EVALUATION OF STATISTICAL MODELS FOR CLEAR-AIR SCINTILLATION PREDICTION USING OLYMPUS SATELLITE MEASUREMENTS

Abstract

The objective of this study is to evaluate and to compare some of the statistical models for the monthly prediction of clear-air scintillation variance and amplitude from ground meteorological measurements. Two new statistical methods, namely the direct and the modelled physical-statistical prediction models, are also introduced and discussed. They are both based on simulated data of received scintillation power derived from a large historical radiosounding set, acquired in a mid-latitudue site. The long-term predictions derived from each model are compared with measurements from the Olympus satellite beacons at the Louvain-la-Neuve site at 12·5 and 29·7 GHz and at the Milan site at 19·77 GHz during 1992. The model intercomparison is carried out by checking the assumed best-fitting probability density function for the variance and log-amplitude fluctuations and analysing the proposed relationships between scintillation parameters and ground meteorological measurements. Results are discussed in order to understand the potentials and the limits of each prediction model within this case study. The agreement with Olympus measurements is found to be mainly dependent on the proper parametrization of prediction models to the radiometeorological variables along the earth–satellite path. ©1997 by John Wiley & Sons, Ltd.