SUMMARYThe prediction of satellite link attenuation is generally based on the point rainfall rate for 0.01 per cent of the year. The physical quantity used to determine attenuation is rain rate, whereas the 0.01 per cent point is a more complex parameter, not obviously related to the dynamics of rain structure. The dynamics of the horizontal and vertical structure of rain are directly related to rain rate, rainfall volume, local geology and climate. If attenuation is to be predicted from rain rate it would be better to use a single parameter that remains substantially the same in most environments. This is not the case with the 0.01 per cent point. It is argued in this paper that a breakpoint in the rain-rate exceedances occurs close to 105 mm/h. Rain can be divided at the breakpoint into two broad classes: below this point it is more uniformly distributed over the area of rainfall; above, it has an increasingly more complex horizontal varaiability, with intense rain columns embedded in a background of less intense rain. By replacing the climatic zone dependent 0.01 per cent point rain rate with the "xed rain rate of 105 mm/h, a model for attenuation calculation can be derived that relates the mechanics of the physical processes to rain rate and elevation angle. The model has been developed and tested against the ITU-R model using the full measured rain-rate exceedances for each site in the ITU-R data bank. Attenuation exceedances have been shown to depend on the shape of the rain-rate exceedance curve, whereas the ITU-R model generates the same curve shape for all sites, because it uses only one rain rate, R . The new model performs at least as well as the ITU-R model for temperate climates and considerably better for tropical climates. For sites with no measured rain rate it is recommended that a generalized rain-rate exceedance curve be used, especially for tropical regions.

SUMMARYThe ITALSAT satellite experiment started in January 1991 and ended in January 2001, permitting an extensive program of propagation measurements at 18.7, 39.6 and 49.5 GHz. In these frequency bands, upand down-links experience severe signal attenuation due to meteorological effects such as those due to gas (oxygen and water vapour), clouds, turbulence and, especially, rain. The propagation measurement campaigns aim mainly at assessing and at modelling the appropriate fade margin compensating propagation attenuation in the design of satellite communication systems. This margin depends significantly on the season and on the time of the day, due to variations of meteorological conditions. This paper reports the results obtained from copolar signal measurements carried out at the Earth station of Spino d'Adda, near Milano (North Italy), at the three frequencies during 7 years (from 1994 to 2000). The measured cumulative distribution functions of total attenuation are compared to ITU-R models' prediction. Moreover the statistics conditioned to single months of the year, seasons and 4 h contiguous periods of the day are also shown. These statistics can be useful for telecommunication systems whose service quality and design must be matched to the season of the year or the time of the day.

No abstract

No abstract

SUMMARYThe utilization of high frequencies, such as Ka-band and beyond, necessary to avoid the highly congested lower satellite frequencies and to get larger bandwidth availability is considered for many developing satellite systems.The new satellite low-margin systems in Ka-band will need to be designed using fade countermeasures to counteract rain attenuation. One of these techniques foresees the possibility of switching the communication link among different Earth stations spread on a very large territory to reduce the system outage time to the joint outage time of all the stations.The design of such systems depends on the probability that the Earth stations simultaneously exceed their margins.In this paper, a well-assessed model is utilized for the prediction of joint statistics of rain attenuation in multiple locations, using Monte Carlo simulation. The model is based on a pair of multi-variate normal processes whose parameters are related to those characterizing the single-location statistics and whose covariance matrices are assumed to depend only on the distances between locations.The main results concerning both the probability and margin improvement will be presented and discussed.