Title :
MP based detection of synchronized motions across the instrumental climate record
Author_Institution :
Appl. Marine & Limnic Studies (IAMARIS), Hamburg-Berlin-Muncheberg Climate Dynamics & Signal Anal. Project, Berlin, Germany
Abstract :
Univariate, adaptive atomic decompositions of representative time series of the global climate system are obtained using a Matching Pursuit (MP) algorithm with a large, overcomplete dictionary of Gaussian logons, i.e. frequency-modulated (FM) extensions to the Gabor atom. Among the types of synchronized motion across the instrumental climate record (1870-1997) that have been identified in a subsequent multivariate synthesis attempt, phase-frequency ("interdomain") synchrony dominates. It exhibits frequency drift of one oscillator in parallel to the phase evolution of another one - a dynamic information that may hardly be gained from methods of time series analysis as predominantly used in climate research. Deep FM, which shows ubiquitous presence, appears to be both precondition and result of this type organized dynamics under participation of multidecadal to centennial (as far as timescales may be inferred from a 128 year data set) climatic modes.
Keywords :
atmospheric movements; climatology; iterative methods; synchronisation; time series; time-frequency analysis; AD 1870 to 1997; Gaussian logon distribution; MP based detection method; atmospheric circulation; frequency-modulated extension analysis; global climate system; instrumental climate record; matching pursuit algorithm; phase evolution; phase-frequency synchronization; synchronized motion analysis; time series analysis method; time-frequency representation; univariate adaptive atomic decomposition method; Matching Pursuit; frequency modulation; global climate; sparse approximation; synchronized motion; time-frequency representation;
Conference_Titel :
Statistical Signal Processing Workshop (SSP), 2011 IEEE
Conference_Location :
Nice
Print_ISBN :
978-1-4577-0569-4
DOI :
10.1109/SSP.2011.5967758