Parametric time series analysis of cold and hot spells in daily temperature: an application in southern Italy

Fifty multiyear samples of lowest and highest daily temperatures at stations in southern Italy were studied. The statistical characteristics of hot and cold spells can be reasonably specified in relation to a statistical model that is periodic in its deterministic component and Markovian in its stochastic portion. This significantly decreases the dimensionality needed to determine the time behavior of lowest or highest daily temperature series at any station. Analyses of the daily means of the lowest or highest temperature series revealed the emergence of natural groups, further decreasing this required dimensionality. The results permit the combination of homogeneous stations to enhance the statistics with which sample relative of frequencies of run lengths are derived.