An analysis by ARMA stochastic processes of the daily average temperatures collected during 1984 in Udine and Lignano Sabbiadoro

The daily average air temperatures collected during 1984 in Udine and Lignano are studied using stochastic processes. In order to obtain the seasonal component, a computational method has been developed involving the Fourier expansion, while the study of the stationary residual series is carried out by means of ARMA recursive filters. It was found that an autoregressive model of the third order represents the residual data fairly well. This method may also be applied to other meteorological parameters with significant frequencies and with a time behaviour which can be made stationary at least in the wide sense. This work was carried out in the scope of the MICMET program which is aimed at obtaining a detailed analysis of micrometeorological phenomena in the region of Friuli-Venezia Giulia.