Articles | Volume 13
https://doi.org/10.5194/asr-13-121-2016
https://doi.org/10.5194/asr-13-121-2016
19 Jul 2016
 | 19 Jul 2016

On the temporal variability of the surface solar radiation by means of spectral representations

Marc Bengulescu, Philippe Blanc, and Lucien Wald

Abstract. This work deals with the temporal variability of daily means of the global broadband surface solar irradiance (SSI) impinging on a horizontal plane by studying a decennial time-series of high-quality measurements recorded at a BSRN ground station. Since the data have a non-linear and non-stationary character, two time-frequency-energy representations of signal processing are compared in their ability to resolve the temporal variability of the pyranometric signal. First, the continuous wavelet transform is used to construct the wavelet power spectrum of the data. Second, the adaptive, noise-assisted empirical mode decomposition is employed to extract the intrinsic mode functions of the signal, followed by Hilbert spectral analysis. In both spectral representations, the temporal variability of the SSI is portrayed having clearly distinguishable features: a plateau between scales of two days and two-three months that has decreasing power with increasing scale, a large spectral peak corresponding to the annual variability cycle, and a low power regime in between the previous two. It is shown that the data-driven, noise-assisted method yields a somewhat more sparse representation and that it is a suitable tool for inspecting the temporal variability of SSI measurements.

Download
Short summary
The continuous wavelet (CWT) and the Hilbert–Huang transforms (HHT) are compared for the analysis of the temporal variability on ten years of daily means of the surface solar irradiance. In both cases, the variability exhibits a plateau between scales of two days and three months that has decreasing power with increasing scale, a spectral peak corresponding to the annual cycle, and a low power regime in-between. The HHT is shown to be suitable for inspecting the variability of the measurements.