Power Spectrum Analysis of the Time Series of Tabriz Annual Temperature

A spectral analysis is a technique to discover and analyses of obvious and latent different waves in climatic oscillation. The spectrum yields a measure of distribution of variance in a time series over all possible wavelengths. In this paper the periodic cycles of annual air temperature time series of Tabriz city for 1951-2005 (55 years) have been estimated based on harmonic analyses. Accordingly at first the periodogram smoothed, and then confidence interval of smoothed periodogram estimated to find out the significant cycles. The result of this method in Tabriz shows prominent sinusoidal periodicities at 2-2.6 and 4.5-5 years as well as a non – sinusoidal (trend) cycle. The periodic behaviors are proofed in many other studies as an ENSO or solar activities results and In others as a result of zonal winds change, while the 5 year cycles are related to the error of waveforms over relatively short, finite intervals.