle). The 20-GHz fades were noted to be two-four times larger than those measured at 12 GHz. Two frequency scaling techniques were employed for estimating the distribution at one frequency given a measurement at the other. The methods pertained to the frequency scaling formulation of the radiocommunications sector of the International Telecommunications Union and the ratio of attenuations in terms of an equal probability rain rate along an effective path. The latter method gave agreement to within 1 dB when adjusted for antenna-wetting signal degradation.

We characterize multipath fading at low elevation angles for unshadowed line-of-sight propagation over a land-mobile satellite link. Equipment aboard a van recorded a CW signal from INMARSAT's geostationary satellite MARECS-B2 while driving in the states of Utah, Nevada, Washington, and Oregon. Elevation angles ranged from 7/spl deg/ to 14/spl deg/. The van carried a tracking helix antenna with beamwidths in the principal planes of approximately /spl plusmn/18/spl deg/. Multipath fading was observed in flat and hilly terrain and is compared with an example of fading due to tree shadowing. The measurements demonstrate that multipath effects at low elevation angles may result in fades exceeding 7 dB for approximately 1% of the driving distance for a worst case scenario corresponding to driving in rolling, hilly terrain where the satellite is in front or behind of the vehicle. Fading in such a case is typically dominated by a single multipath reflection from a terrain feature. On the other hand, tree shadowing at low elevation (7/spl deg/) may result in fades which exceed 25 dB at the 1% level.>

Employing a multi-year radar database of precipitation in a mid-Atlantic coast region, the authors have characterized rain cell size statistics. The measurements were made with a high resolution, high power radar in which equicircle diameter contours of resolutions of 1 km and greater were identified out to ranges of 100 km from the radar. The rain cell was constructed by its core intensity isopleth and a family of enveloping contours having defined rain rate intervals in the regions between them. The isopleth statistics were extracted from a database of cores with rain level number 7 (18-24 mm/h) and greater.>

Transmissions from the Japanese ETS-V geostationary satellite were measured at L-band (1.5 GHz) in a vehicle driving on roads of south-eastern Australia. The measurements were part of a program designed to characterize propagation effects due to roadside trees and terrain for mobile-satellite-service (MSS). It is shown that the cumulative distributions of fade and nonfade durations follow a lognormal and power law, respectively. At 1% probability, fades last 2-8 m, and nonfades 10-100 m, depending on the degree of shadowing. Phase fluctuations are generally small, allowing the channel characteristics to be estimated from levels only.>