Distribution of Fourier power spectrum of climatic background noise
Zhihua Zhang, John Christopher Moore · LaCRIS (University of Lapland) · 2011
In climatic signal processing, significance tests are used to extract statistically significant feature from randomness. For many climatic signals, an appropriate background noise is red noise. In 1998, Torrence and Compo (1) provided an empirical formula on the distribution of Fourier power spectrum of red noise which is the foundation of significance tests on Fourier analysis of climatic signals. In this paper, we will prove this empirical formula in a rigorous statistical framework, and apply it to significance tests of central England temperatures.