Impacts des changements climatiques sur les modèles de conception des liaisons hertziennes

Abstract

The values of most indices used in the design of radio links depend on the variation of the meteorological parameters in the troposphere. The aim of the present work is to analyze the variations of these indices in Quebec. For this, it is necessary to study, the short-term and long-term variations of the refractivity, N and its vertical gradient, dN1, as well as of the geoclimatic factor, K. The methodologies used in this work include : 1) The method proposed by the international Telecommunication Union (ITU) for estimating the values of the indices. 2) The clustering analysis for obtaining 5 areas from North to South with different values of N, and 3) The direct smoothing forecasting technique for forecasting the values of N.Based on the experimental data of several terrestrial links, we found that there is a seasonal variation of K which is mainly due to the variation of the humidity. The value of K also varies from year to year, indicating the need to update ITU recommendations that currently do not consider climate change for estimating the value of K.The clustering analysis has been applied on the meteorological data collected from 50 stations to obtain 5 areas with different variations of N from North to South. The best input data for the identification is the water vapour pressure. The results show that the value of N increase from North to South. Data collected during one year in the North has been used to estimate N, dN1 et K. It is observed that the more negative values of dN1 lie in the summer season. This is mainly due to the important variations of the temperature and humidity during this season.The study of long-term variation of N and dN1 is performed using data collected for a period of 35 years. The variations of N and dN1 show a progressive increase over the period of study, especially in summer. Such a trend may increase the occurrence of super-refraction. Data collected during 39 years at one station located in the North of the Quebec and another located in the South is used to estimate the values of N. The values of the dry component of N are higher in the North, whereas that the values of the wet component are higher in the South. The highest slope for the variation of N in observed in the northern part. We apply direct smoothing forecasting technique on the data collected in the North to estimate future values of N. The performance of this technique is analysed.

Date	3 Feb 2026
Original language	French
Awarding Institution	École de technologie supérieure
Supervisor	Ammar B. Kouki (Supervisor) & Basile L. Agba (Co-supervisor)