Inter‐annual variability, risk and confidence intervals associated with propagation statistics. Part I: theory of estimation

Abstract

SUMMARYThis set of two companion papers aims at providing a statistical framework to quantify the inter‐annual variability observed on the statistics of rain attenuation or rainfall rate derived from Earth‐space propagation measurements. This part I is more specifically devoted to the theoretical study of the variance of estimation of empirical complementary cumulative distribution functions (ECCDFs) derived from Earth‐space rain attenuation or rainfall rate time series. To focus the analysis on the statistical variability but without loss of generality, synthetic rain attenuation time series are considered. A large variability on the ECCDFs, which depends on the duration of the synthetic data, is first put into evidence. The variance of estimation is then derived from the properties of the statistical estimator. The formulation is validated numerically, by comparison with the ECCDF variances derived from the synthetic data. The variance of the fluctuations around the CCDF is then shown to be dependent on the average of the correlation function of the time series, on the probability level and on the measurement duration. This variance of estimation is needed as a prerequisite in conjunction with the knowledge of the climatic variability to characterize the yearly fluctuations of propagation statistics computed from experimental time series. The extensions from simulations to experiments as well as the application to system planning are detailed in part II. Copyright © 2013 John Wiley & Sons, Ltd.