Statistical risk associated with tropospheric propagation models and measurements

This paper aims at providing an explanation for part of the inter annual variability observed on the statistics of attenuation computed from earth-space propagation measurements. The high variability observed for the low percentages of time is shown to be related to the variance of the statistical estimator used to compute empirical complementary cumulative distribution function. The knowledge of this variance of estimation may be useful to provide guidance for dimensioning observation durations for propagation experiments and also to quantify the risk in terms of availability associated with the use of a given propagation margin.