Seasonal dependence of low-frequency temperature variability in Europe: the wavelet analysis

Temperature variability on timescales 100-102 years is studied by application of Wavelet Transform (WT) to a number of seasonal early-instrumental timeseries in the European domain. The WT indicates that the components of variability on longer than interdecadal timescales are season dependent at each location. In winter and spring temperature variability is dominated by centennial components. The estimated monotonic trends indicate winter warming at all locations; no clear evidence is found of accelerating of rate of warming with the beginning of the industrial era. Summer trends indicate little changes or cooling at most locations. We do not find any periodic region-wide component on multidecadal timescale; an oscillatory signal with a timescale 60-80 years is evident only in summer and autumn temperature series from central Europe; in other seasons or at other locations the multidecadal oscillation appears weak/distorted or absent. The present WT analysis extends earlier results based on multichannel singular spectrum analysis of the same dataset and confirms major role of seasonality in spatio-temporal structure of temperature variability on low-frequency range.