Time-frequency decomposition of sunspot number fluctuations

Time-frequency decomposition has been widely used to analyze those long periodic signals in solar physics. Empirical mode decomposition (EMD) is the popular methods to decompose them from time domain to time-frequency domain. However, the approach brings mode mixing in cycle decomposition and can cause inaccurate results and conclusion in solar physics. In this paper, a new time-frequency decomposition approach based on synchrosqueezing transform (SST) is used. We built a synthetic signal and used an echolocation pulse of a bat to illustrate the performance of the SST approach to the EMD. Finally, the sunspot number fluctuations (SNF) was analyzed with the SST, and several new periods are extracted that fail to found with the EMD approach. The result demonstrates that the SST can accurately decompose those intrinsic cycles of SNF, and better understand the phenomenon of solar activity.