Cycles and trends in the Czech temperature series using wavelet transforms

Temperature variability in the Czech Republic is analysed by means of wavelet transforms. This advanced time-frequency analysis provides information about the nature and time-frequency localization of present oscillations. The data set comprises four mean monthly temperature series from Prague-Klei-nentinum, Brno, Mt Milesovka and the gridded temperature series for the Czech Republic. The results of their wavelet transforms are presented as (a) trend analysis (computed by the inverse wavelet transform) and (b) periodicities and oscillations (distinguishable from the series wavelet power spectra and global wavelet spectra). The results show considerable similarity among individual series. Examples are the parts of the wavelet power spectra in the time period between 1930 and 2001 expressed in all series studied, the pronounced oscillations of about 8 and 12-14 years in all series, and the noticeable increase of temperatures in all cases. As a supplement to this study, wavelet transforms were performed on the variance-adjusted version of combined land and marine temperature anomalies for the Northern Hemisphere from the Hadley Centre. Only some results of this analysis are mentioned and discussed. Copyright (C) 2004 Royal Meteorological Society.