Spectrum: spectral analysis of unevenly spaced paleoclimatic time series

Abstract

A menu-driven PC program (SPECTRUM) is presented that allows the analysis of unevenly spaced time series in the frequency domain. Hence, paleoclimatic data sets, which are usually irregularly spaced in time, can be processed directly. The program is based on the Lomb–Scargle Fourier transform for unevenly spaced data in combination with the Welch-Overlapped-Segment-Averaging procedure. SPECTRUM can perform: (1) harmonic analysis (detection of periodic signal components), (2) spectral analysis of single time series, and (3) cross-spectral analysis (cross-amplitude-, coherency-, and phase-spectrum). Cross-spectral analysis does not require a common time axis of the two processed time series. (4) Analytical results are supplemented by statistical parameters that allow the evaluation of the results. During the analysis, the user is guided by a variety of messages. (5) Results are displayed graphically and can be saved as plain ASCII files. (6) Additional tools for visualizing time series data and sampling intervals, integrating spectra and measuring phase angles facilitate the analysis. Compared to the widely used Blackman–Tukey approach for spectral analysis of paleoclimatic data, the advantage of SPECTRUM is the avoidance of any interpolation of the time series. Generated time series are used to demonstrate that interpolation leads to an underestimation of high-frequency components, independent of the interpolation technique.