APPLICATIONS OF THE CYCLIC SPECTRAL-ANALYSIS TO THE SURFACE-TEMPERATURE FLUCTUATIONS IN A STOCHASTIC CLIMATE MODEL AND A GCM SIMULATION

In this paper, we introduce the cyclostationary processes into climate analysis and undertake a systematic study of the cyclic spectra of su rface temperature fluctuations. The technique is adapted from cyclosta tionarity theory in signal processing. To demonstrate the usefulness o f this technique, a very simple cyclostationary stochastic climate mod el is constructed Our results show that the seasonal cycle strongly mo dulates the amplitudes of the covariance and the spectrum The techniqu e was also applied to the surface temperature fluctuations in a fiftee n-year seasonal run of the National Center for Atmospheric Research (N CAR) Community Climate Model (CCM2, R15) using a zonally symmetric all -land surface as the lower boundary. The results indicate that intrase asonal oscillations localized according to time of year are still pres ent even after the surface temperature fields have been normalized usi ng the commonly used procedure. Both examples suggest that the ''annua l cycle'' cannot be ''removed'' by simply using a normalization proced ure. The climate is not as completely represented when modelled as sta tionary processes.