Abstract
Eight reference stations in south-eastern South America are analysed to derive the auto-covariance and auto-correlations of maximum and minimum temperature. Through the analysis are observed oscillations of periods between 18 and 25 years in the series studied. These periodicities show changes over time, especially in the 1950-1970 period. The exception is Río Gallegos where a periodicity of 18 years is observed throughout the entire record. Persistence is defined as the first correlation coefficient and is a function of the number of negative terms in the auto-covariance. However for some years it is mainly driven by the magnitude of the positive terms given by extreme warm or cold outbreaks. Between 1950 and 1970 it was observed an increased variability in the analysed properties. These changes suggest a variation in the frequency of different circulation patterns having a direct impact on the regional thermal structure. Likewise, this kind of variations can have serious socio-economic impacts as these directly affect the frequency and duration of extreme events. This is relevant in the design of consistency and quality control methods to detect outliers or systematic errors.
Keywords:
climate variability; memory; daily temperature; wavelet analysis
